TheCreditCrunchof2007-2008: ADiscussionoftheBackground, … · 2019-10-15 · FEDERAL RESERVE BANK OF ST. LOUIS REVIEW SEPTEMBER/OCTOBER 2008 531 TheCreditCrunchof2007-2008: ADiscussionoftheBackground,

FEDERAL RESERVE BANK OF ST. LOUIS REVIEW SEPTEMBER/OCTOBER 2008 531

The Credit Crunch of 2007-2008:A Discussion of the Background,

Market Reactions, and Policy ResponsesPaul Mizen

This paper discusses the events surrounding the 2007-08 credit crunch. It highlights the periodof exceptional macrostability, the global savings glut, and financial innovation in mortgage-backedsecurities as the precursors to the crisis. The credit crunch itself occurred when house prices felland subprime mortgage defaults increased. These events caused investors to reappraise the risksof high-yielding securities, bank failures, and sharp increases in the spreads on funds in interbankmarkets. The paper evaluates the actions of the authorities that provided liquidity to the marketsand failing banks and indicates areas where improvements could be made. Similarly, it examinesthe regulation and supervision during this time and argues the need for changes to avoid futurecrises. (JEL E44, G21, G24, G28)

Federal Reserve Bank of St. Louis Review, September/October 2008, 90(5), pp. 531-67.

that the phrase “credit crunch” has been usedin the past to explain curtailment of the creditsupply in response to both (i) a decline in thevalue of bank capital and (ii) conditions imposedby regulators, bank supervisors, or banks them-selves that require banks to hold more capitalthan they previously would have held.

Amilder version of a full-blown credit crunchis sometimes referred to as a “credit squeeze,”and arguably this is what we observed in 2007and early 2008; the term credit crunchwas alreadyin use well before any serious decline in creditsupply was recorded, however. At that time theeffects were restricted to shortage of liquidity inmoney markets and effective closure of certaincapital markets that affected credit availabilitybetween banks. There was even speculation

T he concept of a “credit crunch” has along history reaching as far back as theGreat Depression of the 1930s.1 BenBernanke and Cara Lown’s (1991) classic

article on the credit crunch in the BrookingsPapers documents the decline in the supply ofcredit for the 1990-91 recession, controlling forthe stage of the business cycle, but also considersfive previous recessions going back to the 1960s.The combined effect of the shortage of financialcapital and declining quality of borrowers’ finan-cial health caused banks to cut the loan supplyin the 1990s. Clair and Tucker (1993) document

1 The term is now officially part of the language as one of severalnew words added to the Concise Oxford English Dictionaryin June 2008; also included for the first time is the term“sub-prime.”

Paul Mizen is a professor of monetary economics and director of the Centre for Finance and Credit Markets at the University of Nottinghamand a visiting scholar in the Research Division of the Federal Reserve Bank of St. Louis. This article was originally presented as an invitedlecture to the Groupement de Recherche EuropéenMonnaie Banque Finance XXVth Symposium on Banking andMonetary Economics hostedby the Université du Luxembourg, June 18-20, 2008. The author thanks the organizers—particularly, Eric Girardin, Jen-Bernard Chatelain,and Andrew Mullineux—and Dick Anderson, Mike Artis, Alec Chrystal, Bill Emmons, Bill Gavin, Charles Goodhart, Clemens Kool, DanThornton, David Wheelock, and Geoffrey Wood for helpful comments. The author thanks Faith Weller for excellent research assistance.

© 2008, The Federal Reserve Bank of St. Louis. The views expressed in this article are those of the author(s) and do not necessarily reflect theviews of the Federal Reserve System, the Board of Governors, or the regional Federal Reserve Banks. Articles may be reprinted, reproduced,published, distributed, displayed, and transmitted in their entirety if copyright notice, author name(s), and full citation are included. Abstracts,synopses, and other derivative works may be made only with prior written permission of the Federal Reserve Bank of St. Louis.

whether these conditions would spill over intothe real sector, but there is little doubt now thatthere will be a decline in the terms and availabil-ity of credit for consumers and entrepreneurs.Disorder in financial markets occurred as bankssought to determine the true value of assets thatwere no longer being traded in sufficient volumesto establish a true price; and uncertainty prevailedamong institutions aware of the need for liquiditybut unwilling to offer it except under terms wellabove the risk-free rate. These conditions havenow given way to the start of a credit crunch, andthe restrictions on the credit supply will havenegative real effects.

Well-informed observers, such as MartinWolf,associate editor and chief economics commenta-tor of the Financial Times, are convinced that thecredit crunch of 2007-08 will have a significancesimilar to that of earlier turning points in theworld economy, such as the emerging marketscrises in 1997-98 and the dotcom boom-and-bustin 2000 (Wolf, 2007). Like previous crises, thecredit crunch has global implications becauseinternational investors are involved. The asset-backed securities composed of risky mortgageswere packaged and sold to banks, investors, andpension funds worldwide—as were equities inemerging markets and dotcom companies beforethem.

The 2007-08 credit crunch has been far morecomplex than earlier crunches because financialinnovation has allowed new ways of packagingand reselling assets. It is intertwined with thegrowth of the subprime mortgage market in theUnited States—which offered nonstandard mort-gages to individuals with nonstandard incomeor credit profiles—but it is really a crisis thatoccurred because of the mispricing of the risk ofthese products. New assets were developed basedon subprime and other mortgages, which werethen sold to investors in the form of repackageddebt securities of increasing sophistication. Thesereceived high ratings and were considered safe;they also provided good returns compared withmore conventional asset classes. However, theywere not as safe as the ratings suggested, becausetheir value was closely tied tomovements in houseprices. While house prices were rising, these

assets offered relatively high returns comparedwith other assets with similar risk ratings; but,when house prices began to fall, foreclosures onmortgages increased. To make matters worseinvestors had concentrated risks by leveragingtheir holdings of mortgages in securitized assets,so their losses were multiplied. Investors realizedthat they had not fully understood the scale ofthe likely losses on these assets, which sent shockwaves through financial markets, and financialinstitutions struggled to determine the degree oftheir exposure to potential losses. Banks failedand the financial system was strained for anextended period. The banking system as a wholewas strong enough to take these entities onto itsbalance sheet in 2007-08, but the effect on thedemand for liquidity had a serious impact on theoperation of the money markets.

The episode tested authorities such as centralbanks, which were responsible for providing liq-uidity to the markets, and regulators and super-visors of the financial systems, who monitor theactivities of financial institutions. Only now arelessons being learned that will alter future oper-ations of the financial system to eliminate weak-nesses in the process of regulation and supervisionof financial institutions and the response of centralbanks to crisis conditions. These lessons includethe need to create incentives that ensure thecharacteristics of assets “originated and distrib-uted” are fully understood and communicated toend-investors. These changes will involve mini-mum information standards and improvements toboth the modeling of risks and the ratings process.Central banks will review their treatment of liquid-ity crises by evaluating the effectiveness of theirprocedures to inject liquidity into the markets attimes of crisis and their response to funding crisesin individual banks. Regulators will need to con-sider the capital requirements for banks and off-balance sheet entities that are sponsored or ownedby banks, evaluate the scope of regulation neces-sary for ratings agencies, and review the useful-ness of stress testing and “fair value” accountingmethods.

This article consists of two parts: an outlineof events and an evaluation. The first part dis-cusses the background to the events of the past

Mizen

532 SEPTEMBER/OCTOBER 2008 FEDERAL RESERVE BANK OF ST. LOUIS REVIEW

year to discover how and why credit marketshave expanded in recent years due to an environ-ment of remarkably stable macroeconomic condi-tions, the global savings glut, and the developmentof new financial products. These conditions wereconducive to the expansion of credit without dueregard to the risks. It then describes the marketresponses to the deteriorating conditions and theresponse of the authorities to the crisis. The sec-ond part discusses how the structure and incen-tives of the new financial assets created conditionslikely to trigger a crisis. It also evaluates theactions of the authorities and the regulators withsome recommendations for reform.

EVENTSBackground: The Origins of the Crisis

The beginnings of what is now referred to asthe 2007-08 credit crunch appeared in early 2007to be localized problems among lower-qualityU.S. mortgage lenders. An increase in subprimemortgage defaults in February 2007 had causedsome excitement in the markets, but this hadsettled by March. However, in April New CenturyFinancial, a subprime specialist, had filed forChapter 11 bankruptcy and laid off half itsemployees; and in early May 2007, the Swiss-owned investment bank UBS had closed theDillon Reed hedge fund after incurring $125 mil-lion in subprime mortgage–related losses.2 Thisalso might have seemed an isolated incident, butthat month Moody’s announced it was reviewingthe ratings of 62 asset groups (known as tranches)based on 21 U.S. subprime mortgage securitiza-tions. This pattern of downgrades and losseswas to repeat itself many times over the next fewmonths. In June 2007 Bear Stearns supported twofailing hedge funds, and in June and July 2007three ratings agencies—Fitch Ratings, Standard& Poor’s, andMoody’s—all downgraded subprime-related mortgage products from their “safe” AAA

status. Shortly thereafter Countrywide, a U.S.mortgage bank, experienced large losses, and inAugust two European banks, IKB (German) andBNP Paribas (French), closed hedge funds introubled circumstances. These events were todevelop into the full-scale credit crunch of 2007-08. Before discussing the details, we need to askwhy the credit crunch happened and why now?Two important developments in the late 1990sand early twenty-first century provided a sup-portive environment for credit expansion. First,extraordinarily tranquil macroeconomic condi-tions (known as the “Great Moderation”) coupledwith a flow of global savings from emerging andoil-exporting countries resulted in lower long-term interest rates and reduced macroeconomicvolatility. Second, an expansion of securitizationin subprime mortgage– backed assets producedsophisticated financial assets with relatively highyields and good credit ratings.The Great Moderation and the Global

Savings Glut. The “Great Moderation” in theUnited States (and the “Great Stability” in theUnited Kingdom) saw a remarkable period of lowinflation and low nominal short-term interestrates and steady growth. Many economists con-sider this the reason for credit expansion. Forexample, Dell’Ariccia, Igan, and Leavan (2008)suggest that lending was excessive—what theycall “credit booms”—in the past five years. Beoriand Guiso (2008) argue that the seeds of thecredit booms were sown by Alan Greenspan whenhe cut short-term interest rates in response to the9/11 attacks and the dotcom bubble, which is aplausible hypothesis, but this is unlikely to be themain reason for the expansion of credit. Short-term rates elsewhere, notably the euro area andthe United Kingdom, were not as low as theywere in the United States, but credit grew there,too. When U.S. short-term interest rates steadilyrose from 2004 to 2006, credit continued to grow.It is certainly true that the low real short-terminterest rates, rising house prices, and stableeconomic conditions of the Great Moderationcreated strong incentives for credit growth onthe demand and supply side. However, anotherimportant driving force of the growth in lendingwas found in the global savings glut flowing fromChina, Japan, Germany, and the oil exporters

Mizen


2 As we will explain in more detail, defaults on subprime mortgagesincreased, causing losses; but, because investors “scaled up” therisks by leveraging their positions with borrowed funds, whichwere themselves funded with short-term loans, these small losseswere magnified into larger ones.

that kept long-term interest rates down, as then-Governor Bernanke noted in 2005 in a speechentitled, “The Global Saving Glut and the U.S.Current Account Deficit.”

After the Asian crisis of 1997, many affectedcountries made determined efforts to accumu-late official reserves denominated in currenciesunlikely to be affected by speculative behavior,which could be used to defend the currencyregime should events repeat themselves. (Withlarger reserves, of course, those events wereunlikely to be repeated.) Strong demand for U.S.Treasuries and bonds raised their prices andlowered the long-term interest rate. Large savingsflows from emerging markets funded the growingdeficits in the industrialized countries for a time,and significant imbalances emerged betweencountries with large current account surplusesand deficits. These could not be sustained indef-initely; but, while they lasted and long-term inter-est rates were low, they encouraged the growthof credit.

Figures 1 and 2 show that saving ratiosdeclined and borrowing relative to incomeincreased for industrialized countries from 1993to 2006. The U.S. saving ratio fell from 6 percentof disposable income to below 1 percent in littleover a decade, and at the same time the total debt–to–disposable income ratio rose from 75 percentto 120 percent, according to figures produced bythe Organisation for Economic Co-operation andDevelopment (OECD). The United Kingdom andCanada show similar patterns in saving and debt-to-income ratios, as does the euro area—but thesaving ratio is higher and the debt-to-incomeratio is lower than in other countries.

Similar experiences were observed in othercountries. Revolving debt in the form of creditcard borrowing increased significantly, and asprices in housing markets across the globeincreased faster than income, lenders offeredmortgages at ever higher multiples (in relationto income), raising the level of secured debt toincome. Credit and housing bubbles reinforced

Mizen


1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

20

18

16

14

12

10

8

6

4

2

0

CanadaUnited KingdomUnited StatesEuro Area

Percent of Disposable Income

Figure 1

Saving Ratios

SOURCE: OECD Economic Outlook and ECB Monthly Bulletin.

each other. Borrowers continued to seek funds togain a foothold on the housing ladder, reassuredby the fact that the values of the properties theywere buying were rising andwere expected to con-tinue to rise. Lenders assumed that house priceswould continue to rise in the face of strongdemand. In some cases, lenders offered in excessof 100 percent of the value of the property. Con-ditions in housing markets were favorable toincreased lending with what appeared to be lim-ited risk; lenders were prepared to extend thescope of lending to include lower-quality mort-gages, known as subprime mortgages.Growth in the Subprime Mortgage Market.

In the United States mortgages comprise fourcategories, defined as follows:

(i) prime conforming mortgages are made togood-quality borrowers and meet require-ments that enable originators to sell themto government-sponsored enterprises (GSEs,such as Fannie Mae and Freddie Mac);

(ii) jumbo mortgages exceed the limits set byFannie Mae and Freddie Mac (the 2008limit set by Congress is a maximum of$729,750 in the continental United States,but a loan cannot be more than 125 percentof the county average house value; the limitis higher in Alaska, Hawaii, and the U.S.Virgin Islands), but are otherwise standard;

(iii) Alt-A mortgages do not conform to theFannie Mae and Freddie Mac definitions,perhaps because a mortgagee has a higherloan-to-income ratio, higher loan-to-valueratio, or some other characteristic thatincreases the risk of default; and

(iv) subprime mortgages lie below Alt-A mort-gages and typically, but not always, repre-sent mortgages to individuals with poorcredit histories.

Subprime mortgages are nevertheless difficultto define (see Sengupta and Emmons, 2007). Oneapproach is to consider the originators of mort-

Mizen


1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

180

160

140

120

100

80

60

CanadaUnited KingdomUnited StatesEuro Area

Liabilities in Percent of Disposable Income

Figure 2

Debt to Income Ratios

SOURCE: OECD Economic Outlook and ECB/Haver Analytics.

gages: The U.S. Department of Housing andUrban Development (HUD) uses Home MortgageDisclosure Act (HMDA) data to identify subprimespecialists with fewer originations, a higher pro-portion of loans that are refinanced, and, becausesubprime mortgages are nonconforming, thosethat sell a smaller share of their mortgages to theGSEs. A second approach is to identify the mort-gages by borrower characteristics: The Board ofGovernors of the Federal Reserve System, theOffice of the Comptroller of the Currency, theFederal Deposit Insurance Corporation, and theOffice of Thrift Supervision list a previous recordof delinquency, foreclosure, or bankruptcy; acredit score of 580 or below on the Fair, Isaacand Company (FICO) scale; and a debt service-to-income ratio of 50 percent or greater as sub-prime borrowers. Subprime products also existin other countries where they may be marketed asinterest-only, 100 percent loan-to-value, or self-certification mortgages, but they are not as preva-lent as in the United States.

The main differences between a prime mort-gage and a subprimemortgage from the borrower’sperspective are higher up-front fees (such as appli-cation and appraisal fees), higher insurance costs,fines for late payment or delinquency, and higherinterest rates. Therefore, the penalty for borrowingin the subprime market, when the prime marketis inaccessible, is a higher cost in the form of loanarrangement fees and charges for failing to meetpayment terms. The main difference from thelender’s perspective is the higher probability oftermination through prepayment (often due torefinancing) or default. The lender sets an interestrate dependent on a loan grade assigned in lightof the borrower’s previous payment history, bank-ruptcies, debt-to-income ratio, and a limited loan-to-value ratio, although this can be breached bypiggyback lending. The lender offers a subprimeborrower a mortgage with an interest premiumover prime mortgage rates to cover the higher riskof default given these characteristics. Many otherterms are attached to subprime mortgages, whichsometimes benefit the borrower by grantingallowances (e.g., to vary the payments throughtime), but the terms often also protect the lender(e.g., prepayment conditions that make it easier

for the lender to resell the mortgage loan as asecuritized product).

The market for subprime mortgages grewvery fast. Jaffee (2008) documents two periods ofexceptional subprime mortgage growth. The firstexpansion occurred during the late 1990s, whenthe volume of subprime lending rose to $150 bil-lion, totalling some 13 percent of total annualmortgage originations. This expansion came toa halt with the dotcom crisis of 2001. A secondexpansion phase was from 2002 until 2006(Figure 3), when the subprime component ofmortgage originations rose from $160 billion in2001 to $600 billion by 2006 (see Calomiris, 2008),representing more than 20 percent of total annualmortgage originations. Chomsisengphet andPennington-Cross (2006) argue that these expan-sions occurred because changes in the law allowedmortgage lending at high interest rates and fees,and tax advantages were available for securedborrowing versus unsecured borrowing.3 Anotherstrong influence was the desire of mortgage origi-nators to maintain the volume of new mortgagesfor securitization by expanding lending activityinto previously untapped markets. Subprimeloans were heavily concentrated in urban areasof certain U.S. cities —Detroit, Miami, Riverside,Orlando, Las Vegas, and Phoenix—where home-ownership had not previously been common—as well as economically depressed areas of Ohio,Michigan, and Indiana, where prime borrowersthat faced financial difficulties switched fromprime to subprime mortgages.Securitization and “Originate and Distribute”

Banking. Securitization was popularized in theUnited States when the Government NationalMortgage Association (Ginnie Mae) securitizedmortgages composed of Federal HousingAdministration and Veterans Administration(FHA/VA) mortgages backed by the “full faith

3 Chomsisengphet and Pennington-Cross (2006) indicate that theDepository Institutions Deregulation and Monetary Control Act(1980) allowed borrowers to obtain loans from states other thanthe state in which they lived, effectively rendering interest ratecaps at the state level ineffective. The Alternative MortgageTransaction Parity Act (1982) allowed variable-rate mortgages, andthe Tax Reform Act (1986) ended tax deduction for interest onforms of borrowing other than mortgages. These changes occurredwell before the growth in subprime mortgage originations, but theyput in place conditions that would allow for that growth.

Mizen


and credit” of the U.S. government for resale ina secondary market in 1968.4 In 1981, the FederalNational Mortgage Association (Fannie Mae)began issuing mortgage-backed securities (MBSs),and soon after new “private-label” securitizedproducts emerged for prime loans without thebacking of the government.5 The European assetsecuritization market emerged later, in the 1990s,and picked up considerably in 2004. The origina-tions occurred mainly in the Netherlands, Spain,and Italy (much less so in Germany, France andPortugal), but they were widely sold: More thanhalf were sold outside the euro area, with one-third sold to U.K. institutions in 2005-06.

Securitization was undertaken by commercialand investment banks through special purposevehicles (SPVs), which are financial entities cre-ated for a specific purpose—usually to engage ininvestment activities using assets conferred onthem by banks, but at arm’s length and, impor-tantly, not under the direct control of the banks.The advantage of their off-balance sheet statusallows them to make use of assets for investmentpurposes without incurring risks of bankruptcyto the parent organization (see Gorton andSouleles, 2005). SPVs were established to createnew asset-backed securities from complex mix-tures of residential MBSs, credit card, and otherdebt receivables that they sold to investors else-where. By separating asset-backed securities intotranches (senior, mezzanine, and equity levels),the SPVs offering asset-backed securities couldsell the products with different risk ratings foreach level. In the event of default by a proportionof the borrowers, the equity tranche would bethe first to incur losses, followed by mezzanine

Mizen


4 Ginnie Mae is a government-owned corporation within theDepartment of Housing and Urban Development (HUD) that wasoriginally established in 1934 to offer “affordable” housing loans.In 1968 it was allowed by Congress to issue MBSs to finance itshome loans.

5 Private-label MBSs dated back to the 1980s, but the process ofrepackaging and selling on auto loan receivables and credit cardreceivables goes back much farther—to the 1970s.

Billions of U.S. $700

600

500

400

300

200

100

02001 2002 2003 2004 2005

Figure 3

Subprime Mortgage Originations, Annual Volume

SOURCE: Data are from Inside Mortgage Finance, as published in the 2006 Mortgage Market Statistical Annual, Vol. 1.

and finally by senior tranches. Senior trancheswere rated AAA—equivalent to government debt.In addition, they were protected by third-partyinsurance from monoline insurers that undertookto protect holders from losses, which improvedtheir ratings.

A market for collateralized debt obligations(CDOs) composed of asset-backed securitiesemerged; these instruments also had claims ofdifferent seniority offering varying payments.Banks held asset-backed securities in “ware-houses” before reconstituting them as CDOs, soalthough they were intermediating credit to end-investors, they held some risky assets on theirbalance sheets in the interim. Some tranches ofCDOs were then pooled and resold as CDOs ofCDOs (the so-called CDOs-squared); CDOs-squaredwere even repackaged into CDOs-cubed. Thesewere effectively funds-of-funds based on the orig-inal mortgage loans, pooled into asset-backedsecurities, the lower tranches of which were thenpooled again into CDOs, and so forth. As the OECDexplains, the process involved several stepswhereby “[the] underlying credit risk is firstunbundled and then repackaged, tiered, securi-tised, and distributed to end investors. Variousentities participate in this process at variousstages in the chain running from origination tofinal distribution. They include primary lenders,mortgage brokers, bond insurers, and credit rat-ing agencies” (OECD, 2008).

Some purchasers were structured investmentvehicles (SIVs)—off balance-sheet entities createdby banks to hold these assets that could evadecapital control requirements that applied to banksunder Basel I capital adequacy rules. Others werebought by conduits—organizations similar to SIVsbut backed by banks and owned by them. Thescale of these purchases was large; de la Dehasa(2008) suggests that the volumes for conduitswas around $600 billion for U.S. banks and $500billion for European banks. The global market inasset-backed securities was estimated by the Bankof England at $10.7 trillion at the end of 2006.Ironically, many of the purchasers were off-balance-sheet institutions owned by the verybanks that had originally sold the securitizedproducts. This was not recognized at the time but

would later come home to roost as losses on theseassets required the banks to bring off-balance-sheet vehicles back onto the balance sheet.

A well-publicized aspect of the developmentof the mortgage securitization process was thedevelopment of residential MBSs composed ofmany different types of mortgages, including sub-prime mortgages. Unlike the earlier securitizedofferings of the government-sponsored agencyGinnie Mae, which were subject to zero-defaultrisk, these private-label MBSs were subject tosignificant default risk. Securitization of sub-prime mortgages started in the mid-1990s, bywhich time markets had become accustomed tothe properties of securitized primemortgage prod-ucts that had emerged in the 1980s, but unlikegovernment or prime private-label securities, theunderlying assets in the subprime category werequite diverse.

The complexity of new products issued bythe private sector was much greater, introducingmore variable cash flow, greater default risk forthe mortgages themselves, and considerable het-erogeneity in the tranches. In an earlier issue ofthis Review, Chomsisengphet and Pennington-Cross (2006) show that the subprime mortgageshad a wide range of loan and default risk charac-teristics. There were loans with options to deferpayments, loans that converted from fixed to flexi-ble (adjustable-rate) interest rates after a givenperiod, low-documentation mortgages—all ofwhich were supposedly designed to help buyersenter the housing market when (i) their credit orincome histories were poor or (ii) they had expec-tations of a highly variable or rising income streamover time. Not all the mortgages offered as sub-prime were of low credit quality, but among thepool were many low-quality loans to borrowerswho relied on rising house prices to allow refi-nancing of the loan to ensure that they could affordto maintain payments. The link between defaultrisk and the movement of house prices was notfully appreciated by investors who provided aready market for such securitized mortgages inthe search for higher yields in the low-interest-rateenvironment. These included banks, insurancecompanies, asset managers, and hedge funds.Developments in the securitized subprime mort-

Mizen


gage market were the trigger for the credit crunch.For this reason, the crisis is often referred to as a“subprime crisis.” In fact, as we shall see, anynumber of high-yield asset markets could havetriggered the crisis.

Subprime as a Trigger for the CreditCrunch

Conditions in the housing and credit marketshelped fuel the developing “crisis.” Credit scoresof subprime borrowers through the decade 1995-2005 were rising; loan amounts on average weregreater, with the largest increases to those borrow-ers with higher credit scores; and loan-to-valueratios were also rising (see Chomsisengphet andPennington-Cross, 2006). The use of brokers andagents on commission driven by “quantity notquality” added to the problem, but provided themortgagees did not default in large numbers (trig-gering clauses in contracts that might require theoriginator to take back the debts), there wasmoneyto bemade. Mortgages were offered at low “teaser”rates that presented borrowers affordable, but notsustainable, interest rates, which were designedto increase. Jaffee (2008) suggested that the sheerrange of the embedded options in the mortgageproducts made the decision about the best pack-age for the borrower a complex one. Not all con-ditions were in the borrower’s best interests; forexample, prepayment conditions that limit thefaster payment of the loan and interest other thanaccording to the agreed schedule often were evenless favorable than the terms offered to primeborrowers. These conditions were designed todeter a borrower from refinancing the loan withanother mortgage provider, and they also made iteasier for the lender to sell the loan in a securitizedform. In addition, brokers were not motivated asmuch by their future reputations as by the feeincome generated by arranging a loan; in someinstances, brokers fraudulently reported infor-mation to ensure the arrangement occurred.

Policymakers, regulators, markets, and thepublic began to realize that subprime mortgageswere very high-risk instruments when defaultrates mounted in 2006. It soon became apparentthat the risks were not necessarily reduced by

pooling the products into securitized assetsbecause the defaults were positively correlated.This position worsened because subprime mort-gage investors concentrated the risks by leverag-ing their positions with borrowed funds, whichthemselves were funded with short-term loans.Leverage of 20:1 transforms a 5 percent realizedloss into a 100 percent loss of initial capital;thus, an investor holding a highly leveragedasset could lose all its capital even when defaultrates were low.6

U.S. residential subprime mortgage delin-quency rates have been consistently higherthan rates on prime mortgages for many years.Chomsisengphet and Pennington-Cross (2006)record figures from the Mortgage BankersAssociation with delinquencies 5½ times higherthan for prime rates and foreclosures 10 timeshigher in the previous peak in 2001-02 duringthe U.S. recession. More recently, delinquencyrates have risen to about 18 percent of all sub-prime mortgages (Figure 4).

Figure 4 shows the effects of the housingdownturn from 2005—when borrowers seekingto refinance to avoid the higher rates found theywere unable to do so.7 As a consequence, sub-prime mortgages accounted for a substantial pro-portion of foreclosures in the United States from2006 (more than 50 percent in recent years) andare concentrated among certain mortgage origi-nators. A worrying characteristic of loans in thissector is the number of borrowers who defaultedwithin the first three to five months after receivinga home loan and the high correlation betweenthe defaults on individual mortgage loans.

Why did subprime mortgages, which com-prise a small proportion of total U.S. mortgages,transmit the credit crunch globally? The growthin the scale of subprime lending in the UnitedStates was compounded by the relative ease withwhich these loans could be originated and thereturns that could be generated by securitizing

Mizen


6 This is why Fannie Mae and Freddie Mac faced difficulties in July2008, because small mortgage defaults amounted to large losseswhen they were highly leveraged.

7 In the United States the process of obtaining a new mortgage topay off an existing mortgage is known as “refinancing,” whereasin Europe this is often referred to as “remortgaging.”

the loans with (apparently) very little risk to theoriginating institutions. Some originators usedtechnological improvements such as automaticunderwriting and outsourcing of credit scoringto meet the requirements of downstream pur-chasers of the mortgage debt, but there is anec-dotal evidence that the originators cared littleabout the quality of the loans provided they metthe minimum requirements for mortgages to berepackaged and sold. The demand was strong forhigh-yielding assets, as the Governor of the Bankof England explained in 2007 (King, 2007):

[I]nterest rates…were considerably below thelevels to which most investors had becomeaccustomed in their working lives. Dissatis-factionwith these rates gave birth to the “searchfor yield.” This desire for higher yields couldnot be met by traditional investment opportu-nities. So it led to a demand for innovative, andinevitably riskier, financial instruments andfor greater leverage. And the financial sectorresponded to the challenge by providing ever

more sophisticated ways of increasing yieldsby taking more risk.

Much of this demand was satisfied by resi-dential MBSs and CDOs, whichwere sold globally,but as a consequence the inherent risks in thesubprime sector spread to international investorswith no experience or knowledge of U.S. realestate practices. When the lenders foreclosed, theclaims on the underlying assets were not clearlydefined—ex ante it had not been deemed impor-tant. Unlike in most European countries wherethere is a property register that can be used toidentify—and repossess—the assets to sell themto recoup a fraction of the losses, the United Stateshas no property register that allows the lender torepossess the property. As a consequence, oncethe loans had been pooled, repackaged, and soldwithout much effort to define ownership of theunderlying asset, it was difficult to determine whoowned the property. Moreover, differences in thevarious state laws meant that the rules permitting

Mizen


1998 1999 2000 2001 2002 2003 2004 2005 2006

20

18

16

14

12

10

8

6

4

2

0

All MortgagesSubprimePrime

Percent

2007 2008

Figure 4

U.S. Residential Mortgage Delinquency Rates

SOURCE: Mortgage Bankers Association/Haver Analytics.

the lender to pursue the assets of the borrowerwere not uniform across the country.

It has been commonly asserted that the rootof the problem lies with the subprime mortgagemarket in the United States, but this is not thefull story. Subprime was the trigger for the crisis,but mispricing of risk was widespread, and anynumber of other high-yield asset classes couldhave provided the trigger (e.g., hedge funds, pri-vate equity, emerging market equity). Originatorswere willing to sell and investors were willing tobuy securitized products in subprime mortgagemarketswith complex characteristics because ofthe high returns. High yields on these productsmade them attractive to international investors,and the crisis spread internationally, influencingmany other financial markets. Fundamentally,sellers of subprime mortgage securities mispricedrisks by using models that assumed house priceswould continue to rise, while interest ratesremained low. The investment climate of the

timemeant risks of many kinds were underpriced,with unrealistic assumptions about rising valua-tions of underlying assets or commodities. There-fore any number of other high-yielding assetclasses could have started the crisis—it so hap-pened that the subprime market soured first.

The complexity of the structured productsincreased the difficulty of assessing the expo-sure to subprime and other low-quality loans.Even after the credit crunch influenced the capi-tal markets in August 2007, many banks spentmonths rather than weeks evaluating the extentof their losses. The doubts about the scale ofthese losses created considerable uncertainty inthe interbank market, and banks soon becamereluctant to lend to each other unless they werecompensated with larger risk premiums.

The Response in the Markets

Capital and Money Market Paralysis. Theeffects of the subprime mortgage defaults created

Mizen


Jan-0

4

Mar

-04

May

-04Ju

l-04

Sep-0

4

Nov-04

Jan-0

5

Mar

-05

May

-05Ju

l-05

Sep-0

5

Nov-05

Jan-0

6

Mar

-06

May

-06Ju

l-06

Sep-0

6

Nov-06

Jan-0

7

Mar

-07

May

-07Ju

l-07

Sep-0

7

Nov-07

Jan-0

8

Mar

-08

May

-08

Billions1,300

1,200

1,100

1,000

900

800

700

600

ABCP

Non-ABCP

Figure 5

Commercial Paper

SOURCE: Federal Reserve Board/Haver Analytics.

a reappraisal of the hazards of all types of riskyassets. The first effect was seen in capital markets.In June and July 2007, many assets backed bysubprime residential MBS products were down-graded by the ratings agencies from AAA to A+(four notches down)—an unusually large down-grade given that downgrades normally occur insingle notches. The OECD described these down-grades as “unexpected” and indicated that this“exposed ratings agencies to considerable criti-cism” (OECD, 2007). The ratings agencies beganto reassess their ratings procedures for theseproducts, thereby introducing further uncertaintyabout the reliability of their ratings.

Conduits and SIVs had funded their purchasesof CDOs and other securitized assets by issuingtheir own asset-backed commercial paper (ABCP)at short maturities. The expansion of mortgage-related ABCP issuance accounted for half thegrowth in the commercial paper market in recent

years. The ABCP needed to roll over periodically,usually monthly, but as investors were less will-ing to purchase short-term paper in the capitalmarkets, these entities could not obtain the nec-essary short-term funding from these markets.Figure 5 shows that ABCP issuance peaked inJuly 2007 and fell sharply in subsequent months.

As a result of these developments, Bear Stearnswarned investors on July 18 that they would losemoney held by hedge funds in subprime-relatedassets and an IKB Deutsche Industriebank AGconduit incurred losses and was not able to rollover its ABCP; it drew on a credit line from itsparent bank but this was insufficient and IKB wasbailed out through a fund organized by its majorshareholder, KfW Bankengruppe, on August 7,2007. Two days later, BNP Paribas suspendedwithdrawals from three hedge funds heavilyinvested in CDOs that it was unable to value. OnAugust 17, Sachsen LB, a German bank, had failedto provide enough liquidity to support its conduitOrmond Quay, and Sachsen LB was taken overby Landesbank Baden-Württenberg (LBBW) atthe end of August. The need for rollover fundingby conduits and SIVs created pressure on banks’liquidity, giving them little incentive to lend onthe interbank market to other banks or to investin short-term paper. The spread between the ABCPrate and the overnight interest swap rate (the rateon overnight lending converted to the same matu-rity as the ABCP assets using a fixed-rate swaprate), which measures the default and liquidityrisk of ABCP, rose substantially by more than100 basis points in August 2007.8

Banks hoarded liquidity to cover any lossestheymight experience on their own books throughconduits, or those of their SIVs, which might needto be taken back onto their balance sheets. Theselosses turned out to be substantial and involvelarge investment banks, such as UBS, MerrillLynch, and Citigroup (Table 1), whose CEOswouldpay the price by resigning as losses were revealed.

The uncertainty associated with the scale ofthe losses that banks might face created a disloca-tion in the interbank markets. Banks would not

Mizen


Table 1Top Corporate Writedowns

WritedownsBank (billion U.S.$)

Citigroup 46.40

Merrill Lynch 36.80

UBS 36.70

AIG 20.23

HSBC 18.70

RBS 16.50

IKB Deutsche 14.73

Bank of America 14.60

Morgan Stanley 11.70

Deutsche Bank 11.40

Ambac 9.22

Barclays 9.20

Wachovia 8.90

MBIA 8.41

Credit Suisse 8.13

Washington Mutual 8.10

HBOS 7.50

SOURCE: Reuters.

8 1 basis point (bp) = 1/100 percentage point.

lend to other banks for fear of the scale of counter-party risk. If borrowing banks had unrevealedlosses they might not repay the funds that theyborrowed from other banks. The market responsewas demonstrated by two other interest ratespreads shown in Figure 6: the LIBOR-OIS spread(the London Interbank Offered Rate [LIBOR]minus the overnight index swap rate [OIS]) andthe Treasury-eurodollar (TED) spread. The firstspread reflects the difference between the rate atwhich banks will lend to each other, say for oneor three months, compared with the overnight

indexed swap (OIS) rate, which jumped 100 basispoints.9 Secondly, the TED spread, which is thedifference between the U.S. Treasury bill rate andthe eurodollar rate, widened even more. Thisreflected the desire to shift into safe U.S. Treasuriesand the desire to obtain Treasuries as collateral.These effects were observed in the LIBOR andEURIBORmarkets, as well as in the United States,resulting in a global freeze in capital and moneymarkets.

The growing concern caused a sharp drop inthe issuance of asset-backed securities, particu-larly those of lower quality, in August 2007. Alltypes of asset-backed securities and CDOs wereadversely affected from September 2007, subprimeresidential MBSs and CDOs of asset-backed secu-rities issues shrank, and even prime residentialMBSs were substantially lower (Figure 7).Investors realized that the assets were riskier thanhad previously been thought, and the cost ofinsurance to cover default risk using credit defaultswaps (CDS) also had become much more expen-

Mizen


4/13

/07

4/30

/07

5/15

/07

5/30

/07

6/14

/07

6/29

/07

6/17

/07

7/31

/07

8/15

/07

8/30

/07

9/14

/07

10/0

1/07

10/1

6/07

10/3

1/07

11/1

5/07

11/3

0/07

12/1

7/07

1/01

/08

1/16

/08

31/0

1/08

2/15

/08

3/03

/08

3/18

/08

4/02

/08

4/17

/08

5/02

/08

5/19

/08

6/03

/08

6/18

/08

7/03

/08

7/18

/08

Percentage Points

1.5

1.0

0.5

0

–0.5

–1.0

–1.5

–2.0

T-Bill–Eurodollar

Libor-OIS

–2.5

Figure 6

Interest Rate Spreads

SOURCE: Federal Reserve Board, Financial Times, Reuters, and Haver Analytics.

9 The LIBOR-OIS spread is the spread most often used by centralbanks to describe the increase in the cost of interbank lending,reflecting credit and liquidity risk. See Arain and Song (2008, p. 2)and Bank of England (2008, p. 15). LIBOR is set by the BritishBanker’s Association in London. The LIBOR is fixed by establishingthe trimmed average of rates offered by contributor banks on thebasis of reputation and scale of activity in the London interbankmarkets. There is also a dollar LIBOR that determines rates at whichbanks offer U.S. dollars to other banks. EURIBOR is calculated ina similar way for prime European banks by Reuters, with a fewminor differences.

Mizen


Mar-05 Sep-05 Mar-06 Sep-06 Mar-07 Sep-07 Mar-08 Sep-08

CLOs

Other ABS

Other CDOs

CDOs and ABS

CMBS

Subprime RMBS

Prime RMBS

900

800

700

600

500

400

300

200

100

0

$ (billions)

Figure 7

Global Issuance of Asset-Backed Securities and CDOs

SOURCE: Bank of England, Dealogic, and Sifma.

0

30

60

90

120

150

180

210

240

270

300

330

Jan-07 Mar-07 May-07 Jul-07 Sep-07 Nov-07 Jan-08 Mar-08

Basis points

U.S. Securities HousesU.S. Commercial BanksMajor U.K. BanksEuropean LCFIs

Figure 8

Credit Default Swap Premia

NOTE: Data are valid through close of business April 22, 2008. “Premia” indicates asset-weighted average five-year premia.

SOURCE: This figure is reprinted with permission from the Bank of England’s April 2008 Financial Stability Report, Chart 2.18, p. 35.Data are from Markit Group Ltd, Thomson Datastream, published accounts, and Bank of England calculations.

sive.10 Figure 8 indicates that CDSmarkets peakedin August, making insurance costly, and asset-backed securities issues were therefore more diffi-cult to sell. Since that August they have reachedfurther highs, culminating in the peak of March2008 before the Bear Stearns rescue.

The upshot of these events had two importantimplications. First, because the capital marketswere effectively closed for certain types of asset-backed securities, particularly the riskiest types,it became difficult if not impossible for banks toevaluate their exposure to these products andquantify their losses. In the absence of a liquidmarket for these products fromwhich to determinea current price, the best possible solution was toattempt to predict prices—so mark-to-market wasreplaced by mark-to-model, but it was not possibleto establish whether these prices were accurate.Under U.S. accounting standard FASB 157 (onfair value measurement), banks are required tovalue their assets according to a hierarchy of threelevels. Level 1 uses market prices, level 2 usesmarket-based inputs including interest rates orcredit spreads, and level 3 values assets using onlymodel information, relying on assumptions andextrapolations, not market data. As secondarymarkets for many asset-backed securities andCDOs dried up, the valuation of portfolios andlosses stepped down from level 2 to level 3.

The second implication in August 2007 wasthat the LIBOR-OIS spreads increased markedlyas the supply of funds dwindled but did notreturn to normal.11 The widening spreads werefar from a temporary phenomenon; these spreadswere high for an extended period, which had anadverse effect on certain financial institutionsthat depended on the markets for their fundingand on their depositors. Commercial banks with

funding models that relied on short-term com-mercial paper found that they could not obtainfunds to provide new loans. Similarly, investmentbanks that had relied on short-term paper to pur-chase asset-backed securities were unable to makepayments when they were due. The result of thedislocation in the capital and money marketswould lead to the Northern Rock bank run in theUnited Kingdom and the threat of bankruptcy forBear Stearns in the United States (these topics arediscuss in greater detail later), but the actions ofthe authorities to provide more liquidity in themarkets are considered first.

The Need for Market and FundingLiquidity

Market Liquidity. Central banks providedfunding liquidity for distressed institutions andmarket liquidity.12 The actions of the Fed, theBank of England, and the European Central Bank(ECB) were initially different, but there wasconvergence as the crisis evolved. On August 17,2007, the Fed extended its normal lending periodto 30 days and cut the interest rate offered tobanks at the discount window by 50 basis points,acting swiftly and decisively. This was followedby cuts to the federal funds target rate of 50 basispoints on September 18 and two cuts of 25 basispoints in quick succession on October 31 andDecember 11. The ECB also acted quickly to stemthe crisis by moving forward auctions for liquid-ity by injecting €94.8 billion, with more opera-tions totalling €108.7 billion in the followingweeks, to “frontload” the liquidity operationsinto the first part of the maintenance period.13

Mizen


10 A financial institution buying a claim to a package of mortgages orloans can insure itself against default on the underlying repaymentsthrough the credit default swap market (CDS). A fixed premium isexchanged for payment in the event of default. As the probabilityof default rises, so do the premia. There is a primary market forCDS and a secondary market known as the CDX (Commercial DataExchange) market in the United States and iTraxx in Europe.

11 McAndrews, Sarkar, and Wang (2008) indicate that “rates of inter-bank loans with maturity terms of one-month or longer rose tounusually high levels”; they also add that “borrowers reportedlycould not obtain funds at posted rates.”

12 “Funding liquidity” refers to the ease of access to external financeand depends on the characteristics of the borrower. When a bor-rower is not regarded as creditworthy, it may face higher borrowingcosts and quantity restrictions that present a funding problem; thiswill need to be resolved by borrowing from nonmarket sources, andin the case of a bank, from the central bank. Market liquidity is aproperty of the relative ease with whichmarkets clear at a fair value.When markets become very thin, the authorities may intervene toensure they are able to clear, by for example “making the market”by accepting certain assets in exchange for more liquid ones.

13 Central banks may require commercial banks to hold a certainproportion of their deposits at the central bank; the proportion iscalculated over a “maintenance period.” The proportion may bemandated or voluntary, but once set it is usually enforced onaverage over the relevant period.

But it kept interest rates steady. The Bank ofEngland started to respond to the money marketshortage later than other central banks. In August2007 when approached by the commercial banksto provide further liquidity at no penalty to theborrower, it refused. As a consequence, the com-mercial banks increased their reserves targets by6 percent in the maintenance period beginningSeptember 6, 2007.14 The Bank responded bypromising to supply an additional 25 percentof the reserves target if interbank markets didnot normalize, and when they did not do so,on September 13, they increased the supply ofreserves. Ultimately the Bank of Englandincreased liquidity provision by 42 percent fromAugust 2007 to April 2008.

Central banks found that they had to be inno-vative in issuing liquidity directly to the mosttroubled parts of the financial system by develop-ing term lending. The problem for the centralbanks was that although there was plenty of liq-uidity in overnight markets, there was a shortageof funds at 1-, 3- and 6-month maturities wherethe banks needed it, causing the cost of funds atthese maturities to rise. The standard tools didnot work well in dealing with this problem.Although central banks would normally have usedstanding facilities to provide more liquidity tothe markets, recourse to borrowing from the cen-tral bank through standing facilities was seen asan indicator of weakness that carried with it acertain stigma. In the United Kingdom, Barclaysbank experienced repercussions in the equity mar-kets when it borrowed from the Bank of Englandin August 2007. For this reason, commercial banksin the United States bypassed the discount win-dow and borrowed instead for one-month termsfrom the markets, because rates were almost equalon average to the expected discount rate and didnot carry any stigma (see Armantier, Krieger, andMcAndrews, 2008, p. 4). Banks also increased

borrowing from the Federal Home Loan Banks.15

The FHL system provided $200 billion of addi-tional lending in the second half of 2007.16

Central banks found it very hard to keepshort-term market interest rates on 1-month and3-month LIBOR (the interbank lending rate) closeto OIS rates at the same maturity despite the factthat overnight rates were kept at their desiredlevels. The disparity at 1- and 3-month maturitiesreflected banks’ anticipation of the need for fund-ing at that maturity that they could no longereasily obtain from these markets. Standing facili-ties were not addressing the problem because ofstigma in the markets, so there were moves todevelop term lending. A significant feature ofthe response to the credit crunch has been therecognition that the markets needed liquidity atmaturities longer than overnight. The develop-ment of term lending has been the means adoptedby central banks to provide liquidity at terms of1 month, 3 months, and 6 months. Outside theUnited States this has also involved extendingthe types of collateral that they are willing toaccept (i.e., non-government-asset-backed secu-rities such as AAA-rated private sector securitiesincluding residential MBSs).17

The ECB was the first institution to lend atlonger maturities, thereby offering help toEuropean banks by lending against a wide rangeof collateral, including mortgage securities. Itinitiated a supplementary liquidity-providinglonger-term refinancing operation with a maturityof 3 months for an amount of €40 billion onAugust 22, 2007, and a second operation on

14 The Bank of England’s money market operations mechanism allowseligible banks to choose a target level of positive balances (voluntaryreserves) that they will be required to hold with the Bank on aver-age over a maintenance period lasting from one monetary policymeeting to the next. Reserves held are remunerated at Bank rate.The Bank is able to set ceilings on individual institutions’ reservestargets when demand for reserves is high.

Mizen


15 There are 12 Federal Home Loan (FHL) Banks, which are ownedby 8,100 member financial institutions in the United States. Theirpurpose is to provide stable home loan funding to their memberinstitutions. The FHL Banks issue AAA-rated debt through the U.S.Office of Finance to fund their loans. Financial institutions wereable to obtain funds from the FHL Banks by exchanging assets suchas residential MBSs for liquid assets such as U.S. Treasuries. TheFHL Banks’ members historically have been smaller banks andthrifts, but this has been changing in recent years and the lendingof the FHL Banks has broadened to include many larger banks.

16 See www.fhlb-of.com/specialinterest/financialframe.html forinformation on these additional loans.

17 The Federal Reserve Open Market Desk has accepted only U.S.Treasuries, government-sponsored agencies debt, and theirmortgage-backed securities, but at the discount window they haveaccepted a much broader range of collateral.

September 6, 2007, without a specified limit,again at a 3-month maturity. This move wasquickly followed on September 19, 2007, by theBank of England’s announced plans for an auctionof £10 billion at a 3-month maturity against a widerange of collateral, including mortgage collateral,with three further auctions offering £10 billion atweekly intervals. The Bank of England recordedin April 2008 that three-quarters of its lending wasat terms of 3 months or longer, up from about one-third since the beginning of the credit crunch.

On December 12, 2007, the Federal Reserveannounced a term auction facility (TAF) to allowU.S. banks to bid anonymously for a predeter-mined amount of one-month money direct fromthe Fed to ensure an efficient distribution of fundsto banks to augment the stigma-ridden discountwindow. The TAF was designed to reduce thepremium in interest rate spreads for liquidity riskbymaking liquidity available at thematurity termsrequired by the financial system. The TAF had anumber of new features that combined attributesof open market operations and discount windowlending. Distributions of funds were arrangedthrough auctions of fixed amounts (as were openmarket operations). This allowed the FederalReserve to (i) determine how much and whenfunds would be injected into the markets, (ii)ensure that the process of obtaining funds wascompetitive (and therefore not subject to stigma),and (iii), broadly based, offer funds to a largernumber of banks. Similar to discount windowlending, the lending was on a collateralized basisusing collateral that was acceptable for discountwindow lending. A bidder for funds through theTAFwould be required to offer a bid above a mini-mum market-determined rate; the Fed wouldimpose a cap on the size of the bid at 10 percentof the total auction size and would distributefunds at a single-price once the auction was com-pleted.18 The first TAF auction of $20 billion wasscheduled to provide 28-day-term funds andincluded facilities to swap dollars for euros; there

have been 16 auctions for amounts varying from$20 billion to $75 billion up to July 2008.19

In March 2008, the Federal Reserve estab-lished two further facilities: a primary dealer creditfacility intended to improve the ability of primarydealers to provide financing to non-bank partici-pants in securitization markets and promote theorderly functioning of financial markets moregenerally, and a weekly term securities lendingfacility to offer Treasury securities on a one-month loan to investment banks against eligiblecollateral such as residential MBSs. Totaling allthe sources of new liquidity made available bythe Federal Reserve, Cecchetti (2008c) estimatedin April 2008 that the liquidity committed so faramounts to nearly $500 billion ($100 billion tothe TAF; $100 billion in 28-day repurchases ofMBSs; $200 billion to the term securities lendingfacility; $36 billion in foreign exchange swapswith the ECB; $29 billion to facilitate acquisitionby JPMorgan Chase of Bear Stearns; and $30 billionto the primary dealer credit facility). There havebeen larger TAF auctions of $150 billion sinceApril, but term securities lending and primarydealer credit have been lower, at $143 billion and$18 billion, respectively. The Federal Reserve hastaken major steps to intervene in the markets toensure that banks can obtain funds efficiently, butin doing so it has offered Treasuries in exchangefor eligible collateral, not cash, and these provideliquidity in the sense they have a well-functioningmarket for their exchange into cash.

The Bank of England also injected marketableassets into the banking system through a newlydevised special liquidity scheme implementedApril 21, 2008 (see Bank of England, 2008). Thisprovides long-term asset swaps to any bank orbuilding society eligible to borrow from the Bankusing its standing facilities. Under the swaparrangement the Bank stands willing to exchangeexisting AAA-rated private sector securities thatwere issued before December 2007 for governmentsecurities for up to a year, with the provision toroll over the swaps for up to three years. The priceof the swaps is determined by the riskiness ofthe underlying assets and does not release 100percent of the face value of the private securitiesbeing exchanged, but it injects a substantial

Mizen


18 The minimum rate is the OIS one-month swap rate and the agreedprice for the distribution is the “stop-out rate”; see McAndrews,Sarkar, and Wang (2008).

19 See www.federalreserve.gov/monetarypolicy/taf.htm for furtherdetails of the TAF auction dates and amounts.

amount of marketable government securities intothe markets that can be exchanged on markets toprovide the vital additional liquidity required.When the scheme was unveiled, the value of theswaps was expected to be up to £50 billion.

Funding LiquidityThe Northern Rock Bank Run. The paper by

Alistair Milne and Geoffrey Wood in this issueof the Review details many of the developmentsin the Northern Rock bank run, so the discussionhere is brief. Northern Rock had adopted a busi-ness model that relied very heavily on wholesalefunding and securitization of its mortgages (Houseof Commons Treasury Committee, 2008a,b,c,).Funding from the increase in retail deposits wasonly 12 percent of total sources of new funding.Of the wholesale borrowing it undertook, 50percent was short-term, at less than one year tomaturity, and among the securitized bonds itissued £6 billion were purchased by its mastertrust Granite and funded using ABCP with matu-rities of one to three months. The funding modeldepended on regular access to both capital andmoney markets to fund the bank’s activities.Although Northern Rock had adequate liquidityto cover shortages of wholesale funds for briefperiods (as evidenced by the 9/11 episode when,according to its then-chairmen giving evidencebefore a Parliamentary committee, it rode outthe liquidity shortage that lasted for a few days),it could not endure a long freeze in money mar-kets. The problem for Northern Rock was that ithad not envisaged a simultaneous freeze of allits sources of short-term finance, and it had nottaken insurance against this eventuality (Houseof Commons Treasury Committee, 2008a,b,c).

As the possibility of funding problemsemerged, the Bank of England, the FinancialServices Authority, and the HM Treasury, whichwere jointly responsible for financial stability,considered three options: (i) to allow NorthernRock to resolve its funding problems in the mar-kets, (ii) to seek a liquid buyer from among U.K.banks, or (iii) to rescue the bank using publicmoney through a support operation by the Bankof England backed by the Treasury. Initially, theauthorities opted for a support operation, but aleak of the details by the broadcast media before

an official announcement could be made pre-cipitated a run on the bank between Friday,September 14, and Monday, September 17, afterwhich the Treasury announced a guarantee in fullof the deposits in Northern Rock. Subsequentefforts to find a liquid buyer were attempted butfailed and the bank was brought into publicownership at a cost of £25 billion in loans fromthe Bank of England and other guarantees fromHM Treasury.

Milne and Wood (2008) note that it was thefirst run since the nineteenth century on a Britishbank of any significance in the British bankingsystem, and Brunnermeier (2008) rightly consid-ers Northern Rock to be a classic bank run, butthese events were highly unusual for two reasons.First, the run was triggered by the leak of informa-tion about an operation planned by the authoritiesto support the bank in its difficulties. Second, itwas entirely contained within just one institutionand did not spread to other banks. On the contrary,depositors withdrawing money redeposited theircash in other banks, and the change in bankdeposits by individuals in 2007:Q3 rose by £9.1billion and continued to grow in 2007:Q4. Thissuggests that the banking model of Northern Rockwas largely to blame, but also that the unfortunaterevelation of support procedures intended torescue an institution in trouble before an officialannouncement could be made resulted in anadverse signal to the markets—the opposite ofwhat was intended. The banking system itselfwas not distrusted, just Northern Rock.

The run on Northern Rock occurred becauseit used a business model that was inherently riskyif the financing of its mortgages, held for sale asMBSs by Granite through the issue of short-termasset-backed paper, could not be rolled over. Asimilar failure occurred in the United States whenHome State Savings Bank of Cincinnati, Ohio,failed.20 Home State Savings had about $700 mil-lion in deposits in 1985 when it ran into troublebecause a rapidly expanded new businessfinanced by the issue of short-term paper failed.Home State Savings Bank had bought Ginnie Mae

20 I thank Dick Anderson, who observed firsthand both the HomeState loan run in Columbus, Ohio, in 1985 and the Northern Rockrun in Birmingham, United Kingdom, in 2007.

Mizen


MBSs and U.S. Treasuries from E.S.M. of FortLauderdale, Florida. It had financed the purchaseby issuing its own short-term paper with a one-year maturity, which it sold back to E.S.M. WhenE.S.M. collapsed, Home State Savings’ lossesthreatened its banking business. This precipitateda bank run that threatened to spread to other insti-tutions because the losses of Home State Bankabsorbed almost all of the Ohio state deposit insur-ance fund, leaving all other savings and loanscompanies effectively without deposit insurance.The governor of Ohio closed 71 institutions untilthey were able to obtain federal deposit insurance.The nature of this run was very similar to that ofNorthern Rock inasmuch as it resulted from arapidly expanded new business that the regulatorsand the bank itself failed to recognize as highlyrisky, which subsequently caused the institutionto fail.

Bear Stearns. The response of the U.K. gov-ernment to the Northern Rock run recognizedthe need to protect commercial bank depositorsfrom the fallout in the financial system followinga funding problem. The move in recent months bythe Federal Reserve to rescue the private sectorinvestment bank Bear Stearns has been an attemptto limit the damage of the crunch on settlementin the financial system more generally. BearStearns’s hedge funds had invested heavily instructured finance products because these allowedthe actual leverage ratio to be much higher thanthe reported leverage ratios on funds undermanagement.21 Concerns had mounted over thedegree of leverage and the quality of the MBSsin which Bear Stearns had invested. Reportedly,Goldman Sachs had provided indications to thehedge fund Hayman Capital that it would not takeexposure to Bear Stearns. As news spread ofthis warning, an investment bank run occurred,reducing Bear Stearns’ ability to finance its activ-ities. These had been funded by the sale of short-term ABCP assets and had been rolled over reg-ularly, but on Friday, March 14, 2008, it became

clear that Bear Stearns would not be able to rollover the assets as normal and as a result wouldfail to meet payments due on Monday, March 17.To avoid the costly unraveling of over-the-counterinterest rate, exchange rate, and credit defaultderivatives—for which Bear Stearns was a coun-terparty—that might threaten to bring into bank-ruptcy other financial institutions, includingJPMorgan Chase, Bear Stearns’ banker, the FederalReserve Bank of New York stepped in to supportthe institution with a 28-day loan via JPMorganChase. Analysis over the weekend revealed thata takeover would be necessary, and this wasarranged through a shares purchase by JPMorganChase initially set at $2 per share, but laterincreased to $10 per share to placate shareholdersand ensure the deal would be accepted, combinedwith a $29 billion loan from the Federal Reserve,and with JPMorgan Chase taking on the first $1billion of losses to Bear Stearns. The actionsaverted a financial system crisis that might haveresulted in what Brunnermeier (2008) refers toas “network and gridlock risk,” and interventionappears to have prevented this from occurring.

Freddie Mac and Fannie Mae. In differentcircumstances than those of Bear Stearns, FreddieMac and Fannie Mae received support from theU.S. Treasury following advice from the FederalReserve Bank and the Securities and ExchangeCommission (SEC) in July 2008.22 Confidencein the institutions’ ability to raise $3 billion ofnew funds through an auction in the marketswas fragile. Freddie Mac and Fannie Mae heldMBSs that they had issued in their own nameor bought to encourage “affordable” loans at thebehest of HUD. Many of these were subprimemortgages, which were affected by the downturnin house prices, and rising delinquencies ontheir own mortgages or those they insured forothers pointed to further financial problemsahead. A fall of 20 percent in the value of theequity of the institutions in mid July 2008reflected the fears of lower future profitability

Mizen


21 Brunnermeier (2008) reports that Bear Stearns’ Asset ManagementFund reported leverage ratios of 2:1 and 3:1 on, respectively, High-Grade Structured Credit Strategies Fund and its Enhanced LeverageFund, but CDO investments would have increased these leverageratios considerably.

22 Freddie Mac and Fannie Mae are government-sponsored mortgageagencies with debts of $1.5 trillion, direct guarantees to mortgagesto the value of $5 trillion, and insurance for a further $2 trillion ofother institutions’ mortgages, which means, directly or indirectly,they support more than half of the $12 trillion U.S. mortgagemarket.

and the circulation of suspicions by LehmanBrothers that between them they would need toraise $75 billion in additional funding, whichcould dilute ownership. The scale of the capitalrequired was small in relation to the size of thecompanies, but failure to obtain a relatively smallamount of funding would question the credibil-ity of the institutions and if that meant the debtsecurities issued by the mortgage agencies mightdecline in value, greater problems would thenoccur for other financial institutions. Many banks,money market funds, and pension funds holdFreddie Mac and Fannie Mae debt securities andused them as collateral for borrowing. The pos-sibility that agencies’ government-sponsoredMBSs might be sold off by investors was a majorconcern. The proposal put forward by U.S.Treasury Secretary Hank Paulson on Sunday,July 14, 2008, involved a credit line of $300 bil-lion as a temporary measure; the Housing andEconomic Recovery Act of 2008 passed byCongress in late July approves the plan to allowthe Treasury to purchase debt securities andshares in the agencies with the agreement ofthe companies until December 31, 2009, whenthe authority expires. Once again, failure of theinstitutions to continue to operate as normalwould have resulted in a severe dislocation inthe financial system.

EVALUATIONThe Problems with “Originate andDistribute” Banking

A number of commentators, includingAlexandre Lamfalussy and Willem Buiter,23 havenoted that banks have replaced their traditional“originate and hold” model of lending long andborrowing short, with an “originate and distribute”model, in which they lend and then sell the claimsto someone else. They argue that the widespreadadoption of an “originate and distribute” modelwas responsible for the crisis. It is difficult to dis-

agree, but securitization has been operating for40 years without associated crises, so somethingmore is at work. The change in the past decadehas been the growth in residential MBSs backedby subprime mortgages with a larger number ofsteps between originator and holder and, as aconsequence, greater opacity. This has contributedto the mispricing of risk that was not properlyappraised. The result is twofold: Investors are farremoved from the underlying assets both physi-cally (due to the global market for these assets)and financially (since they often have little ideaabout the true quality and structure of the under-lying assets several links back in the chain). TheInternational Monetary Fund has referred to thisas an arm’s-length financial system in itsWorldEconomic Outlook for 2006, andMonacelli (2008)calls it an “atomistic” model. Equity and bondmarkets can have these features too, but structuredfinancial products are far more complex instru-ments. The extension of originate and distributebanking to subprime mortgage securities has cre-ated an asset class with an opaque ownershipstructure and therefore imprecision concerningwho holds the underlying risks. This feature hasdistorted the incentive structure at every step inthe process and greatly complicated the assess-ment of risks because few investors understandthe structure from top to bottom. Ultimately thisis responsible for the crisis.Poor Incentive Structures Under “Originate

and Distribute” Banking. The problem with theextended originate and distribute banking modellies in its weak incentives to measure risk accu-rately at any stage in the process. There mayhave been control measures in place, but thesewere allowed to slip. The model had six badlydesigned incentive mechanisms as illustratedby the experience in the period leading up tothe crisis.

First, brokers and agents of banks sellingmortgages were motivated by up-front fee incomeunadjusted for borrower quality. The bonusesrewarded growth of business over a short timescale (typically a yearly cycle) with no penaltiesif subsequent developments revealed a lack ofdue care and attention in the origination processor losses to the originator. There is evidence ofmanipulation of data, in some cases amounting

23 Respectively, they are the former general manager of the Bank forInternational Settlements and former chief economist of theEuropean Bank for Reconstruction and Development and U.K.Monetary Policy Committee member.

Mizen


to fraud, by brokers, who—with the exception oftheir appointed appraisers of property—were thesole point of contact with the borrower.24 Thesebrokers and agents were often not employees ofthe mortgage origination companies; therefore,they were strictly speaking outside the regulators’reach (see de la Dehesa, 2008).

Second, originators had no greater incentiveto look more carefully than brokers at borrowerquality. The incentives for the originators of theloans, faced with the knowledge that the productswould be combined in complex ways and sold,were different from those for an originator whointended to hold the assets to maturity. This fun-damentally altered the incentives of the seller. Inthe years before the crisis occurred, the originationof subprime mortgages increased rapidly becausemortgage originators needed new loans to pack-age and sell to investors; in the rush to providemore loans for securitization underwriting stan-dards were allowed to slip as uncritical use ofautomated underwriting systems and validatorswere introduced to ease the burden.25 In July 2008,the attorney general of Illinois, Lisa Madigan, fileda civil action against Countrywide for deceitfulconduct and lax standards in subprime mortgagelending with hidden fees and risky terms. More-over, Countrywide is accused of having “usedegregiously unfair and deceptive lending practicesto steer borrowers into loans that were destinedto fail.” This first action against Countrywide bya public prosecutor has been brought on behalfof thousands of borrowers.

Third, the profits from securitization createdincentives for originators to obtain new loansregardless of their quality provided they met mini-mum standards for resale.26 As the quantity ofnew borrowers declined, lenders reduced theirstandards to maintain the volume of loans feedinginto the securitization market. This generated anincreasing share of “NINJA” loans—so calledbecause the recipients had No verified INcome,Job, or Assets—and piggyback loans that com-bined two mortgages to cover the purchase of asingle residence. Anderson (2007) reports thatbetween 2003 and 2006 the market share of theNINJA loans doubled and the piggyback loansquadrupled. Later-stage securitized loans weretherefore much riskier than the earlier ones:Defaults on 2006 and 2007 vintages of subprimeloans are projected to be higher than default forearlier vintages.

Fourth, tranching enabled the SPVs to con-struct products with ratings suitable for certaintypes of investors. The senior tranche wouldobtain a AAA rating, suitable for pension funds;the next tranche would obtain BBB, suitable forconduits and SIVs; and so on. Equity tranches alsocould be rebundled with other equity tranchesinto CDOs with higher credit ratings, despitethe fact that they were complex combinations ofpoorer-quality mortgages in a more highly lever-aged form.

Fifth, ratings agencies made a large share oftheir profits from rating structured finance prod-ucts; for example, Portes (2008) reports Moody’sgenerated 44 percent of its revenues from theseactivities. There was scope for conflict of interestwithin ratings agencies because they were paidan up-front fee by the issuer to provide a ratingof the assets. At the same time, though, the samebusiness would sell advice to clients (for anotherfee) on how to improve those ratings, identifying“tranching attachment points” to make sure thesecuritized assets just attained the required ratingfor the intended investor group.

Mizen


26 If a certain proportion of the underlying mortgages defaulted, therewas often a clause that required the originator to take back therepackaged assets; but, provided the seller met some fairly minimalstandards to ensure the predicted default risk was acceptable tothe buyer, the originator could sell the mortgages at a profit.

24 The November 28, 2007, Fitch Ratings special report on “The Impactof Poor Underwriting Practices and Fraud in Subprime RMBSPerformance” cites BasePoint Analytics LLC, a fraud analyticsconsulting firm, which “analyzed over 3 million loans originatedbetween 1997 and 2006…including 16,000 examples of non-performing loans that had evidence of fraudulent misrepresentationin the original specifications. Their research found as much as 70%of early payment defaults contained fraud misrepresentation onthe application” (p. 1). Fraud might include occupancy misrepre-sentation, incorrect calculations of debt-to-income ratios, artificiallyhigh credit scores (based on authorized use of someone else’s credithistory), questionable stated income or employment, and so on.

25 When the scale of the early payment defaults became known in2007, the Fitch Ratings report urged mortgage originators to bemore vigilant regarding verification of stated income, credit scores,property valuation, underwriting standards, and internal audit.(See Fitch Ratings “The Impact of Poor Underwriting Practices andFraud in Subprime RMBS Performance,” p. 7. The special reportis available atwww.securitization.net/pdf/Fitch/FraudReport_28Nov07.pdf).

Sixth, fund managers, like brokers, weremotivated by bonuses and usually on a competi-tive basis relative to their peers. CDOs offered asimple means to enhance portfolio performance,which generated bigger bonuses and improvedthe performance of funds offered to the public.Greater leverage could be obtained through CDOsthat had embedded leverage in their structure, andthis offered better returns. Pricing of the least-liquid tranches could be based on mark-to-modelvaluations that depended on critical assumptionssuch as the correlation structure of the underlyingassets made by the managers themselves (seeBrunnermeier, 2008). As Chuck Prince, formerchief executive officer of Citigroup, commentedconcerning the incentives facing the investmentbanks: “as long as the music is playing, you’vegot to get up and dance. We’re still dancing.”(Nakamoto and Wighton, 2007). This statementabove all others suggests that fund managers andinvestment bank executives were fully aware thata bubble was inflating but until it burst there wasmoney to be made.

Some economists argue the incentives pre-sented a classic example of a principal-agentproblem in a world of asymmetric information,in which incentives to different parties were sub-stantially at variance with one another. Here weargue there is reason to believe that the incentivesof brokers, originators, SPVs, rating agencies, andfund managers were very much aligned. At everystage, profits could be made by providing assetswith characteristics that the buyer required, andproviding there was another buyer farther up thechain, the risk considerations were not paramount.Even end-investors were satisfied because theassets met the conditions in the “search for yield.”The regulators should have ensured originators,arrangers, and fund managers focused on theconflicts of interest more carefully, because thecomplexity and length of the chain between sellerand buyer meant poor-quality mortgages securi-ties encouraged the improper consideration ofthe risks, but this was not done.27

The incentive structure contributed to whatGiovannini and Spaventa (2008) call “the infor-mation gap” between the originator and theinvestor, but there was another issue: complexityin the assessment of risk.Provision of Information. In many respects,

the provision of information and the regulationsconcerning information lie at the root of the2007-08 credit crunch. The observed change inbanking practice toward originate and distributemodels has greatly altered the incentives facingthe originators of loans, and information aboutthe risks associated with the assets was lackingbut regulators and investors were slow to pickthis up. Not only does a lender who intends tosell the securitized loans face less incentive todiligently examine the quality of the borrower,or the collateral against which the loan is made,but there is an information asymmetry betweenthe seller of the securitized assets and buyerthat cannot easily be overcome by organizationssuch as the ratings agencies. Willem Buiter(2008a) has argued information may not havebeen collected at all, or if it was collected, itmay have been neglected during the process oftransferring assets from originator to buyer. Thisdiffers from a standard information asymmetrymodel where true information cannot be observedby the lender and must be taken on trust fromthe borrower or obtained by incurring a monitor-ing cost (e.g., the information asymmetry facinga bank and a customer, when only the customerknows the true value of an investment project).In this case of an investor-seller relationship,information that could be made known is notrevealed—not because the investor could notknow it or incurs a cost of obtaining it—butbecause the investor does not specifically requireit to be revealed by the seller. While there werecases of sellers fabricating or adjusting data onmortgage applications, in many more cases trueinformation on the financial condition of themortgagee was not passed up the chain becauseit was not required. Originators and arrangersprovided just enough information to satisfy theinvestor at the next stage of the process and nomore. This problem occurred at every link inthe chain as products were combined, splitinto tranches, and resold. Figure 9 shows that

27 Regulators were not sufficiently aware of the dangers offered byincentives set at the time, and some of the agents were outsidetheir jurisdiction in any case.

Mizen


“information gaps” exist at all points betweenthe seller and buyer.

In a speech to the European Parliament onJanuary 23, 2008, Jean-Claude Trichet, presidentof the ECB, commented that there were “lessonsto be drawn in terms of the structure of incentivesin all stages of the securitisation process and the‘originate to distribute’ model. All the relevantplayers—including originators of loans, arrangersof securitised products, rating agencies, conduitsand SIVs, and final investors—should have theright incentives to undertake a proper assessmentand monitoring of risks” (ECB, 2008).

A report of the U.S. President’sWorking Groupon Financial Markets (2008, italics in original)explains that the incentives and the informationgap are related:

Originators, underwriters, asset managers,credit rating agencies, and investors failed toobtain sufficient information or to conduct

comprehensive risk assessments on instru-ments that often were quite complex. Investorsrelied excessively on credit ratings, whichcontributed to their complacency about therisks they were assuming in pursuit of higherreturns. Although market participants hadeconomic incentives to conduct due diligenceand evaluate risk-adjusted returns, the stepsthey took were insufficient, resulting in a sig-nificant erosion of market discipline.

An important challenge for policymakers isto consider the options governing informationrequirements on originators and subsequent sell-ers of these highly engineered products. Alteringthe rules over the provision of information willgo a long way to making the products transparentand reducing the information gap. This in no waydiminishes the institutions’ own responsibilitiesto change the incentives offered to mortgageoriginators, agents, brokers, and fund managers.

Mizen


Figure 9

Information Asymmetries

SOURCE: Bank of England Financial Stability Report, October 2007.

Borrower

Broker

Originator

Arranger/Issuer

Trust/SPV

Asset Funds/SIVs, etc.

End-Investors

Credit RatingAgencies

Rating Agency FacesInformation Problems

Rating Dependencies

Rating Securities

InformationGaps

Complexity in the Assessment of Risk. Itseems surprising that investment banks accus-tomed to dealing with complex assets could beconvinced that AAA-rated assets could commandreturns that had such large spreads over risk-freeassets such as Treasuries without being inherentlymore risky. Perhaps Chuck Prince was right thatinvestment banks knew the risks but were pre-pared to continue to “dance” while money wasbeing made. For a less sophisticated class of end-investors, several factors made risk assessmentmore complex and difficult.

The Development of Structured FinanceProducts for Mortgages of Differing Quality.The process of combining these financial prod-ucts made evaluation of their riskiness extremelydifficult. The purchaser believed that develop-ment of structured finance allowed for diversifi-cation of risks and at every stage the benefits ofdiversification would reduce the risks comparedwith those on the underlying mortgages. But theembedded leverage in these products meant thatend-investors were often buying assets with muchgreater risk characteristics compared with theunderlying pool of mortgages, credit card debts,or loans than they might suppose. With highleverage ratios, a level of defaults that mightaffect a small proportion of an investor’s capitalcould quickly multiply to threaten to eliminateit all. Despite these dangers the returns on struc-tured finance products were good, and manyinvestors were persuaded that the risks werelow because the ratings were good.

Reliance on Ratings to Assess Asset Quality.Given the complexity of the products offered,investors relied on ratings provided by ratingsagencies such as Moody’s, Standard & Poor’s,and Fitch. These ratings indicate the likelihoodof default on the product, and for the highestratings—AAA—the likelihood was equivalent togovernment debt default for developed economies(i.e., negligible). The granting of AAA ratings toasset-backed securities meant many investorsbelieved they were buying very safe assets, andcertain organizations such as pension funds,which face restrictions on the assets they are per-mitted to purchase, were able to buy these assets.These risks were not properly priced becausethey did not anticipate the potential for lower

house prices or the potential effects house pricedeclines would have on subprime default rates.In addition, there is a widespread view that thecomplexity of the products offered created adependence on ratings agencies to evaluate therisk of these types of assets, without (much) fur-ther due diligence undertaken by the investor.There is then the question of the risks being rated.In their defense, ratings agencies argue that thepurchasers of their services requested defaultratings and not ratings of market or liquidity risk,partly because these were more expensive tocompute because of the increased work involved.Although the ratings agencies offered assessmentsof default risk, the ratings themselves were(mis)interpreted by some end-investors as indi-cators of all three types of risk.

The Belief that Tranching Reduced theRisk to the Senior Holders of Asset-BackedSecurities. Ratings agencies were able to providehigh ratings because they believed at the timethat residential MBSs and CDOs were financiallyengineered to reduce the risk of default. Modelsof the default risk suggested the top tranches werevery safe, but the models relied on a poolingprocess, wherein a large number of individuallyrisky loans were assumed to have a reducedrisk of default when combined into a package.Because the ratings agencies believed the seniortranches were very safe, CDOs in the seniortranche would be assigned AA or AAA ratings,mezzanine tranches would be assigned BBB rat-ings, and equity would be BBB to CCC or lower.Whether the risks in the senior tranches wereas low as the AAA ratings suggest is difficult togauge, but with the great benefit of hindsight, itappears unlikely. The loans were low quality,and were not as independent as the models ofthe risk characteristics had assumed. Delinquen-cies on the individual loans began to rise togetherwhen the housing market slowed; they weremuch riskier than ratings agencies or end-investors supposed.

Actions by the Central Banks andGovernment

Market Liquidity. Although opinions differedamong central banks on how to manage the crisis

Mizen


at first, the views have converged considerablysince September 2007. The schemes introducedby the Federal Reserve, the Bank of England,and the European Central Bank all widen therange of high-quality collateral the central bankwill accept and extend the lending term. Thesechanges merit further consideration.

First, the central banks have all made liquidityavailable overnight for 28 days, but terms of threemonths or longer also are available. This changewas designed to inject cash at longer maturities,but this also has an effect on shorter rates; so thatthe change in the composition of the liquidityoperations does not affect overnight rates, wherenecessary, central banks may have to absorb theexcess liquidity by withdrawing cash overnight.28

This process is somewhat reminiscent of OperationTwist, the action of the Federal Reserve under theKennedy administration in the 1960s when itoperated at various maturities to twist the yieldcurve (see Holland, 1967).29 The policy objectiveat the time was to raise short rates that needed tobe high relative to short rates of other countriesto deal with the balance of payments problemswhile lowering long rates that needed to be lowto encourage economic growth. The effectivenessof Operation Twist divides the academic com-munity, but then-Governor Bernanke discussedthe possibility of such an operation in the con-text of a speech on deflation in November 2002(Bernanke, 2002). As Chairman, he has come torely on it to deliver the term lending to financialinstitutions while still keeping the federal fundsrate at its target value.

Second, the TAF operations and similar activ-ities of the Bank of England and the ECB have notjust extended the term of the liquidity operationsthat central banks offer to the markets, they alsohave altered the collateral they accept. In thisrespect, the latest operations are different fromOperation Twist, and the move to accept a vari-ety of collateral that previously was not eligiblehas been critical for the present crisis. Markets forMBSs had dried up as banks were not prepared

to purchase the short-term assets issued by thepurchasers of MBSs and withheld liquidity tocover their own needs; therefore, borrowing overterms longer than overnight was restricted bythese developments. Central banks engaged in aswap of collateral—government-backed securitiesin exchange for riskier MBSs—with appropriateconditions to ensure markets had collateral witha market-determined value that could be used toobtain liquidity at the required maturities. AsBuiter (2008b) points out, the central banks havein effect become “market makers of the last resort.”Once the market had been made by the centralbanks to swap the private sector securities forgovernment securities, it was hoped the marketswould normalize. The fact that this has not beenthe case, as indicated by spreads between three-month LIBOR and the expected overnight ratesthat are still wider than usual, creates a puzzle.Why is there still a larger spread than in previousyears? The scale of the operations by central bankshas been vast, and it is unlikely that a shortageof funds is the reason for the spread. One answerto this puzzle is that the spreads were unusuallycompressed in recent years and have widenedbecause they were previously abnormally nar-row—many supervisory institutions warnedthat risk had been mispriced in the run up to thecrunch. A second response is that considerableuncertainty remains about the ability of financialinstitutions to obtain funding in the future, andthe injection of liquidity has eased the marketsbut not eliminated the uncertainty about the futurefunding. If the first answer is correct, then thecentral banks should not be concerned about thesustained spreads in the markets: There has beena correction for the true degree of credit risk. If thesecond answer is correct, the central banks shouldconsider further what can be done to reduce mar-ket uncertainty arising from liquidity risk.30

Third, the central banks collaborated to alle-viate the shortage of liquidity. When the need for

Mizen


28 Whether banks need to “mop up” liquidity depends on the size ofthe operation they intend to carry out.

29 I am grateful to Charles Goodhart for pointing out this connection.

30 McAndrews, Sarkar, andWang (2008) report research at the FederalReserve Bank of New York that seeks to determine the effectivenessof the TAF on the spreads inmoneymarkets by observing the spreadagainst announcements and operations of the TAF by the Fed. Theyconclude the TAF had a negative effect on spreads. The effective-ness of the control of the central bank on the liquidity risk premiumin money markets is a vital area of research.

liquidity was first identified in late 2007, a jointeffort by the central banks of all the major indus-trialized countries infused liquidity into themarkets. The Federal Reserve provided access todollars for the European banks via a currency-swap arrangement, with the ECB and the SwissNational Bank (SNB) acting as conduits, inDecember 2007 and January 2008 in amounts upto $20 billion and $4 billion to the ECB and theSNB, respectively, on both occasions, which effec-tively increased lending in another major currencythrough the currency-swap market. In May 2008,a further injection of liquidity provided dollarsin amounts of up to $50 billion and $12 billionto the ECB and the SNB, respectively. The movetoward common solutions to the liquidity prob-lem is part of an ongoing process that is likely tolead to further collaboration on types of eligiblecollateral and market operations.Funding Liquidity. The questions that need

to be answered are whether the authorities shouldhave provided funding liquidity and whetherthey should have provided it in this way. Thesetwo questions address concern about the conductof the authorities in the rescue of illiquid banks—first for Northern Rock in the United Kingdom(whether this is a bailout depends on what even-tually happens to the shareholders) and later forBear Stearns in the United States—and are ofconsiderable interest. All these rescues involveda considerable amount of public money. Theguarantee offered by the British governmentbacked the deposits in the Northern Rock notcovered by the deposit insurance scheme oper-ated by the Financial Services Authority. Thesubsequent decision to nationalize Northern Rockon February 18, 2008, involved £25 billion ofpublic money plus the state guarantees to thebank itself. In the United States, the rescue ofBear Stearns on March 17, 2008, involved a loanof $29 billion, and the credit line offered to thegovernment-sponsored mortgage agencies onJuly 14, 2008, involved $300 billion of U.S.Treasury support, and Congressional approval forfunding (the size of which is unknown at thispoint). The question is whether the authoritiesshould have offered funding liquidity and whetherit was done in a timely and efficient way.

At a much earlier stage, Mervyn King hadvoiced concern that central banks should notprovide liquidity too freely to institutions facingdifficulties to avoid moral hazard. In a letter tothe Chairman of the Treasury Select Committeeon September 12, 2007, he outlined his views(reported in the Fifth Report of the Committee,House of Commons Treasury Committee, 2008)as follows:

[T]he Governor pointed out that he did notagree with the suggestions for additional meas-ures that others believed the Bank of Englandshould undertake: lending at longer maturities,removing the penalty rate or increasing therange of collateral against which the Bankwould be prepared to lend. In the letter, hegave three reasons for his position. First, hestated that “the banking system as a whole isstrong enough to withstand the impact of takingonto the balance sheet the assets of conduitsand other vehicles.” Second, “the private sectorwill gradually re-establish valuations of mostasset backed securities, thus allowing liquidityin those markets to build up.” Third, therewould be a risk of “moral hazard.” In essence,this “moral hazard” argument is that, shouldthe central bank act, and effectively provideextra liquidity at different maturities againstweaker collateral, markets would, especiallyif the liquidity were provided at little or nopenalty, take it as a signal that the central bankwould always rescue them should they takeexcessive risk and get into difficulties.

In examining these arguments it is clear thatthe banking system as a whole could sustain thelosses incurred, but individual institutions, likeNorthern Rock and Bear Stearns, could not. Take-over by the private sector was the preferred optioneven in the United Kingdom, although it was notpossible to find a satisfactory resolution with aprivate buyer. King’s confidence in the marketsto reestablish valuations and liquidity in capitalmarkets now seems optimistic, although mostcommentators at the time would have expectedthe markets to settle. Despite actions to providemarket liquidity, they markets still have not“normalized,” and “normal” is difficult to define.The actions that Governor King sought to avoid—lending at longer maturities against a wider

Mizen


definition of collateral—have, in fact, beenimplemented in the Special Liquidity Scheme,but not without a penalty rate, and under swaparrangements that aim to minimize the risk thatthe Bank accepts on its balance sheet. It seemsthat the Bank’s reading of the crisis, at this stage,was later to be revised.

What then about King’s third point regardingmoral hazard? Moral hazard occurs when provi-sion of emergency funding for an institution introuble today encourages banks to take more risksin the future. King sought to avoid moral hazardby providing a plentiful supply of liquiditythrough existing schemes—not through specialarrangements requested by the banks that carriedno penalty rate of interest.31 If banks were to acton the knowledge that the central bank stoodready to rescue them, collectively or individually,in the event of another crisis, public moneywould insure activities of the banks and encour-age excessive risk-taking. This is the cost of liq-uidity provision and needs to be avoided, but abalance needs to be struck between making pro-vision for market and funding liquidity to dealwith a bank in crisis and withholding provisionto avoid future moral hazard.

Arguably, the crisis in Northern Rock occurredbecause all banks held far fewer liquid assets inrecent years than they did, say, 40 years ago (see,by way of comparison, Goodhart, 2008a). Effec-tively, banks had been allowed to insure some oftheir funding risk with central bank money forsome time. The gradual move toward fundingmodels using short-term paper entailed risks thatmight require the authorities to provide liquiditywhenmarkets were unable or unwilling to do so.32

The run on Northern Rock happened because ithad taken this process a step further than otherU.K. banks and gambled that it would not face a

funding problem on all short-termmoney marketssimultaneously. Although the U.K. authoritiesmight have wished to see the markets and thebanking system resolve the crisis on their own, inthe end they needed to support a private sectorfinancial institution with a loan and governmentguarantees. That they were willing to do so forNorthern Rock was explained by its share of theU.K. mortgage loan market and the limited scopeof the U.K. deposit insurance scheme.33 A signifi-cant number of depositors stood to face a drawn-out process before a fraction of their assets werereturned to them under deposit insurance rules,and this is surely something that needs to beaddressed to avoid bank runs in the future. Theposition of the Governor that moral hazard shouldbe avoided by refusing to lend freely in the eventof a crisis was a last-ditch attempt to manage theincentives facing banks, but the banks had beenallowed to develop funding models that trans-ferred funding liquidity risk to the central bankfor many years. Paul De Grauwe (2008) is right toargue that

[A] new equilibrium must be found in whichtighter regulation is reintroduced, aimed atreducing the propensities of too many in themarkets to take on excessive risks. The needto re-regulate financial markets is enhanced bythe fact that central banks, backed by govern-ments, provide an insurance against liquidityrisks. Such insurance inevitably leads to moralhazard and excessive risk-taking.

In the case of Bear Stearns, the Fed steppedin—despite the fact that Bear Stearns was not adepository institution—because the importanceof its role as a counterparty to international deriva-tives trades afforded it strategic importance inthe financial markets that made it too embeddedto fail. This was emphasized by Christopher Cox,chairman of the SEC, the U.S. regulator, andTimothy Geithner, president of the Federal ReserveBank of New York. The need to do so was alsothe result of risky management that adopted abusiness model too heavily reliant on short-termrollover funding frommarkets. Bear Stearns stood

Mizen


31 The ECB view was very different. It regarded the crisis as primarilya crisis of confidence, and therefore moral hazard considerationswere not a high priority. The ECB argument is about the appropri-ateness of market liquidity to restore confidence—not fundingliquidity to save a failing institution—therefore, this argumentaddresses a different issue.

32 Warnings in 2007 by the FSA and the Bank of England seem toindicate awareness of the former but not necessarily the hazards ofthe latter. Perhaps, like the banks themselves, they did not envisageall markets for short-term funds being closed simultaneously.

33 The deposit insurance scheme in the United Kingdom insuresonly the first £2,000 and 90 percent of the next £35,000.

to gain from the high returns that the businessmodels generated, but these returns also involvedlarge risks, and given the scale of the potentiallosses implied bankruptcy for the institutionsconcerned unless the government intervened.The risks taken by the managers of these institu-tions were much larger than the shareholders orthe investors would have accepted if they hadbeen aware of them. In this respect, evaluatingwhether the problem constituted moral hazardagain centers on whether Bearn Stearns know-ingly took more risks than would have been thecase if the losses had been borne entirely by theowners. It is difficult to believe that Bear Stearnsdid not know it was taking large risks to obtainhigh returns, but there may have been a failure toappreciate just how large would be the potentiallosses given default. The shareholders experi-enced losses when shares were sold at $10 pershare to JPMorgan Chase compared with valua-tions of $150 per share a year earlier.

The second question concerns the effective-ness of the response mechanism of the authoritiesin the United States and the United Kingdom.The decisions to defend vital elements of thebanking and financial system were made in realtime but on the basis of prearranged strategiesfor crisis resolution. This raises another question:Were the systems well structured to make thesedecisions when they needed to be made? In theUnited Kingdom the investigation into the runon the Rock by Parliament concluded that thetripartite arrangement in place in Britain did notresolve the bank run in a smooth fashion (Houseof Commons Treasury Committee, 2008, p. 107).The origins of the tripartite arrangement forresolving bank crises are found in the separationof monetary policy and financial stability respon-sibilities when the Bank of England was grantedindependence in May 1997. Financial regulationand supervision, which had been the Bank’sresponsibility, was separated and given to theFinancial Supervision Authority (FSA). Theresponsibilities in the case of a crisis were thensplit among the Treasury, the Bank, and the FSAas documented in a Memorandum of Understand-ing, which had been reviewed and revised as lateas March 2006 (House of Commons Treasury

Committee, p. 104). The weakness of the tripartitesystem stemmed from the difficulty of knowingwho was ultimately “in charge” when eventswere moving at a swift pace (House of CommonsTreasury Committee, pp. 109-10). The ability tomake the political decision to involve publicmoney in a rescue implied the Treasury had to beinvolved, but if there were to be a lender-of-last-resort operation, this would engage the Bank ofEngland in consultation with the banking super-visors at the FSA. Cecchetti (2008a) argues thatseparation of the liquidity provider from thesupervisor was bound to stress the system at atime of crisis, but in testimony to ParliamentMervyn King expressed no desire to take backthese responsibilities. Eventually all three insti-tutions were involved in the decisionmakingprocess concerning the rescue of Northern Rock.

Could the Bank of England have balanced itsresponsibilities between monetary policymakingand financial stability? The question is about con-flicts of interest between monetary policy andfinancial stability objectives. Cecchetti (2008a)argues that the liquidity provider should havesome supervisory responsibility and implies thata central bank is capable of trading off its respon-sibilities internally. At a time when the financialturmoil requires liquidity to be supplied to themarkets but the inflation outlook requires a tight-ening of monetary policy, the Bank of Englandhas had to innovate to provide term lending andhold rates to control inflation simultaneously,but it has done so quite successfully. Reforms tothe arrangements are inevitable, with a strength-ening of the financial stability role of the Bank ofEngland proposed. Buiter (2008b) discusses theseproposals in detail. The most important issue isto see that action is more effective by establishinga clear line of communication and control forfuture crises.

The Bear Stearns crisis resolution processappears to have delivered what the FederalReserve set out to achieve. The crisis was dealtwith swiftly, and as a result the financial systemdid not face the settlements equivalent to a “pay-ments problem.” The owners and fund managersof the investment bank were effectively punished

Mizen


for taking risky strategies, regardless of whetherthey were aware of their scale, and the creditorsof the institution were able to pass the debts ofthe company to its acquirer. The financial systemis now aware of the dangers of highly leveragedinvestments funded by issue of short-term paper.However, the cost to the Federal Reserve, whichprovided a $29 billion loan against collateral ofquestionable value, depends on the scale of thelosses Bear Stearns will incur and where the Fedstands in the line of creditors. The best-case sce-nario is that JPMorgan Chase will provide suffi-cient funding to cover all the losses and the loan.The worst case is that the Fed has accepted poor-quality collateral for a loan that will not be repaidand the Fed will be a long way down the list ofcreditors.

The key question, however, is why BearStearns had been allowed to take such risks, whileunder the oversight of the SEC, that then requiredthe Federal Reserve to step in when it faced diffi-culties that could have become systemic.34 BearStearns had met the SEC requirements untilMarch 10, 2008, but when it failed it did so quickly.The option to ask the private sector to rescue BearStearns by liquidating its positions without Fedsupport was considered not to be an option for tworeasons. First, it could not be arranged quicklyenough without Fed coordination and support,and second, the resilience of the markets for BearStearns’ assets was not great—a sell-off of assetswould have depressed prices and forced BearStearns into insolvency. This seems to have beena situation that the Fed and the SEC did not fore-see coming.

In the case of the GSEs Freddie Mac andFannie Mae, the line of credit from the U.S.Treasury and Congress is much larger. It is notyet clear whether the promise of a large sum willbe sufficient to restore confidence, and perhapsthe GSEs will not need it, but it may allow the

institutions to continue to operate under the samerules as before. This case seems to involve thelargest moral hazard. GSEs have always operatedunder an implicit government guarantee; this hasnow been made explicit. But the moral hazardexists because the incentive structure facing GSEexecutives and the business model they operatecontinues as before while the extent of the sup-port from the government is unlimited. It has tobe hoped that changes to the regulatory environ-ment—the Federal Housing and Economic ReformAct of 2008 provides a regulator for GSEs and theFederal Home Loan Banks—will offset these dan-gers. The regulator will be able to establish capitalstandards, prudential management standards,enforce its orders and remove officers, put theagencies into receivership, and review/ approveany new products that they may develop. The keyissue is how aggressively these powers are used.

Regulation, Supervision, and AccountingConventions

Amajor concern throughout the credit crunchhas been the role of supervisors and regulators inthe process—a rather obvious conclusion now—but an issue that still needs to be addressed. Wenote here particular areas where regulators’ atten-tion should be concentrated.Regulation of Originators and Brokers. The

first concern is the regulation of the mortgageoriginators and the subsequent producers ofstructured finance products. Although U.S. mort-gage banks are subject to regulation by federaland state agencies, Jaffee (2008) acknowledgesthat regulators’ benign practices exacerbated thecrisis. The process by which mortgages wereoriginated without much attention to borrowerquality is an issue that now seems fundamentallyimportant. The accommodating environmentprovided by the regulations is well documented(U.S. Treasury, 2008, Bernanke, 2007, and Angelland Rowley, 2006).

A second and equally fundamental issue isthe conduct of the originators: Did they act inthe borrowers’ best interests? Some have arguedthat lenders were in fact engaged in “predatorylending”: the selling of loans not in the best inter-

Mizen


34 Jaffee and Perlow (2008) indicate that the five largest investmentbanks, including Bear Stearns, submitted to voluntary supervisionby the SEC to satisfy European Union requirements for regulation.These “consolidated supervised entities” were required to maintaina 10 percent capital ratio, similar to the Fed’s standard for well-managed bank holding companies, and they were required to holdcash and securities of a high quality to cover all their liquidityrequirements.

ests of the borrower.35 Jaffee (2008) argues thatpredatory lending occurred not because consumerprotection legislation was lacking but because itwas not enforced. The existence (or non-existence)of predatory lending is controversial, and othercommentators have suggested that incentives wereinfluenced by the Community Reinvestment Act(CRA, 1977), which required lenders to offercredit, including home ownership opportunities,of their entire community and not just wealthysubsectors.36 HUD required Freddie Mac andFannie Mae to purchase “affordable” home loansecurities in the mid-1990s, and the purchasesof these securities increased again from 2004 to2006. HUD expected the agencies to imposehigher standards on lenders, but ironically,because Freddie and Fannie Mae bought subprimeMBSs, they provided additional incentives to theoriginators of the loans without having a directinfluence on their lending standards (Leonnig,2008).

Agents operating on behalf of financial insti-tutions sold mortgages without establishing thefinancial position of the borrowers, relying insteadon the appreciation of the housing asset to ensurerepayments could be met out of capital gains.The Federal Reserve moved in July 2008 to estab-lish rules to prevent mortgages being sold with-out verification of income, and financial assetsto ensure repayment is possible without relyingsolely on the appreciation in the value of thehouse purchase. Other practices labeled as“unscrupulous” include the imposition of pre-payment conditions that prevent a borrower fromrepaying the loan at a faster rate than scheduled,often on worse terms for subprime than for primeborrowers; these also will be regulated. The Fedhas announced a new rule to provide protectionsfor a newly defined category of “higher-pricedmortgage loans” (Federal Reserve Board, 2008).

These rules

• prohibit a lender from making a loan with-out regard to borrowers’ ability to repay theloan from income and assets other thanthe home’s value;

• require creditors to verify the income andassets they rely upon to determine repay-ment ability;

• ban any prepayment penalty if the paymentcan change in the initial four years; forother higher-priced loans, a prepaymentpenalty period cannot last for more thantwo years;

• require creditors to establish escrowaccounts for property taxes and home-owner’s insurance for all first-lien mortgageloans.

The regulation of the U.S. mortgage market isset to improve with the new rules, but setting of“gold standards” for originators to match products(e.g., alternative mortgages) offered by the GSEsor minimum borrower standards would also help.If sellers were required to offer the alternativeand see that minimum standards were met, itwould help ensure that the selling of mortgagesdoes not revert to previous bad practice. Similarregulations should be created for non-U.S. banksoperating in other places to prevent the problemsspreading to other countries.Regulation of Off-Balance-Sheet Vehicles and

Banks’ Obligations to Them. A further questionfor regulators is the extent to which banks shouldbe allowed to avoid regulation by using off-balance-sheet vehicles to conduct business instructured finance products. In the United States,the Financial and Accounting Standards Boardis reconsidering FASB Statement 140, whichallows banks to transfer assets and liabilities toSPVs. The difficulty here is that under Basel Irules for capital adequacy requirements, banksare required to hold 8 percent of their capitalagainst loans, while off-balance-sheet vehiclesof banks—the SIVs and conduits—are not. Thisarrangement offers banks a clear incentive tominimize the capital requirements by creatingoff-balance-sheet vehicles to hold assets and make

Mizen


35 A New York Times exposé of Countrywide Financial revealed thatagents were offered incentives in the form of fees based on profits,not on the best interests of the borrowers.

36 Mortgages lenders are not eligible for credit from the CRA, butbanks that purchase loans made by mortgage brokers are eligibleif the loans are extended to underrepresented communities.

loans because the regulatory hurdle is lower foroff-balance-sheet vehicles. Under Basel II rulesthis anomaly will be removed; banks and theiroff-balance-sheet entities will be treated in muchthe same way, removing the incentive for banksto arbitrage the capital requirements. It is unfor-tunate that these rules were not in place inEurope until January 1, 2007; by then the SPVs,SIVs, and conduits had created or bought largepools of securitized mortgage products.37 Thedanger with the separation of on-balance-sheetactivities from those of vehicles that are off-balance sheet is that it creates a false picture ofbank stability. When these off-balance-sheetinstitutions need funds, they turn to banks forliquidity. Requiring banks to reveal the extentof their liquidity commitments to off-balance-sheet vehicles—and the scale of their activitiesshould these entities need to be brought backonto the balance sheet—would resolve the prob-lem. The banking system as a whole was strongenough to take these entities onto its balancesheet in 2007-08, but the effect on the demandfor liquidity seriously affected the operation ofthe money markets.

The Financial Stability Forum of the Bankfor International Settlements (a committee ofglobal regulators and supervisors) has proposedin a report “Enhancing Market and InstitutionalResilience” (Bank for International Settlements,2008) that rules should be changed to make theholding of asset-backed securities and CDOs morecostly for banks. This will be accomplished bythe following means: (i) raising capital require-ments, under the Basel II capital adequacy rules,for complex structured finance vehicles; (ii) intro-ducing additional requirements for warehousedassets on banks’ balance sheets awaiting sale; and(iii) strengthening the capital requirements forliquidity buffers offered to conduits by banks.This step is a welcome development, but asWyplosz (2008) has pointed out, intrinsicallywhere risk-taking is concerned, regulation canhelp squeeze risk out of a segment of the market,but it typically reappears elsewhere. When banks

are regulated, non-bank vehicles emerge to assumethe risk in an unregulated environment, and ifregulation is imposed on them, new means willbe discovered to avoid the regulations. Financialmarkets have strong incentives to innovate, so theregulators need to invest more effort into aware-ness of the areas in which risk is being taken inpursuit of high returns to keep in step with thefinancial institutions they are regulating. Thisshould be done without stifling the financial inter-mediation process altogether. One way that regu-lators can offer incentives to the markets is torequire them to hold the riskiest segments (theequity tranches) of their structured financeproducts on their own books (Buiter, 2008a,de la Dehesa, 2008). Regulators need to evaluatethe bigger picture at a level beyond the financialinstitution, because it is the externalities of exces-sive risk-taking that matter. Regulators need toask questions about an institution’s own assess-ment of the risk being carried, but they also needto consider the systemic risks that arise when theactions of an individual bank impinge on otherbanks or the markets.38

Regulation of Ratings Agencies. A third con-cern is the regulation of rating agencies. Ratingsagencies have been subject to a great deal of criti-cism because their primary purpose is to evaluatethe risks of the products or entities that they rate.They seem to have done badly in rating struc-tured finance products, and the agencies them-selves are reviewing their processes. A majorworry is the potential conflict of interest theyface because rating agencies are well rewardedfor rating structured finance products. Buiter(2008a) and Portes (2008) have suggested thatChinese walls within organizations are notenough to prevent these conflicts of interest;they argue that ratings agencies should be single-product firms selling one thing: ratings. Incen-tives are only one reason why ratings agencieshave not done a good job, and why they mayneed to be regulated. Another side of the story

Mizen


38 A narrow definition of the regulators’ range is to protect the tax-payer from excessive risk-taking by financial institutions. A broaderdefinition covers protection of the financial system, includingpayments, settlements, and even the reputation of the financialindustry (when it constitutes a major economic sector).

37 Basel II is yet to be implemented in the United States; thereforeU.S. banks are still able to arbitrage the regulations on capitalwhile their European counterparts cannot.

is the extent to which the ratings agencies hadmodels appropriate for rating structured financeproducts. The modeling exercises involved areformidable but even taking this into account,Giovanni and Spaventa (2008) note that themodels for structured finance products werecalibrated using short spans of data over a benignperiod of moderation in financial markets andrising house prices.39 They simply had not expe-rienced turbulence or falling house prices toevaluate whether the models might prove unre-liable. These problems were compounded whenrating CDOs and CDOs-squared because theseproducts were given too much benefit for com-bining lower tranche residential MBSs, when infact the default risks were more highly correlatedthan the models assumed and were prone to acommon shock—a fall in house prices. Funda-mentally, is a rating metric suitable for sovereignbonds, investment, and sub–investment-grade cor-porate bonds, or project finance also suitable forstructured financial products? The InternationalOrganization of Securities Commissions (IOSCO)has suggested that agencies should introducenew ratings for mortgage-backed or structuredfinance products because of the perception thatthey behave differently than other financialinstruments in times of stress.

Whether regulation should be extended toratings agencies is not a new topic of debate, butthe recent experience will mean it has a new leaseof life. With its influence through the presidencyof the European Union in 2008, France has encour-aged the European Commission to propose thatratings agencies be registered and subject togreater regulation if they wish to operate in Europe.This follows the recommendation of IOSCO andthe Financial Stability Forum. Ratings agencieshave long argued that they publish their opinions,underpinned by their reputations, and the use towhich they are put is not something for which

they are answerable. In the United States the SECconfers a “nationally recognized statistical ratingagency” status on certain qualifying ratings agen-cies, allowing their ratings to be used for regula-tory purposes by others, but it stops short ofregulating their methodologies. This deters entryinto the ratings business because ratings havevalue to the purchaser when they can be used toreduce capital; conferring this status on someagencies and not others creates barriers to entry.Similarly, Basel II makes provision for credit rat-ings agencies to be used to evaluate bank capital,and the same argument applies. Whether regula-tion should be extended to ratings agencies is aquestion that needs to be addressed, since atpresent there is no regulation of their procedures.The Financial Stability Forum, through the IOSCO,offers a code of conduct fundamentals for creditrating agencies that it recommends but does notrequire agencies to adopt.Regulation and Stress Testing. A fourth issue

concerns the evaluations of risk by banks them-selves. A number of early warnings had signaleddifficulties ahead for financial institutions undercertain risk scenarios; for example, in Londonthe supervisory agency of the United Kingdom,the Financial Services Authority, had been con-cerned for some time about complexity and liq-uidity of financial markets, stating in January2007 that “Financial markets have becomeincreasingly complex since the last financialstability crisis, which implies that transmissionmechanisms for shocks have also become morecomplicated and possibly more rapid…It is stillimportant for market participants to considerhow they would operate in an environmentwhere liquidity is restricted (Financial ServicesAuthority, 2007). The Bank of England was evenmore direct in its Financial Stability Report,stating “Financial institutions can become moredependent on sustained market liquidity both toallow them to distribute the risks they originateor securitise and to allow them to adjust theirportfolio and hedges in the face of movementsin market prices. If it becomes impossible orexpensive to find counterparties, financial insti-tutions could be left holding unplanned creditrisk exposures in their ‘warehouses’ awaiting

Mizen


39 In simple terms the ratings agencies assessed the probabilities ofdefault for individual mortgages, considered the correlationsbetween individual loans, used these to assess the probability ofdefault for the securitized products, and then rated the tranchesaccordingly. Where they failed to calculate defaults correctly wasin assessing the proper weight to be attached to falling house priceson the defaults of individual loans, the interdependence betweenloan defaults, and the likelihood of falling house prices occurring.

distribution or find it difficult to close out posi-tions” (Bank of England, 2007). In other coun-tries, similar cautions were sounded but despitethese early signals, most investment and com-mercial banks regarded themselves as adequatelyprotected against likely shocks based on theirown stress testing evaluations.40 Banks did notappear to heed these warnings because theythought they were sound.

The use of stress testing assumes new signifi-cance under Basel II terms. If banks can demon-strate that they are robust to a battery of shocksthat might conceivably happen, then they canreduce the capital they are required to hold. Aswith ratings agencies, however, the modelingprocess should be subject to investigation.Northern Rock in Britain had received approvalfor a Basel II waiver on the basis of its internalstress testing processes on July 27, 2007, just sixweeks before the bank run. This did not stop thatbank’s shortage of funds as markets seized up.The nature of the models used and the range ofstress tests that they must meet need carefulscrutiny. Mark-to-model methodologies do notnecessarily correspond well with reality, and theweakness of the underlying assumptions of themodels becomes apparent when a crisis occurs.It is likely that stress tests conducted by banks todetermine their resilience to liquidity shortageswill need to be respecified to account for longer-lasting liquidity crises of the type experiencedduring 2007-08. The difficulty here lies in prepar-ing for the next crisis, not the last one. By defini-tion the nature of any future shock (or combinationof shocks) likely to trouble the financial systemis difficult to predict, and should be part of anongoing research agenda. When knowledge of thetypes of shocks is difficult to determine, requiringmore capital to be held as a buffer is the obvioussolution.

Fair Value Accounting. This raises a furtherissue: How should institutions determine thecapital to be held against assets? This is a ques-tion of accounting as much as regulation. Since1998 the Financial Accounting Standards Board(FASB) in the U.S. Statement of FinancialAccounting Standard (SFAS) No. 133, hasrequired fair value accounting for derivatives,and European institutions followed suit since2005. There is a general vision to have all finan-cial instruments accounted for at fair values, andwhile it has the advantage of presenting currentvaluations on assets and liabilities of banks ratherthan historic cost valuations, it also has somenegative implications. The main concern is thatas asset prices decline, fair value accounting booklosses associated with illiquid assets are imme-diately revealed and banks are then required toreduce leverage in order to meet capital ratiosunder BIS rules. The banks may not intend tosell the assets but their low current valuation—and the more illiquid the asset the more difficultit is to determine the accuracy of current valua-tions—may enforce it. Forced sales can driveprices down creating a vicious circle. In thepresent conditions current value of residentialMBSs and CDOs may not be fair value at all.Some of these points have been addressed bythe FASB thorough its three levels approach inreporting valuations. Level 1 has a market valuefrom market inputs, Level 2 has some marketinputs, and level 3 has none. These do not elimi-nate subjectivity of fair value prices but they doreveal where assumptions affect asset valuations.Another approach has been to reflect the inten-tion of the asset holder: Hence, an asset holdercan report financial assets at historic cost if theyintend to hold them to maturity, but report themat fair value if they are either “available for sale,”in which case any variation in fair value bypassesthe income statement and is applied directly tothe firm’s equity, or “due to be traded,” in whichcase the variation is included in an income state-ment. Many institutions use an intents model butinvestment banks are an exception. Investmentbanks argue that opponents of fair accountingcannot have it both ways. Fair accounting cannotbe opposed during financial crises but adopted

Mizen


40 It is interesting to note that opinions differed between banks:Goldman’s bank-wide risk committee reportedly forced the saleof most mortgage derivatives, believing them to be too risky, andBarclays also sold a fair share of these assets. Citi, UBS, andMerrillLynch retained large holdings on the balance sheets, and thesehave been the institutions most affected by write downs and creditlosses. Bear Stearns shed the assets into subsidiaries—hedge fundswhich were off balance sheet—and marketed the shares in thesefunds aggressively.

at other times. There is a need for consistency.They also argue that when the current marketvaluation is low there will be a buyer at thatprice seeking value, and should the bank choosenot to sell, it stands to gain if assets subsequentlyappreciate. The problem for commercial banksis that they must maintain capital ratios to com-ply with regulatory requirements, and thereforeat times have little choice whether to hold orsell. This is particularly acute when many assetsmarkets experience falling current valuationssimultaneously.

It has been suggested that the lack of coordi-nation between regulators has been detrimentalto effective regulation in all these areas. In Europein particular the different stance taken by nationalregulators did not ensure that financial institu-tions were well regulated. Similar problems hademerged in regulation in the United States, wherea combination of federal and state-level regulationdid not provide a consistent response to changingpractices of financial institutions. Buiter (2008a)calls for a Europe-wide regulator to avoid theintercountry differences in the approach towardregulation of structured finance. Internationalefforts to coordinate regulatory practice are boundto be helpful, coordinated by the FinancialStability Forum of the Bank for InternationalSettlements. Goodhart (2008b) also endorsesco-ordinated action within countries and acrossinternational borders and covers topics that havenot been discussed in detail here. These includedeposit insurance schemes; bank insolvencyregimes often also referred to as “prompt correc-tive action”; and the inherent procyclicality ofcapital adequacy requirements under Basel IIwith the difficulties it creates when bank crisesoccur. Clearly, many changes will need to beimplemented in regulation and supervision inlight of the 2007-08 credit crunch.

CONCLUSIONThis paper has argued that a number of factors

provided conducive conditions for a credit crunch.First, there was a period of exceptional stabilitywith very low long-term interest rates supportedby the global savings glut flowing from emerging

industrialized economies. Second, financialinnovation had developed well-understood finan-cial products such as MBSs and introducedgreater complexity, higher leverage, and weakerunderlying assets based on subprime mortgages.Third, no one anticipated that house prices wouldfall nationwide in the United States—these con-ditions were not built into the models used toassess risk—but house prices did fall and whenthey did so defaults increased in the subprimesector, which proved a trigger for the crisis asinvestors reappraised the risks associated withthe high-yielding residential MBSs and CDOscomposed of these assets. Any number of otherhigh-yield asset markets might have provided thetrigger for the 2007-08 credit crunch, includinghedge funds, private equity, and emerging marketequities; it just happened to be the subprime crisisthat occurred first. The failure of a number ofbanks then spurred a reaction in the markets forshort-term paper and banks of all kinds withdrewfrom lending in money markets. The authoritiesdecided to act to provide liquidity to the marketsand funding liquidity for failing banks such asNorthern Rock, Bear Stearns, and government-sponsored enterprises. Central banks handledthe crisis well from the perspective of providingliquidity to the markets, but spreads remain largerthan before the crunch. They did less well in pro-viding funding liquidity for failing institutions andthe consequences of these actions for the taxpayerthat are unquantifiable at this stage. Finally, reg-ulation and supervision needs to be enhanced inthe face of rapid financial innovation—the scopeof regulation will need to increase to ensure sys-temic risks are minimized in the future.

REFERENCESAdrian, Tobias and Shin, Hyun Song. “Liquidity,Monetary Policy, and Financial Cycles.” FederalReserve Bank of New York Current Issues inEconomics and Finance, January/February 2008,14(1), p. 2.

Anderson, Richard. “What Were They Thinking?”Federal Reserve Bank of St. Louis Central Banker,Winter 2007, p. 5.

Mizen


http://www.newyorkfed.org/research/current_issues/ci14-1.html


http://www.stlouisfed.org/publications/cb/2007/d/pages/economists.html

Angell, Cynthia, and Clare Rowley. “Breaking NewGround in U.S. Mortgage Lending.” FDIC Outlook,Summer 2006.

Armantier, Olivier; Krieger, Sandra and McAndrews,James. “The Federal Reserve’s Term AuctionFacility.” Federal Reserve Bank of New YorkCurrent Issues in Economics and Finance, July2008, 14(5).

Bank for International Settlements. “Report of theFinancial Stability Forum on Enhancing Marketand Institutional Resilience” April 7, 2008;http://www.fsforum.org/publications/r_0804.pdf.

Bank of England. Financial Stability Report, April2007, p 7.

Bank of England. Special Liquidity Scheme, April 21,2007; www.bankofengland.co.uk/markets/ sls/sls-information.pdf.

Bernanke, Benjamin S. “Deflation: Making Sure ‘It’Doesn’t Happen Here.” Remarks before the NationalEconomists Club, Washington, DC, November 21,2002;www.federalreserve.gov/BOARDDOCS/SPEECHES/2002/20021121/default.htm.

Bernanke, Benjamin S. “The Global Saving Glut andthe U.S. Current Account Deficit,” Homer JonesLecture, St. Louis, Missouri, April 14, 2005;www.federalreserve.gov/boarddocs/speeches/2005/20050414/default.htm.

Bernanke, Benjamin S. and Lown, Cara S. “TheCredit Crunch.” Brookings Papers on EconomicActivity, 1991, (2), pp. 205-47.

Board of Governors of the Federal Reserve System.Press release July 14, 2008; www.federalreserve.gov/newsevents/press/bcreg/20080714a.htm.

Boeri, Tito and Guiso, Luigi. “The Subprime Crisis:Greenspan’s Legacy, in Andrew Felton and CarmenReinhart, eds., The First Global Financial Crisis ofthe the 21st Century, 2008; www.voxeu.org/index.php?q=node/1352.

Brunnermeier, Markus K. “Deciphering the 2007-08Liquidity and Credit Crunch.” Journal of EconomicPerspectives, 2008 (forthcoming).

Buiter, Willem. “Lessons from the North AtlanticFinancial Crisis,” in Andrew Felton and CarmenReinhart, eds., The First Global Financial Crisis ofthe the 21st Century, 2008a; www.voxeu.org/index.php?q=node/830.

Buiter, Willem. “Lessons from Northern Rock: Howto Handle Failure,” in Andrew Felton and CarmenReinhart, eds., The First Global Financial Crisis ofthe 21st Century, 2008b; www.voxeu.org/index.php?q=node/964.

Calomiris, Charles W. “Not (Yet) a ‘Minsky Moment,’”in Andrew Felton and Carmen Reinhart, eds., TheFirst Global Financial Crisis of the 21st Century,2008, www.voxeu.org/index.php?q=node/739.

Cecchetti, Stephen G. “Subprime Series: Part 3: WhyCentral Banks Should Be Financial Supervisors,”in Andrew Felton and Carmen Reinhart, eds., TheFirst Global Financial Crisis of the 21st Century,2008a; www.voxeu.org/index.php?q=node/755available at voxeu.org.

Cecchetti, Stephen G. “Subprime Series: Part 4: DoesWell-Designed Monetary Policy Encourage Risk-Taking?” in Andrew Felton and Carmen Reinhart,eds., The First Global Financial Crisis of the 21stCentury, 2008b;www.voxeu.org/index.php?q=node/758.

Cecchetti, Stephen G. “Federal Reserve PolicyResponses to the Crisis of 2007–8: A Summary,” inAndrew Felton and Carmen Reinhart, eds., TheFirst Global Financial Crisis of the 21st Century,2008c; www.voxeu.org/index.php?q=node/1048.

Chomsisengphet, Suphela and Pennington-Cross,Anthony. “The Evolution of the Subprime MortgageMarket.” Federal Reserve Bank of St. Louis Review,January/February 2006, 88(1), pp. 31-56.

Clair, Robert T. and Tucker, Paula. “Six Causes ofthe Credit Crunch (Or, Why Is It So Hard to Get aLoan?),” Federal Reserve Bank of Dallas EconomicReview, Third Quarter 1993, pp. 1-19;www.dallasfed.org/research/er/1993/er9303a.pdf.

Mizen




http://research.stlouisfed.org/publications/review/06/01/ChomPennCross.pdf

http://research.stlouisfed.org/publications/review/06/01/ChomPennCross.pdf

De Grauwe, Paul. “Cherished Myths Fall Victim toEconomic Reality. Financial Times, August 23,2008; www.ft.com/cms/s/0/b89eb5b2-5804-11dd-b02f-000077b07658.html.

de la Dehesa, Guillermo. “How to Avoid FurtherCredit and Liquidity Confidence Crises,” in AndrewFelton and Carmen Reinhart, eds., The First GlobalFinancial Crisis of the 21st Century, 2008;www.voxeu.org/index.php?q=node/657.

Dell’Ariccia, Giovanni; Igan, Deniz and Luc Laeven.“Credit Booms and Lending Standards: Evidencefrom the Subprime Mortgage Market” CEPRDiscussion Paper No. 6683, Centre for EconomicPolicy Research, 2008.

European Central Bank. “Enhancing the EUArrangements for Financial Stability,” Centre forEconomic Policy Research, Keynote address byJean-Claude Trichet at the session “EuropeanFinancial Supervision, Crisis Management onFinancial Markets,” European Parliament, Brussels,January 23, 2008; www.ecb.eu/press/key/date/2008/html/sp080123.en.html.

Financial Services Authority. Financial Risk Outlook2007, London, FSA, January 2007, p 29.

Giovannini, Alberto and Spaventa, Luigi. “Filling theInformation Gap,” in Andrew Felton and CarmenReinhart, eds., The First Global Financial Crisis ofthe 21st Century, 2008,www.voxeu.org/index.php?q=node/690.

Goodhart, Charles A. E. “The Background to the2007 Financial Crisis.” International Economicsand Economic Policy, 2008a, 4(4), pp. 331-46.

Goodhart, Charles A.E. “The Regulatory Response tothe Crisis.” FMG Special Papers No. 177, 2008b.

Gorton, Gary B. and Souleles, Nick. “Special PurposeVehicles and Securitization.” Working Paper No.05-21, Federal Reserve Bank of Philadelphia,September 2005.

Holland, Thomas E. “‘Operation Twist’ and theMovement of Interest Rates and Related EconomicTime Series.” International Economic Review,October 1969, 10(3), pp. 260-65.

House of Commons Treasury Committee. The run onthe Rock. Fifth report of Session 2007-08; London:The Stationery Office Limited, 2008a;www.publications.parliament.uk/pa/cm200708/cmselect/cmtreasy/56/5602.htm.

House of Commons Treasury Committee. The run onthe Rock. Fifth report of Session 2007-08, Volume I.Report, together with formal minutes (HC 56-I).London: The Stationery Office Limited, January 24,2008b; www.publications.parliament.uk/pa/cm200708/cmselect/cmtreasy/56/56i.pdf.

House of Commons Treasury Committee. The run onthe Rock. Fifth report of Session 2007-08, VolumeII. Oral and written evidence (HC 56-II). London:The Stationery Office Limited, January 24, 2008c;www.parliament.the-stationery-office.co.uk/pa/cm200708/cmselect/cmtreasy/56/56ii.pdf.

International Money Fund. World Economic Outlook,September 2006.

Jaffee, Dwight, W. “The U.S. Subprime MortgageCrisis: Issues Raised and Lessons Learned.”Working Paper No. 28, Commission on Growthand Development, 2008.

Jaffee, Dwight, W. and Perlow, Mark. “InvestmentBank Regulation After the Bear Stearns Rescue.”Central Banking, 2008, 18, pp. 38-44.

King, Mervyn. Speech by the Governor of the Bankof England at the Northern Ireland Chamber ofCommerce and Industry, Belfast, October 2007,pp. 2-3; www.bankofengland.co.uk/publications/speeches/2007/speech324.pdf.

Leonnig, Carol D. “How HUD Mortgage Policy Fedthe Crisis—Subprime Loans Labeled ‘Affordable,’”Washington Post, June 10, 2008, p. A01;http://www.washingtonpost.com/wp-dyn/content/article/2008/06/09/AR2008060902626.html.

McAndrews, James; Sarkar, Asani and Wang, Zhenyu.“The Effect of the Term Auction Facility on theLondon Inter-Bank Offered Rate.” Staff Report No.335, Federal Reserve Bank of New York, 2008.

Mizen


http://www.newyorkfed.org/research/staff_reports/sr335.html

http://www.newyorkfed.org/research/staff_reports/sr335.html

Milne, Alistair and Wood, Geoffrey. “Banking CrisisSolutions Old and New.” Federal Reserve Bank ofSt. Louis Review, September/October 2008, 90(5),pp. 517-30.

Monacelli, Tommaso. “An Extensive but BenignCrisis?” in Andrew Felton and Carmen Reinhart,eds., The First Global Financial Crisis of the 21stCentury, 2008; economistsview.typepad.com/economistsview/2007/08/tommaso-monacel.html.

Nakamoto, Michiyo, and Wighton, David. “CitigroupChief Stays Bullish on Buy-outs.” Financial Times,July 9, 2007.

Organisation for Economic Co-operation andDevelopment. Financial Market Trends, November2007, 93(2), p. 21.

Organisation for Economic Co-operation andDevelopment. Financial Market Trends, May 2008,94(1), p. 13.

Portes, Richard. “Ratings Agency Reform,” inAndrew Felton and Carmen Reinhart, eds., TheFirst Global Financial Crisis of the 21st Century,2008; www.voxeu.org/index.php?q=node/887.

President’s Working Group on Financial Markets.“Policy Statement on Financial Market

Developments,” March 13, 2008, Washington, DC;http://online.wsj.com/documents/pwgpolicystatement20080313.pdf.

Sengupta, Rajdeep and Emmons, William R. “WhatIs Subprime Lending?” Federal Reserve Bank ofSt. Louis Monetary Trends, June 2007;http://research.stlouisfed.org/publications/mt/20070601/cover.pdf.

U.S. Department of the Treasury. “Blueprint for aModernized Financial Regulatory Structure.”Washington, DC: Office of Domestic Finance,March 2008; www.ustreas.gov/offices/domestic-finance/regulatory-blueprint/.

Wolf, Martin. “Why the Credit Squeeze Is theTurning Point for the World.” Financial Times,December 11, 2007; www.ft.com/cms/s/0/90126fca-a810-11dc-9485-0000779fd2ac.html.

Wyplosz, Charles. “Subprime ‘Crisis’: Observationson the Emerging Debate,” in Andrew Felton andCarmen Reinhart, eds., The First Global FinancialCrisis of the 21st Century, 2008;www.voxeu.org/index.php?q=node/471.

Mizen


http://research.stlouisfed.org/publications/review/08/09/Milne.pdf

http://research.stlouisfed.org/publications/review/08/09/Milne.pdf


Documents

TheCreditCrunchof2007-2008: ADiscussionoftheBackground, … · 2019-10-15 · FEDERAL RESERVE BANK OF ST. LOUIS REVIEW SEPTEMBER/OCTOBER 2008 531 TheCreditCrunchof2007-2008: ADiscussionoftheBackground,