Private Placement

166 The Economics of the Private Placement Market

Mark Carey, Stephen Prowse, John Rea, and Gregory UdellStaff, Board of Governors

The staff members of the Board of Governors ofthe Federal Reserve System and of the FederalReserve Banks undertake studies that cover awide range of economic and financial subjects.From time to time the studies that are of generalinterest are published in the Staff Studies seriesand summarized in the Federal Reserve Bulletin.

The following paper, which is summarized inthe Bulletin for January 1994, was prepared inthe spring of 1993. The analyses and conclusionsset forth are those of the authors and do notnecessarily indicate concurrence by the Board ofGovernors, the Federal Reserve Banks, ormembers of their staffs.

Board of Governors of the Federal Reserve SystemWashington, DC 20551

December 1993

Preface

This study would not have been possible withoutthe generous assistance of an enormous number oforganizations and individuals. The opinionsexpressed in the study are not necessarily those ofthe Board of Governors, of other members of itsstaff, nor of those who assisted us.

Without implying that they agree with or areresponsible for the contents of the study, we aregrateful to the following organizations for theirparticipation in formal interviews: Aetna, Bank ofAmerica, Bankers Trust, Chemical Bank, Citibank,Continental Bank, First National Bank of Chicago,The First Boston Corporation, Fitch InvestorsService, Goldman Sachs and Company,J.P. Morgan, John Hancock Mutual Life, LehmanBrothers, Massachusetts Mutual Life, MerrillLynch, Mesirow Financial, Metropolitan Life,New York Life, Oppenheimer & Co., The Pruden-tial, Standard & Poor’s, State of WisconsinInvestment Board, Teachers Insurance AnnuityAssociation, The Travelers Company, andWilliam Blair Mezzanine Capital Partners.

We are also grateful to the following individualsfor the assistance they provided:Loren S. Archibald, Adrian Banky, Steven M.Bavaria, Donald A. Bendernagel, Allen N. Berger,Salvatore J. Bommarito, Sanford Bragg, Philip E.Brown, John E. Cartland, III, Robert E.Chappell, Jr., Gerald Clark, Robert Clement,Richard J. Cobb, Jr., Tim Conway, Patrick J.Corcoran, Sally Corning, Paul R. Crotty, Dean C.Crowe, Andrew Danzig, Charles E. Engros, EileenFox, Steven Galante, Andrew B. Hanson, ThomasB. Harker, William Hogue, Gwendolyn S. IIoani,Gregory Johnson, David S. Jones, John Joyce,Gilbey Kamens, Thomas Keavency, Michael B.

Kessler, Robert B. Lindstrom, Charles M. Lucas,Kathryn C. Maney, Terrence P. Martin,Dennis P. McCrary, Thomas Messmore,Michael G. Meyers, Stephen T. Monahan, Jr.,Anthony Napolitano, Roger Nastou, Scott J.Nelson, R. Gregory Nelson, Kevin Newman,Ramon de Oliveira-Cezar, Barbara Paige,Patrick M. Parkinson, Dale R. Paulshock,Paul Reardon, Steven H. Reiff, David L. Roberts,Clayton S. Rose, Barry M. Sabloff, Martha S.Scanlon, Karl A. Scheld, John E. Schumacher,Terrance M. Shipp, Eric A. Simonson,Bram Smith, Dewain A. Sparrgrove, James W.Stevens, Maleyne M. Syracuse, Drew M. Thomas,Thomas D. Thomson, Richard TB. Trask,William F. Treacy, Bruce Tuckman, Tracy Turner,James C. Tyree, Marc J. Walfish, Russell S. Ward,Richard S. Wilson, Richard A. Yorks, andRobert L. Zobel.

We are especially grateful to those who offeredspecial assistance or who were helpful to an extentfar beyond the call of duty or public spiritedness:to S. Ellen Dykes for editorial and productionassistance; to Edward C. Ettin and Myron L.Kwast, who inspired and supported the study; toLloyd Campbell, Nathaniel S. Coolidge,Leland Crabbe, George Fenn, Mitchell A. Post,and Stephen A. Sharpe for their many helpfulcomments; to David Blood, Simon Jawitz, andRobert E. Joyal for special assistance; and toJalal Akhavein, Curtis Atkisson, Dana Cogswell,William Gerhardt, and John Leusner for expertresearch assistance.

Finally, we are grateful for the support of theBoard of Governors, without which this studywould not have been possible.

Contents

Introduction ...................................................................................................................... vii

Part 1: An Economic Analysis of the Traditional Market for Privately Placed Debt .............. 1

1. Overview of the Traditional Private Placement Market ...................................... 1

Principal Themes of Part 1 and Key Definitions .................................................... 2Organization of Part 1 and Summary of Findings .................................................. 4

2. Terms of Privately Placed Debt Contracts ......................................................... 5

Issue Size ........................................................................................................ 5Maturity and Prepayment Penalties ...................................................................... 7Types of Payment Stream and Yields .................................................................. 10Variety of Securities .......................................................................................... 11Covenants ....................................................................................................... 11Covenants and Renegotiation .............................................................................. 13Collateral ........................................................................................................ 14Summary ......................................................................................................... 15

3. Borrowers in the Private Placement Market ...................................................... 15

Issuers in the Private Placement Market ............................................................... 17Differences among Firms Issuing in the Public, Private, and Bank Loan Markets ........ 21Companies Issuing in Both the Public and the Private Markets ................................ 25Summary ......................................................................................................... 26

4. Lenders in the Private Placement Market ......................................................... 27

Life Insurance Companies .................................................................................. 28Finance Companies ........................................................................................... 30Pension Funds .................................................................................................. 30Banks ............................................................................................................. 31Other Investors ................................................................................................. 31Summary ......................................................................................................... 31

5. Private Placements, the Theory of Financial Intermediation,and the Structure of Capital Markets ............................................................... 32

Asymmetric Information, Contracting, and the Theory of Covenants ......................... 34Information-based Theories of Financial Intermediation .......................................... 36The Covenant–Monitoring–Renegotiation Paradigm ............................................... 36Private Placements in a Theory of Credit Market Specialization ............................... 38Other Empirical Evidence Revelant to the Theory of Credit Market Specialization ...... 39Summary of Part 1 ........................................................................................... 41

Part 2: Secondary Trading, the New Market for Rule 144A Private Placements,and the Role of Agents ......................................................................................... 43

1. The Rule 144A Market .................................................................................... 43

Features of the Market ...................................................................................... 44Prospects for Development ................................................................................. 47

2. The Role of Agents .......................................................................................... 47

Who Are the Agents? ........................................................................................ 48The Stages of a Private Placement Transaction ...................................................... 48

Information Flows ............................................................................................ 57Price Determination .......................................................................................... 59Agents’ Fees and Other Costs of Issuance ............................................................ 59Private Market Efficiency ................................................................................... 60Private Placements without an Agent ................................................................... 60Agent Operations under Rule 144A ..................................................................... 61Summary ......................................................................................................... 61

Part 3: Special Topics ...................................................................................................... 63

1. The Recent Credit Crunch in the Private Placement Market .............................. 63

Definition of Credit Crunch ................................................................................ 63Evidence That a Credit Crunch Occurred ............................................................. 64Sources of the Credit Crunch .............................................................................. 66Prospects for an Easing of the Crunch ................................................................. 68Effects on and Alternatives of Borrowers ............................................................. 69Conclusion ...................................................................................................... 70

2. The Current and Prospective Roles of Commercial Banks .................................. 70

Banks as Agents and Brokers ............................................................................. 70Why Do Banks Act as Agents, or Why Is the List of Bank Agents So Short? ............ 71Prospective Changes in Market Share of U.S. and Foreign Banks ............................. 72The Role of Regulation ..................................................................................... 72Banks as Issuers of Equity ................................................................................. 74Banks as Issuers of Debt .................................................................................... 75Banks as Buyers of Private Placements ................................................................ 75Banks as Competitors of Private Placement Lenders .............................................. 76

Appendixes

A. Definition of Private Placement, Resales of Private Placements,and Additional Information about Rule 144A .................................................... 77

B. The Market for Privately Placed Equity Securities ............................................ 81

C. Legal and Regulatory Restrictions on Bank Participationin the Private Placement Market ...................................................................... 84

D. Historical Data on Issuance of Private Debt ...................................................... 89

E. A Review of the Empirical Evidence on Covenants and Renegotiation ................. 91

F. An Example of a Private Placement Assisted by an Agent .................................. 95

G. Estimates of Issue-Size and Maturity Distributions ............................................ 98

H. Borrower Substitution between the Public and Private Markets .......................... 101

References ........................................................................................................................ 105

Introduction

The private placement market is an importantsource of long-term funds for U.S. corporations.Between 1987 and 1992, for example, the grossvolume of bonds issued in the private placementmarket by nonfinancial corporations was morethan 60 percent of that issued in the publiccorporate bond market. Furthermore, at the end of1992, outstanding privately placed debt of nonfi-nancial corporations was more than half as largeas outstanding bank loans to such corporations.

Despite its significance, the private placementmarket has received relatively little attention in thefinancial press or the academic literature. This lackof attention is due partly to the nature of theinstrument itself. A private placement is a debt orequity security issued in the United States that isexempt from registration with the Securities andExchange Commission (SEC) by virtue of beingissued in transactions ‘‘not involving any publicoffering.’’1 Thus, information about privatetransactions is often limited, and following andanalyzing developments in the market are difficult.The last major study of the private placementmarket was published in 1972, and only a fewarticles have appeared in economics and financejournals since then.2

This study examines the economic foundationsof the market for privately placed debt, analyzesits role in corporate finance, and determines itsrelation to other corporate debt markets. Themarket for privately placed equity is brieflydescribed in appendix B. In the remainder of thestudy, the term private placement refers only toprivately placed debt.

There seem to be two widespread mispercep-tions about the nature of the private placementmarket. One is the belief that it is mainly asubstitute for the public bond market: that is,issuers use it mainly to avoid fixed costs associ-ated with SEC registration, and lenders closelyresemble buyers of publicly issued bonds. Thismisperception may have arisen because privateplacements are securities and because the defini-tion of a private placement focuses on its exemp-

tion from registration. Regulatory considerationsand lower transaction costs do cause some issuersto use the private market. Principally, however, itis an information-intensive market, meaning thatlenders must on their own obtain informationabout borrowers through due diligence and loanmonitoring. Many borrowers are smaller, less-well-known companies or have complex financings, andthus they can be served only by lenders thatperform extensive credit analyses. Such borrowerseffectively have no access to the public bondmarket, which provides funding primarily to large,well-known firms posing credit risks that can beevaluated and monitored with publicly availableinformation. In this respect, private marketlenders, which are mainly life insurance compa-nies, resemble banks more than they resemblebuyers of publicly issued corporate debt. Even ifregistration of public securities were not required,something resembling the private placementmarket would continue to exist.

The second misperception is that the privateplacement market is identical to the bank loanmarket in its economic fundamentals. Thismisperception may have been fostered by thetendency of some recent studies of information-intensive lending to group all business loans notextended through public security markets underthe rubric ‘‘private debt.’’ Included in this categoryare bank loans, private placements, financecompany loans, mezzanine finance, venturecapital, and other kinds of nonpublic debt. Aprincipal finding of this study, however, is that allinformation-intensive lending is not the same. Inparticular, the severity of the information problemthat a borrower poses for lenders is an importantdeterminant of the markets in which the companyborrows and of the terms under which credit isavailable.

Besides dispelling these misperceptions, thestudy describes in detail the nature and operationof the private placement market. It also offersempirical support for the proposition that theprivate placement market is information intensiveand that private market lenders and borrowers aredifferent from lenders and borrowers in othermarkets. It provides a theoretical explanation forthe existing structure of business debt markets thatbuilds upon recent theories of financial intermedia-tion, covenants, debt contract renegotiation, anddebt maturity. Finally, it analyzes some recent

1. See appendix A for a more detailed definition of ‘‘privateplacement.’’ Some securities issued in other countries are alsoreferred to as ‘‘private placements.’’ This study focuses only onsecurities issued in the United States.

2. Shapiro and Wolf (1972).

developments in the private placement market,including a credit crunch, the effect of the SEC’sRule 144A, and changes in the roles that banksplay.

Organization of the Study

The information-intensive nature of the privateplacement market is the theme of part 1 of thestudy. This part compares the terms of privateplacements with those of public bonds and bankloans and considers borrowers’ characteristics andtheir motivations for using the private market, aswell as the operations of lenders. An explanationgrounded in theories of financial intermediationand financial contracting is given for the structureof the market and for the differences between theprivate market and other markets for capital.

Part 2 focuses on the process of private issuanceand completes our basic analysis of the privateplacement market by considering the role ofagents and the effect of Rule 144A. Agents areinvolved in most private placements: They advisethe issuer and assist in distributing securities. Inthe process, they gather and disseminate informa-tion, an important task for a market in whichinformation is scarce.

In 1990, the SEC adopted Rule 144A to reviseand clarify the circumstances under which aprivately placed security could be resold. Privateplacements are often described as illiquid securi-ties, but this perception is not entirely accurate.A relatively small secondary market for privateplacements has existed for years, although thelegal basis for secondary trading was somewhatuncertain. 3

Rule 144A has led to the development of amarket segment for private placements that are notinformation intensive. This new segment is thusfundamentally different from the traditional privatemarket and has many characteristics of the publicbond market. Its primary attraction for borrowershas been the availability of funds at interest ratesonly slightly higher than those in the public

market without the burden of registration require-ments. Though still developing, the new markethas attracted a significant volume of issuance andthus could be a major step toward the integrationof U.S. and foreign bond markets.

Part 3 analyzes two special topics. One is therecent credit crunch in the below-investment-gradesegment of the private debt market. Life insurancecompanies had been the primary buyers oflow-rated private placements, but most havestopped buying such issues. Many medium-sizedborrowers have been left with few alternatives forlong-term debt financing. Our explanation for thecrunch, which emphasizes a confluence of marketand regulatory events, highlights the fragility ofinformation-intensive markets.

The other special topic is the role of commer-cial banks in the private market, both as agentsand as providers of loans that compete somewhatwith private placements. The prospect for asubstantial increase in competition between thebank loan and private placement markets isconsidered, as is the prospect for a substantialchange in banks’ roles as agents.

Sources of Information

Any analysis of the private placement market ishandicapped by a lack of readily available infor-mation. Because the securities are not registeredwith the SEC, only limited data about transactionsare publicly available, and most participantsdisclose relatively little about their operations.Also, relatively little has been written about themarket.

In conducting this study, we have relied onpublic sources to the extent possible, but we havealso held extensive interviews with marketparticipants. Our interviewees are active partici-pants in the market and include staff members oflife insurance companies, pension funds, invest-ment banks, commercial banks, and rating agen-cies. The information obtained from these inter-views is an important part of our analysis,although our conclusions are based, not on anysingle test or source of information but rather onthe weight of the evidence from extant studies,from new empirical results and theoretical argu-ments presented here, and from the remarks ofmarket participants.

3. For practical purposes, private placements may be legallytraded among institutional investors with a reliance on thesame assurances and exemptions that are employed in thenew-issue market or on Rule 144A, which provides a non-exclusive safe harbor for certain secondary market transactionsin private placements among certain institutional investors.Trading that relies on the traditional assurances and exemptionsis relatively infrequent because the process is cumbersome andbecause secondary-market buyers, unless they are alreadymembers of a syndicate, must often conduct due diligence justas in the new-issue market.

Part 1: An Economic Analysis of the Traditional Market for Privately Placed Debt

1. Overview of the Traditional PrivatePlacement Market

Part 1 of this study describes and analyzes what isnow called the traditional market for privatelyplaced debt. Until the development of the Rule144A market in 1990, it was the entire market forprivate debt. It continues to be the larger of thetwo markets. Unless otherwise noted, in part 1 theterms private placement and private debt referonly to debt securities issued in the traditionalmarket, and the term private market refers only tothe traditional market for privately placed debt. 4

Taken as a whole, the traditional and the 144Aprivate placement markets are a significant sourceof funds for U.S. corporations. Their importancecan be seen by comparing gross offerings bynonfinancial corporations of private and publicbonds (chart 1). Between 1986 and 1992, forexample, gross annual issuance of private place-ments by such corporations averaged $61 billionper year, or more than 60 percent of average

issuance in the public market (table 1). 5 In 1988and 1989, private issuance actually exceededpublic issuance, as the financing of acquisitionsand employee stock ownership plans boostedprivate offerings. However, public issuance surgedin 1991–92, partly because of the refinancing ofoutstanding debt, and private issuance fell. Thepunitive prepayment penalties normally attached toprivately placed debt make refinancing unattractiveto issuers even when interest rates are falling.

A similar comparison of private placementswith bank loans, another major source of corporatefinancing, is difficult because of a lack of data onthe gross volume of new bank loans and becauseof differences in maturity. Comparing outstandingbank loans with estimates of outstanding privateplacements is possible, however. At the end of1992, bank loans to U.S. nonfinancial corporationswere $519 billion, whereas outstanding privateplacements of nonfinancial corporations wereapproximately $300 billion, or somewhat morethan half of bank loans. At the same time, out-standings of public bonds issued by nonfinancial

4. Some recent academic studies have used the term privatedebt to refer to any debt not issued in the public bond market(and similar public markets)—for example, bank loans—andthe term private market to refer to all nonpublic debt markets.In this study, the terms refer only to private placements andtheir market.

5. Data for gross issuance of private placements are fromIDD Information Services and Securities Data Corporation,which obtain the data from a survey of investment banks andcommercial banks serving as agents in placing the securities.Data for private placements that do not involve an agent arenot included. Consequently, reported totals probably understategross issuance of private placements.

Cohan (1967) presents evidence that a shift from the publicto the private market occurred during the 1930s. He found thatprivate placements represented about 3 percent of debt issuancebetween 1900 and 1934 but averaged about 46 percent from1934 to 1965. As noted by Smith and Warner (1979), therelative growth in private issuance partly reflects passage of theSecurities Act of 1933, the Securities Exchange Act of 1934,and the Trust Indenture Act of 1939, all of which raised thecost of public debt relative to private placements.

1. Gross issuance of publicly offered andprivately placed bonds by nonfinancialcorporations, 1975–92

Billions of dollars

150

125

100

75

50

25

1975 1980 1985 1990

Public bonds

Private placements

Sources: Federal Reserve Board and IDD Information Services.

1. Average gross issuance of publicly offeredand privately placed bonds by nonfinancialcorporations, 1975–92Billions of dollars, annual rate

Type of issuance 1975–80 1981–85 1986–92

Public offerings . . . . . . . . . 21.0 35.6 97.0Private placements . . . . . . 14.7 19.8 60.5

Source. Federal Reserve Board and IDD InformationServices.

corporations stood at $775 billion.6 In short, theprivate placement market has provided a substan-tial fraction of corporate finance in the UnitedStates.

Most private placements are fixed-rate,intermediate- to long-term securities and are issuedin amounts between $10 million and $100 million.Borrowers vary greatly in their characteristics, butmost are corporations falling into one of threegroups: mid-sized firms wishing to borrow for along term and at a fixed rate, large corporationswishing to issue securities with complex ornonstandard features, and firms wishing to issuequickly or with minimal disclosure. Investors arealmost always financial institutions. Life insurancecompanies buy the great majority of privateplacements of debt.

Principal Themes of Part 1and Key Definitions

As noted in the introduction, previous studies havetended to characterize private placements as closesubstitutes for either publicly issued corporatebonds or for bank loans. Besides providing adetailed description of the market, part 1 developsthe theme that neither of these views is correct.Private placements have some of the characteris-tics of bank loans and public bonds, as well assome unique characteristics.

Studies characterizing private placements assimilar to public bonds note that both are securi-ties and both tend to have long maturities andfixed rates. Such studies focus on regulatory andissuance costs as the factors that motivate borrow-ers to issue privately rather than publicly. In theseexplanations, some issuers choose the privatemarket to avoid delays and disclosure associatedwith SEC regulations. Other, relatively smallissues are said to be done in the private marketbecause fixed costs of issuance are smaller there,offsetting interest rates that are somewhat higherthan in the public market. Large issues are said tobe sold in the public market because fixed costsare spread over a larger base, making lower ratesthe dominant consideration for issuers.

Although regulatory and issuance costs canaffect a borrower’s choice of market, othereconomic forces are of greater importance. Thetraditional private placement market is fundamen-tally an information-intensive market. Privatemarket borrowers or their issues are informationproblematic, and so a key activity of privatemarket lenders is the gathering or production ofinformation about borrower credit quality. Theitalicized terms are drawn from theories thatemphasize the asymmetry of information that oftenexists between borrowers and lenders. Manyborrowers have better information about theirprospects than lenders, and they can often takeactions once a loan is made to reduce the likeli-hood of its repayment. To determine the interestrate at which to lend to such borrowers, lendersmust engage in due diligence during origination;and to control moral hazard risk once a loan ismade, they must engage in loan monitoring. 7

Lenders in information-intensive markets aregenerally financial intermediaries. Because duediligence and loan monitoring involve fixed costs,it is economically efficient that only one or a fewlenders lend to an information-problematicborrower, rather than the large number of smalllenders of a prototypical theoretical securitiesmarket. In theoretical models of information-intensive lending, atomistic lenders (small savers)lend to an intermediary, and the intermediary inturn lends to the ultimate borrowers and isresponsible for due diligence and monitoring.Real-world information-intensive intermediariesdiffer from other intermediaries, such as moneymarket mutual funds, in that they have developedthe capabilities required for lender due diligenceand monitoring.8

6. Outstandings of public bonds of nonfinancial corporationsare the sum of bonds rated by Moody’s Investors Service andpublicly issued medium-term notes. Private placements areestimated by subtracting the figure for public bonds fromoutstandings of all corporate bonds reported in the flow offunds accounts. Data for bank loans are from the flow of fundsaccounts.

7. In some contexts, due diligence refers specifically toactivities directed toward compliance with SEC regulations. Inthis study, the term refers to all credit analysis performed bylenders before and during origination or issuance. Moralhazard risk refers not so much to the risk of fraud or unethicalactions as to the risk that a firm’s shareholders or managerswill take actions that increase the risks borne by bondholders.

8. A few words of explanation of this terminology may behelpful. In common parlance and in the traditional academicliterature, financial intermediary refers to an institution thatgathers funds from many (often small) savers and then lends ata profit. Intermediary also sometimes refers to an institution ora person that brings together lenders and borrowers in directmarkets, for example an underwriter in the public bond market.In some recent academic literature, however, intermediary hascome to mean an institution that lends to information-problematic borrowers. We use the terms information-producing lenders or information-intensive lending instead ofintermediary and intermediation because such a lender need

2

Firms that issue bonds publicly are generallynot information problematic. Public marketinvestors rely mainly on reports by rating agenciesand other publicly available information forevaluations of credit risk at the time of issuanceand for monitoring.

Information problems are conceptually separatefrom observable credit risk. For example, asubordinated loan to a large, highly leveragedmanufacturer of auto parts may be quite risky, butlenders’ evaluation and monitoring of the risk maybe a relatively straightforward exercise involvingpublicly available information (financial state-ments, bond ratings, and some knowledge of theauto industry). In contrast, a loan to a smallmanufacturer of specialized composite materialsmay have low risk but require extensive duediligence by lenders to evaluate and price the riskand considerable monitoring to keep the risk undercontrol. The loan may be low risk because thefirm has recently received a large, stable defensesubcontract and requires additional financing onlyto support a highly profitable increase in produc-tion. These facts, however, are unlikely to bewidely known and must be discovered and verifiedby lenders.

Although information problems and observablecredit risk are conceptually separate, they arecorrelated with one another and with firm size. Forexample, small firms tend both to be riskier and topose more information problems for lenders.Market participants sometimes use a firm’s size asan index of its access to different credit markets:A large firm has access to all markets, a medium-sized firm has access to the private placementmarket but not to the public market, and a smallfirm lacks access to either market. Firm size isoften a good indicator because of its correlationwith information problems, but the extent of theinformation problems that a firm poses for lenders

not be an intermediary in the traditional sense (someinformation-producing lenders are wealthy individuals) and alsobecause many intermediaries, such as money market mutualfunds, do little credit analysis.

Recent theoretical literature has also not always clearlydistinguished different types and circumstances of creditanalysis. The terms credit evaluation and monitoring oftenrefer to analyses done both before and after a debt contract issigned. We refer to that done before as due diligence and tothat done after as loan monitoring. A distinction between thetwo is important to our analysis.

usually is the primary determinant of the marketsin which the firm may borrow. In many instances,for example, large firms with outstanding publicdebt have borrowed in the private placementmarket when their transactions involved complexi-ties that public market investors were not preparedto evaluate.

To be information problematic, a loan mustimpose more costs on lenders during the initialdue diligence stage or the loan monitoring stage,but not necessarily at both stages. For example,the cost of due diligence for a public issue by alarge, complex corporation may be greater thanthat for a private placement by a medium-sizedfirm. However, the private placement mightstill be information problematic because itincluded many more covenants than the publicissue and required more monitoring by lendersthan public investors are prepared to undertake.Similarly, a private placement by a large, well-known firm that included few covenants andrequired little monitoring might still be informa-tion problematic if it were a very complex ornovel issue. In such a case, public lenders wouldbe unprepared to perform the necessary duediligence; only information-intensive lenderswould be prepared to do so.

The traditional private placement market thushas much in common with the bank loan market,even though it is a market for securities. Bankborrowers are often small or medium-sizedfirms for which publicly available information islimited. The prospect for loan repayment isdiscovered by loan underwriting procedures thatare broadly similar to due diligence proceduresin the private placement market, and bank bor-rowers are typically monitored after loans havebeen made.

Because of these similarities, some studies havegrouped bank loans, private placements, and otherinformation-intensive loans under the heading ofprivate debt, in some cases implying that allvarieties of such debt are fundamentally thesame. However, all information-intensive lendingis not the same. Most important, borrowers in thebank loan market are, on average, substantiallymore information problematic than borrowersin the private placement market. Also, privateplacements have mainly long terms and fixedrates whereas bank loans have mainly shortterms and floating rates; other differences aswell exist among the various nonpublic marketsfor debt.

3

Organization of Part 1and Summary of Findings

The remainder of part 1 describes and analyzes thetraditional market for privately placed debt andexplains differences between the private, public,and bank loan markets. Section 2 describes theterms of privately placed debt contracts andcompares them with terms of bank loans andpublicly issued bonds, including issue size,maturity, rates, covenants, and other terms. As inthe information-intensive bank loan market (and incontrast to the public bond market), borrowers andlenders typically negotiate the terms of privateplacements, especially any covenants that restrictthe actions of the borrower. Covenants are animportant part of the technology of loan monitor-ing. Both bank loans and private placements ofteninclude financial covenants, such as minimuminterest-coverage ratios, that can trigger renegotia-tion of the loan terms if the borrower’s character-istics change.9 Such covenants are very rare inpublicly issued securities.

Private market borrowers, described in sec-tion 3, issue long-term, fixed-rate debt privatelyfor several reasons. Many are information prob-lematic, and their issues would not be readilyaccepted in the public bond market. These borrow-ers are, on average, smaller than issuers in thepublic market and larger than those that borrowonly from banks. Borrowers that are not informa-tion problematic generally find total costs to belower in the public bond market, unless thesecurities they issue have novel or complexfeatures requiring extensive due diligence bylenders. Many new types of security have beenintroduced in the private market, but after theirfeatures are widely understood have come to beissued mainly in the public market. Some borrow-ers also use the private market to issue quickly orto avoid disclosures associated with SEC regula-tions. Finally, some nonproblematic borrowerswith small-sized issues use the private marketbecause fixed costs of issuance are lower.

The operations of lenders in the private market,described in section 4, are typical of information-intensive lenders. Life insurance companies, theprincipal lenders, evaluate and monitor theplacements they buy in a manner that is generallysimilar to that of commercial banks’ loan under-

writing and monitoring operations. They usuallyhave loan officers, loan committees, and creditanalysts. Some even have specialized workoutgroups.10 These characteristics differ from thoseof typical public bond buyers. 11 Although somebuyers of publicly issued debt perform some duediligence and monitoring, their efforts are muchless extensive than those in information-intensivemarkets. The activities of public market borrowersare often followed rather closely by credit ratingagencies and investment banks.

Most private market lenders attempt to buildand maintain reputations for reasonableness inrenegotiations of debt contracts. Covenants ininformation-intensive debt contracts are frequentlyviolated, triggering renegotiations. In somerenegotiations, a lender is in a position to extractconsiderable rents from a borrower. Borrowersthus prefer to contract initially only with lendersthat have a reputation for fair dealing. Thispreference is especially strong in the privateplacement market, where loans typically are forlong terms and for substantial amounts, and carrypunitive prepayment penalties. Life insurancecompanies may be especially adept at building andmaintaining such reputations because doing so isespecially important in some of their other lines ofbusiness.

Asset–liability management considerations makeprivate placements particularly attractive tointermediaries with long-term, fixed-rate liabilities,such as life insurance companies. By the sametoken, these features of private placements areunattractive to banks, which must bear the costs ofswapping fixed-rate payment streams to matchtheir floating-rate liabilities. Conversely, banks aremore likely than insurance companies to findshort-term, floating-rate loans to be profitable.Such economies of scope are probably the mainreason for the observed division of lendingbetween banks and insurance companies. Thereasons for the limited participations of otherkinds of intermediaries, such as finance compa-

9. Covenants are usually designed to trigger renegotiationwhen a borrower’s credit quality deteriorates, but most cove-nant violations occur for other reasons, such as borrowergrowth.

10. In contrast to banks, an insurance company’s relation-ship with a private placement issuer is usually one-dimensional: the life insurance company typically providesonly the loan, not other services such as transaction accounts orinsurance policies.

11. Insurance companies buy many assets other than privateplacements, of course, including publicly issued corporate debt.When we refer to public bond buyers we mean the groupswithin a financial intermediary responsible for purchasingpublic bonds, and private lenders are the groups responsible forpurchasing and monitoring private placements. The operationsof these groups tend to be different, with only the privateplacement groups performing substantial amounts of duediligence and loan monitoring.

4

nies, mutual funds, and pension funds, are dis-cussed in section 4.

In the final section of part 1, facts and ideasfrom earlier sections are combined with financialtheory to produce a descriptive theory explainingaspects of the current structure of the bank loan,private placement, and public bond markets. Thetheory emphasizes that information-problematicborrowers choose information-intensive marketsbecause they can, on the whole, obtain betterterms there. Flexible renegotiation of contracts inthe event of covenant violations is an importantpart of the mechanism supporting better terms forborrowers, and the mechanics of covenants andrenegotiation influence the identity and operationsof lenders. The theory offers several reasons thatinformation-intensive lenders are usually financialintermediaries. It reveals links between the extentto which borrowers are information problematicand the maturity of the loans they will tend toobtain. These links imply that lenders’ decisions toserve particular classes of borrowers and to investin particular varieties of due diligence and moni-toring capacity will be influenced by the nature oftheir liabilities.

The theory also helps explain why information-intensive lending seldom occurs in the public bondmarket. In principle, the public bond market mightwell have developed the capacity to lend toinformation-problematic borrowers. However, threefeatures of private placements make them a bettervehicle than public bonds for lending toinformation-problematic borrowers: limitedliquidity, the usually small number of investors inany given placement, and lower barriers to theflow of information from borrowers to lenders.Debt contracts that are vehicles for information-intensive lending are typically illiquid and held byonly a few investors. A borrower prefers that adebt contract with many restrictive covenants anda high probability of being renegotiated remainwith the lenders in the original negotiations. Thoseare the lenders whose reputations for fairness theborrower originally determined to be adequate.A borrower also prefers that the number of lendersremain small because renegotiation is less costly.Also, flows of certain information, such asborrowers’ projections of future performance, aremore difficult to manage in the public market thanin the private market because of legal issuesrelated to SEC registration.12

The argument that the private placement marketis information intensive does not imply thatregulatory and issuance costs are unimportant.As noted, some issuers that are not informationproblematic borrow in the private placementmarket because fixed costs are smaller, issuance isless time consuming, or disclosure can be avoided.However, the remarks of market participants andevidence presented in the body of the studyindicate that these factors are less important thanthe information-intensive nature of private marketlending as determinants of its structure andoperation. Even if registration requirements werelifted, something resembling the traditional privatemarket would continue to exist. Information-problematic firms would still need long-term,fixed-rate loans, and life insurance companieswould still have long-term, fixed-rate liabilities.As information-problematic borrowers tend to bemedium-sized or small, and thus tend to issuesmaller amounts, lower fixed costs of issuancereinforce the appropriateness of private placementsas a vehicle for information-intensive lending.

2. Terms of Privately Placed DebtContracts

Private placements generally have fixed interestrates, intermediate- to long-term maturities, andmoderately large issue sizes. Their contractsfrequently include restrictive covenants. Theseterms differ from those found in other markets fordebt, for example, the markets for bank loans andpublicly issued bonds.

Issue Size

On average, private placements are larger thanbank loans and smaller than public bonds. In1989, the median new commercial and industrial(C&I) bank loan was for about $50,000; morethan 96 percent were less than $10 million(chart 2). 13 When loan size distributions were

legal liability of issuers encourages issuers when making apublic offering to disseminate only information for which thehistorical foundation is clearly demonstrable.

13. The year 1989 was chosen because, as described in thesection on the credit crunch (part 3, section 1), 1990–92 mayhave been unusual years in the private placement market. Thenonfinancial subset of all new loans and issues was chosenbecause data on other types of bank loans are not available.Sources of data and details of the calculations that producedthe charts are in appendix G.

12. Although law and regulation do not prohibit dissemina-tion of such information, the pattern of court rulings regarding

5

Distribution of size of debt instruments, 19891

By number of issues By volume

2. Loans 3. LoansPercent of issues Percent of volume

0

10

20

30

40

50

60

45.5

24.1

12.9 13.9

2.5 1.0 .03

4. Private placements 5. Private placements

6. Public bonds 7. Public bonds

1. The samples of private placements and publicly issued bonds onwhich the charts are based include only issues by nonfinancial

corporations and exclude medium-term notes, convertible and exchange-able debt, and asset-backed securities.

Numbers may not sum to 100 because of rounding.

0

10

20

30

40

50

60

.6 1.94.5

35.4

26.8 28.3

2.5

0

10

20

30

40

50

60

.9

10.9

27.2

41.7

13.4

3.7 2.20

10

20

30

40

50

60

.01 0.7

5.8

27.424.8

16.3

24.9

0

10

20

30

40

50

60

2.4

12.5

60.6

20.9

3.7

Lessthan.05

.05–.25

.25–1

1–10

10–25

25–100

100–250

250–500

500and

more

Size of loan (millions of dollars)

0

10

20

30

40

50

60

.24.3

45.7

36.3

13.5

Lessthan.05

.05–.25

.25–1

1–10

10–25

25–100

100–250

250–500

500and

more

Size of loan (millions of dollars)

6

computed by volume rather than number, largeloans naturally accounted for a larger share(chart 3). The mean loan size was about $1 mil-lion. The 3.6 percent of loans for $10 million ormore accounted for 58 percent of total loanvolume. Although most are small, loans for asmuch as $100 million are not extraordinary.

In contrast, the median private placement issuedby nonfinancial corporations in 1989 was $32 mil-lion, and the mean was $76 million (charts 4 and5). None was less than $250,000 (compared with70 percent of bank loans in that category). Mostprivate placements were for amounts between$10 million and $100 million.14

The median public issue was $150 million, andthe mean public issue was $181 million. Mostpublic issues were larger than $100 million(charts 6 and 7). None was smaller than $10 mil-lion, and only 15 percent were smaller than$100 million.15

In interviews, market participants oftenremarked that the private market is cost-effectivemainly for issues larger than $10 million, whereasthe public market is cost-effective for issues largerthan $100 million. The data are consistent withthis assertion, as only 10 percent to 15 percent ofprivate placements and underwritten public issues(excluding medium-term note issues) fall belowthe respective boundaries.

These cross-market patterns in size of financingare often attributed to economies of scale in issuesize, that is, to declining costs to the issuer,including fees and interest costs, as issue sizeincreases. 16 Such arguments are usually based on aperception that, holding all else constant, interestrates are lowest in the public market and highestin the bank loan market and on a perception thatfixed costs of issuance are highest in the publicmarket, smaller in the private market, and lowestin the bank loan market. 17

An alternative, possibly overlapping explanationis that the three markets specialize in providingdifferent kinds of financing to different kinds ofborrowers and that relevant borrower characteris-tics are associated with issue size. In particular,borrowers of large amounts are often big andwell-established firms that require relatively littleinitial due diligence and loan monitoring bylenders, whereas those borrowing small amountsoften require much due diligence and monitoring.Thus, borrowers of small-to-moderate amountsusually must borrow in the private placement orbank loan markets, where lenders are organized toserve information-problematic borrowers, whereasthose borrowing larger amounts usually can issuein the public market because they are not informa-tion problematic. As we show later in part 1, bothexplanations are important, but the second expla-nation is probably more important in determiningthe market in which a borrower issues debt.

Maturity and Prepayment Penalties

According to their maturity distributions, commer-cial and industrial bank loans tend to haverelatively short maturities, private placements tendto have intermediate- to long-term maturities, andpublic bonds have the highest proportion of longmaturities. In 1989, the median bank loan had amaturity of just over three months, and the meanmaturity was around nine months (charts 8 and9). 18 Almost 80 percent of loans had maturities ofless than one year. When weighted by loan size,two-thirds of loans had maturities shorter than onemonth. In interviews, market participants oftenstated that banks seldom lend long term, evenwhen the loan interest rate floats. They stated thatloans in the three- to five-year range are notuncommon, five- to seven-year loans are lesscommon, and loans longer than seven years arerare. These remarks are supported by the charts.

The distributions in the charts are for a nonran-dom sample of new loans, not for loans on the

tion ‘‘Type of Payment Stream and Yields.’’ Another problemis that empirical evidence of a relation between yield and issuesize within the public market is weak.

Interest rates may be higher in the private and bank loanmarkets for various reasons, one of which is that lenders mustbe compensated for the fixed costs of due diligence andmonitoring they perform. Lenders charging no fees mustdemand a higher yield on smaller loans to recover such fixedcosts.

18. Sources of data and details of the calculations thatproduced the maturity distributions appear in appendix G.

14. The nonfinancial straight debt subsample represented bychart 4 is fairly representative of all private placements,including convertible, mortgage-backed, and medium-term noteissues. See appendix G.

15. These statistics do not imply that the total number ofprivate placements or public issues exceeds the total number ofbank loans larger than, say, $10 million. The number of newbank loans in any year is very large, so even a small fractionof new loans can be substantial.

16. See Bhagat and Frost (1986), Ederington (1975), andKessel (1971). For a comprehensive list of studies on thepatterns of underwriting fees, see Pugel and White (1985).

17. One problem with this explanation is that interest costsare not always lowest in the public market for all classes ofborrower. This issue is discussed in more detail in the subsec-

7

Distribution of maturities of debt instruments, 19891

By number of issues By volume

8. Loans 9. LoansPercent of issues Percent of volume

0

10

20

30

40

50

60

23.0

35.1

19.5

12.6

8.0 1.0 .4 .04 .04

10. Private placements 11. Private placements

12. Public bonds 13. Public bonds

1. The samples of private placements and publicly issued bonds onwhich the charts are based include only issues by nonfinancial

corporations and exclude medium-term notes, convertible and exchange-able debt, and asset-backed securities.

Numbers may not sum to 100 because of rounding.

0

10

20

30

40

50

60

67.2

15.3

6.6 4.9 4.01.8 .1 0 .01

0

10

20

30

40

50

60

6

2327 27

107

0

10

20

30

40

50

60

5

14

28 29

16

8

0

10

20

30

40

50

60

9 810

45

7

21

Lessthan

1/12

–1/12

1/2

–11/2

1–3 3–7 7–10 10–15

15–20

20and

more

Term (years)

0

10

20

30

40

50

60

108

11

43

7

20

Lessthan

1/12

–1/12

1/2

–11/2

1–3 3–7 7–10 10–15

15–20

20and

more

Term (years)

8

books. Because very short-term loans stay on thebooks for only a short time, a maturity distributionfor a bank’s portfolio of loans at a specific timewould be less skewed toward the short end. Sucha distribution, however, would probably still showbanks to have relatively few loans with maturitieslonger than seven years.

Private placements are generally intermediate tolong term (charts 10 and 11). In the sample, themedian nonfinancial private placement had amaturity of nine years, and the mean maturity wasalso about nine years. No private placements hadmaturities shorter than one year. 19 A moderatefraction had intermediate maturities, but abouttwo-thirds had maturities of seven years orlonger. 20 The median average life of privateplacements is between six and seven years; manyprivate placements include sinking fund provisionsthat cause their average lives to be significantlyshorter than their maturity (chart 14). 21

Nonfinancial corporate bonds issued in thepublic market tend to have long maturities(charts 12 and 13). The median maturity of oursample of bonds issued in 1989 was ten years, andthe mean maturity almost thirteen years. Only17 percent had a maturity of less than seven years.The median average life of public bonds wasaround ten years.

From the standpoint of financial theory, thiscross-market pattern of maturity distributions is abit of a puzzle. Even if long-term borrowers havea strong preference for fixed rates, banks could inprinciple make long-term, fixed-rate loans andexecute swaps to obtain payment streams matchingtheir floating rate liabilities. Apparently, however,they seldom do so. One explanation may be thatthe cost of swaps and other hedges is sufficient tomake such loans unattractive to banks. Anotherpossibility is that the different markets tend toserve borrowers that require different amounts ofcredit evaluation and monitoring and that inequilibrium such differences are responsible for

cross-market patterns in many contract terms,including maturities.

Privately placed debt contracts almost alwaysinclude strong call protection in the form ofpunitive prepayment penalties. 22 As discussed insection 4, buyers of private placements usuallyfund their purchases with long-term, fixed-rateliabilities, and call protection is an important partof their strategy for controlling interest rate risk.Prepayment penalties in the private marketgenerally require the issuer to pay the presentvalue of the remaining payment stream (principalplus interest at the contracted rate) at a discountrate equal to the Treasury rate plus some spread,frequently 50 basis points, but sometimes evenzero. The discount rate for a nonpunitive call-protection provision includes a risk premium

19. A few private placements may have maturities shorterthan a year. The methods used to collect the sample may havecaused private placements with such maturities to be omitted.

20. The maturity distribution was similar when all privateplacements were included in the sample (see appendix G).

21. Descriptive information included with a sample ofprivate placements obtained from Loan Pricing Corporationindicated that about 45 percent of the sample placements hadamortizing features that made their average lives shorter thantheir maturity. This estimate of the fraction of private place-ments that amortize is probably low because other placementsin the sample may have been amortizing but not recorded assuch. About 11 percent of the volume of publicly issued bondsin 1989 was amortizing.

22. For a 1991 sample of private placement commitmentsmade by life insurance companies, 20 percent of privatelyplaced bonds were noncallable, and another 70 percentincluded punitive prepayment penalties. Statistics presented inKwan and Carleton (1993) indicate prepayment penalties mayhave appeared in private placements only recently. Their dataindicate that as recently as 1985–86, only a small percentage ofprivate placements carried prepayment penalties. However,during periods when prepayment penalties were not common,most private placements were noncallable until their averagelife was reached. Prepayment penalties reportedly became morecommon at the behest of investors, who profit from prepay-ments by borrowers wishing to escape the confines of restric-tive covenants.

14. Distribution of average lives of fixed-rateprivate placement commitments measuredas a percentage of the total value of newprivate commitments by major life insurancecompanies, January 1990–July 1992

Percent

25

20

15

10

5

0Lessthan

3

3–5 5–7 7–10 10–15 Morethan15

Average life (years)

Source. American Council of Life Insurance.

9

similar to that of the security itself (that is,associated with the credit quality of the security).When the discount rate fails to include a suffi-ciently high premium, the lender realizes aneconomic gain if the security is prepaid, even ifthe security is matched with liabilities of equalduration.

In the past decade, publicly issued bonds haveincluded increasing call protection. Crabbe (1991a)presents statistics indicating that 78 percent ofpublic bonds issued in 1990 were noncallable forlife, whereas only 5 percent of those issued in1980 were noncallable. 23 Bank loans are typicallyprepayable at any time at par.

Types of Payment Stream and Yields

Most bank loans carry floating interest rates,whereas most private and public bond issues carryfixed rates. Only 3 percent of commitments bymajor life insurance companies to purchase privateplacements from January 1990 to July 1992carried floating rates. Only 95 of the 1,588 privateplacements of debt recorded in the InvestmentDealers Digest (IDD) database for 1989 are listedas having floating rates. 24 The 95 represented6 percent of issues and accounted for 14.3 percentof volume. However, many of the floating-ratefinancings in the IDD sample may, in effect, havebeen bank loans, so the latter statistics probablysubstantially overstate the fraction of privateplacements with floating rates. 25 About 5 percentof the volume of public bonds issued in 1989 hadvariable rates.

Publicly available data on private placementyields in recent years are limited.26 However,many market participants stated that the yieldspreads over Treasuries on traditional investment-grade private placements are higher than thespreads for publicly issued bonds with similarcredit risk. The average differential betweenprivate and public spreads varies over time, butparticipants spoke of a range of 10 to 40 basispoints. The differential is often called a liquiditypremium, but it must also compensate lenders forany costs of credit evaluation and monitoring. Theterm credit analysis premium might be moreappropriate.

Some market participants noted that spreads oninvestment-grade private placements are occasion-ally lower than those on comparable publics forvery brief periods, up to a few days. They attrib-uted this difference to slower adjustment of theprivate market to changes in the yield curve.

Spreads on below-investment-grade privateplacements have often been below those oncomparable public junk bonds. Investors maydemand larger risk premiums on public junk bondsbecause employing the risk control technologies oflender due diligence and loan monitoring is moredifficult in the public markets or because compara-tively rated public issues actually are riskier.

Several researchers have examined the relationbetween issuer quality and yield spread in alterna-tive markets by focusing on the difference betweenthe private placement and the public bondmarkets. For the 1951–61 period, Cohan (1967)found that the spread between yields on privateplacements and yields on public bonds rose as thecredit quality of the issuer increased. Thus, theprivate placement market was relatively moreattractive for lower quality credits. In a study thatcontrolled for the restrictiveness of covenants,Hawkins (1982) confirmed this result for theperiod 1975–77.27 These results are consistent

23. Most of the change in callability occurred forinvestment-grade bonds. During 1987–91, about 90 percent ofnew issues of below-investment-grade bonds were callable atsome time or under some circumstances. See Crabbe andHelwege (1993) for more details. Crabbe (1991a) found thatpublic bond yields were negatively related to the degree of callprotection.

24. These statistics are for all placements in the IDDdatabase, not just issues of nonfinancial corporations. If thesame sample of nonfinancial business nonconvertible debt thatwas the basis of issue-size and maturity statistics is used,4.5 percent of the number and 13.8 percent of the volume havefloating rates.

25. The IDD database was obtained from IDD InformationServices. The data on insurance company commitments arefrom the American Council of Life Insurance. In a database ofbank loans and private placements produced by Loan PricingCorporation, many of the transactions listed as private place-ments and as having floating rates involved only commercialbanks as lenders, providing further evidence that the IDDsample overstates the fraction of placements with floating rates.

26. See part 3, section 1, for charts of a few yield series.27. Shapiro and Wolf (1972) argue that the relationship

between the private–public spread and quality is positivebecause private securities have more restrictive covenants,particularly at the lower quality levels. However, Hawkins(1982) found no relation between covenant restrictiveness andquality for his sample of private placements. Hawkins’s resultis an anomaly; other research and our interviews support theview of a positive relationship between the private–publicspread and credit quality, at least until the recent credit crunchin the below-investment-grade sector of the private placementmarket. However, no empirical test has been adequate tosupport or disprove Shapiro and Wolf’s contention that thepositive relationship was due strictly to differences in covenant

10

with our discussions with market participants, whoindicated that the public market tended to haverelatively little appetite for small-sized, low-quality issues. However, this statement does notnecessarily hold for larger-sized, low-qualitysecurities. The development of the junk bondmarket in the 1980s produced a competitive publicmarket for large, non-investment-grade bonds.Thus, Cohan’s and Hawkins’s findings may nothold for larger issues in the second half of the1980s. Moreover, the credit crunch in the below-investment-grade sector of the private placementmarket since mid-1990 has led to a significantincrease in the average spreads for below-investment-grade private placements of all sizes.

Variety of Securities

A wide variety of securities, including secured,unsecured, asset-backed, senior, and subordinated,is issued in the private placement market. Table 2lists the different types appearing in the IDDdatabase for 1989, with the number of issues andthe volume for each type.

Covenants

Loans to information-problematic borrowers,which are typically medium-sized or smallerborrowers, generally have tighter covenants thanloans to less-information-problematic borrowers.Covenants are one mechanism that lenders can useto reduce the likelihood of borrowers’ takingactions that might lead to an expropriation ofwealth from lenders. In the absence of covenantrestrictions, smaller borrowers are, on average,more likely to attempt such expropriations. Theyoften have less to lose in terms of reputation andare typically more information-problematic so thatdetection and control of expropriation attempts aremore difficult for lenders. Thus, the more informa-tion problematic the borrower, the larger thenumber and the tighter the nature of covenants bylenders. Stated differently, lenders offer smaller,more problematic borrowers lower interest rates inreturn for tighter covenants, and thus such borrow-ers are more willing to negotiate debt contractsthat include tight covenants. Moreover, withoutsuch covenants, lenders might refuse to make

protection between the markets. Our discussions with marketparticipants suggest that, until the recent crunch, the privateplacement market was generally more receptive than the publicmarket to small-sized, lower-grade issuers.

loans to such borrowers regardless of the interestrate.

Covenants in any debt contract are eitheraffirmative covenants, negative covenants, orfinancial covenants (which are a subset of negativecovenants). Affirmative covenants require aborrower to meet certain standards of behavior.They include requirements that the firm stay in thesame business and meet its legal and contractualobligations. They are common in public bonds,private placements, and bank loans. Negativecovenants restrain the borrowing firm from takingactions that would be detrimental to the bondhold-ers. They include restrictions on capital expendi-tures, on the sale of assets, on dividends and otherpayments, on the types of investments that thefirm can make, on the amount of additional debtthat the firm can incur, on liens that the firm cangive to other lenders, and on merger and acquisi-tion activity. 28

Financial covenants restrict measurable financialvariables and can stipulate, for example, mini-mums to be maintained on capital, the ratio ofassets to liabilities, working capital, current ratio(current assets/current liabilities), or the ratio ofearnings to fixed charges. 29 A financial covenantcan be either a maintenance covenant or anincurrence covenant. With a maintenance cove-nant, the criterion set forth in the covenant mustbe met on a regular basis, say at the end of eachquarter. With an incurrence covenant, the criterionmust be met at the time of a prespecified event,such as the firm’s making an acquisition orincurring additional debt.

The number and the tightness of negative andespecially financial covenants in private place-ments are associated with the quality of the issuer,that is, with the degree of both its informationproblems and its observable risk. Tightness refersto the likelihood that a particular covenant will bebinding in the future. Private placements forlower-quality issuers often include many financialcovenants. 30 Contracts for moderately risky issuersoften include only one or two financial covenantswith minimum values farther from current values

28. Appendix E contains a review of the empirical economicliterature on covenants.

29. Fixed charges are interest expense plus rental payments,required repayments of indebtedness, and preferred stockdividends.

30. Issuance of private placements involves several legaldocuments, including the securities themselves and a compan-ion ‘‘securities purchase agreement.’’ Many of the terms of thetransaction, including covenants, are specified in this agree-ment. See the section on agents (part 2, section 2) for a morecomplete list of the documents involved in a transaction.

11

and thus less likely to be violated. Highly ratedissues (A or better) usually have no financialcovenants, unless their average life exceeds sevenyears, in which case an incurrence covenant on adebt ratio is often included. Most financialcovenants in private placements are incurrencecovenants, although occasionally one or twomaintenance covenants may be included, espe-cially when these are designed to match mainte-nance covenants in other debt of the issuer, suchas bank loans.

Bank loan agreements typically contain onlymaintenance covenants. Financial covenants inbank loan agreements are reportedly generallytighter than in private placements, even for

borrowers with the same characteristics. As withprivate placements, the number and the tightnessof bank loan financial covenants depend on thequality of the issuer. Loans to small and medium-sized borrowers typically include many financialcovenants. Very large companies, however,generally obtain bank loan facilities, frequently inthe form of unfunded loan commitments, withoutmeaningful financial covenants.

Indentures in publicly traded bonds, even forbelow-investment-grade bonds, generally containno financial covenants. Beginning in 1992,however, some public junk bonds includedfinancial covenants, especially debt ratio andinterest coverage covenants. Market participants

2. Types of private placement debt issue in the 1989 IDD database

TypeNumberissued

Distributionby number

issued(percent) 1

Volume(millions of dollars)

Distributionby volume(percent) 1

Adjustable-rate notes . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 .13 611.7 .53Bonds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 1.20 532.9 .46Capital bonds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 .19 102.6 .09Capital notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 .19 185.0 .16Collateralized mortgage bonds . . . . . . . . . . . . . . . . . 8 .50 299.0 .26

Collateralized notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 .06 155.0 .14Conditional sale agreement . . . . . . . . . . . . . . . . . . . . . 2 .13 76.4 .07Convertible subordinated debenture . . . . . . . . . . . . 2 .13 68.0 .06Convertible subordinated notes . . . . . . . . . . . . . . . . . 4 .25 20.0 .02Debentures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 .25 102.0 .09

Debt securities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169 10.64 11,587.0 10.10Discount debentures . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 .06 264.3 .23Equipment trust certificates . . . . . . . . . . . . . . . . . . . . . 18 1.13 1,017.5 .89First mortgage bonds . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 3.90 3,017.3 2.63First mortgage financing . . . . . . . . . . . . . . . . . . . . . . . . 6 .38 499.8 .44

First mortgage notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 1.01 986.6 .86Floating-rate notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 3.21 5,933.6 5.17Floating-rate secured notes . . . . . . . . . . . . . . . . . . . . . 1 .06 26.5 .02Floating-rate senior notes . . . . . . . . . . . . . . . . . . . . . . . 3 .19 349.0 .30General mortgage bonds . . . . . . . . . . . . . . . . . . . . . . . . 4 .25 38.8 .03

Guaranteed bonds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 .06 5.0 .00Guaranteed notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 .76 2,986.4 2.60Guaranteed participation certificates . . . . . . . . . . . 4 .25 106.4 .09Guaranteed pass-through certificates . . . . . . . . . . . 4 .25 69.5 .06Guaranteed secured notes . . . . . . . . . . . . . . . . . . . . . . 2 .13 113.5 .10

Guaranteed senior notes . . . . . . . . . . . . . . . . . . . . . . . . 5 .31 941.5 .82Guaranteed subordinated notes . . . . . . . . . . . . . . . . . 1 .06 150.0 .13Industrial development bonds . . . . . . . . . . . . . . . . . . 5 .31 25.0 .02Industrial revenue bonds . . . . . . . . . . . . . . . . . . . . . . . 1 .06 10.0 .01Junior subordinated notes . . . . . . . . . . . . . . . . . . . . . . 7 .44 173.1 .15

Lease-backed notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 .19 359.6 .31Lease certificates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 .13 161.9 .14Lease financing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 .94 1,220.2 1.06Leveraged lease financing . . . . . . . . . . . . . . . . . . . . . . 27 1.70 1,938.6 1.69Leveraged lease notes . . . . . . . . . . . . . . . . . . . . . . . . . . 3 .19 523.1 .46

Medium-term notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 1.26 535.7 .47Mortgage-backed bonds . . . . . . . . . . . . . . . . . . . . . . . . 1 .06 225.0 .20Mortgage-backed notes . . . . . . . . . . . . . . . . . . . . . . . . . 4 .25 631.0 .55Mortgage bonds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 .19 79.4 .07Mortgage financing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 .69 1,005.7 .88

12

disagree on whether this development is perma-nent or transitory. Some participants assert thatsuch issues were bought by investors that did notfully understand the nature of the monitoring andrenegotiation activities associated with theirpurchases and that these investors will stop buyingsuch issues at some future time. Others assert thatsuch issues are, in effect, illiquid and were boughtby mutual funds with staffs of credit analysts,making the instruments functionally equivalent tobelow-investment-grade private placements. Thisdifference of opinion may not be resolved untilsome of the securities deteriorate in quality andmust be renegotiated.

Covenants and Renegotiation

Covenants can limit a borrowing firm’s flexibilityin financial and strategic policymaking. Theconstraint on flexibility can, however, be relaxedthrough implicit or explicit provisions for contractrenegotiation, thus increasing borrowers’ willing-ness to accept tight covenants. For example, if thepursuit of a new strategy, such as the acquisitionof another firm, would violate an existing cove-nant, the borrower may request that the debtcontract be renegotiated. It might, for example,request a waiver of the covenant. The lenderanalyzes the effect of the new strategy, and if the

2. Continued

TypeNumberissued

Distributionby number

issued(percent) 1

Volume(millions of dollars)

Distributionby volume(percent) 1

Mortgage notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 2.33 2,451.4 2.14Nonrecourse secured notes . . . . . . . . . . . . . . . . . . . . . 6 .38 143.2 .12Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 8.19 14,832.7 12.94Other certificates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 .19 64.7 .06Participating certificates . . . . . . . . . . . . . . . . . . . . . . . . 8 .50 241.6 .21

Participating notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 .06 97.0 .08Pass-through certificates . . . . . . . . . . . . . . . . . . . . . . . . 26 1.64 1,171.9 1.02Project financing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 .44 799.2 .70Promissory notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 .13 125.9 .11Receivable-backed certificates . . . . . . . . . . . . . . . . . . 15 .94 2,959.5 2.58

Sale-leaseback financing . . . . . . . . . . . . . . . . . . . . . . . . 1 .06 146.5 .13Second mortgage financing . . . . . . . . . . . . . . . . . . . . . 5 .31 73.8 .06Secured bonds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 .13 195.0 .17Secured loan certificates . . . . . . . . . . . . . . . . . . . . . . . . 3 .19 118.2 .10Secured notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 6.99 6,366.4 5.55

Secured promissory notes . . . . . . . . . . . . . . . . . . . . . . 1 .06 1.4 .00Secured term loan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 .06 55.5 .05Senior debentures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 .38 505.7 .44Senior extendable notes . . . . . . . . . . . . . . . . . . . . . . . . 1 .06 100.0 .09Senior mortgage notes . . . . . . . . . . . . . . . . . . . . . . . . . . 1 .06 10.0 .01

Senior notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 481 30.29 27,026.4 23.57Senior secured bonds . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 .31 491.3 .43Senior secured notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 2.90 3,705.1 3.23Senior subordinated extendable notes . . . . . . . . . . 1 .06 180.3 .16Senior subordinated variable-rate notes . . . . . . . . 6 .37 1,307.0 1.14Senior subordinated debentures . . . . . . . . . . . . . . . . 6 .38 842.5 .73

Senior subordinated notes . . . . . . . . . . . . . . . . . . . . . . 77 4.85 4,624.2 4.03Subordinated bonds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 .06 16.0 .01Subordinated capital debentures . . . . . . . . . . . . . . . . 1 .06 15.0 .01Subordinated debentures . . . . . . . . . . . . . . . . . . . . . . . . 10 .63 655.6 .57

Subordinated floating notes . . . . . . . . . . . . . . . . . . . . . 1 .06 105.0 .09Subordinated notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 4.53 2,967.7 2.59Subordinated secured notes . . . . . . . . . . . . . . . . . . . . . 1 .06 50.0 .04Subordinated variable-rate notes . . . . . . . . . . . . . . . 2 .13 5,000.0 4.36Variable-rate notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 .06 89.1 .08

Total . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1,588 99.94 114,668.4 99.96

1. Numbers may not sum to 100 because of rounding.

13

lender can establish that it will improve theprospects of the firm without increasing the riskto the lender, the lender may agree to waive oradjust the covenant. Even if the new strategyincreases the risk of the loan as it is presentlystructured, the lender may grant a waiver if theborrowing firm agrees to adjust other terms of thedebt contract. In effect, banks, insurance compa-nies, and other lenders to information-problematicborrowers offer contracts that limit borrowerincentives to take risks and still permit flexibilitythrough contract renegotiation. They can offerflexible contracts because of their ability tomonitor and analyze borrowers.

One reason information-problematic firmsseldom borrow in the public market is that thebenefits of covenants are hard to capture therebecause diffuse ownership makes them difficult torenegotiate. Knowing that renegotiation with manylenders is very costly, public bond issuers arewilling to include at most a few loose covenants.Because many covenants are not feasible in publicdebt, much of the monitoring technology ofinformation-intensive lenders is not useful forpublic debt, as public bond buyers may have nolegal mechanism for controlling excessively riskyborrower behavior even if they detect it. Thusmany information-problematic firms are unable toborrow in the public market.

This discussion implies that bank loans, privateplacements, and public bonds will differ not onlyin the number and the tightness of covenants, butalso in the frequency with which the covenants arerenegotiated. As noted, the covenants in bankloans are often relatively tight, implying a highfrequency of renegotiation because bank borrowersstart closer to the limits in their covenants. 31

Those covenants that do appear in publiclyissued bonds are relatively loose, implying alow frequency of covenant renegotiation. Privateplacement covenants and renegotiation rates fallbetween the two extremes but are generallycloser to those of bank loans. The covenants ina private placement are typically violated severaltimes during the life of the security, requiringseveral waivers or other renegotiations of terms

(see Zinbarg, 1975, and Kwan and Carleton,1993). 32

Including extensive, customized covenants ispossible in private placements and commercialloans partly because both are negotiated debtinstruments. Issuers and lenders can tailor contractterms in a way that satisfies the objectives of bothas much as possible. 33 Publicly issued bonds,which are underwritten without any direct negotia-tion between the issuer and the investors, areseldom customized.

Collateral

Some private placements are asset-backed securi-ties, such as leveraged leases, collateralized trustcertificates, and collateralized mortgage obliga-tions. Also, a significant fraction of traditionalprivate placements of straight and subordinateddebt, such as first and second mortgage bonds, aresecured. Approximately one-third of the privateplacements in Kwan and Carleton’s sample weresecured. Similarly, 6 percent of the volume ofprivate issues in 1989 was asset-backed, and21 percent was otherwise secured, for a total of27 percent secured (see table 2). 34 Asset-backedsecurities are more common in the public market,whereas collateral is much less common in otherforms of public debt. In 1989, 24 percent of publicissuance was asset-backed, and only 4 percent wasotherwise secured.35 A much larger fraction ofbank loans is secured. Statistics from the FederalReserve’s Survey of Terms of Bank Lending andthe Federal Reserve/Small Business Administra-

31. Here we refer to the typical middle-market commercialbank loan, in which only one bank or a few banks are involvedin the credit. The large syndicated bank loans, which mayinvolve many banks, may be much more like public securitieswith respect to covenant tightness and renegotiation (seeEl-Gazzar and Pastena, 1990).

32. In the final section of part 1, we argue that the fre-quency of renegotiation is not determined simply by the degreeof covenant tightness. The combination of renegotiation andcovenant tightness may be related to borrower quality.

Kwan and Carleton present evidence that roughly half of asample of private placements were modified at least once; mostmodifications occurred while the loans were in good standing.

33. All of the terms, including the rate, prepayment penalty,take-down provisions, maturity, and covenants, are typicallynegotiated in a traditional private placement; however, a majorfocus of most negotiations is the nature of the covenants.

34. The fraction secured rises to 31 percent if floating andvariable rate instruments, ‘‘loans,’’ industrial revenue ordevelopment bonds, and medium-term notes are omitted fromthe computations. The IDD sample results probably understatethe percentage of private placements with collateral attached,because collateral status must be inferred from listed securitytype. Kwan and Carleton’s finding that about one-third ofprivate placements have collateral is probably the best availableestimate.

35. Federal National Mortgage Corporation (Fannie Mae)and Federal Home Loan Mortgage Corporation (Freddie Mac)obligations were not included in the computations yieldingpublic market percentages.

14

tion’s National Survey of Small Business Financeindicate that about two-thirds of commercial bankloans to nonfinancial businesses are secured.36

Conventional wisdom suggests that bank loansfrequently involve collateral because bank borrow-ers are relatively risky; collateral is less often usedin the private placement market because privateplacements tend to be less risky on average thanbank loans; and collateral is infrequently used inthe public debt market because of the high qualityof the average issuer. As we argue in the lastsection of part 1, collateral is useful not only forcontrolling observable risks but also for solvinginformation problems. Collateral in debt contractshelps minimize the incentives of firm owners toact in ways that are detrimental to lenders.Because these incentives are more acute insmaller, more information-problematic firms,collateral is widespread in the bank loan marketbut rare in the public bond market. 37

Summary

Bank loans typically have floating rates and short-to intermediate-term maturities and are relativelysmall and prepayable at par. They tend to includerelatively tight financial covenants and thus mustfrequently be renegotiated.

Private placements typically have fixed ratesand intermediate- to long-term maturities, aremoderately large, and include punitive prepaymentpenalties. Many include financial covenants.Though these covenants are usually looser thanthose in bank loans, and thus are less easilyviolated, a typical private placement is renegoti-ated at some point. A significant number ofprivate placements include no financial covenants,and thus renegotiation is less frequent for them.

Publicly issued bonds are typically fixed-rate,long-term, large loans. The presence of prepay-ment penalties and other call protection has variedover time. They seldom include financial cove-nants and are seldom renegotiated.

Individual lenders and borrowers take thiscross-market pattern of terms as given and choosethe market(s) with preferable terms. The nextsection explains why borrowers choose the privateplacement market, and section 4 explains whylenders do so.

3. Borrowers in the Private PlacementMarket

Borrowers in the private placement market gen-erally fall into one or more categories (table 3).Most are information-problematic firms or, if theyare not, their financings are complex enough thatonly information-intensive lenders will be willingto buy them. Others have specialized needs thatare a disincentive to public issuance, such as adesire to avoid the disclosure associated withregistration. Finally, some have issues too small tobe done cost-effectively in the public market. 38

Firms that are not information problematic andthat want to issue nonproblematic securities inlarge amounts generally borrow in the publicmarkets. Those wishing to borrow for short termsor at floating rates generally borrow from banks(or similar intermediaries, like finance companies)or issue commercial paper. Some firms with apreference for long-term and fixed-rate funds,other things equal, may nevertheless end upborrowing for short terms and at floating ratesfrom banks.

In describing U.S. capital markets, marketparticipants often speak of a hierarchy of borrow-ers and debt markets based on a concept ofborrower access. In this hierarchy, nonproblematicfirms with nonproblematic issues can borrow inany market; and, for any given financing, theychoose the market offering the best terms.Information-problematic firms or issues, however,effectively have no access to the public markets,because public market lenders are not prepared toperform the necessary due diligence and monitor-ing. Moderately problematic firms may borrow ineither the bank or the private placement market,whereas very information-problematic firms mustuse the bank loan market or cannot issue anyoutside debt (that is, they may be able to borrowonly from those with ownership interest).

36. A distinction should be made between bank creditfacilities extended to large and those extended to small firms.Large firms with access to the commercial paper market andthe public bond market obtain their bank credit facilities(usually lines of credit backing up their commercial paper) onan unsecured basis. However, of those borrowers that dependon commercial banks for their funding, the majority borrow ona secured basis. These borrowers drive the statistics reflected inthe Survey of Terms of Bank Lending and the National Surveyof Small Business Finance.

37. There is also empirical evidence that within the bankloan market riskier borrowers are more likely to pledgecollateral (see Berger and Udell, 1990, 1993a, and 1993b).

38. Table 3 is intended as a summary of our characterizationof private market borrowers; in no way is it intended as acomplete representation of all borrowers or capital markets.For example, it does not include the decisions of borrowersdesiring short-term, fixed-rate loans or long-term, floating-rateloans.

15

From a broad economic perspective, thishierarchy and the differential access of borrowersare not exogenous restrictions on borrowers’actions but are features of an economic equilib-rium that is the outcome of choices by bothborrowers and lenders. For example, in principleinformation-problematic borrowers could issuesecurities publicly, and public bond market lenderscould acquire the expertise needed to perform duediligence and loan monitoring. In reality, however,the choices of lenders and borrowers have resultedin an equilibrium in which information-problematic firms and financings rarely appear inthe public markets (see section 5 for an analysis ofthe economic forces resulting in this equilibrium).In this section, we employ the concepts of accessand of a hierarchy of borrowers because they arepractical and simplify exposition when the focus ison borrowers alone, taking lenders and the broadmarket structure as given. We emphasize,however, that the current pattern of access is notset in stone but could change if the economicfundamentals changed.

The set of firms with access to the privatemarket but not to the public market is not thesame as the set of private market borrowers. Someprivate issues are by companies that have access

to the public market but choose the private marketfor special reasons. Similarly, by asserting thatvery information-problematic firms typically mustborrow from banks, we do not mean to imply thatall bank borrowers are problematic. In fact, banksserve a wide variety of borrowers.

In the remainder of this section, we explain thetaxonomy in table 3 in more detail and thenpresent supporting empirical evidence. Theevidence suggests that, as a group, firms withaccess only to the bank loan and private placementmarkets differ in several respects from those thathave access to the public bond market. Mostnotably, the average borrower in the former groupis significantly smaller than the average issuer inthe public bond market. Smaller-sized issuers areoften more information problematic and thus mustborrow in an information-intensive market.Similarly, firms with access only to the bank loanmarket are significantly smaller and more informa-tion problematic than those having access to theprivate placement market. Another difference isthat the private placements of companies issuingin both the public and the private markets tend tobe considerably larger and more complex thanprivate placements issued by companies thatborrow only in the private market.

3. A taxonomy of market choices of borrowers

Type of borrower and issue

Type of loan borrower wants 1

Long-term,fixed-rate

Short-term,floating-rate 2

Information-problematic firmModerately problematic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Private placement Bank loanVery problematic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bank loan 2,3 Bank loan

Non-information-problematic firm withinformation-problematic issue or transaction . . . . . . . . . Private placement Bank loan

Firm with specialized needs (e.g., speed) . . . . . . . . . . . . . . . . . . Private placement or bank loan 2,3 Bank loan

Non-information-problematic firm withsmall nonproblematic issue . . . . . . . . . . . . . . . . . . . . . . . . . . . Private placement or bank loan 2,4 Bank loan

Non-information-problematic firm withlarge nonproblematic issue . . . . . . . . . . . . . . . . . . . . . . . . . . . . Public bond Bank loan or commercial paper

1. Though a borrower may prefer a long-term, fixed-rateloan, in some cases it may choose or be forced to accept ashort-term, floating-rate loan. This situation is especially likelyfor very information-problematic borrowers.

2. ‘‘Bank loan’’ includes any short-to-intermediate-term,floating-rate loan by any of a number of information-intensiveintermediaries, such as commercial banks or financecompanies.

3. Very problematic borrowers may be forced to choose ashort-term or floating-rate loan because they lack access to the

private placement market, even though in principle they mayprefer a long-term, fixed-rate financing. Firms with veryspecialized needs may find even the private placement marketunable to meet those needs and may turn to the bank loanmarket.

4. Firms wishing to borrow small amounts (less than around$10 million) may choose a bank loan instead of a privateplacement to avoid fixed costs of issuance associated with theplacement.

16

Our principal explanation for these factsinvolves economic theories centered on asymmet-ric information, but at least two other explanationsare possible. One is that small firms tend to issuein small amounts and differential fixed costs ofissuance make the net cost of obtaining funds forrelatively small issues lower in the private market.Another possibility is that smaller firms tend tohave higher observable risk (defined in part 1,section 1) and different classes of lending institu-tion may have different incentives to take risks.Mispriced deposit insurance may give banks thelargest incentives to take risks, whereas theabsence of any guarantees may give public bondbuyers the smallest. State guaranty associations forlife insurance companies, which offer policyhold-ers some protection if their insurer fails, provideintermediate incentives. 39

The three explanations of market choice are notnecessarily mutually exclusive. The evidenceoffers most support for the explanation centered ondifferences in information problems across firms,some support for differential fixed costs ofissuance as a decisive factor in some cases, andlittle support for the explanation centered ondifferences in observable risk across firms. Thetwo most important weaknesses of the thirdexplanation are that contract terms (especiallycovenants) are systematically different in thepublic and private markets for firms with the samebond rating and that enlarging the set of lendinginstitutions under consideration reveals inconsis-tencies. 40 Finance companies, for example, enjoyno guarantees similar to deposit insurance and yetreportedly lend mainly to high-risk borrowers. Allof the evidence is consistent with the view that theprivate market normally receives issues thatrequire lender due diligence or loan monitoring.Our characterization of and explanation for thehierarchy thus focuses on differences in informa-tion problems.

Issuers in the Private Placement Market

Most private placements carry fixed interest ratesand are of intermediate- to long-term maturity.Because firms generally find short-term and

floating-rate loans no harder to obtain thanlong-term, fixed-rate loans (for reasons describedin section 5), we infer that private issuers prefer afixed rate and a long term.41 In this study, we donot analyze firms’ reasons for seeking long-term,fixed-rate debt financing. Commonly cited motiva-tions include a desire to reduce the uncertaintyassociated with interest rate fluctuations or withfunding long-term investments with short-termloans.

Broad Industry Types of Issuers

Most issuers of private placements are nonfinan-cial businesses or financial institutions (table 4).In 1989, businesses accounted for 61 percent ofthe total volume of private placements andfinancial institutions for 30 percent. State and localgovernments were responsible for only thirty-oneissues in 1989, and only four were for more than$25 million.

Information-problematic Firms

Borrowers that are information problematic haveaccess to the bank loan market for working capitaland intermediate-term loans, but normally they

39. The cross-guarantee arrangements differ by state. SeeBrewer, Mondschean, and Strahan (1993) for more details.

40. If bond ratings are a measure of observable risk andobservable risk is the key factor in market access, the contractterms for debt with the same bond rating should be similaracross markets.

41. Bank borrowers, however, may not necessarily prefer ashort term and a floating rate. Quite information-problematicborrowers may prefer a private placement to a bank loan atterms apparently generally available in the two markets, butthey may be able to issue privately only on terms much worsethan average because control of moral hazard risks becomesmore difficult the longer the term of the loan. Effectively, suchfirms lack access to the private market and, in spite of theirpreference for long terms and fixed rates, must borrow in thebank loan market.

4. Distribution of private issuers, by type ofindustry, 19891

Percent

Industry type

Distribution

By volume ofissuance

By number ofissues

Nonfinancial . . . . . . . . . . . . . 55 50Financial . . . . . . . . . . . . . . . . 30 30Utilities . . . . . . . . . . . . . . . . . . 6 6Government . . . . . . . . . . . . . 1 2Unknown . . . . . . . . . . . . . . . . 8 11

1. Numbers may not sum to 100 because of rounding.Source. IDD Information Services.

17

cannot obtain longer-term financing in the publicbond market, as buyers of publicly offered bondsgenerally do not devote staff and other resourcesto the credit analysis required for investment inthese companies. Investors in private placements,however, have developed the necessary capacityfor initial due diligence and loan monitoring andhave achieved economies of scale enabling themto offer favorable borrowing terms to information-problematic firms.

The information problems that borrowers posefor lenders span a spectrum. A firm’s position onthis spectrum tends to be correlated with both itssize and its observable credit risk. Informationproblems posed for lenders tend to increase asborrower size decreases partly because smallerfirms enter into fewer externally visible contractswith employees, customers, and suppliers. Largerfirms enter into more contracts and larger dollarvolumes of contracts. The terms of these contracts,and the large firms’ performance under them, aregenerally observable at relatively low cost; forexample, they are often reported in the financialpress. Facts about contract performance revealinformation about a firm’s likely future perfor-mance, and when such facts are widely available,a firm will find building a reputation for goodperformance easier. In general, the larger the coststo a firm of losing its good reputation, the smallerthe agency problems that must be managed by itslenders. 42

Size may also be related to information prob-lems because size is correlated with age. Youngerfirms, which tend to be smaller, generally have notyet had time to acquire a reputation.43 Similarly,observable credit risk may be positively correlatedwith information problems because risk is corre-lated with age.44 Younger firms tend to be riskier

because they may not yet have achieved organiza-tional stability and the marketability of theirproduct lines may not be well established. Riskmay also be associated with information problemsbecause the incentive to engage in behavior thatexpropriates wealth from lenders is more acute inobservably riskier firms.45

Most issuers of private placements are medium-sized firms and can be described as only moder-ately problematic. Very problematic, typicallysmall borrowers usually lack access to the privatemarket, where lenders’ capacity for due diligenceand especially for monitoring is often not as highas that of banks and some other lenders. Suchborrowers may also be able to obtain better termsin the bank market. A bank loan generally con-tains more restrictive covenants than a privateplacement, has a considerably shorter maturity,and involves more monitoring by the bank.Consequently, smaller companies borrowing frombanks are, in effect, issuing a safer security thanthey would have issued in the private placementmarket and can thus obtain a lower rate. 46 Theshorter maturities, tighter covenants, and floatingrates may make bank loans less-than-perfectsubstitutes for private placements for such compa-nies, but such terms may be preferable to no loanat all or to a loan with a very high interest rate.

Extremely problematic borrowers, such asstart-up or very small firms, may be unable toissue outside debt, especially straight debt, andmay be forced to rely on equity financing. Sourcesof long-term funding for such companies includeequity funds, mezzanine debt funds, and venturecapital funds. These sources are particularlyattractive to firms that are unable to providecollateral for an intermediate-term bank loan.Equity and mezzanine debt funds typically extendfinancing through a combination of subordinateddebt and equity. The principal difference betweenthe two is that equity funds usually require alarger equity interest—often in excess of20–25 percent. Venture capital funds typicallyinvest in developing companies and require an

42. Shockley and Thakor (1993) provide evidence on therelation between firm size and information problems. Theyexamined the announcement effects of bank loan commitmentsobtained by publicly traded firms and found that positiveabnormal returns were higher for smaller firms. They interpretthis result as evidence that the value of information producedby the bank decreased as borrower size increased, implyingthat smaller firms are more information problematic.

43. Other reasons for a relation between risk-taking behaviorand the size of borrowing firms may exist. Recent research infinance implies that, because they tend to have diffusely heldstock, larger firms are controlled more by their managers thanby their shareholders. Because managers’ human capital tendsto be undiversified, they tend to adopt strategies that areless risky than would maximize shareholder wealth. Such atendency offers some protection to bondholders as well.

44. Berger and Udell (1993b) found empirical evidenceassociating firm age and risk. In particular, they found that therisk premium on commercial loans is negatively associatedwith firm age.

45. See Boot, Thakor, and Udell (1991) for a model inwhich the acuteness of moral hazard is positively related to thelevel of observable firm risk.

46. In a world of perfect information, borrowers would beindifferent between a safer bank loan with shorter maturity andstrict covenants and a riskier private placement with longermaturity, looser covenants, and a higher rate. However, thepoint of indifference may not be obtainable when borrowershave better information about their credit quality than lenders.In this circumstance, smaller borrowers may prefer the moremonitoring-intensive credit offered by commercial banks tocredit from insurance companies (see section 5).

18

equity interest. Again, these alternative sources,like bank loans, are not perfect substitutes forstandard private placements, as they require theborrower to give up an equity interest in the firm.For many smaller, owner-managed firms, this maybe a drawback.47 However, equity funds may bethe only source of financing for those firms toosmall or too risky even for the bank loan market.

Firms with Information-problematic Financings

Large, non-information-problematic firms withcomplex financing requirements have often usedthe private placement market. Such companiestend to issue straight debt in the public bondmarket but turn to the private placement marketfor complex transactions that public marketinvestors are not well prepared to evaluate. Privateplacement investors have developed the special-ized skills for analyzing the credit risk of thesetransactions and can command loan spreadssufficient to provide a satisfactory return on theirservices. Examples of such transactions are projectfinancings, capitalized equipment leases, jointventures, and new types of asset-backed securities.The private placement market often serves as atesting ground for new types of securities, whichmay eventually move to the public market asinvestors become more familiar with their struc-ture and the methods for analyzing their creditrisk. One frequently cited example is asset-backedsecurities, which reportedly originated in theprivate market but are now issued in the publicmarket as well.

Firms with Specialized Needs

Another category of firms using the privateplacement market consists of borrowers that couldissue in the public bond market—and in someinstances have done so—but turn to the privatemarket for reasons unrelated to the complexity oftheir financings. Included in this group areprivately held U.S. companies and foreign compa-nies that wish to preserve their privacy. Foreignissuers in the U.S. private placement market alsoavoid the conformance to U.S. generally acceptedaccounting principles that would be required ifthey issued in the public debt market. Corpora-tions contemplating acquisitions or takeovers also

have often relied upon the private placementmarket to protect the confidentiality of theirtransactions and thus decrease the likelihood ofcompeting offers.

Many large companies have used the privateplacement market to raise funds when time is afactor. For example, when in 1989 the Congresssignificantly curtailed the tax advantages of issuingdebt for Employee Stock Ownership Plans(ESOPs), many large firms sold large ESOP-related issues just before the new tax laws becameeffective (July of that year). More than $7 billionof ESOP notes were issued in the private marketin June 1989. More generally, corporations haverelied upon the private market when funds wereneeded before a time-consuming public registra-tion could be completed.48 Often these transactionsare to finance acquisitions, and in many instancesthe issues are sold with registration rights, whichplaces in interest rate penalty on the issuer if thesecurities are not registered publicly within aspecified period of time.49

Another special circumstance leading firms touse the private market involves financings requir-ing nonstandard or customized features, such asdelayed disbursements or staggered takedowns.In general, selling securities with such specializedterms in the public market is not possible, butinvestors in private placements often have theflexibility to accommodate issuers’ preferences.

Firms with privately placed, medium-term noteprograms may also be considered a group thatissues in the private market for reasons relatedmainly to regulatory and practical restrictions inthe public markets. Medium-term notes have madeup an increasing share of total private placementissuance over the past four years. In 1991, forexample, medium-term note issuance totaled$6.2 billion, representing 8.3 percent of totalprivate bond issuance. However, this amount wassmall relative to public medium-term note issuancein 1991, which totaled $73.5 billion. Most firmsthat have private, medium-term note programs areeither private or foreign firms that issue no publicsecurities or public firms that issue privately whilewaiting to establish a public program.

47. Another source of funding for smaller firms is an initialpublic offering (IPO). Again, this type of funding means givingup some ownership of the firm.

48. To a large degree, shelf registration, which has beenpossible since 1982, has eliminated this motivation to issueprivately. For securities not sold under a shelf registration,however, the time to bring the offering to market is consider-ably longer than that for a private placement.

49. For this reason, these securities are often sold to typicalpublic market lenders rather than to private market lenders.

19

Issue Size, Fixed Costs of Issuance,and Choice of Market

Besides information problems and regulatoryrequirements, fixed costs of issuance can affect aborrower’s choice of market. 50 As noted in part 1,section 2, most private placements are for amountsbetween $10 million and $100 million. Focusingfirst on the tradeoff that can be decisive for issuesaround $100 million in size, issuance expenses aregenerally lower for private than for public securi-ties, primarily because they are not registered withthe SEC and because they are not are underwrit-ten. Public issuers incur both registration andunderwriting expenses. For large issues that arenot information problematic, however, the higherfixed costs of a public offering are often offset bythe availability of lower interest rates, whichreflect the greater liquidity of public bonds and thesmaller costs of credit analysis that public lendersbear. Consequently, a company that could issue ineither market would find, all else being equal, thatthe choice hinged upon the size of the offering.For issues smaller than some size cutoff, lowerissuance costs make the private market lessexpensive; for larger issues, lower yields make thepublic market less expensive. Currently, marketparticipants place the break-even point for the twomarkets between $75 million and $100 million.51

At the other end of the spectrum, privateplacements below $10 million are relativelyuncommon for three reasons. First, privateplacements involve some fixed costs of issuance,which can make total costs of small private issueshigh. Also, most buyers of private placementswould demand high interest rates on small issuesto cover their fixed costs of due diligence and loanmonitoring. Finally, prospective issuers of smallamounts tend to be smaller than the average pri-vate market borrower. Such issuers may be too

information problematic for private marketlenders, whose monitoring capacity is not so highas that of banks and some other lenders. Conse-quently, as noted above, small companies tend torely on other sources of funds, one being the bankloan market. As in the private placement market,fees can cause the effective interest rates on bankloans to vary inversely with loan size; nonetheless,for most small borrowers, bank loans are prefera-ble to private placements. 52

Because mainly small and medium-sizedcompanies are information problematic andbecause such companies typically borrow small ormoderate amounts, differential fixed costs ofissuance as well as the need for an information-intensive lender lead such companies to borrow inthe private placement or bank loan markets ratherthan the public market. The most important factorin determining the market in which a firm issues,however, seems to be the extent of the informationproblems the firm poses for lenders.

Other Factors Influencing Market Choice

Apart from gaining access to credit marketsthrough financial intermediaries, information-problematic firms often gain other advantagesfrom issuing private placements. Borrowers havethe opportunity to establish relationships withlenders, the terms of the securities can be tailoredto some degree to suit the borrowers’ needs, theadvancement of funds can be staggered ordelayed, and confidentiality concerning theborrowers’ financial condition and businessoperations can be maintained. Restrictive

50. Fixed costs of issuance include fees paid to an agent orunderwriter, legal and printing costs, and costs of registration(if any). Private issuers often hire agents to assist them withplacements and must pay the agents’ fees, but such fees aretypically smaller than fees for a comparable underwrittenpublic issue.

51. A thorough examination of economies of scale in theprivate placement market has not yet appeared. Blackwell andKidwell (1988) found no evidence of economies of scale in theprivate market, but their study had several limitations (seeappendix H). They also found no evidence of economies ofscale in the public market. This finding stands in sharp contrastto research by Kessel (1971), Ederington (1975), and Bhagatand Frost (1986) and to conventional wisdom in the investmentbanking community. For a comprehensive list of studies on thepatterns of underwriting fees, see Pugel and White (1985).

52. There is empirical evidence that such economies of scalein loan size exist in the commercial bank loan market. Bergerand Udell (1990) suggest that the difference in pricing attribut-able to loan size between a $100,000 and a $1,000,000 com-mercial loan is 190 basis points. However, this result should beviewed as an upper limit because loan size in their model maybe a proxy for risk not controlled for by other variables. In asubsequent study, Berger and Udell (1993b) found no evidencethat size was a statistically significant predictor of loan priceswhen firm characteristics and contract terms were controlledfor. However, the data set for that study was small and limitedto firms with fewer than 500 employees that had relativelysmall loans. Several interpretations can be offered to reconcilethese apparently conflicting results. Because the sample inBerger and Udell (1993b) was truncated, there may not havebeen enough variation to yield significance. Alternatively,economies of scale in loan size may be driven principally bylarge loans that were excluded from that study. Specific studieson the production function shed further light on the issue.Udell (1989) examined the loan review component of commer-cial bank loan department operations and found evidence ofsignificant economies of scale in loan size.

20

covenants, however, impose costly restrictions onborrowers and thus are seen as a disadvantage. Inaddition, prepayment penalties eliminate borrow-ers’ opportunity to refinance the bonds at a costsaving, regardless of the level of interest rates.Nevertheless, medium-sized or hard-to-understandborrowers in search of long-term, fixed-rate fundsare often willing to trade off the risk controlfeatures of private bonds against their perceivedbenefits.

Evidence from Stock Prices

Previous studies of the reaction of stock prices toannouncements that firms had placed bondsprivately support the hypothesis that the privateplacement market is information intensive. In onestudy, Szewczyk and Varma (1991) hypothesizethat, if a company is information problematic,its stock price should rise in response to theannouncement of a private placement. Stockinvestors might view the private placement as asignal that the firm is more creditworthy inasmuchas institutions with access to private informationare willing to invest in the firm. If stock investorsview the successful placement of private debt as asignal that the firm is engaging in value-enhancingprojects, they are likely to bid up the price of thefirm’s stock. In addition, stock investors mayrealize that the private placement probably resultsin the monitoring of the firm’s management byadditional lenders.

For a sample of public utility companies issuingprivate placements between 1963 and 1986,Szewczyk and Varma found that their stock prices,on average, significantly exceeded the predictedchange after the announcement of a privateplacement. Moreover, the greatest positiveresponse was shown by utilities that had notissued debt publicly, that is, those for which theleast amount of public information would havebeen available. As a check on the results,Szewczyk and Varma also examined stock pricesof utilities that had not placed debt privately.In response to the utilities’ announcements ofpublic debt offerings, the changes in their stockprices fell short, on average, of predicted changes.

Research by Bailey and Mullineaux (1989) andVora (1991) also supports a conclusion that privateplacement issuers tend to be information problem-atic. In contrast, James (1987) and Banning andJames (1989) find a negative stock price response,but it comes for private placements used to paydown bank loans. In such situations, the number

of lenders monitoring management may notincrease, and the intensity of monitoring mightdecrease. Taken as a whole, the results support aconclusion that private issuers are informationproblematic, but not as problematic on average asbank borrowers.

Differences among Firms Issuing in thePublic, Private, and Bank Loan Markets

To summarize the preceding discussion, borrow-ers’ access to debt markets is apparently closelyrelated to firm size, with size mainly a proxy forthe degree of information problems that borrowerspose for lenders. Broadly speaking, veryinformation-problematic companies withoutcollateral may be unable to borrow even from aninformation-intensive lender. 53 Such companies,which are typically small, may be forced to relyon venture capital or on other forms of equityfinance. Small firms that are less informationproblematic or those that can provide collateral areconfined largely to the bank and finance companyloan markets for debt financing. Even less prob-lematic firms, which are typically medium-sized,also have access to the private placement market.Large corporations can borrow in any of thesemarkets and in the public bond market. Besidessize of the firm, other characteristics, especiallythose related to the nature and size of the financ-ing, are important in determining a firm’s choiceof credit market.

Empirical evidence supports these assertions.We analyzed the characteristics of firms classifiedaccording to a hierarchy of access to the public,private, and bank loan markets and found a patternof firm sizes and other characteristics consistentwith the explanation of borrowers’ choice ofmarket that focuses on the different informationproblems posed by different firms. However,borrower size is also correlated with issue size andwith observable borrower risk, so the observeddifference in sizes of firms with different levels ofaccess is also potentially consistent with explana-tions based on issuance costs or risk. To evaluatethe relative importance of the three explanations,we looked at several other firm characteristics thatare plausibly correlated either with the degree ofinformation problems or with observable risk.

53. The taking of collateral can be viewed as anothermechanism (like covenants) that lenders use to control risksassociated with information problems.

21

We employed an indirect approach in identify-ing the access of actual firms to the three markets,since access is not directly observable. Wecombined information on corporations in COM-PUSTAT with data on private placements from theIDD database.54 Corporations in COMPUSTATwith a long-term credit rating are assumed to haveaccess to the public bond market, inasmuch asthey must have issued corporate bonds at sometime to have received a bond rating; those withouta rating are assumed to lack access to the publicmarket. 55 In 1989, 1,149 corporations in COM-PUSTAT had ratings and thus constitute the publicmarket group, that is, those corporations with theability to raise funds in public debt markets. 56 Toform a group of firms with access to the privateplacement market but not to the public market,companies listed in the IDD private placementdatabase as issuing in 1989 were matched withthose in COMPUSTAT that had no credit rating.The cross-matching of the two databases yielded atotal of 113 such companies, which make up whatis called the private market group. Those firms inCOMPUSTAT that in 1989 had neither a creditrating nor outstanding long-term debt but that didhave some short-term debt outstanding wereassumed to be constrained to borrow only frombanks (or other, bank-like intermediaries such asfinance companies); this collection of firms iscalled the bank group and contains 472 members.Finally, those firms in COMPUSTAT that hadneither a credit rating nor any outstanding debt(short or long term, except for trade debt) in 1989were assumed to be shut out of all three debtmarkets. This collection of firms is called theequity group and consists of 613 firms.

This method of classifying firms is far fromperfect for various reasons. First, and perhapsmost important, implicit in the definition of eachgroup is an assumption that a company cannot tapa particular debt market if it has not actually done

so. This assumption is clearly not correct in allcases. For example, several firms classified in thebank group probably could have issued in theprivate or public bond markets on standard termsbut simply chose not to do so. Firms that issuedprivate placements before 1989 but not in 1989are less likely to fall in the bank group becausesuch firms probably still showed long-term debt ontheir balance sheets in 1989. Second, according tothe bank group definition, the presence of short-term debt on the balance sheet indicates the firm’sability to tap the bank loan market. However,COMPUSTAT’s definition of short-term debtincludes loans from various lenders: loans payableto stockholders, officers of the company, parents,subsidiaries, and brokerage companies as well asloans payable to banks, finance companies, andother intermediaries. Our aim is to include in thebank group all firms that have access to banks orbank-like intermediaries, but several firms withoutsuch access were probably misclassified (theyshould be in the equity group) because they hadloans outstanding from stockholders or othernon-intermediary sources. Third, many equitygroup firms may have had bank lines of credit thatwere simply unused at the end of their 1989 fiscalyears. 57 Fourth, the presence of a credit rating inCOMPUSTAT implies only that a firm once hadaccess to the public bond market, not that it hadaccess in 1989.

The private market, bank, and equity groups arealso undoubtedly biased selections of firmsbecause only those firms that appear on theCOMPUSTAT tapes have been selected.COMPUSTAT’s bias toward large firms meansthat the firms in these three groups are likelylarger on average than corresponding groups offirms for the economy as a whole. Other charac-teristics may show some bias as well. However,the bias probably makes observed differencesacross groups less dramatic. Consequently, anydifferences found in the analysis are unlikely to bethe result of this sampling bias.

Finally, the criteria used to define the fourgroups focus on the characteristics of the firm, noton the characteristics of the debt issue. As men-tioned earlier, some firms that could readily issuestraight debt in the public market may be con-strained to the private market for more compli-cated issues such as some leases or project

54. COMPUSTAT provides no information on the types oflong-term debt on balance sheets. For information on the IDDdatabase, see appendix G.

55. In COMPUSTAT, the ratings are by Standard and Poor’s(S&P). Virtually all investment-grade firms and almost allbelow-investment-grade firms with public debt outstandinghave a rating from S&P. Further, in 1989 S&P rarely provideda debt rating for a firm with some private or bank loan debtbut no public debt outstanding.

56. The year 1989 was chosen to avoid distortions causedby the credit crunch in the private placement market in1990–92, which is described in part 3, section 1. Also, since1989, S&P has rated an increasing number of private place-ments, so our method of identifying public market group firmswould be less reliable for those years.

57. Many smaller firms reportedly choose a date for the endof their fiscal year that is at a point in their annual cycle atwhich debt is at a minimum in an attempt to window-dresstheir year-end balance sheets.

22

financings. We address this issue later in thissection.

Despite these classification problems and biases,we believe our method of classifying firms ison the whole roughly accurate and that thedistinctions that are revealed are economicallymeaningful.

The firm characteristics examined include thesize of the firm, measured by total assets, sales,and market value of equity. We also looked at thethree-year growth rate of sales, return on assets,(measured by operating income before deprecia-tion divided by total assets), research and develop-ment (R&D) expenditures as a percentage of sales,the fixed-asset ratio, the ratio of total debt toassets, and the interest coverage ratio.

Differences in firm size across groups, measuredby total assets, total sales, or market value ofequity, are pronounced (table 5). Firms in thepublic market group are much larger than firms inthe private market group, which in turn are verymuch larger than firms in the bank or equitygroups. For example, mean assets of companies inthe public market group are $6.3 billion, consider-ably larger than the mean of $3.4 billion for firmsin the private market group. The means for thebank and equity groups are even smaller at$40 million. These differences in means are allstatistically significant at the 1 percent level. Themedians have a similar relationship among thethree groups.58

Table 5 presents statistics for three othervariables that are plausibly correlated with thedegree of information problems posed by firms:the ratio of R&D expenditures to sales, thefixed-asset ratio, and a three-year average growthrate for sales. Many economists have used R&Dexpenditures as a proxy for the potential severityof agency problems between shareholders anddebtholders. 59 The risk implicit in research anddevelopment cannot be easily monitored byoutsiders, including debtholders, as a firm withlarge R&D expenditures has wide scope fordiscretionary behavior. For example, such a firmmay require intensive monitoring by debtholdersto ensure that it is working on a mundane researchproject with a moderate but fairly sure payoffrather than a longshot with a high payoff. Inten-sive monitoring may be required to ensure that thefirm is not underinvesting in projects with positive

net present values (Myers, 1977). R&D-intensivecompanies, being inherently more informationproblematic than other firms, may therefore findbanks more receptive to providing financingbecause banks can monitor more intensively thanlenders in the public markets. The evidenceprovided by this variable on the intensity ofmonitoring in the private placement marketgenerally conforms with our hypothesis aboutdifferences in the degree of information problemsacross the four groups. Mean R&D intensity ishigher in the private placement market than in thepublic market, although the medians are about thesame. The significantly higher R&D intensity forthe bank and equity groups than that for theprivate market group indicates that issuers in theformer groups tend to require significantly moremonitoring by lenders than do issuers of privateplacements.

A similar hierarchy of information problems issuggested by the fixed-asset ratios. Firms with alarge percentage of fixed assets may have fewerinformation problems than other firms for tworeasons. First, they may be able to offer some oftheir fixed plant and equipment as collateral topotential creditors. Second, monitoring the sale offixed assets or their transformation from one useto another may be easier than it is for more liquidassets. The more of a firm’s assets that are fixed,therefore, the smaller may be the scope forshareholders to engage in wealth-transferringinvestment projects.

As one moves from the public to the private tothe bank and finally to the equity group, thedecline in fixed-asset ratios implies that informa-tion problems increase. The higher fixed-assetratio for the bank group compared with that forthe equity group suggests that a small firm’sability to provide fixed assets as collateral may bea factor in its ability to obtain bank loans.

Sales growth rates may also be correlated withinformation problems in that high growth may bea sign of entry into new lines of business or ofbeing in lines of business that are in rapidlydeveloping markets. Both situations offer morescope for agency problems to surface during thelife of a debt contract. The evidence from thisvariable, however, is weaker than that from R&Dintensity and the fixed-asset ratio: The mean issignificantly smaller for firms in the public groupthan for those in the private group, a findingconsistent with private issuers requiring moremonitoring; the median is smaller as well. Valuesfor the private group do not differ significantlyfrom those for the bank and equity groups,

58. Easterwood and Kadapakkam (1991) also find thatindustrial firms using the private market are smaller than thoseusing the public market.

59. See Prowse (1990) and references therein.

23

however, and the medians display an unevenpattern.

On the whole, the results for the three variablesconform with our hypothesis about the differingdegree of information problems posed by the fourgroups of firms. They also accord with theremarks of market participants, who asserted thatbuyers of private placements, especially the largerlife insurance companies, engage in organized andactive monitoring, although their monitoringprograms are typically not so intensive as those ofbanks.

Average return on assets and two measures ofleverage, total-debt-to-asset ratios and interest-

coverage ratios, are indicators of observable creditrisk. As noted in part 1, section 1, informationproblems and observable credit risk are separateconcepts, and in principle there is no reason thatthe pattern of credit risk should be different ininformation-intensive and non-information-intensive markets. In practice, however, both arerelated to borrower size.

Caution should be used in interpreting thedifferences between the bank and equity groupsand the other groups in the measures of leverage,as firms in the former groups either had nolong-term debt outstanding or no debt at all ontheir balance sheets (according to COMPUSTAT

5. Mean characteristics of firms with access to the public, private, bank loan, and equity markets1

Variable

Group of firms

Public(1)

Private(2)

Bank(3)

Equity only(4)

Billions of dollars

Total assets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3 2 3.4 2 .04 .04(1.5) (.5) (.05) (.09)

Total sales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 2 1.0 2 .04 .04(1.0) (.4) (.03) (.03)

Market value of equity . . . . . . . . . . . . . . . . . . . . . . . . . 1.8 2 .7 2 .10 .07(.3) (.06) (.01) (.02)

Percent

Three-year average sales growth . . . . . . . . . . . . . . . 6.2 3 13.9 14.4 19.3(4.9) (8.1) (.9) (5.0)

Ratio

Ratio of R&D expenditures to sales . . . . . . . . . . . .03 .07 2 .38 .39(.01) (.009) (.04) (.05)

Fixed-asset ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .49 .42 2 .31 .28(.46) (.40) (.21) (.19)

Return on assets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .08 .06 2 −.17 −.04(.08) (.07) (−.05) (−.01)

Ratio of total debt to assets . . . . . . . . . . . . . . . . . . . . .40 .40 2 .74 2 0(.30) (.40) (.15) (0)

Interest coverage ratio . . . . . . . . . . . . . . . . . . . . . . . . . . 3.50 2 2.70 2 1.30 2 40.40(2.10) (1.80) (−.02) (15.10)

Memo:Number of firms in group . . . . . . . . . . . . . . . . . . . . . . 1,149 113 472 613

1. Numbers in parentheses are medians. Public firms arethose with access to the public, private, and bank debt markets.Private firms are those with access to the private and bank debtmarkets. Bank firms are those with access to the bank loanmarket only. Equity firms are those with no access to the bankloan, private placement, or public bond markets.

2. Mean of group is significantly different from mean ofgroup in column to the right at the 1 percent level.

3. Mean of group is significantly different from mean ofgroup in column to the right at the 5 percent level.

24

and ignoring trade debt). Thus, zeros will appearin either the numerator or the denominator of theratios for many equity group firms, making theratios poor measures of the riskiness of these firmsand influencing the mean and median values forthe groups.

A comparison of ratios for the public andprivate placement groups indicates that differencesin credit risk may not be as great as differences ininformation problems. Both the mean and mediandebt-to-asset ratios and the return on assets aresimilar for the two groups. Median interest-coverage ratios are also similar, but the meaninterest-coverage ratio is significantly higher forthe public group. The implication is that privateplacements issuers may be somewhat riskier as aclass, but not a great deal riskier, than public bondissuers. Comparing ratios for the private placementand bank groups, the means of the three ratiosdiffer significantly; the medians also differ aspredicted except for the debt-to-asset ratio. Itappears that members of the bank group poselarger observable credit risks for lenders. 60

On the whole, these results accord well with theremarks of market participants, who oftendescribed private issuers as ‘‘solid companies’’that have taken a major step in ‘‘graduating’’ fromhaving access only to the bank loan market butthat are typically ‘‘not quite ready’’ to issue in thepublic bond market. Some investors also indicatedthat their historical experience of loss on privateplacements and public bonds was virtuallyidentical within credit-rating categories. Thestatistics presented here and the remarks ofparticipants offer little support for a hypothesisthat low observable credit risk is the primaryrequirement for a borrower to have access to thepublic market, instead of only the private place-ment and bank loan markets. The existence of thepublic junk bond market and the fact that contractterms, especially covenants, and lender duediligence and monitoring activities differ acrossthe public and private markets for borrowers withthe same bond ratings also imply that informationproblems are a more important determinant ofmarket access than observable credit risk.

In sum, if the groups of firms analyzed here arerepresentative of borrowers’ access to debtmarkets, then their characteristics are broadly

consistent with our explanation of the factorsinfluencing borrowers’ choice of debt market.Corporations able to borrow in the public marketstend to be large and to pose relatively fewinformation problems for lenders; thus they canborrow from a wide variety of lenders. Companiesissuing in the private but not the public market aresmaller and appear to be more informationproblematic; however, they apparently do notrepresent substantially greater observable creditrisks. Such companies must be served byinformation-intensive lenders. The companiesconfined to the bank loan market or to equitymarkets are much smaller, are more informationproblematic, and pose larger pure credit risks.Consequently, they require the greatest degree ofdue diligence and loan monitoring by lenders, orthey are unable to issue debt at all. The informa-tion problems associated with smaller andmedium-sized firms and their increased need forinformation-intensive lenders appear to be themajor reasons for the size pattern observed amongthe three groups and for the differential access offirms to credit markets.

Companies Issuing in Both the Publicand the Private Markets

As mentioned earlier, some firms that couldreadily issue straight debt in the public marketmay be constrained to the private market for morecomplicated issues, such as leases or projectfinancings. To obtain evidence regarding thishypothesis, we examined differences in privateissues between our private market group and afifth group of firms that issue in the private marketeven though they have previously tapped thepublic market for funds. This group, called thepublic–private group, consists of those firms thatare listed in the IDD database as having issued aprivate placement in 1989 and listed on theCOMPUSTAT tape as having a bond rating. Itcomprises 109 firms, with 175 issues of privatedebt in 1989.

Several differences exist between the privatedebt issues of firms in the private market groupand those in the public–private group (table 6).The much larger average size of private placementissues by the public–private firms than that of theprivate market group firms reflects the much largersize of firm in the former group. In addition, themix of securities issued by the private marketfirms differs significantly from that of the public–private group in terms of their credit analysis

60. The median debt-to-asset ratio may be lower for thebank group because firms with access only to banks may relymore on trade credit than do public or private placement groupfirms and trade debt is not included in COMPUSTAT’s debtmeasures.

25

requirements. We define ‘‘complex’’ securities tobe equipment trusts, lease-backed bonds, leveragedleases, receivables-backed bonds, and variable andfloating rate notes. Complex securities appear inthe public market, but in many cases they requireinvestors to engage in sophisticated and intensivecredit analysis. We define ‘‘simple’’ debt securitiesto be senior securities, secured notes, mortgage-backed notes, debentures, and medium-term notes.Simple securities likely require less in the way ofdue diligence and monitoring. Measured by thenumber of issuers and by the dollar amountissued, the percentage of total private issuance inthe form of complex securities was much higher in1989 for public–private group firms than forprivate market group firms. Conversely, a muchhigher percentage of total private issuance byprivate market firms in 1989 was in the form of

simple debt. 61 This evidence supports the hypothe-sis that firms with access to the public market maychoose to issue more complex securities in theprivate market, where the capacity of investors forcredit analysis is greater.

The average size of simple and complex issuesfor the two groups of firms is consistent with theproposition that issuance cost is of secondaryimportance in determining market choice byborrowers (last two rows of table 6). The averageissue size for complex private placements was$576.5 million, suggesting that on the basis ofissuance costs alone the public market would havebeen the appropriate choice. That they were issuedin the private market indicates that due diligenceand loan monitoring requirements were such thatonly information-intensive lenders would buy theissues.

Simple securities issued by the public–privategroup could be issued in either the public or theprivate market because they require relatively lowlevels of due diligence and monitoring by lenders.In this case, issuance costs are likely to be adominant consideration. The average issue size of$50.8 million for the simple securities issued bythe public–private group in the private market isconsistent with this notion, because the privatemarket reportedly offers lower total costs forissues of that size.

Summary

The marked differences between firm characteris-tics and loan characteristics for the various groupssupport the hypothesis that firms have differentialaccess to the three markets according to theinformation problems they pose for lenders. Atone end of the scale are small, relatively unknownfirms posing significant information problems thatrequire extensive due diligence or loan monitoringby lenders. These firms tend to have access onlyto relatively short-term loans provided by banksand other bank-like intermediaries, which have thestaff and expertise to undertake information-intensive lending and which limit borrowers’risk-taking through tight covenants or collateral inloan agreements.

Somewhat less information-problematic,typically larger borrowers can issue in the private

61. This pattern appears robust to plausible variations in thedefinitions of simple and complex securities. In particular, thepattern persists if all secured bonds (including mortgage bonds)are defined as complex.

6. Characteristics of the private placements offirms with and without access to the publicdebt market, 1989

Characteristics

Group 1

Private Public–private

Number of issues . . . . . . . . . . . . . . 140 175

Percent

Firms issuingSimple debt 2 . . . . . . . . . . . . . . . . 96.5 87.6Complex debt 2 . . . . . . . . . . . . . . 3.5 12.4

Issuance in form ofSimple securities . . . . . . . . . . . . 91.9 74.2Complex securities . . . . . . . . . . 8.1 25.8

Millions of dollars

Private issue sizeMean . . . . . . . . . . . . . . . . . . . . . . . . . 100.1 184.5Median . . . . . . . . . . . . . . . . . . . . . . 48.5 60.0

Mean size ofSimple issues 2 . . . . . . . . . . . . . . . 86.7 50.8Complex issues 2 . . . . . . . . . . . . 236.0 576.5

1. The private group comprises firms that issued a privateplacement in 1989 and have no access to the public debtmarket. The public–private group comprises firms that issued aprivate placement in 1989 and have access to the public debtmarket. Access to the public debt market is defined by theexistence of a public debt rating.

2. Simple debt includes senior securities, secured notes,mortgage-backed notes, debentures, and medium-term notes.Complex debt includes lease-backed bonds, leveraged leases,receivable-backed bonds, and variable and floating-rate notes.

26

placement market. These borrowers must still beserved by an information-intensive lender, but theypose fewer problems than the average bankborrower. They can issue longer-term debt withsomewhat looser covenants than those in bankloans.

Finally, well-known, typically larger firms thatare not information problematic and that havestraightforward financings can issue in the publicdebt markets, where lenders perform little duediligence and loan monitoring and where cove-nants are relatively few in number and loose innature.

The reasons for this equilibrium pattern ofborrower characteristics are discussed in part 1,section 5. The pattern is evidence that the variousdebt markets are imperfect substitutes for oneanother, which implies that breakdowns or failuresin one market may have material effects on firmsthat rely on that market for a major part of theirfinancing needs, even if other markets are func-tioning normally. An example of such a break-down is discussed in part 3, section 1.

4. Lenders in the Private PlacementMarket

Although various institutions hold some traditionalprivate placements in their portfolios, life insur-ance companies purchase the great majority ofthem. For example, for a sample of 351 place-ments issued during 1990–92, life insurancecompanies purchased 83 percent of dollar volume,whereas the next largest type of investor, foreignbanks, purchased only 3.6 percent (table 7). 62

Lending in the private placement market is alsoconcentrated in the hands of a relatively fewlenders. Although the sample lists 315 separateinvestors, most participated in only one deal or ina few deals and bought only small amounts. Thetop twenty investors were life insurance companiesand accounted for 56 percent of dollar volume.

The concentration of private placement lendingin the hands of a relatively few lenders and a fewtypes of lender has probably occurred for fourreasons. First, the large proportion of information-problematic borrowers in the traditional privatemarket necessitates that major buyers of privateplacements be intermediaries. Intermediaries cancapture economies of scale in due diligence andmonitoring and can also build and maintain overlong periods the reputations for fair dealing thatare important when debt contracts must includecovenants.

Second, financial intermediaries tend to special-ize in a few liability-side lines of business (forexample, banks mainly take deposits) at leastpartly because of regulatory restrictions. Givensuch specialization, the natural tendency of lendersto seek superior risk-adjusted returns will lead tospecialization on the asset side. Different debtinstruments are associated with different patternsof risks, and different lenders have differentabilities to implement a cost-effective and appro-priate set of risk control measures in order to earnsuperior risk-adjusted returns on any given type ofasset. For example, banks’ short-term depositliabilities lead them to make short-term loans,whereas insurance companies’ longer-termliabilities lead them to purchase longer-termassets.

62. The sample was drawn from Loan Pricing Corporation’sDealscan database. An effort was made to include onlytraditional private placements, but some Rule 144A issues mayhave been included.

The shares shown in the table should be viewed as roughapproximations for several reasons. First, the sample may notrepresent the population of private placements issued duringthe period. Second, the sample includes some issues thatappear to be bank loans, not traditional private placements, ineffect. Removal of these would reduce the shares of U.S. andforeign banks and of U.S. savings and loans and mutualsavings banks. Finally, the sample period is unusual in that itinvolves a severe credit crunch in the below-investment-gradesegment of the market (described in part 3, section 1). Becausepurchases of private placements by finance companies havetraditionally been below-investment-grade securities, the lowshare of finance companies may not be representative of otherperiods nor of their current share of all outstanding placements.Thus, the types of lender are listed in table 7 in the order ofimportance as indicated by anecdotal evidence, not in the orderof their share of the sample.

7. Lender shares of the market for traditionalprivate placements, 1990–92Percent

Type of lender Share of volume

Life insurance companies . . . . . . . . . . . . . . . . . 82.6Pension, endowment, and trust funds . . . . . 1.7Finance companies . . . . . . . . . . . . . . . . . . . . . . . . 1.4Mutual funds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7Casualty insurance companies . . . . . . . . . . . . 1.4U.S. commercial banks . . . . . . . . . . . . . . . . . . . . 3.3Foreign banks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6U.S. savings and loans and

mutual savings banks . . . . . . . . . . . . . . . . .7U.S. investment banks . . . . . . . . . . . . . . . . . . . . . .9Unknown . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.7

Source. Calculations based on data from Loan PricingCorporation.

27

The risks most commonly associated withtraditional private placements of debt are creditrisk, asset concentration risk, interest rate risk, andliquidity risk. Extensive credit evaluation andmonitoring are required to control credit risk inprivate placements, whereas appropriate diversifi-cation can control asset concentration risk. Interestrate risk may be controlled by matching privateplacements with liabilities of similar duration, orother hedges. With regard to liquidity risk, if alender holds private placements, its liabilities mustnot be redeemable on demand, or other parts of itsportfolio must be sufficiently liquid to meet anylikely withdrawals. The relative efficiency withwhich different classes of financial intermediarycan undertake to control these risks, as well aslegal and regulatory constraints, determines theinstitutional pattern of investments in privateplacements. Although many financial intermediar-ies can effectively control the credit and assetconcentration risks associated with private place-ments, life insurance companies are especiallywell positioned to control the liquidity and interestrate risks. 63

A third reason for the concentration of privateplacement lending is the concentrated structure ofthe insurance and related markets. At the end of1991, the twenty largest life insurance companiesheld 51 percent of industry assets. Because thesecompanies have a large volume of funds to invest,their domination of the private placement marketis natural. A final reason for concentration is thatlarge lenders have an advantage in obtainingprivate placements because their large volume ofinvestments permits them to participate in themarket continuously, giving them up-to-dateinformation about the state of the market (seepart 2, section 2).

Apart from the statistics shown in table 7 andsome data for the life insurance industry that arediscussed in parts 2 and 3, little detailed informa-tion on investors in private placements is publiclyavailable. Consequently, much of our discussion isbased on interviews with market participants. Tosummarize this information, life insurers buy abroad spectrum of private placements, but many ofthem focus on senior, unsecured debt. Financecompanies are also said to be significant buyers ofprivate debt, but they tend to specialize in high-risk investments and, consequently, require that

borrowers provide collateral and equity kickers,such as warrants or convertible bonds. They havedeveloped special expertise in due diligence andmonitoring involving collateral and equity fea-tures. Though commercial banks have the capabili-ties for credit analysis, they are not significantbuyers of private placements, probably becausetheir short-term, liquid, floating-rate liabilities arenot well matched by private bonds. Regulatoryand other restraints prevent or discourage majorinvestors in public bonds, such as most pensionfunds and mutual funds, from investing heavily inprivate bonds.

Life Insurance Companies

Market participants estimate that life insurerspurchase between 50 percent and 80 percent ofnew issue volume each year (table 7 supportsestimates at the high end of that range). Atyear-end 1991, life insurers held $212 billion ofprivate placements in their general accounts,representing 26 percent of their total bond hold-ings and 16 percent of their general accountassets. 64

The twenty largest insurance companies, asmeasured by total assets, accounted for 68 percentof industry holdings of private placements at theend of 1992. Furthermore, for this group, privateplacements were 39 percent of total bond holdingsand 22 percent of general account assets. The nexteighty largest insurers account for most of theremaining industry holdings of private placements,and within this group, several companies havesizable portfolios.

Some idea of how the life insurance industryallocates its funds among different classes ofprivate bonds can be obtained from the ACLIInvestment Bulletin, which provides data on thecomposition of new commitments of funds toprivate placements by major insurance companies.Life insurance companies strongly prefer fixed-rateprivate placements: In 1992, more than 97 percentof their commitments were fixed rate. Securitizedinstruments, mainly mortgage-backed securities,were 13 percent of commitments although, asdiscussed in part 1, section 2, a much largerfraction probably carried collateral. Insurers investprimarily in medium- to long-term maturities; lessthan 10 percent of their 1992 commitments had an

63. Though a lender with floating-rate liabilities mightcontrol interest rate risk with swaps or other hedges, one withshort-term liabilities might find the risks associated with majorinvestments in long-term, illiquid assets difficult to manage.

64. Information on private placements held in separateaccounts is not available.

28

average life of three years or less, with more thanhalf having average lives between five and tenyears.

This concentration on medium- to long-term,fixed-rate debt is sensible because such securitiescan easily be matched with the life insuranceindustry’s long-term, fixed-rate liabilities. Manyprivate placements also have sinking fund provi-sions that further enable insurers to match the cashflow of their investments with that of theirliabilities. The strong call protection that is typicalof private placements also facilitates matching.65

Life insurance companies buy private placementsfrom firms in all sectors of the economy. Mosttend to diversify across a broad range of indus-tries, although many have favorite industries inwhich they have a particular expertise. In 1992,78 percent of their total commitments went to thenonfinancial sector, with just over 30 percentgoing to manufacturing, 8 percent to the oil, gas,and mining industries, and another 20 percent tothe utilities, communication, and transportationsectors. Life insurance companies have sharplyincreased their purchases of securities issued byforeign companies, or U.S. subsidiaries of foreigncompanies, to over 7 percent in 1992 from lessthan 3 percent of total commitments in 1990.

The large insurers’ investment in risk-controltechnology is extensive.66 Most of these insurershave large staffs of credit analysts, who evaluatethe credit quality of potential issuers and monitorthe health of firms to which credit has beenextended. Most conduct a quarterly review of eachprivate bond held in their portfolios, with a moreformal annual or semiannual review. Violations ofcovenants or requests for waivers of covenantsgenerate further reviews. The costs of risk-controloperations are covered by the higher risk-adjustedyield of private placements relative to publicbonds, which require little or no active monitoringby security holders. 67 The private market providesborrowers willing to compensate the lender forthese risk-control services.

The large investment in credit evaluation andmonitoring leads most large insurance companies

to concentrate on more complex credits; however,strategies vary even among these companies.Besides dominating the straight debt sector of themarket, life insurers buy other types of privatesecurities, such as convertible debt or asset-backedbonds, though their share of these sectors issomewhat lower. In terms of credit quality,insurers focus primarily on securities rated A andBBB (chart 15). At the end of 1992, around17 percent of total private bonds held by thetwenty largest companies were rated belowinvestment grade; however, substantial variationexists, with some companies having up to 38 per-cent of their private portfolio in below-investment-grade bonds and others having almost none atall. 68 Securities in this credit range, particularlythose rated just below investment grade (whichinsurers often refer to as Baa4 securities), arefavored by those insurance companies attemptingto gain maximum advantage from their creditanalysis and monitoring skills. These insurancecompanies like to take advantage of the largedifference in yields between investment-grade andbelow-investment-grade credits by lending tostrong BB-rated companies. However, others aremore conservative and focus solely on issues ratedA or higher. 69

65. See part 1, section 2, for statistics on call protection inprivate placements.

66. See Travelers (1992) for a description of the credit-monitoring practices at insurance companies.

67. The premium on private bonds as compared with that onpublic bonds is often characterized as reflecting the fact thatprivate bonds are typically less liquid than public bonds. Webelieve that the premium is due more to a requirement tocompensate investors in private bonds for their intermediationservices than to any differences in liquidity.

68. There are regulatory restrictions on the amount ofbelow-investment-grade bonds a life insurer can hold.

69. Over the past two years, in response to regulatorypressures and concerns about their financial condition, insurershave withdrawn substantially from the below-investment-gradesector of the private market. See part 3, section 1.

15. Distribution of credit ratings of privateplacements held in the general accounts of lifeinsurance companies, December 31, 1992

Percent

0

10

20

30

40

A orhigher

BBB BB B orlower

Credit rating

Source: National Association of Insurance Commissioners.

29

According to market participants, smallerinsurers typically have much less extensiverisk-control technology at their disposal. Theytherefore tend to concentrate on higher-quality,less-complex credits. They also may participate indeals that larger insurance companies have alreadycommitted to, using the presence of these largerinsurers as a signal that the deal is a favorableone.

Most insurance companies rely heavily onagents for prospective transactions, although somedirect lending occurs between an insurer and itsexisting borrowers. Only the very largest insurancecompanies originate new transactions on a regularbasis, and only one insurer syndicates privatebonds. The largest insurers generally prefer to bethe sole source of funds for an issuer. However,many issues are larger than the maximum amountthat individual insurers permit to be lent to oneborrower; a typical issue may have up to a halfdozen insurance companies funding it. Insurerstypically fund between 5 and 20 percent of thedeals that are marketed to them.

Most large insurers invest in both public andprivate bonds, and they have allocation mecha-nisms to alter the flow of money into thesemarkets as spreads change in the two markets.Until recently, the groups within large insurancecompanies responsible for purchasing private andpublic bonds were usually separated; however,some companies have recently combined thegroups. Market participants report that manymedium-sized insurers have for some time used asingle group to make all investments in bonds.

Finance Companies

Finance companies have traditionally participatedin the lower-rated or mezzanine sector of theprivate bond market, specializing in collateralizeddebt or debt with equity kickers. Rates in thissector of the market may be fixed or floating.Finance companies’ choice of this market sectorfollows naturally from their historical concentra-tion in secured or asset-based lending. Returns onprivate placements required by finance companiesare generally well in excess of the yields on theless risky, straight bonds purchased by insurancecompanies.

According to market participants, the participa-tion of finance companies in the private market ismuch more concentrated than that of insurancecompanies. Among the twenty largest finance

companies, only a half dozen or so provide asignificant volume of funds, although some othersare attempting to expand their presence in themarket. Outside the top twenty, few financecompanies participate at all.

Pension Funds

Pension funds, which are significant investors inpublicly issued corporate bonds, have not been bigbuyers of private placements, except for a fewstate pension funds. Market participants suggestseveral reasons.70 First, many pension funds havecharters preventing them from investing inbelow-investment-grade or illiquid assets.Although in practice some higher-rated privatebonds may be more liquid than some publicbonds, market participants generally considerprivate placements to be illiquid. Second, few stateor corporate pension funds are currently staffedwith the credit analysts and other personnel thatwould allow them to become direct investors inprivate placements. Instead, staffing is directedtoward public market investments, which requiremuch less credit analysis. A decision to hire thenecessary staff and install the expensive internalmonitoring systems to support direct investment inprivate placements would require a long-termcommitment to the private market by the pensionmanager. Few pension fund managers thus farhave been willing to so commit. Even if theyshould wish to do so, state pension funds faceproblems in hiring the necessary personnel. Staffsize and salaries are generally controlled by thestate legislatures, and increasing the size of creditanalysis staffs is thus cumbersome and time-consuming.

As an alternative to direct investment, somepension funds have turned to money managers,often insurance companies. Indirect investments,however, are on a fairly small scale, no doubtpartly because pension fund managers are reluctantto invest even indirectly in a market with whichthey are unfamiliar. The private market operateslargely in conformance with unwritten, informalrules enforced by the desire of the major agents

70. Pension funds appear to be the main suppliers of fundsin the private equity market, which they finance indirectlythrough investments in limited partnership investment funds(see appendix B). Their preference for private equity overprivate debt appears mainly to stem from a desire to earn themuch higher returns that are potentially available in the privateequity market.

30

and buyers involved to maintain their reputations.To investors that are outsiders, the way the marketoperates may thus be hard to understand, whichmay inhibit them from risking their money there.Also, insurance companies themselves, who wouldbe the primary source of the managerial resourcesnecessary for any large-scale activity in this area,have been reluctant to set up separate accountprivate placement funds financed with institutionalmoney.71 They apparently see little investorinterest in such funds or do not wish to interruptthe flow of private placements to the companyitself. 72 Furthermore, market participants reportthat investor experience with at least one separateaccount fund has not been good because themanaging insurance company, lacking a stake inthe separate account investments, did not performadequate monitoring.

Banks

Banks, which are information-intensive lenders,might also be expected to have interest in thetypes of securities offered in the private market.However, for several reasons they seldom buyprivate placements. First, banks’ liabilities are notlong term and are not as well matched withprivate bonds on the asset side as they are withshort-term, floating-rate loans. Of course, the swapmarket can be used to turn fixed-rate assets intofloating-rate, but longer-term swaps are expensive.Second, the looser covenants on private place-ments relative to bank loans may make somebanks uncomfortable.

Bank purchases of private placements aresubject to some regulatory restrictions, which aredescribed in appendix C. Bank holding companiesmay purchase privately placed debt securitieswithout restriction. Banks themselves may alsopurchase them but must place them in a loanaccount and follow traditional underwritingprocedures. The latter requirement means thatbanks must evaluate and document the credit-

worthiness of the borrower as they would with anybank loan. As credit analysis is the norm in theprivate placement market, such evaluation anddocumentation do not appear to be onerousrequirements. Some issuers attempt to createinterest among banks and life insurance companiesby constructing offerings that include both privatebonds and loans, which are identical in their termsexcept for the classification of the instrument.

Other Investors

Other investors in private bonds include mutualfunds, foreign banks, endowment funds, and somevery wealthy individuals, but the combined marketshare of these participants is quite small. Mutualfunds are restricted to holding no more than15 percent of their assets in the form of illiquidsecurities. An exception exists for private place-ments purchased pursuant to Rule 144A. For suchsecurities, the mutual funds’ boards of directorsmay classify the securities as liquid if theydetermine that the securities are generally as liquidas comparable publicly traded bonds.73 Mutualfunds have recently increased their investments inprivate placements, especially underwrittenRule 144A securities, so current restrictions mayin the future be constraints. In the mid-1980s,Japanese banks aggressively bought private bonds,but since then they have disappeared from themarket.

Summary

A capacity for due diligence and loan monitoringis a prerequisite for a significant volume of directinvestment in private placements by a lender. Lifeinsurance companies, finance companies, banks,and a few other financial institutions have thiscapability. However, life insurers dominate theprivate debt market, partly because they have largepools of funds suitable for investment in longer-term, fixed-rate, illiquid securities. Insurance

71. Insurance company separate accounts operate much likemutual funds, in that buyers of liabilities associated withseparate accounts bear the risk of investments, whereasliabilities associated with the general account of an insurergenerally offer fixed payoffs backed by the insurer’s capital.

72. Insurance companies have recently had a strong appetitefor investment-grade private placements. Because of theirwithdrawal from the below-investment-grade sector of themarket, however, they appear to have excess capacity toanalyze and monitor lower-quality credits (see part 3,section 1).

73. Rule 144A securities are described in detail in part 2,section 1. After life insurance companies, mutual funds havebeen the largest buyers of 144A private bonds. According tothe SEC staff report on Rule 144A (September 1991), insur-ance companies bought just over two-thirds of the privatebonds issued under Rule 144A in the eighteen months follow-ing the rule’s adoption, mutual funds bought 15 percent,pension funds bought 5 percent, and banks and thrifts bought4 percent. More recently the share held by mutual funds hasincreased.

31

companies also have a long history of lendingdirectly to middle-market firms that has allowedthem to develop expertise and cost-effectiverisk-control technologies. This expertise mayconstitute a barrier to entry for other financialinstitutions, including most pension funds, whichmight otherwise seem to be suited to lending inthis market. Regulatory and other obstacles alsodiscourage pension funds and mutual funds fromparticipating heavily in the market. Banks have thenecessary expertise in credit monitoring but forseveral reasons have not found private placementsto be suitable investments. As in other creditmarkets, finance companies have carved out aniche in the private market for higher-risk borrow-ers. This segment constitutes a small part of theoverall market, but it is one in which the insur-ance companies have little interest.

Some market participants feel that, over thelong term, pension funds will overcome theobstacles that have precluded their large-scaleparticipation to date and will be much moreimportant providers of funds in this market, muchas they have replaced life insurance companies asthe major source of finance in the private equitymarket. The immense growth of their assetsprojected for the future may force pension plans toconsider investments in markets new to them.However, the information-intensive nature of thetraditional private market is unlikely to change; soif pension funds are to be a larger source offinance, they will likely become so throughindirect investments in funds managed by insur-ance companies. The alternative is for pensionfunds themselves to acquire the capacity forconducting due diligence and monitoring.

5. Private Placements, the Theory ofFinancial Intermediation, and theStructure of Capital Markets

As previously discussed, contract terms andborrower and lender characteristics differ systemat-ically across major debt markets (see table 8).Privately and publicly issued bonds tend to havelong terms and fixed rates, whereas bank loanstend to have short terms and floating rates. Publicissues and issuers are the largest on average, andbank loans and bank borrowers are the smallest.On average, public issues are the least risky,private placements are riskier, and bank loans areriskier still. Public issuers tend to be well known;private placement issuers tend to be less wellknown; and bank borrowers tend to be companies

for which relatively little information is availablepublicly. 74 Public issues rarely include collateraland have few restrictive covenants. In traditionalprivate placements, collateral is not uncommon,and covenants often impose significant restrictionson borrowers. Bank loans, in contrast, tend both tobe secured and to have tight covenants. The termsof public issues are rarely renegotiated, whereasthose of most private placements are renegotiatedat least once, and those of bank loans are fre-quently renegotiated. Public issues are typicallyliquid, whereas most private placements and bankloans are illiquid. Investors in public securitiescarry out relatively little due diligence andmonitoring of borrowers. Investors in bank loansand private placements perform significantamounts of due diligence and loan monitoring.Most private placement lending is done by asingle type of financial intermediary, life insurancecompanies.

This section offers an integrated explanation forthese patterns, elements of which have beenmentioned in previous sections. The explanation iscentered on hypotheses that borrowers pose aspectrum of information problems for lenders andthat lenders address such problems through duediligence at loan origination and loan monitoringthereafter. Firms that are not information problem-atic can borrow in any market but generally findcosts to be lowest in the public bond (and com-mercial paper) markets. Information-problematicfirms find it optimal to negotiate debt contractsthat include certain kinds of covenants andcollateral and to deal with lenders having acapacity for due diligence and loan monitoring.Such lenders also can flexibly renegotiate thecontracts, which is efficient since covenants arefrequently violated.

Such contracts are not well suited to the publicmarkets that exist today; instead they are issued inthe bank loan and private placement markets. 75

Lenders in these markets are almost alwaysfinancial intermediaries, and they tend to focustheir investments in assets that match the rate andmaturity structure of their liabilities. Correlationsamong several factors—the degree of information

74. This statement refers to the average information-problematic borrower. As noted earlier, banks provide large,well-known companies with lines of credit to finance workingcapital or to back commercial paper.

75. Of four major markets, two are for nonproblematicborrowers (public bond and commercial paper), and two are forproblematic borrowers (bank loan and private placement). Oneof each pair of markets is for short-term, floating-rate debt; theother of each pair is for long-term, fixed-rate debt.

32

problems posed by borrowers, the borrowers’ size,their risk, and the size of the loan—account forborrowers being smaller and riskier on averageand loans smaller on average in such information-intensive markets than those in the public markets.

The differences between the average borrowerfrom banks and the average issuer of privateplacements arise mainly because monitoring andrisk control mechanisms involving covenants andcollateral are less reliable the longer the averagelife of a loan is. Such mechanisms are mostimportant in loans to very information-problematicborrowers; these borrowers can obtain long-termloans only at high rates, if at all. Thus, they tendto borrow in the shorter-term market, causing theaverage severity of information problems posed byborrowers to be highest there.

This explanation accounts for more of thefeatures of the U.S. financial system than dotraditional explanations that focus mainly onregulation and considerations of asset–liabilitymatching as causal factors. It raises many newquestions, however. Why must lenders toinformation-problematic borrowers be intermediar-ies? How do due diligence and loan monitoringmitigate risks associated with information prob-

lems? What is the role of covenants and collat-eral? Why are these risk-control mechanisms lesseffective for long-term loans? Why would aborrower agree to a contract with tighter ratherthan looser covenants? Why are covenantsfrequently violated and renegotiated, and why is alender’s reputation for flexibility in renegotiationimportant? Why is information-intensive debtilliquid? Why is the public market ill-suited toinformation-intensive lending (what is to preventpublic market lenders from acquiring capacity indue diligence and loan monitoring)? Whatcomplex of characteristics is required to make alender competitive in an information-intensive debtmarket?

Most of these questions have been addressed atleast to some extent by existing financial theory.In the rest of this section, we review and extendrelevant areas of financial theory to answer thesequestions and to provide a sense of the founda-tions of this study. We find existing individualtheories of covenants and financial intermediationto be inadequate as a basis for a theory of finan-cial structure. We propose a merging and anextension of the two bodies of theory in the formof a ‘‘covenant–monitoring–renegotiation’’ (CMR)

8. Credit market characteristics

Characteristic

Market

Bank loan Private placement Public bond

Maturity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Short Long Long

Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Floating Fixed Fixed

Severity of information problems posedby the average borrower . . . . . . . . . . . . . . . . . . . . . . . . High Moderate Small

Average loan size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Small Medium to large Large

Average borrower size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Small Medium to large Large

Average observable risk level . . . . . . . . . . . . . . . . . . . . . . . . High Moderate Lowest

Covenants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Many, tight Fewer, looser Fewest

Collateral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Frequent Less frequent Rare

Renegotiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Frequent Less frequent Infrequent

Lender monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Intense Significant Minimal

Liquidity of loan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Low Low High

Lenders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Intermediaries Intermediaries Various

Principal lender . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Banks Life insurance cos. Various

Lender reputation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Somewhat important Most important Unimportant

33

paradigm in order to answer to the questionsposed earlier. We evaluate the consistency of theparadigm with some recent research in empiricalfinance and graphically relate borrowers andcapital markets on an information continuum.

Asymmetric Information, Contracting,and the Theory of Covenants

Two imperfections of capital markets are at theheart of many of the contracting problems thatshape debt markets. 76 First, the interests ofbondholders and stockholders of borrowing firmsare not always aligned; second, parties to financialcontracts are not likely to be equally informedabout the characteristics of the issuing firm.77 Theinformational advantage borrowers have overlenders leads to two kinds of bondholder–stockholder conflict. First, once a debt contract issigned, borrowers have incentives to expropriatewealth from lenders (moral hazard). Second,before a contract is signed, potential borrowershave incentives to understate the risks they willpose for lenders, including moral hazard risks. Asimple example of moral hazard risk is providedby Black (1976), who noted that ‘‘there is noeasier way for a company to escape the burden ofa debt than to pay out all of its assets in the formof a dividend, and leave the creditors holding anempty shell’’ (p. 7). In the absence of sufficientlypowerful constraints or capacity for lendermonitoring and enforcement capacity, such actionsmay be either unobservable by the firm’s bond-holders or beyond their control. Smith and Warner(1979) identify four major kinds of moral hazardthat lenders must control:

Dividend payment. If a firm issues bonds andthe bonds are priced assuming the firm willmaintain its dividend policy, the value of thebonds is reduced by raising the dividend rate

and financing the increase by reducing invest-ment. At the limit, if the firm sells all its assetsand pays a liquidating dividend to the stock-holders, the bondholders are left with worth-less claims.

Claim dilution. If the firm sells bonds, and thebonds are priced assuming that no additionaldebt will be issued, the value of the bondhold-ers’ claims is reduced by issuing additionaldebt of the same or higher priority.

Asset substitution. If a firm sells bonds for thestated purpose of engaging in low varianceprojects and the bonds are valued at pricescommensurate with that low risk, the value ofthe stockholders’ equity rises and the value ofthe bondholders’ claim is reduced by substitut-ing projects which increase the firm’s variancerate.

Underinvestment. Myers (1977) suggests that asubstantial portion of the value of the firm iscomposed of intangible assets in the form offuture investment opportunities. A firm withoutstanding bonds can have incentives toreject projects which have a positive netpresent value if the benefit from acceptingthe project accrues to the bondholders. 78

Covenants may alter the relationship betweenbondholders and stockholders in two fundamentalways. First, covenants affect the relationship whenthe borrowing firm is in financial distress byproviding lenders with a mechanism for earlyintervention. This intervention may take one ofseveral forms: forced bankruptcy, a renegotiatedrestructuring, or the imposition of additionalconstraints on firm behavior. This can be viewedas the role of covenants ex post, which is to

76. Modigliani and Miller (1958) argued that if capitalmarkets are perfect and there are no taxes, a firm’s capitalstructure is irrelevant—that is, the value of a firm is indepen-dent of the way it is financed. They argued that the structure ofthe right-hand side of the balance sheet will determine the waythe firm’s cash flow will be allocated, but it will not affect theamount of the cash flow. By extension, the structure of thefirm’s financial contracts (that is, the right-hand side claims) isalso irrelevant. For example, pledging the firm’s equipment toone lender as collateral will alter the allocation among creditorsin liquidation but will not alter the amount allocated.

77. In keeping with the literature on contracting, we refer toa borrowing firm’s bank, its private creditors, and its publiccreditors collectively as its bondholders.

78. Smith and Warner (1979), pp. 118–19. Black andScholes (1973) and Merton (1974) have shown that optionpricing theory can be used to value debt and equity. In effect,issuing a bond is equivalent to the owners’ selling the firm’sassets to the bondholders in exchange for a package consistingof the proceeds from the bond issue, a claim on the firm’sdividends, and a European call option on the firm’s assets withan exercise price equal to the face value of the bonds and anexercise date equal to the bond’s maturity. Because stockhold-ers’ equity is essentially a call option, the stockholders’ interestis to increase the riskiness of the firm’s assets—just as theowner of a call option benefits from an increase in the risk ofthe stock on which the option is written. Ceteris paribus, thegain in stockholders’ equity (that is, the European call option)will be offset by the loss in the value of the bonds.

34

permit these interventions after the consequencesof the firm’s actions have been revealed.

Second, and possibly more important, is the roleof covenants ex ante. Debt contracts that includecovenants can effectively constrain the ability ofstockholders to engage in strategies designed toexpropriate wealth from bondholders or otherwiseto engage in actions that are detrimental tobondholders. Smith and Warner document thatcovenants of the kind observed in private place-ments and bank loan contracts can mitigatebondholder–stockholder conflicts. They alsodemonstrate that contracting is not a zero-sumgame. Terms of contracts affect not only thedistribution of wealth between the bondholdersand the stockholders but also the total value of thefirm. Covenants can increase a firm’s value(relative to value under a contract without cove-nants) by providing disincentives to, or restrictionson, exploitive stockholder behavior. For example,asset substitution incentives may be so powerfulthat under a contract without constraints stock-holders are willing to substitute an asset with alower expected return so long as it has a suffi-ciently higher risk than the existing asset. Such asubstitution increases stockholder wealth eventhough it decreases the firm’s total value becausethe bondholders lose more than the stockholdersgain. Rational bondholders, however, anticipatethat some of their claim will be expropriatedthrough asset substitution and price their bondsaccordingly (that is, they demand a higher rate).Thus, in the absence of constraints on assetsubstitution, equilibriums involving debt financingshave two features: First, firms will take more risksthan in the presence of constraints (the incentiveto substitute assets does not disappear just becausethe bondholders’ anticipation of asset substitutionis reflected in the interest rate). 79 Second, a firm’sstockholders will absorb the loss in the firm’svalue that results from the asset substitution.Consequently, any covenant that restricts assetsubstitution (for example, a requirement to stay inthe same business, a restriction on asset sales, orrestrictions on investments, mergers, and acquisi-tions) can increase firm value. Because ultimatelythe stockholders gain from such restrictions in

equilibrium, they will agree to covenants in debtcontracts.

The theory of covenants and renegotiationemphasizes that covenants must be based onmutually observable and verifiable characteristics,actions, or events (see, for example, Berlin andMester, 1992, and Huberman and Kahn, 1988).Covenants cannot, for example, be written oncharacteristics, actions, or events that are observ-able only by the stockholders and not by thebondholders. Covenants also need to be observableand verifiable by third parties, such as a court oflaw.80 Characteristics, actions, or events that areobservable but not verifiable cannot be includedin covenants; however, they may still significantlyaffect an optimal debt contract. For example, abank can refuse to renew a one-year loan on thebasis of a mutually observable but nonverifiablecharacteristic but would have difficulty legallydeclaring a two-year loan in default at the end ofthe first year because of a violation of a covenantwritten on that same characteristic. This examplesuggests that, in many cases, a short-term loanwithout a covenant may dominate a longer-termloan with a covenant (see Berlin, 1991, and Hartand Moore, 1989).

Although covenants can be written only onobservable and verifiable characteristics, they maybe related to nonverifiable and even unobservablecharacteristics. This relation greatly increases thepower of covenants for mitigating bondholder–stockholder conflicts. A relation between observ-ables and unobservables may exist for tworeasons. First, observable, verifiable actions orevents may be correlated with nonverifiable orunobservable actions or events. For example, thetrue risk of a firm, that is, the volatility of itsreturns, may not be observable. However, itscurrent ratio may be correlated with this volatilityand, therefore, serve as a proxy for risk. Second,an observable characteristic, action, or event maybe related to an unobservable characteristic, action,or event through either self-selection or incentiveeffects. For example, a firm’s ability to takeunobservable risks may be much greater inindustry A than in industry B. Consequently, acovenant that restricts a firm to industry B limits

79. Even when bondholders price in anticipation of assetsubstitution, stockholders are still better off substituting assets(that is, switching to the riskier strategy) than they would besticking with the safe strategy. If stockholders stuck with thesafe strategy, the bondholders, having priced their bonds on thebasis of a risky strategy, would enjoy a windfall.

80. It is not difficult to imagine a wide variety of observ-able, but not verifiable, characteristics, actions, or events. Forexample, qualitative attributes of owner–managers wouldgenerally be mutually observable but not verifiable. Somecharacteristics of firms may be too complex to include incovenants.

35

the ability of a firm to alter its (unobservable) riskprofile. A financial covenant may have the sameeffect. For example, a minimum current ratiorequirement may constrain a borrower from sellingon account to slow-paying customers. 81 Selling tosuch customers necessarily increases the observedliquidity risk of the firm because its current ratiodeteriorates. It may also create an incentive toincrease the firm’s unobservable risk, to the extentthat the firm has more ability to sell to unobserv-ably (to the lender) riskier customers if it ispermitted to extend trade credit on longer terms.82

Collateral can also be used to mitigatebondholder–stockholder conflict. For example, alien on firm assets (inside collateral) preventsborrowers from selling those assets without lenderapproval. 83 This limits the firm’s ability toexpropriate lender wealth through asset substitu-tion (see Smith and Warner, 1979). Owners’pledging personal assets as collateral for a corpo-

rate loan (outside collateral) effectively increasestheir equity exposure. Such increased exposuremay have important incentive effects dependingon the owner’s level of risk aversion. Outsidecollateral may also be useful in solving adverseselection problems because a borrowing firm’swillingness to pledge collateral may reveal its truequality (see Chan and Kanatas, 1985), or it maybe useful in solving incentive problems because itmay alter the marginal return to risk shifting (thatis, asset substitution) (see Boot, Thakor and Udell,1991).

Information-based Theories of FinancialIntermediation

Some theories of financial intermediation focus onthe information problems associated with financialcontracting. Such theories emphasize that financialintermediaries enjoy economies of scale inproducing information about borrower qualitybecause of fixed costs of producing informationabout any given borrower. Fixed costs makehaving only one or a few lenders for each bor-rower economical. Many small individual inves-tors can delegate information-production responsi-bility to a single large financial intermediary thatalone bears the fixed costs. 84

Commercial banks and life insurance companiesare financial intermediaries in the spirit of thesemodels. Both types of institution collect fundsfrom many relatively small investors. Theseinvestors (depositors or policyholders) delegatedue diligence and monitoring responsibility to theintermediary.

The Covenant–Monitoring–RenegotiationParadigm

The literature on covenants and that on financialintermediation offer considerable insight intothe ways in which markets address issues ofbondholder–stockholder conflict. Separately,however, they fall short of describing the real-world financial landscape. The literature oncovenants has not adequately addressed theassociation of covenant constraints with informa-tion production—due diligence at the originationstage and monitoring after loan funding. In

81. If a company sells on account to slow-paying customers,its turnover of accounts receivable will slow down (that is, thedays turn, or the average days an invoice is outstanding, willincrease) as its accounts receivable increase. Assuming noincrease in the firm’s capitalization (that is, its stockholders’equity plus long-term debt), this increase in accounts receivablewill have to be financed by an increase in current liabilities.Because the current ratio is defined as current assets/currentliabilities, the current ratio necessarily decreases.

82. A firm’s accounts receivable generate risk because thefirm is extending credit to its customers. It is generallyassumed that slower-paying customers are riskier on averagethan faster-paying customers (ignoring for purposes of thisdiscussion the ability of the firm to affect the payment patternsof any individual customer through discounts and collectionactivity). The firm chooses whether to sell to safe or to riskycustomers based on the risk–return trade-off. This decision willbe reflected in the firm’s turnover of accounts receivable andits current ratio, which can be observed by the bank. However,it can also affect the firm’s unobservable risk. Let us assume,for example, (1) that all customers who pay their trade debts inless than thirty days (fast payers) are low risk, (2) that half ofall potential customers who pay in more than thirty days (slowpayers) are low risk and the other half of the slow payers arehigh risk, and (3) that the risk quality of the slow payers isperfectly observable by the firm extending the trade credit, butonly the accounts receivable turnover and the current ratio areobservable by the bank. Under these assumptions, a constrainton the firm’s trade policies through a minimum current ratiowould effectively limit the ability of the firm to change itsunobservable risk profile because it would truncate the firm’sdecision set.

83. See Berger and Udell (1990) for a discussion of thedistinction between inside and outside collateral. Essentially,inside collateral involves pledging firm assets to a particularlender, creating a creditor preference. Aside from lender controleffects, this type of collateral alters the payoff allocationamong creditors in liquidation but does not affect the aggregateamount of the payoff. Outside collateral involves pledgingnonfirm assets (typically by the firm’s owners) to specificlenders. This type increases the assets available to satisfycreditor claims in liquidation (that is, it increases the amountof the payoff in liquidation).

84. See, for example, Boyd and Prescott (1986), Diamond(1984), and Ramakrishnan and Thakor (1984).

36

addition, although covenant constraints can bevalue-enhancing to the extent that they minimizecosts associated with borrower–stockholderconflict, they may also be value-reducing in thatthey may prevent the borrowing firm frominvesting in positive-value projects. A completetheory must account for the fact that borrowerschoosing contracts with restrictive covenants alsotend to be served by lenders that provide flexiblerenegotiation of the contracts. Borrowers agreeingto contracts with covenants want the option to payoff their loan or the ability to renegotiate thecontract if they are constrained from investing invalue-enhancing projects. Like loan origination,loan renegotiation requires that lenders produceinformation.

The existing information-based theories offinancial intermediation fall short because theygenerally do not capture nor analyze the dynamicnature of intermediated loans: Intermediariesproduce information both at the origination stage(lender due diligence) and on a more-or-lesscontinuous basis after funding (monitoring). 85

Dynamic production of information in conjunctionwith covenant restrictions enables a lender todeclare a loan in default and demand immediaterepayment if necessary while still offering flexibil-ity through renegotiation. The information-basedmodels also generally do not explain why someborrowers are served in intermediated markets andothers in the public debt markets and why thecontracts offered in those markets differ sodramatically. 86 What has been missing in thetheoretical literature until quite recently is a linkbetween the theory of covenants, the mechanismof renegotiation, and the information-based theoryof financial intermediation.

An initial attempt at a link was offered byBerlin and Mester (1992), who developed atheoretical model in which financial intermediariesextend loans that include restrictive covenants toborrowers. In their model, covenants are beneficialbecause they limit the problems discussed earlier.

Berlin and Mester’s financial intermediaries useobservable, but not necessarily verifiable, informa-tion to form the basis for renegotiation; renegotia-tion is beneficial because it enables borrowingfirms to invest in positive-value projects that theyotherwise would have forgone because of covenantrestrictions. 87

In a more general setting than Berlin andMester’s, covenants can be viewed as a mecha-nism for triggering reevaluation of borrowerriskiness by a financial intermediary. A covenantviolation does not necessarily (and, indeed, usuallydoes not) indicate that risk has increased.88 It canoccur, for example, because a borrower wishes toinvest in a new value-enhancing project that wouldtrigger a violation of a covenant restricting newinvestments. Lenders can determine the appropri-ate response to a violation only if they analyze theborrower’s situation, that is, if they produceinformation at the time of the violation. Simplemonitoring during the life of the loan is often oflittle use except insofar as it improves the lender’sability to respond to covenant violations because,in the absence of a violation, lenders typicallycannot change the terms of the loan no matterwhat their monitoring reveals.

Because financial intermediaries have a compar-ative advantage over small individual investors inproducing information about borrower risk and infacilitating renegotiation, loans with covenants,especially financial covenants, are in generalnaturally made by intermediaries. Also, intermedi-aries may have more incentive to consider grant-ing a covenant waiver than individual investors, asindividual investors that do not expect to makemany loans regularly in the future may perceivethat they have little to gain from granting awaiver, whereas intermediaries that regularlyinvest in the market may profit from a reputationfor being constructively flexible. Such a reputationmay give intermediaries another competitiveadvantage over individual investors in conductinginformation-intensive lending.

85. Campbell and Chan’s (1992) model involves informationproduction at both stages but does not consider many of theimplications.

86. Only a few papers have attempted to explain thesimultaneous existence of public debt and intermediated debt.Diamond (1991), for example, developed a model in whichreputation determined whether firms were able move from(monitored) intermediated debt to (unmonitored) public debt.Although this model captures some of the essential features ofthe financial structure that we observe, it does not address thedifferences in the contracts offered in these markets. Moreover,it does not capture the dynamic nature of information produc-tion in conjunction with covenant restrictions, which wasdescribed in part 1, section 2.

87. Also, as pointed out by Smith and Warner (1979),renegotiation with a few well-informed intermediaries is lesscostly than renegotiation with the large number of investors,which is common in the public debt market. El-Gazzar andPastena (1990) found empirically that dispersion of investorownership is positively associated with the looseness ofcovenants.

88. That most renegotiations are not associated with firmdeterioration is consistent with our discussions with marketparticipants. Berlin and Mester (1992) also make this point,and the findings of Lummer and McConnell (1989) are consis-tent with it. The latter study showed that, in a sample of 357revised bank credit facilities from the period 1976–86, 259involved favorable revisions of terms.

37

This view of financial intermediation is ourcovenant–monitoring–renegotiation (CMR)paradigm. In the paradigm, information-intensivefinancial intermediaries serve information-problematic borrowers, not so much because theycan more efficiently produce information at theorigination stage but because they can efficientlyemploy covenants to control bondholder–stockholder conflicts. 89 In equilibrium, lendersentering into debt contracts that include covenantsmust be able to monitor efficiently, that is, mustefficiently produce information throughout the lifeof the contract. Lenders monitor a borrower’sperformance for two reasons: to determine whetherthe borrower is in compliance with covenants andto determine the proper action in the event of aviolation.90 A covenant violation may indicate thatthe firm is in distress or signal that a borrower istaking actions not in the lender’s interest. Cove-nant violations are a noisy signal about a borrow-er’s prospects, however, because they can be basedonly on observable, verifiable information. Todecide whether to liquidate a loan that is intechnical default, to renegotiate its terms, or towaive the covenant, a lender must produce newinformation (including information that may not beverifiable) about the borrower, quite apart fromsimply determining whether the firm is in compli-ance with its covenants. This type of informationproduction is often similar to that which occursduring loan origination.91

Berlin and Mester (1992) demonstrate theoreti-cally that the combination of tight covenants andthe option to renegotiate becomes more valuableas a borrower’s observable quality declines. Theintuition behind this result is straightforward. Forlow-quality firms, information-related problems aremore acute. Therefore, low-quality firms benefitthe most from the inclusion of restrictive cove-

nants in debt contracts because these covenantsprovide a mechanism for credibly committing toabstain from behavior that exploits the firm’slenders. However, restrictive covenants have ahigh probability of being binding in the future.Hence, the option to renegotiate is very valuable,and the reputation of lenders very important.

Covenants may be pareto-improving in any debtcontract because they can constrain borrowerbehavior. Covenants used in conjunction with adebt contract offered by a financial intermediarymay be especially potent, for three reasons. First,fixed costs of information production are keptdown. Second, renegotiations are most feasibleand least costly when the number of lenders issmall. Third, because a borrower is often at abargaining disadvantage in the event of a viola-tion, it will contract initially only with lenderswith a reputation for fair dealing in renegotiations.With their long-term presence in the creditmarkets, intermediaries are most able to build andmaintain such reputations. Tight covenants are notpresent in widely distributed debt because diffuseowners cannot efficiently produce information,renegotiate, or maintain reputations.

Private Placements in a Theoryof Credit Market Specialization

The CMR paradigm illuminates the differencesamong the commercial bank loan market, theprivate placement market, and the public bondmarket. Because their liabilities have short terms,banks prefer to invest in short-term assets. Such apreference naturally leads them to specialize in(among other things) lending to quite information-problematic, generally small firms. The optimalcontract for such borrowers has a short maturitybecause renewal can be based on nonverifiableinformation. It still includes tight covenantsbecause the borrowers are so problematic. Theseare frequently violated for reasons not associatedwith increases in expected losses or risk, and sobank loans tend to be renegotiated frequently.Quite problematic borrowers accept restrictiveterms because banks maintain a reputation for fairdealing and flexibility in renegotiation, because thecovenant constraints have short terms, and becausebank loans can typically be prepaid withoutpenalty. 92

89. Information production in the form of credit evaluationat the origination stage also occurs for traded debt but is notnecessarily performed by the investors in the securities.Investment bankers perform due diligence as part of theirresponsibility as underwriters; the results of their evaluation aredisclosed in the offering prospectus. Rating agencies alsoperform due diligence and reveal its results. Consequently, theCMR paradigm captures the distinguishing feature of interme-diated debt: the role of information production after debtfunding.

90. Debt contracts almost always include provisions requir-ing borrowers to report any violation of covenants, so monitor-ing for compliance is the less important of the two reasons.

91. Using covenants to trigger re-evaluations is bothcost-effective and legally necessary. Continuously conductingfull evaluations would be too costly for lenders. Also, anenforceable mechanism for putting a loan into technical defaultmust be based on information that is observable and verifiableby all parties.

92. See Berlin (1991) and Hart and Moore (1989) for aformal model of the maturity structure of loans and theverifiability of information.

38

Because their liabilities have long terms, lifeinsurance companies prefer to invest in long-termassets such as private placements, with fixedinterest rates and call protection. Since therenewal–refusal mechanism for controlling risk isabsent in such loans, life insurance companies relymore than banks on their ability to demandpayment based on covenant violations, that is, onverifiable events. However, covenants are also lesseffective as a risk-control mechanism in long-termdebt. Thus, in equilibrium, issuers of privateplacements tend to be less problematic, andcovenants in private placements tend to be looserthan in bank loans.93 As a result, private place-ment covenants are less frequently violated andrenegotiated. With less frequent renegotiation,borrowers are more willing to rely on a lender’sreputation for fair dealing, rather than on anability to prepay without penalty if renegotiationsgo sour. Since reputation is important, the equilib-rium can work only if private placements arefairly illiquid so that borrowers are assured ofcontinued dealings with good lenders. 94 Thus thepublic bond market is not well suited toinformation-intensive lending. Although renegotia-tion occurs less frequently than in bank loans, notuncommonly a private placement is renegotiatedseveral times during its life span. Life insurance

companies invest significant resources in monitor-ing capacity (although not so many as banks do).

Public market borrowers pose relatively fewinformation problems for lenders. Thus, publiclyissued bonds can have long terms, and a relativelyfew, loose covenants are adequate. Intensivemonitoring is unnecessary, and renegotiation isinfrequent. Given these characteristics, ownershipof public debt can be diffuse rather than concen-trated, and the contracts can be liquid.95

The CMR paradigm is not inconsistent with thetraditional view of market segmentation, whichfocuses on transactions costs and regulation inexplaining the institutional structure of creditmarkets. The traditional view is simply incom-plete. In a sense, the traditional view emphasizesthe liability side of bank and life insurancecompany balance sheets and largely ignores theasset side. The CMR paradigm focuses on theasset side. Consistent with the traditional view, theCMR paradigm indicates that long-term (short-term) loans appeal to life insurance companies(banks) because they match the maturity of theirliabilities. However, it emphasizes that in equilib-rium long-term and short-term lenders will tend toserve different classes of borrowers and to usesomewhat different risk-control technologies.

Other Empirical Evidence Relevant to theTheory of Credit Market Specialization

The CMR paradigm is consistent with empiricalevidence indicating that financial intermediariesact as specialists in information production. James(1987) found a positive stock-price response to theannouncement of bank credit agreements. Thisresult is consistent with the notion that banksproduce information about firm quality and revealthis information through their credit decisions(an approved bank credit agreement is a positivesignal to the market); it contrasts with the results

93. Of course, private placement borrowers typically obtaintheir short-term working capital from commercial banks. Theymay also have other short-term credit facilities with commer-cial banks.

As noted, there are differences between the bank debt andthe private placement contracts of private placement issuers(bank debt contracts have more restrictive maintenance cove-nants). Such differences may arise from specialization byintermediaries. However, a short-term callable bank loan is notcomparable to a long-term noncallable private placementbecause the bank loan can always be paid off and refundedwhereas a private placement locks in a borrower for a substan-tially longer time. Therefore, a private placement that has thesame covenants as a bank loan will be much more restrictive,in effect, than the bank loan because it is noncallable and has alonger maturity. The issue of simultaneously outstanding bankdebt and private placements notwithstanding, the principaldistinction we are drawing in the CMR paradigm is betweenthose borrowers that depend strictly on the bank loan market(and have no access to long-term debt in the private placementmarket) and those firms that have access to the private place-ment market. That is, we are principally comparing the bankdebt contract of bank-dependent borrowers with the privateplacement contracts of borrowers who are not bank dependent.

94. There are additional reasons that information-intensivedebt is illiquid. When selling such debt contracts, originatorsmust do so at a discount because buyers in the secondarymarket have to be compensated for their due diligence at thetime of purchase and such compensation cannot come fromfees charged to the borrower. Also, borrowers may be lesscooperative in assisting due diligence at resale than atorigination.

95. Berlin and Loeys (1988) demonstrate theoretically thatlower-quality firms (that is, firms with a higher probability ofdeteriorating) are likely to prefer an intermediated loan withtight covenants because the incremental value of hiring adelegated monitor to produce information about their truecondition is higher. Monitoring is inefficient, however, if debtof a lower-quality firm is publicly held because each bond-holder will have an inadequate incentive to monitor afterweighing the private gains from monitoring against benefits.That is, holders of public bonds do not enjoy the economies ofscale of information production available to a financial interme-diary. Consequently, publicly issued debt tends to be mostattractive to issuers of high quality and to firms about whichmuch information related to their financial condition is publiclyavailable.

39

of numerous studies documenting a negativestock-price reaction to the issuance of publicsecurities. 96 One study subsequent to James (1987)indicates that the positive stock price response isconfined to renewals (Lummer and McConnell,1989), but another finds an effect for both newand renewed loans (Billet, Flannery, and Garfinkel,1993). Wansley, Elayan, and Collins (1991) findthat the availability of other signals of firm qualityis important. All of these studies conclude that theuniqueness of bank loans stems from the ability ofbanks, as financial intermediaries, to produceinformation not otherwise available in the market.Bailey and Mullineaux (1989) and Szewczyk andVarma (1991) document a similar positive stock-price response to the announcement of a privateplacement arrangement, suggesting that lifeinsurance companies perform the same type ofinformation production that commercial banks do.

Also consistent with the CMR paradigm isevidence that banks may have an advantage overinsurance companies in the production of informa-tion about their borrowers. Besides helping toexplain banks’ preference for short-term lending,such evidence helps explain why banks lend to amore problematic group of borrowers. Nakamura(1993), for example, argues that banks have aspecial advantage over other financial intermediar-ies because they obtain information from borrow-ers’ checking accounts. This information isvaluable because patterns in checking accountactivity can signal changes in a firm’s quality.Udell (1986) and Allen, Saunders, and Udell(1991) show theoretically and empirically thatbanks can sort borrowers by manipulating theprices of their multiple services, including demanddeposits and loans. The more intensive informa-tion production by banks may also explain thecontradiction between results found by Bailey andMullineaux (1989) and Szewczyk and Varma(1991), which show a positive stock response toprivate placements, and other studies. James(1987) and Banning and James (1989) found anegative response, mostly associated with privateplacements that were used to repay bank debt.Vora (1991) found a positive response but only forunrated firms.97

The CMR paradigm is consistent with empiricalevidence on corporate restructuring and bank-ruptcy. Gilson, John, and Lang (1990) found thatthe probability that a firm would be restructuredprivately (versus entering formal bankruptcy) waspositively related to the ratio of private debt (bankloans plus private placements) to total debt. Theyalso found that stock returns (that is, cumulativeabnormal stock returns) were significantly higheron average for announcements of private restruc-turings (for which the returns were positive) thanfor bankruptcy (for which the returns werenegative). One explanation for these results is that,in a private restructuring, firms avoid the directand indirect costs associated with bankruptcy,which may total as much as 20 percent of firmvalue (see Warner, 1977, and Weiss, 1990, ondirect costs; and Altman, 1984, Cutler andSummers, 1988, and Lang and Stultz, 1991, forindirect costs). As noted earlier, one advantage tointermediated debt is that it facilitates renegotia-tion. Hence, lower-quality firms with a higherprob-ability of future distress value the renegotiationmechanism offered by financial intermediariesmore than do higher-quality firms.98 Other thingsbeing equal, such firms will thus prefer to issueprivate rather than public debt. Another explana-tion for the higher cumulative stock returnsassociated with private restructurings is thepossibility that relatively higher-quality firmssignal their value by choosing to restructureprivately.

Gilson, John, and Lang (1990) also examinedstock returns at the time that the market firstlearned that a firm was in financial distress. Theyfound that those firms subsequently enteringbankruptcy proceedings suffered negative cumula-tive returns on average when the market firstlearned of their financial distress, whereas thosefirms subsequently restructured privately sufferedno negative cumulative returns.

Taken together, the Gilson, John, and Langresults are generally consistent with the CMRparadigm. Financial intermediaries can useinformation produced through borrower monitoringin conjunction with restrictive covenants to beginnegotiations leading to a restructuring before afirm deteriorates beyond a point of no return. Thatis, financial intermediaries may be able to inter-vene at the earlier stages of firm distress becauseof three characteristics of intermediated debtcontracts: covenant restrictions, monitoring by

96. See Smith (1986) for a survey of this literature.97. Alternatively, the methodology employed in these

studies may be too weak to capture the empirical relationshipbetween stock returns and announcement effects in privateplacements. One problem may be identifying when informationabout a private placement is released to the market. The longtime involved in agenting a private placement may makeidentifying an appropriate event window difficult.

98. Lower-quality firms also value covenant restrictivenesswhen combined with renegotiation flexibility.

40

lenders, and the flexibility in renegotiation that isassociated with a limited number of lenders.Therefore, among those firms that suffer distress,those with intermediated debt are more likely torestructure privately. Firms without intermediateddebt, however, are likely to suffer more deteriora-tion before negotiations begin and are more likelyto enter bankruptcy. This finding is also consistentwith the results of Franks and Torous (1990), whofound that firms filing for bankruptcy are generallyin poorer condition than those restructuringprivately. In particular, bankrupt firms are lessliquid and less solvent than those that work outtheir debt in private restructurings.

Summary of Part 1

The arguments and evidence presented in part 1 ofthis study imply that, as shown in the following

diagram, firms can be placed on an informationcontinuum corresponding to their access todifferent debt markets. At one end of the contin-uum are small, new, extremely information-problematic firms that require a prohibitive amountof evaluation and monitoring and that have littleor no collateral to offer prospective lenders. Suchfirms must either use internally generated funds orobtain outside equity financing (perhaps fromventure capitalists). 99 Slightly less problematic,larger firms migrate to commercial financecompanies and commercial banks, which provideshort-term loans with tight covenants, intensive

99. Venture capitalists can be viewed as agents who, actingas insiders, produce information about the prospects of newfirms. They design tailored contracts that combine a highmeasure of control with a risky claim on the success of thefirm. See Chan (1983) or Chan, Siegal, and Thakor (1987) fora formal model of the role of venture capitalists in aninformation-theoretic setting.

Graphical Summary of Part 1

FIRM CONTINUUM

Firm size — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — —

Information availability — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — —

Very small firm, possiblywith no collateral andno track record

Small firms, possiblywith high growthpotential but oftenwith limited trackrecord.

Medium-sized firms.some track record.Collateral available,if necessary.

Large firms ofknown risk andtrack record.

Sources of Capital

Insider seed money

Commercial paper

Short-term commercial loans

Intermediate-term commercial loansMedium-

termnotes

Mezzaninefund

financing

Private placements

Public debt

Venture capital Public equity

41

monitoring, and the renegotiation option. Thesefirms tend to be risky and often borrow on asecured basis. 100 Somewhat larger firms may beable to obtain intermediate-term bank financing orsubordinated debt financing from mezzanine debtfunds or equity funds. Like bank loan officers,

mezzanine fund and equity managers intensivelymonitor their borrowers. They also controlinformation-related contracting problems partly byexercising some control through their share of theborrower’s equity. Somewhat stronger borrowersobtain bank credit on an unsecured basis fromcommercial banks. Even less information-problematic firms have access to the privateplacement market. These firms still have enoughinformation problems to require the services of anintermediary, but they are generally not so prob-lematic as commercial bank borrowers. Thus theycan issue long-term debt with looser covenantsthan those that exist in the bank loan market.Finally, firms that pose minimal informationproblems for lenders can issue in the public debtmarkets.

100. Several theoretical papers have shown that collateralmay be a powerful tool in solving information-related problemsassociated with debt contracting (see Chan and Kanatas, 1985,Chan and Thakor, 1987, and Besanko and Thakor, 1987a and1987b). Finance companies and commercial banks frequentlyrequire collateral as part of their loan contract (see Berger andUdell, 1990 and 1993b). Much evidence suggests that securedlending tends to be associated with riskier borrowers (seeBerger and Udell, 1990, 1993a, and 1993b, Boot, Thakor, andUdell, 1991, and Swary and Udell, 1988).

42

Part 2: Secondary Trading, the New Market for Rule 144A Private Placements,and the Role of Agents

In focusing on an economic analysis of thetraditional market for privately placed debt, part 1ignored two important features of that market: theeffects of Rule 144A and the role of agents.

Though resale of private placements is some-times thought to be prohibited, in fact a smallsecondary market for them has existed fordecades. Rule 144A, however, has created a newmarket for private placements. Adopted in April1990 by the SEC, this rule establishes conditionsunder which private placements may be freelytraded among certain classes of institutionalinvestors. The rule has spawned the developmentof a market for underwritten private placements,which has characteristics—such as not beinginformation intensive—more like those of thepublic bond market than like those of the tradi-tional private market. Part 2, section 1, analyzesthe Rule 144A market.101

The great majority of new private issues areassisted by an agent, which offers many of theadvisory and distribution services of a public bondunderwriter but does not actually perform afirm-commitment underwriting, except withunderwritten Rule 144A issues. Agents are at thenexus of many private market information flowsand thus play an important role. Section 2describes their role.

1. The Rule 144A Market

Rule 144A gave securities firms the opportunity tounderwrite private placements, allowing newissues of private debt to be distributed in much thesame way as issues in the public bond market.Securities firms have taken advantage of thisopportunity by providing public-like borrowers analternative to the public market and the traditionalprivate placement market. The 144A market thusbridges a gap between the two existing markets bymaking available to large corporations, not havingthe information problems of the typical issuer ofprivate debt, a more efficient means of placingdebt in the private market.

Although Rule 144A applies only to certainsecondary market transactions, it has implications

for the distribution of private placements. The rulepermits sophisticated financial institutions, desig-nated in the rule as qualified institutional buyers(QIBs), to trade private placements freely amongthemselves without jeopardizing the exemption ofthe securities from SEC registration. In any privateplacement transaction, whether in the primary orin the secondary market, the seller must ensurethat the sale does not constitute a public offering,which would violate the basis for exemption.Before the adoption of Rule 144A, securities firmsdid not underwrite private placements becausesales of securities to investors as part of anunderwritten distribution might be construed as apublic offering. Rule 144A, however, takes theview that QIBs are not part of the public; conse-quently, transactions between QIBs cannot involvea public distribution. Most securities firms areQIBs, and thus they can purchase private place-ments from issuers and resell them to other QIBswithout violating the private placement exemption.

The SEC justified this treatment of QIBs on thegrounds that the Congress had never consideredsophisticated, institutional investors to need theprotection offered by the registration of securities.The purpose of registration was to protect unso-phisticated, individual investors. The SEC there-fore concluded that, if secondary transactionsinvolved only sophisticated investors, such trans-actions would not constitute a public distributionand thus could be effected without restriction.102

The SEC had two basic purposes in adoptingRule 144A. One was to increase liquidity in theprivate placement market and thus to lower thedifferential between private and public yields. Theother was to make the private placement marketmore attractive to foreign issuers. Foreign compa-nies had been infrequent issuers in the publicmarkets, primarily because they found the registra-tion requirements expensive and burdensome,especially the stipulation that financial statementsbe reconciled with generally accepted accountingprinciples in the United States.103 Althoughforeign companies have long been able to bypassthese obstacles by issuing private placements, theyhad not done so to any great extent, partly because

101. Rule 144A applies to both debt and equity securities,but the discussion in this section focuses only on debtsecurities.

102. SEC (1988), pp. 97–102.103. Despite appearances, the burden of registration and

disclosure requirements may be less than many potentialforeign issuers have perceived it to be. See Engros (1992),pp. 5–9.

of the higher yields in the private market thanthose in the public market. The negotiation ofterms and frequent inclusion of restrictive cove-nants in private debt contracts also made theprivate placement market unattractive to foreigncompanies.

As defined in Rule 144A, QIBs are financialinstitutions, corporations, and partnerships thatown and invest on a discretionary basis at least$100 million in securities. 104 This definition isbroad enough to include life insurance companies,pension funds, investment companies, foreign anddomestic commercial banks, master and collectivebank trusts, and savings and loan associations.Besides meeting the securities test, banks andsavings and loans must have net worth of at least$25 million. The SEC imposed this conditionbecause it believed that securities holdings alonedid not necessarily reflect the appropriate degreeof investor sophistication for institutions havinginsured deposits. 105 In contrast to other institu-tional investors, a broker–dealer must own only$10 million in securities to qualify as a QIB. TheSEC chose a lower amount to avoid excluding asignificant number of broker–dealers that wereactively participating in the private placementmarket. 106

Besides confining transactions to QIBs,Rule 144A stipulates three other conditions. First,to ensure that a minimum amount of informationis available, an issuer must provide buyers withcopies of its recent financial statements and basicinformation about its business. Second, whenissued, privately placed securities cannot be of thesame class as any of the issuer’s securities alreadytraded on a U.S. stock exchange or on theNASDAQ system. This requirement is intended toprevent the development of an institutional marketin publicly traded securities. Third, the seller of144A securities must take ‘‘ reasonable’’ steps toinform the buyer that the sale is occurring pursu-ant to Rule 144A.107

Features of the Market

Although the SEC adopted Rule 144A only in1990, the 144A market has developed so that it iseasily distinguished from the traditional privateplacement market. In our view, the essentialfeature of the new market is that it is not informa-tion intensive, which is to say that it has taken onthe main features of the public bond market. Themost visible and discussed similarity to the publicmarket has been the underwriting of 144A offer-ings. Indeed, this aspect serves as the basis for ourdefinitions of a 144A security and the 144Amarket. 108

Nature and Size

Measuring the development of the underwritten144A market is especially difficult because manymarket participants, as well as the informationservices that collect data on the private placementmarket, consider a 144A security to be any privateplacement that relies upon the documentationrequired for a financing pursuant to Rule 144A.Unfortunately, this definition includes privateplacements that are, other than the documentation,no different from traditional private placements.Thus, relying upon these data, for which we haveno alternative, necessarily leads to an overstate-ment of the size of the underwritten 144A market.

Using the broad definition, gross issuance of144A securities has expanded rapidly since theinception of the 144A market in 1990. Thevolume of offerings in 1992 was about $33 billion,almost double that in 1991 (the first full year therule was in effect) and nearly two-thirds of thevolume in the traditional market (table 9).

The difficult question to answer is, How muchof the broad measure of 144A issuance has beenunderwritten? No direct estimates have been made,but an indirect estimate of underwritten issuancecan be obtained by assuming that issues with twoor more credit ratings have been underwritten.Underwritten offerings, whether in the publicmarket or the 144A market, typically have at least

104. Bank deposit notes and certificates of deposit, loanparticipations, repurchase agreements, and currency and interestrate swaps are excluded. When it adopted Rule 144A, the SECexcluded U.S. government and agency securities as well;amendments to the rule in October 1992 removed theexclusion.

105. SEC (1990a), pp. 17–20.106. SEC (1990a), p. 21.107. Further details on the provisions of Rule 144A, the

SEC’s reasons for adopting the rule, and its justification are inappendix A.

108. In this regard, some have argued that underwriting isnot a meaningful distinction because most underwrittensecurities have been sold before the formal offering. Althoughthis situation may be true, our view is that underwriting ischaracteristic of a non-information-intensive market, which inturn is the critical feature of the 144A market. The focus onunderwriting is partly a matter of convenience, but it alsocoincides with a view held by many market participants thatunderwriting is the distinctive feature of the 144A market.

44

two ratings because the underwriters otherwiseincur significantly higher regulatory capitalcharges. Available information from the SECshows $4.4 billion of 144A issues with at leasttwo ratings in 1991 and $6.0 billion in the firsteleven months of 1992.109 These figures areroughly in line with market estimates, which placeunderwritten issuance in 1991 at slightly morethan $3 billion and in the first half of 1992 atroughly double that pace.110 Even the largerfigures from the SEC suggest that the underwrittenmarket is still in an early stage of development.

Characteristics of Underwritten 144ASecurities

Besides being underwritten, 144A securities haveassumed many other features of publicly offeredbonds. The terms and documents generallyconform to the standards used in the publicmarket; in particular, bonds have ‘‘ public style’’covenants, which are fewer and considerably lessrestrictive than those found in traditional privateplacements. Underwriters charge roughly the samefees as those for a public offering, but the issueravoids the considerable expenses associated withpublic registration. The underwritten 144A securi-ties also generally have two credit ratings; and,in many instances, the offering memorandum isstyled like a prospectus in a public offering. Also,

144A offerings are usually transferred through thebook-entry system operated by the DepositoryTrust Company. All of these features are a part ofunderwriters’ efforts to market 144A privateplacements to traditional public market investors,such as mutual funds, pension funds, and groupswithin life insurance companies responsible forpublic bond investments. 111 Furthermore, under-written private placements have been comparablein size more to public offerings than to traditionalprivate placements: In 1991, for example, theaverage issue for 144A securities, broadly defined,was $92 million, nearly double that for non-144Aplacements. Finally, the terms of the securities arerarely negotiated with investors but are typicallyset before the offering.

Despite this similarity to public bonds, under-written 144A securities generally have not yetachieved the same degree of liquidity as publicbonds, and thus their yields contain a premium.112

In the first year of the market, the premium wasreported to be about the same as that on traditionalprivate placements. More recent reports suggest,however, that the liquidity of 144A securities hasincreased and that the premium has decreased, asmajor dealers have allocated capital and traders tomaking markets for 144A securities. 113

Foreign Issuers

Thus far, the proportion of foreign issuance hasbeen greater in the 144A market than in either thetraditional private or the public bond market.Based upon the broad measure of 144A issuance,approximately one-third of the total volume of144A offerings in 1991 and 1992 was accountedfor by foreign issuers, including U.S. subsidiariesof foreign companies. In contrast, 17 percent ofthe traditional private placements and 6 percent ofthe public offerings were by foreign issuers.

Several factors lie behind foreign use of the144A market. One is that the adoption ofRule 144A itself served to publicize the alreadyexisting advantages of the private placementmarket to foreign companies. Thus, the effect ofthe rule has been to alter foreigners’ perceptionthat all offerings in the United States are subject

109. SEC (1993), appendix A. The report does not cover allthe 144A issues used to compute the totals in table 9. Thereport examined issues totaling $7.6 billion in 1991 and$8.0 billion in the first eleven months of 1992.

110. See Investment Dealers’ Digest (1992), pp. 13–14, andKeefe (1992), pp. 1 and 10.

111. In this regard, a major underwriter noted that 70–95percent of 144A placements during the first half of 1992 hadbeen sold to public investors. See Vachon (1992b), pp. 23–24.

112. An additional reason for the premium is that investorstypically demand a slightly higher rate from foreign issuers andfrom first-time issuers.

113. Keefe (1992), p. 10.

9. Gross issuance of public and private debtsecurities in U.S. markets, 1989–92Billions of dollars

Issuance 1989 1990 1991 1992

Rule 144A privateplacements . . . . . . . . . . . . 2.2 16.7 33.3

By foreign issuers . . . . . . . .4 5.5 10.5

Non-Rule 144A privateplacements . . . . . . . . . 134.8 101.0 75.8 52.4

By foreign issuers . . . . 20.3 15.8 12.5 9.4

Public bonds . . . . . . . . . . . . 188.9 203.6 307.1 401.8By foreign issuers . . . . 9.2 14.8 20.2 24.1

Source. IDD Information Services and Securities DataCorporation.

45

to excessive regulatory burdens. Indeed, marketparticipants concede that some of the foreignissuance done under Rule 144A could have beenas easily accomplished before the rule’s adop-tion.114 Moreover, since the rule’s adoption,investment banks have devoted greater effort tobringing foreign issuers to the private placementmarket. A second factor boosting foreign issuancehas been the low interest rates in the United Statesrelative to those in European countries. Theincrease in 1991 in foreign issuance in the publicbond market and the record pace of offerings in1992 attest to the yield advantage in U.S. markets.A final factor is that the premium in yields onforeign bonds issued in the private placementmarket has declined.

Among other aspects of foreign issuance in the144A market, many foreign issuers have beenwell-known corporations, but at the same time,about 20 percent of the issues have come fromfirst-time borrowers in the United States. 115 Themajor sources of issuance from abroad have beenthe United Kingdom and Mexico. ThroughNovember 1992, more than half of the foreignissues studied by the SEC were involved in globalofferings, and virtually all the global offeringsoriginated with an offshore entity. In contrast,about half of those foreign-related offeringsconfined solely to the 144A market involved U.S.subsidiaries of foreign corporations. 116 About50 percent of the volume of foreign 144A securi-ties in 1991–92 came from financial institutions,and most of that was in medium-term notes.

Domestic Issuers

Despite the attention given to foreign use of the144A market, U.S. companies have accounted fornearly 70 percent of the volume through 1992.Domestic issuers in the 144A market havetypically been those companies with specialcircumstances that preclude issuing in the publicbond market, where yields are lower. In somecases, the companies have not wanted to spend thetime nor incur the expense required to register thesecurities with the SEC. Among these have beenprivate companies that, in the past, have borrowedin the traditional market but have now found morefavorable pricing in the 144A market. Alsoincluded are nonregistered subsidiaries of publicly

registered parents that have issued debt in thesubsidiaries’ names. In other cases, companieswith outstanding public securities have turned tothe underwritten 144A market to protect theconfidentiality of the specific circumstancesleading to the borrowing.

Another group of domestic companies has usedthe 144A market as a temporary alternative to thepublic bond market. These companies normallyissue public securities but have turned to the144A market to avoid any delays arising duringthe registration process that could cause issuers tomiss favorable financing opportunities. The144A private placements sold under these circum-stances have included registration rights, whichobligate the issuer to register the bonds with theSEC within a specified time. Failure to do soresults in the bonds’ carrying higher coupon rates.Most companies selling these types of 144A secu-rities have been rated below investment grade.

Investors

During the first two years after the adoption ofRule 144A, life insurance companies were thelargest group of investors in 144A securities. Asthe 144A market has developed features of thepublic bond market, however, the compositionof investors has shifted toward those, such asmutual funds and pension funds, that generallyconcentrate investments in public securities.Information on buyers of 144A securities from asample of new issues studied by the SEC impliesthat the share of life insurers’ purchases of straightdebt fell from roughly 75 percent betweenApril 1990 and August 1991 to 60 percentbetween September 1991 and April 1992(SEC, 1993). Over the same two periods, thecombined share of mutual funds and pension fundsrose from a little over 10 percent to nearly40 percent. Market participants indicate that thecomposition of buyers has continued to shifttoward mutual funds and pension funds and, inaddition, that many life insurance companies haveshifted responsibility for investing in 144A securi-ties from their private placement groups to theirpublic market groups. Thus, the dominance of thelife insurance companies in the later period of theSEC study likely understates the growing signifi-cance of public market investors in the 144Amarket.

Public market investors are attracted to the144A market because its public-like features suittheir investment style. In contrast to the buy-and-

114. Engros (1992), p. 7.115. Private Placement Reporter (1992a), p. 10.116. SEC (1993), appendix A.

46

hold strategy of investors in traditional privateplacements, many public market investors followa total-return strategy in which they attempt toincrease the return beyond the security’s couponrate of interest. To do so, these investors look forundervalued securities offering the potential forcapital gains. 117 Such investors require liquidity,because they do not expect to hold the securitiesto maturity. From this perspective, public marketinvestors have found the liquidity in the 144Amarket to be sufficient.

In contrast to the move of public marketinvestors to the 144A market, buyers of traditionalprivate placements are unlikely over time to findthis market attractive. The comparative advantageof traditional market investors is in credit analysisand credit monitoring, neither of which is requiredextensively in the 144A and public markets. And,in the buy-and-hold strategy of traditional inves-tors, liquidity is of little importance.

Prospects for Development

Because it has filled a gap in U.S. capital markets,the underwritten 144A market appears likely toundergo further development and growth. Beforethe adoption of Rule 144A, no market existed thatcould accommodate large issues that were unsuitedfor the public market but did not require aninformation-intensive market. Issuers of thisnature, whether domestic or foreign, had no choicein U.S. markets but to accept the terms of theprivate market. Although such issuers often didnot have to tolerate restrictive covenants, they hadto pay a premium over public bond rates becauseof the lack of liquidity in the private placementmarket. By increasing liquidity, Rule 144A hasreduced the premium and has thus increasedofferings by such issuers.

In being both non-information-intensive andprivate, the 144A market represents a new bondmarket. Whether the need for such a marketextends much beyond current levels of activity isan open question. The midsized, information-problematic firms, which issue in the traditional

market, will probably not move to the 144A mar-ket. They must borrow from a financial intermedi-ary and often do not want their issues to be tradedin a liquid market to investors that might notunderstand their particular circumstances.

Perhaps, the greatest potential for the144A market lies in its use by foreign issuers,inasmuch as they represent the largest group ofborrowers with no previous satisfactory alternativein the United States. If foreign issuance expandssignificantly, Rule 144A may prove helpful inintegrating world capital markets. Borrowing bylarge, domestic corporations with specializedrequirements seems to offer much less potential,as such borrowing constitutes a small share of thecredit needs of large corporations. If, however, theliquidity of the 144A market increases so thatyields in the public and 144A markets are roughlythe same, a considerable portion of public marketborrowing may shift to the 144A market, whichwould offer lower borrowing costs overall becauseof the absence of registration costs.

2. The Role of Agents

Almost all new public issues of bonds aremanaged by an underwriter on the basis of a firmcommitment. New issues of private placements,however, are often assisted by an agent oradviser. 118 Agents provide various services toissuers, including advice about the structure,pricing, and timing of financings; assistance inlocating investors; and help in negotiating withthem. Agents assist traditional private issues ona best-efforts basis, but many Rule 144A trans-actions are firm-commitment underwritings.Although no quantitative evidence is available,remarks by market participants indicate that anagent assists in about two-thirds of traditionalprivate issues; the rest of these issues involvedirect contacts between issuers and investors.Apparently, although lenders and borrowers in theprivate placement market might be able to findeach other and write contracts by themselves,such a process would be costly; in many cases,employment of a third-party agent is moreefficient.

The role of agents in the private placementmarket is somewhat more complicated than the

117. One element of this strategy is identifying companieslikely to undergo a credit-rating upgrade. Credit analysis isused for this purpose but is not essential for ensuring thelong-run value of the security, as in investing in traditionalprivate placements. Consequently, public market investorsperform much less extensive credit analysis and monitoringthan investors in traditional private placements. Public marketinvestors also tend to rely more upon the research of invest-ment banks and other outside credit analysts.

118. Technically, an agent has the power to commit theissuer, whereas an adviser does not. We use the word agent torefer to both.

47

previous paragraph may imply. Like the privatemarket itself, the agent industry exists primarily tosolve problems associated with costly and asym-metric information. Agents add value in severalways:

• They reduce search costs for both borrow-ers and lenders by maintaining informationabout lenders’ preferences and by screening outunqualified borrowers. 119

• They have knowledge of prevailing marketprices and the tradeoff rates between prices andother contract terms. Borrowers need suchinformation for both search and negotiation,and buying it from an agent is often cheaperthan gathering it.

• They provide technical advice and otherassistance to borrowers during negotiations,helping them obtain better terms.

• They enforce informal bargaining conven-tions that reduce bargaining costs for everyone.

The private market is thus broader and deeper thanit would be without agents: More borrowers areserved, and more competition exists amonglenders.

The structure of the agent industry is influencedby economies of scale and scope, by limitedstrategic relationships between agents and lenders,and to some extent by specialization. The primaryeconomy of scope is with the provision of othercorporate financial services: Agents tend toflourish in those large commercial banks andinvestment banks that sell a large volume andvariety of corporate finance products, such asloans or underwritings. The relationship officersof such banks can refer significant numbers ofpotential clients to the private placement agentswithin the organization. Economies of scope alsoexist with public-issue underwriting, in that salesforces for public securities can distribute someprivate placements.

The primary economy of scale is related to thecosts of gathering information. These costs aresmaller for high-volume agents for two reasons.First, the fixed costs of gathering information canbe spread over many clients. Second, an agentacquires information as a byproduct of assistingindividual transactions, both reducing the amountof information it must gather by other means andproviding more to trade in the information

marketplace. Agents and lenders gather informa-tion through informal sharing arrangements witheach other, and high-volume participants are moresought after as partners in such arrangements.

Economies of scale and scope influence anagent’s style of providing services as well as thedegree of concentration of the industry. Althoughmost agents are in large measure generalists, theyhave some variety in the technologies they canchoose when conducting their business, especiallywith regard to the distribution of securities. Theyalso tend to specialize somewhat in the technolo-gies best suited to the kinds of client their hostorganization’s relationship officers tend to refer. 120

Although large agents may have advantages,competition appears substantial because entry andexit costs are relatively low and the roster ofagents is constantly changing.

Who Are the Agents?

According to a database supplied by the publishersof the Investment Dealers Digest, thirty investmentbanks and commercial banks were responsible for96 percent of the volume of all agented privatelyplaced debt transactions from 1989 through 1991(see table 10). Each of these agents placed at least$1 billion of debt securities during at least one ofthose three years. The database, however, does notinclude all new private issues. Possibly, a tablebased on a complete list of transactions wouldchange the ranking somewhat and would addentries to the list. 121

The Stages of a Private PlacementTransaction

This subsection describes the role of the agent ateach stage of private placement issuance, empha-sizing the ways in which agents add economicvalue to the transaction. Readers not already

119. For example, some lenders may offer better terms thanothers to borrowers in a particular industry, perhaps becausethey have particular expertise in lending to that industry.

120. For example, some agents assist mainly large place-ments (say, more than $100 million in face value), some servemainly investment-grade borrowers, and others serve mainlybelow-investment-grade borrowers. Some agents may get adisproportionate share of a given industry’s business.

121. The total number of agents of private placements isunknown because many banks, investment banks, and ‘‘ bou-tiques’’ act as agents for relatively small volumes of issuance.The IDD data files for 1989–91 list 173 organizations asagents, many of them for only one or a few transactions in asingle year. Many agents with a relatively small volume ofbusiness do not report their transactions to IDD. Marketparticipants’ off-the-cuff estimates of the total number ofcurrently active agents range from 100 to 300.

48

familiar with the details of private issuance mayfind the description of a sample private placementtransaction that appears in appendix F helpful atthis point. The example provides a sense of theflow of the process that may be useful backgroundfor the analysis in this section.

As shown in the following diagram, a dealpasses through five major stages. During theprospecting stage, agents identify potential issuersand compete with each other to gain the issuer’sbusiness. Issuers decide whether to place a privateissue or to use another vehicle for financing andwhether to hire an agent or to issue withoutassistance.

During the contract design stage, and sometimesduring prospecting, agents analyze in detail anissuer’s condition, operations, and plans (duediligence) and use this information to set major

debt contract terms. They summarize the terms ona term sheet and write an offering memorandumdescribing the issuer, which is somewhat similar toa prospectus. The memorandum and term sheet areoften packaged together and called ‘‘ the book.’’ Ifnecessary, agents seek a rating of the issue. Theythen choose an initial strategy for distribution and,in some cases, carry out preliminary inquiries ofinvestors.

During the distribution stage, which is coinci-dent with the design stage for many deals, theagent seeks investors. Negotiations that change theterm sheet often occur. In some cases, the agentfirst seeks a lead lender (traditionally, the investorthat buys the largest fraction of the placement) andconducts most negotiations with it; only after thelead has committed to the deal does the agentattempt a broader distribution. In other cases,

10. Major agents of U.S. private placements of debt, 1989–91

Agent

Rank according to business volume 1 Three-year agent volume

Privateplacementagenting

Investmentbanking

Commercialbanking

Amount(millions of

dollars)Percent of

total

Goldman Sachs . . . . . . . . . . . . . . . . . . . . . . 1 2 . . . 37,469 11.7First Boston . . . . . . . . . . . . . . . . . . . . . . . . . . 2 4 . . . 32,143 10.0Salomon Brothers . . . . . . . . . . . . . . . . . . . . 3 7 . . . 25,811 8.1J.P. Morgan . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 12 8 22,075 6.9Merrill Lynch . . . . . . . . . . . . . . . . . . . . . . . . 5 1 . . . 19,574 6.1Lehman Brothers . . . . . . . . . . . . . . . . . . . . . 6 3 . . . 16,635 5.2

Citicorp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 ( 2) 1 14,485 4.5Chase Manhattan . . . . . . . . . . . . . . . . . . . . . 8 . . . 5 13,264 4.1Morgan Stanley . . . . . . . . . . . . . . . . . . . . . . 9 6 . . . 12,908 4.0Drexel Burnham . . . . . . . . . . . . . . . . . . . . . 10 n.a. . . . 12,246 3.8PaineWebber . . . . . . . . . . . . . . . . . . . . . . . . . 11 11 . . . 11,726 3.7FNB Chicago . . . . . . . . . . . . . . . . . . . . . . . . 12 . . . 10 11,009 3.4

Chemical Bank . . . . . . . . . . . . . . . . . . . . . . . 13 . . . 2 3 10,708 3.3Bankers Trust . . . . . . . . . . . . . . . . . . . . . . . . 14 . . . 20 10,167 3.2Kidder Peabody . . . . . . . . . . . . . . . . . . . . . . 15 5 . . . 8,387 2.6Continental Bank . . . . . . . . . . . . . . . . . . . . 16 . . . 13 6,460 2.0Bank of America . . . . . . . . . . . . . . . . . . . . . 17 . . . 3 6,164 1.9Manufacturers Hanover . . . . . . . . . . . . . . 18 . . . 2 3 5,740 1.8

Donaldson Lufkin . . . . . . . . . . . . . . . . . . . . 19 10 . . . 4,022 1.3NationsBank/NCNB . . . . . . . . . . . . . . . . . 20 . . . 4 3,714 1.2Dillon Read . . . . . . . . . . . . . . . . . . . . . . . . . . 21 ( 2) . . . 3,517 1.1Bear Stearns . . . . . . . . . . . . . . . . . . . . . . . . . . 22 8 . . . 3,125 1.0Smith Barney . . . . . . . . . . . . . . . . . . . . . . . . 23 14 . . . 2,997 .9Capstar Partners . . . . . . . . . . . . . . . . . . . . . . 24 . . . 2,634 .8

Prudential-Bache . . . . . . . . . . . . . . . . . . . . . 25 9 . . . 2,566 .8Dean Witter . . . . . . . . . . . . . . . . . . . . . . . . . . 26 15 . . . 1,929 .6Wertheim Schroder . . . . . . . . . . . . . . . . . . 27 . . . . . . 1,802 .6Lazard Freres . . . . . . . . . . . . . . . . . . . . . . . . 28 . . . . . . 1,184 .4Barclays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 . . . . . . 1,159 .4Alex Brown . . . . . . . . . . . . . . . . . . . . . . . . . . 30 13 . . . 1,091 .3

1. Investment and commercial banking ranks were deter-mined using underwriting and loan volumes respectively.

2. Dillon Read and Citibank were the eleventh andtwelfth ranked investment banks for investment grade debtrespectively.

3. End-of-1991 consolidated loans were combined forChemical and Manufacturers Hanover in arriving at a bankranking.

Sources. Noted in text.

49

the agent attempts a broad distribution from thebeginning. An initial commitment by a lender isknown as ‘‘ circling’’ the deal. Such a commitmentis contingent on approval by the lender’s invest-ment committee and on due diligence by thelender that produces satisfactory verification of theinformation in the offering memorandum. Negotia-tions about price are conducted in terms of spreadsover Treasuries of comparable average life until adeal is fully subscribed, at which time couponrates are set. 122 If necessary to attract additionalinvestors, the coupon rate may be increased afterit has been set, but it may not be reduced even ifTreasury rates fall between rate-setting andclosing. Similarly, if Treasury rates rise, bytradition the lenders may not demand a highercoupon.

The contract design and distribution stagestypically require one to two months. The process

of obtaining a rating is the most important sourceof delays. 123

The penultimate stage, due diligence by lenders,begins when a deal is fully subscribed. Beforecircling, lenders carry out a significant amount ofcredit analysis, which often involves gatheringsome information not found in the offeringmemorandum. During the due diligence stage,lenders verify the information in the offeringmemorandum and, if satisfied, present the deal toinvestment committees for approval. Rarely doinvestment committees reject a deal for anythingbut unsatisfactory due diligence. Rejection aftercircling imposes large costs on other members ofthe lending syndicate and on agents and borrow-ers. Agents are less likely to bring deals to alender with a history of such behavior, and otherlenders are less willing to join it in syndicates.Rejections thus in the long run affect a lender’sability to invest in private placements on favorableterms.

122. As is described further below, initial term sheets varygreatly in the extent of their detail. Most commonly, a termsheet will initially include suggestions regarding covenants butno spread. Interested investors respond to an initial offer byreturning the sheet with acceptable covenants circled, modifica-tions noted, a spread they will accept, and the volume they willbuy at that spread given that their modifications to other termsare included.

123. Given life insurance companies’ recent aversion tobelow-investment-grade placements, delays associated with therating process are especially likely for potential issuers nearthe borderline between an investment-grade and a below-investment-grade rating.

Stages of private placement issuance

Stage Activities and events

Prospecting Client (issuer) identification

Competition among agents

Due diligence by agents

Design of major contract terms Writing of offer memo and term sheet

Pre-rating by NAIC or rating agency

Decision about strategy of distribution

Distribution Solicitation of investors

Circling by investors

Due diligence by lenders Due diligence by lenders

Investment committee approvals

Formal letters of commitment

Contract writing Negotiations on exact language and terms

Closing

50

In the final stage of private issuance, lawyershammer out the language of the debt contract,which involves several documents besides thenotes themselves (see table 11 for the majordocuments). 124 The lenders are represented by abond counsel, which is by tradition chosen by thelead lender but paid by the borrower. The bor-rower is often represented by its own counsel andis usually assisted by the agent. Transactions canunravel at this point when interpretations of termsheets differ, but such unraveling is relatively rare.Although it varies, the time required for the finalstage is usually a few weeks. Once all parties signthe contract (closing), funds can be disbursed tothe borrower.

The remainder of this subsection describes andanalyzes each of the stages in more detail.

Prospecting, Initial Advice, and Inter-AgentCompetition

Commercial banks and investment banks obtainmost of their private placement clients through

contacts initiated by relationship officers, who aretraditional bank loan officers, investment bankersresponsible primarily for maintaining relationshipswith clients, and hybrids of the two. Relationshipofficers call on current or prospective clients oftheir organization, attempt to learn about the broadspectrum of client needs for capital and financialservices, and in the process often help clients torecognize opportunities and incipient problems.These officers are also able to identify opportuni-ties to sell specific products.

Relationship officers consult their privateplacement group when they recognize that aprivate placement may be an appropriate way fora client to raise funds. When several differentborrowing strategies might serve a client’sinterests, some organizations arrange presentationsto the client by different groups within theorganization, for example, the private placementgroup and the loan syndication group.

The prospecting process sometimes departs fromthis description at some commercial banks wheremost customer contact is by traditional loanofficers and where the loan officers’ compensationis determined by success in originating loans. Thistype of compensation scheme may deter loanofficers from recommending a private placementover a commercial loan. According to marketparticipants, commercial banks are losing thisweakness as they change their organizationalstructures and compensation schemes.

Agents may also obtain clients through requestsby previous private placement clients for help withnew transactions. Such requests are sometimesmade directly to the agent group, as the clientalready knows them. Direct requests are alsoreceived from potential issuers who want competi-tive bids from different agents. Relatively fewagenting jobs for first-time clients result fromprospecting by the private placement group itself.

Agents compete for the right to assist particularprivate placements, with the degree of competitiondepending both on expected profits and on theextent to which a borrower seeks multiple bids.Some agents specialize in particular types oftransactions, and thus their explicit costs andopportunity costs differ across transaction types,so a given borrower can be quoted a variety offees. Competition exists also along dimensionsother than fees, as borrowers must estimate boththe likelihood that a given agent can successfullydistribute the securities and the interest rate andother loan terms that the agent can obtain. Bor-rowers do not typically possess the informationrequired to make such estimates with precision,124. See Engros (1992) for a complete list of documents.

11. Major documents in private placementissuance

Document Purpose

Offering memorandum Describes the issuer. Similar to aprospectus, but the information itmay contain is not restricted.

Term sheet Lists terms of debt contract. Initially,the rate is often not included. Thisdocument is the focus of initialnegotiations. Often bundled with theoffering memorandum in a ‘‘ book.’’

Securities purchaseagreement

Details the representations, warran-ties, covenants, and other provisionsestablishing the legal relationshipbetween the borrower and lender. Asecurities purchase agreement isentered into with each investor.

Securities The notes or other instruments ofindebtedness.

Placement agentagreement

Specifies the obligations of theissuer and the agent. May limit theactions the agent can take, forexample, may rule out solicitation ofcertain classes of investor, such asindividuals.

Closing opinionsand miscellaneousclosing documents

A variety of of documents settingout opinions of counsel and stipula-tions by the issuer are often requiredat closing.

Source. Engros (1992)

51

so they must rely, at least to some extent, onreputations and on the claims made by agents insales presentations. Agents from an organizationwith which a borrower has a satisfactory, ongoingrelationship thus have a significant advantage incompeting for that borrower’s private placementbusiness. 125

Value Added. A considerable amount of economicvalue is added by agents during the prospecting,advice, and competition stage of a transaction.Some borrowers know little or nothing about theprivate market and may not consider it as a sourceof funds unless it is suggested by a relationshipofficer. Even if they are somewhat informed,borrowers will usually not commit to bear theopportunity costs associated with a private marketoffering without first comparing the opportunitiesthere with those in other markets. Such a compari-son can be done only with reasonably current andcomplete information about the operation of theprivate market and the terms available there. Thecosts of gathering such information are muchhigher for the private placement market than forthe bank loan and public debt markets, especiallyif the borrower has never issued a private place-ment. Either directly or through their organiza-tion’s relationship officers, agents provide suchinformation to potential borrowers as part of theirmarketing efforts and thus improve the efficiencyof financial markets.

Economies of Scale and Scope. Although avail-able data do not support precise measurement, theremarks of market participants imply that econo-mies of scale and scope at the prospecting, advice,and inter-agent competition stage of transactionsstrongly influence the structure of the market foragent services. An agent organization need not belarge, but it must bear the staff and overhead costsof near-continuous gathering of information aboutprivate market conditions and of maintainingrelations with lenders. Thus, the number ofrelationship officers calling on clients likely toissue private placements must be sufficient to yieldclients paying fees that at least cover costs.Although the organization as a whole is notabsolutely required to be large, commercial banksand investment banks that serve many corporate

clients of medium to large size are more likely toprovide a large flow of private placement pros-pects to their agent groups. Such organizations canthus spread the overhead costs of informationgathering over a broader base of revenues. Inother words, scope economies may exist betweenagenting and providing other financial services tomedium and large corporations. Commercial banksthat focus mainly on small business lending,mortgage loans, or consumer lending will havedifficulty making a profit on private placementagenting.

Indirect evidence of economies of scope can beseen in the rankings of the thirty major agentsaccording to their volume of commercial bankingand investment banking business (table 10). Bankholding companies were ranked by the totalconsolidated volume of commercial and industrialloans on their books at the end of 1991.126

Investment banks were ranked according to thetotal volume of domestic securities issues of allkinds for which they acted as lead manager. 127

As with the ranking of agents, we claim not thatthe order of rankings is entirely accurate orimportant but only that a significant rankingindicates a large volume of activity in the capitalmarkets.

The top twenty-six agents rank among the toptwenty commercial banks or the top fifteeninvestment banks, or both. All of the top fifteeninvestment banks are major agents, as are all ofthe top five commercial banks. Fifteen of the toptwenty commercial banks reportedly acted as agentat least once. This predominance of large commer-cial and investment banks in the agenting industryis consistent with the existence of significanteconomies of scale and scope in agenting.128

The economies of scale and scope realized atthe prospecting, advice, and competition stageinfluence an agent’s strategy and specialization.An agent within a commercial or investment bankthat serves mainly Fortune 500 and large interna-tional corporations will naturally find most of itsclients coming from those groups. As is discussedfurther below, design and distribution of the

125. Occasionally a private placement will involve morethan one agent. Sometimes a small agent with a client wantinga relatively complicated placement will bring in another agenthaving the necessary expertise. Sometimes a client will ask thattwo or more agents work together.

126. Commercial and industrial (C&I) loan volume waschosen as a ranking criterion because, among all groups ofbank clients, C&I loan customers appear most likely to issueprivate placements. Data for the rankings were drawn from theDecember 31, 1991, Y-9 reporting form filed by bank holdingcompanies.

127. Rankings were taken from reports in CorporateFinancing Week and the Investment Dealers Digest.

128. The top twenty-six agents advised 94 percent of thevolume of transactions recorded in the IDD database.

52

private issues of such borrowers is typicallydifferent from that for middle-market borrowers,and it is efficient for the agent to gather somewhatdifferent information and to maintain somewhatdifferent relationships with lenders than an agentspecializing in serving middle-market borrowers.

Design of Major Contract Termsand Distribution of Securities

Having won an issuer’s business, an agent beginsdesigning and perhaps distributing the securi-ties. 129 Design involves setting the terms of thesecurities, including payment amounts, timing, andcovenants. Distribution involves finding lendersthat will buy the securities. In contrast to thephases of public issuance, the line between thedesign and distribution phases is blurred and insome cases does not exist because design of theterms of privately placed securities often involvesnegotiations between lenders and borrowers. Thenegotiations may be implicit or explicit and maytake place either before or during the period whenthe securities are offered to lenders. The natureand the timing of the negotiations depend to alarge extent on the style of distribution chosen bythe agent, which in turn depends on the identityof the agent, the characteristics of the borrowerand the loan, and market conditions.

At one extreme, the process can resemble abest-efforts public underwriting. Here the agentuses its knowledge of market conditions andlenders’ preferences to design terms that are likelyto satisfy lenders, including an interest rate spread.The securities are then offered to many potentialinvestors on a take-it-or-leave-it basis. If the issuecannot be fully sold, the interest rate may beincreased or other terms may be changed. There isoften no lead lender in the usual sense, althoughone lender may be designated as lead.

At the other extreme, the agent may contact oneor a few potential lenders immediately uponreceiving a mandate from the issuer and informthem of the identity of the borrower and the likelyamount of the loan. Reactions of the lenders andensuing negotiations influence the terms of thesecurities. By the time the term sheet is finalized,distribution may be pro forma because all oralmost all of the lenders may have made informal

commitments. Any unsold portion is madeavailable to investors at large, although they haveno opportunity to negotiate the terms.

Between these extremes is a continuum ofstyles. One part of the design phase, however,does not vary much across styles: due diligence.

Due Diligence. Agents of traditional privateplacements do not bear the market price risksassociated with public underwriting, as non-underwritten placements never appear on agents’books. Agents are nevertheless at risk, in threeways. First, they are paid only for successfulplacements, and thus their investment in a particu-lar transaction of staff time and other resources isat risk until closing. Deals can unravel for manyreasons; one is a lender’s discovery after circlingbut before formal commitment that the offeringmemorandum misrepresented the borrower’scircumstances.

Second, the agent’s reputation with lenders is atrisk. Lenders also invest time and resources inevaluating potential loans, and the semiformal loancommitment that circling a deal represents isbased mainly on the information in the offeringmemorandum and term sheet. If in performing itsown due diligence a lender finds an offer memo tobe materially incomplete or inaccurate, it will beless likely in the future to expend resources inconsidering transactions proposed by that agent.Also, if an agent is associated with too manyplacements that later decline in credit quality orgo into default, lenders will be less likely to dealwith that agent.

Third, private placement agents have beennamed as parties in some lender-liability lawsuits.Agents must thus take the potential costs associ-ated with such suits into account when estimatingthe profits from assisting a transaction.

Agents control these risks by conducting a closeexamination of a borrower’s business, financialposition, and plans. They perform this duediligence immediately after they receive a mandateto assist a borrower’s placement and, to someextent, before that. This examination resembles thedue diligence performed by lenders and usuallyincludes a visit to the borrower’s headquarters orother relevant sites. Besides controlling risks, theexamination provides the agent with informationneeded to write the offer memo and term sheet.

Some commercial banks and investment banksare sufficiently concerned about these risks thatprivate placement agenting jobs must be approvedby a credit committee. Some market participants

129. Winning an issuer’s business is known as getting amandate; it involves a contract between the issuer and agentknown as a placement agent agreement.

53

stated that their committees reject a substantialfraction of agenting jobs.

Value Added from Due Diligence by Agents.Two ways in which agents add value are bypre-screening borrowers and by gathering informa-tion needed by potential lenders. Each of the largeprivate market lenders is offered hundreds ofplacements in a typical year and refuses all but asmall fraction.130 At the typical large lender, aninitial evaluation occurs when the agent offers thetransaction. This evaluation is based mainly uponinformation in the offering memorandum and termsheet. Some proposed transactions can be quicklyrejected, because they fail to meet the investor’scredit criteria, its yield objective, or its diversifica-tion requirements. Others require more extensiveevaluation, but this is still based on information inthe offering memorandum and any additionalinformation communicated during negotiations.Lenders typically perform their own due diligenceto verify the information in the offering memoran-dum only after circling a deal.

The typical placement is offered to manypotential lenders. The process would be inefficientif each of them gathered all the informationrequired either to reject or to circle a deal and ifeach had to weed out obviously unqualifiedborrowers. In such a situation, the aggregate staffcosts associated with private placement lendingwould be much larger.

Agents improve the efficiency of the intermedia-tion process by performing these two functions. Todo so, they must perform due diligence similar tothat done by lenders during the verification stage.As noted, such examinations of borrowers beginduring the prospecting, advice, and interagentcompetition stages. 131 At this point, many poten-tial borrowers that are not actually able to issueare weeded out on the basis of a modest amountof information-gathering and effort by the agent.Resources are saved because only one organizationprocesses and rejects the ‘‘ applications’’ of suchborrowers and because only one organizationgathers the information that appears in the offeringmemorandum.132

This division of labor works because agents thatdo not perform adequate due diligence willquickly acquire a bad reputation.133 Lenders donot actually commit funds based only on anagent’s due diligence, but they are willing to incurthe costs of initial evaluations. If they later findthat the agent did not conduct a thorough evalua-tion or misrepresented the facts, they can preventfurther losses by backing out of the deal. In therelatively small community of private placementprofessionals, the agent’s reputation will betarnished, not only with that lender but with otherlenders as well. The agent will then be at acompetitive disadvantage, as lenders will be lesswilling to consider placements offered by it in thefuture. Thus, the incentives of agents (with regardto due diligence) are kept closely enough in linewith those of lenders that the efficiencies ofhaving agents perform much of the pre-screeningcan be captured.

Determinants of the Style of Design and Distribu-tion. The terms of a private placement aredetermined mainly by market conditions and therisks associated with lending to the borrower.Securities issued by risky or information-problematic borrowers must include more cove-nants or a higher rate of interest or both. However,the process by which the terms are determinedmay influence the nature of the terms and thecosts associated with issuance. The processincludes the negotiating strategies adopted by theissuer and agent and the way in which lenders areidentified.

For example, an agent may be uncertainwhether or not lenders will insist on a covenantrestricting a borrower’s interest coverage ratio. Ifthe agent makes preliminary inquiries, the lenderswill know that such a covenant is negotiable andwill be more likely to insist on it. The agent mayoffer securities without the covenant to lenderssequentially, hoping to find some that makecounteroffers not including the covenant. 134 But asequential offering runs the risk that some lendersthat would enter negotiations if they saw thecovenant on the term sheet will reject the dealentirely. Returning to such lenders after complet-ing the sequence is difficult. Also, sequentialnegotiations can be time-consuming and costly,

130. Many market participants spoke of rejection rates of 80or 90 percent.

131. Depending on the agent and the nature of competitionamong agents for a borrower’s business, the prescreeningevaluation may be done before the agent receives a mandatefrom the borrower.

132. More than one agent may have to weed out an unquali-fied potential issuer if it approaches several agents.

133. However, the factual accuracy of the offering memo-randum is technically the responsibility of the issuer, not theagent.

134. A sequential offering may involve sending a book, witha request for counteroffers, to half a dozen lenders and then toadditional lenders as needed.

54

and in a long-run equilibrium agents’ fees mustreflect costs. Thus, competitive pressures oftenmilitate against sequential offerings. Instead, theagent may offer securities to many lenderssimultaneously, on a first-come, first-served basis.If the issue is not fully subscribed, terms can bechanged in response to lenders’ counteroffers andanother offering made. However, a simultaneousoffering can be more expensive than a sequentialoffering that is quickly subscribed, as more lendersare involved. Also, for placements that require alead lender, a simultaneous offering to the universeof lenders may be infeasible because smallerlenders will not consider some deals until a leadlender has circled.

In cooperation with the borrower, an agentmakes decisions on four matters in determiningthe style of a distribution:

1. The terms included in the initial term sheet2. The extent to which the initial terms will be

represented as non-negotiable3. Whether to seek a lead lender as the first

step in distribution4. The manner of solicitation of lenders

(sequential or simultaneous) and the number andidentity of those solicited.

Decisions are aimed at obtaining good terms whilelimiting the agent’s costs of design and distribu-tion.135 At the outset, the agent commits to assistthe issuer for a fee equal to a fixed percentage ofthe loan, and thus the agent’s profits are directlyrelated to its costs. Agents usually avoid high-riskstrategies because they collect fees only forsuccessful distributions. They also consider theeffects of a strategy on their reputations andrelationships with lenders. Negotiating strategiesthat annoy lenders may hamper an agent’s abilityto do business in the future.

In this context, several factors appear to be theprimary determinants of the decisions that aremade. One is the complexity or severity of theinformation problems posed by the borrower’sbusiness, financial structure, and corporate struc-ture and by the complexity of the financing inprogress. Complexities force potential lenders toinvest more resources in credit analysis and, insome cases, not all lenders will have the necessaryexpertise. There is an incentive to find a lead

lender for such placements, as the agent can usethe lead’s commitment as a signal to otherinvestors that necessary analyses have been doneand that the terms are satisfactory. There is also anincentive to offer the placement initially to onlyone or to a few potential lead lenders, as they willbe more likely to invest in the necessary analysisif they know that competition to buy the place-ment will be limited until the terms are set. 136

A second factor is the rating of the borrowerand any prospective changes in its condition.Because default risk varies much more acrossB-rated borrowers than across A-rated borrowers,lenders must do much more analysis of lower-rated borrowers before they can negotiate terms.Here, again, an incentive exists to find a leadlender and to negotiate initially with only a fewpotential leads. Lenders, being also more reluctantto lend to borrowers that appear to be headeddownhill, insist on more stringent covenants tocontrol risk. They will be more likely to enternegotiations if the initial term sheet includes astrong covenant package, as it is a signal that theborrower recognizes the problem and will notimpose unusually large negotiating costs on thelender over the term of the loan.

The distribution facilities available to the agentare a third factor affecting distribution strategy.When a financing is highly rated and straightfor-ward, requiring relatively little analysis by lenders,a lead lender may be unnecessary, and offering theplacement simultaneously to the universe ofbuyers of private placements may be possible.Some large investment banks use their fixed-income sales forces to make such offers. Becausethese sales forces already bear the fixed costs ofstaying in communication with a large group ofbuyers, this method can be cheaper to implementthan distributions made solely by the less special-ized members of the private placement group.Thus, other things being equal, agents with suchdistribution channels at their disposal are morelikely to offer a placement simultaneously to manybuyers.

A widespread distribution may not always befeasible. Besides the reasons already given, if aborrower wants to maintain confidentiality about

135. Here terms include not only coupon rate and covenantsbut also in some cases confidentiality, as some borrowers wantto issue quietly, or the establishment of a relationship withparticular lenders.

136. In equilibrium, lead lenders must be compensated forthe costs of analysis of complex placements, and this compen-sation must be in the form of more favorable terms. Lendersrelying on the lead’s signal will have fewer costs of analysis,and thus they can earn excess returns and should be eager tobuy such placements. However, the follow-on lenders must becompensated for the risk that the lead lender did not conduct agood analysis.

55

the transaction, the offering is likely to be shownto a limited number of lenders. Lenders canextract a premium from such borrowers, of course,as breaking off negotiations and turning to anotherpotential lender are costly to the borrower. Aninexperienced or uninformed agent is more likelyto offer a placement to a few lenders at a time andsolicit counteroffers from them than to offer toseveral lenders on a take-it-or-leave-it basis. Suchan agent may lack the knowledge required tochoose an optimal set of terms and may also haverelationships with only a few lenders. A distribu-tion may also be limited if a borrower wishes toestablish a relationship with a particular set ofinvestors. Finally, although in principle a broaddistribution by a fixed-income sales force may bedone quickly for some standard placements, incases where rapid progress on negotiations andapprovals is required the number of lenders oftenmust be small.

Market conditions, too, may influence distribu-tion strategies. When demand is high for place-ments in general or for particular kinds of place-ments, agents are more likely to write initial termsheets with fewer and looser covenants and tosuggest rates slightly below market. 137

This framework is a basis for describing thespectrum of placement design and distributionstyles already mentioned. Agents are most likelyto choose a style similar to a best-efforts publicunderwriting (involving an offering to manylenders on a take-it-or-leave-it basis) when theplacement has a fairly high rating and standardterms, when the issuer is relatively well knownand has no unusual corporate or financial struc-ture, when the issuer does not insist on confidenti-ality or unusual speed, and when the agent has themeans to distribute broadly at low cost. 138

The style at the other end of the spectrum,negotiating terms with one or a few lenders, ismost likely for placements that are highly complexor that require confidentiality, speed, or that aremotivated in part by the borrower’s desire toestablish a relationship.

A common hybrid style involves initial negotia-tions with one or a few potential lead lenders,followed by an offering to many lenders once a

lead has been obtained. This style is mostcommon for placements with some complexity, sothat the signal provided by the lead’s commitmentis important, but in which the borrower does notinsist on confidentiality nor on speed.

In general, the choice of design and distributionstyle is the outcome of the complex decisionproblem previously described. Styles vary widelybecause the circumstances surrounding individualprivate placements vary widely. The examplesgiven here hint at, but do not fully capture, thediversity of styles.

Value Added by Agents’ Design and Distribution.Agents are used primarily because they have theknowledge, expertise, and organization to placesecurities on terms more favorable (even aftersubtracting their fees) than the borrower itselfcould obtain. Some borrowers acting alone mightlocate willing lenders at only moderate cost, butthey could be at a disadvantage in negotiationsbecause the lenders might assume that, shouldnegotiations break down, the borrower would findlocating additional lenders costly. Agents’ activi-ties increase the efficiency of capital marketsbecause, in effect, they heighten competitionamong lenders and reduce the total costs ofborrowing.

Strategic Implications of Distribution Methods forAgents. As we have argued, some agents mayspecialize in serving certain kinds of privateplacement clients (for example, middle-marketcompanies) because their organizations’ relation-ship officers, the primary source of clients,specialize in serving those clients. To some extent,agents also specialize in styles of distribution.Such specialization both influences and is influ-enced by specialization in types of clients.

All private placement agents can perform thestandard varieties of design and distribution, inwhich they send offer memos and term sheets tosome number of potential lenders and thennegotiate with those lenders. One avenue ofspecialization involves the identity of the lendersan agent ordinarily deals with. Because largeinsurance companies often find focusing theirlimited staff time on large or complex placementsmore profitable, agents that advise on mainlysmaller issues may find maintaining close relation-ships with midsized and smaller lenders moreprofitable. Conversely, agents that tend to adviseon large and complex placements may deal mainlywith the largest life insurance companies. Asophisticated borrower surveying the field of

137. During the past two or three years, insurance compa-nies have shifted funds from commercial mortgages andbelow-investment-grade securities toward investment-gradesecurities. Market participants indicated that this shift hasresulted in tighter spreads and more flexible covenants forinvestment-grade placements.

138. Agents’ fees as a percentage of the offering are, onaverage, smallest for this variety of placement.

56

agents may find it most advantageous to chooseone that frequently deals with appropriate lenders.

A more recent variety of specialization involvesthe use of public bond sales forces to offer privateplacements on a take-it-or-leave-it basis to a largenumber of potential buyers. At present, only a fewagents use this method and only for some of theplacements on which they work. The relationshipofficers of these agents provide a steady stream ofclients issuing the kind of highly rated, relativelystandard placements that are most amenable todistribution on a take-it-or-leave-it basis. Accord-ing to market participants, such agents apparentlyare mainly large investment banks. Few, if any,commercial banks appear to use the method at thistime.139

Economies of scope between agenting andpublic debt underwriting do not appear to beenormous. All of the top ten private debt agentslisted in table 10 are either investment banks orcommercial banks with agents located in securitiessubsidiaries with debt underwriting powers.However, five of the agents ranked in the next tierof ten had either no securities subsidiary or onewith limited powers. Thus, an organization canhave a substantial agenting business without alsobeing able to act as underwriter.

Lender Due Diligence and Contract Writing

After enough lenders have circled a deal to makeit fully subscribed, the final phases of the privateissuance process begin. First, lenders that circledverify the information on which they based theircommitments. Large lenders conduct relativelyextensive investigations that include trips to theborrower’s facility (small lenders may again relyon the lead). If the investigations are satisfactory,formal letters of commitment to lend are dis-patched. If lenders find material omissions ormisrepresentations, either the deal falls apart ornegotiations are reopened.

Following formal commitments, by conventionthe lead lender nominates a bond counsel to act asthe lenders’ representative in negotiating thedetailed language of the debt contract. The bondcounsel is paid by the borrower, which retains itsown counsel to assist in negotiations. The agentoften also assists in negotiations.

Closing or settlement concludes the process ofissuance. The documents are signed, and funds aredisbursed to the borrower and the agent.

Information Flows

The private placement market is rife with informa-tion problems. As noted in part 1, the risks oflending to private market borrowers are often hardto observe and to control because relatively littlepublic information may be available about themand because their businesses, corporate structures,or financings may be complex.

The lack of publicly available, timely informa-tion about the terms of private debt, includingprices and other market conditions, is anotherinformation problem. Such information is valuable,and the collection, processing, and sale of it toborrowers is the primary business of agents.Lenders, however, also need such information, andagents are involved in transmission of informationto them as well.

How Agents Gather Market Information

Agents can learn about current market conditionsin four major ways: by observing deals in whichthey participate, by asking lenders, by asking otheragents, and by subscribing to newsletters and otherinformation clearinghouses.

Observation of deals in which an agent partici-pates is most reliable, as the agent sees all offersand counteroffers and knows all details of theinitial and final terms of the debt contract.However, a large flow of deals with a variety ofcredit ratings and levels of complexity is requiredto support a constant reading of current prices andterms for the spectrum of private placementcontracts. According to indications from marketparticipants, even agents with very large volumesof business rely on multiple sources of informa-tion, not just on their own deals.

Agents also ask lenders about the terms of dealsin progress and about completed deals. Suchinquiries are perhaps the primary way that smallagents keep up with market conditions. Lendershave mixed incentives to share information. Onthe one hand, judicious limits on the flow ofinformation to agents may give lenders an advan-tage in negotiations. On the other hand, lendersalso want information from agents and thus willenter into informal sharing arrangements withthem. Lenders also cultivate agents, especially

139. Only a few commercial banks possess securitiessubsidiaries (section 20 subsidiaries) with full debt-underwriting powers, and thus only they among banks wouldpossess public security sales forces.

57

those doing a large volume of business, becausethey want to be offered securities and to be placedat the beginning of the queue in sequentialdistributions. 140 Lenders can reward agents byresponding promptly to offers, by not imposingnuisance costs while deals are in progress, and bysharing information. Because they have the mostto gain by cultivating large agents, which haveboth the largest flow of deals to offer and the bestinformation, lenders are most likely to shareinformation with them. Apparently, agents seldomshare information with one another, perhapsbecause they are in competition.

In recent years, several newsletters and otherpublicly available sources of information aboutprivate market deals have appeared. None offers acomplete picture of the market, and some offerinformation that is slightly dated. However, marketparticipants indicated that they do gather informa-tion from these sources and find it useful. Thenewsletters themselves gather information byasking lenders and agents (and sometimes borrow-ers) about deals recently completed and those inprogress.

Interestingly, some lenders reportedly seldomshare information with the newsletters. Thissituation is consistent with their incentives to shareinformation only with agents from which theyexpect favors in return. Agents also have incen-tives to limit information flows, but these are notso strong as the incentives of lenders. At themargin, the interest of agents may be to increasethe efficiency of the private market, as improve-ments in terms available to borrowers (due toimproved information flows) may increase theflow of deals. However, large agents may losesome of their informational advantage from suchan improvement in efficiency.

How Lenders Gather Information

Lenders’ sources of information are similar tothose of agents, but lenders have an advantage inthat they observe not only the terms of debtcontracts that they buy but also at least the initialterms of all contracts they are offered. Many ofthe larger private market lenders we interviewedstated that they are offered many more than 500deals in a typical year but that they purchase only

a small percentage of them. They could reducetheir prescreening costs by specifying moreprecisely to agents the kinds of deals they willbuy; however, doing so would reduce the sizeof their window on current market conditions.Several market participants mentioned that privatemarket lenders actively lobby agents to offer themevery deal and are unhappy with agents that fail todo so.141

Lenders also gather information from agents,typically by inquiring about the final terms ofdeals they were offered but did not participate in.They may also make such inquiries of otherlenders, though the sense of market participants’comments was that these inquiries are lessfrequent. Newsletters do not appear to be aprimary source of information about marketconditions.

Economies of Scale

Besides being able to spread fixed costs ofperforming agent operations over a larger volumeof business, large agents (and large lenders) havean advantage in gathering the information requiredto operate in the private placement market. Notonly are they able to glean more informationdirectly from deals they participate in, but theyhave more to trade when making inquiries of otherlenders and agents. Such economies of scale maytranslate into larger profits. They may also act as abarrier to entry of new agents, as such agents willtypically have neither large deal flows nor infor-mation to trade. The effect on the profit differen-tial between large and small lenders may be lesssignificant, because large lenders tend to be leadlenders and small lenders can free-ride by buyingpieces of the deals the large lenders commit tobuy.

Since data on the costs and profitability ofagents are not available, quantitative evidence ofeconomies of scale and on the competitiveness ofthe agent market is limited. However, economiesof scale often foster concentration of an industry,and the agenting industry is somewhat concen-trated. In 1991, the top five agents of debt had41 percent of the market by volume, the top tenhad 65 percent, and the top twenty had 89 percent.Of course, as discussed earlier, such concentration

140. One market participant’s revealing comment was that,at general social events, the private placement market lenders,not the agents, pay the tab. By contrast, in the public market,the underwriters, not the investors, pay the tab.

141. Some lenders do implicitly specify broad parametersfor deals they want to see. For example, a few lenders are wellknown to buy only investment-grade placements, and thus theyare less interested in conditions in the below-investment-gradesegment of the private market.

58

could result from a combination of economies ofscope and concentration in the markets for otherfinancial services.

Price Determination

As noted previously, the prices of private marketsecurities are determined primarily by negotiation.In the case of securities distributed on a non-negotiable basis by fixed-income (public bond)sales forces, the negotiations are implicit in thatthe agent uses information about market conditionsto set a price. This section briefly discusses themechanics of price determination and the methodsthat agents and lenders use to set initial andreservation prices.

In most cases, term sheets for private offeringsdo not include a price or a rate spread overTreasury securities of comparable maturity. 142

When they send a term sheet, agents often orallysuggest a price range to potential lenders. Lendersthat circle the deal will circle the terms theyaccept on the term sheet, suggest alternatives forthose they do not accept, and state a rate spreadand a quantity they will purchase at that spread.The spread and terms may then become thesubject of negotiations, or the agent may simplyreject or accept the counteroffer. The agentcollects counteroffers (the circles) and negotiatesuntil it and the issuer decide that the deal is fullysubscribed, at which point investors are notifiedwhether they are in or out of the deal and acoupon rate is set (based on that day’s Treasuryyield curve and the largest spread among thecounteroffers to be accepted). Lenders are thusexposed to a form of interest rate risk during theperiod between notification of acceptance of theircircle and closing. If they hedge risks associatedwith a circled deal and the deal falls through, theyare left with the risk associated with the hedge.Clearly lenders will sometimes have an incentiveto back out of a deal during the period betweencircling and commitment (if interest rates rise), butconventions in the market discourage this action.In general, lenders can pull out of a circled dealwithout damage to their reputations only if theydiscover discrepancies when performing theirown due diligence.

Agents determine initial prices by variousmethods. An obvious method is to use spreads forrecently issued private placements of comparablerisk and maturity. However, partly because privateplacements are often tailored contracts, the privatemarket is thin enough for some risk levels andmaturities so that there may be no comparablerecently issued privates. Thus, agents often lookfor comparable publicly issued corporate debt(especially in investment-grade deals), marking upspreads by their estimate of the public–privatedifferential. Participants’ estimates of the averagedifferential are in the range of 10 to 40 basispoints for investment-grade securities. 143 A fewagents use formal pricing models in their exer-cises, but comments made in interviews suggestthat these are generally used as supplements ratherthan as primary determinants of prices.

Lenders conduct similar exercises to determinemarket prices but also must determine reservationprices. At some insurance companies, this determi-nation is effectively done by portfolio managers ina part of the organization separate from thatresponsible for buying privates. In some cases,portfolio managers mainly compare the returnsavailable from different classes of investments,taking diversification into account. In other cases,they compute required levels of risk-adjustedreturn on equity and then specify some form ofdemand schedule to the private placement group.A demand schedule may be as simple as a targetvolume of private placement purchases in eachrisk class for a given year, at the best availablemarket prices, or as complicated as explicitrequired rate of return on equity with quantityconstraints attached.

Agents’ Fees and Other Costs of Issuance

Issuers generally agree in advance to pay the agenta fixed percentage of the face amount of an issueat closing. The fee is thus contingent on successfulissuance.

We have little quantitative evidence about fees.Market participants agreed that fees vary with thequality and complexity of a financing. Low-ratedor complex deals require more analysis and aremore difficult to distribute and shepherd throughthe lender due diligence and final negotiationstages. Also, percentages vary inversely with deal

142. This statement is true for most offerings of traditionalprivate placements. In some cases in which a placement issimultaneously offered to the universe of potential lenders byan agent’s fixed-income sales force, a spread is specified, andinvestors take it or leave it.

143. The differential for underwritten Rule 144A privateplacements is smaller, perhaps 5 to 15 basis points at this time.

59

size. Agents’ costs have a large fixed componentthat is independent of deal size, and thus agentsmust earn a larger percentage of small deals.

For a $100 million straightforward A-ratedprivate issue, market participants gave fee esti-mates that ranged from 3⁄8 to 5⁄8 percentage point,with the most common answer being 50 basispoints. Estimates ranged widely for complex orsmall issues, up to several percentage points.Many participants stated that fees have fallenslightly in recent years.

Issuers bear other fixed costs of issuance.Besides the opportunity costs of cooperating withdue diligence by agents and lenders, issuers mustpay the lenders’ bond counsel and typically mustalso retain their own counsel and pay othermiscellaneous costs associated with negotiations.Market participants’ estimates of these costs variedwidely, but for straightforward issues were oftenbetween $50,000 and $125,000, or 5 to 13 basispoints for a $100 million issue.

Private Market Efficiency

In considering the efficiency of the privateplacement market, we focus on whether lenders oragents earn either subnormal or supranormalprofits. Quantitative data on which precise judg-ments might be based are not available, but thecomments of market participants suggest that themarket is relatively efficient.

With regard to lenders’ profits, one majorinsurance company stated recently in a publicforum that interest rates on its private originationsduring 1989–91 were, on average, 31 basis pointshigher than rates on comparable public issues andthat 18 basis points of this differential were spenton costs of origination and monitoring. Thesenumbers leave 13 basis points for profit and forcompensation for the reduced liquidity of privateplacements relative to that of publicly issuedbonds. Another major company displayed propri-etary data during interviews indicating that recenthistorical net loss rates due to defaults on privateplacements have been similar to loss rates oncomparably rated public issues.

Presuming that these data are accurate andreasonably typical of private market lenders’experience, and assuming that lenders do not makesubnormal or supranormal profits on their publicbond market activities, the data place roughboundaries on the degree of private marketinefficiency that may exist. The key question is thesize of the differential required to compensate

lenders for the relative illiquidity of privateplacements. If this differential is near zero, thenprivate lenders may be making modest excessprofits. 144 If the differential is near 13 basis points,then lenders are taking a modest loss at themargin. Regardless, the dollar sums involvedapparently cannot be large enough to representextraordinary inefficiencies that would be a majorconcern to policymakers.

Agents’ profits are even harder to estimate, asno information is available about their costs.Based on market participants’ remarks about feesand staff sizes and on publicly available informa-tion about the volume of issues assisted byparticular agents, the largest agents may beearning substantial marginal profits on the staffand overhead costs of their private placementgroups alone. However, portions of these profitsmust be attributed to the actions of relationshipofficers and other divisions of commercial banksand investment banks, so actual profit rates maynot be unusual.

Smaller agents may also be able to make profitsif their flow of business is reasonably steady. Asnoted, smaller agents will find maintaining theirknowledge of market conditions more expensiveand difficult, and they will face minimum fixedcosts of maintaining a staff.

We have no reason to think that agents makelarge excess profits, and many market participantsremarked on the substantial competition thatexists. On the whole, the private placement marketappears to be reasonably efficient, although it maynot always react quickly to changes in conditions.

Private Placements without an Agent

Data are not available on the volume of privateplacements issued without an agent’s assistance,but it is probably substantial. Estimates by majorprivate market lenders suggest that as much asone-third of total private issuance is done withoutan agent. In most cases, such issues are sold by acompany that has previously borrowed in theprivate market and sold to investors that boughtparts of the previous placements.

144. Computing profit rates is difficult because the capital atrisk is hard to identify. If at the margin the only capital at riskis the staff and overhead costs of origination and monitoring,lenders’ marginal rate of return on equity in private marketoperations may be as high as 70 percent. But that estimate isalmost surely far too high because private market lending is, onthe whole, probably riskier than buying public bonds, so moreequity must be allocated to such lending than to public bondmarket lending.

60

In such cases, some of the services that agentsprovide are not relevant. For example, duediligence by the agent adds little or no value,as monitoring by the lenders since the previousissuance has kept them informed about theborrower. Locating appropriate potential lenders isalso virtually costless for the borrower. Apparentlythe other services provided by the agent—notably,help in negotiating terms—are thought by someissuers not to be worth the fee. Many repeatborrowers do use an agent, however, so eithercircumstances or opinions differ across repeatborrowers.

Agent Operations under Rule 144A

As noted in part 2, section 1, the market for manynew private issues made under Rule 144A oper-ates much more like the market for new publicissues than like the traditional private placementmarket. Some securities involved in transactionsexempt from registration under Rule 144A havebeen distributed by agents in the fashion describedabove. Others, especially those of well-knownU.S. or foreign companies, have been formallyunderwritten.

Agent prospecting, advice, competition, and duediligence are much the same for both underwrittenand traditional privates, but the distribution ofunderwritten securities is usually similar to thatseen in the public market. Underwritten securitiesare often sold to typical buyers of public issues.For example, many life insurance companies buysuch issues through their public bond investmentgroups, not through their private placementinvestment groups.

When there is no firm-commitment underwrit-ing, some Rule 144A offerings are made on atake-it-or-leave-it basis by the agent organization’s

fixed-income sales force. Thus, Rule 144A place-ment distributions are often at the public-like endof the spectrum of private market distributionstyles. Agents that are proficient at this style ofdistribution have a distinct competitive advantagein assisting Rule 144A placements.

Summary

Agents are a key part of the market for privatelyplaced debt. They gather, process, and sell infor-mation that would be prohibitively expensive formany issuers themselves to collect. They helpenforce norms of behavior for borrowers andlenders that make the private market functionmore efficiently.

Agenting appears to be associated with econo-mies of scale and scope that confer a distinctadvantage on the large commercial banks andinvestment banks that specialize in serving thecorporate finance needs of middle-market andlarge companies. Economies of scope of agentingapparently occur with other corporate financeservice activities, in that bank and investmentbank relationship officers can provide a stream ofclients to agents while selling other products.Economies of scale arise from fixed costs ofmaintaining a staff of agents and from the infor-mation sources in the private market, which aresuch that costs of collecting information fall as thevolume of an agent’s business rises.

That agenting appears to be a competitivebusiness with low barriers to entry implies that theprofits available to new or small agents are notlarge. Slow trends of falling information costs andincreasing information flows will likely increasecompetition among agents even more and willimprove the efficiency of the private placementmarket as a whole.

61

Part 3: Special Topics

Part 3 focuses on two special topics. One is arecent credit crunch that cut off the access of mostbelow-investment-grade companies to the privateplacement market. The second is the current andprospective role of commercial banks in theprivate placement market.

Credit crunches have long been an interestingand controversial topic, partly because producingcompelling evidence that a crunch occurred isoften difficult. For the recent private placementcredit crunch, relatively extensive evidence isavailable. The causes of the crunch are intertwinedwith the intermediated and information-intensivenature of the private market and are somewhatdifferent from the mechanisms said to be responsi-ble for a possible concurrent crunch in the bankloan market. The story of the private placementcredit crunch sheds additional light on the eco-nomics of the private market and of financialintermediation.

The role of banks in the capital markets haschanged substantially during the past twenty years:The rise of the commercial paper market and othermarkets is associated with a decline in the share ofbank loans in all debt financings. As the bank loanand the private placement market are information-intensive and as medium-sized companies areresponsible for a large share of borrowings in bothmarkets, the two markets may be in competition,and one may come to dominate. However, we findthe latter possibility unlikely. Because the focus ofbanks on relatively short-term lending appears toresult from the maturity of their liabilities, theyprobably will not eclipse the private market as asource of long-term loans to information-problematic borrowers unless the structure of theirliabilities changes in a major way. Repeal of thelaws governing the separation of banking andother forms of commerce seems to be only a firststep in such a change. For similar reasons,traditional buyers of private placements appearunlikely to become major short-term lenders.Finally, neither commercial banks nor investmentbanks seem to possess a competitive advantagethat would allow them to dominate the market forprivate placement agent services.

1. The Recent Credit Crunch in thePrivate Placement Market

Since the middle of 1990, issuers of below-investment-grade securities have encountered a

sharp contraction in the availability of credit in theprivate placement market. A sharp rise in interestrate spreads on these securities indicates that thereduction in supply has been larger than anydecline in credit demand associated with the weakeconomy. This credit crunch has resulted mainlyfrom a greater reluctance of life insurance compa-nies to assume below-investment-grade credit risk.This reluctance is due mostly to concerns thathigh balance sheet proportions of such investmentscould lead to a runoff (or even a run) of liabilitiesand threaten the profitability and, perhaps, eventhe survival of insurance companies. Asset qualityproblems at many life insurance companies,regulatory changes, and runs at a few insurancecompanies have contributed to the reluctance ofinsurance companies to buy below-investment-grade private placements.

The reduced availability of credit from lifeinsurance companies has likely adversely affectedthe ability of below-investment-grade companiesto obtain financing. Few alternative lenders haveentered or expanded their presence in the below-investment-grade sector of the private market tofill the void. The reason appears to center on thehigh start-up costs that potential lenders mustincur to enter the private market. Also, the numberof alternatives to private placements is limited.Although they may be the main practical alterna-tive, bank loans are far from perfect substitutes,and some firms shut out of the private market mayhave found banks to be reluctant lenders.

Definition of Credit Crunch

Many definitions of the term credit crunch haveappeared in the literature. 145 In our view, a creditcrunch occurs when, for a given price of credit,lenders substantially reduce the volume of creditprovided to a group of borrowers whose risk isessentially unchanged. That is, a credit crunch iscaused by a reduction in lenders’ willingness tomake risky investments or by a ‘‘flight to quality’’by lenders. In terms of a standard supply anddemand diagram, a credit crunch is a substantialdecline in the volume of credit caused mainly by aleftward shift of the credit supply curve, when the

145. See Owens and Schreft (1992) for a review ofdefinitions.

shift is not due principally to an increase in theriskiness of borrowers. This definition is similar inspirit to that of Bernanke and Lown (1991), whodefine a crunch as ‘‘a significant leftward shift inthe supply for bank loans, holding constant boththe safe real interest rate and the quality ofpotential borrowers.’’

A contraction of supply alone does not neces-sarily imply a credit crunch, as credit availabilitymay decrease and lending terms tighten because ofan increase in the riskiness of borrowers. Thus ourdefinition of a credit crunch does not include areduction in supply that is a normal response to arecession or an economic slowdown. In suchcircumstances, the riskiness of borrowers normallyincreases, and lenders demand compensation eitherin higher interest rates or in tighter nonprice termsof loans. Although borrowers might characterizesuch a reduction in credit supply as a creditcrunch, such a characterization would not beappropriate because the decrease in credit is anormal response of lenders to changing economicconditions. Cantor and Wenninger (1993) refer tothis situation as a ‘‘credit slowdown.’’ 146

Our definition of credit crunch differs fromsome, notably that of Owens and Schreft (1992),in that it does not require that the reduction incredit be accomplished by nonprice rationing. Thereduction may be effected entirely by an increasein the relative price of credit, as would normallyoccur in response to a leftward shift of a supplycurve, or by some combination of price increaseand nonprice rationing.

Evidence That a Credit Crunch Occurred

Recent events in the below-investment-gradesegment of the private placement market qualifyas a credit crunch because gross issuance ororiginations for below-investment-grade debtdeclined substantially and spreads on such debtincreased sharply, whereas spreads on investment-grade private placements held steady or declined.A general increase in the riskiness of borrowerscannot account for these phenomena. The declineof issuance may have been accomplished partly bynonprice rationing, but we have no quantitativeevidence to support such a claim, and marketparticipants’ remarks about nonprice rationingwere mixed.

Data from three separate sources confirm areduction in issuance of below-investment-gradeprivate placements. First, gross issuance bybelow-investment-grade, nonfinancial corporationsfell more than 50 percent in 1991, a much steeperdrop than that seen in issuance by investment-grade corporations (table 12). 147 As a percentageof gross offerings, below-investment-gradeissuance declined from 16 percent in 1990 to9 percent in 1991. Data for 1992 indicate thatissuance remained depressed, although the percent-age was slightly above that in 1991. Second,although total commitments by major life insur-ance companies to purchase private placementsremained roughly constant from early 1990through mid-1992, the proportion of below-

146. Cantor and Wenninger’s definition of credit slowdownwould also include a reduction in credit due to a reduction indemand.

147. Estimates of issuance were constructed from dataobtained from IDD Information Services. Gross issuanceexcludes offerings to finance employee stock ownership plans(ESOP) and restructurings. Underlying developments are moreevident with their exclusion, as both were heavy in 1989 butfell off sharply in 1990 and 1991. Before 1990, ratings reflectthe judgment of agents supplying information on transactionsthey placed. Thereafter, ratings assigned by the NationalAssociation of Insurance Commissioners are available.

12. Gross issuance of private placements by nonfinancial corporations, 1989–921

Billions of dollars except as noted

Type of issuance 1989 1990 1991 1992

Total issuance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54.7 49.9 42.9 29.5Below-investment-grade . . . . . . . . . . . . . . . . . . . . . 6.6 8.1 3.8 3.2

MemoRatio of below-investment-grade

to total (percent) . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.1 16.2 8.9 10.8

1. Excludes restructuring-related issues in excess of $250million and issues to finance employee stock ownership plans.

Source. IDD Information Services.

64

investment-grade issues dropped sharply in themiddle of 1990, from 21 percent in the first halfof 1990 to 11 percent in the second half of thatyear. Since then, the percentage has variedbetween 31⁄2 percent and 71⁄2 percent (chart 16). 148

Third, the reduced rate of gross purchases indi-cated by the survey is also evident in insurancecompanies’ holdings of below-investment-gradesecurities. Holdings of such securities at all lifeinsurers fell 11 percent in 1991, whereas holdingsof investment-grade securities rose nearly 12 per-cent. As a result, speculative-grade private bondsas a percentage of all private placements ininsurance company portfolios declined from19.8 percent in 1990 to 16.7 percent in 1991. Thelow rate of commitments to purchase below-gradeprivate placements in 1992 led to a further declinein their share to 15.3 percent last year.

Accompanying the decline in gross issuance andoutstandings has been a sharp increase in yieldspreads on below-investment-grade private

placements. According to market reports, before1990 the difference between yields on BB- andBBB-rated private placements with comparableterms was about 100 basis points; since then, thedifference has been as high as 250 basis points. 149

Although data are unavailable for periods before1990, the relative movement in yields on BB andBBB private bonds is confirmed in the spreadsreported in the ACLI survey (charts 17 and 18). 150

During the first half of 1990, the spread betweenyields on BB private placements and comparableTreasury securities was about 300 basis points,compared with 190 basis points on BBB privateplacements. From that time, the spread on BBbonds moved up to almost 425 basis points in thesecond quarter of 1991, but more recently it hasretreated to around 350 basis points. During thesame period the BBB spread drifted down to180 basis points. Similarly, the spread on A-ratedprivate placements varied little over the past threeyears. 151

The substantial increase in spreads over Trea-suries for BB private placements cannot plausiblybe attributed to a general increase in risk associ-ated with the slowdown in economic activitybecause such an increase in risk should have alsoled to an increase in BBB spreads. In fact, thosespreads declined. Similarly, although the slow-down might have caused issues to be moreconcentrated at the low-quality end of the riskrange that each rating category spans, leading toan increase in average spreads for each ratingcategory, such a mechanism should have affectedboth BB and BBB spreads. The data thus indicatethat, within the below-investment-grade segment

148. Commitment data are from a survey of major lifeinsurance companies by the American Council of Life Insur-ance (ACLI). Respondents to the survey hold approximatelytwo-thirds of all private placements in the general accounts oflife insurance companies. The survey began in 1990, so earlierdata are not available for comparison. However, at year-end1990, the twenty largest life insurance companies reported that20.1 percent of their private placements were below investmentgrade. Hence, the 21 percent share of private placementcommitments going to below-investment-grade bonds in thefirst half of 1990 probably was similar to earlier rates ofacquisition of such securities.

149. BBB-rated bonds are investment grade, whereas thoserated BB are below investment grade.

150. Care must be used in interpreting the reported spreads.Although they are transaction prices, they do not reflect astandardized security. The nonprice terms of private placementscan differ widely for bonds carrying the same credit rating, andthe terms affect the yields. For example, at any given moment,the difference in spreads between the highest-risk BB issue andthe lowest-risk BB issue may be as much as 150 basis points.Under normal circumstances, averaging spreads within a ratingcategory produces a representative spread for that rating.However, as most of the BB bonds issued since mid-1990probably were at the least risky end of the BB risk range, theincrease in the BB spread shown in chart 17 probably under-states the actual increase.

151. In the public high-yield bond market, spreads increasedsharply from mid-1989 through 1990 but have since fallensignificantly, though they remain above the levels that prevailedin early 1989. Issuance of public junk bonds stopped almostcompletely during 1990 and most of 1991 but surged in 1992to the second highest level ever. Thus, experience in the publicjunk bond market has been significantly different from that inthe market for below-grade private debt.

16. New commitments to purchase below-investment-grade private placements as apercentage of total commitments by major lifeinsurance companies, 1990–92

Percent

0

5

10

15

20

1990:H1 1990:H2 1991:H1 1991:H2 1992:H1 1992:H2

Source. American Council of Life Insurance.

65

of the private placement market, for a given levelof risk loan prices went up whereas the volume ofloans went down. These facts support our asser-tion that a credit crunch occurred within thatmarket segment.

Sources of the Credit Crunch

A credit crunch can occur for several reasons. Itmay result from actions taken by regulators thataffect lenders’ ability or incentive to assumecertain risks. It may result also from internaldevelopments at lending institutions, such asunexpectedly large loan losses, that cause portfolio

rebalancings involving greater conservatism inlending. For lenders that are financial intermedi-aries, a credit crunch may result from liabilityholders’ becoming concerned about the intermedi-aries’ financial condition. The ability of intermedi-aries to raise funds to support their investmentactivity may be adversely affected in such circum-stances and may lead to their adoption of moreconservative investment strategies to restore publicconfidence. The latter mechanism appears to havebeen primarily responsible for the crunch in theprivate placement market. Problems of assetquality at life insurance companies, a change inregulatory reporting requirements, and runs on afew insurers combined to raise doubts about thesolvency and liquidity of insurance companies andto focus the public’s and the rating agencies’attention on the proportion of an insurer’s assetsinvested in below-investment-grade securities as asignal of its solvency.

Publicity about high proportions of poorlyperforming commercial mortgages in insurancecompany portfolios was one event raising doubtsamong the public about the solvency of insurers.Commercial mortgages make up 25 percent ofgeneral account assets at the twenty largestinsurance companies, which include most of themajor participants in the private placement market.Additional exposure to commercial real estaterisks comes from direct real estate investments,which at many life insurance companies consistprimarily of real-estate-related limited partner-ships. As the press has widely reported, delin-quency and foreclosure rates on these commercialreal estate investments have risen sharply over thepast few years. These problems heightened publicawareness of the financial problems of lifeinsurance companies and thus added to thepressure on those with significant holdings ofcommercial real estate loans to shift out of alllower-quality assets. Also, since even soundcommercial real estate loans turned out to beriskier than anticipated when they were made, lifeinsurance companies shifted investments towardhigh-quality assets.

Publicity about losses on some publicly issuedjunk bonds also raised concerns about the qualityof below-investment-grade securities in general,and a change in regulatory reporting requirementsmade insurance companies’ holdings of suchassets seem to have increased. In June 1990, theNational Association of Insurance Commissioners(NAIC) introduced finer distinctions in its creditratings of corporate bonds, including privateplacements. Under the old rating system, many

17. Yield spreads on privately placed corporatebonds, 1990–921

Basis points

350

250

150

BB

BBB

A

1990 1991 1992

1. Quarterly weighted averages.Source: American Council of Life Insurance.

18. Difference between BB spread and BBBspread, 1990–921

Basis points

200

100

150

1990 1991 1992

1. See chart 17 for notes and source.

66

securities, especially public bonds, with creditquality equivalent to BB or B received aninvestment-grade rating. To correct this shortcom-ing, the NAIC adopted a system with categoriesmore closely aligned with those in the publicmarket (table 13). NAIC-1, the top rating, wasgiven to securities rated AAA to A; NAIC-2 toBBB securities; NAIC-3 to BB securities; andNAIC-4 to B securities. Although insurers’ actualholdings were probably little changed, the reclassi-fication resulting from the new system causedinsurers’ reported holdings of below-investment-grade bonds, both private and public, to risebetween 1989 and 1990 from 15 percent of totalbond holdings to 21 percent. The level of reportedholdings of high-yield bonds jumped more than40 percent.

The sudden appearance of a much increasedpercentage of below-investment-grade securitieson the balance sheets of life insurance companiesfocused the attention of policyholders and otherholders of insurance company liabilities on thecomposition of insurers’ bond holdings. Asevidence of increased public sensitivity, a recentstudy by Fenn and Cole (forthcoming) found thatstock prices of insurance companies with highconcentrations of junk bonds were adverselyaffected in early 1990 by the publicity surroundingthe financial problems of First Executive, whoseinsurance units subsequently failed because oflosses on junk bonds. In contrast, stock prices ofinsurance companies with little exposure to junkbonds were not affected. The public’s greatersensitivity to the quality of life insurance compa-nies’ assets discouraged many insurers from

purchasing lower-quality private placements fromfear of losing insurance business to competitorswith lower proportions of below-investment-gradebonds in their portfolios.

That public fears regarding below-investment-grade private placements were warranted is notclear, as market participants report that loss rateson those securities have not been unusual. Lossrates on such securities may be expected to differfrom those on similarly rated public junk bondsbecause private placements typically containcovenants or collateral and because only a fewinformation-intensive lenders are involved; thuscorrective actions are more timely, and workoutsare less difficult. Because nonparticipants lack aclear understanding of the private market,however, the public has a tendency to equatebelow-investment-grade private placements withpublic junk bonds.

Another development pressuring insurancecompanies to restrict purchases of below-investment-grade private placements has been theconcern of credit rating agencies about the lack ofliquidity of private placements, especially thosethat are below investment grade. This concernappears to be a consequence of the July 1991collapse of Mutual Benefit, which lacked theliquidity needed to meet heavy redemptions bypolicyholders. Driven by a fear of being down-graded, insurance companies have sought moreliquidity in their bond portfolios by concentratingon higher-grade credits, which are more readilysold in the secondary market. 152

Another regulatory move by the NAIC appearsnot to have been a significant cause of the crunch.This move involved changes in the mandatorysecurities valuation reserves (MSVR) held againstbonds in life insurance company portfolios. Forbonds that would have been rated investmentgrade under the old rating system, but fell toNAIC-3 or NAIC-4 under the new system,required reserves jumped from 2 percent of thebonds’ statement values to 5 percent for NAIC-3and 10 percent for NAIC-4.153 Also, the timeallowed to reach the mandatory reserve levels was

152. Some market participants reported an increase ofsecondary market sales of private placements by life insurancecompanies during 1991. The sales were done discreetly toavoid raising concerns and causing the price of the securities tofall, as they usually do after appearing on a bid list. Somemarket participants interpreted the increase in secondary marketactivity as an attempt by the sellers to increase the liquidity oftheir portfolios. Others interpreted it as an attempt to demon-strate the liquidity of private placements to the rating agencies.

153. Mandatory reserve levels for NAIC-1 bonds werereduced, while those for NAIC-2 bonds were unchanged.

13. NAIC credit ratings

NAIC rating designationEquivalent rating-agency designation

Old system 1

Yes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AAA to BNo* . . . . . . . . . . . . . . . . . . . . . . . . . . . . . BB, BNo** . . . . . . . . . . . . . . . . . . . . . . . . . . . CCC or lowerNo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . In or near default

New system 2

1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AAA to A2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . BBB3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . BB4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CCC or lower6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . In or near deafult

1. The asterisks appended to the ‘‘No’’ ratings are part ofthe rating designation.

2. Effective December 31, 1990.Source. Securities Valuation Office, National Association of

Insurance Commissioners.

67

shortened. At year-end 1991, however, all of thetwenty largest life insurance companies hadMSVRs that were more than adequate to meet thefully phased-in standards.

The individual importance of these factors ascauses of the credit crunch is hard to isolate. Theyare, however, interrelated. For example, the effectof the new NAIC rating system probably wouldhave been much smaller had insurance companiesnot experienced problems with commercial realestate loans. Futhermore, the new rating system,combined with the failure of First Executive,focused public attention on below-investment-grade private placements as an asset that couldadd to the industry’s financial problems. In anycase, the main impetus behind the credit crunchhas been life insurance companies’ fears thatliability holders might lose confidence in them andredeem insurance policies, annuities, and guaran-teed investment contracts should they exhibitabove-average holdings of below-investment-gradesecurities.

Prospects for an Easing of the Crunch

As a group, life insurance companies are unlikelyto resume investing in below-investment-gradeprivate placements at pre-1990 levels until theirasset problems have improved and public concernabout the health of the industry has appreciablydiminished. As this improvement hinges mainlyon a recovery of the commercial real estatemarket, many analysts expect that insurers will,for the foreseeable future, remain reluctant toprovide funds to the low-grade sector of theprivate market. This prospect has already ledsome insurers to cut staff and to reduce resourcesdevoted to credit evaluation and monitoring. If thecutbacks become widespread, the long-run abilityof the insurance industry to supply credit tomedium-sized, below-investment-grade companiescould be impaired.

Risk-based capital standards, which becomeeffective at the end of 1993, could reinforce thereluctance of insurance companies to buy below-investment-grade securities. The new standards areaimed at measuring the prudential adequacy ofinsurers’ capital as a means of distinguishingbetween weakly capitalized and strongly capital-ized companies. To this end, insurance companieswill report the ratios of their book capital to levelsof capital that are adjusted for risk. As an insur-er’s ratio falls progressively below one, succes-sively stronger regulatory actions will be triggered.

In the current environment, most insurers willprobably attempt to achieve ratios in excess ofone. One way they can raise their risk-basedcapital ratios is to shift into low-risk assets. In thisregard, below-investment-grade securities carryrisk weights much higher than those oninvestment-grade bonds and even those on com-mercial mortgages. Over time, however, as thefinancial condition of insurance companiesimproves and public concern about their healthrecedes, insurers will be more inclined to considerrisk-adjusted returns in reaching investmentdecisions and thus may allocate a greater propor-tion of assets to higher-risk categories, such asbelow-investment-grade bonds.

Despite the almost three-year absence ofinsurance companies from the below-investment-grade sector and the persistence of unusually highspreads, new lenders have not picked up much ofthe slack in the private placement market, pri-marily because of the high start-up costs ofentering the market. Long-term investments inexpensive internal monitoring systems and staffsof credit analysts, lawyers, and workout specialistsare required. Also, the market operates largely onthe basis of unwritten, informal rules enforced bythe desire of major agents and buyers to maintaintheir reputations. Thus, to an outsider, the way themarket operates may be hard to understand. Beinga newcomer to the market with no establishedreputation may involve costs. These factors mayinhibit outside investors from risking their moneyin this market.

State and large corporate pension funds arenatural candidates to fill the gap left by theinsurance companies in the private market becauseof their demand for fixed-rate investments. Manypension funds, however, have charters that preventthem from investing in below-investment-grade orilliquid assets. Most pension fund managers arealso reportedly reluctant to invest in an unfamiliarmarket. Because pension funds generally lack thenecessary capabilities for due diligence andmonitoring, their managers have difficulty famil-iarizing themselves with the private market bymaking small initial investments. A decision toinvest in below-investment-grade private place-ments involves a significant long-term commit-ment of resources that few pension fund managersappear to find attractive. In the case of statepension funds, even if they wished to invest, manywould face problems in hiring the necessarypersonnel because state legislatures generallycontrol staff sizes and salaries. Any attempt bystate pension funds to hire large numbers of credit

68

analysts thus could run into political obstacles.Pension funds (and others) might quickly enter

the private market by investing in funds managedby professional private placement investors.Several funds have been formed in the past twoyears, but they are unlikely to operate on a scalesufficient to fill the void left by the insurancecompanies. Pension fund managers appear reluc-tant to invest even indirectly in a market withwhich they are unfamiliar. In addition, some areconcerned that fund managers would not monitorborrowers with sufficient diligence. Also, insur-ance companies, which would be the primarysource of the managerial resources necessary foroperating of managed private placement funds,have thus far not set up funds on a large scale,even though some companies currently haveexcess capacity to analyze and monitor lower-quality credits. Some are unwilling to make along-term commitment of resources to this effortbecause they expect eventually to resume investingin below-investment-grade private placements fortheir own accounts. Finally, most institutionalinvestors would expect insurance companies actingas investment managers to purchase some of thesecurities for their own accounts. Such a require-ment lessens the incentive to establish managedfunds because of insurers’ current aversion topurchasing below-investment-grade bonds.

Finance companies face much smaller start-upcosts than pension funds do, but their participationhas traditionally been in the highest-risk segmentof the private placement market, a segment inwhich life insurance companies have not generallybeen active. Insurers typically have madeunsecured loans, mainly to the highest-qualityspeculative-grade borrowers. In contrast, financecompanies specialize in secured lending, normallywith equity features attached. Thus, the risk–returnprofile of the typical insurance company borrowerdoes not suit finance companies, nor would suchborrowers generally find finance companies’ termsattractive. In addition, several finance companiesthat were significant lenders in the private markethave reduced their lending to low-rated firmsbecause they have been faced with credit problemsof their own.

Marginal increases in the number of lenders andin their commitments to below-investment-gradeprivate placements may not have much effect onthe credit crunch. With only a few lenders remain-ing in this segment of the market, and with mostof these willing to lend only a limited amount toany one borrower, agents often have difficultyputting together a syndicate of lenders sufficient to

purchase even medium-sized issues. Because theagents must incur fixed costs before a deal can beproved viable, and because they are paid onlyupon success, most agents have also withdrawnfrom the below-investment-grade segment of themarket. This situation explains an apparentparadox: Those few remaining, willing lenderssometimes complain that not enough prospectiveissues are coming to market to permit themto lend all their funds available for below-investment-grade borrowers. Thus the crunchmay disappear only with a wholesale return oflife insurance companies to this market segmentor with the entry of a significant number of newlenders.

One development that may have eased thecrunch for a few borrowers is the increasedfrequency of ratings of private placements bymajor rating agencies. Issuers on the cusp betweena NAIC-2 and NAIC-3 often obtain ratings fromone of the agencies before seeking ratings fromthe NAIC. Because the agencies charge higherfees for ratings than does the NAIC and are lessoverworked, they can often gather more informa-tion and conduct more extensive analyses, whichsometimes justify investment-grade ratings.154 TheNAIC generally accepts such ratings but reservesthe right to overrule them.

Effects on and Alternatives of Borrowers

The effect of this credit crunch on the economicactivity of potential borrowers is impossible toassess with any precision. As private placementsare seldom the vehicle for providing day-to-dayworking capital, it seems unlikely that many

154. Although the NAIC does consider covenants andcollateral in rating an issue, the agencies may be able to givemore consideration to these factors. The appropriate focus of arating procedure is somewhat different for public bonds thanfor private placements. Investors in public bonds tend to bepassive and ill-prepared to work through instances of borrowerdistress and thus are interested mainly in the likelihood ofdefault, which may be relatively insensitive to covenants andcollateral (which, in any case, are rare in public bonds).Investors in private placements, however, are prepared to dealwith distress and are interested primarily in the likelihood ofloss rather than default. Methods of rating public bonds thatfocus on distress may thus produce ratings of private place-ments that are too low on average, as they do not considercovenants and collateral. Thus, most issuers seeking a ratinghave gone to agencies whose ratings do measure likelihood ofloss. Of the four rating agencies whose ratings are accepted bythe NAIC, Fitch and Duff & Phelps have produced such ratingsfor some time, and Standard & Poor’s has recently developed arating system specifically designed for private placements thatfocuses on likelihood of loss. Moody’s is the fourth approvedagency.

69

potential borrowers have failed because of a lackof financing. Private placements often providefunds for expansion, however, and the growth ofsome medium-sized businesses possibly has beenconstrained by this credit crunch. According tomarket participants, one rationale for privateissuance is not only to lengthen the maturity oftheir debt but also to loosen constraints imposedby the collateral requirements typical of bankloans. Many medium-sized borrowers can obtainbank loans only in amounts up to 50 percent offinished inventory and 80 percent of eligiblereceivables. Often, upon reaching those limits,borrowers have issued an unsecured privateplacement, used part of the proceeds to pay downthe bank debt, and used the remaining proceedsand new bank debt to finance expansion.

With that course no longer open, low-ratedborrowers must attempt to find other sources ofcapital. The bank loan market seems to be the firstalternative for many lower-rated borrowers.Although market participants disagree somewhat,most report that the credit problems at commercialbanks have caused these banks to limit lending, totighten terms as lines have come up for renewal,or even to eliminate lines of credit. This view isconfirmed by the surveys of the lending terms oflarge banks periodically undertaken by the FederalReserve System.155 Furthermore, some insurancecompanies have reportedly had to increase theirloans to existing borrowers whose credit lineshave been cut by their commercial banks.156

Some low-rated companies have taken advan-tage of favorable stock market conditions in 1991and 1992 and issued equity. In some cases, thereduced leverage resulting from equity injectionshas raised issuers’ credit ratings to investmentgrade, and has given them renewed access to theprivate bond market. Alternatively, some firmshave attached credit enhancements to their privateplacements to move up to an investment-graderating. The public junk bond market, despite itsrevival in the latter half of 1991, has been asource of funds for only a few companies, as the

typical below-investment-grade private issue isgenerally too small and too complex a credit forthe public market.

Conclusion

The market for privately placed debt is served bylenders that are financial intermediaries. As such,the market is vulnerable to breakdowns, whichoccur when those who provide funds to thefinancial intermediaries are no longer willing to doso or when intermediaries become sensitive to thethreat of such a withdrawal. This mechanismappears to be the main one behind the recentcredit crunch for below-investment-gradeborrowers.

The conditions causing the breakdown infinancial intermediation at life insurance compa-nies appear unlikely to ease significantly in thenear future. With other lenders and markets unableto fully accommodate the financing needs of themedium-sized, below-investment-grade companiesthat are most affected, those companies may forseveral more years have more difficulty than usualin financing expansions.

2. The Current and Prospective Roles ofCommercial Banks

Commercial banks participate in the privateplacement market as issuers, buyers, and agents.They also compete with private market lenders inproviding credit. Drawing on parts 1 and 2, thissection describes the current role of banks in theprivate placement market and speculates abouttheir role in the future.

Banks as Agents and Brokers

U.S. commercial banks have recently been strongcompetitors in the market for private placementagenting services. Of the 5,550 private place-ments of debt appearing in the IDD database for1989–91, U.S. commercial banks were either soleagent or co-agent for 1,944, or 35 percent. Theirshare of volume was 32 percent. 157 Foreign bankshad a 1 percent share of all volume.158 In the

155. Board of Governors of the Federal Reserve System,‘‘Senior Loan Officer Opinion Survey,’’ various issues.

156. Interestingly, some of the movement of borrowersbetween banks and insurance companies seems to have been afunction of the different ways in which regulators and ratingagencies classify high-risk credits. Some credits (admittedlyfew in number) that carried a highly leveraged transaction(HLT) designation, yet were rated NAIC-2, have found a muchwarmer reception in the private market than at the banks.Conversely, some issues rated NAIC-3 or below by the NAICbut not carrying the HLT status reportedly have satisfiedtheir financing needs at banks rather than at the insurancecompanies.

157. The amount of a co-agented issue was split equallyamong co-agents in computing shares of volume.

158. Any subsidiary, branch, or bank owned by a foreignbank was classified as a foreign bank agent.

70

market for private equity agenting, U.S. banks hada 14 percent share of volume during 1990–91,whereas foreign banks had a 6 percent share. 159

During 1975–77, U.S. banks had only about a7 percent share of the total private placementagenting market (Board of Governors, 1977).Their share has clearly grown substantially duringthe ensuing fifteen years.

Table 14 lists the twenty-five U.S. banks thatappear as agents in the IDD database for theperiod 1989–91, along with the number andvolume of assisted placements of both debt andequity. Two things about the list are striking. First,only ten banks accounted for 98 percent of theknown volume of new issues assisted by banks.Second, the list is relatively short when comparedwith the list of more than 10,000 commercial

banks in the United States. The table is surelyincomplete, as some banks that act as agents maynot report their transactions to IDD; however, itdoes show that apparently only a small fraction ofbanks act as agents.

As a group, commercial banks do not appearto specialize in assisting types of transactions orissuers in industries that are different from thoseassisted by investment banks.

Regulatory restrictions may to some extentreduce banks’ ability to compete in the agentingmarket. In particular, the few banks possessingsection 20 subsidiaries with full debt and equityunderwriting powers may have a competitiveadvantage over banks having no such powers.

Why Do Banks Act as Agents,or Why Is the List of Bank Agents So Short?

Banks appear to enter the private placementagenting business for two reasons. First, suchbusiness can generate profitable fee income.As noted previously, almost no data are availableon agents’ fee income, costs, or profits. On thebasis of scanty knowledge about staff sizes and feerates gleaned from interviews, we speculate thatagenting is quite profitable for those banks doing ahigh volume of business. For those that assist inonly a few transactions, and thus cannot captureeconomies of scale, agenting may be only margin-ally profitable.

Second, banks may act as agents as part of astrategy of offering a broad array of corporatefinancial services, not just loans. In section 2 ofpart 2 we argued that economies of scope existbetween private placement agenting and otherlines of capital market business, such as makingloans or underwriting securities. The relationshipofficers of commercial and investment banks arethe primary sources of prospective clients forprivate placement agenting. An institution mustprovide financial services to many corporateclients to generate a flow of agenting businesssufficient to justify maintaining an agenting group.Table 15 provides evidence in support of thisassertion. It ranks the top twenty-five U.S. bankholding companies by volume of commercial andindustrial loans on the books at the end of 1991,and gives the known private placement agentingvolume (debt only) for such banks during1989–91. As tables 14 and 15 reveal, all thetop ten bank agents were among the top twenty-five holders of commercial and industrial loans,and the majority of the top lenders also acted asprivate placement agents.

159. Our database does not include private equity issuedduring 1989.

14. U.S. bank agents of private placements,1989–911

AgentDeals

(number)

Volume(millions of

dollars)

J.P. Morgan . . . . . . . . . . . . . . . . . . . . . . 289 24,299Citicorp . . . . . . . . . . . . . . . . . . . . . . . . . . . 184 14,577Chase Manhattan . . . . . . . . . . . . . . . . 301 13,621First Nat’l Bank of Chicago . . . . . 346 11,126Bankers Trust . . . . . . . . . . . . . . . . . . . . 206 10,988

Chemical Bank . . . . . . . . . . . . . . . . . . . 239 10,927Continental Bank . . . . . . . . . . . . . . . . 160 6,811Bank of America . . . . . . . . . . . . . . . . 133 6,399Manufacturers Hanover . . . . . . . . . . 100 5,768NationsBank/NCNB . . . . . . . . . . . . . 55 3,875

Mellon Bank . . . . . . . . . . . . . . . . . . . . . 94 1,012Security Pacific . . . . . . . . . . . . . . . . . . 19 598PNC Financial Corp . . . . . . . . . . . . . 2 127First Continental Bancshares . . . . 1 75First National Bank of Boston . . 5 75

Texas Commerce Bank . . . . . . . . . . 2 40Corestates . . . . . . . . . . . . . . . . . . . . . . . . 1 40Huntington National Bank . . . . . . . 1 25NBD Bank . . . . . . . . . . . . . . . . . . . . . . . 2 23Northern Trust . . . . . . . . . . . . . . . . . . . 1 13

Shawmut . . . . . . . . . . . . . . . . . . . . . . . . . 2 10First California . . . . . . . . . . . . . . . . . . . 1 7State Street Bank & Trust . . . . . . . 1 7Fleet National Bank . . . . . . . . . . . . . 1 4Banc One . . . . . . . . . . . . . . . . . . . . . . . . 1 2

Total . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,147 110,449

1. Number of deals and volume include placements of bothdebt and equity. The list of banks is surely incomplete because(1) some banks may not report agent activity to IDD, and(2) some that do report may not be identifiable as banks fromthe information in the IDD database.

Source. Computations using data from IDD InformationServices.

71

Prospective Changes in Market Shareof U.S. and Foreign Banks

As noted, the agenting market share of U.S. bankshas increased substantially during the past fifteenyears. During interviews, market participantsoffered two explanations. First, as banks have lostcommercial and industrial loan business to otherlenders or markets, they have become increasinglyinterested in selling a broad array of financialservices to corporations. Many have also reorga-nized their operations, converting loan officers intorelationship officers that operate more on theinvestment bank model of customer relationshipmanagement. This reorganization has increasedbanks’ efficiency at identifying potential clients forprivate placement agenting and at winning theirbusiness.

Second, according to some participants, invest-ment banks had placed a lower priority on their

private placement businesses during the mid-1980sand instead emphasized lines of business related tomergers and acquisitions. If true, this change mayhave provided banks with a window of competi-tive opportunity that they exploited.

Foreign banks began entering the agentingmarket only during the past few years. Theirentrance was coincident with two events: anincrease in issues of private placements by foreignborrowers and a substantial increase of foreignbanks’ share of the market for commercial andindustrial loans. Foreign bank agents may havean advantage in winning the business of foreignborrowers. Relationship officers of foreign bankscan probably market private placement agentingservices in much the same way, and with muchthe same effectiveness, as relationship officers ofU.S. banks.

Prospective changes in market share are difficultto assess. Having learned to exploit their agentingopportunities more efficiently, banks are unlikelyto lose expertise or to abandon the private market.U.S. banks may gradually lose market share iftheir share of all corporate financial servicesdeclines. They may gain market share if theirefficiency continues to increase. Foreign banksseem likely to continue to have some presencein the agenting market, but beyond that theirprospects are impossible to assess. Banks willprobably not come to dominate agenting becauseinvestment banks are intent on remainingcompetitive.

The Role of Regulation

Banks and their subsidiaries may engage inagenting without prior permission; they are subjectonly to prudential supervision that focuses onensuring disclosure of possible conflicts ofinterest. Bank holding companies and theirnonbank subsidiaries, including section 20 (securi-ties) subsidiaries, must obtain permission from theFederal Reserve Board to act as agents, and suchagents are subject to various restrictions. Seeappendix C for a detailed description of legal andregulatory restrictions on the private placementagent activities of banks.

Regulatory restrictions that focus on agentingitself do not appear so far to have imposed manycompetitive disadvantages on banks. Limits onbanks’ general securities powers, however, mayhave imposed two disadvantages. First, banks (butnot section 20 subsidiaries) are effectively pre-vented from acting as brokers or dealers in the

15. Top twenty-five bank holding companiesby commercial and industrial loans,and agenting volumeBillions of dollars

Bank holding company C&I loans

3-yearagentingvolume

Citibank . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.9 14.6Bank of America . . . . . . . . . . . . . . . . . . . 21.3 6.4Chase Manhattan . . . . . . . . . . . . . . . . . . . 18.6 13.6Manufacturers Hanover . . . . . . . . . . . . . 16.6 5.8Bank of New York . . . . . . . . . . . . . . . . . 14.4 . . .

NCNB (TX, FL, and NC) . . . . . . . . . . 13.0 3.9Chemical Bank . . . . . . . . . . . . . . . . . . . . . . 11.7 10.9First National Bank of Boston . . . . . 11.2 .1Morgan Guaranty . . . . . . . . . . . . . . . . . . . 10.5 24.3Security Pacific . . . . . . . . . . . . . . . . . . . . . 10.2 .6

Wells Fargo . . . . . . . . . . . . . . . . . . . . . . . . . 9.8 . . .Continental . . . . . . . . . . . . . . . . . . . . . . . . . . 8.9 6.8First National Bank of Chicago . . . . 8.2 11.1Mellon Bank . . . . . . . . . . . . . . . . . . . . . . . . 8.1 1.0National Westminster USA . . . . . . . . . 6.3 .3

NBD Bank . . . . . . . . . . . . . . . . . . . . . . . . . . 5.9 .2Bankers Trust . . . . . . . . . . . . . . . . . . . . . . . 5.5 11.0Union Bank . . . . . . . . . . . . . . . . . . . . . . . . . 4.8 . . .Corestates Bank . . . . . . . . . . . . . . . . . . . . . 4.5 .4Pittsburgh National Bank . . . . . . . . . . . 3.8 . . .

Marine Midland . . . . . . . . . . . . . . . . . . . . . 3.4 . . .Wachovia Bank (North Carolina) . . 3.3 . . .Manufacturers Bank . . . . . . . . . . . . . . . . 3.3 . . .Texas Commerce Bank . . . . . . . . . . . . . 3.0 .4First Union National Bank . . . . . . . . . 3.0 . . .

1. C&I loan holdings are as of December 31, 1991. Thethree-year agenting volume is for 1989–91: it includes onlyplacements of debt.

. . . Bank does not appear in the IDD database for 1989–91.Source. Computations using data from IDD Information

Services and regulatory filings.

72

secondary market for private placements becausethey cannot buy and sell restricted securities fortheir own account. As the secondary market forprivate placements has been relatively small todate and banks may act as riskless principals, thisdisadvantage has probably been minor.

Perhaps more important are Glass–Steagalrestrictions on bank underwriting of new issues ofpublic securities. Because of economies of scopebetween public underwriting and the distributionstage of private placement agenting, in somecases, public security sales forces can distributeprivate placements more efficiently than can aprivate placement agenting group. Only bankholding companies possessing section 20 subsidi-aries with full debt powers (and full equitypowers, for private equity issues) will possesssuch sales forces and be able to capture the costefficiencies. Competitive pressures will causeinvestment banks or commercial banks withsection 20 subsidiaries to win the mandate toassist most such issues.

As market participants indicated, underwritingpowers may convey another, more subtle advan-tage. Part of the service that a financial institution

typically provides is advice that leads a borrowerto issue in the private market. Such advice oftenincludes an analysis of the relative benefits ofraising funds in various of markets, including thebank loan and public security markets. The adviceof an institution capable of assisting financing inall the relevant markets is likely to be affordedmore credibility than the advice of one that canassist only in the market it is recommending.Credibility of advice is an important factor in theminds of many issuers as they choose an agent.Thus banks with full securities powers actuallyhave an advantage in this regard over investmentbanks that do not make nor syndicate loans, assuch banks can assist in three markets (loan,private, and public), while such investment bankscan assist in only two (private and public).Conversely, banks without securities powers mayin some situations be at a disadvantage.

Table 16 lists U.S. bank holding companies thathad received Federal Reserve Board permission tohave section 20 subsidiaries as of May 1992, thepowers of those subsidiaries, and the locationwithin the banking organization of the privateplacement agenting group, if any. All the banks

16. Section 20 subsidiaries of U.S. bank holding companies and the location of agents in the corporatestructure, as of May 1992

Company

Powers

Agent locationBasic 1 Full debt Debt and equity

Banc One Corp. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yes . . . . . . Sec. 20Bankers Trust NY Corp. . . . . . . . . . . . . . . . . . . . Yes Yes . . . Sec. 20Barnett Banks Inc. 2 . . . . . . . . . . . . . . . . . . . . . . . . . Yes . . . . . . . . .Chase Manhattan Corp. . . . . . . . . . . . . . . . . . . . . . Yes Yes . . . Sec. 20Chemical NY Corp. (and MHT) . . . . . . . . . . . Yes . . . . . . Bank 3

Citicorp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yes . . . Yes Sec. 20

Dauphin Deposit Corp. . . . . . . . . . . . . . . . . . . . . . Yes . . . Yes n.a.First Chicago Corp. . . . . . . . . . . . . . . . . . . . . . . . . . Yes . . . . . . Bank 3

First Union Corp. . . . . . . . . . . . . . . . . . . . . . . . . . . . Yes . . . . . . n.a.Fleet/Norstar Financial Corp. . . . . . . . . . . . . . . . Yes . . . . . . BankJ.P. Morgan & Co. . . . . . . . . . . . . . . . . . . . . . . . . . . Yes . . . Yes Sec. 20Liberty National Bancorp 2 . . . . . . . . . . . . . . . . . Yes . . . . . . . . .

NationsBank Corp. . . . . . . . . . . . . . . . . . . . . . . . . . Yes . . . . . . BankNorwest Corp. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yes . . . . . . n.a.PNC Financial Corp. . . . . . . . . . . . . . . . . . . . . . . . Yes . . . . . . BankSecurity Pacific Corp. (now BofA) . . . . . . . . Yes . . . . . . BankSouthtrust Corp. 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . Yes . . . . . . . . .Synovus Financial Corp. . . . . . . . . . . . . . . . . . . . Yes . . . . . . n.a.

1. Subsidiaries authorized to underwrite and deal in certainmunicipal revenue bonds, mortgage-related securities, commer-cial paper, and asset-backed securities.

2. As of May 1992, did not yet have permission to act asagent for private placements in the section 20 subsidiary.

3. Some fees earned on agenting of private placements bythe section 20 subsidiary were reported to the Federal Reservebut not enough to account for agenting volume listed in IDD.Anecdotal evidence indicates that these banks may have more

than one agenting group, with groups specializing, and thatsome banks with agents in the bank perform distributionsof some placements through the subsidiary sales force andbook some income in the subsidiary.

n.a. No signs of agenting activity observed in IDD or inregulatory filings.

Source. Federal Reserve Bulletin and miscellaneousregulatory filings.

73

with full securities powers have chosen to locatetheir agents (if any) in the section 20 subsidiarywhereas, to our knowledge, only one of those withpartial securities powers has chosen to do so. Thisdifference may occur for two reasons. First, theadvantages that full securities powers confer onagents may outweigh costs of the additionalregulatory restrictions that are imposed when theyare located in a section 20 subsidiary. Second, andperhaps more important, regulations limiting thefraction of revenue a section 20 subsidiary mayearn from ineligible underwriting activity encour-age the holding company to move eligible activi-ties (which include agenting of private placements)into the section 20 subsidiary to prevent thelimitations from binding.

The three largest bank agents are located insection 20 subsidiaries with full powers. However,other banks without full powers do a substantialagenting business. Thus, lack of securities powersdoes not seem to be an absolute barrier to agent-ing of private placements.

Banks as Issuers of Equity

The private placement market appears not to be animportant source of equity capital for U.S. banks.Table 17 lists the private equity issues of U.S.banks during 1990–91 that appear in the IDDInformation Services data base.160 U.S. banksissued about $2 billion of equity in the privateplacement market during 1990–91, but $1.25 bil-lion was in a single placement of convertiblepreferred stock by Citibank with a foreign inves-tor. Only twelve individual issues appear on thelist, and several of the issuers are relatively wellknown and presumably could issue in the publicmarkets without great difficulty. During this periodthe number and total volume of issues by foreignbanks was also not large (table 18). 161

The legal separation of banking and commercein the United States may be one reason banks donot issue much private equity. The Bank HoldingCompany Act of 1956, the amendments of 1970 tothat act, and Federal Reserve Board rulingsprevent nonbank corporations from owning orcontrolling banks or bank holding companies.

Acquisition of more than 5 percent of the votingstock of a bank or bank holding company requiresFederal Reserve Board approval. As appendix Bnotes, most private equity is purchased by institu-tional investors, especially pension funds, whichtend to take large blocks of individual offerings.When a purchase would amount to more than5 percent of a bank’s total capital, costs ofobtaining regulatory approval would reduce theissue’s attractiveness for purchasers.

160. No issues of equity by savings and loans appear in thedata base for this period.

161. The tables are surely an incomplete representation ofbanks’ issuance. The method by which IDD collects informa-tion (voluntary reporting by agents) favors the reporting oflarger transactions assisted by relatively high volume agents.Many small transactions likely are missed.

17. Private placements of equity by U.S. banks,1990–91

Issuer

Amount(millions of

dollars) Date

Citicorp 1 . . . . . . . . . . . . . . . . . . . . . . . . . 1,250.0 3/91Team Bank . . . . . . . . . . . . . . . . . . . . . . . 200.0 1/90Manufacturers Hanover Trust 1 . . 200.0 5/91Bank of New England . . . . . . . . . . . 150.0 10/90LaSalle National . . . . . . . . . . . . . . . . . 60.0 1/90AmeriTrust . . . . . . . . . . . . . . . . . . . . . . . 60.0 3/90

NCNB Texas National Bank . . . . 56.0 1/90SouthTrust . . . . . . . . . . . . . . . . . . . . . . . 16.3 12/91Larimar Bancorporation . . . . . . . . . 16.5 5/91North Fork Bancorporation . . . . . . 11.1 6/91First Commercial Bancorp . . . . . . 11.0 2/91Banc Plus . . . . . . . . . . . . . . . . . . . . . . . . 20.0 10/91

Total . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,050.9

1. The Citicorp and Manufacturers Hanover issues were ofconvertible preferred stock and were Rule 144A issues. Detailsof the other issues are not known.

Source. IDD Information Services.

18. Private placements of equity by foreign banks,1990–91

Issuer

Amount(millions of

dollars) Date Rule 144A

Indosuez Holdings . . . . . . 150.0 91 YesGrupo Financiero

Bancomer . . . . . . . . . . 121.0 11/91 . . .Toronto Dominion

Bank . . . . . . . . . . . . . . . 64.8 3/91 . . .Banque National

de Paris . . . . . . . . . . . . 52.5 3/90 . . .NMB Postbanken . . . . . . . 48.0 1/90 . . .Barclays . . . . . . . . . . . . . . . . . 50.0 4/90 . . .Banco Hispano

Americano . . . . . . . . . 20.0 7/90 . . .Espirito Santo Financial

Holding . . . . . . . . . . . . 15.7 7/90 YesCredito Italiano . . . . . . . . . 9.1 91 YesBanco Exterior

International . . . . . . . .8 2/90 . . .Thai Farmers Bank . . . . . .7 91 Yes

Total . . . . . . . . . . . . . . . . . . . . 532.6

Source. IDD Information Services.

74

In general, the private equity market appears toserve, directly or indirectly, mainly those start-upor other high-risk issuers that promise highreturns. Such returns compensate investors for theilliquidity and monitoring costs associated withprivate equity. As a mature and highly regulatedindustry, banking may be unattractive to suchinvestors.

Banks as Issuers of Debt

U.S. banks and bank holding companies are moreactive issuers in the private debt market. The IDDdatabase lists 174 private debt issues by themduring 1990–91, 97 unsecured and 77 secured, fora total of $8.43 billion (table 19). These issueswere about 5 percent of all private debt issues inthe database for the period.

About a third of banks’ issuance was asset-backed debt, such as mortgage-backed orreceivable-backed notes. According to remarks bymarket participants, many such issues would havebeen difficult to issue publicly. Either they were anew form of instrument (for example, somereceivable-backed bonds) or they required a buyerto engage in extensive due diligence and moni-toring (many second-mortgage-backed bonds).As noted in part 1, the private placement market isa proving ground for new types of instrument. Ifno problems surface with a new instrument (andif extensive monitoring is not required), public

market investors may eventually be willing to buyit. Thus the private placement market is importantto banks and other financial institutions as anarena for testing some of their financialinnovations.

Table 20 summarizes private debt issuance byforeign banks. The totals for unsecured debt aresimilar to those for U.S. banks, but foreign banksissued very little secured or asset-backed debt.

Banks as Buyers of Private Placements

Appendix C describes regulatory restrictions onbank purchases of private placements. To summa-rize, and ignoring minor exceptions, banks maynot buy privately placed equity, but they may buyprivate debt so long as it is booked as a loan forregulatory purposes. Bank holding companies maybuy limited amounts of equity and may buyprivately placed debt without restriction.

Almost no data on the share of new privateissues that is purchased by banks are publiclyavailable. As part of a staff report on Rule 144Adated September 30, 1991, the SEC collectedinformation on initial purchasers of sixty-nineRule 144A placements issued from April 1990through July 1991. Banks and savings and loaninstitutions (which were grouped together in thereport) were initial purchasers of only $232 mil-lion of the $6.75 billion, or 3.44 percent ofplacements in the sample. They purchased only4.16 percent of sample placements of straight debt,

19. Private placements of debt issued by U.S.banks, 1990–91

Nature of instruments

Issues

Number

Value(millions of

dollars)

Miscellaneous unsecured bonds,notes, and debentures . . . . . 97 5,268.7

Known senior . . . . . . . . . . . . . . . . 33 1,595.9Known subordinated . . . . . . . . . 5 218.5Priority unknown . . . . . . . . . . . . . 59 3,454.3

Asset-backed debt . . . . . . . . . . . . . . 77 8,429.6Mortgage-backed . . . . . . . . . . . . . 54 1,973.2Receivable-backed . . . . . . . . . . . 6 925.7Lease financing . . . . . . . . . . . . . . . 2 86.4Collateral unknown . . . . . . . . . . 15 175.6

Total issues 1 . . . . . . . . . . . . . . . . . . . . 174 8,429.6

1. One of 174 issues, for $114.6 million, was underRule 144A. Sizes of individual issues ranged from $0.9 millionto $400 million.

Source. IDD Information Services.

20. Private placements of debt issued by foreignbanks, 1990–91

Nature of instruments

Issues

Number

Value(millions of

dollars)

Miscellaneous unsecured bonds,notes, and debentures . . . . . 92 4,880.0

Known senior . . . . . . . . . . . . . . . . 25 772.2Known subordinated . . . . . . . . . 15 1,316.9Priority unknown . . . . . . . . . . . . . 52 2,790.9

Asset-backed debt . . . . . . . . . . . . . . 4 5,175.2Mortgage-backed . . . . . . . . . . . . . 2 80.0Receivable-backed . . . . . . . . . . . 1 130.2Collateral unknown . . . . . . . . . . 1 85.0

Total issues 1 . . . . . . . . . . . . . . . . . . . . 96 5,175.2

1. Five of ninety-six issues, for $98.5 million, were underRule 144A. Sizes of individual issues ranged from $0.6 millionto $306.7 million.

Source. IDD Information Services.

75

7.42 percent of placements of asset-backedsecurities, and 0.7 percent of placements ofcommon and preferred equity. Similarly, U.S.commercial banks purchased only about 3 percentof a different sample private placements (seetable 7).

The extent to which these samples are represen-tative of all private placements is not known. Asnoted in appendix C, we have some reason tobelieve that banks may be less willing to purchaseRule 144A placements because performing normalloan-underwriting due diligence for those may bemore difficult than for traditional placements.Anecdotal evidence obtained in interviews,however, confirms that banks are infrequent buyersof private placements. According to marketparticipants, most banks prefer investments ofshorter duration than the average private place-ment. Relationships of issuers with buyers ofplacements also tend to be less close than relation-ships of loan borrowers with bank lenders, andbanks like the opportunity to sell other servicesthat close relationships provide. Market partici-pants suggested that a bank is most likely to buypart of a placement when it already has a relation-ship with the issuer (since costs of due diligenceare small).

All-in returns on private placements of debtmay also be smaller for banks than for insurancecompanies. Most banks would probably swap thefixed-rate payment stream of a placement to matchthe repricing pattern of floating rate liabilities, andthe cost of such swaps may make most placementsunattractive to banks.

In summary, no major regulatory barriers tobank purchases of private placements of debtappear to exist, but various economic consider-ations may make placements less attractive thanother investments to banks. Banks seem unlikelyto become major buyers of private placements inthe near to medium term.

Banks as Competitors of PrivatePlacement Lenders

Although banks do not buy many private place-ments of debt, they do compete with buyers of

private debt. Firms that can borrow in the privateplacement market can also typically borrow in thebank loan market. The two sources of funding arenot perfect substitutes in that terms typically differ,but a sufficient cost differential can persuadeborrowers to choose the alternative with otherwiseless-attractive terms.

Available data do not support an empiricalanalysis of the extent of substitution between thetwo markets. Market participants indicated that thecompetition is greatest for borrowings at interme-diate maturities of three to seven years. Privatemarket lenders do not usually offer competitiveterms at shorter maturities, and banks are notusually competitive at longer maturities. Thisdifference in competitive advantage is most likelydue to economies of scope between lending andthe differing liabilities for the two classes oflenders, as part 1 discussed. Typical privatemarket lenders have long-term, often fixed-rateliabilities and thus can make long-term loans morecheaply than can banks, which must bear the costof swaps and other hedges. Banks are naturallymost competitive for short-term, floating-rateloans.

According to market participants, at intermedi-ate maturities, the decision between a bank loanand a placement depends on a borrower’s prefer-ence for a fixed or floating rate, on the currentcost of interest rate swaps, and on prevailing ratesin the two markets. Rates and swap costs varyenough that a borrower’s decision may be differentat different times. In other words, the maturitybeyond which a qualified borrower desiring a fixedrate is almost certain to go to the private place-ment market varies with market conditions.

No substantial change in the nature of thiscompetition or in the comparative advantages ofbanks and private market lenders appears in theoffing. The average maturity of insurance companyliabilities has grown shorter during the last decade,making the private placement market morecompetitive at shorter intermediate maturities.Differences in the maturity structures of banks andprivate market lenders are likely to persist,however, at least until the legal restrictions thatseparate banking and other forms of commerce areremoved.

76

Appendix A. Definition of Private Placement, Resales of Private Placements,and Additional Information about Rule 144A

A private placement is a security that is issued inthe United States but is exempt from registrationwith the Securities and Exchange Commission as aresult of being issued in transactions not involvingany public offering. This definition is based uponlegal criteria and not upon the economic character-istics of the financial instrument. For example, inan economic sense, commercial loans to busi-nesses and private placements of debt are similar,but in a legal sense a loan is not a security andthus is not a private placement. 162 Similarly, bankdeposits are not private placements even thoughsome are called notes and are distributed bydealers. 163

The legal status of a financial instrument can beeconomically important, however, because securi-ties fall under the jurisdiction of securities lawwhereas other instruments are covered by commer-cial law. A significant practical difference betweenthe two codes is that the Securities Act of 1933(sections 12(2) and 17) provides for civil liabilitiesand criminal sanctions for fraud in the sale of asecurity. As a consequence, buyers of securitieshave greater protection and recourse in the eventof fraud than those who make loans.

Exemptions from Registrationfor Private Placment Issuance

The Securities Act of 1933 requires that all offersand sales of securities be made through a registra-tion statement filed with the Securities andExchange Commission (SEC) unless an exemption

from registration is available. Section 3 of the actexempts from registration certain types of securi-ties, such as U.S. Treasury securities and commer-cial paper, and section 4 exempts certain securitiestransactions, such as the issuance of privateplacements and resales of registered securities.More specifically, section 4(2) of the act exempts‘‘transactions by an issuer not involving anypublic offering.’’ 164 This exemption is based onthe premise that sophisticated buyers of securitiesdo not need the protection afforded by registrationas they should be capable of obtaining informationabout the issuer on their own.

Initially issuers based exemptions under sec-tion 4(2) on interpretations by the SEC and caselaw.165 Between 1974 and 1980, the SEC adoptedthree rules to clarify the conditions for exemptionsfor offerings. In 1982, the SEC issued Regula-tion D, which provides a different formal basis fora private placement exemption by combining andexpanding these rules. Regulation D is a non-exclusive safe harbor, however, and most issuershave continued to rely on section 4(2), althoughtheir offerings satisfy most conditions of Rule 506of Regulation D.166

Rule 506 offers and sales may be for anyamount and to any number of accredited investors,but the issuer or its agent may not engage in anygeneral solicitation or advertising, the investorsmust purchase securities for their own accountsand not for distribution to the public, and nomore than thirty-five unaccredited investors maypurchase the securities. Accredited investors aremainly institutional investors, but the category alsoincludes individuals with sufficient net worth or

162. Legal definitions of security and loan are economicallyvague. For example, the Securities Exchange Act of 1934states that a security is

any note, stock, treasury stock, bond, debenture, certificate ofinterest or participation in any profit-sharing agreement . . . ,investment contract, . . . or in general, any instrumentcommonly known as a ‘‘security’’; . . . (Section 3(a)(10) ofthe Securities Exchange Act of 1934 [15 U.S.C. Section78c(a)(10)])

Thus, an instrument’s categorization is to some extent based ontradition rather than on its economic characteristics. Loans,including commercial loans, mortgage loans, consumer loans,and other instruments that fall under the jurisdiction of com-mercial law are not securities.

163. Opinions of market participants regarding the definitionof private placement vary somewhat. For example, a fewparticipants may include certain bank certificates of deposit inthe definition.

164. As private placements are not an exempted type, thecircumstances of their issuance must support the issuer’sdecision not to register the securities.

165. Case law on section 4(2) generally requires that theissuer provide investors with reasonable access to the informa-tion needed to evaluate the risks in the transaction, thatinvestors be sophisticated and capable of evaluating those risks,that investors purchase the securities for investment and notwith the intention of distributing them to the public, and thatthe issuer take steps to restrict the resale of the securitiesunless they are registered or are sold in transactions exemptfrom registration (Carlson, Raymond, and Keen, 1992).

166. A nonexclusive safe harbor specifies a set of conditionsunder which an action is legal but does not rule out thepossibility that such action would be legal under other condi-tions specified by other laws or regulations.

Rule 506 incorporates, by reference, Rules 501–503. Prelimi-nary note 3 of Regulation D states that an issuer’s failure tosatisfy all the terms and conditions of Rule 506 does notpreclude the availability of the section 4(2) exemption.

77

income.167 The issuer must reasonably believe thatunaccredited investors are capable of evaluatingthe investment. The rule does not require publicdisclosure of information about the issuer if onlyaccredited investors are involved. However, theissuer must disclose to investors that the securitieshave not been registered with the SEC and thatthey cannot be resold unless they have beenregistered or the resale transaction is exempt.Finally, the issuer must file a notice with the SECwithin fifteen days of the first sale. 168 Issuersrelying on section 4(2) for an exemption generallyattempt to conform to these conditions, except thatthey do not file a notice with the SEC.169

Some private placements are issued underexemptions other than those offered by Regula-tion D and section 4(2). For example, section 4(6)of the 1933 act exempts issues totaling less than$5 million at one time (instead of over a twelve-month period) under some circumstances, andsection 3(a)(11) exempts securities that are issuedby a resident of a state who is in business in thatstate and that are sold only to residents of thestate. The relative volumes of placements issuedunder the several possible exemptions are notknown, but remarks by market participantsindicate that most of total volume relies forexemption on section 4(2).

Exemptions from Registrationfor Resales of Private Placements

Sections 4(2) and Regulation D provide exemp-tions from registration only for issuers of privateplacements. Those wishing to resell registeredsecurities can rely on section 4(1) of the SecuritiesAct, which exempts transactions by parties otherthan issuers, underwriters, and dealers. Thisexemption is not directly available to investors

in private placements, however, because suchinvestors are technically regarded as underwriters;if they were not, private placements could beindirectly distributed to the public through resalesby investors, thereby circumventing the registra-tion requirements of the Securities Act (Davis,Polk, and Wardwell, 1990).

Exemptions for resales of private placements areprovided by the informal guidelines of theso-called section 4(11⁄2) exemption and by SECRules 144 and 144A. The section 4(11⁄2) exemp-tion combines sections 4(1) and 4(2) of theSecurities Act, neither of which by itself issufficient to support an exemption. This exemp-tion, based upon SEC no-action letters and marketpractice, assumes that resales of private place-ments are permissible without registration so longas they generally satisfy the conditions necessaryfor an issuer to justify an exemption undersection 4(2). 170 That is, if buyers are of the sameclass of investors eligible to purchase privateplacements from an issuer and if they indicatetheir intention to hold the securities for investmentpurposes, then the seller in the transaction can beviewed as not being an underwriter and thus canrely on section 4(1) for an exemption (Carlson,Raymond, and Keen, 1992). Resales based onsection 4(1⁄2) are somewhat cumbersome in thatthey involve letters of intent (to hold for invest-ment purposes) from buyers to sellers.

Rule 144 permits investors to resell privateplacements after two years from the date of thesecurities’ issuance, subject to certain limitations,and to sell without limitation after three years. 171

Rule 144A

Rule 144A, adopted by the SEC in April 1990,provides an exemption from registration forsecondary market transactions in private place-ments in which the buyer is a sophisticatedfinancial institution, defined in the rule as aqualified institutional buyer (QIB). The ruleapplies only minimal restrictions to qualifying

167. Accredited investors include (a) banks, savings andloan associations, broker–dealers, insurance companies,registered investment companies, small business investmentcompanies, and private development companies; (b) corpora-tions, partnerships, tax-exempt organizations, trusts, andemployee benefit plans; (c) individuals with net worth inexcess of $1 million or annual income over the past two yearsin excess of $200,000 (individually) or $300,000 (jointly);(d) directors, executive officers, and general partners of theissuer; and (e) any other entity in which all of the equityowners are accredited investors. Because of the burden ofdocumenting non-institutional investors’ status (and theassociated legal liability), many agents prefer to arrange privateplacements only with institutional investors.

168. See Carlson, Raymond, and Keen (1992).169. Rules 504 and 505 offer exemptions involving some-

what different restrictions for issues totaling less than $1 mil-lion and $5 million in a twelve-month period.

170. A no-action letter is a letter from the SEC staff inresponse to a letter from a party contemplating a transaction. Inthe no-action letter, SEC staff indicates that it neither agreesnor disagrees with the reasons offered for the contemplatedtransaction and that it plans to take no action to preventcompletion of the transaction.

171. One limitation is that prospective buyers must haveaccess to sufficient public information. Another is on thevolume of securities that may be sold. Also, the securities mustbe sold through a broker that has not solicited buy orders inanticipation of the sale.

78

transactions. 172 QIBs are a subset of accreditedinvestors; but, in any case, most private place-ments are purchased by QIBs, and thus the rulemakes underwriting of new issues and activesecondary trading feasible.

As defined by Rule 144A, QIBs are financialinstitutions, corporations, and partnerships thatown and invest on a discretionary basis at least$100 million of securities. 173 The scope of thisdefinition is broad enough to include the majorinvestors in private placements, such as lifeinsurance companies, pension funds, investmentcompanies, foreign and domestic banks, savingsand loan associations, and master and collectivetrusts. Besides meeting the securities requirement,banks and savings and loan associations must havenet worth of at least $25 million. In contrast toother institutional investors, broker–dealers mustown only $10 million of securities to qualify as aQIB. Moreover, if acting solely as an agent or ariskless principal, a broker–dealer does not haveto be a QIB to place the securities.

The legal underpinning that the SEC used forRule 144A is from the legislative history of theSecurities Act. 174 The SEC concluded that theCongress never considered sophisticated institu-tional investors to need the protection offered bythe registration of securities. Rather, the purposeof registration was to protect unsophisticated,individual investors. Thus, transactions in privateplacements involving only institutional investorsmay be legally separated from those involvingindividual investors. The implication of thisargument, as embodied in Rule 144A, is that byestablishing a class of private placement transac-tions confined solely to a well-defined class ofsophisticated institutional investors, such transac-tions would be exempt from registration becausethey do not constitute an offering to the public.In essence, QIBs are not considered part of thepublic; therefore, sales to them involve no public

offering, and sellers are not considered to beunderwriters. QIBs can thus rely on section 4(1)of the Securities Act, which exempts secondarytransactions not involving a dealer, underwriter, orissuer, and dealers can rely on section 4(3), whichexempts transactions conducted as a part of dealermarketmaking (SEC, 1988).

Underwriting of Private Placements

The dealer exemption indirectly carries with it theauthority to underwrite new issues of privateplacements. Generally speaking, an underwritingoccurs when an investment bank purchasessecurities from the issuer with a view to resellingor distributing those securities to other parties.Under Rule 144A, as long as the resales are toQIBs, the activity is interpreted not as an under-writing or distribution but rather as a secondarymarket transaction. Consequently, an issuer maysell the securities to an ‘‘underwriter’’ usingsection 4(2) or Regulation D for an exemptionfrom registration, and the underwriter may thenresell the securities to eligible institutional inves-tors relying on Rule 144A.175

When a Rule 144A exemption is available, thebuyer does not need to provide a letter stating thatthe purchase of the securities is for investmentpurposes, as it does with a traditional privateplacement. Rather, the buyer supplies informationconfirming its eligibility to purchase the securitiesunder Rule 144A. The buyer also must agree notto resell the securities without having an exemp-tion (Weigley, 1991).

SEC’s Reasons for Adopting Rule 144A

The SEC adopted Rule 144A for three reasons(SEC, 1988). One was to formalize marketpractice regarding resales of private placements byeliminating the uncertainty surrounding the use ofthe section 4(11⁄2) exemption. A second was toincrease the liquidity of private placements.Although a significant volume of trading thatrelied upon the section 4(11⁄2) exemption andRule 144 had ocurred during the 1980s, most

172. Rule 144A imposes three major restrictions. First, toensure that at least a minimum amount of information isprovided, the issuer of the securities involved in the transactionmust provide buyers with copies of its recent financial state-ments and basic information about its business. Also, at thetime of issuance, the securities in the private placement maynot be of the same class of securities already traded on a U.S.stock exchange or quoted on the NASDAQ system. Thepurpose of this requirement is to prevent the development of aninstitutional market in publicly traded securities. Finally, theseller of the private placements must take ‘‘reasonable’’ stepsto inform the buyer that the sale is occurring pursuant toRule 144A.

173. Bank deposit notes, certificates of deposit, loan partici-pations, repurchase agreements, and swaps are excluded.

174. The rule has not yet been tested in court.

175. Some investment banks also perform what is known asa modified 144A offering. Most of the securities are sold toinstitutional investors eligible to purchase them underRule 144A. Some sales, however, may be made to buyers thatdo not qualify under Rule 144A; in these circumstances, theinvestment banks rely on the section 4(11⁄2) exemption. Modi-fied 144A offerings are done to broaden the investor base. SeeWeigley (1991).

79

market participants felt that the conditions of thoseexemptions resulted in reduced liquidity in theprimary and secondary markets.

Finally, the SEC hoped that Rule 144A wouldmake the private placement market more attractiveto foreign corporations. The SEC was motivated inpart by the growing desire of many U.S. investorsto purchase foreign securities in the United Stateswithout incurring the costs of obtaining them inforeign markets. Even though U.S. markets offeredmany advantages to foreign issuers—including abroad investor base, the opportunity to diversifyfunding sources, virtually the only source oflong-term, fixed-rate funds, and financing fornon-investment-grade companies—foreign corpora-tions had not been a significant presence in U.S.markets, and thus their securities were not readilyavailable to domestic investors.

Foreign corporations are reluctant to issue inthe public and private markets for several reasons.In the public market, they are discouraged by theexpense and the time involved in registering thesecurities with the SEC and in satisfying thecontinuing requirements for reporting. Particularlyburdensome in this regard is the requirement thatfinancial statements conform to U.S. generallyaccepted accounting principles. Some potentialissuers are also loathe to disclose more informa-tion about their operations than that required intheir home countries. Finally, the potential forlitigation, brought by either the SEC or investors,that accompanies registration is a significantdeterrent for many foreign corporations (Engros,1992; Gurwitz, 1990).

Despite appearances, the burden of registrationand disclosure requirements may not be as great asperceived by many potential foreign issuers. Forexample, relative to domestic issuers, the SECrequires much less disclosure for foreign corpora-tions with limited business interests in the UnitedStates or limited ownership by U.S. residents,although the financial statements must, in part, stillbe reconciled with U.S. generally acceptedaccounting principles. Also, if shares of a foreigncorporation are not listed on a U.S. stock exchangeor quoted on the NASDAQ system, the SECrequires under Rule 12g3-2(b) only that thecorporation provide information made public in itshome country. 176 Nevertheless, outsiders have heldthe view that the disclosure requirements associ-ated with public offerings in the United States areburdensome.

Although privately placed securities are a meansfor foreign corporations to avoid registration andpublic disclosure altogether, prior to adoption ofRule 144A foreign corporations had not issuedextensively in the private market either. This waspartly the result of the higher yields of traditionalprivate placements. In addition, the greaterfrequency of restrictive covenants in privateplacements than found in securities issued inforeign markets and the negotiation of termscaused many potential issuers to shy away fromthe private market.

Regulation S

In adopting Rule 144A, the SEC hoped to lowerone of the barriers to foreign issuance caused bythe illiquidity of private placements. Also, byadopting Regulation S at the same time asRule 144A, the SEC facilitated sales of overseasofferings by foreign and U.S. issuers in the privateplacement market. Regulation S stipulates condi-tions under which offshore offerings and resales ofsecurities, whether issued by U.S. corporations orby foreign entities, are not required to be regis-tered with the SEC. Generally, as long as securi-ties transactions take place outside the UnitedStates and no effort is directed toward selling topersons within the United States, offshore transac-tions are exempt from registration with theSEC.177 Under Regulation S, selling activitiesinvolved in the distribution of private placementsand in resales pursuant to Rule 144A are notconsidered to be directed selling efforts. Thus, incontemporaneous offerings of securities inside andoutside the United States, for which the securitiessold in the United States are privately placed, theexemption from registration provided by Regula-tion S is preserved. Moreover, in strictly offshoreofferings, Regulation S generally allows securitiesto be sold in the United States to QIBs withoutregistration (SEC, 1990b; Morison, 1990). Takentogether, Rule 144A and Regulation S have thuseased the way for foreign corporations engaged inoffshore offerings to enter the 144A market.

176. Engros (1992), pp. 5–6.

177. Additional conditions apply to certain types of issuersor securities, generally when the transaction is likely to havesubstantial interest from U.S. investors. The primary restrictionis that the securities remain outside the United States for atleast forty days after the initial offshore offering to ensure thatan indirect offering does not take place in the United States.

80

Appendix B. The Market for Privately Placed Equity Securities

The private placement market for corporate equitysecurities consists of all equity securities notregistered with the Securities and ExchangeCommission. The private equity market overlapsto some extent with that segment of the privatedebt market that includes debt with equity kickers;indeed, private market participants specializing inthe debt or equity side often view this segment aspart of ‘‘their’’ market. For this study, however,we have included bonds with equity kickers aspart of the debt market. Even so, the privateequity market consists of a wide range offinancings—straight equity, venture capital, andequity for mergers and acquisitions—funded by avariety of investors.

Apart from the type of security, the private debtand equity markets differ in several other respects.Insurance companies do not dominate the privateequity market as they do the market for privatelyplaced debt. State and large corporate pensionfunds, endowment funds, finance companies,corporations, and individual investors are allimportant sources of funds in the private equitymarket. In addition, the average annual volume ofissuance of private debt substantially exceeds thatof private equity. Finally, a large amount offinancing in the equity market is conductedthrough private limited partnerships.

Recent Trends in Issuance

Gross issuance of private equity is a fraction ofthat of private debt. Since peaking in 1989, totalissuance of privately placed equity has fallenconsiderably (table B.1). However, the dollarvolume of private equity financing, even at itspeak, was less than 25 percent of that in theprivate debt market.

Much of the recent decline in private equityissuance reflects changes in the activity of limitedpartnerships, which raise funds through the sale ofpartnership interests and make equity investments

in companies. Sales of these interests, which arethemselves treated as private placements, appear tomake up the bulk of gross issuance by the finan-cial sector. In 1990, for example, almost 90 per-cent of total financial sector issuance was in theother financial category, which is dominated bylimited-partnership investment funds. The volumeof limited-partnership issuance appears to beparticularly sensitive to the pace of merger activityin the economy. The sharp decline in mergers andacquisitions in 1991 was reflected in the dramaticdecrease in issuance by the other financial sectorfrom $9.4 billion in 1990 to $2.6 billion in 1991.

Chart B.1 shows the relative sizes of grossissuance in the public and private equity marketssince 1986 for nonfinancial firms. Until 1991,when firms issued unprecedented amounts of newpublic equity, the private market made up ahealthy share of total gross equity financing byU.S. firms. One reason for the 1991 decline mayhave been the very high price–earnings ratios inthe public market that siphoned issuance from theprivate market. Indeed, movement on the marginbetween the public and the private equity marketsby smaller firms is likely to be much greater thanthat between the public and the private debtmarkets; whereas few public debt issues are forless than $100 million, many public equityofferings are.

Another reason for the 1991 decline may havebeen the further slowing of merger and acquisitionactivity among firms. Although firms may havepared their acquisition plans in 1991 in responseto the recession, the reluctance of domestic banksto provide senior loans for merger deals may alsohave contributed to the lower volume of equityissuance by smaller firms. Also, insurance compa-nies cut back their investments in the privateequity market in 1991. Market participants note,however, that pension funds, some foreign banks,and a few finance companies may have somewhatstepped up their presence in the market.

B.1. Gross offerings of privately placed equity by U.S. corporationsBillions of dollars

1986 1987 1988 1989 1990 1991 1992

Nonfinancial. . . . . . . . . . . . . . . . . 3.7 7.1 6.2 10.9 6.1 5.2 6.1Financial . . . . . . . . . . . . . . . . . . . . 2.9 6.0 9.1 14.8 10.6 4.9 3.8

Total . . . . . . . . . . . . . . . . . . . . . . . . 6.6 13.2 15.3 25.6 16.7 10.1 9.9

81

Investors, Issuers, and Terms of Issuance

Twenty to thirty years ago, life insurance compa-nies were the major buyers in the private equitymarket. Not only did they invest directly incompanies themselves, they also provided thelion’s share of the funds for the limited partner-ships, which first appeared in the early 1970s.These investment funds raise capital from institu-tional investors and provide private equity financ-ing. The funds target small firms with growthpotential that, at their current stage of develop-ment, are shut out of the public equity market.The funds typically have a life of five to ten years;at dissolution, the general partners (the managersof the fund) take their cut, and the remainingreturns are distributed to the limited partners (thecontributing investors). These funds have grown innumber and size over the past twenty years.

Since 1989, however, the major sources offinance in the market have been corporate andstate pension funds. The pension funds investprimarily through limited partnerships as they donot have the expertise to invest directly in compa-nies themselves. Market participants estimate thatpension funds now provide more than half of allfinancing for the partnership funds. Financecompanies, endowment funds, corporate investors,and individuals provide the rest. In the early1990s, insurance companies were only minorcontributors of new capital to the market, no doubtbecause of their reallocation of funds toward less-risky borrowers in all markets, including theprivate debt market.

The structure of private equity investments canvary significantly, from simple common stock to

preferred stock with a plethora of restrictivefeatures that allow investors to maintain controlover the company’s direction. Such featurestypically include the right to elect directors, votingrights for major transactions contemplated bymanagement, antidilution protection, and Boardcontrol at the option of investors if certainperformance criteria are not met. Market partici-pants expect more complex forms of equitysecurities to evolve in response to the growth ofmore specialized partnership funds.

Typical issuers in the market include those firmstoo small to tap the public markets for financing.Large troubled companies also often look to theprivate market to obtain equity infusions fromfinancial institutions or wealthy individuals thatwould be hard to obtain in a widely distributedpublic offering. In 1991, for example, Manufac-turers Hanover and Citicorp raised a total of$1.5 billion in private offerings of preferred stock.In February 1992, Chrysler raised $400 million ofequity in a private offering.

The traditional forms of ‘‘ exit’’ for the privateequity investor are either an initial public offering(IPO) or the sale of the company to another(typically corporate) investor. The IPO has beenparticularly popular in the past couple of years asinvestors in the public equity market have beenvery receptive to private companies seeking initialpublic equity. Market participants also point to thefuture possibility of selling private equity securi-ties in the secondary market, should that marketbecome sufficiently liquid. Here some disagree-ment has arisen over the potential effect ofRule 144A. Some feel that the rule will eventuallyincrease liquidity in the secondary market,whereas others feel that it has merely formalizedprevailing market practice and consequently willnot significantly affect the market. There is generalagreement, however, that the market will becomemore liquid over time—with or without Rule144A—simply because increasing participation ofpension funds in the market will expand theamount of funds invested in equity privateplacements.

Market participants feel that the share offinancing from pension funds could easily grow to70 or 75 percent and that this growth could be amajor factor in the growth of the share of privateequity financing that limited-partnership fundsprovide. They also expect, however, that somelarge corporate pension funds and some insurancecompanies will continue to invest directly incompanies. Pension funds may be willing toallocate increasing shares of their portfolios to

B.1. Public and private equity issuance bynonfinancial U.S. corporations1

Billions of dollars

0

10

20

30

40

50

60

30.227.9

13.7 12.9 12.3

44.848.4

1986 1987 1988 1989 1990 1991 1992

Private

Public

3.77.1 6.2

10.9

6.1 5.2 6.1

1. Figures include common and preferred equity issued inthe United States.

82

the private equity market for several reasons. First,the private market offers the opportunity to diver-sify assets outside the traditional public stockand bond markets. Second, to the extent that theprivate market is less efficient than its publiccounterpart, it offers investors the chance to makesuperior returns. Finally, by placing funds with

active investors (the limited partnerships) that takecontrolling positions in companies and monitorand sometimes change management, pension fundscan participate in the increased returns generatedby the turning around of poorly managedcompanies.

83

Appendix C. Legal and Regulatory Restrictions on Bank Participationin the Private Placement Market

Commercial banks may participate in the privateplacement market as issuers, buyers, agents, andbrokers, but they are subject to legal and regula-tory restrictions in some of these activities.

Banks as Issuers

No restrictions on issuance of privately placedsecurities apply specifically to commercial banksor bank holding companies. Like other issuers,these entities must comply with securities laws.

Summary of Bank Powers To Buy PrivatePlacements

Table C.1 summarizes bank powers to buy privateplacements, which vary with the nature of thesecurity and the type of buyer. Briefly, andignoring exceptions detailed below, banks may notbuy privately placed equity, but they may buyprivate debt, so long as it is booked for regulatorypurposes as a loan. Bank holding companies maybuy limited amounts of equity and may buyprivately placed debt without restriction.

Bank Purchases of Privately Placed Debt

Two complexities prevent a simple yes-or-noanswer to the question ‘‘May banks buy privatelyplaced debt for their own accounts?’’178 First,regulatory authority over banks operating in theUnited States is divided. National banks are

regulated primarily by the Office of the Comptrol-ler of the Currency (OCC), state-chartered banksthat are members of the Federal Reserve Systemand foreign banks are regulated by the FederalReserve, and state-chartered nonmember banks(with FDIC insurance) by the FDIC. State-chartered banks must also comply with restrictionsimposed by state authorities.

The second complication is that financialinstruments that are securities for some legal andregulatory purposes may be loans for other legaland regulatory purposes.

National bank activities involving securities aresubject to the Glass–Steagal Act (12 USC §24(7))and to the investment securities regulation of theOCC (12 CFR Part 1). Glass–Steagal specificallyauthorizes national banks to purchase for theirown account ‘‘investment securities,’’ which theOCC defines to be ‘‘a marketable obligation in theform of a bond, note, or debenture which iscommonly regarded as an investment security.[They are not] investments which are predomi-nantly speculative in nature.’’ 179 To date, the OCChas taken the position that private placements arenot ‘‘marketable’’ (because of the absence of apublic market for such securities), and thus suchsecurities are not eligible for purchase as ‘‘invest-ment securities’’ by national banks.180 However,they may be purchased and classified as loans forregulatory purposes, so long as normal loanunderwriting procedures are followed. That is,private placements may be booked as loans so

178. Banks may buy private placements for trust and othermanaged accounts, so long as they follow policies to ensureavoidance of conflicts of interest.

179. 12 CFR §1.3(b).180. The OCC Handbook for National Bank Examiners

(§203.1, p. 1) states that a security ‘‘is marketable if it may besold quickly at a price commensurate with its yield andquality.’’ Based solely on this definition, many privately placedsecurities would qualify as marketable, but in practice the OCCdoes not consider them such.

C.1. Bank powers to buy private placements

Nature of security Banks 1 Bank holding companies

Private debt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . May purchase, but must place in loanaccount and follow underwritingprocedures.

May purchase without restriction.

Private equity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Generally may not purchase, except forsome specific types of stock.

May purchase, in limited quantities.

1. This table provides only a rough summary of banks’powers. Limitations on bank powers vary by chartering

authority and insured status, that is, according to the identity ofthe regulators.

84

long as the creditworthiness of the borrower(issuer) is evaluated and documented.181

Regulatory treatment of securities as loans isnot unusual. Banks have long been permitted toclassify as loans their purchases of commercialpaper, which is commonly recognized as asecurity. The OCC has explicitly stated that theclassification of an instrument as a security for thepurposes of securities law does not necessarilymean it must be a security for purposes of bankinglaw.182

The Federal Reserve Act makes state-charteredmember banks subject to Glass–Steagal restrictionson securities activities. 183 The Federal Reserve’spractice has to date generally followed that of theOCC in matters of investment security regula-tion.184 Thus state-chartered member banks alsomay not book private placements as investmentsecurities, but they may classify private place-ments as loans if proper underwriting proceduresare followed. In the same vein as OCC writingson this subject, the Federal Reserve Board’sCommercial Bank Examination Manual states:

Occasionally, examiners will have difficulty distinguish-ing between a loan and a security. Loans result fromdirect negotiations between a borrower and a lender.A bank will refuse to grant a loan unless the borroweragrees to its terms. A security, on the other hand, isusually acquired through a third party, a broker ordealer in securities. Most securities have standardizedterms which can be compared to the terms of othermarket offerings. Because the terms of most loans donot lend themselves to such comparison, the averageinvestor may not accept the terms of the lendingarrangement. Thus, an individual loan cannot beregarded as a readily marketable security. 185

The securities investments of state nonmemberbanks are subject to the restrictions of relevantstate laws, rather than to those of the Glass–Steagal Act. Banks with FDIC deposit insurance(that is, almost all banks) are also subject to FDICregulations. The FDIC has generally followedOCC practice in this area, with one possibleexception, but recent legislation may lead to some

departure from OCC practice. A recent amend-ment to the FDI Act (12 USC 1831a), which waspart of the FDIC Improvement Act (FDICIA),prohibits any insured state bank from engaging asprincipal in any activity that is not permissible fora national bank unless the state bank meets itscapital requirements and the FDIC consents. Since‘‘activity’’ includes making any investment, aninsured state bank is now able to ask the FDIC forpermission to place in its investment account debtsecurities that do not qualify as investmentsecurities.

The other possible departure from OCC practiceinvolves the fact that some states grant banksauthority to make ‘‘leeway’’ investments, that is,to buy limited quantities of certain securities thatare otherwise ineligible. Except for securities indefault, the FDIC will not criticize such invest-ments so long as they are permitted by applicablestate law, the total of all such investments doesnot exceed 10 percent of equity capital andsurplus, and the investments have been approvedby the bank’s board of directors or trustees asleeway securities. 186

The types of securities that qualify as leewaysecurities vary by state, but those acceptable to theFDIC as leeway investments are limited mainly tosecurities that state or local governments issue orguarantee, some of which may also qualify asprivate placements. We speculate that if anyprivate placements of debt have been purchased bybanks under this authority, the quantity has beeninsignificant.

The distinction between buying a privateplacement for an investment security account anda loan account has little economic meaning.187

Since due diligence similar to that in commercialloan underwriting is the norm in the privateplacement market, in practice commercial banksappear not to be generally restricted from purchas-ing private placements. One perhaps unintendedeffect of current regulations is possible, however.Some private placement agents have in recentyears developed distribution channels that employ

181. Loan-to-one-borrower limits must also be observed.182. See especially OCC Interpretive Letter, No. 329

(3/4/85). Also see the passages, ‘‘For purposes of banking law,an [instrument] may be a loan, a security or an investmentsecurity. Those distinctions are made on a case-by-case basis’’(OCC Interpretive Letter, No. 182 [(3/10/81]), and ‘‘creditanalysis and risk are what typifies [a] transaction as a bankingpractice’’ (OCC Interpretive Letter, No. 338 [5/2/85]).

183. 12 USC §335.184. See Fein (1991), p. 4-3, for a discussion of a case in

which the Federal Reserve did not agree with the OCC.185. Section 203.1, p. 1 (1984 edition).

186. See the FDIC’s DOS Manual of Examination Policies,Revision 9-91, p. 3.2-13. See also Fein (1991), p. 4-15.

187. Different limits on volume of loans to one borrower doapply, however, for national and state-chartered member banks.The limit for loans is 15 percent of capital and surplus,whereas that for securities is 10 percent. Insured state nonmem-ber banks are not subject to any federal loan-to-one-borrowerlimit, and limits imposed by states vary. Thus a requirementthat private placements be booked as loans does not impose amore stringent limit under federal guidelines but may do sounder some state guidelines.

85

public security sales forces. These channels aremost often used for the placements, especiallyRule 144A placements, of highly rated or well-known borrowers. Some of these channels circum-scribe the ability of buyers to perform the in-depthdue diligence that is normal for traditional privateplacements. Under current regulations, banks maybe restricted from buying such placements becausethey may be unable to provide the underwritingdocumentation necessary to their classification asloans. If such restriction actually occurs, currentregulations will unintentionally discourage bankinvestments in some higher-quality and more-liquid private placements and permit the purchaseof riskier and less-liquid placements.

Bank Purchases of Privately Placed Equity

To a first approximation, national and statemember banks may not purchase equity for theirown accounts. The exceptions are numerous,however. Those for state member banks (regulatedby the Federal Reserve) are detailed in table C.2,which is a copy of table 1 in section 203.1 of theBoard’s Commercial Bank Examination Manual.The manner of issue of the equity securities,public or private, is immaterial.

FDICIA extended the Glass–Steagal limitationson equity investments to state-chartered nonmem-ber banks: Insured state banks are now generallyprohibited from acquiring or retaining any equitysecurity that is not permissible for a national bank.Some exceptions exist, however, for certain kindsof equity investments and for banks that had madesuch investments during a given period, providedthat the FDIC does not object and a capitallimitation is not breached.188

Bank Holding Company Purchasesof Privately Placed Debt and Equity

Bank holding companies, which are regulated bythe Federal Reserve Board under the BankHolding Company Act of 1956, may purchasedebt securities, whether publicly issued or pri-vately placed. Regulation of equity purchases alsodoes not regard the manner of issuance. Holdingcompanies may purchase up to 5 percent of thevoting stock of any nonbank corporation withoutprior Board approval, though such investments

must be passive. They may purchase up to24.9 percent of a nonbank firm’s total capital,including subordinated debt and nonvoting stock;again, the investment must be passive.

Banks as Agents

Current law and regulation allows banks to act asagents for issuers of private placements, butrestrictions on their activities differ somewhatdepending on whether the activity is performed ina bank, in a securities affiliate of a bank holdingcompany (section 20 subsidiary), or in anothernonbank subsidiary of a bank holding company.Fein (1991) discusses the legal history of banksecurities powers, which involved legal challengesby the Securities Industry Association andothers. 189

Agent activities have few restrictions when theyare performed in a bank.190 A bank need notobtain permission to act as an agent. However, itshould structure its relationships with issuers sothat it acts as adviser (without power to committhe issuer formally) rather than as agent (with suchpower). (In this study, the word agent refers toboth agents and advisers.) Issues for whichnational and state-chartered member banks act asagent may be placed in the bank’s own accounts,in trust accounts or other managed accounts, orwith other affiliates or the parent holding company(if these exist). 191 A bank’s main obligation, andthe main focus of examinations of bank privateplacement agent activities, is to fully discloseinformation about the bank’s interests to all partiesinvolved in a transaction. This disclosure permitsparties to assess the risk flowing from any poten-tial conflicts of interest on the part of the bank. Inparticular, but not exclusively, the bank mustdisclose any lending relationships with issuers orwith potential or actual buyers of the securities

188. See FDIC rule adopted October 27, 1992.

189. See OCC Interpretive Letter, No. 32 (December 9,1977) for a relatively early explicit ruling that banks may actas agents.

190. OCC Interpretive Letter, No. 463 (December 27, 1988)states that ‘‘national banks may use an operating subsidiary toconduct any activity which is permissible for the bank,’’ thoughthe letter also notes that restrictions flowing from bank holdingcompany law may apply. If no holding company restrictionsapply, an operating subsidiary faces the same (minimal)restrictions on private placement agent activities that a bankdoes.

191. The FDIC strongly discourages such purchases:‘‘Policy constraints should prohibit placing private issues withfunds that the bank manages in a fiduciary capacity, especiallywhen the issuer is a bank loan customer.’’ DOS Manual ofExamination Policies, Revision 9-91, section 3.2, p. 26.

86

and the nature of any compensation it will receivefor assisting the transaction.

Bank holding companies wishing to serve asagents of private placements either in the holdingcompany or in a nonbank affiliate must first obtainthe Board’s permission. When it is so located, theactivity is subject to additional restrictions:

• An issue may be placed with nonbankaffiliates only up to 50 percent of the amount ofthe placement, and no issue may be placed withbank affiliates.

• Loans that are the functional equivalents ofpurchasing for the account of an affiliate and loansto cover unsold portions of an issue cannot be

C.2. Permitted stock holdings by member banks

Type of stock Authorizing statute and limitation

Federal Reserve Bank Sections 2 and 9—Federal Reserve Act (12 USC 282 and 321), and Regulation 1(12 CFR 209)—Subscription must equal 6 perent of the bank’s capital and surplus,3 percent paid in.

Safe deposit corporation 12 USC 24—15 percent of capital and surplus.

Corporation holding bank premises Section 24A—Federal Reserve Act (12 USC 371(d))—100 percent of capital stock.Limitation includes total direct and indirect investment in bank premises in any form(e.g., loans, etc.). Maximum limitation may be exceeded with permission of theFederal Reserve Bank for state member banks and the Comptroller of the Currencyfor national banks.

Small business investment company 15 USC 682(b) (Section 302(b)—Small Business Investment Act of August 21,1958)—Banks are prohibited from acquiring shares of such a corporation if, uponmaking the acquisition, the aggregate amount of shares in small business investmentcompanies then held by the bank would exceed 5 percent of its capital and surplus.

Edge and Agreement corporations,and foreign banks

Sections 25 and 25(a)—Federal Reserve Act (12 USC 601 and 618)—The aggregateamount of stock held in all such corporations may not exceed 10 percent of themember bank’s capital and surplus. Also, the member bank must possess a capitaland surplus of $1 million or more prior to acquiring investments pursuant toSection 25.

Banking Service Corporation 12 USC 1861 and 1862 (Section 2(a) of the Bank Service Corporation Act of1958)—10 percent of capital and surplus. Limitation includes total direct and indirectinvestment in any form. No insured bank may invest more than 5 percent of its totalassets.

Federal National Mortgage Association 12 USC 1718(f)(Section 303(f), National Housing Act of 1934)—No limit.

Bank’s own stock 12 USC 83—Shares of the bank’s own stock may not be acquired or taken assecurity for loans, except as necessary to prevent loss from a debt previously con-tracted in good faith. Stock, so acquired, must be disposed of with six months of thedate of acquisition.

Corporate stock acquired throughdebts previously contracted (DPC)

Case law has established that stock of any corporation may be acquired to preventloss from a debt previously contracted in good faith. See Oppenheimer v. HarrimanNational Bank & Trust Co. of the City of New York, 301 US 206 (1937). However, ifthe stock is not disposed of within a reasonable time period, it loses its status as aDPC transaction and becomes a prohibited holding under 12 USC 24(7).

Operations subsidiaries Permitted if the subsidiary is to perform, at locations at which the bank is authorizedto engage in business, functions that the bank is empowered to perform directly(12 CFR 250.141).

State Housing Corporation incorporatedin the state in which the bank is located

12 USC 24—5 percent of its capital stock, paid in and unimpaired, plus 5 percent ofits unimpaired surplus fund when considered together with loans and commitmentsmade to the corporation.

Agricultural Credit Corporation 12 USC 24—20 percent of capital and surplus unless the bank owns over 80 percent.No limit if the bank owns 80 percent or more.

Government National MortgageAssociation

12 USC 24—No limit.

Student Loan Marketing Association 12 USC 24—No limit.

Bankers’ banks 12 USC 24—10 percent of capital stock and paid in and unimpaired surplus. Bank-ers’ bank must be insured by the FDIC, owned exclusively by depository institutions,and engaged solely in providing banking services to other depository institutions andtheir officers, directors or employees. Ownership shall not result in any bank’sacquiring more than 5 percent of any class of voting securities of the bankers’ bank.

Source. Commercial Bank Examination Manual, Section 203.1 (March, 1984), pp. 2–3.

87

made to issuers. Holding companies must be ableto document that any credit to an issuer wasextended under different terms, for differentpurposes, and at different times than were thesecurities being placed.

• Loans cannot be made to the issuer to coverprincipal and interest payments until at least threeyears have passed since issuance.

• Issues may not be placed with accountsmanaged by affiliated bank trust departments norwith other accounts advised or managed byaffiliates.

• No lines of credit or other guarantees may beprovided to support privately placed issues advisedby the affiliate. For example, no affiliate of theholding company may provide a backup line ofcredit to support a placement of commercial paperadvised by a nonbank affiliate.

• The notes for an issue must be in denomina-tions of at least $100,000.

• All lending relationships of the consolidatedholding company with the issuer must be disclosedto actual and potential purchasers, and no invest-ment advice may be provided to purchasers.

• Securities may be placed only with accreditedinvestors.

• The securities may not be registered.• The issue must comply with relevant securi-

ties laws, for example, there can be no publicsolicitation nor offering.

• When acting as agent for a private placement,section 20 subsidiaries must comply with someadditional restrictions that apply to public under-writings. Non–section 20 subsidiaries need notdo so.

Several foreign banks have received permissionto conduct agent activities in securities subsidi-aries, subject to the above restrictions, somedetails of which differ because of the banks’

foreign status. U.S. branches of foreign banks andU.S. banks owned by foreign banks may also actas agents in the private market, in which case theyare subject to the regulations of the relevant bankregulator.

These restrictions are more stringent than thosefaced by banks. For example, in its No ObjectionLetter 87-3 (March 24, 1987) and its InterpretiveLetter 496 (December 18, 1989), the OCCpermitted banks to act as agents in the privateplacement of registered securities of their holdingcompanies or subsidiaries. Banks may providelines of credit to issuers they advise and mayplace advised issues with affiliates and in trust ormanaged accounts, subject to guidance fromregulators.

Bank Activities in the Secondary Marketfor Private Placements

In general, banks and non–section 20 subsidiariesmay act as traders of securities, regardless of thenature of their issuance, but not as brokers ordealers. Section 20 subsidiaries may act as brokersor dealers. In practice, the ability of banks toactively trade private placements is limitedbecause such placements must be booked as loans(as noted above) and normal loan underwritingstandards apply. The requirement that analyses ofcreditworthiness be performed may not be asubstantial hindrance in the secondary market fortraditional private placements, in which mostbuyers intend to hold purchases for some time andfor which due diligence is the norm. However,extensive credit analyses are more unusual in theRule 144A secondary market, which operatesmuch like the public bond market. Thus, exceptfor section 20 subsidiaries, banks may havedifficulty participating in this market.

88

Appendix D. Historical Data on Issuance of Private Debt

Table D.1 displays time series of gross issuance ofprivate placements and public issues of debt since1935, drawn from historical databases at theFederal Reserve Board. For the private placementsthe sources of the data are as follows: Before1978, the source is the Securities and ExchangeCommission. In 1978, the Federal Reserve begancomputing estimates, for use in the flow of fundsaccounts, of gross private issuance based on dataobtained from IDD Information Services. Thetable reports these estimates for 1978 and there-after. Data on private issuance have been unavail-able from the SEC since 1982. For the years thatboth series are available, the series do not agree.

That derived from IDD data is larger than thatestimated by the SEC, especially in the early1980s. We do not claim that the series in table D.1are accurate (indeed, they are almost certainlyinaccurate as some private transactions are neverreported anywhere), nor do we have any informa-tion about systematic variation in estimation errorsover time.

The public and private gross issuance series areplotted in chart D.1 on a log scale, and the ratio ofprivate to public issuance is plotted in chart D.2.The reasons for the variations in the relativeimportance of the private and public markets are asubject for future research.

D.1. Gross issuance of publicly offered andprivately placed bonds

Billions of dollars, log scale

1

–0+

1

2

3

4

5

6

1940 1950 1960 1970 1980 1990

Public

Private

Sources: SEC and IDD Informational Services.

D.2. Ratio of privately placed bonds tototal bonds

Percent

20

40

60

80

1940 1950 1960 1970 1980 1990

Sources: SEC and IDD Informational Services.

89

D.1. Corporate bond issuance: public offerings and private placements, 1935–921

Billions of dollars, except as noted

Year Total Public

Private

AmountPercentage

of total

1935 . . . . . . . . . . . . . . . . . . . . . . 2.50 2.12 .39 15.41936 . . . . . . . . . . . . . . . . . . . . . . 4.40 4.03 .37 8.41937. . . . . . . . . . . . . . . . . . . . . . . 1.95 1.62 .33 16.81938 . . . . . . . . . . . . . . . . . . . . . . 2.73 2.04 .69 25.31939 . . . . . . . . . . . . . . . . . . . . . . 2.68 1.98 .70 26.2

1940 . . . . . . . . . . . . . . . . . . . . . . 3.15 2.39 .76 24.11941 . . . . . . . . . . . . . . . . . . . . . . 3.20 2.39 .81 25.41942 . . . . . . . . . . . . . . . . . . . . . . 1.33 .92 .41 30.91943 . . . . . . . . . . . . . . . . . . . . . . 1.36 .99 .37 27.11944 . . . . . . . . . . . . . . . . . . . . . . 3.45 2.67 .78 22.71945 . . . . . . . . . . . . . . . . . . . . . . 5.86 4.86 1.00 17.11946 . . . . . . . . . . . . . . . . . . . . . . 6.75 4.88 1.86 27.61947 . . . . . . . . . . . . . . . . . . . . . . 7.18 5.04 2.15 29.91948 . . . . . . . . . . . . . . . . . . . . . . 8.98 5.97 3.01 33.51949 . . . . . . . . . . . . . . . . . . . . . . 4.78 2.32 2.45 51.4

1950 . . . . . . . . . . . . . . . . . . . . . . 4.92 2.36 2.56 52.01951 . . . . . . . . . . . . . . . . . . . . . . 5.57 2.25 3.33 59.71952 . . . . . . . . . . . . . . . . . . . . . . 7.60 3.65 3.96 52.01953 . . . . . . . . . . . . . . . . . . . . . . 7.08 3.86 3.23 45.61954 . . . . . . . . . . . . . . . . . . . . . . 7.49 4.01 3.48 46.51955 . . . . . . . . . . . . . . . . . . . . . . 7.42 4.12 3.30 44.51956 . . . . . . . . . . . . . . . . . . . . . . 8.00 4.23 3.78 47.21957 . . . . . . . . . . . . . . . . . . . . . . 9.96 6.12 3.84 38.51958. . . . . . . . . . . . . . . . . . . . . . . 9.65 6.33 3.32 34.41959 . . . . . . . . . . . . . . . . . . . . . . 7.19 3.56 3.63 50.5

1960 . . . . . . . . . . . . . . . . . . . . . . 8.08 4.81 3.28 40.51961 . . . . . . . . . . . . . . . . . . . . . . 9.42 4.70 4.72 50.11962 . . . . . . . . . . . . . . . . . . . . . . 8.97 4.44 4.53 50.51963 . . . . . . . . . . . . . . . . . . . . . . 10.86 4.71 6.14 56.61964 . . . . . . . . . . . . . . . . . . . . . . 10.87 3.62 7.24 66.71965 . . . . . . . . . . . . . . . . . . . . . . 13.72 5.57 8.15 59.41966 . . . . . . . . . . . . . . . . . . . . . . 15.56 8.02 7.54 48.51967 . . . . . . . . . . . . . . . . . . . . . . 21.96 14.99 6.96 31.71968 . . . . . . . . . . . . . . . . . . . . . . 17.38 10.73 6.65 38.31969 . . . . . . . . . . . . . . . . . . . . . . 18.35 12.73 5.61 30.6

1970 . . . . . . . . . . . . . . . . . . . . . . 29.03 24.37 4.66 16.01971 . . . . . . . . . . . . . . . . . . . . . . 30.06 23.29 6.77 22.51972 . . . . . . . . . . . . . . . . . . . . . . 26.13 17.43 8.71 33.31973 . . . . . . . . . . . . . . . . . . . . . . 21.05 13.24 7.80 37.11974 . . . . . . . . . . . . . . . . . . . . . . 32.06 25.90 6.16 19.21975 . . . . . . . . . . . . . . . . . . . . . . 43.22 32.58 10.64 24.71976 . . . . . . . . . . . . . . . . . . . . . . 42.67 26.45 16.22 38.01977. . . . . . . . . . . . . . . . . . . . . . . 45.22 24.07 21.15 46.81978 . . . . . . . . . . . . . . . . . . . . . . 39.57 19.82 19.76 49.91979 . . . . . . . . . . . . . . . . . . . . . . 44.05 25.81 18.24 41.4

1980 . . . . . . . . . . . . . . . . . . . . . . 55.20 41.62 13.57 24.61981 . . . . . . . . . . . . . . . . . . . . . . 52.54 38.33 14.21 27.01982 . . . . . . . . . . . . . . . . . . . . . . 59.58 44.28 15.30 25.71983 . . . . . . . . . . . . . . . . . . . . . . 68.27 47.14 21.13 30.91984 . . . . . . . . . . . . . . . . . . . . . . 110.41 74.08 36.32 32.91985 . . . . . . . . . . . . . . . . . . . . . . 165.76 119.56 46.20 27.91986 . . . . . . . . . . . . . . . . . . . . . . 313.36 232.60 80.76 25.81987 . . . . . . . . . . . . . . . . . . . . . . 299.75 207.68 92.08 30.71988 . . . . . . . . . . . . . . . . . . . . . . 328.86 201.16 127.70 38.81989 . . . . . . . . . . . . . . . . . . . . . . 297.12 179.70 117.42 39.5

1990 . . . . . . . . . . . . . . . . . . . . . . 275.77 188.78 86.99 31.51991 . . . . . . . . . . . . . . . . . . . . . . 361.97 287.04 74.93 20.71992 . . . . . . . . . . . . . . . . . . . . . . 443.55 377.69 65.86 14.8

1. Details may not sum to totals because of rounding.

90

Appendix E: A Review of the Empirical Evidence on Covenants and Renegotiation

The Nature of Covenantsin Private Placements

Laber (1992) has examined the frequency ofdifferent types of covenants in private placements.From a sample of twenty-five private placementsissued between 1989 and mid-1991, he found thatall had covenants that (1) related to mergers andconsolidations, (2) restricted the sale of assets,(3) restricted liens given to other creditors, and(4) restricted either payments (dividends, stockrepurchases, and payments to preferred stock) ornet worth. Eighty-eight percent had covenants thatset a maximum leverage ratio; 72 percent hadcovenants that restricted investments; and 48 per-cent had covenants that required a minimuminterest coverage ratio. Laber did not report on thefrequency of covenants related to working capital,although such covenants do appear in privateplacements. 192

According to market participants, most financialcovenants in private placements are incurrencecovenants; occasionally one or two maintenancecovenants may be included, especially when theseare designed to match maintenance covenants inother debt of the issuer, such as bank loans.

The Relationship between Covenant Tightnessand Issuer Quality

Our discussion with market participants indicatedthat the number and tightness of financial cove-nants in private placements, just as in other debtmarkets, are a function of the quality of the issuer.Securities purchase agreements for lower-qualityissuers often include many of the financialcovenants seen by Laber, and the covenants aretight in that stipulated minimum values for ratiosare close to current values. Contracts for moder-ately risky issuers often include only one or twofinancial covenants with minimum values setfurther from current values. Highly rated issuesusually have no financial covenants, althoughA-rated issues may have a debt-incurrencecovenant if their maturity is beyond seven years.

In contrast, Hawkins (1982) found in a sampleof fifty securities issued in the mid-1970s norelationship between the restrictiveness of cove-nants in private placements and the quality of theissuer, for three types of financial covenants:working capital restrictions, cash payout restric-tions, and debt restrictions. One reason forHawkins’s findings is that the market may havechanged in the 1980s. Indeed, many observershave argued that covenants in private placementshave become less restrictive over the past decade(Asquith and Wizman, 1990; Brealey and Myers,1991; Brook, 1990; and McDaniel, 1988). Ifrestrictiveness has decreased more for higher-quality issues than for lower-quality issues, thenthe difference would be accounted for. Along thisline, Asquith and Wizman found for public bondsthat covenant restrictions on debt financing anddividends fell more for A-rated bonds than forlower-rated bonds. Fitch (1991), however, statesthat the decline in the strength of public bondindentures is characteristic of the public marketand not the private placement market. The declineis also inconsistent with the interview results ofBrook (1990) and the market descriptions ofChemical Bank (1992) and Travelers InsuranceCompany (1992). 193

Cross-Market Differences in Covenants

Market participants indicated that bank loanscontain roughly the same types of covenants asthose in the private placement market, with twodifferences. First, financial covenants in bank loansare typically maintenance covenants, whereas mostcovenants in private placements are incurrencecovenants. Second, bank loan covenants areusually tighter. In some cases, we were informed,private placement covenants are set by looseningthe covenants in an issuer’s existing bank loan.

Although the types of financial covenants areroughly the same in private placements and bank

192. Engros (1992) and Hawkins (1982) report that cove-nants specifying working capital ratios are common in privateplacements. Vachon (1992a) also lists the maintenance ofworking capital above some minimum level in his taxonomy ofcovenants in private placements. Travelers (1992), however,reports that working capital requirements are common in bankloans but not in private placements.

193. In its description of ‘‘representative terms and cove-nants,’’ Chemical Bank (1992) notes that covenant tightnessdecreases as issue quality increases. It reports that a representa-tive A-rated issue would tend to have a minimum net worthcovenant, lien limitation covenant, and change of control/con-solidation, merger, or sale covenant. A representative BBB-rated issue would tend to have all of the above plus a restrictedpayments covenant and maximum long-term debt covenant.A representative BB-rated issuer would tend to have all of thecovenants in a representative BBB-rated issue plus a covenantrestricting an interest coverage ratio or fixed charge coverageratio.

91

loans, evidence points to subtle differences in theway they are implemented. Travelers (1992)observes that bank loan covenants tend to reflect alending philosophy different from that motivatingprivate placement covenants. It argues that banks,as short-term lenders, emphasize liquidity orworking capital. This approach is reflected in theinclusion of working capital covenants. Banks alsoappear to be more sensitive to the relation betweentotal liabilities and net worth; therefore, bankloans tend to restrict the ratio of total liabilities tonet worth. In contrast, private placement investorstend to emphasize the importance of a firm’slong-term assets and its long-term debt becausethey are long-term investors. As a result, accord-ing to Travelers (1992), private placements seldominclude working capital covenants, and they tendto restrict long-term liabilities rather than totalliabilities. 194

Empirical Evidenceon Cross-Market Differences

Several studies that focused on the differencesbetween covenants in privately placed debt andthose in public debt indicate that private placementcovenants are tighter than public bond covenants.Smith and Warner (1979) observed this fact froma 1971 American Bar Foundation study of bondindentures. Laber’s (1992) conclusion that cove-nants in the private placement market are morerestrictive than those in the public market wasbased on a comparison of his private placementdata with the findings of other studies of cove-nants in the public market. DeAngelo, DeAngelo,and Skinner (1990) also presented evidence thatthe covenants of private debt contracts are tighterthan those of public debt contracts. El-Gazzar andPastena (1990) had similar results and alsoreported that private placement covenants tend tobe tighter than those in large syndicated bankloans. This finding, however, probably reflects thepredominance of large syndicated bank facilities intheir sample as opposed to loans to the smallerbank-dependent borrowers that are the focus ofour comparison.195 Because they did not appropri-ately control for issuer size and quality in their

analysis, their results are not inconsistent with thegeneral proposition that bank loan covenantsassociated with bank-dependent borrowers aretighter than private placement covenants, whichare in turn tighter than public bond covenants,controlling for the size and quality of the bor-rower.

From interviews with staff of investment banks,rating agencies, corporate general counsel depart-ments, and law firms, Brook (1990) found thatprivate placements typically had tighter covenantsthan public bonds had but that the pattern differedby type of covenant. For three classes ofcovenants—disposition of assets or maintenance ofcapital, limitations on debt, and restrictions ondividends—private placements were the mostrestrictive, non-investment-grade public bondswere somewhat less restrictive, and investment-grade public bonds were the least restrictive.

Brook also found that restrictions on invest-ments were not present in public bonds but werecommon in private placements. However, negativepledge clauses and restrictions on sale-leasebacktransactions were common to all three.196 Anti-merger provisions were present in all three typesof securities. Some interviewees in the Brookstudy, however, indicated that antimerger cove-nants tended to be somewhat stronger in thenon-investment-grade market and stronger yet inthe private market. He found that in the publicbond market the typical anti-merger provision didnot prevent a merger but only required that theacquirer assume the acquiree’s debt.

The Value of Covenant Protection

Another empirical issue is the value of covenantprotection. Unfortunately, data limitations and thelack of market prices make examination of thisissue problematic for the bank loan and the pri-vate placement markets. However, several studieshave been conducted on the degree of protection

for borrower size and quality when comparing covenanttightness across markets. The non-highly-leveraged transaction(non-HLT) syndicated bank loan market provides short-termand intermediate-term credit to large companies, often in theform of unfunded loan commitments. These are used forvarious purposes, including working capital, takeovers, recapi-talization, leveraged buyouts, and commercial paper backups.These loans are often sold in the primary and secondary loansales market. Much of the non-HLT volume in this marketcomes from Fortune 500 borrowers, which obtain bank loanfacilities with minimal covenant constraints.

196. A negative pledge clause restricts the issuer fromconveying a lien on company assets to other creditors.

194. This finding is somewhat inconsistent with othersources, which indicate that working-capital-related covenantsare characteristic of private placements (Engros, 1992; Hawk-ins, 1982; and Vachon, 1992a).

195. This statement suggests the importance of distinguish-ing between the syndicated bank loan market and the so-calledmiddle market for commercial bank loans and of controlling

92

provided by covenants in public bonds for lever-aged buyouts. These studies focus on the relationbetween covenants and bondholder returns inleveraged buyouts (LBOs). (A loss, or negativereturn, to existing bondholders due to an LBOindicates a lack of covenant protection.) Marais,Schipper, and Smith (1989) found that existingbondholders did not suffer losses in LBOs.However, their results contrast with anecdotalevidence, such as the RJR–Nabisco leveragedbuyout, as well as other academic studies, such asWarga and Welch (1990). Asquith and Wizman’s(1990) results are particularly relevant: They foundthat, although existing bondholders on averageincurred significant losses in LBOs, these losseswere related to the strength of the covenants. Theyfound that ‘‘bonds with strong covenant protectiongain value, whereas those with weak or noprotection lose value.’’ Similar results wereobtained by Cook, Easterwood, and Martin (1992).Crabbe (1991b) analyzed the value of covenants inthe public market ex ante by examining thepricing of super poison puts. 197 He found thatpublic bonds with super poison put covenants paida lower yield than those without. It is difficult,however, to extrapolate from these studies of thevalue of event risk protection in the public marketto the value of credit quality protection providedby covenants in the private market. Of course, theubiquity of covenants in private placements andcommercial bank loans itself suggests that they arevalued in these markets.

Empirical Evidence on Renegotiation

While virtually all sources, including marketparticipants interviewed for this study, agree on

the ranking of markets with respect to covenanttightness and renegotiation, some evidencesuggests that public bond covenants can provide ameasure of protection and that the cost of renego-tiation in the public market is not necessarilyprohibitive. Kahan and Tuckman (1992) studiedcovenant renegotiation in public bonds. Theyfound a sample of sixty-nine firms that sought torenegotiate bond covenants during 1988 and 1989.The authors argue that their finding so many firmsseeking renegotiation is inconsistent with theassumption of most researchers (for example,Berlin and Loeys, 1988; Bulow and Shoven, 1978;and Lummer and McConnell, 1989) that renegotia-tion is limited to the information-intensive markets(for example, commercial bank loans and privateplacements) and is prohibitively expensive in thepublic market. Covenant renegotiation in thepublic bond market involves the issuer’s sending a‘‘consent solicitation’’ to each bondholder request-ing an alteration in one or more of the covenantsin the bond indenture agreement. Except for analteration of interest and principal provisions, mostindenture agreements require a two-thirds majorityof the outstanding face value of the bond issue.(The Trust Indenture Act of 1939 requires theconsent of all bondholders when principal andinterest are to be modified.) Kahan and Tuckman(1992) found that most solicitations are ultimatelysuccessful but noted that many of the solicitationshave a coercive element in that the consentingbondholders receive a fee (typically $10 per$1,000 of face value) if the solicitation is success-ful, whereas those not consenting receive no fee.They also found, however, that bondholdersenjoyed, on average, significant positive abnormalreturns around the announcement of ‘‘potentially’’coercive solicitations. This finding suggests thatissuing firms ‘‘cannot, or do not, exploit thecoercive nature of their solicitations.’’

The evidence presented by Asquith and Wizman(1990) and Kahan and Tuckman (1992) mayappear inconsistent with the view that covenantsare not binding in the public market, but it is notnecessarily so. The results in both of these studieswere driven by extraordinary events. In theAsquith and Wizman study, the events wereLBOs. Similarly, the largest category by far in theKahan and Tuckman sample also involved firmsthat were targets of an LBO (and most othersinvolved relatively unusual events). The results ofthese studies suggest that covenants in publicbonds can impose meaningful limitations on eventrisk. When more routine types of actions by firmsare likely to trigger bank or private placement

197. Super poison puts usually give bondholders the right toredeem their bonds at face value (or the nominal cash flowsdiscounted at U.S. Treasuries plus some spread if the contractincludes a prepayment penalty) if a designated event occursand if it is accompanied by a downgrading in the borrower’sbond rating from investment grade to non-investment grade.The RJR–Nabisco LBO spawned the use of super poison puts.Previously some bonds offered poison puts, which essentiallyprovided protection against hostile takeovers but not againsttakeovers approved by the target’s board. Super poison putsprovide protection against a wider range of events than dopoison puts, including target-board-approved takeovers (seeBrook, 1990). In 1989, Standard & Poor’s began rating theamount of event protection provided in public bonds separatelyfrom the rating of the bond itself. Moody’s, however, adjustsits bond ratings to reflect event risk rather than providing aseparate rating for event risk. Crabbe (1991b) found thatsuper poison puts may have reduced interest costs to issuers byabout 20 to 30 basis points. Recently, super poison puts haveappeared less frequently in investment-grade public bonds, butthey are still found in below-investment-grade public bonds.

93

covenants, however, public bond covenants arenormally not binding.198 In a sample of 128 firms

that violated accounting-based covenants, Chenand Wei (1991) found that only 4 involvedpublic debt. The remainder were privatelyplaced securities. 199

198. As noted by Brook (1990), a distinction should bemade between the public junk-bond market and the publicinvestment-grade market. Brook reports that meaningfulcovenant restrictions on asset disposition, maintenance ofcapital, and dividend payouts are found in the public junk-bondmarket but not typically in the public investment-grade market.

199. Chen and Wei looked only at actual violations and notat requests for covenant waivers.

94

Appendix F. An Example of a Private Placement Assisted by an Agent

This appendix describes a private placementtransaction involving an agent from start to finish.So much variety exists in the private market thatno single transaction can be called typical. Thisexample introduces terminology, concepts, and asense of the stages that deals must pass through.It is based partly on publicly available data aboutan actual transaction, but names and amountshave been changed, and many of the details arefictional.

In late spring 1991, Acme Stores, Inc., openednegotiations with its primary bank, BigBank, torenew a revolving credit agreement that had beenin force for three years. Acme Stores is a publiclyheld company with $1 billion in annual sales,operating in a highly competitive area of the retailsales sector. BigBank is a bank affiliate of a largeU.S. bank holding company.

The occasion inspired wide-ranging discussionsbetween officers of Acme and those of BigBankabout Acme’s current and prospective capitalneeds. BigBank’s officers adopted the ‘‘corporatefinance’’ perspective now coming into fashionamong loan officers at many of the large banks, inwhich the loan officer advises on financial strate-gies involving a wide range of instruments insteadof focusing on the sale of bank loans. BigBanknoted that the history of takedowns and repay-ments for the existing revolver indicated that about$30 million of the $45 million now outstandingwas essentially a term loan that probably wouldnot be repaid in the near future. Traditionally,BigBank noted, an expiring revolver would havebeen rolled over into a combination of a term loanfor the longer-term component of the balance anda line of credit for the remaining, seasonalcomponent. BigBank was not devoted to traditionand said that it would be happy to arrange a newrevolver if that was desired, as Acme was arelatively high-quality borrower. BigBank stated,however, that investors often look more favorablyon middle-market companies that have obtainedlong-term debt financing and that this point in theinterest cycle appeared to be a good time toborrow at a fixed rate.

Acme was interested in obtaining long-term,fixed-rate financing. BigBank noted that a floating-rate term loan could be swapped into fixed rate,but the maximum maturity that BigBank couldoffer would be about five years. Acme was inprinciple interested in a longer term. BigBankadvised that the combination of Acme’s size,business profile, financial condition, and the

amount of the term loan (around $30 million) wassuch that a public offering of debt was infeasiblebut that a private placement including somefinancial covenants might be an attractive option.At this point, BigBank’s private placement agentwas called in for consultation. As a major corpo-rate lender, BigBank had found setting up its ownprivate placement agent organization profitable.

Based on a quick review of Acme and itsfinancial position, the agent estimated that aprivate placement of fixed-rate, senior, unsecurednotes with a maturity of about ten years would befeasible. Current private market conditions forborrowers like Acme were such that the loanwould have to amortize; a bullet loan probablycould not be placed with investors. Furtherdiscussion yielded a plan for an issue of $32 mil-lion of eight-year notes, with annual principalrepayments of $4 million.

The agent explained the best-efforts nature ofthe process. In cooperation with Acme, BigBankwould design an initial set of terms for thesecurities and then seek investors, which wouldlikely be life insurance companies. BigBank wouldnot guarantee the pricing of the issue. Termsmight change during negotiations with the inves-tors, and the possibility of no deal being struckwas real. BigBank’s fee would be 3⁄4 of 1 percentof the face amount of the placement, but the feewould be collected only if the placement weresuccessful. After signing an agent agreement withBigBank, obtaining commitments from investorswould take about two months, and funds could bedisbursed another month or so after that. 200

Acme sought bids and advice from other privateplacement agents and found BigBank’s fees to becompetitive. It also found that some agents werenot interested in the transaction. They explained toAcme that $32 million was a relatively smallplacement, and that the staff time and other fixedcosts to do a placement were about the sameregardless of a deal’s size. Thus, some agentspreferred to concentrate only on larger transactionsfor which the fees were larger.

200. The two-month lead time was required partly becauseAcme had to be rated by the NAIC. A rating was especiallyimportant because Acme was on the borderline between a BBand BBB rating (NAIC-3 and NAIC-2). Under current condi-tions, investors would be hard to find for a BB issue, and thecoupon rate would be so high that Acme would be better offwith a five-year bank loan. With a BBB rating, a coupon ratein the vicinity of 200 basis points over comparable Treasurieswas likely. See part 3, section 1, for a discussion of the specialcircumstances currently surrounding BB private placements.

95

Acme also knew that it had to renew itsrevolver before funds from the private placementwould be available. If the private placement didnot succeed, Acme would need either a term loanor a larger revolver, which would be easiest tonegotiate if BigBank were the agent that failed tocomplete the private placement. Thus Acme gavea mandate to BigBank’s agent organization toarrange the transaction.

The agent’s first action was to conduct duediligence, meaning it made a close examination ofAcme’s business, financial position, and plans.This examination was similar to the one thatlenders would later conduct, and it included a visitto Acme’s facility. Performance of due diligencewas relatively easy for BigBank because itsrelationship officers already had substantialinformation about Acme.

Having gathered much information about Acme,the agent began writing the offering memorandum(offer memo) and term sheet. An offer memodescribes the borrower and is functionally similarto a prospectus for a public offering. However,it usually includes more information, such asforecasts, than does a prospectus. The term sheetis a description of the major terms of the securitiesto be offered, including covenants. An interest rate(expressed as a spread over Treasuries of compa-rable maturity) is generally not included initially.Lenders use the offer memo and term sheet indeciding whether or not to buy the securities.Initial term sheets vary in their content and detail,depending on the nature and quality of theborrower, the complexity of the transaction, andthe distribution style of the agent. In this case, theBigBank agent wrote a relatively detailed termsheet, which included covenants restrictingAcme’s financial ratios. These covenants weresimilar to those in the new revolving creditagreement concurrently being negotiated withBigBank. A substantial penalty for prepayment ofthe note was also included. Although Acme wasquite unhappy about this penalty, the agent notedthat the securities could not be placed withoutit. 201 The ‘‘final’’ versions of both the offer memoand term sheet were produced after extensive

consultations with Acme, about two weeks afterBigBank received the mandate to go ahead.

With the offer memo and term sheet in hand,the agent initiated the process of getting a ‘‘pre-rating’’ of the securities. 202 The NAIC does suchratings only at the request of an insurancecompany. The agent used a contact to make arequest, reimbursing the insurer for the NAIC’sfee. This insurance company made no commitmentto buy the securities. With securities issued byvery-high-quality borrowers, beginning andsometimes even completing distribution with norating in hand is possible; but with a borderlineborrower like Acme and under the market condi-tions of the time, no lender would seriouslyconsider a purchase without knowing the rating.The NAIC rating process takes at least threeweeks.

Upon obtaining the rating, and finding it to be aNAIC-2 (or BBB) as expected, the agent beganlooking for investors who would buy the securi-ties. The style in which private placement distribu-tions are conducted varies widely across bothagents and transactions. In this case, BigBank’sagent noted that the transaction was for a rela-tively small amount of senior unsecured debt andthat the borrower was moderately risky by thestandards of the private placement market.

These facts had several implications. First, oneor more lead lenders were required. A lead lenderis one that commits to buy a significant fraction ofthe placement and that has the necessary creditevaluation and monitoring capacity to assess theloan’s risk, negotiate terms, and monitor perfor-mance during the life of the loan. Relativelysmall lenders with limited evaluation and moni-toring capacity will rely to some extent on thequality signal implicit in a lead’s commitmentwhen making their own decision to purchase aplacement. Second, lead buyers of senior unse-cured private debt are usually life insurancecompanies. Third, the largest life insurancecompanies were less likely to be interested ina transaction of relatively small size.

Thus BigBank’s agent began distributing theplacement by sending the term sheet and offermemo to several insurance companies, moderate tolarge in size, verbally suggesting an interest rate201. Acme objected to the prepayment penalty for two

reasons. First, it effectively removed Acme’s option to reducefinancing costs in the future should interest rates fall. Second,should Acme wish to pursue a strategy in the future that wouldcause one of the covenants to be violated, Acme would haveeither to persuade the lenders to grant a waiver or to prepaythe bonds (incurring the penalty) to escape the covenantrestrictions. Acme reluctantly agreed to the prepayment penaltyprovision only after BigBank assured it that lenders with ahistory of flexibility in waiving covenants could be found.

202. Virtually all securities bought by life insurance compa-nies are rated by the end of the year of purchase by the NAIC.As these ratings influence capital requirements and determinereserves that must be held against placements, many insurersprefer to obtain ratings before purchase, known as a ‘‘pre-ratings.’’

96

of about 190 basis points over Treasuries. Theseinsurers were known to the agent to be receptiveto deals with borrowers in Acme’s industry andrisk category.203 The securities were offered tothese lenders on a first-come, first-served basis.

Although private market lenders often call theagent to obtain information not in the offer memo,and in rarer cases may perform substantialamounts of independent due diligence beforemaking a commitment, the norm in the privatemarket is for lenders to make semiformal purchasecommitments based on analysis of the informationin the offer memo and term sheet and to do sofairly rapidly, within a week or two.204 Potentiallead lenders generally indicate interest by makinga counteroffer in which they state that they willpurchase a given amount at a given rate and withgiven changes in the other terms.205

A life insurance company of moderate sizecircled (agreed to buy) about half of the placementat a spread of 200 basis points over Treasuries ofcomparable maturity, and two others followed thislead in committing to another 40 percent or so,

one at 190 basis points over and one at 195 overTreasuries. Left with only $3 million of notes toplace, the agent turned to several smaller insur-ance companies, offering them all or part of theremaining notes on the terms negotiated with thelead lender. These companies rely most on thesignal implicit in the lead lender’s decision, asthey are offered a take-it-or-leave-it propositionand given only two or three days to decide.

Time required for distribution varies. In thiscase distribution took about two weeks.

Having fully subscribed the placement, Big-Bank’s agent informed the lenders that they werein the syndicate and set the coupon rate (at 200basis points over Treasuries on that day) and thenturned to the next stage, lenders’ due diligence.Each of the major lenders conducted an investiga-tion of the borrowing company, including visits tothe Acme’s facility, that was similar to that doneby Bigbank at the beginning of the process. Duediligence is done promptly, and on its completionthe lenders’ committees pass formal judgment onthe loan and dispatch formal commitment letters.

In the final stage of the private issuance,lawyers hammered out the language of the debtcontract, which involved several documentsbesides the notes themselves. The lenders wererepresented by a bond counsel that was chosenby the lead lender but paid by Acme. Acme wasrepresented by its own counsel with assistancefrom BigBank. The process of working thecontract took three weeks.

Closing ends the process of issuance and theagent’s role. In this case, BigBank collected a feeof $240,000. Acme paid down a $30 millionbridge portion of its new revolver, which wentinto effect while the private placement transactionwas in progress, and put $2 million into itstreasury.

203. Some insurers avoid some industries or risk categories,and the nature of insurance company preferences is constantlychanging. Preferences for particular kinds of restrictive cove-nants also vary across private market lenders.

204. When demand for new issues is very heavy, thedecision period may be compressed to a few days.

205. In this sort of lead-lender distribution, the coupon rateis fixed at the time the deal is fully subscribed. If a lender thatcircled (agreed to buy) withdraws after the coupon rate is set,the coupon rate may be revised upward if the agent finds doingso necessary for completing the distribution, but it may not berevised downward. Circling is a semiformal commitment thatcan be (but rarely is) overturned by an insurance company loancommittee. Withdrawals of commitments occur mainly becausea lender finds on further investigation that the offer memo andterm sheet did not accurately represent the borrower or securi-ties. Withdrawals for such reasons are fairly infrequent.

97

Appendix G. Estimates of Issue-Size and Maturity Distributions

This appendix explains how the issue-size andmaturity distributions described in part 1, sec-tion 2, were computed and presents distributionsfor all private placements (those in the text are forplacements by only nonfinancial corporations).Distributions for commercial and industrial loansmade by U.S. commercial banks, private place-ments issued by nonfinancial corporations, andpublic debt issues by nonfinancial corporations areshown in the text for 1989. We chose this yearbecause it was the last year before the creditcrunch in the below-investment-grade segment ofthe private placement market. The data anddistributions for each of the three types of instru-ment are described separately.

Private Placements

Both issue-size and maturity distributions for newissues of private placements were produced from adatabase of 1989 private placement issues obtainedfrom IDD Information Services. This firm isassociated with the publisher of InvestmentDealers Digest. 206 This database includes informa-tion about new private issues of both debt andequity, including issue size (amount); issue date;the name of the issuer; the name(s) of the agent(s)involved; the nationality of the issuer; the type ofsecurity involved; an industry code for the issuer(an IDD designation, not an SIC code); a maturitydate where applicable; an indication of whetherthe issue was junk, lease-related, or part of anacquisition-related financing; and, in some cases,a coupon rate or a number of shares issued and afew words of descriptive comments. Collectedprimarily from reports that agents sent to IDD, thedata include few or no deals not involving anagent and omit many agent-assisted deals thatwere unreported. The accuracy of the data has notbeen verified, and we have reason to believe thatat least some of the data are unreliable (forexample, the junk designation is usually ‘‘no’’ forissues assisted by Drexel, which does not squarewith Drexel’s reputation for specializing inbelow-investment-grade issues).

Issue-size distributions were produced using theissue-size (amount) field in the database. Maturitydistributions were produced by computing maturityat issue from the issue date and maturity date

appearing in the database. This computation ofteninvolved some approximation, since in many casesonly the month and year (not the day) of issuanceand maturity appear in the database, and in somecases only the year of maturity is specified. Inthese cases we assumed that the month and day ofmaturity were the same as the month and day ofissuance; in effect, we may have rounded up somematurities to the next year.

Both issue-size and maturity distributions wereproduced from a sample limited to 1989 issues ofbonds by U.S. issuers. Two subsamples wereanalyzed. The one shown in the text (charts 4, 5,10, and 11) was for issues by nonfinancial corpo-rate borrowers and excluded medium-term notes,convertible or exchangeable debt, and mortgage-backed securities. 207 This subsample totaled 1,020issues (maturity dates were available for only901).

The second subsample excluded only medium-term notes and contained 1,620 issues (maturitydates were available for only 1,373 of them).Charts G.1 through G.4 display the distributionsfor this subsample. These distributions are similarto those for the other subsample.

The distributions for new issues shown here arenot necessarily representative of the distributionsfor outstanding private placements. Adequate dataon outstandings are not available. Kwan andCarleton (1993) report that the average term tomaturity for a sample of 563 private placementsissued between 1985 and 1992 was 11.12 years,a finding that is in rough agreement with thedistributions displayed here. 208

Publicly Issued Bonds

Issue-size and maturity distributions for new issuesof public bonds by U.S. issuers were producedfrom a sample issued during 1989 by nonfinancialcorporations and collected by the Federal ReserveBoard from public announcements, proxy state-ments, and other sources. The subsample analyzedhere included no government issues, no medium-term notes, no convertible or exchangeable debt,and no mortgage-backed or other asset-backedsecurities. An unusual issue related to the

206. IDD has since sold its data services operations toSecurities Data Corporation.

207. That is, borrowers with an IDD industry designation of‘‘financial’’ or ‘‘government’’ were excluded.

208. Kwan and Carleton’s sample was from the portfolio ofthe Teacher’s Insurance and Annuity Association, a largeinsurance company.

98

RJR–Nabisco merger was also omitted. Thissample included 297 issues.

Commercial and Industrial Bank Loans

Issue-size and maturity distributions for new orrenewed bank loans were produced using datafrom the Quarterly Survey of Terms of BankLending to Business. This survey is of a stratifiedrandom sample of approximately 350 banksrepresenting insured commercial banks in theUnited States. Large banks are oversampled.

Participating banks report data on all loans tobusinesses made during the first full week inFebruary, May, August, and November (largerbanks report data only for loans made on two orthree days of those weeks and only for a subset ofbranches). Individual survey responses are confi-dential and include for each loan an amount andmaturity date as well as other information. Nomaturity date is recorded for demand loans withno specific maturity date, and rules for reportingloans made under a revolving credit agreement arecomplicated, in many cases requiring that nomaturity date be reported. Construction and land

G.1. Size distribution of all private placementsby number of issues, 1989

Percent of issues

G.3. Maturity distribution of all privateplacements, by number of issues, 1989

Percent of issues

0

10

20

30

40

.31 .82

6.11 7.24

25.94

40.11

13.54

3.842.08

.05– .25– 1– 5– 10– 25– 100– 250– 500–.25 .1 5 10 25 100 250 500 and

more

Size of issue (millions of dollars)

G.2. Size distribution of all private placementsby volume, 1989

Percent of volume

G.4. Maturity distribution of all privateplacements, by volume, 1989

Percent of volume

0

10

20

30

40

4 5 6 5

96

13

5 4

24

96

3

1– 2– 3– 4– 5– 6– 7– 8– 9– 10– 15– 20– 302 3 4 5 6 7 8 9 10 15 20 30 and

more

Term (years)

0

10

20

30

40

.01 .23 .69

5.76

27.14 26.38

17.10

22.68

.05– .25– 1– 5– 10– 25– 100– 250– 500–.25 .1 5 10 25 100 250 500 and

more

Size of issue (millions of dollars)

0

10

20

30

40

4 4 42

8

4

9 96

29

13

6

2

1– 2– 3– 4– 5– 6– 7– 8– 9– 10– 15– 20– 302 3 4 5 6 7 8 9 10 15 20 30 and

more

Term (years)

99

development loans were omitted from the samplewe analyzed.

Several steps were required to arrive at distribu-tions representative of commercial and industrialloans by all banks. First, subsamples of loans bybanks reporting for less than five days in a surveyweek or for less than all branches were blown upto make them comparable to subsamples offive-day, all-branch reporters. Then sample bankswere stratified according to size, using theirfour-quarter averages of total C&I loans outstand-ing at the end of each quarter (computed fromCall Report data). We pooled loans for banks ineach stratum and computed size and maturitydistributions for each stratum. Essentially, weassumed that the sample of loans made by surveybanks in a stratum was representative of thepopulation of 1989 C&I loans made by all banksin that stratum. Distributions were computed foreach bank size stratum because loan sizes and loanmaturities tend to be related to the size of thelending bank, with larger banks making morelarge and more long-term loans.

We arrived at loan-size and maturity distribu-tions for all bank loans by taking a weightedaverage of the distributions for the differentbank-size strata. The weighting variables were thefractions of all C&I loans outstanding that allbanks in a stratum showed on end-of-quarter Call

Reports (averaging these outstandings for all of1989). That is, having achieved (we hope) repre-sentative distributions for each size class of banks(as described in the previous paragraph), weweighted the distributions according to the shareof total outstanding loans accounted for by eachsize class. Because the largest banks make adisproportionate share of C&I loans, the distribu-tions for the large-bank strata had more influenceon the final, representative distributions than didthe distributions for small-bank strata.

As noted, the distributions are for newlyoriginated or renewed loans, not for the populationof loans outstanding. Distributions for outstandingsmay differ substantially from those for origina-tions, for two principal reasons. First, short-termloans may essentially be overweighted by ourmethod of constructing distributions although asensitivity analysis indicated that any suchoverweighting probably has little effect on theestimated distributions. Second, bank loans areoften prepaid.

Our distributions have other drawbacks. Mostnotably, the data available from the survey forrevolving credit agreements are limited, and theseconstitute a significant share of all bank loans.However, the proper maturity date definition toapply to revolvers in comparing them to privateplacements and public bonds is not clear.

100

Appendix H. Borrower Substitution between the Public and Private Markets

Empirical research on the relation between publicand private bond markets has focused primarily onthe ability of borrowers to switch between thetwo. The research in this regard is limited, but ithas covered a considerable period of time. In oneof the earliest studies, which covered 1951 to1961, Cohan (1967) concluded that the twomarkets were segmented. This conclusion wasbased on a finding that yields, adjusted forissuance costs, were generally higher on privateplacements than on public offerings. Cohanattributed the lack of similar borrowing costs tothe structure of the markets: The public bondmarket served primarily large, better-knowncorporations, and the private market servedsmaller, less-well-known corporations. Thepreference of the public market for better-knowncorporations prevented smaller companies fromshifting out of the private market into the publicmarket to take advantage of the lower borrowingcosts.

Cohan’s research did not examine the sensitivityof the demand for funds in the two markets tochanges in yields. Wolf (1974) corrected thisshortcoming by estimating demand equations over1956–70 for two groups of firms: One includedFortune 500 companies, and the other included allother industrial corporations. 209

For each group, the dependent variable in aregression equation was the ratio of gross issuancein the private market to total gross issuance inboth the private and the public markets; theexplanatory variables included the spread betweenyields in the private and public markets, a measureof the supply of funds in the private market, andaverage issue size. With smaller borrowers havingno access to the public market, Wolf hypothesizedthat only the Fortune 500 companies would besensitive to the yield spread. The empirical resultswere consistent with this hypothesis; the coeffi-cient on the yield spread was statistically signifi-cantly different from zero only in the equations forthe Fortune 500 companies. Moreover, whenthese companies were further disaggregated, the

estimated elasticity of demand with respect to theyield spread for the top 100 corporations was, inmagnitude, nearly three times that for the top 250corporations.

Although Wolf’s approach was an improvementover Cohan’s, the demand equations were mis-specified because other relevant explanatoryvariables, such as issuance costs, credit quality,and characteristics of the instruments, were notincluded. In this regard, Blackwell and Kidwell’s(1988) study improved on Wolf’s by incorporatingmore institutional features of the markets into ananalysis of a borrower’s choice of the public orthe private market. They assumed that, holding allelse constant, a borrower chooses to issue in theleast expensive market. The basic elements of thedecision can be illustrated by assuming theborrower is contemplating issuing a one-year bondwith one annual interest payment. The all-in costper dollar borrowed, r, can be computed from theexpression

S − K = (1 + c)S / (1 + r),

where S is the issue price, K is the fixed cost ofissuance, and c is the coupon rate of interest. Thebond is assumed to be issued at par; thus, theissue price and face value are equal. Variable costscan be ignored as they do not produce economiesof scale in issue size. 210 From this expression, theall-in cost can be derived as

(1) r = (1 + c) / [1 − (K /S)] − 1.

As issue size S increases, the all-in cost r falls,thereby producing the economies of scale.

Two conditions are necessary for a borrower tofind that the private market is less expensive forsmall issues and more expensive for large issues.One is that the fixed issuance cost must be lowerin the private market, and the other is that thecoupon rate must be lower in the public market.Assuming both are satisfied, the break-even issuesize S*, the point at which the all-in cost is thesame in both markets, is

(2) S* = Kpr + (Kpu − Kpr)(1 + cpr) / (cpr − cpu),

where the subscript pr indicates the private marketand the subscript pu, the public market. For an

209. Wolf also considered a third group that includednon-industrial corporations. The heterogeneity of the groupmade the estimated demand equations difficult to interpret. Ofthese corporations, public utilities were restricted primarily tothe public market for legal and regulatory reasons, real estatecompanies financed almost exclusively in the private marketbecause of the complicated nature of their transactions, andlarge finance companies used both markets although theygenerally confined their issuance of subordinated bonds to theprivate market.

210. S − K measures the net proceeds of the issue receivedby the borrower.

101

issue less than S*, the borrower would choose theprivate market; for an issue greater than S*, theborrower would choose the public market.

If either of the two conditions is not satisfied,borrowing in only one market is always cheaper.Issuance costs are lower in the private marketbecause of the absence of both SEC registrationexpenses and underwriting fees. However, thecoupon rate condition is not necessarily satisfied.On the one hand, for some borrowers the differ-ence in coupon rates in the two markets may befully accounted for by an illiquidity premium thatmakes the coupon rate in the private market higherthan that in the public market. In this case, issuesize determines the choice of the market. On theother hand, information-problematic borrowersshould find that the coupon rate is higher in thepublic market, because of the unwillingness ofpublic-market investors to undertake the creditanalysis and monitoring required to lend to theseborrowers. Under this circumstance, these borrow-ers, in effect, have no choice but to issue in theprivate market.

Blackwell and Kidwell provide some weakempirical evidence that information problems aremore important than issuance costs for smallcompanies using the private market. To examinethis issue, Blackwell and Kidwell first estimated alinear version of equation 1 for a sample of publicbonds issued by large companies that had usedonly the public market. The explanatory variablesin the regression estimating equation 1 includedissue size, characteristics of the issuer, features ofthe securities, and market conditions. Based uponthis regression, Blackwell and Kidwell thenpredicted the all-in cost in the public market for asample of private bonds issued by small compa-nies that had used only the private market. Thepredicted value was then compared with the actualall-in cost of the private placements to determinewhat, if any, cost saving was achieved in theprivate placement market.

The estimated cost saving from using theprivate placement market averaged 132 basispoints. The average saving was 301 basis pointsfor below-investment-grade companies but was−16 basis points for investment-grade companies.This finding is consistent with the cost savingreflecting the information-problematic nature ofthe companies issuing the private placements asbelow-investment-grade firms are more likely(though not certain) to be information problematic.Because such companies require the most duediligence and monitoring, they must pay thelargest premium to borrow in the public market,

where investors generally forgo the risk controlinherent in due diligence and monitoring. Somemarket participants have indicated also thatissuance costs tend to be higher for such compa-nies, but they are unlikely to be as large asBlackwell and Kidwell’s estimates.

Several problems temper Blackwell and Kid-well’s results. One is that the issue size variablein the regression estimates was not significant,indicating an absence of economies of scale inissue size in the public market. Another problemis that the values of the explanatory variables forthe private issues are generally well outside theranges of the variables used to estimate the model,leading to extremely imprecise estimates of thecost savings. Finally, the negative estimate for theinvestment-grade issues, even though statisticallyinsignificant, implies that nonprice factors notincluded in the regression model dominate priceconsiderations for high-grade borrowers. Other-wise, it makes no sense for these companies toborrow in the private market when the publicmarket is the cheaper alternative. The omission ofrelevant variables in the regression model couldalso lead to biased estimates of the cost saving inthe private market.

In a second regression, Blackwell and Kidwellproduce results that can be interpreted as implyingthat those companies with access to the publicmarket utilize the private market primarily becauseof special borrowing needs and not because ofcost considerations. This regression involvedestimating equation 1 for a sample containing bothprivate and public issues; the public issues werethose described above, whereas the private issueswere those of companies that had issued in boththe public and private markets. To distinguishbetween the public and the private issues, theregression included dummy variables that allowedthe intercept and the coefficients of severalexplanatory variables to shift with the market.

The dummy variables were generally insignifi-cant, indicating that the all-in cost of issuing inthe two markets was the same. This finding isinconsistent with the underlying model of bor-rower choice and certainly is at odds with marketparticipants’ description of the cost differentialsbetween the two markets. The modeling strategymay well account for this result and, in fact, maylack the power to detect differences in issuancecosts. One problem concerns those firms that hadissued in both markets, the so-called switch hitters.They are large relative to companies issuing onlyprivate placements, and consequently their obser-vations cluster around the break-even issue size

102

S*. Thus the estimated equation provides unreli-able estimates of issuance costs for small issues(which are generally by small firms), and preciselyfor such issues the differences in the all-in costbetween the private and public markets should bemost pronounced. An even more critical problemstemming from the presence of switch hitters inthe sample is their tendency to use the privatemarket for non-price-related reasons, such asspeed or complexity of a particular issue, ratherthan for cost alone. Thus, their use of the privateplacement market may provide little informationabout relative issuance costs in the two markets

and, indeed, could lead to a finding that costminimization is irrelevant. That is, the negativeresults for issuance cost actually point to theimportance of nonprice factors.

In general, Blackwell and Kidwell’s work pointsto the need for a richer model of borrower choice.Choosing is not simply a matter of selecting themarket with the lower all-in cost of borrowing. Asour analysis in part 1 emphasizes, nonprice terms,such as disclosure requirements, covenants andtheir renegotiation, and speed, can matter andshould be incorporated into models of borrowers’choice of markets.

103

References

Allen, Linda, Anthony Saunders, and Gregory F.Udell. ‘‘The Pricing of Retail Deposits: Concen-tration and Information,’’ Journal of FinancialIntermediation, vol. 1 (December 1991),pp. 335–61.

Altman, Edward I. ‘‘A Further Empirical Investi-gation of the Bankruptcy Cost Question,’’Journal of Finance, vol. 39 (September 1984),pp. 1067–89.

American Bar Foundation. ‘‘Commentaries onModel Debenture Indenture Provisions, 1965;Model Debenture Indenture Provisions,All Registered Issues, 1967; and CertainNegotiable Provisions Which May Be Includedin a Particular Incorporating Indenture.’’Chicago, Ill.: ABA, 1971.

American Council of Life Insurance. ‘‘InvestmentBulletin: Survey of Private Placement Commit-ments.’’ Washington, D.C.: ACLI, variousissues.

Asquith, Paul, and Thierry A. Wizman. ‘‘EventRisk, Covenants, and Bondholder Returns inLeveraged Buyouts,’’ Journal of FinancialEconomics, vol. 27 (September 1990),pp. 195–213.

Bailey, Dianna C., and Donald J. Mullineaux.‘‘The Nonuniqueness of Bank Loans.’’ WorkingPaper. Louisville, Ky.: University of Louisville,1989.

Banning, Peter S., and Christopher James. ‘‘AnAnalysis of Private Nonbank Loan Agree-ments.’’ Working Paper. Gainesville, Fla.:University of Florida, August 1989.

Benston, George J. ‘‘The Effectiveness and Effectsof the SEC’s Accounting Disclosure Require-ments,’’ in Henry G. Manne, ed., EconomicPolicy and the Regulation Corporate Securities.Washington: American Enterprise Institute,1969, pp. 23–79.

Berger, Allen N., and Gregory F. Udell. ‘‘Collat-eral, Loan Quality, and Bank Risk,’’ Journal ofMonetary Economics, vol. 25 (January 1990),pp. 21–42.

, and . ‘‘Securitization,Risk, and the Liquidity Problem in Banking,’’ inMichael Klausner and Lawrence J. White, eds.,Structural Change in Banking. Homewood, Ill.:Irwin, 1993a.

, and . ‘‘Lines of Credit,Collateral, and Relationship Lending in SmallFirm Finance.’’ Finance and Economics Discus-sion Series, 93-9. Washington: Board ofGovernors of the Federal Reserve System,March 1993b.

Berlin, Mitchel. ‘‘Loan Reschedulings and theSize of the Banking Sector.’’ Working Paper.New York: New York University, Stern Schoolof Business, 1991.

, and Jan Loeys. ‘‘Bond Covenantsand Delegated Monitoring,’’ Journal of Finance,vol. 43 (June 1988), pp. 397–412.

, and L.J. Mester. ‘‘Debt Covenantsand Renegotiation,’’ Journal of FinancialIntermediation, vol. 2 (June 1992), pp. 95–133.

Bernanke, Ben S., and Cara S. Lown. ‘‘The CreditCrunch,’’ Brookings Papers on EconomicActivity (1991, no. 2), pp. 205–247.

Besanko, David, and Anjan V. Thakor. ‘‘Collateraland Rationing: Sorting Equilibria in Monopolis-tic and Competitive Credit Markets,’’ Interna-tional Economic Review, vol. 28 (October1987a), pp. 671–89.

, and . ‘‘CompetitiveEquilibrium in the Credit Market under Asym-metric Information,’’ Journal of EconomicTheory, vol. 42 (June 1987b), pp. 167–82.

Bester, Helmut. ‘‘Screening versus Rationing inCredit Markets with Imperfect Information,’’American Economic Review, vol. 75 (Septem-ber 1985), pp. 850–55.

Bhagat, Sanjai, and Peter A. Frost. ‘‘Issuing Coststo Existing Shareholders in Competitive andNegotiated Underwritten Public Utility EquityOfferings,’’ Journal of Financial Economics,vol. 15 (January-February 1986), pp. 233–59.

105

Billett, Matthew Thayer, Mark J. Flannery, andJon A. Garfinkel. ‘‘The Effect of LenderIdentity on a Borrowing Firm’s Equity Return.’’Working Paper. Gainesville, Fla.: University ofFlorida, July 1993.

Black, Fischer. ‘‘The Dividend Puzzle,’’ Journal ofPortfolio Management, vol. 2 (Winter 1976),pp. 5–8.

, and Myron Scholes. ‘‘The Pricing ofOptions and Corporate Liabilities,’’ Journal ofPolitical Economy, vol. 81 (May-June 1973),pp. 637–54.

Blackwell, David W., and David S. Kidwell. ‘‘AnInvestigation of Cost Differences BetweenPublic Sales and Private Placements of Debt,’’Journal of Financial Economics, vol. 22(December 1988), pp. 253–78.

Board of Governors of the Federal ReserveSystem. Commercial Bank Private PlacementActivities. Staff Study. Washington: Board ofGovernors, June 1977.

Boot, Arnoud W.A., Anjan V. Thakor, andGregory F. Udell. ‘‘Secured Lending andDefault Risk: Equilibrium Analysis, PolicyImplications and Empirical Results,’’ EconomicJournal, vol. 101 (May 1991), pp. 458–72.

Boyd, John H., and Edward C. Prescott. ‘‘Finan-cial Intermediary-Coalitions,’’ Journal ofEconomic Theory, vol. 38 (April 1986),pp. 211–32.

Brealey, Richard A., and Stewart C. Myers.Principles of Corporate Finance, 4th ed.New York: McGraw Hill, 1991.

Brewer, Elijah III, Thomas H. Mondschean, andPhilip E. Strahan. ‘‘The Effect of Capital onPortfolio Risk at Life Insurance Companies,’’ inProceedings of the 29th Annual Conference onBank Structure and Competition. Chicago:Federal Reserve Bank, 1993.

Brook, Richard. ‘‘Debt Covenants and Event Risk:The Practitioner as a Source of Evidence.’’Working Paper. New York: Columbia Univer-sity, The Center for Law and Economic Studies,October 1990.

Bulow, Jeremy, and John B. Shoven. ‘‘TheBankruptcy Decision,’’ Bell Journal of Econom-ics, vol. 9 (Autumn 1978), pp. 437–56.

Campbell, Tim S., and Yuk-Shee Chan. ‘‘OptimalFinancial Contracting with Ex Post and Ex AnteObservability Problems,’’ Quarterly Journal ofEconomics, vol. 107 (May 1992), pp. 785–95.

Cantor, Richard, and John Wenninger. ‘‘Perspec-tive on the Credit Slowdown,’’ Federal ReserveBank of New York Quarterly Review, vol. 18(Spring 1993), pp. 3–36.

Carey, Mark S., and others. ‘‘Recent Develop-ments in the Market for Privately Placed Debt,’’Federal Reserve Bulletin, vol. 79 (February1993), pp. 77–92.

, and . ‘‘The Economics ofPrivate Placements: A New Look,’’ FinancialMarkets, Institutions & Instruments, vol. 2(August 1993), pp. 1–66.

Carlson, James B., Elizabeth A. Raymond, andStephen A. Keen. ‘‘Securities Law,’’ in JasonH.P. Kravitt, ed., Securitization of FinancialAssets, Volume 1. Englewood Cliffs, N.J.:Prentice-Hall Law & Business, 1992.

Chan, Yuk-Shee. ‘‘On the Positive Role ofFinancial Intermediation in Allocation ofVenture Capital in a Market with ImperfectInformation,’’ Journal of Finance, vol. 38(December 1983), pp. 1543–68.

, and George Kanatas. ‘‘AsymmetricValuations and the Role of Collateral in LoanAgreements,’’ Journal of Money, Credit, andBanking, vol. 17 (February 1985), pp. 84–95.

, Daniel Seigal, and Anjan V. Thakor.‘‘Learning, Corporate Control and PerformanceRequirements in Venture Capital Contracts.’’Working Paper 36. Evanston, Ill.: NorthwesternUniversity, 1987.

, and Anjan V. Thakor. ‘‘Collateral andCompetitive Equilibria with Moral Hazard andPrivate Information,’’ Journal of Finance,vol. 42 (June 1987), pp. 345–63.

106

Chari, V.V., and R. Jagannathan. ‘‘Banking Panics,Information, and Rational Expectations Equilib-rium,’’ Journal of Finance, vol. 43 (July 1988),pp. 749–61.

Chemical Bank Structured Finance Division.‘‘Private Placements.’’ New York, 1992.

Chen, Kevin C.W., and K.C. John Wei. ‘‘TheEconomic Consequences of ViolatingAccounting-based Debt Covenants and Credi-tors’ Waiver Decisions.’’ Working Paper.Bloomington, Ind.: Indiana University,October 1991.

Cohan, Avery B. ‘‘Yields on Corporate DebtDirectly Placed.’’ New York: Columbia Univer-sity Press for the National Bureau of EconomicResearch, 1967.

Cook, Douglas O., John C. Easterwood, andJohn D. Martin. ‘‘Bondholder Wealth Effects ofManagement Buyouts,’’ Financial Management,vol. 21 (Spring 1992), pp. 102–13.

Corcoran, Patrick. ‘‘The Credit Slowdown of1989–1991: The Role of Supply and Demand,’’in Proceedings of the 28th Annual Conferenceon Bank Structure and Competition. Chicago:Federal Reserve Bank, 1992.

Corporate Financing Week. ‘‘Underwritten 144ADeals Surge,’’ CFW, August 31, 1992.

Crabbe, Leland. ‘‘Callable Corporate Bonds:A Vanishing Breed.’’ Finance and EconomicsDiscussion Series, no. 155. Washington: Boardof Governors of the Federal Reserve System,March 1991a.

. ‘‘Event Risk: An Analysis of Lossesto Bondholders and ‘Super Poison Put’ BondCovenants,’’ Journal of Finance, vol. 46(June 1991b), pp. 689–706.

, and Jean Helwege. ‘‘Alternative Testsof Agency Theories of Callable CorporateBonds.’’ Finance and Economics DiscussionSeries, 93-26. Washington: Board of Governorsof the Federal Reserve System, July 1993.

Davis Polk & Wardwell. ‘‘Adoption of Rule 144Aand Amendments to Rule 144A,’’ in Edward F.Greene and others, eds., Rule 144A: TheExpanded Private Placement Market. Engle-wood Cliffs, N.J.: Prentice-Hall Law & Busi-ness, 1990.

DeAngelo, H., L. DeAngelo, and D.J. Skinner.‘‘An Empirical Investigation of the Relationbetween Accounting Choice and DividendPolicy in Troubled Companies.’’ WorkingPaper 90-07. Ann Arbor, Mich.: University ofMichigan, School of Business Administration,1990.

Diamond, Douglas W. ‘‘Monitoring and Reputa-tion: The Choice between Bank Loans andDirectly Placed Debt,’’ Journal of PoliticalEconomy, vol. 99 (August 1991), pp. 689–721.

. ‘‘Financial Intermediation andDelegated Monitoring,’’ Review of EconomicStudies, vol. 51 (July 1984), pp. 393–414.

, and P.H. Dybvig. ‘‘Bank Runs,Deposit Insurance, and Liquidity,’’ Journal ofPolitical Economy, vol. 91 (June 1983),pp. 401–19.

Easterwood, John C., and Palani-Rajan Kadapak-kam. ‘‘The Role of Private and Public Debt inCorporate Capital Structure,’’ Financial Man-agement, vol. 20 (Autumn 1991), pp. 49–57.

Ederington, Louis H. ‘‘Uncertainty, Competition,and Costs in Corporation Bond Underwriting,’’Journal of Financial Economics, vol. 2(March 1975), pp. 71–94.

El-Gazzar, S., and V. Pastena. ‘‘NegotiatedAccounting Rules in Private Financial Con-tracts,’’ Journal of Accounting and Economics,vol. 12 (1990), pp. 381–96.

Engros, Charles E. Jr. ‘‘United States PrivatePlacements,’’ Client Memorandum. New York:Lord Day & Lord, Barrett Smith, January 1992.

Fama, Eugene. ‘‘What’s Different about Banks?,’’Journal of Monetary Economics, vol. 15(January 1985), pp. 29–36.

Fein, Melanie. Securities Activities of Banks.Englewood Cliffs, N.J.: Prentice-Hall, 1991.

107

Fenn, George, and Rebel Cole. ‘‘Announcementsof Asset-Quality Problems and ContagionEffects in the Life Insurance Industry,’’ Journalof Financial Economics, forthcoming, 1994.

Fitch Investors Service, Inc. ‘‘Covenants EnhancePrivate Placements.’’ Corporate FinanceIndustrials: Special Report. New York: FIS,October 28, 1991.

Franks, Julian R., and Walter N. Torous. ‘‘HowFirms Fare in Workouts and Chapter 11 Reorga-nization.’’ Working Paper. Los Angeles: Univer-sity of California, August 1990.

Gilson, Stuart C., Kose John, and Larry H.P. Lang.‘‘Troubled Debt Restructurings: An EmpiricalStudy of Private Reorganization of Firms inDefault,’’ Journal of Financial Economics,vol. 27, October 1990, pp. 315–53.

Gompers, Paul A. ‘‘Grandstanding in the VentureCapital Industry.’’ Working Paper. Boston,Mass.: Harvard University, School of BusinessAdministration, January 1993.

Gurwitz, Aaron. ‘‘SEC Rule 144A and Regula-tion S: Impact on Global Fixed IncomeMarkets.’’ New York: Goldman Sachs FixedIncome Research, September 1990.

Hart, Oliver, and John Moore. ‘‘Default andRenegotiation: A Dynamic Model of Debt.’’Discussion Paper 57. London School of Eco-nomics, June 1989.

Hawkins, Gregory D. ‘‘Essays on Non-publiclyIssued Debt: Revolving Credit Agreements andthe Pricing of Privately Placed Debt.’’ Ph.D.dissertation. Massachusetts Institute of Tech-nology, August 1982.

Huberman, Gur, and Charles M. Kahn. ‘‘LimitedContract Enforcement and Strategic Renegotia-tion,’’ American Economic Review, vol. 78(June 1988), pp. 471–84.

Investment Dealers’ Digest, ‘‘Underwritten IssuesCould Spur Expansion of Rule 144A Market,’’IDD, January 6, 1992.

Jacklin, Charles J., and Sudipto Bhattacharya.‘‘Distinguishing Panics and Information-basedBank Runs: Welfare and Policy Implications,’’

Journal of Political Economy, vol. 96(June 1988), pp. 568–92.

James, Christopher. ‘‘Some Evidence on theUniqueness of Bank Loans,’’ Journal of Finan-cial Economics, vol. 19 (December 1987),pp. 217–35.

Kahan, Marcel, and Bruce Tuckman. ‘‘Do Bond-holders Lose from Junk Bond CovenantChanges?’’ Working Paper. New York:New York University, Salomon Center,Leonard N. Stern School of Business,February 1992.

Kalay, Avner. ‘‘Stockholder–Bondholder Conflictand Dividend Constraints,’’ in Michael C.Jensen and Clifford W. Smith, Jr., eds., TheModern Theory of Corporate Finance, 1st ed.New York: McGraw Hill, 1984, pp. 751–73.

Keefe, Victoria. ‘‘Underwritten 144A DealsSurge,’’ Corporate Financing Week, August 31,1992, pp. 1 and 10.

Kessel, Reuben. ‘‘A Study of the Effects ofCompetition on the Tax-exempt Bond Market,’’Journal of Political Economy, vol. 79 (July-August 1971), pp. 706–38.

Kwan, Simon, and Willard T. Carleton. ‘‘TheStructure and Pricing of Private PlacementCorporate Loans.’’ Working Paper. Tucson,Ariz.: University of Arizona, College ofBusiness and Public Administration, Febru-ary 1993.

Laber, Gene. ‘‘Bond Covenants and FinancialManagement: Some Comparisons of Public andPrivate Debt.’’ Working Paper. Burlington, Vt.:University of Vermont, School of Business,January 1992.

Lang, Larry H.P., and Rene M. Stulz. ‘‘Intra-Industry Competition and Contagion Effects ofBankruptcy Announcements.’’ Working Paper.New York: New York University, Leonard N.Stern School of Business, 1991.

Lummer, Scott L., and John J. McConnell.‘‘Further Evidence on the Bank Lending Processand the Capital Market Response to Bank LoanAgreements,’’ Journal of Financial Economics,vol. 25 (November 1989), pp. 99–122.

108

Marais, Laurentius, Katharine Schipper, and AbbieSmith. ‘‘Wealth Effects of Going Private forSenior Securities,’’ Journal of FinancialEconomics, vol. 23 (June 1989), pp. 155–91.

McDaniel, M.W. ‘‘Bondholders and Stockholders,’’Journal of Corporation Law, vol. 13 (Winter1988), pp. 206–13.

Merton, Robert C. ‘‘On the Pricing of CorporateDebt: The Risk Structure of Interest Rates,’’Journal of Finance, vol. 29 (May 1974),pp. 449–70.

Modigliani, Franco, and Merton Miller. ‘‘The Costof Capital, Corporation Finance, and the Theoryof Investment,’’ American Economic Review,vol. 48 (June 1958), pp. 261–97.

Morison, Francis J. ‘‘Regulation S and Its Relationto Rule 144A,’’ in Edward F. Greene and others,eds., Rule 144A: The Expanded Private Place-ment Market. Englewood Cliffs, N.J.: Prentice-Hall Law & Business, 1990, pp. 143–49.

Myers, Stewart C. ‘‘Determinants of CorporateBorrowing,’’ Journal of Financial Economics,vol. 5 (November 1977), pp. 147–75.

Nakamura, Leonard I. ‘‘Commercial Bank Infor-mation: Implications for the Structure ofBanking,’’ in Michael Klausner and Lawrence J.White, eds., Structural Change in Banking.Homewood, Ill.: Irwin, 1993.

Osborne, Alfred E. ‘‘Rule 144 Volume Limitationsand the Sale of Restricted Stock in the Over-the-Counter Market,’’ Journal of Finance,vol. 37 (May 1982), pp. 505–17.

Owens, Raymond E., and Stacey L. Schreft.‘‘Identifying Credit Crunches.’’ Working Paper.Richmond, Va.: Federal Reserve Bank,June 1992.

Petersen, Mitchell A., and Raghuram G. Ragan.‘‘The Benefits of Firm–Creditor Relationships:Evidence from Small Business Data.’’ WorkingPaper. Chicago, Ill.: University of Chicago,Graduate School of Business, 1992.

Press, Eric G., and Joseph B. Weintrop.‘‘Accounting-Based Constraints in Public and

Private Debt Agreements,’’ Journal of Account-ing and Economics, vol. 12 (January 1990),pp. 65–95.

Private Placement Reporter. ‘‘Industry Notes 144ATrends,’’ PPR, March 9, 1992a.

. ‘‘Fund Managers Petition SEC forShare of the 144A Wealth,’’ PPR, August 10,1992b.

Prowse, Stephen D. ‘‘Institutional InvestmentPatterns and Corporate Financial Behavior inthe United States and Japan,’’ Journal ofFinancial Economics, vol. 27 (September 1990),pp. 43–66.

Pugel, Thomas A., and Lawrence J. White. ‘‘AnAnalysis of the Competitive Effects of AllowingCommercial Bank Affiliates To UnderwriteCorporate Securities,’’ in Ingo Walter, ed.,Deregulating Wall Street. New York: JohnWiley & Sons, 1985, pp. 93–139.

Ramakrishnan, Ram T.S., and Anjan V. Thakor.‘‘Information Reliability and a Theory ofFinancial Intermediation,’’ Review of EconomicStudies, vol. 51 (July 1984), pp. 415–32.

Shapiro, Eli, and Charles Wolf. The Role ofPrivate Placements in Corporate Finance.Boston, Mass.: Harvard University, School ofBusiness Administration, Division of Research,1972.

Sharpe, Stephen A. ‘‘Asymmetric Information,Bank Lending and Implicit Contracts: AStylized Model of Customer Relations,’’ Journalof Finance, vol. 45 (September 1990),pp. 1069–87.

Shockley, Richard, and Anjan V. Thakor. ‘‘Infor-mation Content of Commitments to Lend in theFuture: Theory and Evidence on the Gains fromRelationship Banking.’’ Working Paper. Bloom-ington, Ind.: Indiana University, January 1993.

. ‘‘Credit Rationing, ConcessionaryLending, and Debt Maturity,’’ Journal ofBanking and Finance, vol. 15 (June 1991),pp. 581–604.

109

Smith, Clifford W. Jr. ‘‘Investment Banking andthe Capital Acquisition Process,’’ Journal ofFinancial Economics, vol. 15 (January-February 1986), pp. 3–29.

, and Jerold B. Warner. ‘‘On FinancialContracting: An Analysis of Bond Covenants,’’Journal of Financial Economics, vol. 7(June 1979), pp. 117-61.

Summers, Lawrence, and David M. Cutler. ‘‘TheCost of Conflict and Financial Distress: Evi-dence from the Texaco–Pennzoil Litigation,’’Rand Journal of Economics, vol. 19 (Summer1988), pp. 157–72.

Swary, Itzhak, and Gregory F. Udell. ‘‘InformationProduction and the Secured Line of Credit.’’Working Paper. New York: New York Univer-sity, Leonard N. Stern School of Business,1988.

Szewczyk, Samuel H., and Raj Varma. ‘‘RaisingCapital with Private Placements of Debt,’’Journal of Financial Research, vol. 14(Spring 1991), pp. 1–13.

The Travelers Securities Department, PrivatePlacement Division. ‘‘Private Placements,’’1992.

Treitler, Betsy. ‘‘Unlocking the World of PrivatePlacements,’’ Investment Dealers’ Digest,March 1, 1993, pp. 14–20.

Udell, Gregory F. ‘‘Pricing Returned CheckCharges under Asymmetric Information,’’Journal of Money, Credit, and Banking, vol. 18(November 1986), pp. 495–505.

. ‘‘Loan Quality, Commercial LoanReview and Loan Officer Contracting,’’ Journalof Banking and Finance, vol. 13 (July 1989),pp. 367–82.

U.S. Securities and Exchange Commission. ‘‘AnAnalysis of Regulation D,’’ 1984.

. SEC Docket, 42, Washington,November 1988.

. Resale of Restricted Securities;Changes to Method of Determining HoldingPeriod of Restricted Securities under Rules 144

and 145: Final Rule, Rule Amendments, andSolicitation of Comments, April 23, 1990a.

. ‘‘Offshore Offers and Sales: FinalRules, Amendments and Solicitation of Com-ments,’’ in Federal Securities Law Reports,May 2, 1990b.

. ‘‘Staff Report on Rule 144A,’’September 30, 1991.

. ‘‘Private Resales of Securities toInstitutions: Proposed Rule Amendments,’’July 16, 1992.

. ‘‘Staff Report on Rule 144A,’’January 23, 1993.

Vachon, Michael. ‘‘Coming to America,’’ Invest-ment Dealers’ Digest, May 11, 1992a,pp. 18–25.

. ‘‘Too Much Cash, Too Few Deals,’’Investment Dealers’ Digest, August 31, 1992b,pp. 23–24.

Vora, Gautam. ‘‘Stock Price Movements UponPrivate Placements of Corporate Bonds.’’Working Paper. Albuquerque, N.M.: Universityof New Mexico, Anderson School of Manage-ment, May 1991.

Wansley, James W., Fayez A. Elayan, and M. CaryCollins. ‘‘Firm Quality and Investment Opportu-nities: An Empirical Investigation of theInformation in Bank Lines of Credit.’’ WorkingPaper. Knoxville, Tenn.: University of Tennes-see, May 1991.

Warga, Arthur, and Ivo Welch. ‘‘BondholderLosses in Leveraged Buyouts.’’ First BostonWorking Paper Series. New York: CS FirstBoston Corporation, May 1990.

Warner, Jerold. ‘‘Bankruptcy Costs: Some Evi-dence,’’ Journal of Finance, vol. 32(May 1977), pp. 337–47.

Weigley, Roger D. ‘‘Underwritten OfferingsGaining Under Rule 144A.’’ Letter to theSecurities and Exchange Commission, PrivatePlacement Reporter, December 16, 1991.

Weiss, Lawrence A. ‘‘Bankruptcy Resolution:Direct Costs and Violation of Priority of

110

Claims,’’ Journal of Financial Economics,vol. 27 (October 1990), pp. 285–314.

Wolf, Charles R. ‘‘The Demand for Funds in thePublic and Private Corporate Bond Markets,’’Review of Economics and Statistics, vol. 56(February 1974), pp. 23–29.

Zinbarg, Edward D. ‘‘The Private Placement LoanAgreement,’’ Financial Analysts Journal,July-August 1975, pp. 33–35 and 52.

Zwick, Burton. ‘‘Yields on Privately PlacedCorporate Bonds,’’ Journal of Finance, vol. 35(March 1980), pp. 23–29.

FRB1/1–2200–12/93

111