Econ 492: Comparative Financial Crises Lecture 2 19 September 2012 David Longworth This material is copyrighted and is for the sole use of students registered.

Econ 492: Comparative Financial Crises

Lecture 219 September 2012

David LongworthThis material is copyrighted and is for the sole use of students registered in ECON 492. This material

shall not be distributed or disseminated to anyone other than students registered on ECON 492. Failure to abide by these conditions is a breach of copyright, and may also constitute a breach of

academic integrity under the University Senate’s Academic Integrity Policy Statement.

Economics 492 Lecture 2 2

• U.S. Federal Open Market Committee decision last Thursday:– “…agreed today to increase policy accommodation

by purchasing additional agency mortgage-backed securities at a pace of $40 billion per month.”

– “…anticipates that exceptionally low levels for the federal funds rate are likely to be warranted at least through mid-2015.”

– Result: “Central banks fuel a questionable euphoria” (G&M, Monday) (S&P 500, TSX, MSCI, commodities)

Financial Crises in the News:


• IMF Global Financial Stability Report Update (July, 2012)– http://

www.imf.org/External/Pubs/FT/fmu/eng/2012/02/index.htm – Place to click there for Excel data on peripheral country and

financial credit spreads• Mendoza and Terrones (2012) “An Anatomy of Credit

Booms and Their Demise,” NBER WP 18379. – “banking crises, currency crises or Sudden Stops often follow

credit booms”– Uses a “threshold method” for identifying booms

New Reference on Crises:

http://www.imf.org/External/Pubs/FT/fmu/eng/2012/02/index.htm

http://www.imf.org/External/Pubs/FT/fmu/eng/2012/02/index.htm


Overview

I. TransmissionII. Policy Response During the Crisis III. Prevention

Note: AG indicates Franklin Allen and Douglas Gale (2009), Understanding Financial Crises. KA indicates Charles P. Kindleberger and Robert Aliber (2005), Manias, Panics, and Crashes. RR indicates Carmen Reinhart and Kenneth Rogoff (2009), This Time is Different.


l. Transmission

• Outline of Course

(Prediction)

Transmission

Policy ResponsePrevention

Causes


I. Transmission

• Illiquidity (“and all its friends”)– Bank runs– Margin and liquidity spirals– Fire Sales (Cash-in-the-market pricing)

• Interconnectedness and contagion• Decline in wealth of private sector: effects on output and

employment• Zero bound on nominal interest rates takes away

conventional monetary policy channel• Effect on sovereign debt crises and vice versa• Longer-run effects on GDP growth


I. Transmission

• Illiquidity (“and all its friends”)– Recall that there is both “funding liquidity” and

“market liquidity”– “all its friends” include (according to Tirole):• Market freezes• Fire sales• Contagion• Ultimately, insolvencies and bailouts• I would include “bank runs” (as one cause) and market

liquidity spirals


I. Transmission

• Bank Runs (AG 3, Diamond & Dybvig)– Banks have liquid liabilities, illiquid assets– So banks are susceptible to unexpected liquidity

demands (bank runs)– Model this by having a liquid asset (short asset) that

doesn’t pay interest, and an illiquid asset that does– Banks (intermediation) solve mismatch between time

preference of customers and asset maturity– Typically, markets are incomplete and so can’t provide

an efficient solution to this mismatch problem


I. Transmission

• Bank Runs (AG 3, Diamond & Dybvig)– In a two-period model with no aggregate uncertainty about

liquidity withdrawals, there is an equilibrium in which the bank provides withdrawals (consumption) c(1) to its depositors at time 1 and c(2) to its depositors at time 2, invests x in the long asset and y in the short asset

– In the same model, if the bank can sell the long asset early (period 1), taking a discount, a bank run will also be an equilibrium. This is because, if all depositors, whether they would normally withdraw to consume at time 1 or time 2, decide to withdraw at time 1, the bank cannot possibly pay them all off.


I. Transmission

• Bank Runs (AG 3, Diamond & Dybvig)– Critics of this type of model have argued that

suspension of convertibility of deposits into cash could stave off bank runs

– But Diamond and Dybvig have shown that a sequential payout by bank tellers would mean that they would not find out until too late that a run was in progress.


I. Transmission

• Bank Runs (AG 3, Diamond & Dybvig)– Equilibrium bank runs:• Impossible to predict• Coordination among individuals facilitated by

“sunspots” (extraneous variables, not “fundamental”)• If “the probability of a bank run is sufficiently small,

there will exist an equilibrium in which the bank is willing to risk a run because the cost of avoiding the run outweighs the benefit.” (AG, p.82)


I. Transmission

• Bank Runs (AG 3, Diamond & Dybvig)– Are bank runs associated with the business cycle (and not

“sunspots”)? Potential paper. Also, how correlated with the leverage cycle (C/Y)? How correlated is the leverage cycle with the business cycle (Y)?• Some support for a yes answer: Gorton’s 1988 study of U.S.

(1865-1914)• Indeed, if bank runs are part of transmission of crises, and

crises are typically associated with the credit cycle, which is highly correlated with the business cycle, no surprise

• Many suspect that liquidity problems are associated with fears of credit problems and perhaps actual credit problems


I. Transmission

• Bank Runs (AG 3, Diamond & Dybvig)– Runs aren’t just from banks (like Northern Rock in

the U.K., and Greek banks recently)• But from “shadow banking system” as well– Canadian asset-backed commercial paper, money

market mutual funds, U.S. financial commercial paper, structured investment vehicles (SIVs), repo market, etc.


I. Transmission

• Margin and Liquidity Spirals– Financial institutions (and large investors) engage in

securities financing transactions• Repos (sales and repurchase agreements)

– A “haircut” determines the fraction of the market value that can be borrowed

• Securities borrowing– A “haircut” again determines what collateral must be

posted• As well, they engage in derivatives transactions

– Except for large highly-rated banks and securities dealers, “initial margin” must be posted


I. Transmission

5 per cent haircut

95 per cent loan

$100 Million Bond


I. Transmission

• Margin and Liquidity Spirals– When market liquidity becomes lower, it is typically

associated with higher market volatility– But higher market volatility means that collateral

coverage for a given “haircut” or “initial margin” is less: haircuts and margins tend to rise in the market

– One tends to get the type of liquidity and margin spiral shown in the following diagram• Spiral can work in the opposite direction in boom periods


Liquidity/Margin Spiral

lower market liquidityfunding problems

less market making

higher margins

losses on existing positions

Liquidity Spiral

Adapted from Brunnermeier & Pederson (2009) and presentations by Mark Carney and David Longworth


I. Transmission

• Fire Sales (Cash-in-the-market pricing)(AG 4,5)– First, assume a model with markets only, no banks– Limited market participation: not everyone

participates in every market (fixed set-up cost)– Market liquidity depends on amount of cash held by

market participants– If there is a lack of cash in the market, small shocks

have large effects on prices• Then prices are not determined by expected present values,

but by ratio of available liquidity to amount of asset supplied


I. Transmission

• Fire Sales (Cash-in-the-market pricing)(AG 4,5)– Amount of cash in market depends on

participants’ liquidity preference, which will determine the average level of the short-term asset held

– Changes in liquidity demand relative to liquidity supply determines price volatility


I. Transmission

• Fire Sales (Cash-in-the-market pricing)(AG 4,5)– Now add banks to the model• Small events (e.g., small liquidity shocks) can have a

large impact on the financial system because of how banks and markets interact: can lead to systemic crises• If banks have to provide liquidity to customers, they

may have to sell much-less-liquid assets (if they are running out of liquid ones)


I. Transmission

• Fire Sales (Cash-in-the-market pricing)(AG 4,5)– With banks added to the model:• Prices in those markets may be determined by cash in

the market• The resulting “fire sale prices” may be quite low• Banks have to mark assets held in their trading book to

market. At the end of the quarter, these losses will show up in the calculation of profits/losses and thus affect the bank’s capital– The market anticipates this effects even before

quarterly statements are released.


I. Transmission

• Interconnectedness and contagion(KA8, AG10)– Interconnectedness: banks hold many liabilities of

other banks (short-term deposits—including those for settling payments, shares, repos, derivative instruments)• Therefore the failure or weakness of one bank could

translate into the failure or weakness of other banks• As well, the failure of one bank may lead to loss-sharing

arrangements being invoked in payments systems and central counterparties (for repos or OTC derivatives)– By their current design, such losses should be limited


I. Transmission

• Interconnectedness and contagion(KA8, AG10)– Banking contagion (usually used only for across

regions or across countries) can arise from a number of factors:• Interconnectedness as described above• Concern about common exposures, with fire sales

potentially driving down prices


I. Transmission

Contagion (usually used only for across regions or across countries) can arise from a number of factors:• Contagion of bubbles: “when money flows from one

country to another and adjustments automatically occur both in the countries that receive these funds and in the countries that are the sources of them.” (KA, p.143)– Example from KA, pp. 142-3(: From real estate and stock

market bubble in Japan (late 1980s) to real estate and stock market bubbles in Nordic countries (late 1980s) and to markets in south-east Asia (mid 1990s) and to tech stocks in the U.S. (late 1990s)


I. Transmission

• Interconnectedness and contagion Topics:– What is the analogue in the most recent crisis and how

did it compare with previous crises? – How were various emerging market economies

affected in the current crisis when international banks cut back in foreign lending, particularly in trade finance (why, and what were the effects?)

– Why was there more banking contagion from the U.S. to continental Europe and the U.K. than to other regions? Why was there financial contagion at all to countries such as Japan and Canada?


I. Transmission

• Interconnectedness and contagion(KA8, AG10)– AG have a model of U.S. regional contagion

• “even though the initial shock occurs only in one region, which can be an arbitrarily small part of the economy, it can nevertheless cause banks in all regions to go bankrupt.”

• Results depend on the nature of the network of interbank deposits across institutions

– AG cite a number of references to studies of the actual nature of interbank relationships in certain countries.


I. Transmission

• Contagion across countries (stock markets)– K. Forbes, Jackson Hole, 2012– Interdependence (high correlations across equity

markets) has increased over time, especially in euro area

– Contagion, spillovers from extreme negative events, is more common when country has a more levered banking system, greater trade exposure, weaker macro fundamentals, and large international liabilities


I. Transmission

• Decline in wealth in private sector: effects on income and employment– Lower wealth arises from fire sales, bursting of bubbles,

lower valuation of financial sector firms– Wealth effects on consumption (standard consumption

function)– Through financial accelerator, lower collateral means can

borrow less, so lower consumption and housing expenditure (and investment by businesses)

– Through bank capital channel, less lending by banks, which means less consumption, housing, and investment expenditure (but large corps. can go to bond market)


I. Transmission

• Decline in wealth in private sector: effects on income and employment– Spreads increase between interest rates on

loans/market debt and government yields (even separately from bank capital channel), lowering housing and investment spending

– In New Keynesian models, lower aggregate demand leads to lower employment

– Spillovers across borders from lower import demand in countries suffering declines in wealth and income


I. Transmission

• Zero bound on nominal interest rates takes away conventional monetary policy channel– Normally, the response of monetary policy authorities to the

decline in wealth, income, and employment would be to lower the policy interest rate because of the downward pressure on inflation

– When the policy interest rate gets to zero (or near zero), that option is no longer available

– Central bank must turn to unconventional policy instruments (discussed in the next section)

– ZLB in history: BoJ; recent crisis: Fed, BoE, BoJ, BoC, ECB, Riksbank


I. Transmission

• Effect on sovereign debt crises and vice versa– Government bailouts or payouts to insured

depositors increase sovereign debt– Fall in GDP leads to decline in government

revenue and increase in sovereign debt– If sovereign debt was high before banking crisis, a

sovereign debt crisis may occur– Banks hold lots of sovereign debt, so a sovereign

debt crisis can lead to a banking crisis


I. Transmission• What happens to GDP growth after the crisis?

– Cecchetti et al. (2010)• High debt to GDP ratio is bad for growth—beyond 85% for

government debt

– Jorda et al. (2011)• More credit-intensive booms tend to be followed by deeper

recessions and deeper recoveries

– Reinharts (2010)• For major shocks, GDP growth lower and unemployment higher in 10

years after than in 10 years before

– Reinharts and Rogoff (2012)• When government debt/GDP over 90 per cent, low growth is likely—

typically for more than a decade

• Topics: role of capital? Role of cross-country effects?


II. Policy Response During the Crisis

(Prediction)

Transmission

Policy Response

Prevention

Causes


II. Policy Response During a Crisis

• Guarantees and closures• Domestic lender of last resort: liquidity policy• Expansionary Monetary Policy• Expansionary Fiscal Policy• International lender of last resort: IMF, EU, etc



Crisis Policy Advanced Economies Emerging EconomiesExpansionary fiscal policy 90 50

Expansionary monetary pol. 80 65

Recapitalization 95 80

Nationalizations 65 60

Liquidity support 95 85

Guarantees on bank liabilities 90 35

Deposit freeze 0 15

Differences in the Mix of Banking Crisis Policies (% Use to nearest 5%) (Source: Laeven and Valencia, IMF, 2012)



• Guarantees and closures– Deposit insurance introduced or limits increased– Bank bond debt guaranteed (e.g., Ireland)– Bank holiday (cannot withdraw funds: deposit freeze)– Markets closed (especially stock markets)– Short-selling of bank stocks banned temporarily– Resolution of bank (range of possibilities)

• Government injects capital (recapitalization) • Government nationalizes (with or without paying)• Bank taken over by deposit insurance fund to be wound down (only

insured depositors paid off in first instance, then other creditors)

– Issues: effectiveness, moral hazard, benefit/cost



• Domestic lender of last resort: liquidity policy– Central bank policy existing before recent crisis

• “Discount window” lending against good collateral (bonds, paper) with haircut (reduction from market value) and small penalty rate

• Repo (purchase and resale agreement) of good bonds and paper with haircut

• These provided additional liquidity for banks needing it

– Broad (ECB) vs. narrow (BoC, Fed) in normal times• Potential topic: Does a broad list of collateral in normal times lead to

moral hazard and to major problems in crisis times?

– Expansion in recent crisis was initially in: frequency of repo operations, size of operations, length of period, and range of eligible collateral



• Domestic lender of last resort: liquidity policy– Because of “stigma” attached to discount window in U.S., a

Term Auction Facility was introduced that had a wider range of collateral than repo operations. In Canada, the non-mortgage loan portfolio of banks was eligible for a TAF-like facility• BoE has changed auctions of liquidity so that they always happen

—this is to avoid stigma in a crisis

– Central banks also introduced liquidity facilities to deal with problems in specific financial markets (as opposed to financial institutions). The Fed did this in particular for the commercial paper market.



• Domestic lender of last resort: liquidity policy– Making foreign currency liquidity available: central

bank FX swap lines– Potential topic: Why did the range of special

liquidity facilities vary across countries? Why were special liquidity facilities not needed in previous crises?



• Monetary Policy– Conventional monetary policy, reducing policy

interest rate (incentive to get to ZLB quickly in some instances)

– Unconventional monetary policy• Conditional or unconditional commitment regarding

future policy interest rate


II. Policy Response During a Crisis– Unconventional monetary policy

• Expansion of excess bank reserves (settlement balances) purchasing government debt or repos – Untargeted version is pure quantitative easing (Japan early last

decade) (many modern theories would discount effectiveness)» “helicopter drop” variant: finances government payments to

individuals (fiscal policy)– If particular maturities of government debt is purchased, it is also a

form of debt-management policy. (So is selling long-term debt to buy short-term debt)

– If private sector debt (e.g. private MBS) is purchased, it is also a form of fiscal policy (credit policy)

– In these latter three cases, there is a question of governance/coordination

– Effects: expectations of future interest rates; portfolio-balance.



• Monetary Policy– Potential topic: What was the effectiveness of

unconventional monetary policy in the recent crisis (e.g., QE2 vs. QE1 in the U.S.)? What was the announcement effect from recent Fed announcements?



• International Lender of Last Resort– IMF, or EU, or bilateral sovereign loans– Typically in an exchange crisis (fixed exchange rates)– But also could be in cases where there is extreme

pressure on exchange rates; or significant associated fiscal problems

– The history of IMF loans in the last 30 years has been about the appropriate “conditionality” of loans

– Current episode: EU and IMF loans: Greece, Ireland, Portugal and …. (will there be more?)


III. Prevention

(Prediction)

Transmission

Policy Response

Prevention

Causes


III. Prevention

• Macroprudential policy• Contingent capital and bail-in debt• Monetary policy


III. Prevention

• Macroprudential policy– Focuses on the safety and soundness of the financial

system as a whole, as opposed to the safety and soundness of individual financial institutions

– Macroprudential tools: deal with market failures associated with procyclicality of aspects of the financial system, as well as the interconnections and similar exposures across financial institutions (cross-sectional aspect) (recall market failures from last week)• Possible topic: How do financial cycles compare with “real cycles

and inflation cycles” across countries? How should one measure a “financial cycle”? Implications?

Powers

Activities: Data Collection, Surveillance, Analysis, Risk Assessment, Stress Testing

Policy Instruments: Macroprudential instruments, advice on policies, warnings

Goals: dampening procyclicality and reducing potential effects of contagion

Objective: avoiding significant financial instability

Framework

47

III. PreventionProximate Object of Concern:

Macroprudential Instrument:

Excessive Credit Creation

InsufficientLiquidity

Continuation of a Bank

Capital Requirements

√ (total or sectoral)

√ (maturity mismatch)

√ (Contingent capital)

Pigouvian Taxes √ √ (on non-core deposits)

√

Constraints on quantities, or on credit conditions

√ (Cred; RR on assets)√ (Haircuts, LTV)

√ (BCBS Liquidity)√ (Haircuts, LTV)

48


III. Prevention

• Macroprudential policy– Capital requirements, leverage requirements, and

liquidity requirements are being dealt with in Basel III • Higher capital requirements, capital buffer built up,

countercyclical requirements (typically linked to credit) (Note that effects overall depends on extent to which Modigliani-Miller theorem regarding capital cost and capital structure is violated.)• Study of how systemically important institutions should

have higher capital requirements


III. Prevention

• Macroprudential policy– Capital requirements, leverage requirements, and

liquidity requirements are being dealt with in Basel III • Two types of liquidity requirements

– Liquidity coverage ratio (liquid assets vs. liabilities)» “Sufficiently high quality liquid assets to survive a

significant stress scenario lasting one month” (BCBS)– Net stable funding ratio (liability structure)

» “Incentive for banks to fund their activities with more stable sources of funding” (BCBS)


III. Prevention

• Macroprudential policy– A Study Group that I chaired for the Committee on

the Global Financial System proposed regulating margin requirements on derivatives and haircuts on repo transactions on a “through the cycle basis” to reduce the procyclicality of the margin cycle


III. Prevention

• Macroprudential policy– Several Asian countries regulate loan-to-value ratios for

mortgages (particularly on residential properties) in an active manner to reduce the cycle in property prices• There are several aspects of requirements for mortgage

insurance that could be examined for more active regulation in Canada (LTV ratios, debt-service-to-income ratio, home equity loan ratio, amortization period); constant level or varying countercyclically. Potential topic: How should macroprudential policy connected to mortgages and housing prices be carried out, i.e., what should be the proximate goal and the tools? How would this have worked in previous housing bubbles?


III. Prevention

• Macroprudential policy– Some consider “through the cycle provisioning” for

loan losses (as was used in Spain) to be a macroprudential policy instrument

– There are other possible macroprudential policy instruments, such as reserve requirements on assets and levies (taxes) on non-core deposits• There is some evidence that in modern financial systems

rapid credit growth has as its counterpart the growth in non-core short-term deposit liabilities (wholesale deposits, commercial paper, repos, etc.)


III. Prevention

• Macroprudential policy– Potential topics: Would macroprudential policy “x”

have prevented the recent and other financial crises? When household debt is high relative to personal disposable income (e.g., Canada, New Zealand, Sweden), should the authorities respond in order to prevent future crises and, if so, how?


III. Prevention

• Contingent capital and bail-in debt– To deal with moral hazard of “too big or complex

to fail” as well as the practical issue of having time to wind down a large institution or to change its owners

– “Contingent capital is a subordinated security, such as a preferred share or subordinated debenture, that converts to common equity under certain conditions.” (BoC FSR, Dec 2010, p.52)


III. Prevention

• Contingent capital (CC) and bail-in debt– Gone-concern CC converts when supervisor judges that

bank is no longer viable– Going-concern CC converts well before, for modest

erosions of capital– Bail-in debt applies to senior debt as well– Conceptually, “the sum of common equity plus

contingent capital and bail-in senior debt could be subject to an overall minimum requirement, chosen to provide for the restoration of prudential capital requirements” (BoC FSR, Dec 2010, p.54)


III. Prevention

• Monetary policy– Giving monetary policy a full-fledged financial

stability objective in addition to, but secondary to, its price stability objective• Inflation targeters would implement this by sometimes

returning inflation to target over a longer time period• Use of reserve requirements (on short-term risky

source of funding or on risky assets) to mimic Pigouvian taxes. New reference: Anil Kashyap and Jeremy Stein (2012), “The Optimal Conduct of Monetary Policy with Interest on Reserves,” AEJMacro, vol.4(1), pp 266-282.


III. Prevention

• Monetary policy– Or, having monetary policy play a supporting role

where not in conflict with its price stability objective:• Choice of target inflation rate (e.g., inclusion of house

prices)• Price level target versus inflation target (Carney, 2009)• Making very prominent the uncertainty about the

future interest rate path (note related criticism of Fed in past decade)


III. Prevention

• Monetary policy– Or, having monetary policy play a supporting role

where not in conflict with its price stability objective:• Choice of target inflation rate (e.g., inclusion of house

prices)• Price level target versus inflation target (Carney, 2009)• Making very prominent the uncertainty about the

future interest rate path (note related criticism of Fed in past decade)


III. Prevention

• Monetary policy– Potential topic: What should the role of monetary

policy (or the relative roles of monetary policy and macroprudential policy) be in maintaining financial stability? Could monetary policy have prevented the recent crisis in some countries? At what cost?


This Week• Prepare two-paragraph topic description for next class

(typed paper copy necessary)– What is hypothesis or question to be answered?• Which crises or countries are being compared?

– What is main method of analysis?– Note at top of page: cause, prediction, transmission, policy

response, prevention• Reference list on course web site should be helpful; also

Lectures 1 and 2• I have office hours this afternoon and tomorrow

morning; or can e-mail me

Econ 492: Comparative Financial Crises Lecture 2 19 September 2012 David Longworth This material is copyrighted and is for the sole use of students registered.

Documents

booms economics

transmission bank

comparative financial

commodities economics

currency crises

banking crises

friends bank

transmission illiquidity