Financing bank stock ownership

Journal of Banking and Finance 4 (1980) 111-124. © North-Holland Publishing Company

FINANCING BANK STOCK OWNERSHIP

A Question of Conflict of Interest*

Larry G. MEEKER and Forest E. MYERS Federal Reserve Bank of Kansas City, Kansas City, MO 64198, USA

Received January 1979, final version received April 1979

The bank stock loan conflict of interest question arises when compensating balances are intermingled with a bank's correspondent balances for the benefit of those bank stockholders seeking a bank stock loan. This study attempts to determine if this practice exists using two- stage least square regression analysis and cross-sectional data obtained from one-bank holding company applications in the Tenth Federal Reserve District. Our results suggest that bankers with established correspondent banking relationships capitalize on their correspondent balances to obtain favorable interest rates on bank stock loans.

1. Introduction

Borrowing against bank stock is a relatively common practice by the controlling owners of the Nation's smaller banks. It is a practice that tends to increase the marketability and liquidity of large controlling blocks of closely held bank shares by making it easier for new owners to buy those shares and for current owners to raise cash by refinancing their shares. The practice, however, has been subject to critical review because it is not completely separable from other official bank business. Those controlling owners receiving bank stock loans are often officers and directors who have the opportunity to obtain favorable loan terms by directing bank correspondent business to the lender. Hence, the opportunity exists for intermingling personal and bank business. This raises questions of conflicts of interest both from a regulatory perspective and from the perspective of minority shareholders who generally cannot obtain bank stock loans with terms comparable to those received by the bank's controlling shareholders.

In the correspondent banking relationship, small banks (respondent banks) usually seek a variety of services including check clearing, loan participations, and financial advice from larger, more centrally located banks (correspondent banks). Check clearing and many operating requirements as

*The views expressed in this paper are those of the authors and do not necessarily reflect those of the Federal Reserve Bank of Kansas City or the Federal Reserve System.

112 L.G. Meeker and F.E. Myers, Financing bank stock ownership

well as state reserve requirements necessitate correspondent balances. The compensation for these correspondent services usually takes the form of non- interest bearing deposits. Competition for these interest free deposits among correspondent banks provides an incentive for offering interest rate reductions to stockholders on bank stock loans. The economic advantage of bank stock loan rate reductions as opposed to other methods of payment, such as an increase in other free services, is two-fold. First, since an increase in free services would increase bank profitability, maintenance of idle correspondent balances to support an interest rate reduction on a bank stock loan shifts potentially taxable bank profits from the bank directly to the bank's shareholders thus avoiding the corporate taxation of those revenues. Hence, while service related costs may make the lender indifferent to the form of the remuneration, the borrower is not indifferent because of tax considerations. Second, the controlling shareholders in a bank, because they frequently manage the bank and its correspondent accounts, have an opportunity to further augment their gain by obtaining compensation for interest free correspondent balances through favorable bank stock loan terms. Minority shareholders are not in a position to utilize the correspondent balances because they have no control over them. As a consequence, benefits derived from correspondent balances flow only to controlling shareholders and exclude minority shareholders. All bank stock loans in this study involve controlling blocks of shares in excess of 50 percent of the bank's outstanding common shares.

This study attempts to determine if correspondent balances influence rates on bank stock loans to controlling shareholders using two-stage least square regression analysis and cross-sectional data obtained from one-bank holding company applications in the Tenth Federal Reserve District. The results indicate that bankers with established correspondent banking relationships capitalize on their correspondent balances to obtain favorable interest rates on bank stock loans.

2. Previous studies

The bank stock lending issue has received more political and accusatory attention than empirical study. However, the field is not without some empirical work. In 1964, the United States House Committee on Banking and Currency initiated a study in which it was noted that 'the rate of interest charged on bank stock loans approaches and in some instances may go below the prime rate, and the loan-to-value ratio . . . may run from 90 to 100 percent'.1

In another study by Joy and Meeker involving banks in the Tenth Federal Reserve District, it was reported that 'on average for all years (1964-1975),

1U.S. Congress (1964, p. 12).

L.G. Meeker and F.E. Myers, Financing bank stock ownership 113

the ratio of the total dollar amount of debt used to purchase bank stock to the total dollar amount of bank stock purchased was about 87 percent'. 2

More recently, a survey done by the Federal Deposit Insurance Corporation, the Federal Reserve Board, and the Comptroller of the Currency for the U.S. Senate Committee on Banking, Housing, and Urban Affairs indicated that about 81 percent of all bank stock loans of $25,000 or more were undertaken to acquire bank or bank holding company stock and that about 82 percent of the loans were to insiders. 3 Furthermore, their data indicated that with the exception of 1974 when the prime rate averaged over 10 percent, bank stock loan rates generally exceeded the prime rate. 4 All of these studies, however, have been descriptive in nature and have not reflected directly on the conflict of interest question.

The conflict of interest question was addressed most directly by Peterson and McLaughlin who attempted to answer three questions: '(1) Do stock loans affect the proportion of correspondent balances held? (2) Does the existence of stock loans lead to lower bank profits? (3) Are banks with stock loans deficient in serving their local communities? '5 Their answer to the conflict of interest question was indecisive even by their own standards. 'The evidence indicates that banks with stock loans tend to have larger correspondent balances than banks without them; but, on the average, this does not affect their cash and due ratios, profit rates or dividend payments, nor does it reduce their participation in the loan market. In addition, the effective interest cost argument (that low interest loans are backed by correspondent balances of such size as to bring the returns on the loan up to market rates) is not supported by the regression analysis results. '6

In summary, the conflict of interest question has received little analytical attention and certainly has not been resolved.

3. Data

The data for this study were compiled from three sources: one-bank holding company application files at the Federal Reserve Bank of Kansas City, bank examination reports, and bank Reports of Condition. Data from 80 one-bank holding company applications that involved a bank stock loan from another bank were obtained. The subject banks ranged in asset size from $2.8 to $299.9 million with the mean bank having about $19 million in assets. These observations included all banks that applied for one-bank holding company status that met several criteria. First, the observations had

2Joy and Meeker (1979, p. 3). 3U.S. Congress (1978, p. 1). 4U.S. Congress (1978, p. 4). SPeterson and McLaughlin (1974, p. 8). 6Peterson and McLaughlin (1974, p. 12).


to be informationally complete with respect to a loan commitment date, a loan interest rate, loan amount, number of shares of stock purchased, stock purchase price, and collateral pledged against the loan. Second, observations prior to 1973 and after early 1976 were not used because of lack of easy access to data for those time periods.

Third, observations were excluded if the lending bank could not be identified as a correspondent bank of the bank whose stock was hypothecated at the time of the loan commitment. This exclusion criteria was necessary because the model is designed to detect relationships between correspondent balances and bank stock loan rates where existing relationships exist and not where new relationships are being formed. Very few bank stock loan observations were excluded for this reason because bank stock loans made to bank holding companies are almost exclusively transfers of debt from individuals to the holding company. Finally, it should be noted that in almost all instances, the bank loan rates in our sample were renewable annually and none exceeded a two year renewal period. Had the loan rates not been frequently renewable, we would have had to invoke some exclusion criteria for longer renewal periods since our objective is to examine current bank stock lending practices where lenders have the option of changing or setting loan rates.

4. Methodology

The question of conflict of interest in bank stock lending practices is considered in this study from the perspective of the lender. A bank stock loan is treated as any other risky investment where return is assumed to be a function of a risk free interest rate plus a risk premium. In mathematical terms, this can be expressed as

Rt .=I+P, (1)

where RL is the rate of return on a loan, I is the risk free rate of interest, and P is a risk premium. With many bank business loans, R L may typically consist of two parts: an explicit interest payment, Re, and an implicit interest payment, Ri, tied to compensating balances. Incorporating this observation into eq. (1) yields the expression

Re+ Ri= I + P,

which can be rearranged as follows:

R e = I + P - R i.

(2)

(3)


Eq. (3) expresses the explicit interest rate on a loan as a function of the risk free interest rate, a risk premium, and any implicit return payments made in the form of compensating balances.

Since implicit interest payments for bank stock loans, if they exist, are neither formally documented nor acknowledged, one of the major tasks in exploring the conflict of interest question in bank stock lending is identifying a proxy variable for the Ri term in eq. (3) which would capture any effects of compensating balances. The obvious choice is some measure of correspondent balances since they could serve as compensating balances. This proxy variable choice, however, can present some problems because one of the determinants of correspondent balances may be the loan rate consideration given on a bank stock loan. The solution to the problem is the employment of two stage least squares regression analysis whereby the jointly determined variables, explicit bank stock loan rates and correspondent balances, can be simultaneously determined.

A bank's level of correspondent balances is determined by its use of correspondent services, S, the banking environment 7 of the state in which the bank operates, E, and the explicit interest payment, Re, on any bank stock loan from the correspondent bank. Our data consists primarily of loan commitments that involve transferring bank stock debt from principals in a bank to a bank holding company. As a result, the explicit interest rate on a bank stock loan, R e, may be a relevant factor in explaining a bank's correspondent balances at the time of the loan commitment to the holding company. Functionally, correspondent balances, CB, can be expressed as

CB=f(S,E, Re). (4)

Now, replacing Ri in eq. (3) by its functional equivalent, g(CB), yields the expression

Re=I+P-g(CB). (5)

Expressions (4) and (5) are thus the nucleus of the two-stage least square problem where R e and CB are jointly determined variables. All else being equal, higher interest rates on bank stock loans should lessen the need for compensating loan balances and hence correspondent balances; and higher correspondent balance levels should make possible lower interest rates on bank stock loans.

7A state's banking environment is a product of both regulatory and economic factors. Included among the former are reserve requirements, definition of reservable assets, and restrictions on chartering, mergers, and other methods of expansion: Included among the latter are distance to financial centers, competition for correspondent services, and a variety of other considerations related to geography, mix of economic activities, and so forth.


5. The regression equations

Neither eq. (4) nor (5) can be estimated directly because certain of the variables in each equation are not observable in their expressed form. As a result, proxy variables must be defined for those variables. The regression equations used in this study are

Re=A o +A~ T-Bill+, A2LN/COL+ A3E/INT+A4DDF/DD , (6)

and

DDF/DD = B 0 + B 1TD + B2DD/TD + B 3 VARDD + B4MBR + BsCA A

+ B60K + BTNE + BaCOLO + BgWYO + B1 oMO + B11Re,

(7)

where A o and B o are constants designed to absorb the effects of any omitted variables. The variables in the interest rate equation, eq. (6), are defined as follows:

T-Bill

LN/COL - -

E / INT

DDF/DD

The one year Treasury Bill rate at the time of the bank stock loan commitment. The ratio of the amount of the bank stock loan to the market value of the collateral pledged against the loan. In all cases, the collateral consisted of controlling shares of the purchased bank's stock. Market value of the stock was taken to be the value placed on the stock in the holding company acquisition and was generally the purchase price of the stock. The ratio of the bank's past yearly earnings before taxes to the first year's interest expense on the bank stock loan.

- - T h e ratio of a bank's 'demand balances with banks in the U.S.' (commonly called demand due from, DDF) to its demand deposits, DD. DDF includes demand balances held with all domestic commercial banks and not just the bank granting the bank stock loan. Both DDF and DD are computed from four quarter averages to compensate for seasonality and random- ness. The averages are centered on the loan commitment date, i.e., data from two quarters prior to the loan commitment date and two quarters after the loan commitment date are used to compute DDF and DD. The hat over DDF/DD in eq. (6) indicates that DDF/DD is an estimated value obtained from regression eq. (7) in the two-stage least squares estimation procedure.


The one year Treasury Bill rate is a proxy for the risk free interest rate in eq. (5). A one year rate was chosen because all of the bank stock loans in the sample were renewable annually. Hence, the Treasury Bill rate proxies the return on a risk-free investment similar in other respects to the riskier bank stock loan. Its coefficient, At, is expected to have a positive sign in that higher risk free interest rates should necessitate higher returns on risky investments.

The risk variable in eq. (5) is proxied by two variables in eq. (6), LN/COL and E/INT. These variables represent two important types of lending risk, collateral risk and loan performance risk, respectively.

Collateral risk concerns the likelihood of the lender recovering his investment plus interest should foreclosure be necessary. Performance risk, on the other hand, is associated with deviations in scheduled interest and principal payments that hinder the flexibility of the lender tomake forward commitments to other borrowers from expected loan proceeds. The coefficient of the collateral variable, A2, is expected to be positive while the coefficient of the performance variable, A3, is expected to be negative so that in both cases higher risk is reflected in higher bank stock loan rates, all else being equal.

Finally, the correspondent balance measure, DDF/DD, in eq. (6) is a proxy for the implicit interest payments made on bank stock loans. Correspondent balances with all correspondents rather than with the correspondent bank granting the bank stock loan are used because we cannot specifically identify the correspondent services obtained from each correspondent bank. In almost all instances, however, the bank stock loan was obtained from the primary correspondent. Additionally, it was assumed that the ratio of collected to uncollected balances was constant across the entire sample. The coefficient of DDF/DD, A4, should be negative if correspondent balances serve as compensating balances for bank stock loans. If not, A4 should not be statistically different from zero.

The variables in the correspondent balance equation, eq. (7), are defined as follows:

TD - - Total deposits of the respondent bank at year-end prior to the loan commitment date.

DD/TD - - T h e ratio of demand deposits to total deposits for the respondent bank at year-end prior to the loan commitment date.

VARDD - - T h e coefficient of variation of the respondent bank's demand deposits computed from quarterly figures four quarters prior to the loan commitment date.

MBR - - A dummy variable which is one if the respondent bank is a member of the Federal Reserve System and zero otherwise.

CA - - The number of correspondent banks the respondent I bank has accounts with.


OK

NE

COLO

WYO

MO

- - A dummy variable which is one when the respondent bank is located in Oklahoma and zero otherwise.

- - A dummy variable which is one when the respondent bank is located in Nebraska and zero otherwise.

- - A dummy variable which is one when the respondent bank is located in Colorado and zero otherwise. A dummy variable which is one when the respondent bank is located in Wyoming and zero otherwise. A dummy variable which is one when the respondent bank is located in Missouri and zero otherwise.

Our dependent variable is a scaled version of correspondent balances, as necessitated in eq. (6), incorporating the respondent bank's total demand deposits, DD, a variable considered to be highly important in determining a bank's correspondent balances [see Lawrence and Lougee (1970) and Meinster and Mohindru (1975)]. Hence, the objective in regression eq. (7) is to identify factors that influence correspondent balances relative to demand depisits and, more importantly, to separate those service factors that influence correspondent balances from any loan agreement factors that might influence balances.

Total deposits, TD, is entered in the regression to reflect size characteristics of the respondent bank. Large banks generally require more correspondent services than small banks. However, a bank's capacity for performing services for itself that could be done by a correspondent increases with bank size. The net effect of these factors should be reflected in lower ratios of correspondent service usage relative to bank size, DDF/DD, as bank size increases. As a result, B1 is expected to be negative.

An important service consideration in past studies [Lawrence and Lougee (1970) and Meinster and Mohindru (1975)] is the relationship between demand deposits and total deposits. Generally, the higher the ratio of demand deposits to total deposits, DD/TD, the greater the demand for correspondent services. However, the arguments above indicate that service demands from a correspondent are not expected to increase proportionally with demand deposits [Dewald and Dreese (1970) and Lawrence and Lougee (1970)]. Since our dependent variable is scaled by demand deposits, the expected relationship between DD/TD and DDF/DD is negative. Hence B z should be negative.

Liquidity is also a factor influencing the need for correspondent balances. The more volatile a bank's deposits, the greater the need for correspondent balances to ensure maintaining a minimum correspondent balance level [Dewald and Dreese (1970) and Meinster and Mohindru (1975)]. ~ Were the data readily available, a volatility measure of daily deposit fluctuations


would be much more desirable than our quarterly observations of deposit levels used to compute VARDD. In any event, the coefficient of VARDD, B3, is expected to be positive.

To the extent that banks which are members of the Federal Reserve System can obtain most correspondent services free of charge or for nominal charges, one would expect that member banks would require fewer correspondent services than non-member banks. As a result, the coefficient of the dummy variable MBR, B4, is expected to be negative reflecting lower correspondent balances for member banks. This variable, however, may not have a great influence on the correspondent balances of the banks in our sample because of bank size. For the 80 percent of our data consisting of small banks with under $15million in deposits, both member and non- member banks have similar levels of correspondent balances. This does not appear unusual since the small banks are frequently located away from larger financial centers and find it more profitable, because of traditional clearing patterns and credit availability, to utilize a correspondent bank even when they have access to free services from the Federal Reserve. Hence, Federal Reserve membership is not expected to be a particularly important factor in the correspondent balance regression.

A final factor considered as a measure of relative correspondent balance service needs is the number of correspondent accounts a respondent bank has with different correspondents, CA. In one sense, it is a size measure in that larger banks might be expected to have more correspondent accounts [see Knight (1970) and Lawrence and Lougee (1970)]. To the extent that CA is a proxy for size, its expected relationship with DDF/DD is negative as is our other size measure, total deposits. Alternatively, any fixed costs to correspondent accounts which might not be fully recovered, as might be the case in less active accounts, would tend to cause the relationship between CA and DDF/DD to be positive. As a result, the expectations about this variable are somewhat indeterminant.

To account for differences in the banking environments 8 among the states in which the bank stock loan observations have been drawn, dummy variables have been entered for the states of Oklahoma, Nebraska, Colorado, Wyoming, and Missouri. The state of Kansas is implicitly, defined in the regression equation as the standard against which other states are compared.

Finally, an estimate of the explicit bank stock loan rate obtained in the two-stage least squares process from eq. (6), Re, is entered as a determinant of DDF/DD. Ceteris paribus, lower bank stock loan rates should be associated with higher levels of DDF/DD if correspondent balances are acting as compensating balances for a bank stock loan. Hence, Bl l is expected to

SSee footnote 7 for a list of factors which may vary among states.

120 L.G. Meeker and F~E. Myers, Financing bank stock ownership

be negative. This coefficient and A4 in eq. (5) form the crucial tests of our hypothesis.

One highly important consideration not expressed in the variables previously described is the length of time the recipient of the bank stock loan has owned the bank whose stock is hypothecated. The presence of a new owner in a bank may have several potential influences on the relationship between correspondent balance levels and the interest rate on a bank stock loan. First, the new owner may represent an unproven quantity with respect to bank management and performance, hence, increasing the risk associated with granting a bank stock loan. While this type of risk can be resolved with time, it may cause the initial terms on a bank stock loan to a new customer to be less favorable than they would otherwise be were the risk not present. Second, correspondent account start-up costs make bank stock loan rate concessions more likely on established accounts where those costs have already been absorbed as opposed to new accounts where those costs are still being recovered. Both of these factors are consistent with the banking practice of giving established business customers more favorable loan terms than new customers. Furthermore, they would tend to change the relationship between a bank stock loan rate and the correspondent balance account and would reduce the loan rate concessions for balances maintained. Third, the presence of a new owner may make historical correspondent balance levels less reliable as indicators of future balance levels than if the borrower had owned the bank for a considerable length of time. This is evidenced in our data by large changes in DDF/DD from one year prior to the bank stock loan to one year after the bank stock loan for the newly acquired banks. These changes are statistically highly significant compared to the small changes in DDF/DD that take place in the sample of banks with established owners. Thus, it would be difficult for the lender to utilize the correspondent balance account as part of the compensation for a bank stock loan. Implicit in this argument is the assumption that compensating balances for a bank s tock loan are not an explicit part of the loan agreement, but are rather spread throughout the entire correspondent account. As a result, our data is divided into two sets to reflect the apparently different populations, one consisting of loans to new owners of banks and one consisting of loans to established bank owners. Regressions are run on each data set separately. An ownership period of one year prior to the bank stock loan commitment date was chosen as the dividing line between new and established owners.

6. Empirical results

The empirical results are presented in table 1. Panel A of that table contains the results for the two-stage least squares regressions on the

• L.G. Meeker and F.E. Myers, Financing bank stock ownership 121

established owner data set which consists of 46 observations and Panel B of that table contains the regression results for the new owner data set which consists of 34 observations. Since there were no Wyoming observations in the new owner data set and no Missouri observations in the established

Table 1

Two-s tage least square regression results?

(A) Established owner regressions

(B) New owner regressions

Standard Standard Var iab le Coefficient e r ro r Coefficient er ror

Loan rate regression [eq. (6)'1

In te rcep t 7.361"* 1.101 9.161"* 1.259 T-Bill 0.205" 0.138 0.074 0.153 L N / C O L 0.178 0.736 - 1.726"* 0.848 E / I N T - 0.071" 0.042 0.017 0.074 DDF/DD - 0.056"* 0.028 - 0.050 0.042

R 2 0.18 0.25

F 2.32* 2.36*

Correspondent balance regression [eq. (7)]

In te rcep t 44.740** 21.304 - 9.208 28.187 TD - 0.0001 0.0001 - 0.0002 0.0002

DD/TD - 0.054 0.106 - 0.231 0.214 VARDD 0.423"* 0.185 - 0.579 0.384 M B R - 1.150 2.317 - 2 . 5 2 3 2.947 CA b - 1.084" 0.632 - 1.584 1.279

OK 6.357** 2.917 10.615"* 5.967 N E - 2.565 2.719 - 3.639 3.825 COLO 4.821 4.677 - 12.021"* 6.127 W Y O 8.535** 4.195 - - - - MO - - - - - 3.720 5.728

R e - 3.221 2.863 6.415 5.256

R 2 0.49 0.45

F 3.38** 1.91"

aOne asterisk indicates significance at a 90 percent confidence level. T w o asterisks indicate

s ignif icance at a 95 percent confidence level. bA two tailed test is used for this variable since there were no clear expecta t ions about the

variables sign.

owner data set, no statistical results for these variables are reported in their respective places in table 1.

As can be seen from Panel (A) in table 1, al l of the variables in the established owner regressions have the expected signs and many are


statistically significant. In the loan rate regression, the only variable lacking statistical significance at a 90 percent confidence level is the collateral risk variable, LN/COL. This is not surprising since in practice bank stock loans are considered to be almost riskless as far as capital recovery is concerned. It is interesting, however, that the timely repayment of interest, as reflected in the performance risk variable, E/INT, is significant. This variable might also reflect the risk of having to resort to collateral value in a foreclosure. Finally, and most important, the coefficient of the correspondent balance variable, DDF/DD, is negative and statistically significant at a 95 percent confidence level. This is important evidence that correspondent balances serve, in part, as compensating balances on bank stock loans for established bank owners.

The bottom portion of Panel (A) in table 1 presents the estimated equation for correspondent balances. While all of the variables in this regression have the expected signs, with the possible exception of CA for which we had no clear expectations, several of the variables lacked statistical significance. Total deposits, TD, the ratio of demand deposits to total deposits, DD/TD, Federal Reserve membership, MBR, and the explicit bank stock loan rate, Re, all failed to bei signficant determinants of our correspondent balance measure, D.DF/DD. The number of correspondent banks a respondent bank does business with, CA, is significant and negatively related to DDF/DD, possibly reflecting the CA variable's role as a proxy for bank size. This result is consistent with other research [Knight (1970) and Lawrence and Lougee (1970)-I. Variability of demand deposits, VARDD, is also a significant determinant of correspondent balances as are the state dummy variables for~ Oklahoma and Wyoming. The state dummy variables indicate that banks in Oklahoma and Wyoming keep significantly higher correspondent balances than banks in Kansas, the state used as the basis of comparison.

The new owner regressions, Panel (B) of table 1, produce considerably poorer results than the established owner regressions in terms of variable signs and statistical significance. The collateral risk variable, LN/COL, is the only variable with statistical significance in the loan rate regression. However, its sign is the opposite of that expected. The performance risk variable, E/INT, is not Statistically significant as it was in the established owner regression. This is not surprising since the occurrence of a recent ownership change of the bank makes the use of historical earnings a questionable indicator of ability to meet current and future interest payments. The correspondent balance variable has the correct sign but is not statistically significant. In the correspondent balance regression, the Oklahoma and Colorado dummy variables are the only variables with statistical significance and the Colorado dummy variable has a sign opposite that in the established owner regression.


One major conclusion to be drawn from the comparative results of the regressions on the two data sets is that established bank owners and not new bank owners are able to capitalize on their correspondent banking relationships to obtain interest rate reductions on bank stock loans in return for having higher correspondent balances. This is consistent with the banking practice of allowing established business customers to capitalize on their demand balances when obtaining business loans. Perhaps with the passage of time, the new bank owners will be able to capitalize on their correspondent balances as their correspondent banking relationships become more established.

7. Summary, implications and conclusions

We have found evidence in this study, using a two-stage least squares regression analysis of a loan rate equation and a correspondent balance equation, that established bank owners but not new owners of banks capitalize on their correspondent balances to obtain bank stock loan rate reductions. In a broader context, this is consistent with the bank lending practice of giving special consideration to established customers as a means of securing and maintaining their busindss.

The implications of this study regarding the conflict of interest question in bank stock lending must be carefully drawn. The empirical indication that established owners and not new owners of banks can use correspondent balances to their advantage in obtaining bank stock loan rate reductions implies that there is no universal practice of granting bank stock loan rate reductions for higher correspondent balances. However, as long as some established controlling owners of banks (keep in mind that new owners by definition one day become established owners) can personally benefit from their bank's official business activities at the expense of minority shareholders, the conflict of interest question exists because i t implies that correspondent balances exceed those needed for services provided to the bank.

One factor potentially complicating the conflict of interest question is state reserve requirements. It is conceivable that in some instances correspondent balances kept in 'excess' of those needed for services might be kept to meet reserve requirements. While our study has not dealt directly with the sources of any 'excess' correspondent balances, it seems highly doubtful, given the wide range of ratios of correspondent balances to total assets observed in our sample and the generally lower levels of reserve requirements, that this would be a factor contributing significantly to 'excess' correspondent balances.

Another factor often considered relevant to the conflict of interest question is the actual level of interest rates on bank stock loans relative to other loan

124 L.G. Meeker and F.E. Myers, Finandng bank stock ownership

rates and market interest rates. Our study does not deal with this question. To make such comparisons, one would have to adjust the rates for numerous considerations not encompassed in this study. For example, a common comparison is between bank stock loan rates and the prime rate. However, the prime rate embodies varying, but agreed upon compensating balances of business customers and often includes other banking services. It is thus quite difficult to make reasonable rate comparisons when one must focus on only one portion of a much larger business relationship.

As suggested at the beginning of the paper, competition for interest free correspondent balances provides an incentive for offering interest rate reductions on bank stock loans. Should banking practices change to a system of explicit pricing whereby correspondent services were individually priced and interest were paid on reserves and correspondent balances, bank stock lending practices would likely undergo considerable revision. Under such a market pricing of services and balances, correspondents would have little latitude for offering interest rate reductions on bank stock loans. Although corporate taxes and the existence of minority shareholders would continue to provide an incentive for controlling shareholders to use bank correspondent balances to obtain interest rate reductions on their bank stock loans, the explicit pricing of correspondent services would focus attention on prices and arrangements which deviated significantly from the norm. This would likely attract regulatory attention as well as increase the threat of minority shareholder lawsuits and thus further discourage any misapplications of bank resources.

References

Dewald, W.G. and G.R. Dreese, 1970, Bank behavior with respect to deposit variability, Journal of Finance XXV, Sept., 869-879.

Joy, O.M. and L.G. Meeker, 1979, The market for large blocks of bank stock: Evidence from the Tenth Federal Reserve District, Kansas Business Review 3, no. 2, Oct., 1-7.

Knight, R.E., 1970, Correspondent banking - - Part 1: Balances and services, Federal Reserve Bank of Kansas City Monthly Review, Nov., 3-14.

Lawrence, R.J. and D. Lougee, 1970, Determinants of correspondent banking relationships , Journal of Money, Credit, and Banking 2, Aug., 358-369.

Meinster, D.R. and R.K. Mohindru, 1975, Determinants of the demand for correspondent balances by small and medium sized banks, Journal of Bank Research 6, Spring, 25-34.

Peterson, M.O. and H.S. McLaughlin, 1974, Conflict of interest and the financing of commercial bank stock ownership, Journal of Bank Research 5, Spring, 7-12.

U.S. Congress, 1964, House Subcommittee on Domestic Finance, Committee on Banking and Currency, The structure of ownership of member banks and the pattern of loans made on hypothecated bank stock, Subcommittee Print, 88th Congress, 2nd session.

U.S. Congress, 1978, Senate Committee on Banking, Housing, and Urban Affairs, Special survey of bank stock loans, loans to officials and major shareholders of other banks, insider loans, and overdrafts: Part 2, Committee Print, 95th Congress, 2nd session.

Financing bank stock ownership

Documents