a joint initiative of ludwig-maximilians-universität and ... · tim to the crisis, unscrupulously...

Forum W I N T E R

2008

Spotlight

Trends

Focus

FINANCIAL CRISIS Hans-Werner SinnBarry EichengreenMartin HellwigPatrick HonohanMarek DabrowskiMardi DungeyDavid G. MayesNicolas VéronFriedrich L. SellJohannes MayrChang Woon Nam

HOW TO DEFINE A RECESSION?

STATISTICS UPDATE

A joint initiative of Ludwig-Maximilians-Universität and the Ifo Institute for Economic Research

VOLUME 9, NO. 4

Klaus Abberger andWolfgang Nierhaus

CESifo Forum ISSN 1615-245XA quarterly journal on European economic issuesPublisher and distributor: Ifo Institute for Economic Research e.V.Poschingerstr. 5, D-81679 Munich, GermanyTelephone ++49 89 9224-0, Telefax ++49 89 9224-1461, e-mail [email protected] subscription rate: n50.00Editor: Chang Woon Nam, e-mail [email protected] in EconLitReproduction permitted only if source is stated and copy is sent to the Ifo Institute

www.cesifo.de

Volume 9, Number 4 Winter 2008_____________________________________________________________________________________

FINANCIAL CRISIS

The End of the Wheeling and DealingHans-Werner Sinn 3

Origins and Responses to the Current CrisisBarry Eichengreen 6

The Causes of the Financial CrisisMartin Hellwig 12

Containment and Resolution in the Financial Crisis: Too Little, Too LatePatrick Honohan 22

The Global Financial Crisis: Causes, Anti-Crisis Policies and First LessonsMarek Dabrowski 28

The Tsunami: Measures of Contagion in the 2007–2008 Credit CrunchMardi Dungey 33

Avoiding the Next CrisisDavid G. Mayes 44

Europe’s Banking Challenge: Reregulation without RefragmentationNicolas Véron 51

We Need More OversightFriedrich L. Sell 60

The Financial Crisis in Japan – Are There Similarities to the Current Situation?Johannes Mayr 64

What Happend to Korea Ten Years Ago?Chang Woon Nam 69

How to Define a Recession?Klaus Abberger and Wolfgang Nierhaus 74

Statistics Update 77

Focus

Spotlight

Trends

Forum

CESifo Forum 4/20083

Focus

THE END OF THE WHEELING

AND DEALING

HANS-WERNER SINN*

Now that the countries of the west have agreed to athree-trillion dollar bailout programme to rescuetheir banking systems, it is time to look forward andto draw lessons from the crisis. To do this we mustunderstand the causes of the crisis.The claims that themodel of American capitalism has self-destructed arejust as misguided as putting the blame on the greed ofinvestment bankers and other groups in society. Theyonly touch the surface of the problem.

The core of the crisis lies in the legal provisions oflimited liability: creditors of corporations have noclaims against the personal assets of the owners(shareholders) of these corporations. These liabilityconstraints lead to a systematic disregard of disasterrisks – occurrences with only a slight probabilitybring about gigantic losses. Investors that opt forhigh-risk projects with high potential gains and loss-es instead of safe projects with similar average prof-its can expect to gain, since they only have to bear aportion of the possible losses. If things go well,investors reap the full profit. If things go badly, atworst their losses would be limited to the stock ofequity invested, because claims against private assetshave been ruled out. This asymmetric situationencourages bold behaviour and risk-taking.

The conclusion that the limited-liability constraintshould be eliminated would be too rash, however,because risk-taking also has its merits. Limited liabil-ity was introduced in the nineteenth century in theUnited States and Europe in order to avoid uncon-trollable burdens being placed on equity holders andto enable entrepreneurs to make enterprising eco-nomic decisions that they otherwise would not havehad the courage to make. It brought about the pro-ductive forces that have created the wealth oftoday’s generations.

In times of great economic insecurity, however, lim-

iting liability can become a problem because it in-

duces entrepreneurs to become gamblers. As always

it is a matter of weighing up the advantages and dis-

advantages and finding the proper middle ground.

The problem of gambling is particularly serious

when corporations are allowed to determine the

extent of their liability themselves by choosing the

ratio of equity to business volume as they see fit.

Then they tend to operate with too little equity and

distribute to their shareholders too large as fraction

of their profits as dividends.The five large US invest-

ment banks, of which three have already fallen vic-

tim to the crisis, unscrupulously pursued this strate-

gy, their motto being that you can’t lose what you

don’t have. The risks created incentives to minimize

the stock of equity kept inside the firms, and the

small amount of equity capital in turn created incen-

tives to pursue overly risky operations. The interplay

of these incentives is the actual cause of the crisis –

and this is where reform must begin.

The privilege of limited liability is not a creation of

the market; it was granted by the legislator, and

because this is the case, the legislator himself must

define his real intention. He cannot allow the bene-

ficiaries themselves to make this definition. If they

can, they will define the limitation in such a way that

they assume almost no liability, as we have seen. US

investment banks, which were not subject to

American bank supervision, practised their business

with equity-asset ratios in the region of four percent,

which is much lower than the rate at which private

commercial banks operate. In addition, they carried

out very complex credit operations outside of their

balance sheets, placing them thus away from investor

control.

Some may point out in defence that gambling is pre-

vented by the rating agencies. They argue that rating

agencies give poor ratings when risks are excessive,

forcing the banks to pay higher interest for the

money they themselves have borrowed. In this way,

so the argument, the market corrects itself and cre-

ates the proper amount of caution. The miserable

failure of the rating agencies during the present cri-

FINANCIAL CRISIS

* Ifo Institute for Economic Research.

sis shows, however, how illusory this reasoning is.The agencies did not give sufficient warning, andtheir AAA ratings were only withdrawn when therewas no other alternative. Since they live on the feesthey collect from the financial institutions theyrated and were dependent on their good-will, theycould not afford to tell the truth. The nearly bank-rupt major customers of the rating agencies inAmerica were glamorised while comparably robustbut smaller customers in Europe were downgraded.This is also how the credit packages with claimsagainst American homeowners, which were alreadyin risky territory, were offered to the world farabove value.

The best proof that the rating agencies and otherinformation channels do not function and are notable to reliably inform purchasers of bank bonds andcredit packages about the true circumstances lies inthe fact that on the capital market equity capital isalways more expensive than debt capital. If the pur-chasers of bank bonds had been correctly informedabout the true repayment probability, they wouldhave demanded adequate risk premiums on interestor sufficient reductions in the prices of these bonds,which would have made these liabilities just asexpensive for the banks as equity. Finance theorydesignates this finding as the Modigliani-Miller the-orem, after its authors. But this theorem fails tomatch reality. Everyone uses the leverage effects ofdebt capital up to the limit that the rating agenciesestablish in order to achieve higher yields from equi-ty capital.Whoever does not do this and instead rais-es his equity-asset ratio to increase repayment prob-ability is not rewarded for his virtue by the capitalmarket.

Bank bonds and securitised risks are entwined in acascade of interlinked legal claims at whose endthere is somewhere a real investment project. Theseare products that even specialists cannot properlyappraise. The purchasers are almost never able toassess the true repayment probabilities correctly.Only the sellers that assemble the securitised pack-ages have some idea of what they are selling. In thelanguage of economists, these bank products arelemon goods, that is goods whose quality can onlybe partially assessed by the customers at the time ofpurchase and for this reason are usually offered atinferior quality. The sellers exploit the customers’lack of information by reducing their costs at theexpense of quality, knowing that the customers arenot able to punish them by refusing to purchase or

by demanding price discounts. Quality declinesbelow the quality that would prevail in a market ofinformed customers. In order to prevent lemon mar-kets, most countries have, for example, food regula-tors who set the lower limits for quality in food inthe form of upper limits for unhealthy ingredients.In the case of pharmaceuticals, quality is safeguard-ed by the licensing procedures. The loans given tohomeowners in the United States – and that endedup as mortgaged-backed securities and collateral-ized debt obligations – are lemon products. InAmerica higher rates of indebtedness are morecommon than in Germany. But the banks do nothave the same claims to the private assets or incomeof homeowners than in Germany. If a low-incomeUS homeowner chooses, he can hand over his housekeys to the bank and has no repayment obligation.Conscious of this limited liability, US homeownerswere much too cavalier in taking on real-estatewhich they could only afford if housing prices con-tinued to rise. The real-estate bubble that began toburst one and a half years ago and that gave rise tothe banking crisis arose this way.

Since the Reagan presidency a quarter century ago,Americans have increasingly become indebted toforeign creditors and have made a good life forthemselves. They financed their investments fromthe capital streaming in from foreign countries andinstead of saving relied on the increasing value oftheir real-estate. The deficit on the current accountbalance, that is the surplus of imported products andservices over exports, reached a peak of 5.5 percentof GDP. This was financed with increasingly moresophisticated investment products that were certi-fied with the stamp of the rating agencies – in theend even the last investment manager of theGerman state banks noticed what junk was beingsold here.The wheeling and dealing has now come toan end. No European bank escaped the painful expe-rience that the expensive value-at-risk models of theinvestment bankers were just as worthless as theagency ratings.

Politicians must finally face the task of defining legalliability limitations for corporations by establishingstrict minimum standards for equity capital require-ment for the various business models of the banks,both in America and in Europe. Stricter rules are nota disadvantage for the economy, since the apparent-ly so much more expensive equity capital, whose useis thus made compulsive, is not economically moreexpensive than debt capital, as shown by the burdens

CESifo Forum 4/2008 4

Focus


Focus

that the taxpayers must now bear. Furthermore, noscarcity of funds would arise as a result, since thesavings of the world just suffices – independently ofsuch rules – to finance the investments.

The necessary steps are as follows:

1. The United States must finally participate ininternational agreements on the harmonisationof banking supervision. These agreements can bebased on the Basel II system, which must beunder government control.

2. Europe needs a common system of financialsupervision. Every state must pay for the losses ofits own banks.

3. Investment banks, hedge funds and private equi-ty firms must be subjected to the same rules ascommercial banks.

4. Personal liability limitations for mortgages andother real-estate loans must be lifted in theUnited States and wherever else they exist.

5. Conduits and other constructs for the shifting ofinvestment banking business from the bank bal-ance sheets should be limited in such a way thatthe risks that the banks take on are transparent inthe bank balance sheets.

Free market advocates that argue against theseremedies, without which a market economy cannotsurvive, confuse the market economy with anarchy.The market economy can only function when it issubjected to traffic regulations. Civil codes in manycountries are full of rules that limit private con-tracts. Only a portion of the contracts that an uncon-trolled market economy would develop is allowed,and because of this the system functions. Europeand the world need stricter rules for financial traffic.Such rules do not constitute a systemic break. Theyare vital for the functioning of the financial capitalmarkets.


Focus

ORIGINS AND RESPONSES

TO THE CURRENT CRISIS

BARRY EICHENGREEN*

Nearly two years after the outbreak of the credit cri-sis (which can be dated to March 2007, when majorlosses were announced by the US subprime-basedinvestors Accredited Home Lenders Holding andNew Century Financial), key issues remain to beresolved. At the most basic level the questions aretwo. What caused the crisis? And in light of one’sanswer to this first question, what should be done tominimize the risk of repetition if not of identicalevents then of something similar?

To say that these questions remain to be answered isnot the same as saying that there has been a shortageof attempts. Standard operating procedure starts byrounding up the usual suspects: unethical mortgagebrokers, greedy bankers, naïve homeowners, and ill-informed investors. Lists focusing less on individualsthan mechanisms emphasize agency problemsbetween brokers and banks, the originate-and-dis-tribute model, excessive leverage and short-termfunding, the perverse incentives created by executivecompensation practices, conflicts of interest withinthe rating agencies, permissive monetary policiesand even high Chinese savings rates. These long listsof causes lead to correspondingly long lists ofreforms: regulate mortgage brokers, rating agencies,and executive compensation; force banks to keep aparticipation in any securities they originate; requirebanks to hold more capital; revisit whether monetarypolicy should respond to credit booms and assetbubbles; and revalue the renminbi. This of course isonly a very incomplete summary of a vast and rapid-ly-growing literature.

The limitations of this standard operating procedurewill be apparent. However successful it is at pin-pointing the immediate causes of the crisis, it fails toidentify the deeper conditions that allowed those

immediate causes to arise.While there is no questionthat investment banks and other financial institu-tions relied excessively on leverage and short-termfunding, for example, the deeper question is howthey came to do so and why such practices were sofreely allowed. Similarly, while there is no disputingthat relying on self-regulation in the form of banks’internal models of value at risk and commercialcredit ratings meant inadequate regulation, thedeeper question is how this belief in the efficacy ofself-regulation was allowed to develop. Failing toinquire into deeper forces may lead to regulatoryreforms that address symptoms rather than funda-mental causes, allowing those causes to again mani-fest themselves in the future in different but equallydestructive ways.

A more parsimonious way of putting the point is: ifthe causes of the crisis are so obvious in retrospect,then why did so many smart people fail to appreciatetheir gravity prior to the event?

Major causes

At the most fundamental level, the causes of the cur-rent crisis go back to the Great Depression of the1930s.1 A factor contributing to financial problems inthat period, in conventional analysis, was conflicts ofinterest between the commercial- and investment-banking arms of large financial conglomerates. Itwas the tendency for the investment-banking divi-sion run by individuals with high risk tolerance togamble the funds of small retail depositors, in thisview, that led to the bank runs and failures that con-verted a garden variety recession into the GreatDepression.2 Whether or not this diagnosis is accu-rate is disputed.3 But, matters of historical accuracynotwithstanding, for present purposes this dispute isbeside the point. The point is that the diagnosis led

* University of California, Berkeley.

1 For some authors, including present company, this would seem tobe the locus classicus of all things economic.2 The Pecora Commission established by the US Senate Committeeon Banking, Housing and Urban Affairs alleged a number of spe-cific conflicts of interest, such as the tendency for banks to under-write questionable securities in order to pay off their own badloans. The obvious irony implicit in this analysis is that the roots ofthe current crisis lie not merely in today’s incentive problems butin incentive problems that arose in the 1920s.3 See White (1986) and Kroszner and Rajan (1994, 1997).


Focus

to the Glass-Steagall Act separating investment andcommercial banking and to tight regulation of thefinancial services industry. Without access to retaildeposits and with money market instruments tightlyregulated, investment banks funded themselvesusing their partners’ capital and could therefore beplaced outside the financial safety net.

Tight regulation and populist posturing were under-standable reactions to the financial disaster that wasthe Great Depression. But as memories of thatepisode dimmed, the advantages of financial mar-kets and financial innovation acquired more weightin policy discussions. In the United States the twocritical events marking this trend were the deregula-tion of commissions for stock trading in the 1970s,the removal of Regulation Q placing a ceiling oninterest rates on retail deposits in the 1980s, and theelimination of the Glass-Steagall restrictions on mix-ing commercial and investment banking in the 1990s.In the days of fixed commissions, investment bankscould make a cushy living booking stock trades fortheir customers. Deregulation meant more competi-tion, entry by low-cost brokers like Charles Schwab,and thinner margins. The elimination of Glass-Steagall then allowed commercial banks to encroachon the investment banks’ other traditional pre-serves.4 Forcing commercial banks to compete fordeposits on price in turn left them no choice but topursue these new lines of business.

In response, investment banks, in order to survive,were forced to branch into activities like originatingand distributing collateralized bond obligations(investment-grade bonds backed by pools of junkbonds), whose market collapsed in the second half ofthe 1980s. Next they developed asset-backed securi-ties and mortgage-backed securities, which gained aclientele in the 1990s. They utilized more leverage,funding themselves through the money market, tosustain their profitability. Commercial banks, nowpaying interest on deposits, were anxious to put theirovernight money to work; this rendered themresponsive to the investment banks’ requests forshort-term funding. Thereby arose the first set ofcauses of the crisis: the originate-and-distributemodel of securitization and the extensive use ofleverage.

It is important to note that these were unintendedconsequences of well-motivated policy decisions. It

is hard to defend rules mandating price fixing instock trading and placing ceiling on deposit rates.The deregulation of commissions allowed smallinvestors to trade stocks more cheaply, which madethem better off, other things equal. Similar argu-ments can be made about the removal of ceilingson interest on retail deposits. But other things werenot equal. Investment banks, which were propelledinto riskier activities by these policy changes andtightly connected to other financial institutions bythe interbank market, remained outside the regula-tory net. Allowing commercial banks to branchinto new activities outstripped the capacity of reg-ulators to keep pace, especially when budgetaryimperatives and ideology left the regulatory agen-cies starved of resources. A fragmented regulatoryregime suitable for a segmented financial-servicesindustry became increasingly inadequate as thesharp lines between commercial banking, invest-ment banking and other financial institutions wereerased.

One can imagine how, with sufficient time, many ofthese problems would have been addressed. Wellbefore the gravity of the crisis had become apparent,US Treasury Secretary Henry Paulson tabled a planfor reorganizing and consolidating supervision andregulation of the country’s financial system. One cansimilarly imagine that, sooner or later, federal insur-ance would have been extended to money-marketmutual funds, and investment banks would havebeen brought under the regulatory umbrella. Butchanges in regulatory practice take time.At the mostbasic level the subprime crisis resulted from the ten-dency for financial normalization and innovation torun ahead of financial regulation.

Similarly, eliminating Glass-Steagall’s restrictions onmixing investment and commercial banking was afundamentally sensible choice. Conglomeratizationallows financial institutions to better diversify theirbusiness. Combining with commercial bankingallows investment banks to fund their operationsusing a relatively stable base of deposits rather thanrelying on fickle money markets. This model hasproven its viability in Germany and other Europeancountries over a period of centuries. These advan-tages are evident in the United States even now, withBank of America’s purchase of Merrill Lynch andthe decision of Goldman Sachs and Morgan Stanleyto transform themselves into bank holding compa-nies and access deposits on the wholesale market orby purchasing regional banks.

4 It was not only commercial banks of course, but also insurancecompanies like AIG that did the encroaching.


Focus

Again, however, the problem was that other policieswere not adapted to the new environment.Conglomeratization takes time. In the short run,investment banks were allowed to lever up theirbets. They remained wholly outside the purview ofthe regulators. As self-standing entities fundingthemselves short-term, they were vulnerable toilliquidity and interruptions to their funding. A crisissufficient to threaten the entire financial system ulti-mately precipitated the inevitable conglomeratiza-tion. Presumably there was a better way.

Another element in the crisis was the consumerspending boom in the United States and the result-ing domestic and international imbalances.The BushAdministration cut taxes upon taking office, result-ing in government dissaving. The Federal Reservecut interest rates in response to the 2001 recession.All the while the financial innovations describedabove worked to make credit even cheaper andmore widely available to households. This of courseis just the story, in another guise, of the subprime,negative-amortization and NINJA mortgagespushed by subsidiaries of the like of LehmanBrothers. The result was increased US consumerspending and the decline of measured householdsavings into negative territory.

What should have been done about this continues tobe disputed, even with the benefit of hindsight.5

Some say that the Fed should have reversed its lowinterest rates more rapidly in order to damp downthe lending boom and consumer-spending binge orthat the Congress should have raised taxes to rein inthe twin deficits. But either response would haveslowed and in the worst case interrupted recoveryfrom the 2001 recession.The appropriate response toill-advised lending and unsustainable credit expan-sion was more vigorous regulation of the financialinstitutions extending the lending and providing thecredit. Financial problems are best addressed usingfinancial instruments, macroeconomic problemsusing macroeconomic instruments. But in thisinstance more vigorous regulation ran up against theobstacles enumerated above.

The other side of this story was financial inter-nationalization. Much as with the separation of in-vestment from commercial banking, the GreatDepression led to the imposition of regulatory

restrictions on international capital flows (not somuch by the United States in this case as by othercountries). The gradual relaxation of those restric-tions, which was another logical reaction of policymakers as memories of the Great Depression faded,and which picked up speed in the 1990s, is a well-known tale. The point here is that by facilitating USdependence on foreign finance and feeding thecountry’s credit boom, it helped to set the stage forwhat followed.

The other element helping to set the stage for thecrisis was the rise of China and the decline of invest-ment in Asia following the 1997–98 crisis.With Chinasaving nearly 50 percent of its GNP, all that moneyhad to go somewhere. Much of it went into US trea-suries and the obligations of Fannie Mae andFreddie Mac. This propped up the dollar. It reducedthe cost of borrowing for Americans, on some esti-mates, by as much as 100 basis points, encouragingthem to live beyond their means.6 It created a morebuoyant market for Freddie and Fannie and forfinancial institutions creating close substitutes fortheir agency securities, feeding the originate-and-dis-tribute machine.

Again, these were not exactly policy mistakes.Lifting a billion Chinese out of poverty is arguablythe single most important event in our lifetimes,and it is widely argued that the policy strategy inwhich China exported manufactures in return forhigh-quality financial assets was a singularly suc-cessful growth recipe.7 Similarly, the fact that theFed responded quickly to the collapse of the high-tech bubble prevented the 2001 recession frombecoming worse. But there were unintended conse-quences. Those unintended consequences wereaccentuated by the failure of US regulators to tight-en capital and lending standards when abundantcapital inflows combined with loose Fed policies toignite a credit boom. They were aggravated by thefailure of China to move more quickly to encouragehigher domestic spending commensurate with itshigher incomes.

Now we are paying the price. As financial problemssurface, a bloated financial sector is being forced toretrench. Some cases, like the marriage of Bank ofAmerica with Merrill, are happier than others, likeLehman. But either way there will be downsizing

5 This, of course, is simply the debate over the US contribution tothe problem of global imbalances and what if anything should havebeen done about it.

6 See Warnock and Warnock (2006).7 Assuming that the financial assets in question are indeed highquality – more on this below.


Focus

and consolidation of the financial-services indus-try. Foreign central banks are suffering capital loss-es on their US treasury and agency securities. Asthe People’s Bank of China and other centralbanks absorb losses on their holdings of US trea-sury and agency securities, they will questionwhether the United States is properly regarded asa supplier of high-quality financial assets.8 TheBretton Woods II model where China trades high-quality manufactures for high-quality US financialassets will lose its appeal.9 Capital flows to theUnited States will diminish. The US currentaccount deficit and Asian surplus will shrink. UShouseholds will have to begin saving again. Alower dollar will enable the United States toexport more merchandise, in turn reconciling itsbalance of payments with the more limited avail-ability of foreign financial capital. All this is of apiece. The shame is that it took a recession and full-blown credit crisis in the United States to bringabout the inevitable adjustment.

At a fundamental level, then, the crisis was madepossible by policies of liberalization, domestic andinternational, that – however well intentioned –allowed financial innovation in the form of newproducts (complex derivative securities) and vehi-cles (conduits and structured investment vehicles)to outpace supervision and regulation. But otherinnovations in the realm of risk managementworked in the same direction. In particular, thedevelopment of mathematical methods designed toquantify and hedge risk encouraged commercialbanks, investment banks and hedge funds to usemore leverage. They encouraged their counterpar-ties to provide it. They encouraged everyone tobelieve that the additional leverage was safe, sincethey now possessed rigorous techniques for manag-ing it. These elaborate models, overseen by Ph.D.’sin finance, could not be maintained in-house byevery financial institution, but they were extensive-ly utilized by large ones, notably the leading invest-ment banks. Not incidentally it was these institu-tions that utilized the most leverage and relied mostheavily on sophisticated derivative instruments tomanage their risk.

The problem was that these models were estimatedon data from periods of low volatility and, typically,relatively short intervals, since the instruments

whose returns were being modeled had existed foronly a relatively brief period of time.10 Rare eventslike a sharp drop in housing prices were outside thesample and not captured by the model. Institutionalinvestors convinced themselves on this basis thattheir practices were safe. They convinced theircounterparties. They convinced their regulators,who allowed them to use their own models whendeciding how much capital to hold to provisionagainst risk.11

All the while, as my colleague Robert Anderson putsit, institutional investors were incurring very largecontingent liabilities that would come due in theevent of that sharp fall in housing prices or anothersimilarly rare occurrence not included in the sampleperiod.12 A proper model would have forced thesefirms to book their contingent liabilities. Vigorousregulation would have required them to provisionagainst them. Forcing firms to book and provisionagainst those liabilities would have provided incen-tives to avoid excessive leverage.

The same story can be told about liquidity risk.Insofar as investors modeled the risks of relying onshort-term funding and incurring maturity mis-matches at all, they did so using data from normaltimes, not from those rare occasions when liquiditydried up.As a result, institutions held inadequate liq-uidity to guard against disruptions to the supply ofshort-term funding. This problem first hit NorthernRock, a British building society, in the late summerof 2007, but it eventually infected the entire USinvestment banking industry.

The other point at which financial engineering cameinto play, also emphasized by Anderson, was in fos-tering the neglect of counterparty risk. The rocketscientists designed credit default swaps and othercomplicated insurance contracts that reduced theperceived need to provision by fostering the impres-

8 This is the explanation for global imbalances modeled byCaballero, Fahri and Gourinchas (2008).9 See Dooley, Folkerts-Landau and Garber (2003).

10 This created the temptation to use returns on other assets tomodel the behavior of new ones. For example, the rating agenciesused their accumulated data and modeling experience in rating cor-porate bond defaults to predict defaults on complex derivativesecurities backed by mortgages. See Mason and Rosner (2007).11 Here we academics have something to answer for to the extentthat we trained our students and encouraged our consultancy cus-tomers to take our models literally.12 The problem was not merely the truncated and unrepresentednature of the sample but also misspecification of the model itself(specifically, the tendency to attach unrealistically low probabilitiesto extreme outcomes, also known as “tail risk”). Why this was aproblem is a topic for another day. One explanation would focus onthe limitations of the underlying theory. Another would emphasizeincentive problems within the banks constructing their models(specifically, the incentive for staff to understate tail risk in order tominimize the capital that had to be put aside by management –higher capital requirements squeezing profits and bonuses).


Focus

sion, reflected in models of value at risk, thatinvestors could shed unwanted risks. But insurance isonly as good as the insurer. It was eventually discov-ered that so-called monoline insurance companieslike MBIA and Ambac, which had seemed ade-quately capitalized in good times, had inadequateresources with which to make good on their insur-ance contracts when the housing market turneddown.13 The models had not incorporated this coun-terparty risk. This gap encouraged institutionalinvestors to neglect it when making their allocationdecisions.

Moreover, the fact that the notional value of creditdefault swaps was a large multiple of the value ofthe underlying bonds created the danger of a domi-no effect if one large issuer of default swaps was tofail. These swaps were traded over the counter withlimited liquidity and in the absence of a clearingmechanism. This created the danger that the defaultof one counterparty in this market might have adomino effect. This was why Treasury and theFederal Reserve were reluctant to allow BearStearns to fail in March 2008. It was part of whatmade Lehman Brothers’ failure in September 2008so disruptive. In this case, it would appear, theauthorities underestimated the threat from thecredit default swap market. It was what then forcedthe US government to step in and rescue AmericanInternational Group (AIG).

Prescriptions for reform

The preceding is by no means a complete or com-prehensive explanation for the crisis. Other authorswould emphasize other factors, and they wouldhave some justification.14 The selective emphasishere has the advantage of highlighting the deeperfactors that gave rise to the imprudent practiceshighlighted by other commentators. By focusing onfactors that have changed in recent years, it helps usunderstand why the crisis erupted when it did and

took its particular form. And by emphasizing morefundamental problems, it leads directly to prescrip-tions for reform.

A first prescription is to revisit the Basel II revisionof capital adequacy standards for internationallyactive banks. Two central pillars of Basel II arebanks’ internal models of value at risk and commer-cial credit ratings for banks lacking internal models,both of which have been revealed as fundamentallyflawed. Commercial credit ratings are unreliableunder the best circumstances, and in any case exist-ing circumstances are less than optimal: the ratingagencies suffer from significant conflicts of inter-est.15 Greater reliance on internal models wasdesigned to capture the implications of the correla-tions between different assets for risks to the balancesheets of commercial banks. Basel I had placed dif-ferent assets in different risk buckets and mechani-cally assigned capital requirements to each bucket,ignoring the correlations between them. It followsthat going back to Basel I would throw the baby outwith the bathwater.

A better solution would be to compute capitalrequirements in two ways – an old-fashioned BaselI way where required capital was taken as a simplefraction of the value of assets (and thus not evendistinguish different risk buckets), and the new-fan-gled Basel II way where required capital was a frac-tion of value at risk as computed on the basis ofcomplex models – and to make banks hold the high-er of the two. The Swiss National Bank has pro-posed requiring the banks under its jurisdiction tohold capital amounting to a fraction of assets, pureand simple. The present proposal is a generalizationof their idea.

A second implication is that capital requirementsshould reflect the liquidity of funding and the poten-tial illiquidity of assets and not just estimates of thevolatility of asset returns. Even financial institutionswith relatively conservative portfolios have experi-enced grave difficulties as a result of their inability torenew short-term funding. This has given rise to dis-cussion of the desirability of liquidity requirementsto supplement capital requirements. A morestraightforward response would be to generalize the

13 This was another case where regulatory decisions taken in thepast, under different circumstances, helped to set the stage for thecrisis. The monoline insurance industry arose in the 1980s whendefaults on municipal bonds caused the regulators to conclude thatregular (“multiline”) companies writing mom-and-pop insurancepolicies should not be allowed to dabble in riskier activities likebond insurance. This decision led to the growth of monoline insur-ers to fill the gap. At the time no one had reason to anticipate theexplosive growth of sub-investment-grade debt securities of allkinds, many of which had to be wrapped with bond insurance inorder to be eligible for inclusion in the portfolios of pension fundsand other institutional investors with restrictive covenants, or howthe demand for such insurance would outstrip the capitalization ofthe monolines – which would in turn make use of mechanical mod-els of value at risk in order to gauge their own capital needs.14 Some of my own writings do the same, see Eichengreen (2008).

15 Changes in their fee structure, as demanded by Andrew Cuomo,the Attorney General of New York State, where the agencies domuch of their business, can go some way toward alleviating theseconflicts. Similarly, freer entry into the industry can ratchet up com-petitive pressure and encourage the adoption of better practices.My views of the rating-agency problem are in Eichengreen (2008).


Focus

Basel II capital adequacy ratios to incorporate aweight for liquidity risk.

Emphasizing counterparty risk as a factor in thecrisis also suggests that minimum capital require-ments should be set not on the basis of the riski-ness of a bank’s individual assets and liabilities butwith the potential impact of its difficulties for therest of the financial system in mind. Capitalrequirements should reflect implications for sys-temic stability, not just implications for the subjectinstitution taken in isolation. This suggests thatbanks’ own internal models, which understandablyfocus on the actual and prospective stability of theindividual institution, are not the best basis forthis calibration. The model that regulators developand use may want to raise required capital ratiosin periods when bank balance sheets are expand-ing rapidly, creating additional risks for systemicstability.16

This diagnosis emphasizing counterparty risk alsopoints to the desirability of clearinghouse arrange-ments as an alternative to the over-the-counter mar-ket. Transactions over the counter are only as reli-able as the counterparty. In contrast, an organizedexchange or clearinghouse can net transactionsimmediately and multilaterally, reducing the dangerof cascading defaults owing to the default of a par-ticular market participant. Authors like Cecchetti(2007) have been arguing the case for exchange-based trading for some time. Recent events supporttheir position.

The question is then why exchange-based trading ofderivative securities has developed so slowly. Oneexplanation is path dependence: the over-the-counter market developed first and thereforeremains more liquid and attractive. A second expla-nation is that market participants prefer variety overliquidity: exchange-traded derivatives would bemore liquid, but exchange-based trading wouldencourage uniformity (it would be hard to organizeliquid exchange-traded markets in a wide variety ofpurpose-tailored instruments).A third explanation isthat the big broker dealers have a proprietary inter-est in the over-the-counter market. All three expla-nations are consistent with the view that a policyintervention to encourage exchange-based trading(the application of tax incentives, for example)would not be inappropriate.

Finally, this diagnosis of the crisis, emphasizing theremoval of the Glass-Steagall Act, the tendency forcommercial banks and insurance companies toencroach on the traditional preserves of investmentbanks and the tendency toward financial conglomer-atization, points to the need for consolidated regula-tion. Having different agencies regulate differentactivities when those activities are undertaken by asingle entity is an enticement for regulatory arbi-trage. It may also prevent any single regulator fromfully appreciating the risks to systemic stability (oreven to the stability of the individual institutionitself) of the entire range of activities in which thesubject institution is engaged. Ironic as it may seem,this suggests that Henry Paulson, who tabled a pro-posal for consolidating regulation of the US financialservices industry in the winter of 2007/8, may havebeen pointing in the right direction.

References

Caballero, R., E. Farhi and P.-O. Gourinchas (2008), “An Equi-librium Model of ‘Global Imbalances’ and Low Interest Rates,”American Economic Review 98, 358–393.

Cecchetti, S. (2007),“A Better Way to Organise Securities Markets,”Financial Times (4 October).

Dooley, M., D. Folkerts-Landau and P. Garber (2003), An Essay onthe Revived Bretton Woods System, NBER Working Paper 9971.

Eatwell, J. and A. Persaud (2008), “A Practical Approach to theRegulation of Risk,” Financial Times (25 August).

Eichengreen, B. (2008), Thirteen Questions about the SubprimeCrisis, University of California, Berkeley (January),emlab.berkeley.edu/users~eichengr.

Goodhart, C. and A. Persaud (2008),“How to Avoid the Next Crash,”Financial Times (30 January).

Kroszner, R. and R. Rajan (1994), “Is the Glass-Steagall Act Justi-fied? A Study of U.S. Experience with Universal Banking before1933,” American Economic Review 84, 810-832.

Kroszner, R. and R. Rajan (1997),“Commercial Bank Securities Ac-tivities before the Glass-Steagall Act,” Journal of Monetary Eco-nomics 39, 475-516.

Mason, J. and J. Rosner (2007), Where Did the Risk Go? HowMisapplied Bond Ratings Cause Mortgage Back Securities andCollateralized Debt Obligation Disruptions, Drexel University andGraham Fisher (mimeo).

Paulson, H. et al. (2008), Blueprint for a Modernized Financial Re-gulatory Structure, Washington DC: US Department of the Treasury(March).

Warnock, F. and V. Warnock (2006), International Capital Flows andU.S. Interest Rates, International Finance Discussion Paper 860,International Finance Division, Board of Governors of the FederalReserve System (September).

White, E. (1986), “Before the Glass-Steagall Act: An Analysis of theInvestment Banking Activities of National Banks,” Explorations inEconomic History 23, 33–55.

16 A specific proposal for countercyclically required capital ratios torestrain credit booms is Goodhart and Persaud (2008).

THE CAUSES OF THE

FINANCIAL CRISIS1

MARTIN HELLWIG*

Introduction

For the media in Germany, the cause of the financialcrisis is obvious: Blinded by greed, bank managersthought only about their bonuses and miscalculatedbadly in betting on American subprime mortgageswhen the very name of these securities should havealerted them to their risks. If an economist suggeststhat the matter might be more complicated, he isdenounced as a homo exculpans, a person who willexcuse anything that managers do.2

If we look at the numbers, however, we see that thereis something more to be explained. According to theGlobal Financial Stability Report of the Internation-al Monetary Fund (IMF) of October 2008, losses onnon-prime mortgage-backed securities in US resi-dential real-estate amount to some 500 billion dol-lars. This figure is both too small and too large.

The figure is too small in the sense that losses of500 billion dollars by themselves cannot explain whythe financial system worldwide has been so devastat-ed by the crisis. Around 1990, losses of savings andloans institutions in the United States were said toamount to some 600 to 800 billion dollars. A decadelater, losses on NASDAQ and on the New YorkStock Exchange amounted to 1.6 trillion dollars inthe calendar year 2000, 1.4 trillion dollars in the cal-endar year 2001, and again 2.7 trillion dollars in thecalendar year 2002. Neither episode caused a world-wide financial crisis.

At the same time, the figure of 500 billion dollars oflosses on non-prime mortgage-backed securities is

too large in the sense that it cannot be explained byanticipations of losses in debt service from thesesecurities. According to the IMF’s Global FinancialStability Report, the volume of non-prime mort-gages that have been securitized amounts to about1.1 trillion dollars. Losses of 500 billion dollars wouldcorrespond to a loss rate of 45 percent on thesemortgages. If the debtor’s down payment amountedto 5 percent, a loss rate of 45 percent on the mort-gage would correspond to a depreciation of theproperty by more than 50 percent. In actual fact, res-idential-real-estate prices in the United States onaverage have declined by 19 percent from their peakin the summer of 2006 to the summer of 2008; acrossmetropolitan areas, the maximum for this period wasjust below 33 percent (Phoenix, Tampa, Miami). Tobe sure, this “back-of-the-envelope” calculationneglects correlations; it also neglects the possibilitythat the decline of real-estate prices is still going on.However, this calculation also neglects the fact that,in actual fact, average down payment rates were 6percent for subprime and 12 percent for “Alt-A”, ornear-prime, mortgages, and that about two thirds ofthese mortgages had been granted before 2006, attimes when real-estate prices were significantlybelow their subsequent peaks.

The IMF’s loss estimates are not actually based onprojections of debt service on subprime mortgages.They are based on market prices for mortgage-backed securities. In some cases, where markets arenot functioning any more, they are based on guessesas to what market prices might be if the marketswere functioning. The IMF itself points out thatthese prices may not be good indicators of thereturns that can be expected if one is willing to holdthese securities to maturity. According to the IMF,therefore, market prices at this point are not a goodbasis for taking for long-term, value-maximisingdecisions.

Under Fair Value Accounting, however, these marketprices are used to value the securities in the banks’books. If, over the past year, banks have forever been“discovering” new losses, the reason is not thatbankers are too stupid to know or too devious toreveal what their losses really are. The reason israther that, week by week and month by month,


Focus

* Max-Planck Institute for Research on Collective Goods, Bonn.1 The following text is based on the Jelle Zijlstra Lecture which Igave in Amsterdam on May 27, 2008. The text of the lecture will bepublished by the Netherlands Institute for Advanced Study. A pre-liminary version is available at http://www.coll.mpg.de/pdf_dat/2008_43online.pdf.The Jelle Zijlstra Lecture gives references and sources for all mate-rial in this text.2 Frankfurter Allgemeine Zeitung, October 27, 2008.


Focus

market prices have been going down and the banks’losses have become ever larger as market partici-pants have become ever less willing to hold thesesecurities – or less able to hold them.

The financial crisis is not just a matter of excessivelending in subprime mortgages and excessive securi-tization. To understand the crisis, we need to look atsystemic interdependence, i.e. the mechanisms bywhich the subprime-mortgage crisis spilled over intothe rest of the financial system.

The securitization of real-estate finance

Before I turn to the systemic issues, I briefly want todiscuss the role of securitization itself. I begin withthe proposition that, in principle, the securitizationof real-estate loans is a good thing. It would be prob-lematic if the crisis led us to throw the baby out withthe bathwater and banned this financial innovation.

Many financial crises in the past have been associat-ed with real-estate finance. The crisis of US savingsand loans institutions in the 1980s was initiallycaused by the fact that, as a result of governmentregulation, these institutions had provided too muchmaturity transformation, from short-term deposits tofixed-rate mortgages, and that their assets wereinsufficiently diversified. Excessive real-estatefinance and the subsequent downturns in real-estatemarkets also were a factor in the banking crises thathit the United States, Sweden, Switzerland, Japanand other countries in the late eighties or earlynineties. In these years, Germany did not have abanking crisis, but German banks certainly had theiradventures with excessive real-estate lending.

Real estate is problematic because investments arelumpy, economic lifetimes are long, and, in anyadvanced economy, the total volume outstanding isvery large. In most OECD countries, the aggregatevalue of residential real-estate has the same order ofmagnitude as the aggregate net value of financialassets; indeed, in many countries, it is higher. Theeconomic lifetime of a house by far exceeds theinvestment horizon of the ordinary saver. The dis-crepancy between the economic lifetime of a pieceof real-estate and the investment horizon of mostsavers is a source of risks, refinancing risks if thereal-estate investment is financed by short-term bor-rowing, valuation risks, if the real-estate investmentis financed by long-term securities, and the saverwants to liquidate these securities when he needs themoney. These risks could be avoided if we chose to

live in tents. If we are not willing to live in tents, wemust accept the existence of these risks as a fact oflife. The only question then is who should bear them.

In the past, these risks used to be borne by financialintermediaries.The US savings and loans institutionsfinanced themselves from savings deposits withmaturities of up to seven years and granted fixed-interest mortgages with maturities of up to 40 years.In 1980/81, about two thirds of these institutionswere technically insolvent: market rates of interestfor all maturities were significantly above 10 per-cent.To keep their depositors from leaving them andinvesting in money market funds rather than savingsaccounts, the savings and loans institutions had toraise their deposit rates so that their own debt ser-vice was actually less than the debt service theyreceived from the mortgages that they had issued along time ago. Equivalently, when discounted at pre-vailing interest rates, the present values of debt ser-vice on long-term mortgages that had been grantedin the 1960s were significantly below the nominalvalues of these mortgages.

Given this experience, the 1980s saw the emergenceof adjustable-rate mortgages, under which fluctua-tions in market rates of interest were passed on todebtors. However, when interest rates were highagain, in 1989 and later, lending institutions, in othercountries as well as the United States, made theexperience that increases in mortgage rates underthe given adjustment clauses simply induced debtordefault. Moreover, at the high market rates of inter-est, property values were depressed. In other words,the attempt to have the risk from maturity transfor-mation in real estate be borne by debtors did notturn out to be propitious either.

Securitization of real-estate finance is based on theassessment that neither the financial intermediariesnor the debtors are in a position to carry the risks ofmaturity transformation in real-estate investmentand finance. Securitization provides a basis for pass-ing these risks on to third parties. This makes eco-nomic sense if the third parties are better able tobear the risks that are involved. This is true, forexample, if the third party is a life insurance compa-ny or a pension fund. These institutions usually haveliabilities with very long maturities; in principle,therefore, short-term fluctuations in market valua-tions of assets should matter much less for them thanfor depository institutions.

It also makes sense for some of the risks of real-estate finance in one country to be passed on to

financial institutions worldwide. Such sharing of risksby many institutions in many countries improves theoverall risk allocation by providing for greater diver-sification of risks for each institution. Public discus-sion of the losses of German banks from subprime-mortgage-backed securities in the United Statesoften carries an undertone that a decent bank shouldinvest its funds at home rather than abroad. To someextent, this involves the populist notion that “our”banks should lend to “our” firms; to some extent, itinvolves the notion that, if it invests closer to home,the bank has better information about the risks thatit is incurring. The populist rhetoric about “our”banks raises questions about the semantics of theword “our” in a society in which ownership is private,and both the ownership and the control of firms areprotected by the rule of law. Whether the creditwor-thiness of loan clients is more easily assessed closer tohome is a matter of dispute; such notions may well bethe result of overconfidence bred by familiarity.However, any regulation requiring “our” banks tolend to “our” firms would leave the banks seriouslyunderdiversified. Indeed, non-diversified domesticlending, in particular real-estate lending, played amajor role in the banking crises of the late 1980s andearly 1990s. The most blatant example was that ofsavings and loans institutions in Texas where stateregulation had required that these institutions limittheir real-estate lending to properties in Texas. Whenthe oil price decline of 1985 caused a downturn in theTexan economy, the prices of Texan real estatedropped dramatically, and mortgage borrowersdecided that it was better to default on their loans.These events caused a crisis of savings and loans insti-tutions in Texas – about two years before the generalcrisis of savings and loans institutions in the UnitedStates began to erupt again.

The operations of packaging and tranching that areassociated with securitization also make good eco-nomic sense. By putting many different mortgagesinto one package that serves as collateral for a mort-gage-backed security, one makes the mortgage-backed security somewhat independent of the risksthat are specific to any one mortgage and any oneproperty. The standardization of securities that isthereby achieved provides for their marketability.With packages rather than single mortgages servingas collateral, buyers of mortgage-backed securitieshave fewer reasons to be afraid of information asym-metries concerning the quality of the underlying col-lateral.

Tranching, i.e. the issuance of different kinds of debtsecurities with different priority rankings and equity

as the first loss absorber, can in principle reduceadverse incentive effects from securitization. Theprobability distribution of losses from a portfolio ofstochastically independent loans tends to be highlyskewed; losses above ten percent are exceedinglyunlikely. Debt titles with claims of up to ninety per-cent of the returns on the portfolio may then bedeemed as very safe. Credit risks on the underlyingreal-estate loans affect mainly the equity tranche. Ifthe equity tranche goes back to the bank that initiat-ed the mortgages, this bank has an incentive to useproper care in its creditworthiness assessments. Ifinstead the equity tranche is retained by the securi-tizing bank, this institution will impose minimumstandards on the mortgages it acquires from the ini-tiating banks.3

When mortgage securitization was developed in theUnited States, the initiating banks were not madeliable for the credit risks of the mortgages theyissued. This omission initially did not make much ofa difference. The securitization was carried out byFannie Mae and Freddie Mac, the Federal NationalMortgage Association and the Federal Home LoanMortgage Corporation.These government sponsoredenterprises guaranteed the debt service on the secu-rities that they issued. At the same time, theyimposed minimum standards for creditworthiness ofthe borrowers whose mortgage were to be taken intoa mortgage portfolio for securitization; the termprime mortgage designates mortgages for whichthese minimum standards are fulfilled. For primemortgages and prime-mortgage-backed securities,expected losses and market write-downs are stillquite small: as of October 2008, the IMF assessed theloss rate on these securities at 2 percent.

Because Fannie Mae and Freddie Mac had original-ly been government institutions and because theystill had a privileged position with the US Treasury,many investors believed that they were backed bythe government even though, in fact, they had beenprivatized long ago, and there was no explicit gov-ernment guarantee. Given this belief, the debt ser-vice guarantees that they provided made mortgage-backed securities appear to be very safe.

Following the burst of the stock market bubble in2000 and the subsequent loss of underwriting activi-ties in stocks, private investment banks discoveredthat debt securitization held great promises as a newline of business and aggressively moved forward into


Focus

3 An example for this sort of arrangement is provided by theGerman Pfandbrief, a security that is backed by a portfolio of mort-gages, where the issuing bank retains full liability for the promiseddebt service.


Focus

this field. Unlike Fannie and Freddie they did notprovide any debt service guarantees for the mort-gage-backed securities that were issued under theirauspices, usually through special purpose vehiclesthat had been created for just this purpose.Moreover, they concentrated on mortgages that didnot fulfil the quality requirements of Fannie Maeand Freddie Mac, the so-called subprime mortgages.In particular, they did not impose lower bounds ondown payment rates or upper bounds on debt ser-vice to income ratios. Even for consumer creditscores, their standards were below those of Fannieand Freddie.

In these developments, no attention seems to havebeen paid to the fact that, with the displacement ofFannie and Freddie by the private investment banksand with the expansion mortgage lending and secu-ritization from prime mortgages to subprime mort-gages, there was no more check against a deteriora-tion of borrower creditworthiness. The standardsused for assessing subprime mortgages were easierto manipulate than the standards for prime mort-gages, the private investment banks shunned the sortof liability that Fannie and Freddie had borne, andthe initiating mortgage banks never had any liabilityat all. I suspect that this oversight was partly due tothe fact that, by the time the private investmentbanks entered the field, securitization was seen as aninvestment banking activity and, cultural differencesbetween investment bankers and loan officers beingwhat they are, investment bankers do not have muchof a feeling for credit risk.

The emergence of the “subprime” bubble2003–2006/07

In 2003 to 2006 the securitization activities of privateinvestment banks’ in the area of subprime mortgagesgrew dramatically. By 2006, more than 40 percent ofnewly granted mortgages belonged to this category,as opposed to 9 percent in 2000; the share of sub-prime mortgages in the stock of outstanding mort-gages had risen from 7 percent in 2000 to 14 percentin 2006. During these years, there was a steadydecline in mortgage quality. To some extent, thequality decline concerned observable variables suchas down-payment rates or debt service to incomeratios. However, econometric studies of delinquen-cies suggest that the quality decline also concernedunobservable variables; conditional on all observ-able variables, delinquency rates twelve months afterthe conclusion of mortgage contracts rose steadilyfrom 2001 to 2006. However, in 2004 and 2005, the

reduction in the quality of the borrowers was con-cealed because increases in real-estate pricesinduced significant increases in borrowers’ equityshares within a year of the conclusion of the mort-gage contract.

After a period of stagnation in the 1990s, real-estateprices in the United States had increased by about9 percent per year from 1999 to 2003, then by almost14 percent from 2003 to 2004 and by almost 16 per-cent from 2004 to 2005. It is probably not a coinci-dence that the jump in the rate of real-estate appre-ciation in 2003 occurred at the very time when theprivate investment banks began to move aggressive-ly into the mortgage securitization business.

It is probably also not a coincidence that this expan-sion occurred at a time when monetary policy in theUnited States was very loose and the yield curve wasvery steep. From 2002 to 2004, interest rates in USmoney markets were significantly below 2 percent,as opposed to 6 percent in 2000 and 4 percent in2001. Long-term interest rates had also fallen, butmuch less than short-term rates: the interest for ten-year Treasuries fell from around 6 percent in 2000 tojust over 4 percent in 2003–2005, the mortgage ratefor fixed-rate prime mortgages fell from around8 percent 2000 to just under 6 percent p.a. in2003–2005. The excess of this mortgage rate over theinterest rate in the money market thus moved from200 basis points (2 percentage points) in 2000 to over400 basis points (4 percentage points) in 2003–2004.

The risk premium for fixed-rate subprime mortgageshad been at 300 base points in 2001 and fell to 100base points in 2004. This decline in the risk premiumfor subprime mortgages is all the more remarkablebecause, as mentioned above, it coincided with adecline in the quality of subprime-mortgage borrow-ers.4 At the same time, there was no comparabledecline in the risk premia for lower-rated corporatebonds.

These observations suggest that the entire develop-ment was supply-driven rather than demand-driven.The aggressive move of private investment banksinto the business of securitizing subprime mortgagescontributed to the lowering of risk premia even with-out of any general change in risk appetites. Investorsin search of high yields were happy to make moreand more funds available for housing finance in thesubprime segment of the market.These investors did

4 The number of cases of fraud in connection with new mortgagesgrew fivefold from 1996 to 2005; in 2003 the growth rate was 77 per-cent, in 2004 93 percent.

not impose any “market discipline”, i.e. quality stan-dards that would have forced the securitizing invest-ment banks and the initiating mortgage banks toaddress the problem of creditworthiness of the finalborrowers.

Who were these investors? Three groups are of par-ticular interest:

– Most equity tranches ended up with hedge fundsand investment banks that were hungry for highyields, as the phrase went. Little thought seems tohave been given to the implications of the mar-keting of equity tranches on the originatingand/or the securitizing institutions’ incentives.

– The so-called mezzanine tranches, subordinated-debt tranches, were being acquired by investmentbanks that wanted to use them as collateral in asecond round of securitization, creating the so-called MBS CDOs, collateralized debt obligationsthat were backed by mortgage-backed securities.In this second round of securitization, debt secu-rities with different priority rankings and equityas a first loss absorber would be issued against aportfolio of (mezzanine) mortgage-backed securi-ties. One purpose of this operation was to obtainadditional funds even for subordinated-debttranches of mortgage-backed securities: if thecredit risks in, say a portfolio of BBB-rated mez-zanine securities were deemed to be sufficientlyindependent, the “super-senior tranche”, i.e. thedebt with the highest priority ranking against thisportfolio, might be given a AAA rating and mightthereby be eligible for inclusion in the portfoliosof institutional investors, like certain insurers, thatwere required to invest only in AAA-rated secu-rities. European banks, whose access to the initi-ating mortgage banks was worse than that of theirAmerican counterparts, were particularly activein this second round of securitization, which theysaw as an opportunity to get a share of the action.

– Many of the securities that were produced by thedifferent rounds of securitization were acquiredby special entities, the so-called conduits andstructured investment vehicles (SIVs), whichbanks in Europe as well as the United States wereusing to acquire and hold such securities withouthaving to back them up with equity capital. Thesespecial entities had virtually no equity capital oftheir own. Moreover, they financed themselves byissuing asset-backed commercial paper, debtsecurities with maturities of one year or less.

To me as an outside observer, it is remarkable thatthe different participants seem to have been entirely

focussed on yields, apparently without paying muchattention to risks. Questions about incentives andliabilities in origination and securitization did notreceive much attention. The rating agencies’ assess-ments of the different securities were not ques-tioned. Indeed, there seems to have been no concernthat ratings of AAA on different securities mustmean different things if the interest rates on thesesecurities differed by fifty or so basis points.

The assessments of mortgage-backed securities bythe rating agencies seem to have been fundamental-ly flawed. The agencies seem to have assumed thatproblems of borrower creditworthiness would bedefused by continued increases in real-estate prices –partly because such increases would raise the bor-rowers’ stakes in their properties and partly becausethey would improve the protection that the proper-ties serving as collateral were providing to lenders. Inassuming that real-estate prices would continue torise, they failed to see that some of the reasons forthe increases that had occurred, in particular thelowering of interest rates from 2000 to 2003 and theinflow of funds into these markets that was due tothe development of subprime-mortgage securitiza-tion, were one-time changes that would not berepeated, and that, therefore, an extrapolation frompast and current real-estate price increases into thefuture was unjustified.

The agencies also seem to have neglected the corre-lations of credit risks that were caused by the depen-dence of all mortgage contracts on common factorssuch as changes in market rate of interest andchanges in real-estate prices. Neglect of correlationsis the only reason I can see for why, in the assessmentof MBS CDOs, they would have given an AAA rat-ing to the super-senior tranche on a package ofBBB-rated mezzanine securities. However, correla-tions came into play when interest rates again beganto rise and real-estate prices began to fall.

Systemic risks in the crisis

Beginning in 2005, US monetary policy becamemore restrictive, with some caution at first and thenvery strongly, so that interest rates in US moneymarkets moved back to around 5 percent in 2006 and2007. Real-estate prices continued to rise from 2005to 2006, albeit at a slower rate of 7.5 percent; in thesummer of 2006, they began to fall, first slowly, at3.6 percent from 2006 to 2007 and then at 15.3 per-cent from the summer of 2007 to the summer of2008.


Focus


Focus

When real-estate prices began to fall, delinquencyrates increased dramatically. The impending difficul-ties in subprime mortgages and mortgage-backedsecurities were quickly recognized, but, prior toAugust 2007, hardly anybody appreciated the impli-cations for the overall financial system. In April2007, the Global Financial Stability Report of theInternational Monetary Fund provided a detaileddescription of the crisis in subprime-mortgages andsubprime-mortgage-backed securities, but endedwith an assessment that the crisis was unlikely tospread to other parts of the financial system. In June2007, the Annual Report of the Bank for Inter-national Settlements provided a similar assessment.

In August 2007, however, the crisis of subprimemortgages and subprime mortgage-backed securitiesdid turn into an international financial crisis. Thetriggering event was the downgrading of a large setof mortgage-backed securities by the rating agencies,some of them by three grades at once, somethingthat was almost unheard of for corporate bonds. Thisdowngrading had an immediate impact on the mar-ket prices of these securities. Even more important-ly, it made market participants wake up to the factthat these securities were much less safe than hadbeen thought to be, not just in terms of the underly-ing credit risk but also in terms of market risk asso-ciated with the downgrading and the system’s reac-tion to this downgrading.

This first surprise was almost immediately followedby a second surprise. Lagging debt service on certainmortgage-backed securities and declines in thesesecurities’ prices after the downgrades caused lossesat institutions holding them. For some of these insti-tutions, these losses caused problems for solvency andfor refinancing. Hardest hit were a few hedge fundsand various special entities, the conduits and struc-tured investment vehicles (SIVs), which European andAmerican banks had used as instruments for investingin mortgage-backed securities without increasing theamount of capital that they had to hold in order tomeet regulatory requirements. These entities hadpractically no equity capital; they invested in asset-backed securities and refinanced themselves by issu-ing commercial paper. Given the lack of equity capi-tal, the initial declines in the prices of mortgage-backed securities wiped out their solvency and tookaway their access to the commercial paper market.Pledges of liquidity provision in case of need that theyhad previously been given by the sponsoring bankswere then called in. However, not all of these pledgeswere honoured, some of them were insufficient toentirely refinance the conduit or SIV in question, and,

given the losses that had already occurred, some ofthese pledges exceeded the sponsoring banks’ owncapacities to absorb losses. For Industriekreditbankand Sächsische Landesbank, for instance, liquiditycommitments to conduits amounted to more thanfour times and ten times the bank’s equity!

Whereas the existence and use of conduits and SIVshad always been known, hardly anybody appreciatedthe scope of their activities. Ex post, their holdings ofasset-backed securities have been estimated at 1 tril-lion dollars as of July 2007, almost equal to the totaloutstanding volume of securitized subprime mort-gages (1.1 trillion dollars), over 17 percent of the vol-ume of all securitized residential mortgages in theUnited States. The information that such a large partof the outstanding volume of mortgage-backed secu-rities had been financed by issuing short-term com-mercial paper and that this refinancing mode was nolonger available came as a second surprise. This sec-ond surprise very much magnified the impact of thefirst surprise, the drastic downgrading of mortgage-backed securities by the rating agencies. Either sur-prise on its own would have required a significantadjustment of market prices. Coming together, theirconsequences were all the more dramatic.

The surprises that had just been experienced con-tributed to an atmosphere of mutual mistrust. Thesolvency problems of conduits and SIVs and of thebanks that had sponsored these institutions raisedthe question of who else might be involved. As aresult of such mistrust, interbank lending was muchreduced; indeed, since August 2007, there have beenrecurrent instances when interbank markets stoppedoperating altogether, and central banks were theonly institutions providing liquidity.

Given this mistrust of investors, institutions likeinvestment banks and money market funds that hadbeen used to refinancing themselves by short-termsecurities now had to make provisions for the eventthat their financiers might cease funding them.Money market funds did so by moving out of com-mercial paper and into government securities, fromcash to short-term Treasuries; investment banks alsotried to shift from less liquid to more liquid assets.This left little room for buying mortgage-backedsecurities, even though the prices of these securitiesmight have dropped much more than the discountedpresent values of expected debt service on theunderlying mortgages.

Given this atmosphere of apprehensiveness and mis-trust, in response to the twin surprises of August

2007,5 the international financial system has inex-orably moved in a downward spiral.At times, the spi-ral has been slowed down by central banks providingliquidity to the system. However, as long as the issueof solvency of participating institutions was notaddressed – and the downward spiral itself went onexacerbating this issue, there was no prospect of thespiral ending. The subsidies and guarantees that gov-ernments in the United States and Europe have pro-vided after the fiasco created by the insolvency ofLehman Brothers may provide such a prospect – ifthese guarantees are indeed sufficient to eliminatedoubts about financial institutions.

The downward dynamics have been driven by theinterplay of following factors:

Market Malfunctioning: The prices of many asset-backed securities have gone down much more thanexpectations of future debt service or foreclosureproceeds would seem to warrant. This explains thediscrepancy, mentioned in the introduction, betweenthe estimated losses of 500 billion dollars on thesesecurities and the losses that would be anticipated onthe basis of actual developments in real-estate mar-kets. There are few buyers for these securities; somemarkets have become inactive altogether. Evenwhere markets have remained active, following thetwin surprises of August 2007, maturity premia andrisk premia have risen and prices have declined dra-matically. The lack of buyers is due partly toinvestors feeling too weak to take on new commit-ments, partly to their expecting the price declines tocontinue, and partly to their being afraid of adverseselection. Akerlof’s “lemons” problem, whereby, inthe presence of asymmetric information about qual-ity, the average quality that is put up for sale is worsethan the average quality outstanding, is relevant forused securities as well as used cars.

Accounting Rules: Many financial institutions havetreated asset-backed securities as market risks, ratherthan credit risks, because this allows them to deter-mine capital requirements on the basis of their owninternal models. For these risks, they must follow theprinciple of mark-to-market or fair value accounting.Under this principle, declines in market values mustimmediately enter the banks’ financial statements. Incases where markets have ceased to function, thefinancial statements are based on estimates of what

the market prices would be if the markets did func-tion. This accounting regime is very different fromtraditional accounting for assets in the banks’ creditbooks. If the bank declared that it was going to holdthese assets until maturity, it could disregard fluctua-tions in market values. Write-downs would be deter-mined by doubts about future debt service ratherthan market prices. To the extent that market pricesare deemed to provide measures of value that aremore “objective”, less at the discretion of the banker,this difference may be seen as an advantage of fairvalue accounting over traditional accounting forcredit risks.6 However, if market prices are driven bypanic and fears of a liquidity shortage, fair valueaccounting may be providing the bank with thewrong signals on which to base its strategy.

Insufficiency of “free” equity capital: In their questfor high rates of return on equity, many bankinginstitutions had greatly expanded their operationsrelative to their equity base. They had little or noequity in excess of regulatory requirements. Fairvalue accounting and regulatory requirements beingwhat they are, losses from declines in the marketprices of assets required the banks to react immedi-ately, either by recapitalizing or by “deleveraging”,i.e. by selling assets. By deleveraging, they put addi-tional pressure on asset markets and on other insti-tutions that were faced with further price declines.

Insufficiency of regulatory capital: Banks were alsoeconomizing on regulatory capital. Using the model-based approach to determine regulatory capital formarket risks, some of the most sophisticated institu-tions had managed to reduce their equity to a smallfraction of the balance sheet, in the case of UBSsomewhere between 2 and 3 percent. At such lowlevels, it doesn’t take much of an asset price declineto raise doubts about solvency. Even if solvency isnot of an issue, deleveraging needs are likely to byquite drastic.

It is important to appreciate the role of capital ade-quacy requirements in the downward spiral.Whereas past studies had discussed pro-cyclicaleffects of capital regulation in terms of macroeco-


Focus

5 I am treating the breakdown of maturity transformation by con-duits and SIVs as an “independent” surprise. While this breakdownwas triggered by the downgrades of subprime-mortgage-backedsecurities, financing structures of these institutions were sounhealthy that some other shock, e.g. a general increase in interestrates, would have had the same effect.

6 Thus, for a long time in the nineties, Japanese banks did not writedown their loan portfolios even when it was clear that many of theloans in these portfolios were bad; this failure to acknowledge loss-es delayed the cleanup of the crisis and thus contributed to the pro-longed recession in Japan. In the early eighties, US savings andloans institutions kept long-term, fixed-interest mortgages at facevalue in their books, even though, at the double-digit market ratesof interest, the discounted present values of debt service on thesemortgages were much lower. Thereby they concealed the fact thatthey were technically insolvent and gave themselves the opportu-nity to “gamble for resurrection”, with the consequence that thecleanup after 1990 was much costlier than a cleanup in 1981 wouldhave been.


Focus

nomic flows of aggregate demand, corporate rev-enues, corporate borrowers’ debt service, bank prof-its, new loans and aggregate corporate investment, inthis crisis, we have seen pro-cyclical effects of capitalregulation on stocks of assets and liabilities in thebank balance sheets. Asset prices declines thatinduce write-downs automatically eat into the banks’capital. Under a mechanical regime of capital ade-quacy regulation, the bank is forced to sell asset, pos-sibly even to realize book losses that are notmatched by losses in the present values of futurereturns on the securities in question.The bank’s assetsales induce further declines in asset prices and putadditional pressure on other banks.

Financial developments since August 2007 have beendriven by the interplay of price declines in malfunc-tioning markets, fair-value accounting, capitalrequirements, and deleveraging. At times, the down-ward movement has accelerated, sometimes becauseindividual banks ran into trouble, sometimesbecause, once again, interbank markets broke down.On these occasions, liquidity assistance from centralbanks, to individual institutions or to the market as awhole, has brought some relief, but was unable tostop the downward movement as such. Whenevermarket participants and the media proclaimed thatthe worst of the crisis was over, the next problemwould appear on the horizon.

At last, the insolvency of Lehman Brothers had suchrepercussions for other institutions (AIG) and forasset markets worldwide that governments in theUnited States and Europe and, ultimately, the tax-payers have become involved.We can only hope thatthis government involvement will suffice to stop thedownward spiral and to help rebuild confidenceamong the actors in the financial system. This is byno means certain, however.

In the introduction, I asked why the losses from sub-prime-mortgage lending have disrupted the entirefinancial system worldwide. The answer to this ques-tion lies in the interconnectedness of the system andin the interplay of the factors that I have described.Because of these systemic interactions, the twin sur-prises of August 2007 induced a gradual implosion ofmarkets, prices and institutions that even the centralbanks’ interventions could not bring to a stop.

Whose fault was it?

In assessing these developments, it is important todistinguish between individual misbehaviour and

faulty system design. Individual misbehaviour in-volves choices that end up harming the individual orinstitution that have taken them. By this, I do notmean choices that turned out badly ex post becauseof bad luck, but choices whose flaws should havebeen obvious ex ante. Faulty system design involvesinstitutional arrangements and regulatory rules thatlead to undesirable results, for the institutions thatare involved or even the financial system as a whole,when individuals pursuing their own interests aresubjected to these arrangements and rules.The ques-tion of who was responsible arises for flaws in systemdesign as for individual misbehaviour, but it arises atanother level, the level of regulatory design ratherthan any specific actions.

By this logic, the insufficiency of creditworthinessassessments by originating banks in subprime lend-ing would be deemed to be a matter of faulty systemdesign rather than individual misbehaviour. Becausethe originate-and-distribute system of mortgagesecuritization had been designed so that the origi-nating banks had no liability, these banks had no rea-son to spend more than a minimum of resources oncreditworthiness assessments. The fact that they hadno liability, however, was an instance of faulty systemdesign.

As the genesis and the evolution of the crisis havebeen described above, the following instances of mis-behaviour seem to have played a role:

– Investment bankers focussing on growth andmarket shares in securitization ignored risks. Tobe sure, once they sold the securities they wouldno longer be liable. However, in the crisis, thelosses just from “warehousing” securities in theprocess of securitization turned out to be enor-mous.

– All sorts of investors, individuals, private universi-ties, foundations, German public banks, Americanand Swiss investment banks were so much con-cerned about yields that they neglected the asso-ciated risks and failed to ask why a mortgage-backed security with a rating of AAA was payingmore interest than a corporate bond with thesame rating.

– These investors also failed to think through theimplications of liability rules for incentives inorigination and securitization.

– Risk control and risk management in the largeinvestment banks that were involved in securiti-zation and/or holding mortgage-backed securitiesfailed to provide comprehensive analyses of theirinstitutions’ risk exposures from these securities,

taking into account the joint dependence of secu-ritization as a business and of the returns on secu-rities held on the movements of residential real-estate prices in the United States and on the fac-tors that were driving these movements.

– To the extent that credit risks were hedged, riskcontrol and risk management did not pay atten-tion to the possibility that, because these riskswere highly correlated, their counterparties,monoline insurers or institutions like AIG, mightgo under at the very time when they would becalled upon to substitute for the defaulting bor-rowers.

– The rating agencies also failed to develop an ade-quate, comprehensive and timely understandingof the relevant risks and the correlations betweenthese risks taking into account their joint depen-dence on movements in underlying factors.

– Conduits and SIVs were engaging in excessivematurity transformation. Not having any equityworth mentioning, investing in long-term assets,and refinancing through the money market wouldbe a sure recipe for disaster even if the long-termassets were not subject to credit risk but “only” tointerest rate risk. The risks of such maturity trans-formation should have been known to any pro-fessional banker.

– By the same token, the risks that the pledges ofliquidity assistance to conduits and SIVs imposedon the sponsoring banks should have been obvi-ous to any professional banker. As an outsider, Isuspect that these pledges were not compatiblewith large-exposure regulation limiting lending toany one client. Perhaps, the public prosecutorsshould take a closer look at this.

– Finally, mention must be made of the looseness ofmonetary policy in the years 2002 to 2004. Thelow money market rates and steep yield curvesthat the United States had in these years madeborrowing short to lend long appear to be veryattractive and contributed to the push of USinvestment banks into the business of securitizingsubprime mortgages. The Federal Reserve Bankshould have known that this constellation wasbound to make the financial system vulnerable, sothat a reversal of monetary policy would be diffi-cult to achieve without endangering financialinstitutions. After all, this is what happened whenthe liquidity flush of 1988 was followed by therestrictiveness of 1989 and when the expansion-ary policy of the years after 1990 was ended in1994.

The following flaws in system design were also im-portant:

– The decline in the quality of subprime mortgageswas at least partly caused by a lack of incentivesfor originating mortgage banks to spend resourc-es on creditworthiness assessments. Incentiveswere missing because (i) the originating banks didnot carry any liability and (ii) the securitizing pri-vate investment banks – unlike Fannie Mae andFreddie Mac – did not impose strong quality stan-dards of their own.

– The failure of the securitizing private investmentbanks to impose strong quality standards of theirown was due to their not being liable either –unlike Fannie Mae and Freddie Mac. Their failureto provide guarantees was possible because hedgefunds were willing to buy up equity tranches –presumably satisfying the demands of finalinvestors hungry for yields.

– The failure of the securitizing investment banksto impose strong quality standards was also sup-ported by the willingness of other investmentbanks to buy up mezzanine securities for a secondround of securitization (MBS CDOs). Apparent-ly, these other investment banks cared moreabout market shares in the securitization businessthan about the credit risks in mezzanine securi-ties. Incentives for investment banks in the secondround of securitization to impose quality stan-dards were just as deficient as for investmentbanks in the first round of securitization. For theeconomist, it is not evident that the second andhigher rounds of securitization served any usefulpurpose – except of course, to get additionalAAA ratings so as to attract additional investorsthat were required by regulation to invest only inAAA-rated securities.

– The absence of any regulation or supervision forconduits and SIVs and for their relations with thesponsoring banks implied that the extent of matu-rity transformation by these institutions was byand large unknown. When this became known inAugust 2007, it contributed greatly to the shock,perhaps even more than the unexpectedly largedowngrading of the subprime-mortgage-backedsecurities as such.

– Private-sector banks had significant governanceproblems. Internally, they were unable to subjecttheir investment bankers to effective risk control.Externally, in relations with shareholders, ana-lysts, and the media, the mechanisms that support“market discipline” in order to enhance “share-holder value” were biased towards yields with lit-tle concern for risk. When Deutsche Bank wasclaiming that a 25 percent rate of return on equi-ty was the benchmark for modern banking insti-tutions, it raised protests among labour unions


Focus


Focus

militating against layoffs, but not, it seems, byfinancial analysts suggesting that this benchmarkmight be an indication of undercapitalization anddemanding that the bank provide informationabout the risks associated with this benchmark.Moreover, the mechanisms that support “marketdiscipline” in order to enhance “shareholdervalue” do not pay much attention to risks that willbe borne by others than shareholders, i.e. to risksthat affect the banks’ creditors or the taxpayerwhen he is called upon to save the bank.

– German public banks had even greater gover-nance problems. Whereas private investors andbanks may have been suffering from yield mania,these public banks were caught up in a yieldpanic. When interest rates and intermediationmargins are low, and you do not have much of anestablished business model, when the implicitpublic subsidies that you used to get in the pasthave just been outlawed by the EuropeanCommission, where do you get the returns youneed in order to cover your operating costs and tosatisfy the demands of the politicians? For thesebanks, mortgage-backed securities looked like aGod-sent remedy, especially when refinanced at 1percent in the American money market. For thepoliticians sitting in these banks’ supervisorycommittees.

– The portfolio managers and risk managers ofinstitutional investors cannot be blamed for nothaving taken into consideration the system riskexposure that was created by maturity transfor-mation in conduits and SIVs. After all, they didnot and could not know the extent of this maturi-ty transformation. They can be blamed, however,for not having paid enough attention to the possi-bility that there might be a major risk that theirrisk models had not captured. I see this as a prob-lem of system design rather than any specific mis-behaviour. Common experience suggests thatthere are always matters outside one’s horizon ofanalysis; any system of risk management mustaddress such eventualities and consider how tomake provisions for them.

– The preceding point concerns the system of bank-ing regulation as well as the system of risk man-agement at the level of the individual bank. Whenthey allowed the banks to determine their regula-tory capital for market risks exclusively on thebasis of their own quantitative risk models, theregulators – like the bank managers – neglectedthe possibility that important risks might not havebeen captured by the models.

– The very mechanical approach to capital regula-tion that we have under the Basel Accord has

greatly contributed to the interplay between mal-functioning markets, fair value accounting, capitalrequirements, and deleveraging that has driventhe implosion of the system since August 2007. Ifoverall capital requirements had been higher, themultipliers would have been smaller. If the appli-cation of the regulation would have left morescope for discretion with respect to the speed ofdeleveraging, the systemic impact of the regula-tion would have been cushioned even more.

– Banking regulation and supervision must be criti-cized for their lack of systemic thinking. Theytend to think about the solvency of the individualinstitution and the protection of its investors inisolation. Yet, the survival of the institution alsodepends on its systemic environment. The factthat hedge funds, conduits etc. were not subject toany reporting requirements makes sense if wethink about these institutions in isolation andconsider their investors to be sophisticatedenough and important enough to fend for them-selves. It does not make sense if we think in sys-temic terms about the roles of these institutionsas counterparties in the securitization business orabout the impact of their failure on asset pricesand all the institutions that are thereby affected.Deleveraging, i.e. the sale of assets after a loss, canmake sense as a way of adapting the bank’s riskexposure to its reduced equity,7 but it is counter-productive if the induced decline in asset pricesrequires other banks to deleverage as well, withadditional price effects that hit right back at thefirst bank.

At this point, most politicians seem to be agreed thatfinancial supervision must be expanded andstrengthened. However, there are few signs indicat-ing that the politicians appreciate the extent towhich not just the lack of supervision over hedgefunds, conduits and SIVs, but also the very mechan-ics of the system that is imposed on banks has con-tributed to the financial crisis. If we are to prevent arecurrence of such a systemic implosion, we need toaddress this problem as well. For this purpose, theconceptual foundations of banking regulation andsupervision must be altogether reconsidered.

7 Even here, there is a possibility that the realization of lossesthrough a fire sale in a malfunctioning market may destroy the via-bility of the bank in the medium run.


Focus

CONTAINMENT AND

RESOLUTION IN THE FINANCIAL

CRISIS: TOO LITTLE, TOO LATE

PATRICK HONOHAN*

Despite its origins in countries with the most sophis-ticated and experienced regulatory systems, contain-ment and resolution policy in the financial crisis thatbroke out in August 2007 was not a success. Insteadof the prompt corrective action prescribed byresearch-based good practice manuals (e.g. Honohanand Laeven 2005), regulatory authorities allowedstrained credit conditions reflecting weak capitaliza-tion of major banks to persist for over a year withonly sporadic and case-by-case interventions.

For too long, policy authorities misinterpreted the crisisas chiefly one of liquidity rather than solvency. Primereliance was placed on large scale expansion of liquidi-ty. Guarantee programs were also expanded as ifdepositor panic and contagion were important drivers.

The failure to recognize the scale of the solvency prob-lem was partly due to the complexity and opacity ofthe financial instruments that were being used by mar-ket participants, but partly also to regulatory denialand disbelief that such severe undercapitalizationcould have developed under the radar. As a result,identification of, and intervention in, insolvent andundercapitalized institutions was dilatory. All of thebank failures were precipitated by market withdrawalof liquidity; in no case did regulators act pre-emptive-ly to require the banks concerned to recapitalize, eventhough all or most seem with hindsight to have beeninsolvent when run. The very modest fiscal costs com-mitted even by October 2008 show the extent to whichrestoration of capital was left too late.

Prevention

By now, the causes of the crisis are fairly clear(Honohan 2008). Over-confidence in risk-manage-

ment systems had lured banks into acquiring mis-

priced mortgage-backed securities (MBS) which

were much riskier than they seemed, and whose

value depended on property prices in the United

States remaining at the unprecedented levels to

which they had been bid, and on the performance of

sub-prime borrowers who had been sold mortgages

they could not afford to service. Widespread falls in

US residential property prices from mid-2006 and

alarming increases in the share of delinquent mort-

gages made recovery on these MBS increasingly

doubtful. As rating agencies downgraded debt, non-

bank conduits that had been created to fund MBS

became unable to roll-over their borrowings and

some found that their sponsor banks were unable to

provide the expected back-up liquidity, leading to

the first bank failures.

Although the scale of estimated losses from this ini-

tial shock, if uniformly spread across major banks,

should have been manageable, the difficulty in

determining exactly where the losses lay, and the

actual and feared knock-on effects of such losses on

counterparty risk caused investors and bankers to

reassess their risk appetite. Realizing that their risk-

management models had proved inadequate,

bankers became much more risk-averse and this

restricted credit availability. In a classic debt defla-

tion process, liquidation of assets caused by the

credit crunch lowered equity prices and worsened

the solvency even of intermediaries who had no

exposure to the US residential property market.

Other property bubbles, notably in the United

Kingdom and Ireland, also burst, adding to pressure

on exposed banks in those countries. Market con-

cerns increasingly extended to any opaque and

highly leveraged institutions, including the leading

US investment banks and credit insurers.The scram-

ble for liquidity boosted the international value of

the US dollar, and helped puncture the bubble in

petroleum and other commodity prices, which

peaked around July 2008. Several emerging markets

which had been relying on short-term capital

inflows, or on the revenue from the commodity

boom, found their currencies, stock markets and

banks come under pressure.* Trinity College Dublin.


Focus

Containment and resolution policy

Given the extensive contingency planning that hasbeen undertaken over the past decade since the EastAsia crisis of 1997–98,1 and a considerable increasein the staffing of financial regulatory authorities, theperformance of these authorities in the crisis canonly be considered disappointing. It is not just thatthey failed to prevent the crisis. In addition, contain-ment and resolution policy has been deficient. Suchpolicies should have been speedily put in place assoon as the crisis broke out in August 2007.

The conventional tools of containment and resolu-tion policy include (i) the legal powers to (ii) inter-vene in the management of banks, requiring themto increase capital and to desist from unsafe prac-tices, and if necessary to take control of a failingbank and arrange for a sale, liquidation or financialrestructuring with the use of public funds; (iii) lim-ited deposit insurance is used to insulate smalldepositors from anxiety and loss; (iv) liquidity pro-vision, including lender of last resort facilities, isarranged for banks known to be solvent but whoare, for some reason, unable to source ready fundsin the market. With the benefit of hindsight, theoperation of all four dimensions seems deficient,especially the delay in recognizing the need for andensuring recapitalization.

Legal powers

One of the early failures, i.e. that of Northern Rock(NR), was exacerbated by a surprising lack of deci-siveness on the part of the British authorities.Doubts as to whether the exceptional liquidity sup-port which would have been necessary to keep thebank afloat would be legal delayed this assistance,triggering a retail run on the bank fatal for its sur-vival and making its subsequent nationalization andcostly downsizing all but inevitable. Here was a casewhere unrealistic and counterproductive lender oflast resort fundamentalism, of a type which hadbecome popular in central bank charters and policystatements over previous years, was allowed to standin the way of needed emergency action. Though farfrom the largest of the failing banks, NR’s demisebecame iconic.

Almost exactly one year later, on 15 September 2008,the famous investment bank Lehman Brothers wasallowed to go into bankruptcy when no private sectorrescuer could be found to assume its liabilities(because of the opaqueness and complexity of its busi-ness). Some observers assumed that the authorities’decision to allow this bankruptcy reflected their deter-mination to avoid moral hazard, by showing that evena large investment bank could be allowed fail. Thispoint of view assumed that the market was now betterplaced than it had been in March to work through abankruptcy of this scale and complexity. However, it isalso reported (Financial Times, 12 October 2008) thatthe US authorities simply had no legal way of savingLehman in the time available when the prospect of aprivate sector solution evaporated.

Deposit insurance

The design and operation of deposit insurance alsoproved largely deficient in achieving the convention-al goals of preventing contagious runs, ensuring mar-ket stability, and protecting the assets and peace ofmind of retail depositors.

Once again NR provides a crisp example. The UKdeposit insurance scheme in effect at the time coveredonly the first £2000 in full, but just 90 percent of thenext £33,000.With genuine concern about the survivalof NR, fuelled by vague and ambiguous statements bythe authorities, depositors queued outside the branch-es of this bank to withdraw their funds. Clearly, the co-insurance built into this dysfunctional scheme (osten-sibly in an attempt to minimize moral hazard) com-bined with the absence of procedures for a quick pay-out by the deposit protection fund, precluded thescheme from providing any protection against a retailbank run. Indeed, a few days’ of retail depositors’queuing was enough to trigger an official blanketguarantee for NR, followed within months by an over-haul of the deposit protection scheme, doubling theamounts covered and eliminating the deductible(prompt payouts, US style, were still elusive).2

Deposit insurance schemes continued to underper-form as the crisis deepened. Concerns over the prop-

1 This includes the Basel Core Principles for Effective BankingSupervision, the Basel II bank regulation accord and the IMF-World Bank Financial Sector Assessment Program, which hasreviewed national policies in well over one hundred countries –though not yet the United States or China.

2 Amazingly, despite the US FDIC’s enviable record of prompt pay-outs, even fully covered depositors at the failing California bankIndymac stood in line to withdraw their funds in the days before itwas intervened. Although the intervention was smoothly managed,it quickly proved to have been too late (relative to the standardsenvisaged by the prompt corrective action approach of the US sys-tem), with losses to the deposit insurance fund of $9 billion, orabout 30 percent of the bank’s total assets.


Focus

erty-related portfolio of the Irish banks saw interact-ing weaknesses in their equity prices, their CDSspreads, and their access to interbank borrowing(they had become heavily dependent on foreign-sourced funding). With one bank reportedly facingimminent failure, the Irish government announced,on 29 September 2008, a global guarantee of theIrish-controlled banks’ liabilities, including subordi-nated debt, and their deposits in other jurisdictions.Given that these banks have a substantial presencein the United Kingdom (especially in NorthernIreland), and the uncertainty at the time about the condition of several British banks, immediateissues of competition-distorting aid within theEuropean Union arose; the UK government protest-ed. Eventually, the issue became submerged in theextensive generalized support packages in Europe,but the incident highlighted, and not for the firsttime, unresolved issues with deposit protectionacross Europe.

A final deposit insurance incident arose with respectto the two Icelandic banks operating an extensivedeposit-taking business in the United Kingdom.Both came under market pressure in early October2008 and, when Landsbanki was intervened by theIcelandic government, there were indications thatthe Icelandic government might not be able to makethe payments due under the Icelandic protectionscheme to depositors of its UK branches. The Britishgovernment then, in order to safeguard UK interests(but with scant regard for international coopera-tion), seized assets of the other Icelandic bankKaupthing, triggering its failure.

Liquidity policy and the lender of last resort

The three big central banks involved in advancedeconomies that experienced failure, the EuropeanCentral Bank, the US Federal Reserve and theBank of England certainly took energetic steps inan attempt to ease general liquidity conditions.Volumes of secured lending by these central banksto banks and others have expanded to unprece-dented levels, and each bank has revised its operat-ing procedures to increase the maturity and condi-tions, the range of counterparties or the classes ofcollateral accepted.

Each central bank started from a different position.Reflecting long-standing central banking operatingprocedures in the Deutsche Bundesbank, and to an

extent in the Banque de France, the ECB hadalready formalized an elaborate system of collateral-ized lending, which it was able to bring to bear quick-ly.The Bank of England and the Fed were at first lesswell equipped in this respect. The Fed soon began todraw on its broad enabling powers to lengthen loanmaturities, to lend Treasury securities and not justcash, and to lend to investment banks, money marketmutual funds and directly to the non-financial cor-porate paper market, among other initiatives. TheFed also lowered its target interest rates aggressive-ly from over 5 percent to 1 percent, whereas theECB held its target at 4 percent for most of the firstyear of the crisis – and then increased it by 1/4 per-cent concerned by the commodity-driven increase ininflation – before beginning a lowering trend.Ironically, it was thus the ECB – famously reluctantto abandon monetary aggregates as a reliable indica-tor of inflation pressure – that seemed to decoupletwo policy instruments: quantitative easing wasapplied to the liquidity problems associated withfinancial stress, while the cost of funds was the pre-ferred instrument for controlling inflation.

However, it is noteworthy how careful each has beenin trying to limit its credit exposure. The ECB evenannounced a tightening of its collateral rules in earlySeptember 2008, as it feared that it was being pre-sented with the riskiest qualifying collateral by bor-rowing banks.

As individual banks began to fail, central bankssought to remain aloof from anything that wouldentail significant credit risk. So far, the ECB hassucceeded in doing so, and has not made any specialbank rescue loans so far. The Bank of Englandstarted with the same attitude, but political consid-erations came into play and resulted in the hugeBank of England loan to NR, a bank which couldnot be considered systemically important. (Some£3 billion of this loan has already been assumed bythe UK Treasury and converted into equity in NR,thereby considerably reducing the Bank of Eng-land’s exposure.)

The Fed has been directly involved in two specialcredit arrangements for institutions deemed too sys-temically important to fail, the investment bankBear Stearns and the insurance company AIG. InMarch 2008, Bear Stearns was acquired by the com-mercial bank JP Morgan Chase, but a part of its port-folio valued at $30 billion was removed and placed ina special purpose vehicle with JPM taking a first loss


Focus

investment of $1 billion and the Fed financing therest. In September 2008, the Fed, despite having justallowed the systemically important LehmanBrothers to go into bankruptcy, decided to give AIG3

a special $85 billion loan at a high penalty rate ofinterest, and received warrants for 80 percent of itsequity in return.

Although the central banks may regard themselvesas having acted very energetically in the crisis, it isthus striking that they have largely retained theoperational independence necessary to protect theircurrencies from being debased in this crisis as hashappened in so many crises of the past, especially indeveloping and transition economies in the 1980sand 1990s.

In implementing their liquidity policies, though, thecentral banks have sometimes struggled in this turbu-lent environment to achieve the target short-termmoney market rates with precision. And as has beenwidely discussed, three-month interbank rates forunsecured lending have been persistently higher thanthe market’s expectation of future overnight policyrates for the following three months (by about 1 per-centage point) since August 2007, with the gap jump-ing to about 31/2 percentage points in Septem-ber–October 2008. In this respect, the central banks’liquidity policies cannot be considered wholly suc-cessful either.

Recapitalization

Heavy credit losses alone have meant that, in orderto sustain the previous level of activity, banks havehad to raise additional capital from existing or newshareholders. Furthermore, it is clear that banks willno longer be able to conceal their true leverage andevade capital requirements by pushing business off-balance sheet. A re-evaluation of the riskiness ofmuch of their credit business also implies a need foradditional capital. At first, major banks were able toraise replacement capital, partly because the earlyreported losses from sub-prime related securitiesand the like seemed like a once-for-all event thatneed not imply limited future profitability. But ofsome $600 billion in credit losses reported by majorbanks by the end of September 2008, only about two-thirds had been replaced. Worsening global econom-

ic prospects, and renewed uncertainty as to the loca-tion and scale of hidden credit losses (triggered,among other things, by the Lehman bankruptcy)added to the difficulties of raising new bank capital.Although bank shares had fallen dramatically, theystill seemed risky to such remaining investors asSovereign Wealth Funds. Clearly, a full resolution ofthe crisis was going to require some new sources ofcapital. With the central banks largely declining thatrole, the task fell to governments, and they have beenslow to accept the challenge.

From the start of the crisis some governments hadbeen prepared to step in on a small scale. The twoGerman banks that failed early on were actuallyowned by government entities and eventually it wasthe sponsoring governments that made the maincapital injections to make these entities whole againand allow them to be sold back into the market. Ashas been noted, the British government eventuallynationalized Northern Rock (though the terms ofcompensation for shareholders have not yet beendetermined.) Likewise, the two large government-sponsored mortgage finance companies Fannie Maeand Freddie Mac, victims especially of the falling res-idential property prices, were taken into conserva-torship in early September 2008 with substantialactual and promised injections of government funds;the US government took an 80 percent ownershipstake and fully guaranteed their liabilities, endingtheir ambiguous status.

The last days of September and the beginning ofOctober saw a succession of failures in Europe, each ofwhich was dealt with in an ad hoc manner involvinggovernment capital injections or guarantees. A novelpattern of wholesale bank runs was emerging. Onceidentified as weak, a bank would not only be starved ofwholesale funding, and see its share price plummet-ing,4 but would also experience a spike in the premiumfor CDS on its longer-term borrowings.An alternativeto the long-advocated use of compulsory subordinateddebt issuance as a discipline on bank managements,the CDS swap rates served as a conspicuous form ofmarket discipline – though it remains a matter forfuture research as to how accurate and efficient thismarket has been through the crisis.

It was only at this point that it became evident to theauthorities that ad hoc recapitalizations of systemi-cally or politically important banks were no longer

3 Much of AIG’s difficulties relate to credit default swaps (CDS) –in effect insurance policies against creditor default – which theyhad written.

4 Emergency limitations on short-sales of bank equity do not seemto have had a lasting effect on their stock prices.


Focus

an adequate response. Faced with the prospect thateach of its major banks would take its turn as thefocus of this triple interacting pressure, the pressureon each government to adopt a pre-emptive system-wide approach intensified. The effect of sharpincreases in interbank rates and the inability of manybanks to mobilize funds were also beginning to befelt acutely as a credit crunch in the nonbank sector.Market fears that considerable further lossesremained embedded but hidden in the balancesheets of most banks had to be taken seriously. Onlya restoration of undoubted creditworthiness of allthe main banks could forestall a damaging interrup-tion in the functioning of banking generally in theaffected countries. This was the genesis of theapproach endorsed by the Euro plus UK summit inParis on October 12, 2008, which saw governmentsestablish funds to acquire equity or preferenceshares in banks, as well as offering to guarantee – fora fee – medium-term interbank borrowings.

The introduction of a broad-based scheme of en-couraged and assisted capital increases in many ofthe advanced countries most affected by the crisisare the most novel of the containment and resolu-tion policies of this crisis. Indeed, it has no precedentin the past half century. Alas, it came late in the cri-sis, by which stage a collapse of confidence in finan-cial and non-financial circles had occurred.

If we take the UK scheme as the ur-case, its hall-marks are the announced availability of a very largesum of public money available for capital combinedwith pressure on all major banks to achieve a muchhigher level of capital.Three large banks were quick-ly signed-up for a capital injection, which gave thegovernment a 40 percent equity stake in two of them(HBOS and Lloyds, which had already announcedtheir intention to merge), and a majority 60 percentstake in a third, RBS. The other major UK banksindicated that they would secure additional capital inthe market, thereby avoiding onerous side-condi-tions, including restrictions on dividend payments.The capital scheme was accompanied by theannounced availability of guarantees for bank medi-um-term borrowing.5 The logic of the scheme wasthat it would both (i) restore market confidence inthe solvency and creditworthiness of banks, therebypromising to unfreeze the interbank market amongthe major players, and (ii) rebalance the banks’incentives to re-engage in credit markets but on a

safe and sound basis. The borrowing guarantee wasdesigned to ensure that banks could obtain funds forlending on, even if interbank markets were slow toreturn to normal.

Other European countries followed suit within days,albeit on a much smaller scale. France arranged forstate capital injections into its four largest banks, butthe German scheme saw no immediate take-up bythe largest banks, possibly because of the very limit-ing restrictions on executive pay built into theGerman scheme. Insofar as one goal of this strategywas to make acceptance of government capital freeof stigma, the slow take-up in Germany was a disap-pointment. In the United States, about a third of the$700 billion fund painfully negotiated throughCongress was now earmarked for capital injectionswith the original distressed asset purchase compo-nent temporarily pushed into the background, aspolicymakers recognized that solvency worriesextended beyond the issue of the complexity ofmortgage-backed securities. Four large US banksreceived injections of $25 billion each, in this case inthe form of preference shares, with smaller sumsgoing to numerous other banks.

The systemically encouraged and assisted bank cap-ital program is unusual in that it does not presup-pose that the bank is insolvent or undercapitalized.Accordingly, unlike what is normally recommendedfor bank recapitalizations, and what was done formost of the failing banks up to that point, it doesnot entail a write-down of existing equity orremoval of management (though the CEO of RBSdid not survive). As an exceptional measure inexceptional market conditions, where bank sharesare trading at distress levels way below book value,this may be alright. But the moral hazard entailedin a standing scheme with such features is consider-able indeed.

It is too early to judge what the eventual fiscal costof the assistance to weak and failing banks in thiscrisis will be. For the interventions made before thesystemic programs began, even a relatively pes-simistic calculation arrives at a total sum of about$400 billion, or a little over 1 percent of the com-bined GDP of the European Economic Area andthe United States. This is an extraordinarily low fig-ure, when compared with crises of the past andabroad,6 and confirms how little was done on recap-

5 The fee charged to the banks for these guarantees was varied inproportion to the average CDS spreads experienced by each bank.

6 The median fiscal cost of 78 systemic banking crises has been over15 percent of GDP (Honohan 2008).


Focus

italization during the first 14 months. Since then,the systemic schemes have expanded the grossexposure in respect of guarantees, equity injectionsand special loans in these countries to several tril-lion dollars. Yet, if the underlying premise, namelythat the market has been too pessimistic and thebanks tapping these funds will prove to be solvent,then (considering also the guarantee fees that arebeing charged) they could even prove to be a prof-itable investment. The huge sums being lent by thecentral banks may also have embedded some cred-it losses, but the additional state equity injectedwould be there to absorb those losses ahead of thecentral bank. Of course the crisis has triggered arecession with economic costs that are much larger,but for this crisis it seems that the fiscal authoritiesmay get off relatively lightly from their direct assis-tance to weak and failing institutions.7

Concluding remarks

Fifteen months in, it would be hard to claim that thecrisis had been effectively contained, let aloneresolved. The biggest flaw in the policy response hasbeen to underestimate for too long the capital needsof the major banks.

Policy in recent banking crises across the world hasbeen criticized for generating moral hazard – thecareless behaviour of those who know themselves tobe insured. But the heavy losses incurred by (i) bankshareholders (for example in AIG, Fannie Mae andFreddie Mac, IndyMac, Fortis and several banks thatdid not fail, but saw their share price tumble by80–95 percent), (ii) debtholders (in Wachovia Bankand Lehman Brothers, among others), and even (iii)depositors (in the Icelandic banks) means that thepolicy response on average cannot be considered tohave greatly aggravated moral hazard. Even theblanket guarantee for the Irish banks has come atthe cost of an annual premium paid to governmentwhich amounts in aggregate to about 8 percent ofthe market capitalization of the institutions con-cerned. True, few of the guarantee schemes have fol-

lowed the UK practice of charging a premium dif-ferentiated by risk, which might have a better effecton discipline.

Earlier recapitalization might have shortened theperiod of credit crunch, thereby lessening investoranxieties and forestalling the global recession. Butmaking this happen would have required not only analtered mind-set by regulators, but a way of convinc-ing fiscal authorities to commit such huge resources.The more dramatic events of September 2008 maythus have been a necessary evil to prepare the polit-ical ground for the required policies.

At the time of writing, the credit famine is extendingto developing countries which had become depen-dent on foreign borrowing, including Ukraine,Hungary and Pakistan. Although some banks inthose countries have been put under pressure fromthese events, they were not triggered by the domes-tic banking system, and call for a rather different setof policy responses of the type traditionally negoti-ated by the IMF.

The failure of regulatory authorities to appreciatethe scale and depth of the solvency problem reflectstheir general reliance, shared with the industry, onmechanical risk management models for assessingthe risks to intermediaries. While regulators warnedof vulnerabilities and optimistic risk-pricing, they didnot fully trust their pessimism in this regard. As aresult they were more or less as surprised as theindustry to observe the unfolding events, andremained slow to act to forestall the entrenchmentof the credit crunch.

References

Honohan, P. (2008), “Risk Management and the Costs of theBanking Crisis”, National Institute Economic Review 206, 15–24.

Honohan, P. and L. Laeven (2005, eds.), Systemic Financial Distress:Containment and Resolution, New York: Cambridge University Press.

7 Conspicuously absent from policies adopted so far has been assis-tance for distressed bank borrowers. Given the heavy deadweightcosts of repossession and sale of a mortgaged residential property,several schemes have been suggested for arriving at a mutuallybeneficial write-down of some delinquent mortgage debt. Thiswould require legislation if it is to work for securitized mortgagesbecause of the difficulty of identifying and obtaining agreementfrom all of the relevant creditors. Clearly there are many deservingborrowers who were missold subprime mortgages and deservesome preference here. In general, though, the difficulty of provid-ing borrower relief on a fair but restricted basis militates againstsuch action.

THE GLOBAL FINANCIAL

CRISIS: CAUSES, ANTI-CRISIS

POLICIES AND FIRST LESSONS

MAREK DABROWSKI*

When the US subprime mortgage crisis erupted insummer 2007 few people expected that it could hitthe entire world economy so hard. One year laternobody had already doubts that we faced a financialsystem crisis on the global scale with dramaticmacroeconomic and social consequences for manyregions and countries. Few local analysts and politi-cians would dare claiming that their country isshielded from the financial crisis effects as the stormunfolds worldwide. Left to the unknown is its scale,sequencing, distribution effects between regions andcountries and the consequences for the future archi-tecture of the financial system. As critical as thequality of the day-to-day management of the crisis isthe understanding of what has happened and whatmay happen, whence the need for an interim diagno-sis, even one incomplete and involving a certain mar-gin of misperceptions and wrong interpretations.

This short commentary tries to analyze the extent towhich monetary policy and financial market regula-tion failures bear responsibilityfor the eruption of this crisis andits further spread, the zigzags ofanti-crisis management, crisisimpact on emerging marketsand, finally, to present some ten-tative lessons for policymakers.

Monetary roots of the currentcrisis

The primary causes of the cur-rent financial crisis are imput-

able to lax monetary policies conducted by the USFederal Reserve Board and other major centralbanks (e.g. the Bank of Japan) from the mid-1990s.Enjoying record-low inflation and low inflationaryexpectations, central banks reverted to more inten-sive fine-tuning in order to avoid the smallest risk ofrecession. As a result, the Fed aggressively reducedits interest rate three times over the last ten years(see Figure 1), starting with the series of crises inemerging markets (Mexico, South-East Asia, Russia,the pre-crisis situation in Brazil) and the Long-TermCapital Management troubles in the United States atthe end of 1998. This was followed by the 2001-2002post-9/11 drastic interest rate cuts, down to 1 per-cent, and the bursting of the dot.com bubble. Onboth occasions, the Fed provided relief to troubledfinancial institutions, helping to circumvent (1998)and reduce (2001) the danger of a US recessionwhile fueling global economic growth. The thirdintervention occurred in the wake of the current cri-sis (end of 2007 and beginning of 2008): the federalfunds rate was reduced from 5.25 percent to 2 per-cent within a few months and then to a low of 1 per-cent in November 2008.

Fearing recession and deflation (in the early 2000s),the subsequent tightening of monetary policy alwayscame too late. Such an excessively lax Fed attitudecontributed to a systematic building up of excess liq-


Focus

* The CASE – Center for Social andEconomic Research, Warsaw. This is arevised and updated version ofDabrowski (2008).

0

1

2

3

4

5

6

7

1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008

FED S FEDERAL FUND RATE

in %

Source: Federal Reserve Statistical Release.

,

Figure 1


Focus

uidity both in the United States and the world.Distracted by the supply shock flowing from eco-nomic reforms and market opening in China, Indiaand in other developing and transition countries,many policymakers and analysts were misled by thetemporary lack of visible inflationary consequences.The Uruguay Round, especially the Agreement onCotton and Textiles, and the ensuing liberalization ofworld trade further exerted downward pressure onprices in the manufacturing market.

However, the excess liquidity had not vanished andbrought on three asset bubbles: one in the real estatemarket (primarily in the United States but also inseveral European countries such as the UnitedKingdom, Ireland, Spain, and Iceland, the Balticcountries and Greece), a second in the stock marketand a third in the global commodity markets. Thesebubbles had to burst sooner or later.

Serious macroeconomic concerns had also started tosurface, namely an increasing current account deficitin the United States, and recently, mounting globalinflationary pressures triggered by rising commodityprices (seen as an external price shock by nationalmonetary authorities in individual countries), rapid-ly growing official international reserves in manyemerging market economies and a depreciating USdollar.

Regulatory failures

Monetary policy is not the sole culprit. A large shareof responsibility weighs upon regulations and regu-latory institutions that lagged well behind rapidfinancial market developments. Two major inconsis-tencies are particularly apparent when looking atinstitutional issues:

• the global character of financial markets and thetransnational character of major financial institu-tions as opposed to the national nature of finan-cial supervisory institutions (even inside the EU);

• the increasing role of financial conglomeratesoperating in various sectors of the finance indus-try and the innovative, cross-sectoral financialinstruments versus the sectoral segmentation offinancial supervision; only a few countries canboast of consolidated financial supervision. TheUnited States presents additional institutionalpeculiarities with two levels of responsibilities(federal and state) for financial supervision.

The blame should also be borne by rating agenciesand supervisory authorities that failed to under-stand the nature of innovative financial instrumentsand that provided excessively short-sighted riskassessment by not taking sufficiently into accountthe actual risk distribution in the long intermedia-tion chain between the final borrower and creditor,thus underestimating the actual risk. The same rat-ing agencies which granted excessively positivegrades to financial institutions and individual finan-cial instruments in times of boom hastily started todowngrade their ratings at the time of distress,adding to market panics.

Precautionary regulations, usually meant to enhancethe safety and credibility of financial institutions,such as capital-adequacy ratios (especially whenassets are risk-weighted and mark-to-market priced)or tight accounting standards related to reserve pro-visions against expected losses, also unveiled theirperverse effect as they led to sudden credit stops andmassive fire selling of assets. They proved to bestrongly pro-cyclical, especially as the crisis hadalready erupted.

Zigzags of crisis management

The crisis management roved chaotic and centeredon calming nervous financial markets in the short-term rather than addressing fundamental challengeslike the massive insolvency of financial institutions.The lack of international institutions able to managemacroeconomic and regulatory policy coordinationvery often led to hasty national actions as exempli-fied by the Irish government’s unilateral decision toprovide full deposit guarantees, prompting other EUgovernments to follow suit or by the Iceland-UKconflict over cross-border deposit guarantees.

The drastic cuts in Fed rates at the end of 2007 andat the beginning of 2008 are another instance of ashort-sighted unilateral policy. The cuts added toinflationary pressure and the commodity marketsbubble worldwide. And when combined with theappreciation of the euro and the yen, it exportedthe risk of recession to Europe and Japan while fail-ing to restore domestic US financial market confi-dence, as demonstrated by increasing spreads andperiodic liquidity crunches. The initial diagnosispointing to liquidity rather than solvency as the cri-sis’ raison d’être now appears to have been erro-neous. Indeed, US authorities wasted time and

potential ammunition needed at the moment and inforthcoming months on suboptimal monetary andfiscal interventions (interest rate cuts and broad-based tax rebates) in order to stimulate the econo-my and provide more liquidity rather than concen-trating their resources on fixing the insolvency offinancial institutions.

The belated and costly interventions of some gov-ernments to rescue their financial sector look con-troversial to many. These doubts are at least partlyjustified. On the one hand, rescue plans represent anadditional burden on taxpayers, though part of thecurrent recapitalization costs may be recovered bysubsequent privatizations. Those countries, howev-er, with an already high debt to GDP level must par-ticularly be cognizant of the limits of their fiscalinterventions as they are prone to illiquidity andinsolvency.

Furthermore, the prospect of worldwide economicstagnation/recession adds to potential fiscal stress inmany countries. Thus, fiscal policy requires a verycareful approach. While rescuing large insolventfinancial institutions may be sometimes unavoidableat least in short term (see below), the idea of using alarge-scale fiscal stimulus to overcome reces-sion/stagnation must be treated with a large dose ofskepticism. The experience of Japan, which tried tofight the post-bubble recession in the 1990s withaggressive monetary easing and large-scale fiscalstimulus, should be studied very seriously. Japan’s fis-cal activism failed to overcome stagnation, but con-tributed to building up the large public debt(175 percent of GDP in 2006). In this context, therecent IMF call for global fiscal expansion is contro-versial (IMF 2008).

Whether government intervention is sufficient toguarantee market confidence, considering govern-ments’ failure to avoid the crisis and provide an ade-quate response right from the onset, is a legitimatequestion to ponder. In general, private sector andmarket-oriented solutions, like arranging the take-over of a bank in trouble by a new private investor ifavailable at a given time, will always prove a bettersolution than its nationalization. The bottom line isthat the current crisis cannot serve as the excuse forturning to government interventionism and state(public) ownership of financial institutions as a long-term solution.

On the other hand, as learned from the GreatDepression, governments must intervene in large-

scale financial crises to prevent a systemic bankingcrisis and total collapse of the financial system and aresulting deep recession spiral. This lesson seems tobe well understood by contemporary policymakers(others analogies referring to the early 1930s are notalways correct). The very nature of financial institu-tions – a high level of leverage and mismatchbetween their assets and liabilities (borrowing shortin order to lend long) – makes them extremely vul-nerable in times of distress and confidence crises.The collapse of one large bank or investment fundmay cause a far-going chain reaction as was experi-enced recently after the bankruptcy of LehmanBrothers. Hence, a government rescue of troubledfinancial institutions cannot be compared to the bail-ing out of loss-making non-financial corporations.Regarding moral hazard, it is difficult to expect gov-ernment bail-outs to reward irresponsible bank man-agers and owners because they are already runningout of business.

How is the crisis spreading to emerging markets?

Amidst shattering hopes of ducking the side-effectsof the current financial crisis, emerging marketeconomies are nonetheless feeling its blow. From2006 onwards, these economies have experiencedrising inflationary pressure resulting in numerouseconomic and social problems. This pressure isunlikely to subside quickly, even if the price ofsome commodities has started to decrease and theUS dollar has recovered in recent weeks. Moreover,a slower world economy means a weaker demandfor many commodities, as well as investment andconstruction-related products. Plus, the global cred-it crunch and liquidity problems of many transna-tional corporations have already led to net capitaloutflows from emerging markets, halting newinvestment projects. Finally, banks in many emerg-ing market economies are vulnerable to a globalliquidity crunch due to short-term internationalfinancing exposure and risky lending practices. Putotherwise, new waves of crises in emerging marketeconomies appear rather unavoidable, much moreso in countries that have not built up sufficientinternational reserves or have not run fiscal sur-pluses. Ukraine, Hungary, Belarus and Pakistan, forinstance, have already filed for IMF emergencysupport.1


Focus

1 The same concerns Iceland which does not belong to the group ofemerging-market economies.


Focus

These new crisis cases are very telling. Hungary’stroubles were caused by the combination of bank-ing sector fragility, excessive fiscal deficits and pub-lic debt, and a long-lasting lack of political consen-sus to conduct the necessary fiscal adjustment. Inthe case of Ukraine, there was a sudden collapse ofmetal prices (the main export of this country),excessive exposure of the banking sector to inter-national short-term financing and domestic politi-cal turmoil. Belarus, the country with a closed econ-omy and financial system, became the victim of itsreluctance to introduce market-oriented reforms,which made the country unable to resist a negativeterms-of-trade shock. The same factor – a negativeterms-of-trade shock caused by higher oil and foodprices – combined with a high fiscal deficit anddomestic security problems (conflict in the borderarea with Afghanistan) led to the crisis situation inPakistan.

The number of emerging-market applicants for IMFemergency assistance may be expected to increase incoming weeks and months when the effects of theglobal economic slowdown and a bursting commod-ity bubble will spread to the developing world.Looking ahead, once the crisis is over, “easy” moneycan hardly be expected to return to emerging mar-kets regardless of the quality of their macroeconom-ic policies, business climate and political risk. Toregain credibility and attract investors, the recentlyreceding demand for economic and institutionalreforms may well be back on the agenda.

The first lessons

Although it is too early for final conclusions fromthe ongoing crisis episode, some lessons can alreadybe drawn. Though not a new notion, the first lessonis that monetary policy cannot be excessively pro-active and too much engaged in anti-cyclical fine-tuning. Its involvement must be symmetric, i.e. mon-etary policy must not only stimulate an economyduring difficult periods but it must also be able totighten early enough when the danger of overheat-ing looms on the horizon. Risky ideas such as “therisk management approach” advertized by AlanGreenspan in the early 2000s (Greenspan 2004) andgoing well beyond the classical central bank man-date must be abandoned.

A short-term focus on inflation targeting limited tothe consumer price index is of no avail without mon-

etary authorities tracking more carefully the mone-tary aggregates and asset markets. This means thatconceptual and methodological approaches to thisinnovative and increasingly popular monetary policystrategy require rethinking.

There are two other, more fundamental dilemmasfacing monetary authorities, which should be dis-cussed again in the context of the current crisis expe-rience. First, in an era of globalization, no nationalmonetary policy can be entirely sovereign; even thebiggest central banks (the Federal Reserve Board,the European Central Bank and the Bank of Japan)must take into account the external macroeconomicenvironment and the potential consequences of theirdecisions on others. Thus the question whether coor-dination of monetary policy among major centralbanks, which would go beyond spectacular joint ratecuts (as that of 8 October 2008) or the joint emer-gency liquidity interventions, is possible and desir-able must be discussed again. The closer coordina-tion could perhaps minimize the frequency of globalfinancial shocks, such as sharp changes in exchangerates between major currencies, and secure a stableglobal liquidity management.

While nothing close to a Bretton Woods system or anew monetary order is likely to emerge in the nearfuture, an institutional framework ensuring effectiveinternational cooperation in the sphere of monetarypolicy is urgently called for. The IMF, which couldhave well served this purpose as it did under theBretton Woods system, was downsized and weak-ened recently (obviously prematurely) to the extentof undermining its policy coordination mission inmonetary and fiscal affairs.

The limited sovereignty of national monetary policyis even more obvious in case of small openeconomies where central banks’ ability to “leanagainst wind” is even more restricted. The choice ofthe optimal monetary/exchange rate regime (eitherone of the “corner solutions” or the hybrid one), sohotly discussed at the end of the 1990s and early2000s and then forgotten for a while during an extra-ordinary calm on financial markets, will be placed onthe policy agenda again.

The second fundamental question concerns thedegree of central banks’ responsibility for the stabil-ity of the financial market and banking system. Thecurrent crisis tends to corroborate that shifting toomuch responsibility to central banks, as happened in

the United States and the United Kingdom, compro-mises their anti-inflationary mission. In contrast, aclearer and more straightforward anti-inflationarymandate of the ECB, backed by its strong legal inde-pendence, has limited its involvement in rescue oper-ations of the troubled financial institutions andforced governments of Eurozone countries to takethe lead in fixing financial sector problems.

More thought need to be given to financial regula-tion, financial supervision and rating agencies. Howshould they respond effectively to financial innova-tions, financial conglomerates, and cross-bordertransactions? How should they assess the variouskinds of risks on a long-term rather than a short-term basis? How should they mitigate the creditboom in times of prosperity and the credit crunch intimes of distress?

The question of effective international coordinationof financial regulation and financial supervision iseven more pressing than monetary policy coordina-tion because of the global character of the financialindustry. Again, the IMF can play an important rolehere but this requires a strengthening of its institu-tional mandate and operational capacity.

On a European level, the crisis revealed a similarparadox. In spite of a Single European Market(including its financial sector component) and a sin-gle currency, there is no European financial supervi-sion per se and no fiscal scope for joint rescue oper-ations. Resistance to future financial storms inEurope will depend on how these shortcomings areaddressed.

References

Dabrowski, M. (2008), The Global Financial Crisis: Causes,Channels of Contagion and Potential Lessons, CASE Network E-Briefs 7, /21929649_E-brief_072008.pdf.

Greenspan A. (2004), “Risk and Uncertainty in Monetary Policy”,American Economic Review, Papers and Proceedings 94, 33–40.

International Monetary Fund (IMF), IMF Urges Stimulus as GlobalGrowth Marked Down Sharply, World Economic Outlook Update,6 November,http://www.imf.org/external/pubs/ft/survey/so/2008/NEW110608A.htm.


Focus


Focus

THE TSUNAMI: MEASURES OF

CONTAGION IN THE 2007–2008CREDIT CRUNCH

MARDI DUNGEY*

The credit crunch of 2007–2008 is the most substan-tial financial crisis since those that preceded theGreat Depression in the United States and thebanking crises prior to the First World War. For thefirst year of its development the crisis seemed toremain relatively constrained to developed mar-kets, but after the dramatic events of October 2008it has clearly become a truly global phenomenon,enveloping emerging and developed financial mar-kets and likely to result in recession and substantialslowdowns in most regions. Comparisons with theAsian crisis of 1997–1998 have fallen by the way-side and have made a mockery of the economic pol-icy advice given to developing markets in the throesof financial crises in the last 20 years. TheWashington consensus advised widespread fiscalreform, the liberalization of financial services toexternal investors and the removal of leaders offailing financial institutions. Policy responses to thecredit crunch to date have instead promoted loosercredit conditions, fiscal stimulus, widespreaddeposit guarantees, reregulation, mergers andnationalization of financial institutions. The spreadof the credit crunch has not been caused by a lackof institutions or poor institutional design ormacroeconomic policy, as per standard crisis theory,see for example the overview in Flood and Marion(1999), but rather seems to have originated in poorassessment of risks and less than transparent inter-linkages between financial institutions as a result offinancial innovation. The very products intended todiversify risk now appear to have simply trans-formed it to liquidity and counterparty risk.Network models of financial markets may providegreater insights to the current crises than tradition-al crisis models (see Allen and Babus 2008).

The credit crunch has been deepening for a remark-

ably long time. Arguably a good starting point for

tracking the emergence of the current difficulties is

the problems reported with Bear Stearns hedge

funds in July 2007 (there were some earlier warnings

of this in earnings downgrades for Bear Stearns the

previous month). Liquidity at the short-end of the

market began to evaporate, leading to serious fund-

ing difficulties for a number of financial firms. The

most dramatic indication of the difficulties spreading

beyond the United States was the run on Northern

Rock in September 2007, halted only by the guaran-

tee placed by the UK government on deposits in this

institution. From here things proceeded to worsen,

punctuated by revelations of increasing subprime

mortgage associated portfolio losses and rating

downgrades for financial institutions, the formal

nationalization of Northern Rock in February 2008,

the takeover of Bear Stearns by JP Morgan Chase in

March 2008, repeated central bank interventions to

increase liquidity in the market, the lowering of offi-

cial interest rates, and rights issues to raise funds for

beleaguered financial institutions – for example the

Royal Bank of Scotland and UBS. In September and

October 2008 the pace of events increased, encom-

passing the rescue of Fannie Mae and Freddie Mac

in the United States, the forced sale of Washington

Mutual to JP Morgan, the bailout of Dexia in

Europe, the non-rescue and subsequent bankruptcy

of Lehman Brothers, the takeover of HBOS by

Lloyds TSB in the United Kingdom, of Merrill Lynch

by Bank of America and Wachovia by Citigroup, the

rescue of AIG, the nationalization of Bradford and

Bingley in the United Kingdom, and part-national-

ization of Fortis in Europe.A further consequence of

this turmoil was fear of loss of confidence in the

banking system, prompting the Irish and Australian

governments to guarantee bank deposits, which in

Australia led directly to instability in the non-depos-

itory financial sector with the suspension of investor

access to a number of institutions. This period has

also marked the realization of serious problems in

emerging markets and the re-emergence of interna-

tional lending packages including funding from the

IMF, such as to Hungary in October 2008, and at the

end of October the release of a formal IMF short-* University of Tasmania, University of Cambridge and AustralianNational University.

term liquidity facility to address the needs ofeconomies with otherwise sound policies facing liq-uidity difficulties.

The spread of the credit crunch has been widespreadand dramatic. When crises spread across differentfinancial markets it may be for a number of identifi-able reasons. Common factors may lead to simplyincreased volatility in multiple markets (see Pericoliand Sbracia 2003; Forbes and Rigobon 2001). Tradeand direct economic or financial linkages such ascommon banking structures, lead to the existence ofspillovers or fundamentals-based contagion(Dornbusch, Claessens and Park 2000; Kaminsky andReinhart 2000). In addition, new linkages betweenmarkets may emerge, and it is this transmission ofvolatility in addition to the common factor andspillover channels that is labeled contagion.

Contagion effects have been documented in previous crisesboth between geographical mar-kets and across financial assetclasses. While the majority ofexisting work documents conta-gion across borders for a singleasset class or transmissions withina single country (e.g. Forbes andRigobon 2002; Granger andHuang 2000), a growing body ofwork considers cross-asset classcontagion effects (see Dungeyand Martin 2007; Hartmann,Straetmans and de Vries 2004).

The current paper adapts themethod developed in Dungeyand Martin (2007) to establish ameasure of the extent of conta-gion that has been experiencedthus far in the credit-crunch of2007–2008.

Characterizing the credit crunch

To examine whether the creditcrunch is associated with conta-gion effects, a reference period ischosen beginning 1 July 2004.This captures the point in thebusiness cycle when the Fed hadreturned to a tightening mone-tary policy stance (the Federal

Funds rate was raised by 25 basis points to 1.25 per-

cent on 30 June 2004, following 12 months at its low

point of 1 percent). The period from that point until

the beginning of the credit crunch was relatively

benign and in historical terms was a relatively low

volatility, low inflation, strong growth period. The

credit crunch period is set to date from 17 July 2007,

when Bear Stearns informed investors of its failing

hedge funds (problems at Bear Stearns had been evi-

dent somewhat earlier as a result of downgraded

expected earnings results in May, and their subse-

quent realization in mid-June). The demise of Bear

Stearns in the United States and nationalization of

Northern Rock in the United Kingdom have become

the defining features of the first 6 months of the

credit crunch.

This paper considers the interactions between

money markets and the major equity market indices


Focus

4 000

4 500

5 000

5 500

6 000

6 500

7 000

2004 2005 2006 2007 2008

200

240

280

320

360

400

2004 2005 2006 2007 2008

1 000

1 100

1 200

1 300

1 400

1 500

1 600

2004 2005 2006 2007 2008

3 000

4 000

5 000

6 000

7 000

8 000

9 000

2004 2005 2006 2007 2008

EQUITY MARKET INDICES IN LOCAL CURRENCIES

1 July 2004 to 30 September 2008

3 000

3 500

4 000

4 500

5 000

5 500

6 000

6 500

7 000

2004 2005 2006 2007 2008

Vertical line marks 17 July 2007

Source: Datastream.

SP500 FTSE

DAX NIKKEI

ALLORD

Figure 1


Focus

during the credit crunch compared with the imme-

diate prior reference period. The dataset consists

of 3 month money market rates and major equity

indices representing 5 countries, all data being

drawn from Datastream. The selected economic

regions are the United States, Britain, Europe,

Japan and Australia. The choice of developed mar-

kets reflects the fact that the credit crunch was until

recently relatively confined to these markets. The

United States and Britain were the first markets to

be seriously affected. From 30 June 2007 to

30 September 2008, the FTSE100 declined by

26 percent and the S&P500 by 22 percent. The diffi-

culties in obtaining interbank funding have been

particularly apparent in these economies – leading

to numerous takeovers of insolvent financial institu-

tions by authorities (the nationalization of Northern

Rock, organised takeovers of Bear Stearns and

Countrywide and governmenttakeover of Fannie Mae andFreddie Mac being prominent),the collapse of Lehman Bro-thers worldwide and the strate-gic move of Goldman Sachs andMorgan Stanley to regulatedbank status. Figure 2 shows theyields on 3 month Libor and3 month US Treasury bills overthe period. Excess demand forUS Treasuries has led to thelowest yields since World War IIbeing recorded during Sep-tember 2008.

The situation in Europe from30 June 2007 to end September2008 had been less dire, althoughstill difficult. Problems in theEuropean banking sector seemedinitially more localized, with theSociété Générale problemsrevealed in January 2008 mainlydue to a single rogue trader.However, there have been ongo-ing smaller problems particularlyin the German banking system,resulting in smaller scale bailoutsthan seen in the United States butnevertheless significant – in Au-gust 2007 Sachsen Landesbankwas sold to Landesbank Baden-Württemberg, and in August 2008the European insurer Fortis was

partially nationalized, Daxia was bailed out andIreland acted to guarantee bank deposits, while justoutside the sample period the ECB extended access toits cash auctions to more market participants.Developments in Europe since the end of the datasample have revealed substantial weakness in theEuropean banking system, including the takeover ofLandsbanki by the Icelandic government, problemswith the German giant Hypo Real Estate, and furtherguarantees on bank deposits by the Swedish andDanish governments. The 3 month Euroibor rate andGerman DAX index are used to represent Europeanconditions, and these are shown in Figures 1 and 2.TheDAX fell by 27 percent from 30 June 2007 to30 September 2008.

Australian and Japanese markets have both beenaffected by the global credit crunch, although withless dramatic failings of domestic financial institu-

4.0

4.5

5.0

5.5

6.0

6.5

7.0

2004 2005 2006 2007 2008

0.0

0.2

0.4

0.6

0.8

1.0

1.2

2004 2005 2006 2007 2008

0.0

1.0

2.0

3.0

4.0

5.0

6.0

2004 2005 2006 2007 2008

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

2004 2005 2006 2007 2008

YIELD ON 3 MONTH MONEY MARKET INSTRUMENTS IN

LOCAL CURRENCIES

1 July 2004 to 30 September 2008

5.5

6.0

6.5

7.0

7.5

8.0

8.5

2004 2005 2006 2007 2008

Vertical line marks 17July 2007

Source: Datastream.

US UK

EU Japan

Australia

Figure 2

tions to date. Equity markets in both countries havealso fallen dramatically; in the case of the AustralianASX200 the index reached its lowest level for5 years in early October 2008, losing 26 percentbetween the end of June 2007 and end of September2008. In Japan the Nikkei2000 has fallen 37 percentover the same period. The short term yields for bothcountries are shown in Figure 2.

All of the data cited in this paper are in domesticcurrencies. During the sample period the Australiandollar and the yen have been relatively volatile,greatly influencing the returns to an internationalinvestor – Table 1 shows the variance of the dailybilateral exchange rates for each of Australia,Britain, Japan and Europe against the US dollar inthe reference and credit crunch periods. The compu-tation of returns in US dollars is one means ofaddressing the problems for the internationalinvestor; however, in this case we are more con-cerned with the actual transmission of shocksbetween the various markets, so that incorporatingthe exchange rates using the domestic currency datais more appropriate. Previous studies have foundthat the use of domestic or US-dollar-based datamakes little difference to the qualitative results pro-duced in this type of exercise.

An empirically implementable model for contagion effects

This paper builds on the long-standing latent factormodeling approach adopted, for example, by Solnik(1974). The framework adapts the contagion modelof Dungey and Martin (2007) which captures thepotential for linkages between different financialmarkets to alter during times of stress.This approachhas been shown to nest many of the existing methodsof modeling contagion and to be consistent with thedefinitions of contagion as the opening of new andunanticipated channels of transmission during peri-

ods of stress (see particularly thereview of methodologies inDungey, Fry, Gonzalez-Hermo-sillo and Martin 2005).

In a multi-asset, multi-countryenvironment, the returns, yi,k,t foran asset in asset class i, and coun-try k, at time t can be representedby a four factor model given by

(1)

where Wt represents the common factor affecting allassets, Ck,t is the factor representing effects commonto country k, Mi,t represents effects common to assetclass i, and fi,k,t are idiosyncratic factors unique toeach individual asset.

Each of these latent factors is time varying, thusallowing for the increased common influences thatmight be posited during the credit crunch period andovercoming the problems of correlation-based testingas discussed in Forbes and Rigobon (2002). However,the parameter loadings are fixed over the sample peri-od; for the common, country, market and idiosyncrat-ic factors these are given by θi,k, αi,k, βi,k, φi,k respec-tively. To identify the parameters a standardizingassumption is commonly applied so that each of theunobserved factors is assumed to be drawn from aN(0,1) distribution. To capture the potential for fat-tailed, GARCH and autocorrelation effects in thereturns data, Dungey and Martin (2007) allow thecommon factor to evolve with a GARCH(1,1) struc-ture. In the current work, that refinement is ignored asit has been shown to make little difference to thequantitative results.

In applying these latent factor model approachesto contagion, it has been usually possible toexpress returns on various assets as premiums oversome risk-free rate – the usual procedure being toutilize US money or Treasury bond markets forthat purpose. In the current situation that would beinappropriate, as the source of the shocks to othermarkets in this case stem from the US money mar-kets themselves. To that end we adapt a refinementof the latent factor model approach first posited byMahieu and Schotman (1994) in dealing with bilat-eral exchange rates by modeling changes in each of the currencies involved using latent factors.This creates an extra common factor (called the


Focus

Table 1

Volatility in bilateral exchange rates against the US dollar in the reference

and credit crunch sample periods, measured as the variance of

daily log returns (%)

Period USD/AUD USD/EURO USD/JPY USD/GBP

Reference period:

1 July 2004 to

16 July 2007 0.3503 0.3102 0.2713 0.2556

Credit crunch period:

17 July 2007 to

30 Sept 2008 0.6301 0.2853 0.4117 0.2849

tkikitikitkkitkitki fMCWy,,,,,,,,,,

�� +++=


Focus

“numeraire” factor in exchange rate work) whichwill here relate to the influence of the US moneymarket.

Consider the case of modeling the premium via anobserved asset return for an asset in market i locat-ed in country k, yi,k,t, less the US money market rateused as the risk-free rate and denoted yr,US,t (wherehere there will be two markets, the money market,denoted by r, and the equity market, denoted by q).In a multi-asset multi-country environment theexcess returns yi,k,t – yr,US,t for an asset in asset class i,and country k over the risk-free rate, at time t, can berepresented by using appropriate combinations ofequation (1) for each return. Hence generally, where i,k ≠ r,US

(2)

where θ∼

i,k = θi,k – θr,US. In the special cases where i = r,k ≠ US (that is the assets which are in the sameasset class but different countries)

(3)

and when the assets are in the different asset classesbut the same country, i ≠ r,k = US

(4)

The volatility of the asset returns represented byequation (2) can be used to give a variance decom-position which is invariant to the standardizationassumption used in identification.

In the case where i,k ≠ r,US the variance can be decom-posed into seven separate components as follows:

(5)Portion of volatility due to common effects

(6)Portion of volatility due to country k effects

(7)Portion of volatility due to market i effects

(8)Portion of volatility due to idiosyncratic effects

(9)Proportion of volatility due to US country effects

(10) Proportion of volatility due to bond market effects

(11)Proportion of volatility due to US bond market idiosyncratic effects

In the case where the assets are both from theUnited States, that is k=US, the proportion of volatil-ity (6) and (9) are replaced with an expression with anumerator of (αi,US – αr,US)2. In the case where theassets are both from the same money market, that isi=r, the proportion of volatility due to market effectsreplaces (7) and (10) with an expression with anumerator of (βr,k – βr,US)2.

The decompositions in equations (5) to (11) give therelative importance of the four sources of volatilityduring the reference period.

During an identified period of crisis, there may becontagion effects where these are modeled as extra(either positive or negative) channels which emerge

tikitUSUSrtkkitkitUSrtki MCCWyy,,,,,,,,,,,

~�� +�+=�

tkiUSrtkikitrUSr ffM,,,,,,,,

�� +� ,

itUSUSrtkkitkitUSrtki CCWyy ,,,,,,,,, (~

�� +�+=�

tkiUSrtkikitrUSrki ffM ,,,,,,,,, )( �� +�+

kitUSUSrkitkitUSrtki MCWyy ,,,,,,,,, )(~

�� +�+=�

tkiUSrtkikitrUSrtiki ffMM ,,,,,,,,,, �� +�+

=2

,

2

,

2

,

2

,

2

,

2

,

2

,

2

,

~

~

USrkiUSrkiUSrkiki

ki

��

�

++++++

=2

,

2

,

2

,

2

,

2

,

2

,

2

,

2

,

~USrkiUSrkiUSrkiki

ki

��

�

++++++

=2

,

2

,

2

,

2

,

2

,

2

,

2

,

2

,

~USrkiUSrkiUSrkiki

ki

��

�

++++++

=2

,

2

,

2

,

2

,

2

,

2

,

2

,

2

,

~USrkiUSrkiUSrkiki

ki

��

�

++++++

2

,

2

,

2

,

2

,

2

,

2

,

2

,

2

,

~USrkiUSrkiUSrkiki

USr

��

�

++++++

=2

,

2

,

2

,

2

,

2

,

2

,

2

,

2

,

~USrkiUSrkiUSrkiki

USr

��

�

++++++

2

,

2

,

2

,

2

,

2

,

2

,

2

,

2

,

~USrkiUSrkiUSrkiki

USr

��

�

++++++

only during the crisis period. Here we are primarilyinterested in the effects from the US money marketson other asset classes and in other geographical mar-kets. The model in equation (2) can be modified toallow for the opening of a channel of contagionwhich exists only during the credit crunch period.This is achieved by adding an extra term to accountfor differing effects from the US money market fac-tor during the credit crunch period. So that in thecredit crunch the excess asset returns are given asfollows,

(12)

where It is an indicator variable taking the value ofzero in the reference period and 1 during the creditcrunch period.

During periods of crisis there is the potential for anextra effect from the US money market on theexcess return given by δi,k, and this is the effectlabeled contagion. This is formally equivalent to aChow test – see also Dungey, Fry, Gonzalez-Hermosillo and Martin (2005) for the equivalence ofthe Forbes and Rigobon (2002) test to a Chow testand its multivariate equivalent. Note that the focuson the effects from the US money markets does notpreclude increases in the shocks travelling throughother channels – global, country or market – each ofthe latent factors are time varying and this allows forgreater turmoil in the credit crunch period.However, the impact coefficients for each of thesefactors are deemed to remain the same across thereference and credit crunch periods.

In the credit crunch period the variance decomposi-tion given in equations (5) to (11) are consequentlymodified to have the denominator

,

and the additional component of the decompositionwill be the portion of volatility due to contagioneffects given by

(13)Portion of volatility due to contagion

=

Estimation techniques

The model is estimated by indirect estimation bysimulating the system comprising equations (2) and(12) for a reference and credit crunch period underthe assumption that each of the independent latentfactors evolves as an iid normal process. Indirectestimation minimizes the difference between thesimulated and actual sample data using an auxiliarymodel (see Gourieroux and Monfort 1994). In thiscase the auxiliary model is given by the varianceand covariance properties of the different assets inthe reference and credit crunch periods (see, forexample, Dungey and Martin 2004 and 2007). Theindependence of the latent factors allows theexpression of the variance and covariance condi-tions in analytic form. This is most easily seen byconsidering the different cases that may occur inthis two-markets, multi-country model as shown inAppendix.

The difference between the theoretical secondmoment conditions in the reference and credit crunchperiods is captured by the additional contagion effectsgiven by δ2

i,k. Hence an appropriate test for the pres-ence of contagion from the US money market to anyparticular excess asset return, yi,k,t – yUS,r,t , is that underthe null hypothesis of no contagion δi,k = 0,∀i,k, andthese can be applied using an LR test.

One further complication arises from the differingtime zones of the data. There is no overlap in tradingtime between the Australian and Japanese markethours and those of the United States. The UK andEuropean markets overlap with the US market, par-ticularly for the key 8:30 am EST (1:30 pm GMT)macroeconomic news releases. However, the dataused in this study represent closing prices in eachcase. Consequently, the time-zone problem is dealtwith by lagging US returns by one day – Australian,Japanese, UK and European returns are modeled ascontemporaneous.1

Estimation is carried out in Gauss 6.0 and makes useof the optimization module OPTMUM, with100 simulations (resulting in a total of 110,900 obser-vations). To aid in convergence in estimation, with-out affecting the variance decomposition results, it is


Focus

tikitUSUSrtkkitkitUSrtki MCCWyy,,,,,,,,,,,

~�� +�+=�

tUSrtkitkiUSrtkikitrUSr fIffM,,,,,,,,,,,

�� +�+�

2

,,

22

,

2

,

2

,

2

,

2

,

2

,

~kiUSrkiUSrkiUSrkiki

�� +++++++

2

,,

22

,

2

,

2

,

2

,

2

,

2

,

2

,

~kiUSrkiUSrkiUSrkiki

ki

��

�

+++++++

1 One could argue that this is a misrepresentation of the impact ofEuropean and UK data on Australia and Japan. However, experi-ments with different lag structures did not make a qualitative dif-ference to the outcomes. This is strong supporting evidence of theimportance of the US market in driving the spread of the problemsassociated with the credit crunch.


Focus

convenient to use a 10 variate VAR(1) in the10 financial returns to prefilter the data.

Empirical results

The model given in equations (2) and (12) is appliedto the excess returns data in equity market indexreturns and changes in money market rates acrossthe five different countries during reference andcredit crunch periods. The most interesting form inwhich to examine the results is via the variancedecompositions, as represented in equations (5) to(11) and (13).

In reporting the results of the estimation, Tables 3and 4 focus on the proportions of volatility due to

global common effects, the identified individualcountry and market effects, the US country effectand the common money market effect on each assetthat originates with the US money market, theunique effects for each asset return and the effects ofcontagion in the case of the credit crunch period.Themissing portion of the total volatility report in Tables3 and 4 in each case is the remnant attributable tothe idiosyncratic US money market effect. A simpleaddition of the columns of the Table shows this to berelatively small in all cases.

Table 3 shows the variance decompositions of themoney and equity markets during the reference peri-od. The global common factor contributes around50 percent of volatility to the individual money mar-kets. The majority of the remainder is distributed

Table 2

Selected Descriptive Statistics for the excess returns data during the reference period and credit crunch period

Money Market Equity Market

Country St Dev Skew Kurtosis St Dev Skew Kurtosis

Reference period: 1 July 2004 to 16 July 2007

US 0.0309 0.0270 7.4147 0.006 – 0.2955 4.5678

UK 0.0188 0.6399 10.1829 0.006 – 0.6591 4.3738

EU* 0.007 1.096 47.6277 0.008 – 0.2373 3.8420

Japan 0.006 1.005 64.7907 0.008 – 0.1983 4.0287

Australia 0.018 – 0.0389 15.6141 0.005 – 0.2273 4.0354

Credit crunch period: 17 July 2004 to 30 September 2008

US 0.0784 – 0.3288 28.7389 0.0096 – 1.0798 12.5580

UK 0.0339 – 0.7722 74.6968 0.0094 – 0.5634 8.6215

EU* 0.0101 1.4634 29.5396 0.0098 – 0.6201 6.1989

Japan 0.0068 0.8547 33.1201 0.0101 – 0.2939 4.4433

Australia 0.0282 – 0.2971 13.3085 0.0076 – 0.7457 10.4052

*EU is represented by the DAX equity market index.

Table 3

Variance decomposition of excess returns in money and equity markets for the reference period from

1 July 2004 to 16 July 2007, following equations (5) to (11), from estimation of the model given in equations (2) and (12)

US UK Europe Japan Australia

Money Market

Globalt

W – 51.77 57.35 55.98 49.23

Money Market trM

, – 16.32 18.38 21.59 21.27

Country tiC

, – 14.51 2.30 1.93 5.60

US country tUSC

, 11.82 13.30 14.77 14.03

Unique tkif ,, – 5.56 8.58 5.72 9.86

Equity Market

Globalt

W 1.46 59.04 60.14 62.63 59.59

Equity Market tqM, 19.86 18.99 18.88 19.01 18.54

Money Market trM

, 0.00 18.73 18.57 18.79 18.24

Country tiC

, 8.15 2.78 4.92 2.48 1.09

US Country * 13.63 13.51 13.68 13.27

Unique tkif ,, 70.53 5.55 2.53 2.20 7.51

Numbers are in %. – * This is included in the Country effect for the US.

across the money market factor (ranging from

16 percent in the United Kingdom to almost 22 per-

cent in Japan) and the common US country effect,

which ranged from 12 to 15 percent across the coun-

tries. The unique factor contributions were under

10 percent in each individual money market. This

tends to suggest that money markets were moving in

a way that reflected global conditions and that the

money market is indeed very international at the

short end. There is, however, some evidence that the

UK market was responding to factors pertinent only

to the United Kingdom, with a contribution of the

UK country factor of 14.5 percent.

In the equity markets the global common factor is an

important component of almost all excess returns

volatility, at around 60 percent of volatility for all

markets except the US. In the United States the two

dominant forces shaping equity market volatility are

the equity market effect, at just under 20 percent of

observed volatility, and the unique US equity market

factor, at 70.5 percent of volatility. This points to the

important role of the United States as a market

leader for the other economies. What is global for

the other countries is not driving innovations in the

US market, supporting other research pointing to

the leading role of the United States in absorbing

and distributing news effects to other economies

(see Ehrmann and Fratzscher 2005). In the other

countries, the US country effects are important, as

they were in the money market results – in the equi-

ty markets these effects account for around 13 per-

cent of total volatility. More important in the equity

markets are individual country effects, at around

18 percent of observed volatility in the equity mar-

kets, compared with their relatively low weight in the

majority of the money markets.

The credit crunch period results are shown in

Table 4. The contributions of both the global factors

in both markets are somewhat reduced in all non-

US markets compared with the reference period.

The contribution to volatility of all other factors

also declines slightly. The major change is in the US

equity market effects, where the equity market

effect and the idiosyncratic effects decline substan-

tially to give way to a dominant effect from conta-

gion at 78 percent of observed volatility. In all other

markets the contagion effects lie between 4 and 10

percent, with the greatest contribution in the

Australian money market and the smallest in the

Japanese equity market. The contagion effect from

the US money markets on US equity markets is pro-

nounced. The interpretation of this result is that

there has been a significant and substantial change

in the transmission of US money market conditions


Focus

Table 4

Variance decomposition of excess returns in money and equity markets for the credit crunch period from

17 July 2007 to 30 September 2008, following equations (5) to (11) and (13), from estimation of the model

given in equations (2) and (12)

US UK Europe Japan Australia

Money Market

Globalt

W – 49.27 53.58 51.51 44.65

Money Market trM

, – 15.53 17.17 19.86 19.29

Country tiC

, – 13.81 2.23 1.78 5.08

US Country tUSC

, – 11.25 12.42 13.59 12.72

Unique tkif ,, – 5.29 8.02 5.26 8.94

Contagion – 4.83 6.57 7.99 9.31

Equity Market

Globalt

W 0.31 56.05 56.62 59.72 56.78

Equity Market tqM, 4.27 18.02 17.78 18.12 17.66

Money Market trM

, 0.00 17.78 17.49 17.92 17.38

Country tiC

, 1.75 2.64 4.64 2.37 1.04

US Country tUSC

, * 12.94 12.72 13.04 12.64

Unique tkif ,, 15.18 5.27 2.38 2.10 7.15

Contagion 78.47 5.07 5.85 4.65 4.72

Numbers are in %. – * This is included in the Country effect for the US.


Focus

to the US equity markets. The relative stability ofthe remaining volatility decompositions suggeststhat given that observed volatility has increased dra-matically in each of the markets, the US equity mar-ket is largely acting as the conduit to transmittingthe effects of the changes in the US money marketto the remainder of the countries in the sample.Kaminsky and Reinhart (2007) argue that devel-oped markets such as the US act to transmit crisesto emerging markets. The evidence here supports anextension of this hypothesis also to developed mar-kets. The turmoil in equity markets seems to becoming from increased general turmoil, where indi-vidual effects are being absorbed by the US equitymarket directly and subsequently transmitted viamarket and common effects to other markets.

Comparison with previous crises

The extent of contagion evident in the volatilitydecompositions may seem relatively small. It isworth considering this in the context of evidence oncontagion effects in other major crises. In the Asiancrisis of 1997–1998 contagion effects between thecurrency markets of Korea, Indonesia, Thailand andMalaysia ranged between 9 and 46 percent ofobserved volatility (Dungey and Martin 2004), andin examining cross-market links, contagion betweenequity and currency markets resulted in contribu-tions of contagion to volatility of up to 11 percentin equity markets and 36 percent in currency mar-kets. Interestingly the contagion effects to equitymarkets in that crisis were a greater proportion ofobserved volatility for developed equity marketsthan emerging markets (Dungey and Martin 2007).In the Long-Term Capital Management (LTCM)crisis in 1998, contagion effects from the UnitedStates and Russia on twelve international sovereignbond markets ranged up to 18 percent. Five of thecountries had contagion effects in the 3 to 5 percentrange, consistent with the current results (the high-est effects were felt by Brazil, see Baig andGoldfajn 2001 and Netherlands, see Dungey et al2006). For the same crisis in the equity markets thecontagion effects were a much larger proportion ofvolatility, with only 3 of 10 countries examined hav-ing less than 20 percent of observed volatility dueto contagion effects, and impacts ranging as high as82 percent. In each of the cases, regardless of theproportion of contagion evident in the results, thecontagion effects are always statistically significant.Further detail on these results can be found in

Dungey and Martin (2004 and 2007) and Dungey etal. (2006 and 2007).

The conclusion from this comparison is that the extentof contagion in the current credit crunch is not out ofline with previous substantial crises. Interestingly, theproportion of contagion experienced in both themoney and equity markets (with the exception of theUS equity market) is relatively similar at between 4and 9 percent. In the other papers examining cross-country and cross-asset market linkages in a singlemodel, one market has seemed to receive more conta-gion effects than another. In the East Asian crisis thiswas the currency market (compared with equity mar-kets) and in the LTCM crisis it was the equity market(compared with the bond markets), whereas in thecredit crunch the contagion effects are relatively simi-lar in the equity and money markets.

Conclusion

The credit crunch of 2007–2008 seems in many waysto be quite different from other financial crises of thepast two decades. It initially involved the deepest andmost developed financial markets in the world, thoseof the US and the UK, but has expanded to otherdeveloped markets and post-August 2008 has becomea truly global phenomenon adversely affecting manyemerging economy markets. The underlying difficul-ties associated with the spread of the credit crunchhave been lack of liquidity due to uncertainty aboutthe value of sophisticated products that were initiallydesigned to diversify risk.

This paper has adopted a latent factor model of finan-cial market returns developed by Dungey and Martin(2007) to consider simultaneously the transmission ofshocks across both geographical borders and assetclasses. Usually this model requires a concept ofreturns in excess of a benchmark, risk-free asset. In thecurrent environment the usual risk-free asset, the USgovernment short-term money market, is a primarysource of the liquidity problems, and the modelingframework was adapted to account for this. Empiricalimplementation of this model for short-term(3 month) money markets and equity markets for theUnited States, the United Kingdom, Europe, Japanand Australia supports a number of conclusions.

First, the relative contribution of global commoninfluences on the excess returns over the US moneymarket in these economies and asset markets has notchanged markedly between the reference period

(from 1 July 2004 to 17 July 2007) and the creditcrunch period (from 17 July 2007 to 30 September2008). This means that increases in volatility in glob-al shocks are being transmitted to all markets in thesame manner as during the non-crisis period.

Second, the contribution of contagion to volatility inthe non-US markets in each asset class lie in therange of 4 to 10 percent, which is in line with resultsfound for the contribution of contagion in evidencegathered for previous crises.

Third, the US equity market seems to have a role inabsorbing shocks from the US money market, havingexperienced a far greater degree of contagion thanother markets, and may be interpreted as having actedas the distributor of these shocks to other jurisdictionsvia common effects, extending the hypothesis of cen-tre and periphery behaviour proposed by Kaminskyand Reinhart (2000) beyond that of a developed mar-ket distributing shocks to emerging markets.

Finally, there is little evidence in general for coun-try and market-based idiosyncratic behaviour dri-ving the volatility in individual markets, whichstrongly supports the global nature of this problemand that it is not in fact driven by (although it maybe exacerbated by) inappropriate domestic policyresponse (see Karolyi 2003). Important advances inattempting to explain the transmission of crisessuch as these are being made through the use ofnetworks (see Allen and Babus 2008), which pro-vide a convincing underlying theory for why wemay mistakenly understand interlinked marketsand entities as suffering from contagion when infact the linkages exist in an identifiable way prior tothe crisis occurring but are obscured due to com-plexity (see Kiyotaki and Moore 2002). Contagionitself only exists when new channels arise in periodsof stress.

References

Allen, F. and A. Babus (2008), Networks in Finance, in: Kleindorfer,P. and J. Wind (eds.), Network-based Strategies and Competencies,Philadelphia: Wharton School Publishing (forthcoming).

Baig, T. and I. Goldfajn (2001), The Russian Default and theContagion to Brazil, in: Claessens, S. and K. Forbes (eds.), Inter-national Financial Contagion: How It Spreads and How It Can BeStopped, Washington DC: World Bank.

Dornbusch, R., S. Claessens and Y. C. Park (2000), “Contagion:Understanding How It Spreads”, The World Bank ResearchObserver 15, 177–197.

Dungey, M., R. Fry, B. Gonzalez-Hermosillo and V. L. Martin (2005)“Empirical Modelling of Contagion: A Review of Methodologies”,Quantitative Finance 5, 9–24.

Dungey, M., R. Fry, B. Gonzalez-Hermosillo and V. J. Martin (2006),“Contagion in International Bond Markets during the Russian andLTCM Crises”, Journal of Financial Stability 2, 1–27.

Dungey, M., R. Fry, B. Gonzalez-Hermosillo and V. L. Martin (2007),“Shocks and Systemic Influences: Contagion in Global EquityMarkets in 1998”, North American Journal of Finance and Econo-mics 18, 155-–174.

Dungey, M. and V. L. Martin (2004) “A Multifactor Model ofExchange Rates with Unanticipated Shocks: Measuring Contagionin the East Asian Currency Market”, Journal of Emerging MarketsFinance 3, 305–330.

Dungey, M. and V. L. Martin (2007), “Unravelling Financial MarketLinkages during Crises”, Journal of Applied Econometrics 22,89–119.

Erhmann, M. and M. Fratzscher (2005), “Equal Size, Equal Role?Interest Rate Interdependence between the Euro Area and theUnited States”, The Economic Journal 115, 928–948.

Flood, R. and N. Marion (1999), “Perspectives on the RecentCurrency Crisis Literature”, International Journal of Finance andEconomics 4, 1–26.

Forbes, K. and R. Rigobon (2001), Measuring Contagion:Conceptual and Empirical Issues, in: Claessens, S. and K. Forbes(eds.), International Financial Contagion, Boston: Kluwer AcademicPublishers.

Forbes, K. and R. Rigobon (2002), “No Contagion, OnlyInterdependence: Measuring Stock Market Co-movements”, Jour-nal of Finance, 57, 2223–2261.

Gourieroux, C. and A. Monfort (1994), Simulation BasedEconometric Methods, CORE Discussion Paper 9313.

Granger, C., B. N. Huang, and C. W. Yang (2000), A BivariateCausality between Stock Prices and Exchange Rates in AsianCountries, The Quarterly Review of Economics and Finance 40,337–354.

Hartmann, P., S. Straetmans and C. G. de Vries (2004), “AssetMarket Linkages in Crisis Periods”, Review of Economics andStatistics 86, 313–326.

Kaminsky, G. and C. Reinhart (2000), “On Crises, Contagion andConfusion”, Journal of International Economics 51, 145–68.

Kaminsky, G. and C. Reinhart (2007), “The Centre and thePeriphery:The Globalization of Financial Turmoil”, in: Reinhart, C.,C. Vegh and A. Velasco (eds.), Capital Flows, Crises and Stabili-zation: Essays in Honor of Guillermo Calvo, Cambridge, MA:MIT Press.

Karolyi, G. A (2003), “Does International Contagion ReallyExist?”, International Finance 6, 179–199.

Kiyotaki, N. and J. Moore (2002), “Balance-sheet Contagion”,American Economic Review 92, 46–50.

Mahieu, R. and P. Schotman (1994) “Neglected Common Factors inExchange Rate Volatility”, Journal of Empirical Finance 1, 279–311.

Pericoli, M. and M. Sbracia (2003), “A Primer on FinancialContagion”, Journal of Economic Surveys 17, 571–608.

Solnik, B. (1974), “The International Pricing of Risk: An EmpiricalInvestigation of the World Capital Market Structure”, Journal ofFinance 29, 365–78.


Focus


Focus

AppendixVariance and Covariance Conditions

These are most easily seen to identify the modelwhen divided into a number of specific cases.Expressions A1 to A9 given below refer to covari-ance and variance expressions for the reference peri-od. It is easily determined that the equivalent expres-sions in the credit crunch period are obtained byadding an extra term δ

2i,k to each expression, where

this represents the extra transmission channelemerging from US idiosyncratic shocks.

Variances for:

Covariances between:

A1. Asset excess returns in the money market

tUSrtki yy,,,,

�tUSrtki yy,,,,

� where USk �

2

,

2

,

2

,,

2

,

2

,

2

,,, )()var( USrirUSrirUSririrUSrir yy �� ++�+++=�

A2. Excess asset returns in the non-US equity markets

2

,

2

,

2

,

2

,

2

,

2

,

2

,,, )var( USriqUSriqUSriqiqUSriq yy �� ++�+++=�

A3. Excess asset returns in the US equity market

2

,

2

,

2

,

2

,

2

,,

2

,,, )()var( USrUSqUSrUSqUSrUSqUSqUSrUSq yy �� ++�+++=�

A4. Asset returns both in the money market for different countries (in which case neither can be the US)2

,,,,,

2

,,,,,,, ))((),cov( USrUSrjrUSrirUSrjrirUSrjrUSrir yyyy �� +��++=��

A5. Asset returns both in the same country, one in the money market and one in the equity market

(in which case neither can be the US)2

,,,,,,

2

,,,,,,, )(),cov( USrUSrUSririqirUSriqirUSriqUSrir yyyy �� +�+++=��

A6. Asset returns both in the equity market for different countries (where one of those countries is not the US)2

,

2

,,,

2

,,,,,,, ),cov( USrUSrjqiqUSrjqiqUSrjqUSriq yyyy �� ++++=��

A7. Asset returns both in the equity market for different countries where one of those is the US. 2

,

2

,,,,,,,,,,,, )(),cov( USrUSrUSqiqUSrUSriqUSqiqUSrUSqUSriq yyyy �� +++�+=��

A8. Asset returns with one in the US equity market and the other in non-US money market.2

,,,,,,,,,,,,, )()(),cov( USrUSrUSrirUSrUSrUSqUSqirUSrUSqUSrir yyyy �� +�+�+=��

A9. Asset returns with one in the money market, one in the equity market for different countries (neither is the US)2

,

2

,

2

,,,,,,, ),cov( USrUSrUSrjqirUSrjqUSrir yyyy �� +++=��


Focus

AVOIDING THE NEXT CRISIS

DAVID G. MAYES*

The present crisis is evolving by the minute and any-thing written about it is likely to be out of date by thetime it is published. This note therefore focuses onthe lessons we can learn to avoid such crises in thefuture.

Every crisis is in some sense different, as on the wholesociety is quite good at learning the direct lessonsfrom previous examples. However, in many otherrespects financial crises are very much the same. Inreflecting on the Nordic crises of the 1980s and 1990sMayes et al. (2001) listed seven characteristics:

– a major regulatory change that opens both buyersand sellers to new opportunities and risks ofwhich they have little experience

– a period of rapid economic growth– a rapid rise in asset prices– a weak framework for supervision– a tax regime that encourages borrowing– unsustainable macroeconomic policies– a substantial adverse shock

While this is not an exact description of the lead upto the present crisis, with the “originate and distrib-ute” model of banking representing the new oppor-tunities and risks, it is near enough to pose the ques-tion of ‘Surely we should have seen this coming?’However, with hindsight this is true of virtually allcrises, the build up looks extreme and unsustainable,with various commentators offering notes of cautionon the way up.

In the light of this experience the best that can beoffered for crisis avoidance is a series of characteris-tics in the system that lean against the causes thathave just been outlined. However, that is only halfthe picture.The system can also be designed to makethe reaction to problems less dramatic, as a crisis is acombination of an excessive build up and a major

reversal, where a drastic loss of confidence matches

the excess confidence on the way up. If the problem

can be dealt with efficiently, then the drastic reversal

may be avoided despite serious shocks.This provides

the traditional balance between a well-structured,

prudential framework on the one hand and an effi-

cient safety net on the other. However, in some

respects the two can work against each other. If

financial institutions believed they will be saved if

things go really wrong then they may be less cau-

tious, thereby increasing the chances of a disaster.

There therefore needs to be a very careful incentive

structure to guard against this form of moral hazard,

as set out in Beck (2003), for example, in the case of

deposit insurance.

This note deals with these two aspects in turn, with

the emphasis lying on the construction of an effec-

tive incentive-compatible safety net.

Encouraging prudential behaviour

Most of the lessons for prudential regulation from

the present crisis are being learnt without the need

to make fundamental changes to the framework of

regulation. Both banks and the authorities realise

the mistakes that were made in measuring risks, par-

ticularly with respect to liquidity in its broadest

sense and with regard to the degree to which risks

transferred off the balance sheet still lay with the

originator and were not properly accounted for. As a

result both parties will be much more cautious over

leveraging, will be much more careful in trying to

measure the riskiness of more complex derivative

products and will be much more concerned to take

account of the risk that has been moved from its pri-

mary location. Taken together therefore we can

expect both a substantial contraction of the banking

industry in the short run and a much more cautious

approach in the future until the memory fades.

Insofar as banks have been taken into public owner-

ship or have received substantial injections of tax-

payer funding, particularly in the form of preference

shares, the traditional pattern of external regulation* University of Auckland.


Focus

by the authorities and operation by the private sec-tor is broken.The authorities can now influence suchbanks much more directly. While it is unlikely thatgovernments will want to get involved in detailedmanagement it is inevitable that they will have a sayon the strategy of the institutions and hence a clearstep towards conservatism can be expected. Indeedin the European case, the European Commission(2008) guidelines on how such intervention shouldbe undertaken includes (§27) that there shall besafeguards such as “behavioural constraints ensuringthat beneficiary financial institutions do not engagein aggressive expansion” and that a restructuringplan should be submitted within 6 months.

The existing framework under Basel II gives amplescope to decide on what capital needs to be held andhow risks should be addressed under pillars 1 and 2.Rating agencies and internal rating systems will radi-cally change their procedures to try to avoid commit-ting such large errors in the future. Liquidity bufferswere never addressed with the same vigour as capitalbuffers by the Basel Committee and we can expectthat this lapse will be corrected over the coming fewyears. Since leverage has also proved a problem, onecan expect pressure to introduce leverage ratio con-straints in addition to other capital buffers.

However, it is also clear that Basel II is seriouslydeficient in at least two respects. First, even at thetime, there were strong reservations about its pro-cyclicality. The system of capital buffers systemati-cally allows capital cover to fall as asset valuesinflate and confidence within the system grows in anupturn. Conversely, as asset prices and ratings fall in the downturn, there is a major need to replenishcapital just at the time when it is most difficult and expensive to do so. It is clear that the wholeapproach to buffers has to change, encouraging cap-ital provision in an upturn and permitting weakercapital cover as the risks are revealed in the down-turn. In this way there will be both more cover foradverse events and an easier ability to weatherthem. The means of doing this have not as yet beenfully worked out, but incentives such as dynamicprovisioning in Spain give an indicator of ways thatthey might be achieved. Similarly there is the sug-gestion by Kashyap et al. (2008) that banks might beable to take out capital insurance, which provides amore orderly (and cheaper) way of obtaining capitalin a crisis than a fire sale – provided that the insur-er itself can withstand a whole rush of claims at thesame time.

Second, it is also clear that the process of stress test-ing has to be changed as it seriously underestimatedthe consequence of a general downturn in the econ-omy and in financial confidence. Some of thisrequires a new approach to models. The macroeco-nomic models typically employed were not capableof generating extreme events. DSGE models, forexample, deliberately try to represent the normalequilibrating process across the economic cycle anddo not encompass what is effectively a regime shiftwhen confidence is suddenly eroded. Such regimeswitching models and more sophisticated treatmentsof expectations formation have been developed(Sargent 2001; Branch and Evans 2008) and morerealistic tests can be expected in the future.

However, the main incentive for prudential behav-iour is not going to come from a set of prudentialregulations that comprehensively seek to lay downminimum conditions that banks must apply to avoidendangering the financial stability of the system. Itwill come from the self-interest of those involved infinancial markets themselves.

This form of market discipline will be exercisedthrough a variety of routes (Llewellyn and Mayes2003) but the most prominent are equity and uncol-lateralized (junior) debt. Falling share prices relativeto the industry average or widening premia for bor-rowing from the market are clear signals of disquiet.Banks could be required to issue subordinated debtat regular intervals so that there is a relevant price(Calomiris 1999). However, for these signals to havean effect, management or alternative management inthe sense of a merger or acquisition has to respond(Bliss and Flannery 2002).

It seems to be very difficult to put together a set ofincentives for those running the bank to act directlyin the interests of the shareholders. Frequently theincentives to management can be seen as very one-sided, encouraging the taking of large risks in thehope of increasing the value of share options andhaving generous exit arrangements if a manager isforced out before completion of a contract. Worsethan this, many of the upside gains may have virtu-ally nothing to do with the management’s actionsand may simply reflect the consequence of a gener-al rise in asset prices in the market. They may alsobe too short term and allow a manager to exitbefore the longer-run consequences of his actionsare felt. It is widely thought that the remunerationof the chief management of banks has got out of


Focus

hand in recent years and needs reform (Sinclair etal. 2008), and this has been an explicit target wherebanks have been nationalised or subject to majorpublic sector share purchase.1

However, the market for corporate control itselfdoes not appear to work very well in much of thebanking sector, as many banks have large marketshares in particular markets where takeovers byother banks would fall foul of the competitionauthorities. Where it does not, the resulting leverageto effect the purchase can bring down the institutionin a subsequent, as illustrated by the Royal Bank ofScotland, Fortis and HBOS. Clearly this circum-stance needs to change and market discipline needsto become more effective.The disclosure of informa-tion under Pillar 3 of Basel II appears to do relative-ly little in this regard and although extensive it doesnot cover some of the more obvious indicators ofrisk management as revealed under the NewZealand disclosure regime since 1996.

Lastly there will be considerable pressure to extendthe boundary of regulation. In the United States thishas already expanded to include investment bankssince they have either been closed, purchased bybanks or decided to become banks themselves. InEurope the boundary is already much wider, withmost institutions treated like banks. However, thereis pressure to extend the coverage to hedge fundsand even sovereign wealth funds. Rules governingthe behaviour of agents are also likely to be tight-ened as in some countries they contributed to thegranting of poor quality mortgage advances.

Pushing the boundary ever further out has obviousdisadvantages as it limits the scope for financing morerisky but higher return enterprises by those who canafford to take the risk and it pushes such activities intoobscurer arrangements and limits intermediation.Theoverall consequence would be to reduce the potentialrate of growth of the economy.

In one sense it is easy to define where the boundaryshould be drawn. If the failure of an institution(under the rules that govern it) would threatenfinancial stability then it should be inside the regula-tory boundary. Not only does the social cost justify itbut the advantages accrued by the protection fromdisorderly failure offers a countervailing benefit.

However, recent events have shown that it is notpossible to treat institutions individually and it isnecessary to worry about the aggregate if many getinto difficulty at the same time. Hence it is both theextent of the sector and the size of individual institu-tions within it that will be of concern.

Correcting problems without a crisis

As noted in the previous section, the most impor-tant part of problem correction should be effectedby the usual market mechanisms before banks everbecome a concern to the authorities. However, ifthat fails the safety net needs to have two maincharacteristics:

– a mandatory programme of structured earlyintervention and resolution

– a credible regime for depositor protection.

The principal purpose of a mandatory programme ofstructured early intervention and resolution (SEIR)is to try to ensure that should a problem occur it canbe resolved rapidly without recourse to public fund-ing and without threatening financial stability. But ithas a secondary purpose as an incentive to all thosewith a claim on the bank to act in advance of anymandatory scheme instituted by the authorities. Themandatory scheme will restrict the choices, forceaction to a tight timetable and in general look to theinterests of the creditors and taxpayers.

Prior to the present crisis only a few countries,including the United States, Canada and Mexico hada clear SEIR framework in place. Most countries hadless formal and more discretionary approaches forhandling problems. While it has been clear from therecent problems that the authorities are prepared toact rapidly, it has been far less clear at what stagethey would act and whether many claimants do bet-ter by waiting or seeking a private sector solution atan early stage. Such uncertainty is likely to causelosses. Compared to what was intended in advance,far more of taxpayers’ money has had to be used inpropping up the financial sector.While the safety netseems to be working in the sense of avoiding a dam-aging system-wide collapse, it has not operated asearly and efficiently as hoped. Similarly, whilstalmost all OECD countries (and many others) havehad deposit insurance regimes in place, the currentturmoil has shown that in general they have not beenappropriately structured to handle a serious prob-

1 See, for example, the Banking Bill introduced in the British Houseof Commons on 7 October 2008 for reforming banking regulation– http://www.publications.parliament.uk/pa/cm200708/cmbills/147/2008147.pdf.


Focus

lem. The United Kingdom has seen the first openbank run since 1866, with Northern Rock inSeptember of 2007. A year later the EU has advo-cated doubling the size of protection for individualdepositors and several countries have felt impelledto introduce widespread guarantees to protect allretail deposits.

An SEIR regime

If an effective regime for acting on problem banksearly and resolving them with minimum cost to soci-ety is in place, this in itself can act as an incentive toprudence if it has an appropriate structure. Insteadof allowing the descent to a normal company insol-vency where costs are high and creditors may have towait over a decade for a final pay out (as in the caseof BCCI, for example), such a structure involvesincreasing constraint on shareholders and manage-ment by the authorities as capital falls and ultimate-ly taking the bank away from them while it still hassome value and resolving it immediately in a cost-minimising manner. Shareholders and managementare therefore likely to want to find a “voluntary” pri-vate sector solution before being forced into actionsby the authorities.

The present crisis has emphasised the need for inter-vention early and the need for very rapid action. Thetraditional triggers for intervention that exist in theUnited States, where SEIR was pioneered (Benstonand Kaufman 1994), have been shown to cut in toolate. Liquidity shortage, not capital shortage, precip-itated the collapse of banks – although of course lackof liquidity is normally the immediate cause of bankclosure, as it is liquidity that determines whetherthey can actually meet their obligations on a partic-ular day. Indeed, as Peek and Rosengren (1997)pointed out, action in the United States was in anycase normally triggered earlier by other adverse sig-nals generated by the CAMELS and other assess-ments of banks’ condition. Clearly, therefore, onelesson that has to be learnt from the recent experi-ence is that the triggers need to be earlier andinvolve a wider range of verifiable measures so thatthere can be little scope for debate with the authori-ties (Moe 2008; Mayes 2008).

The second major lesson from the present crisis is thatthe scope for emergency lending by central banks forbanks facing liquidity problems has had to be muchlarger than previously thought. Traditionally the‘lender of last resort’ function involves the central

bank injecting sufficient liquidity into the market thatany problem stemming from the failure of a normalsource of liquidity in the market can be overcome,provided that the banks have adequate collateral tooffer. (As a last resort such lending is normallyoffered at a premium.) In the Northern Rock crisis inthe United Kingdom in August and September 2007,the Governor of the Bank of England claimed that itwould not be possible to provide enough lending tothe market to enable Northern Rock to obtain ade-quate funds (and hence that any lending had to bespecifically to that bank).A year later, however, fund-ing needed to act on a more general problem was pro-vided (Mayes and Wood 2008). The scale of collater-alized lending required has had a major impact oncentral bank balance sheets and required a shift fromshort-term lending for a week or so to longer-termlending and a commitment to provide such facilitiesfor an extended period of time until the crisis playsitself out. It has also become clear that financial sup-port in the event of liquidity problems has had to beextended outside the traditional definition of banks –into investment banks in the United States and insome respects into insurance. More institutions arevital to financial stability than many had thought. Thefact that central banks can do this will provide addedconfidence in the financial system.

However, collateralized lending by central banks hasby no means provided enough support, and govern-ments have had to step in with loans, guarantees andoutright purchase (nationalization) of financial insti-tutions facing insolvency. The nature of these inter-ventions, some of which have involved on the spotinnovations, has shown that the system needs rethink-ing in many countries. It is not that tools cannot befound nor that most of them have been used beforebut simply that a coherent approach is needed if tax-payer costs and any possible moral hazard are to bemore limited in future. The United Kingdom hasaddressed this directly in a new bill placed before par-liament (see footnote 1) where it seeks the power tobe able to direct transfers and takeovers (to willingrecipients), form bridge banks or nationalize banks inorder to provide the smooth resolution of a troubledinstitution without the need for shareholder approval.Similar powers exist in other countries, where the keyingredient is that there has to be a special resolutionregime for banks and not the application of normalcorporate insolvency law.

Such regimes can take many forms (Mayes etal. 2001; Mayes and Liuksila 2003) and the New


Focus

Zealand regime is of particular interest as noted in thenext section as it can be applied across borders. Theessential ingredients are that all banks must be struc-tured in such a way that if they cannot be allowed tocease normal operations, as this would challengefinancial stability; it must be possible for the authori-ties to step in and restructure them into a viable insti-tution within the same working day so that no con-tracts are broken and depositors get uninterruptedaccess to their funds. In New Zealand this occursthrough a process described as “bank creditor recapi-talization”. As soon as the Reserve Bank, as theresponsible supervisor, determines that interventionis needed, a statutory manager is appointed who takesover from the shareholders, and makes a valuation ofthe claims on the bank, which are then written downsufficiently in order of increasing seniority until suc-cessful recapitalization is achieved. The bank thenreopens under statutory management the followingday and the claimants, who include the depositors,receive a tradable claim on the assets reflecting thesize of the writing down.

Nationalization and the formation of bridge banksare different routes to the same end. Nationalizationkeeps creditor claims whole while taking over fromthe shareholders, while a bridge bank is effectively atemporary nationalization of the whole or essentialpart of the failing bank (with the remainder of assetsand liabilities being placed in receivership). Only afew countries place an explicit requirement on theauthorities to minimize some specific form of loss,e.g. to the deposit insurer. Indeed, in the UnitedStates, the choice of resolution method lies with thedeposit insurer and a specific systemic risk exemp-tion has to be invoked in order to be able to takewider losses to society at large into account. Thisrequires the approval of the Federal Reserve, theComptroller of the Currency and the Secretary ofthe Treasury and has not yet been invoked.(Although it probably would have if the $700bn planfor buying toxic assets had not been put in place.)

In order to prevent the same problems in the future wecan expect to see other countries implement explicitobjectives and sufficient powers to implement SEIR,particularly the components relating to resolution.

Deposit insurance and wider guarantees

This financial crisis has revealed that many of the char-acteristics of actual deposit insurance round the worlddo not meet the concerns for which it was set up.

Traditional deposit insurance is designed to protect thenormal size of deposits of ordinary people. However,“normal” deposits are now often above the insurancelimit and hence many depositors are exposed. Depositinsurance is also designed as a confidence boostingdevice for the financial system as a whole, so that peo-ple do not rush to withdraw their deposits from soundbanks at the first hint of trouble as this would bringthem down as well.As the Northern Rock case in par-ticular revealed, even the UK’s relatively high limit(more than double the EU required minimum) onlyprotected around 90 percent of depositors and a muchsmaller proportion of deposits.

Second, it has only now been realised widely that peo-ple want continuing access to their deposits. In the EUinsurers are only required to pay out within threemonths (and this is renewable twice). Given theextent of transactions that cannot be met with thecash that happens to be in people’s pockets at thetime of failure, it makes sense to withdraw one’sdeposit to maintain liquidity even if the eventual pay-out will be in full, hence nullifying the attempt to useinsurance to stop a run. The EU is now advocatingthree days as the desirable limit. While the NewZealand requirement of access the following day maybe ideal, people may be able to accept the loss of ahandful of days (the United States aims at a week).

Between them these difficulties have contributed tothe situation where countries have largely felt unableto allow any banks to fail as their insurance systemswould not succeed in maintaining financial stability.On the one hand, they have bailed out the banksthemselves either by buying preferential shares orbuying or guaranteeing the assets that have been dif-ficult to sell at a reasonable price. On the other, theyhave issued blanket guarantees to creditors includingdepositors to stop them withdrawing or refusing tooffer further credit. Bailing out the bank itself and itsexisting shareholders, even if there are restrictions onfuture transactions offers an unfortunate incentive forthe future.The same could be said for junior creditors.However, for depositors it is unlikely that the problemfor the future will be so large. Most depositorsthought that they were fully protected in the event ofa general crisis anyway or that their deposits werewith a bank that was “too big to fail”.

Deposit insurance is really designed for problemswith individual small banks when there is no generalthreat to the system. Indeed the hope is, in WesternEurope at any rate, that it will hardly be used at all as


Focus

the sorts of banks that get into trouble will normallybe purchased by larger banks. As a result, depositinsurance has not been effectively tested in manycountries. If insurance limits are to be raised and thesystem geared to cope with the simultaneous failureof more than one large bank, it would become impos-sibly expensive if the system is to be funded so that itcan credibly pay out without recourse to the taxpayer.Such a cost would be counter-productive, although itmight be reasonable to offer people the option of pay-ing explicitly for insurance of larger deposits.

Clearly there needs to be a rethink of what consti-tutes a “normal” deposit, and something of the orderof three times GDP per head or $75,000–$100,000seems to fit that bill in the OECD countries(Campbell et al. 2008; Hoelscher et al. 2006). But themajor need is to be able to rescind the blanket guar-antees and move to a revised safety net where banksand their creditors do not rely in future on such guar-antees to bail them out.

Handling cross-border problems

Although there has been a lot of pressure in Europeto try to address the problems of large cross-borderbanks, the response has been slow and there hasbeen a general reluctance to enhance the existingnetwork of national regulation in home and hostcountries with a higher level coordinated approach.Now we are seeing the consequences with the col-lapse of the three largest Icelandic banks,Kaupthing, Landsbanki and Glitnir, threatening thesolvency of Iceland as a country. Other cross-borderbanking problems have so far been easier to resolve,with the French and Belgian authorities supportingDexia and the Belgian, Dutch and Luxembourgauthorities supporting Fortis, including the recentlypurchased ABN-AMRO. In these cases it was possi-ble effectively to divide up the banks into their com-ponent national parts and let the national authori-ties handle them. Creditors of the Icelandic bankshave not been so lucky as Iceland does not have theresources to handle all the claims for which it isliable. Under EU/EEA rules the home country isresponsible for the supervision and deposit insur-ance of the parent bank and branches in all coun-tries, while host countries are responsible for sub-sidiaries in their jurisdiction. The passport principlemeans that host countries cannot object to the struc-ture of the banking group nor to the home country’ssupervisory actions. Not surprisingly there are now

recriminations when one country is having to pay-out for another’s errors. The EU needs to resolvethis if it is to have a workable system in future withmore cross-border banks.

There are two possible ways to go. One is to have aEuropean level supervisor and resolution agency (aEuropean Deposit Insurance Corporation, if onewere to match the US model of the FDIC), while theother is to have much closer cooperation and jointresponsibility particularly for prompt correctiveaction (Mayes et al. 2008). New Zealand has imple-mented a third way out of the problem, which is toinsist that any significant bank operating within itsjurisdiction (assets over $10bnNZ) must not only bea subsidiary so that it is a definable legal entity withadequate capital of its own in New Zealand but thatit should be capable of running itself as a free-stand-ing organisation before the end of the “value” day inwhich a problem strikes. While this is workable andattractive to small countries, it would eliminatesome of the potential gains from economies of scopeand scale that underlie the rationale of cross-borderbanks in the first place.

If the EU is not prepared to move to an EU-levelsystem, major changes are required in the existingcombination of national systems to ensure a com-mon information base, joint responsibility and deci-sion-making, and a similar range of corrective andresolution tools and powers and the right to usethem to the joint instead of the purely national ben-efit. Maybe such changes are now more likely.

References

Beck, T. (2003), The Incentive-Compatible Design of DepositInsurance and Bank Failure Resolution: Concepts and CountryStudies, in: Mayes, D. G. and A. Liuksila, Who Pays for BankInsolvency?, Basingstoke: Palgrave Macmillan.

Benston, G. and G. G. Kaufman (1994), The Intellectual History ofthe Federal Deposit Insurance Corporation Improvement Act of1991, in: Kaufmann, G. G. (ed.), Reforming Financial Institutions andMarkets in the United States, Boston: Kluwer.

Bliss, R. and M. Flannery (2002), “Market Discipline in theGovernance of U.S. Bank Holding Companies: Monitoring vs.Influence”, European Finance Review 6, 361–95.

Branch, W. and G. Evans (2008), Learning about Risk and Return:A Simple Model of Bubbles and Crashes, University of Oregon,http://darkwing.uoregon.edu/~gevans/bubbles.pdf (accessed 3 November, 2008).

Calomiris, C. (1999), The Postmodern Bank Safety Net: Lessonsfrom Developed and Developing Economies, Washington DC:American Enterprise Institute.

Campbell, A, R. LaBrosse, D. G. Mayes and D. Singh (2008),Standards for Deposit Protection, mimeo.

European Commission (2008), The Application of State Aid Rules toMeasures Taken in Relation to Financial Institutions in the Context


Focus

of the Current Global Financial Crisis, Brussels (October)http://ec.europa.eu/comm/competition/state_aid/legislation/ bank-ing_crisis_paper.pdf (accessed 26 October 2008).

Hoelscher, D., M. Taylor and U. Klueh (2006), The Design andImplementation of Deposit Insurance Systems, Washington DC:International Monetary Fund.

Kashyap, A. K., R. G. Rajan and J. C. Stein (2008), RethinkingCapital Regulation, Paper prepared for Federal Reserve Bank ofKansas City Symposium on Maintaining Stability in a ChangingFinancial System, August.

Llewellyn, D. and D. G. Mayes (2003), The Role of MarketDiscipline in Handling Problem Banks, in: Kaufman, G. G. (ed.),Research in Financial Services: Private and Public Policy 15,Amsterdam: Elsevier.

Mayes, D. G. (2008), Early Intervention and Prompt CorrectiveAction in Europe, Paper presented at the CFS-IMF Conference onA Financial Stability Framework for Europe, 26 September,http://www.ifk-cfs.de/fileadmin/downloads/ publications/other/May-es_a.pdf (accessed 3 November 2008).

Mayes, D. G., L. Halme and A. Liuksila (2001), Improving BankingSupervision, Basingstoke: Palgrave.

Mayes, D. G., M. Nieto and L. Wall (2008), “Multiple Safety NetRegulators and Agency Problems in the EU: Is Prompt CorrectiveAction Partly the Solution?”, Journal of Financial Stability 4,232–257.

Mayes, D. G. and G. E. Wood (2008), “Lessons from the NorthernRock Crisis”, Economie Internationale, forthcoming.

Moe, T. G. (2008), “Time for Prompt Corrective Action”, TheFinancial Regulator 13(1), 61–67.

Peek, J and E. Rosengren (1997), “Will Legislated EarlyIntervention Prevent the Next Banking Crisis?”, SouthernEconomic Journal 64, 268–280.

Sargent, T. (2001), The Conquest of American Inflation, PrincetonNJ: Princeton University Press.

Sinclair, P., G. Spicer and T. Skinner (2008), Bonuses and theCredit Crunch, in: Mayes G. D., R. Pringle and M. Taylor (eds.),Towards a New Framework for Financial Stability, London:Central Banking.


Focus

EUROPE’S BANKING

CHALLENGE: REREGULATION

WITHOUT REFRAGMENTATION

NICOLAS VÉRON*

The intervention by governments in October 2008appears to have succeeded in stopping market panicabout Europe’s largest banks – for some time atleast. The massive plan that combines liquidity andcapital injections and guarantees was an appropriatereaction to the urgency of the moment. But the sameurgency had the consequence that the plan was notinformed by consideration of its long-term conse-quences. We can be sure that credit conditions willreturn to normal at some point in time, in a few yearsat the very latest. What remains entirely undeter-mined, however, is the shape the European financialsector will present itself in whenever the dust fromthis crisis settles.

There can be little doubt that the financial industrywill experience significant restructuring in the mean-time: market developments in late September andearly October have made amply clear that its earlierstructure was unsustainable. Now that the entire sec-tor has been brought under the umbrella of govern-ment guarantee, restructuring trends are bound todepend heavily on policy decisions, whether linkedor not to the implementation of the plan itself (seeVéron 2008). In other words, now that politicianshave (albeit temporarily) eclipsed market forces asthe lords of the financial industry, they have becomeresponsible for its future. It would be irresponsiblefor them to much delay some hard thinking aboutwhat this future should look like.

Europe’s banking integration at the crossroads

One key aspect of Europe’s banking industry thatwill be affected by public policy choices is its degreeof cross-border integration. In the past ten years,

cross-border integration has made rapid progress,due to the expansion of Western European banks’activities in Central and Eastern Europe and to alimited number of large cross-border acquisitionssuch as Santander/Abbey (2004) and UniCredit/HVB (2005). Consider the EU’s 15 largest banks (bymarket capitalization, measured on 30 September2008). In 2007, eight of them, or more than half, hadat least one-third of their total European revenueyielded from outside the home countries, againstnone ten years earlier. For four of them (DeutscheBank, ING, Nordea and UniCredit), the proportionwas even more than 50 percent: these banks deriveless revenue from their home country than from therest of their European operations.1

This striking development, however, remains unequaland the degree of cross-border banking integrationvaries significantly across Europe. In four of the fivelargest EU economies (Germany, France, Italy andSpain), foreign-owned banks are generally marginalplayers, the two only significant exceptions beingHVB in Bavaria (part of UniCredit since 2005) andBNL in Italy (part of BNP Paribas since 2006). In theUnited Kingdom, Santander has built a significantretail operation – and of course, banks from all overthe world are present on the wholesale markets – butnone of the domestic banks (Barclays, HBOS, HSBC,Lloyds TSB, RBS, Standard Chartered) has expandedmuch on the continent. Further north, Scandinaviahas seen a vivid intra-regional integration, but thepresence of non-Scandinavian banks there remainslimited.Altogether, Europe’s large banks are general-ly less internationalised than large European compa-nies in other sectors (Véron 2006).

Even within these limits, though, cross-border bank-ing integration appears to have brought significanteconomic benefit where it has been sufficientlydeveloped to have an impact, most strikingly inCentral and Eastern Europe. In these countries, thestrong catch-up growth of the past few years hasbeen fuelled not only by foreign direct investmentbut also by the easily available credit from Western

* Bruegel, Brussels.

1 Author’s calculations based on corporate disclosures. I thankMartin Saldias Zambrana for his highly valuable research assis-tance in assembling these numbers.

banks which themselves benefited from favourablerefinancing conditions (one negative aspect of thiscredit expansion, however, has been the high shareof foreign-denominated lending). Cross-borderbanking integration has also contributed to a rapiddistribution of new financial products and serviceofferings throughout the EU. On the whole, it is like-ly to have contributed to a more efficient allocationof capital, thus allowing better access to finance andbetter returns on investment for companies andhouseholds across the EU, even though quantifica-tion of such effects remains notoriously difficult (seeLondon Economics 2002).

However, the cross-border integration of the pastdecade has taken place in a climate marked by lowcredit spreads and relatively low concern about sys-temic risk. The public policies which have had mostimpact during that phase have been those carried outat European level on competition and the internalmarket.The European Commission has played a pow-erful role as enabler of cross-border consolidation by ensuring that prudential controls over nationalbanking sectors were not abused for protectionistpurposes. Landmark cases have included Santander-Champalimaud in Portugal (1999), takeover bids onAntonveneta and BNL in Italy (2005), and the com-bination of banking assets held by HVB andUniCredit in Poland (2005–06).

Prudential regulation, meanwhile, has not kept pacewith market developments. While many banks wereinternationalising at a rapid pace, supervision hasremained essentially national, governed by nationallaws and carried out by national authorities. A mod-icum of coordination has been brought by the cre-ation in 2004 of the Committee of European BankingSupervisors (CEBS) in the framework of the so-calledLamfalussy architecture for financial regulation, butCEBS has a merely advisory role, no binding decision-making powers of its own, and few permanent staff.Even before the credit crisis erupted in the summer of2007, many voices have expressed concern that thefragmentation of Europe’s supervisory frameworkmade it increasingly difficult to fulfil its financial sta-bility aims (see Véron 2007a; Decressin, Faruquee andFonteyne 2007).2

With the major repricing of risk and widely sharedexpectation of “reregulatory” initiatives resultingfrom the financial crisis since August 2007, the equa-tion of cross-border banking integration in Europe islikely to be fundamentally modified. Especially sincethe brutal degradation of the interbank lending mar-ket in September 2008, financial stability concernshave taken centre stage and banks have been broughton an unprecedented scale under the protection oftheir home country government. In some cases, thishas resulted in the immediate unwinding of previouscross-border combinations, as when the governmentof the Netherlands unexpectedly announced thenationalisation of the Dutch operations of Fortis andFortis’ share of ABN Amro, invoking depositor pro-tection concerns, on 3 October 2008. Similarly, RBShas signalled a downsizing of its business abroad toconcentrate on its core UK retail activity (see Larsenand Tucker, 2008). In an era of scarce capital and gov-ernments’ overwhelming concern of maintainingcredit in their domestic economies, it is likely that theshort-term trend will be a reduction in the interna-tional operations of most European banks comparedwith the domestic ones. Simply put, support fromhome-country taxpayers cannot sustainably be usedto shore up credit operations in other countries.Granted, cross-border consolidation can also resultfrom the turmoil, as when BNP Paribas acquired mostBelgian and Luxembourg operations of Fortis in earlyOctober, but under the current circumstances suchcases will not compensate the powerful drive towardsrenationalisation of European banking.

As a result, policymakers have a Gordian knot to cutnow. On the one hand, the collapse of trust of earlyOctober forced them to take charge of the bankingindustry and will necessarily result in a reregulationdrive. Conducted by national governments, this wouldtend to refragment the European banking industryalong geographical lines, with detrimental economicconsequences. On the other hand, existing cross-bor-der banks need a credible regulatory and supervisoryframework that respects the interests of their stake-holders in the plurality of countries in which theyoperate. The earlier belief that banks could live theirlives somewhat independently of public oversight, andsupport in crisis times, has been proven a myth. Thishumbling lesson applies as much to Europe’s cross-border banks as to the US investment banks, whichhave either disappeared or sought the protection ofUS authorities by becoming regulated banks. In otherterms, the challenge facing policymakers, beyond theshort-term fixes announced in October, is to achieve


Focus

2 See also Alexandre Lamfalussy’s second Pierre Werner lecturedelivered at the Central Bank of Luxembourg, 26 October 2004;European Financial Services Round Table “On the Lead SupervisorModel and the Future of Financial Supervision in the EU”, June 2005;“Towards Further Consolidation in the Financial Services Industry”,European Parliament motion of 8 May 2006; Charlie McCreevy’skeynote address at the conference “Challenges for EU SupervisoryArrangements in an Increasingly Global Financial Environment”organised by the European Commission, 26 June 2007.


Focus

reregulation without refragmentation of Europe’sfinancial services industry.

Policy options and the case for EU level authority

How to square this circle has been discussed exten-sively among EU Member States since the crisiserupted, and even before. However, the discussion hasbumped repeatedly on the familiar obstacles to insti-tution-building at European level. In individual coun-tries, either central banks have a mandate to supervisebanks, or a separate agency, such as the FinancialServices Authority (FSA) in the United Kingdom,carries out the supervisory tasks in liaison with thecentral bank and government. When a bank or afinancial group expands operations to several coun-tries, the approach has traditionally been to introducecloser cooperation between the home-country super-visors and its counterparts in other countries.However, the degree of internationalisation of someEuropean banks, a feature that could only accentuateif cross-border integration does not stop, is difficult toreconcile with any primacy of the home-countrysupervisor in intergovernmental-style arrangements.This is especially the case where foreign-owned bankhave major systemic importance in host countries. Forexample, Fortis, the largest Belgian bank, is nowexpected to shortly become part of BNP Paribas. InFinland the market leader is Nordea, headquarteredin Sweden, and the number three is Sampo, owned byDanske Bank. In Poland, Pekao is the second-largestbank with a market share approaching 20 percent,higher than its parent UniCredit’s market share inItaly. In Romania, the three dominant banks areBCR, BRD and Raiffeisen Romania, respectivelycontrolled by Erste Bank (of Vienna), SociétéGénérale (of Paris) and Raiffeisen (of Vienna).

Having such systemically important local playerssupervised at group level by a foreign authorityaccountable to foreign stakeholders creates severepolitical tension, especially in times of stress. To alle-viate it, the European Commission has proposed tocreate ‘supervisory colleges’ under the auspices ofCEBS, for which an institutional underpinningwould be provided by the Capital RequirementsDirective (CRD), currently undergoing revision. Butthese colleges, which will further add to the alreadydizzying complexity of committees of regulatoryauthorities in Europe, cannot by themselves solvethe underlying tension. Moreover, they will intro-duce an element of competition among regulatory

structures which will mirror the competition amongbanks themselves. For example, in Finland Nordeawould be supervised by a college led by Sweden’sFinansinspektionen, Sampo by another college ledby Denmark’s Finanstilsynet, and OP-Pohjola byFinland’s own supervisor, Rahoitustarkastus. Evenadmitting close cooperation, the incentives to sup-port national players are such that this de facto reg-ulatory competition is likely to result in a ‘race to thebottom’ as often as in a “race to the top”.

Supervisory colleges are thus insufficient to resolvethe challenge of consistent and effective supervisionof Europe’s cross-border financial groups, and theirexpected reinforcement by the revised CRD, intend-ed as a step forward, is therefore in effect likely tocreate more problems than it can solve. CEBS, thecommittee of European banking supervisors, will notbe in a position to credibly ensure such consistencyand effectiveness, because as an advisory committeeto the European Commission it has no binding deci-sion-making authority of its own, and moreover itsmembers are precisely the national supervisors (andcentral banks) on which it would have to exert over-sight. For the same reasons that self-regulation in theprivate financial industry has been found wanting oflate, it is illusory to think that Europe’s nationalsupervisory authorities can be effectively disciplinedby a committee formed by themselves. A crediblesupervisory framework for cross-border financialgroups will require more than coordination mecha-nisms among national authorities. It will necessarilyentail actual delegation by national authorities ofkey tasks and responsibilities: such delegation toanother national authority accountable to its ownnational constituency is in many cases politicallyuntenable, but delegation could be done to a jointlygoverned actor in which each individual country hasa stake. In effect, the clear conclusion of the numer-ous multilateral discussions carried out in the pastfew months is that no sustainable framework for thesupervision of cross-border banks can be definedwithout corresponding new institutions at suprana-tional level.

If some form of EU-level supervisory authority isnecessary to meet today’s challenge of reregulationwithout refragmentation of Europe’s banking indus-try, would it be sufficient? As the events of the pastfew months have illustrated, government interven-tion in banking crises is multifaceted and ofteninvolves significant commitment of public money.There are currently no legally available instruments

to commit significant financial means for bankingrescues at EU level.3 On 1 October 2008, severalprominent economists have together called for theEU to pool financial resources to inject capital intoailing banks,4 and that same week the FrenchPresidency of the EU has briefly appeared to pro-mote similar ideas. However, one lesson learnt fromthe frantic days of early October is that even underconsiderable financial pressure, Member Statesremain very reluctant to share financial resources ina federal framework beyond what is already provid-ed for within the EU budget and the strictly definedmandates of the European Central Bank and theEuropean Investment Bank. This reluctance is cer-tainly linked to the absence of a tax-raising capacityat EU level, a key feature of the Union’s currentinstitutional arrangements. The same reasons makeit politically difficult to envisage “ex-ante” burdensharing arrangements as have been proposed byexperienced observers (Goodhart and Schoenmaker2006), which would provide ready-made formulasfor the sharing of public costs in case of a multina-tional public rescue.5

But while the absence of federal financial arrange-ments undoubtedly makes the management of bank-ing crises more difficult, it does not fatally under-mine the credibility of EU-level banking supervi-sion, assuming an EU-level supervisory authoritycan be created. This is because having centralizedinformation on financial groups would go a long waytowards resolving the coordination problems cur-rently resulting from geographical fragmentation,even in the absence of corresponding centralizationof financial resources. Anecdotal evidence suggeststhat information does not flow easily betweennational supervisors, especially during crises – evenif the official position of authorities is generally todeny the existence of such difficulties, for legitimateface-saving reasons. By contrast, when constrainedby necessity, national governments have provedready to decide quickly to pool financial resources injoint interventions, as has been the case for the res-cues of Fortis (by Belgium, Luxembourg and theNetherlands) and Dexia (by Belgium and France) inthe last days of September 2008. What has provenmost difficult in these two cases has not been thejoint commitment of funds, but the subsequent joint

management of a fast-changing situation, eventuallyleading, in the case of Fortis, to the decision by theNetherlands to nationalise the bank’s Dutch assetsseparately. This experience can certainly not be gen-eralised to all future cross-border banking crises, butit tends to suggest that the absence of ex-ante feder-al budgetary arrangements may not be an insur-mountable obstacle to efficient public action, if ade-quate analysis and recommendations for MemberStates’ decisions were to be provided by an authori-ty at EU level which would be sufficiently account-able to individual countries to be trusted.

The possible features of a European supervisor

Any debate about establishing new institutions atEU level is often emotionally charged, and easilycaricatured as the creation of all-powerful, opaquebureaucratic monsters (see Lodge 2008). A realisticblueprint for a European supervisory authoritywould need to carefully calibrate its mandate andsize in order to limit them as much as is compatiblewith its envisaged mission, in compliance with theEuropean principle of subsidiarity. It can be suspect-ed that were European leaders to decide to foundsuch a body, it would be under difficult conditionslinked to market or economic stress. For this reason,there is a case for a prior debate on what could bethe key features of an EU-level supervisory authori-ty, even at a time when the prospect for such a stepmay appear remote to many observers.

Mission

The organisation of financial regulatory and supervi-sory tasks varies considerably from one MemberState to another in the European Union. In someMember States, the central bank is in charge of mostor all of the supervisory function, as is the case inFrance or Italy, while in Germany supervision isdivided between the central bank and an indepen-dent agency (BaFin). Furthermore, in countries likeBritain supervision of individual financial servicesfirms is entirely separated from the central bank: it isone of the missions of the FSA. In a different direc-tion, the Netherlands have implemented the so-called ‘Twin Peaks’ model, in which conduct-of-busi-ness regulation and prudential supervision are car-ried out by two different agencies even though theyapply to the same regulated entities (see Taylor1995). Countries often reform their arrangements asa consequence of financial crises or financial scan-


Focus

3 Article 119 of the Treaty Establishing the European Community,which has been used to provide assistance to Hungary, cannot beused to intervene in private entities.4 Open Letter to European Leaders on Europe’s Banking Crisis:A Call to Action, http://www.voxeu.org/index.php?q=node/1729.5 See Cross-border Banking: Divided We Stand, The Economist,18 October 2008.


Focus

dals, as Britain did with the creation of the FSA in1997 following the successive failures of BCCI(1991) and Barings (1995). However, it does notseem that any one organisational model is unam-biguously superior to the others. For example, in thepast decade there has been a trend towards separat-ing supervision from central banking, emulating amodel initially pioneered by Scandinavian countriesfollowing the crises of the early 1990s. But the creditcrisis of 2007–08 has tended to shift the emphasisback to the linkages between the two functions.Spain, a country in which the central bank has exten-sive authority to supervise the banks, has been seento have recently avoided some of the mistakes madein neighbouring countries (Tett 2008).

Therefore, there is no single model that wouldimpose itself as the only possible one at Europeanlevel. However, the already mentioned subsidiarityconcern would suggest limiting the European super-visor’s mission to the core function of prudentialsupervision of financial services firms, with noresponsibility for other regulatory functions such asprotecting investors or ensuring market integrity orthe quality of financial disclosures.This does not nec-essarily mean that such other functions shouldalways remain at national level. For example, there isa strong case for transferring the enforcement func-tion for the implementation of InternationalFinancial Reporting Standards (IFRS) by Europeanlisted companies to a supranational entity account-able to national securities regulators (Véron 2007b).But there is no need for this function to be combinedwith that of prudential supervision. On the contrary,trying to bundle several different missions into a“single” European financial regulator would likelybe counterproductive.

The same considerations would weigh againstentrusting the European Central Bank (ECB) withthe role of European supervisor. Beyond the issue ofgeographical scope (see below), an EU-level central-bank-cum-financial-supervisor would arguably com-bine more tasks than the current EU institutionsallow for: debates about such a powerful institution’sautonomy from politicians and democratic account-ability, already heated in the case of the ECB, wouldbe further exacerbated, with the risk of endangeringthe independence of monetary policy. This is accen-tuated by the fact that supervision often has a closerworking relationship with elected governments thanmonetary policy does, especially when it comes tocrisis management in which public funds may be

used.There is no question that supervision should becarried out in close operational relationship with theECB and other relevant central banks, but the insti-tutions should be distinct.

Sectoral scope

The importance of an adequate scope of financial reg-ulation has been highlighted by the credit crisis, espe-cially in the United States where the unregulated“shadow banking system” has been found to interactwith banks in ways that have been sometimes detri-mental to financial stability. This debate is perhapsless acute in Europe, where investment banking oper-ations are generally included in the scope of supervi-sors together with retail banking. However, Europeancountries vary on whether they bring insurance underthe same authority as banking supervision, as hasbeen the general trend of reforms in the past fewyears. On the one hand, insurance companies have nothad the same systemic impact as banks in the devel-opments of the crisis so far; on the other hand, manyfinancial services firms have both banking and insur-ance activities, some of them of systemic relevance ona pan-European scale (such as ING).

A related question is whether currently unregulatedfinancial market participants, such as hedge funds,should be encompassed. This aspect is likely to bethe matter of discussion at international level. Theblueprint for a European supervisor could be corre-spondingly adapted to any enlargement of the sec-toral scope of financial regulation overall.

Geographical scope

The European supervisor should be an EU institu-tion, because only the EU offers a sufficiently stronglegal and political framework for such a key publicfunction to be carried out at supranational level.Thus,its authority could extend to the whole EU but notbeyond. For example, Swiss banks could in no case beunder its jurisdictions, even though bilateral agree-ments could be envisaged with the Swiss authorities,as currently exist between Switzerland and individualEU Member States, or with the United States.

Whether the supervisor’s geographical scope shouldbe less than the entire EU is a matter of political dis-cussion.There is no strong case to limit it to the euroarea, because banking integration and monetaryintegration seem to be two different dynamics. Mostcross-border banking groups in the EU combine sig-

nificant operations inside and outside the euro area:Santander, Nordea, UniCredit or Raiffeisen are typ-ical examples. Moreover, the centrality of London,as Europe’s financial hub and location of a largepart of wholesale operations for many of Europe’slargest banks, means that the benefits of poolingsupervision at supranational level would be signifi-cantly reduced if the United Kingdom were outsidethe geographical scope. As a consequence, there is astrong case for a scope which would include all EUMember States.

Division of scope with national supervisors

The subsidiarity principle dictates that a Europeansupervisor should have jurisdiction over only thosebanks and financial groups that can not adequately besupervised by national authorities acting alone. Underthat principle, only large financial firms with signifi-cant operations in several countries should be super-vised at EU level, leaving the vast majority ofEurope’s 8,000-odd banks with an unchanged, nation-al supervisory framework. The cut-off betweennational and EU-level supervision could be definedeither by compulsory thresholds – any bank above agiven size and a given degree of internationalizationmeasured by key operating parameters such as rev-enue, employees or assets, say, would be included inthe European supervisor’s jurisdiction – or on a vol-untary basis, supposing that the cost savings resultingfrom one single supervisory partner would be a suffi-cient incentive for all relevant groups to submit them-selves to EU-level oversight. The second option, vol-untary registration, would introduce an element ofregulatory competition between the national andEuropean levels, whose possible consequences wouldneed to be carefully weighed. In any event, the num-ber of groups subject to EU-level supervision can beexpected to be no more than a few dozen in total, thuspreserving a major role for most existing nationalauthorities.

Financial groups supervised at EU level may be sub-ject to a specific legal regime in order to establish aproper basis for their supervision and ensure a suffi-ciently level playing field.This could take the form ofa European banking charter, supported by appropri-ate EU legislation (Decressin and Cihak 2007).

Functions and organisation

The core of a European supervisor’s role should bean unconstrained ability to collect and analyse infor-

mation from the supervised entities. To have credi-bility, it should have direct access in principle tosupervised firms to obtain information, includingpossibly by sending inspectors on site within its geo-graphical scope. In practice, however, it should lever-age a working relationship with, and delegation to,national supervisors in order to benefit from theiradvantage of proximity, which can greatly facilitateboth access to information and understanding of it.How concretely and exactly this relationship can beorganised would be an important operational aspectof setting up a European supervisor.

There is also an important debate required on whattools the European supervisor should have beyondaccess to supervisory information and the means toanalyse it. How far to go in entrusting it with coer-cive powers on the behaviour of supervised entities,as well as a formal role in crisis management and res-olution, would also depend on the legal basis for itsestablishment, the discussion of which is beyond thescope of this contribution. But even a role limited tothe collection and processing of information at EUlevel would be a substantial addition to the currentsituation, in which no single public authority canbuild a complete picture of the European operationsof a large, cross-border financial group.

A European supervisor would also have a role toplay if, as could be decided as a consequence of thecrisis, it is decided to introduce a mechanism of“dynamic provisioning” to allow the assessment ofcapital adequacy to take into account the effects ofmulti-year financial cycles. If part of a global supervi-sory framework, such calculation may need toinclude a regional as well as a local/national and pos-sibly a global component, in whose determinationsupervisors could play a role together with otherinstitutions including central banks.

Governance and funding

As already mentioned, compared with monetarypolicy, supervision of financial firms is a role thatinvolves closer interaction with governments andespecially treasury departments, as crisis manage-ment may involve decisions that directly affect indi-viduals and firms as well as the direct use of taxpay-ers’ money – a fact that recent developments of thecredit crisis have amply illustrated. Therefore, thegovernance of a European supervisor should be con-ducive to the establishment of strong and trustfulrelationships with national governments, especially


Focus


Focus

as the prospect of EU-level financial resources forintervention in the banking system remains a remoteand unlikely one at this point of time.

This would mean that the European supervisorshould be accountable to the European Parliamentand Council, but perhaps also build direct account-ability towards national constituencies and especial-ly national parliaments, at least as far as countriesare directly affected by its decisions.

Similarly, one could imagine direct funding of thesupervisor’s yearly budget (which should probablyremain modest in any case, compared to that ofother EU institutions) by Member States. Private-sector funding of supervision, which exists in severalEU Member States, could also be envisaged as acomplementary or alternative route, with due con-sideration of the risk of capture of the supervisor bythe supervised industry that it may entail.

Location

Where to headquarter a European supervisor maysound an ancillary and secondary concern, but experi-ence suggests that it could be a significant aspect of thediscussion about its establishment. While this decisionwill depend on the inevitable horse-trading amongmember state, it should also be made to maximise theeffectiveness and/or legitimacy of supervision.

To that end, three possible locations come to thefore: London, where a significant part of Europeanbanks’ riskiest and most complex activities takeplace; Brussels, the centre of political power atEuropean level, where the European Parliament inparticular holds most of its hearings and whereCouncil committees meet most often; and Frankfurt,seat of the European Central Bank with which aEuropean supervisor would need to establish a closeworking relationship.

Notwithstanding political considerations, no otherplace in Europe comes close to these three in termsof suitability to host a European supervisor. Thechoice among the three is far from obvious and willpartly depend on choices on other aspects of the newbody’s setup.

The political outlook

The decision to establish a European supervisor forEurope’s largest cross-border financial groups, even

with a limited mandate, would be a significant stepcompared with previous developments of Europe’sfinancial regulatory framework.This is not a decisionthat policymakers can take lightly. It may also be thecase, depending on the legal basis chosen for such adecision (an aspect whose discussion is not includedin this paper), that such a decision would requireunanimity of EU Member States.

In the past, opposition to such a project has been dri-ven in part by national authorities defending their tra-ditional turf, and by national reluctance to delegatesovereignty, but also by concerns that financial sectoroversight was too delicate a matter to be dealt with inthe complex context of the European Union, with itsmultiple institutional constraints. By itself, so theargument goes, an EU institution could not be reac-tive, efficient and accountable enough to properlyperform the supervisory tasks in the short timeframesimposed by the marketplace. This is an importantargument which cannot be ignored. However, theobstacle it creates to reform appears less insurmount-able since the ECB, by its deft handling of the finan-cial crisis from the early days of August 2007, hasproved that an EU institution could be as reactive anddecisive as any other in times of stress.

A parallel argument is that EU institutions may bytheir very nature have a more heavy-handed, pre-scriptive, rules-based and politicised approach thannational ones. However, especially in the currentcontext of reregulation, this argument does not haveany strong empirical basis. On the contrary, it couldbe argued that placing supervision of certain finan-cial groups at EU rather than national level reducesthe likelihood of privileged special treatment thatcould be detrimental to overall financial stability,and taxpayers’ interests. This is actually a generalfeature of institutional regulatory arrangements, ashas been described before the eruption of the finan-cial crisis by Adam Posen, in a parallel observationof the American and European experiences: “whenUS economic policy decisions were left to the statelevel, they tended to be reactionary, just as they wereon civil rights. […] the more the central body has hadauthority over economic policy, the greater the liber-alizing influence […]. The alternative to a strongBrussels is not a decentralized free market and min-imal government interference. It is greater politicalcapture of economic policymaking and abuse ofauthority by member states and subnational govern-ments. […] In fact, the European experience showsthat enforcement of market integration, competition

policy, disclosures and transparency are what reallybrought the economic benefits of European Union –and that enforcement came during the periods whenand in areas where the Commission had competenceindependent of the member states’ specific wishes.When Brussels has been strong, it has been liberaliz-ing, at least internally within the European Union,and that has paid off. When the member states havetaken away authority from Brussels, the effect hasbeen reactionary horse-trading and back-scratching[…]” (Posen 2007).

A frequently mentioned concern is that any creationof a new financial authority at EU level, howeverrestricted its initial mandate, would automaticallylead to “mission creep” and undue expansion of thatmandate to an extent that would suffocate and side-line national financial regulation. This is, however, amatter that can be addressed by adequate legal pro-visions. One could imagine, for example, that the spe-cific description of the mandate of the Europeansupervisor in its articles of association (or otherfounding document) could be modified only by asupermajority or even unanimity of the memberstates. The decade-long experience of MemberStates with the ECB illustrates that, even in caseswhen the temptation of mission creep may be pre-sent – as is seen by some to have been the case, say,with the Target 2 Securities project – there can beenough institutional safeguards for institutional driftto be safely prevented.

A different line of argument against the creation of aEuropean supervisor is that it would not be sufficientto meet its policy aims. This argument runs in twostrands: first, that without supranational financialresources the creation of a supranational supervisoryinstrument would have no material effect; second,that cross-border financial integration is happeningat global level and that its challenges cannot be metunless global supervisory institutions are dulyempowered. Both points evidently contain an ele-ment of truth, but not to the extent that they wouldjustify policy inertia.As previously argued, even with-out federal financial resources a European supervi-sor could significantly alter – and improve – thedynamics of decision by providing ‘the necessary ser-vices of multilateral coordination and politicalbuffering’ (Pauly 2007) which no existing structurecan offer in the present setting, and which can con-tribute to substantially reduce the eventual price tagto taxpayers by assessing the risks early enough andaccelerate decision-making among member states.

On the latter item, it must be noted that even if a fewEuropean banks, mainly headquartered in Spain andthe United Kingdom, have very significant retailoperations overseas, the bulk of cross-border expan-sion of Europe’s large banks over the past decade orso has occurred inside the EU. More to the point, theEU offers a framework of law and political account-ability that has no equivalent at global level. And aEuropean supervisor would be in a much better posi-tion than a fragmented collection of national ones tonegotiate internationally, be it in international discus-sions about shared norms (such as the Basel accordon capital adequacy, or IFRS) or in multilateral onesabout particular financial groups, those headquar-tered in Europe with overseas activities as well asthose headquartered outside the EU which havedeveloped significant European operations.

Conclusion

The establishment of a European supervisor for thelargest cross-border financial groups in Europe wouldbe a significant step towards enabling the necessaryreregulation of the European financial system in thewake of the ongoing crisis, without leading to eco-nomically (and politically) harmful financial refrag-mentation. The deep-running scepticism and resis-tance to such a project, which are more marked insome member states than others, need to be balancedwith the present dangers inherent in the status quo.

Would a European supervisor as is advocated in thispaper have made a difference, had it been in placefrom the outset of the current crisis? Certainly not forthe likes of Sachsen LB, IKB, Northern Rock or HypoReal Estate, whose supervisory regime would nothave been affected in view of their predominantlydomestic activity. However, supranational supervisionmay have helped react earlier to the developments atDexia, Fortis, and possibly other banks which have notbeen through changes as dramatic but have had toface significant upheavals nevertheless. More impor-tantly still, it would have allowed more comprehen-sive and comparative information of national policy-makers in the context of the massive intervention plandecided by European governments in mid-October,whose implementation is permanently at risk of run-ning divergent courses in different countries, to prob-lematic overall effect.

But once again, the main outcome at stake is a mid-term one rather than immediate crisis management.


Focus


Focus

The persistence and development of pan-Europeanfinancial groups, in an era of reregulation, willdepend crucially on the existence of a crediblesupervisory framework which cannot be providedin the absence of institution-building at EU level.This is also a matter of global competition. Takeother, non-financial sectors: in the automotiveindustry, often a symbol of economic nationalismbut one in which the European market is genuinelyintegrated, five of the world’s top ten firms6 areheadquartered in Europe. By contrast, in the mediaindustry, which has a long history of fragmentationalong national lines, Europe is home to only two ofthe top ten. Will European banks in the future bemore like its automakers, or its media companies?The answer to this question will depend largely ondecisions made by governments about their super-visory framework.

The challenge in Europe echoes a global debate.Global financial firms have been key agents of theglobalisation of the past ten or twenty years. If theyretract to a national or regional playground, globalisa-tion may be deprived of one of its key supporting struc-tures.The world has remembered the mistake that wasthe Smoot-Hawley Tariff Act of 1930, and is unlikely toraise barriers to trade as a knee-jerk reaction to the cri-sis – but a refragmentation of the global economiclandscape may also take other forms than eightdecades ago (Bowers 2008). Evidently, policymakershave a duty to focus on the short-term aspects of thefinancial and economic crisis. But in so doing, they mustalso keep in mind the long-term consequences.

References

Bowers, D. (2008), “Nationalisation of the Banks Will Only HarmGlobalisation”, Financial Times, 16 October.

Decressin, J. and M. Cihak (2007), The Case for a European BankingCharter, IMF Working Paper WP/07/173.

Decressin, J., H. Faruqee and W. Fonteyne (eds. 2007), IntegratingEurope’s Financial Markets, Washington DC: International Mone-tary Fund.

Goodhart, C. and Schoenmaker, D. (2006), Burden Sharing in aBanking Crisis in Europe, LSE Financial Markets Group SpecialPaper Series 164.

Larsen, P. T. and S. Tucker (2008), “Cash Injections into RBS Puts Future of Overseas Units in Doubt”, Financial Times,14 October.

Lodge, O. (2008), “Watching the Crusade for Super-regulator”,Financial Times, FTfm Supplement, 16 June.

London Economics (2002), Quantification of the MacroeconomicImpact of Integration of EU Financial Markets, Final Report to theEuropean Commission – DG Internal Market, London.

Pauly, L. (2007), Political Authority and Global Finance: CrisisPrevention in Europe and Beyond, Global Economic GovernanceWorking Paper WP 2007/34, Oxford University.

Posen,A. (2007),“Liberalism Needs Central Power”, Financial Times,4 July.

Taylor, M. (1995), “Twin Peaks”: A Regulatory Structure for the NewCentury, London: Centre for the Study of Financial Innovation.

Tett, G. (2008), “Time for Central Bankers to Take Spanish Les-sons”, Financial Times, 30 September.

Véron, N. (2006), Farewell National Champions, Bruegel PolicyBrief 2006/04.

Véron, N. (2007a), Is Europe Ready for a Major Banking Crisis?,Bruegel Policy Brief 2007/03.

Véron, N. (2007b), The Global Accounting Experiment, BruegelBlueprint Series, April.

Véron, N. (2008), The Challenges of Implementing the Brown–Sar-kozy Plan, www.bruegel.org.

6 Ranked by market capitalisation as of 30 September 2008. Source:Financial Times Global 500 database (www.ft.com).


Focus

WE NEED MORE OVERSIGHT

FRIEDRICH L. SELL*

One important international institution that appar-ently was not heard much from, both before and atthe peak of the financial market crisis, is theInternational Monetary Fund (IMF). But this is notentirely fair. Already at the end of March 2008 theIMF forecasted that banks would need to write offone trillion dollars worldwide. The IMF is indispens-able in the opinion of many economists: without theIMF, financial markets cannot be properly super-vised. In the following, we examine how the IMF’srole in the world economy has changed dramaticallysince it was installed as a Bretton Woods institutionin 1944. We also review key elements of the actualworldwide financial and economic crisis and suggestwhy the central banks and or national supervisionentities alone are (and will be) unable to perform thefunction of supervising the financial markets. Thereason lies in the global externalities produced bythe banking sector under stress which go beyond thescope of national or even regional banking supervi-sion institutions.These externalities call for the coor-dinated action of an international governmentalagency, such as the IMF.

Recently, just before the 2008 IMF-World BankAnnual Meetings on October 11–13, IMF ManagingDirector Dominique Strauss-Kahn and his DeputyManaging Director, John Lipsky, issued a statement.The general tenor was cautiously optimistic. Theworld economy needs not necessarily slip into a deeprecession if the central banks continue to succeed inproviding the banks with sufficient liquidity at mod-erate interest rates; if solvent financial institutionsreceive the influx of capital they require withoutbeing nationalised; and if the modified rescue planTARP, approved by the US Congress, is rapidlyimplemented by the US government. The US gov-ernment’s gigantic bail-out plan was not enthusiasti-cally received by the IMF, but as it stated: “a non-perfect plan is better than no plan at all”.

Before looking at the possible contribution of the

IMF to overcoming the worldwide financial market

crisis, a look at the history of this institution is in

order. The IMF as well as the World Bank were cre-

ated at the conference in Bretton Woods. The IMF

then became a pillar of the post-war monetary sys-

tem. In the gold-dollar standard that applied until

1971, parity adjustments vis-à-vis the US dollar was

only allowed in cases of “fundamental disequilibri-

um in the balance of payments”. Parity changes of

more than ten percent had to be approved by the

IMF. Member states were obliged to pay into the

IMF in the form of gold and/or currency reserves in

accordance with their national quotas. The main task

of the IMF consisted in the granting of short-term

credits for financing balance of payments deficits.

That also did not change after 1973 when member

states were given the freedom to determine their

own exchange rate regimes.

As late as the 1990s, when the IMF assembled rescue

packages to overcome the currency and financial

market crises in Mexico (1994), Thailand (1997),

Brazil (1998) and Russia (1999), the IMF seemed to

be indispensable as an international fire brigade.

However, even at the time critics warned that the

sheer existence of such a fire brigade could lead to

fires being set in the affected countries. A telling

point in connection with the present banking crisis in

the United States and Europe is that during the

Asian financial crisis in the late 1990s, touched off by

the bursting of a real-estate bubble, the IMF called

for the rigorous closing of most financial institutes

that had financed the real-estate sector, largely in the

style of American investment banks.

In the new millennium much – though not every-

thing – has changed. Since the collapse of the

Argentine Currency Boards in 2001, there has been

no single emerging economy that has even come

close to encountering a balance of payments crisis.

With the continuously declining share of “semi-

fixed” exchange rate regimes in the world economy,

the likelihood of speculative attacks has clearly

declined. Only very recently a number of currencies

of central and eastern European emerging* University of German Armed Forces Munich.


Focus

economies have come under pressure. They are can-didates for the euro and their likely qualification istested by speculators, among other things, by theircapability to maintain their respective exchangerates in the band provided by EMS II. The currentstep taken by the Argentinean government tonationalise the formerly private system of pensionsundermines trust in the Argentinean currency. All inall, however, it appears that the quite successful com-mercialisation of the balance of payments equilibri-um and the observable changes in behaviour ofimportant players in the world economy may haverendered the old-style IMF (almost, if we think ofthe current problems which Pakistan has to face)superfluous.

The most important emerging economies today – inAsia: China and India, in Latin America: Brazil –have no need for (new) loans from the IMF. To theextent that these countries were at all debtors of theIMF, in the meantime they have paid back theseloans. For years they have been accumulating cur-rency reserves, above all China, to an extent that haslong since exceeded the precautionary holdings ofrisk-averse central banks.

The IMF has decided – finally, one could say – tomonitor more closely the exchange rate policies ofits 185 member states in keeping with its surveillancetasks, and in bilateral ad-hoc consultations to pub-licly denounce manipulations of exchange rates andstrategic currency market interventions. Obviously,as Strauss-Kahn and Lipsky have also determined,inflation has clearly re-emerged in the world econo-my – more strongly in the emerging than in theindustrial economies – and the dramatic increases inthe prices of oil and food have led to clearly wors-ened supply conditions in most member countries, atleast until the middle of 2008. Since then, we observethat the pressure on the markets in concern has sig-nificantly eased. Especially the oil prices have comedown substantially again. An important, if not deci-sive, explanatory factor here is the clear weakeningof economic growth in the world economy in thewake of the financial market crisis. But now OPEC isready to cut back the oil supply. They announcedthat the daily production would be reduced by1.5 million barrels from November 1 onwards. Howlong then can the “baisse” of energy prices truly last?Also, the inflation trend in emerging economiescould paradoxically even facilitate the supervisoryfunction of the IMF: in light of rising inflation rates,China and other important emerging Asian

economies have in the meantime already made theirexchange rate policies more flexible.

This reminds us that the Bretton Woods currencyregime failed, at the beginning of the 1970s, not leastbecause of the worldwide spread of inflation. For thisreason a Bretton Woods II, as proposed several yearsago by experts such as Stanford University’s RonaldI. McKinnon and that has been temporarily attempt-ed by several East Asian emerging economies bymeans of pegging their currencies to the US dollar –albeit at most only an approximation in practice – isno longer to be expected. What the experts intendedwas a renewed hardening of parities of importantcurrencies to each other. It is also not difficult toimagine the conflicts between market and exchange-rate stabilisation as well as additional inflation risksand inefficiencies that would come from exchangerate targets and corresponding currency marketinterventions of the important central banks in thecurrent, by-far-not-resolved financial crisis.

In the future we are more likely to see a scaled-downversion of Robert A. Mundell’s vision of a one-cur-rency world economy.There will be significantly fewercurrencies than at present because more regional cur-rency unions – patterned after the EuropeanMonetary Union – will be established, as can alreadybe observed in efforts in this direction in the MiddleEast, Asia (ASEAN) and in Latin America(Mercosur). The remaining currencies will then main-tain largely flexible exchange rates with each other. Inthis case the possibilities of incurring severe balanceof payments crises will be largely reduced. For thisreason the IMF will have to take on new areas ofresponsibility.

Instead of the traditional monitoring of the perfor-mance of individual countries, it would be advisablefor the IMF to instruct the governments of industrial,emerging and developing economies as to how eco-nomic policy should suitably react to global supplyand demand shocks. Experience from the 1970sdemonstrates that an expansive fiscal policy in con-nection with a (too) restrictive monetary and aggres-sive wage policy was precisely the wrong response tothe two oil-price shocks. When an indispensable pro-duction factor such as energy becomes more expen-sive, it is imperative to restrict the price increase of theother factors if we wish to avoid an extended stagfla-tion. What is needed is a cut-back of government bor-rowing on the capital markets, moderation in wagedemands and possibly even a slightly inflationary


Focus

monetary policy that corrects the insufficiently mod-erate nominal wage agreements because such a policyreduces the increase of real wages.

Everything has become exceedingly more complicat-ed – though we now have breathing space withregard to the above-mentioned worsening on thesupply side – as we now have, in addition, a sharpdecline in the prices of real estate not only in theUnited States but, for example, also in Britain, Spainand Ireland and because the world economy is at thebrink of a huge recession. But the financial marketcrisis is the biggest problem facing economies today.It is still true that monetary policy and wage bar-gaining must respond in the way described above ifa deep and long-lasting recession in the world econ-omy is to be avoided. The gigantic fiscal efforts con-tained in the Paulson-Bernanke plan, TARP, to cor-rect the balance sheets of US banks will boost capi-tal market interest rates in the United States sooneror later. The same applies to the rescue packagesdecided by European governments. The fact thatgovernments seem to be (at least temporarily) nolonger – or not to the same extent as before – con-fronted with the above-described supply shocks willnot alleviate the pressure on wage bargaining andmonetary policy. In addition, there are problematicsignal effects that are sent out by any bail-out. Suchsignal effects affect Europeans directly. TheMaastricht Treaty contains a no bail-out clause vis-à-

vis highly indebted member states in the euro area.Rescue packages like that of Paulson-Bernanke and,more recently, of European governments underminethis credibility.

It remains the responsibility of the IMF to monitorcritically the development of world economic imbal-ances and to provide recommendations for action. Aparticular focal point in recent years has been thegrowing US current account deficit.The key elementin this development is ultimately how the savings-investment gap in the United States in comparisonto the rest of the world develops in the future. It isvery likely that the private sector in the UnitedStates, under the pressure of the financial market cri-sis, will finally undertake significant savings efforts.A decline in private investments is to be expected inthe first place because of the economic downswing,but one can also expect a rise in the financing costsfor the medium- and long-term procurement of cap-ital for US enterprises from outside sources. Afterall, foreign investors should ask for a higher risk pre-mium in light of the experiences made in the new

millennium. That could, in principle, contribute to areduction of the US current account deficit in con-nection with increased savings activity.

A countervailing factor, however, is that the pro-curement of equity capital on the stock market bylarge US banks as well as the huge borrowing of theAmerican government in connection with thePaulson-Bernanke plan will draw much foreign cap-ital into the country. The US financial centre nowcarries with it clearly higher risks, and, as statedabove, foreign creditors will demand significantlyhigher compensation. All in all, a clear decline inimbalances in the world economy in the wake of thefinancial market crisis is hardly to be expected.Indeed, the IMF considers as a possibility an alarm-ing weakening of capital flows to emerging anddeveloping economies.

IMF director Strauss-Kahn recently proposed a“global dialogue” to cope with the financial marketcrisis and the persisting imbalances in the worldeconomy. But such a dialogue will not be enough.Countries whose exchange rates have been manipu-lated should conclude binding target goals with theIMF regarding the swift dismantling of existinginterventions. But the IMF must also take a moreactive role in surveilling the international financialmarkets and its participants. In a global world econ-omy the national central banks and financial marketauthorities alone are clearly not up to this task. WeEuropeans have experienced this first-hand, forexample, when Ireland attempted to go it alone inshoring up its banking system and as other countriesthen followed suit. But it could also be seen that theturmoil on the worldwide financial markets wasfuelled temporarily by the transitory incapability ofEuropean governments to follow quickly the stepstaken by the US government.

The fundamental reason for the inappropriatenessof national or even regional supervision solutions issimple: the reason lies in the global externalities pro-duced by the banking sector under stress which gobeyond the scope of national or even regional bank-ing supervision institutions.And, as the Irish case hasshown, not only the banking sector but alsoautonomous action taken by national governmentsthemselves gives rise to negative spillover effects.These externalities call for the coordinated action ofan international governmental agency such as theIMF. The IMF has a “comparative advantage” in ful-filling this task given the fact that almost all relevant


Focus

countries in the world are members of this institu-tion, that this institution has a successful tradition inmonitoring economic policy around the world (see,for example, its regular publication, the “WorldEconomic Outlook”) and finally because the IMFhas a long-lasting record of multilateralism, which iswhat is most needed in the current crisis.


Focus

THE FINANCIAL CRISIS IN

JAPAN – ARE THERE

SIMILARITIES TO THE

CURRENT SITUATION?

JOHANNES MAYR*

In the 1990s Japan experienced a deep financial cri-sis that lasted for more than a decade and whoseeffects strain the Japanese economy even today. Inthis article, we give a brief overview of the majordevelopments that preceded the crisis and describethe legislative actions that were decided to ease themarkets and to restore confidence in banking.Employing major financial and economic indicators,we compare the developments in Japan to the cur-rent turmoil on the international financial marketsthat originated in the US subprime mortgage mar-ket. The comparison shows that the following phe-nomena played a crucial role in the run-up to bothcrises:

– an expansive monetary policy– a strong increase of prices on the real estate and

stock markets– a fall in savings of private households– high risk propensity and comparable low equity

ratios in the banking sector– the development of new financial products com-

bined with a comparable low degree of marketregulation

The run-up to the crisis

The economic situation in Japan in the 1980s wascharacterized by high GDP growth rates and adecreasing level of inflation (Figure 1). Followingthe Plaza Agreement the Japanese yen appreciatedfirst gradually and later strongly.1 To prevent a fur-ther appreciation and to support the exporting sec-tors of the economy, the Bank of Japan lowered itsinterest rates subsequent to the Louvre Agree-ment to 2.5 percent (Figure 2).2 The expansivemonetary policy and the resulting increase in liq-uidity led to a boost in prices on the real-estateand stock markets.

The safety system of the banking sector was initiallycharacterized by the Convoy System (Hoshi 2002).Originally, the Convoy System was designed torebuild Japan after World War II. Banking supervi-sion and regulation was conducted such that the via-bility of the weakest banks was not undermined. Itwas implicitly understood that the banking sectorwas fail-safe, as the Japanese Ministry of Finance wasexpected to step in to find a remedy for any problemthat could threaten the viability of a bank. In returnfor protection by the financial authorities, bankswere expected to function as financial intermediaries

* Ifo Institute for Economic Research.1 The Plaza Agreement was signed on 22 September 1985 at the PlazaHotel in New York City by 5 nations – France,West Germany, Japan,the United States and the United Kingdom. The five agreed todepreciate the US dollar in relation to the Japanese yen and Germandeutsche mark by intervening in currency markets.2 The Louvre Agreement was signed by the then G-6 (France, WestGermany, Japan, Canada, the United States and the UnitedKingdom) on 22 February, 1987 in Paris. The goal of the LouvreAgreement was to stabilize the international currency markets andhalt the continued decline of the US dollar caused by the PlazaAgreement.

-5

0

5

10

15

20

25

1970

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

JAPAN

Source: IMF International Financial Statistics; OECD Main Economic Indicators.

Annual real GDP growth rate and annual CPI inflationin %

GDP

Real effective exchange rate

0

20

40

60

80

100

120

1970

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

index

CPI inflation

Figure 1


Focus

channelling surplus household savings into theindustrial sector. To guarantee safety for the finan-cial system, the Deposit Insurance Cooperation(DIC) provided a payoff in which a failed bankwould be closed down for liquidation and a deposi-tor with the failed bank would be protected by up to10 million yen per depositor. Furthermore, the DICprovided financial assistance by transferring thesound assets and liabilities to an assuming bank. Aslarger banks generally perceived to be protected bythe Convoy System, the DIC insurance fund had avolume of only 300 billion yen in 1987, what was fartoo little in the case of the failure of a major bank(Nakaso 2001, 3).

As a consequence of the opening of the Japanesefinancial markets for foreign investors in the early1980s, domestic regulation and the by then dominatingmain bank system3 came under increasing competitivepressure. The Japanese banks broadened their shares

of risky portfolio assets in order to compensate losses of marketshares and significantly loweredtheir equity ratio. This resulted in adecrease of interest margins and ofthe interest spreads between loansof different ratings. Neglectingtheir profitability and riskiness,loans were extended even at nega-tive lending spreads and the vol-ume of credits increased steadily(Figure 3). Additionally, the bankswidened the range of acceptedsecurities, especially for mortgageloans on real estate. This develop-ment was intensified by the forma-tion of new mortgage banks out-

side of the traditional banking sector. These so-calledJusen or housing loan corporations were non-bankfinancial institutions that were founded by banks andother financial institutions in the 1970s to complementthe housing loans offered by banks. In the 1980s, theJusen companies shifted their lending towards real-estate developers. However, the banking supervisionwas not adequately adjusted to the new structure ofthe financial markets.

The beginning of the crisis

As a reaction to the overheating of the real estate andstock markets – the Nikkei Index rose from 13,000points (end of 1985) to 39,000 points (end of 1989) –numerous measures were implemented to guaranteea soft landing of the economy. The Bank of Japantightened its monetary policy and successively raisedthe discount rate from May 1989 to August 1990 from

2.5 to 6 percent. Furthermore, atightening of the modalities forloan granting on the real-estatemarket and a tax on speculationgains were introduced. The rise ininterest rates led to a significantincrease in refinancing costs. Togenerate liquidity, investors soldtheir assets, and the prices on the

1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000

0

1

2

3

4

5

6

7

8

9

10

1993 1995 1997 1999 2001 2003 2005 2007

Source: Bank of Japan and Federal Reserve Board.

years

US

Federal funds target rate

1993-2008

INTEREST RATES IN JAPAN AND THE US

Japan

basic discount rate

1978-2001

years

in %

Figure 2

-3

-2

-1

0

1

2

3

-10

-5

0

5

10

15

20

1985 1987 1989 1991 1993 1995 1997 1999 2001

Source: IMF International Financial Statistics.

CREDIT VOLUME AND LENDING SPREAD

LHS: Lending Spreada)

a) Average interest rate on loans - CD quotatins (3-month)

RHS: Annual change in volume of loans

in %in %

Figure 3

3 A main bank relationship means that afirm meets a substantial proportion of itsfinancial needs through the intermedia-tion of one bank. In return for the prefer-ential business that it receives from thefirm, the main bank implicitly undertakesthe monitoring of the firm and bears theresponsibility for organizing expensivedebt workouts in case the firm encountersfinancial distress.


Focus

real-estate and stock markets dropped sharply untilmidyear 1992 by 65 and 75 pecent, respectively(Figure 4).

Thereupon a considerable part of the overall volumeof credit was rated as NPLs (non-performing loans)and important credit rating agencies downgradedthe Japanese banks, which again put further pressureon equity prices.

The height of the crisis

Between 1994 and 1996, a number of major Japanesebanks became insolvent and, following the drop inreal estate prices, a growing number of Jusen compa-nies encountered difficulties and accumulated lossesthat were far beyond the amounts that founderbanks could cover. The financial crisis hit its peak inNovember 1997 when major financial institutionscollapsed almost on a weekly basis. As a result, for-eign financial institutions squeezed their credit limitsto Japanese banks in general and domestic lenderbanks increasingly placed their money with the Bankof Japan instead of offering it on the interbank mar-ket. As a result, liquidity in the interbank marketdried up and interest rates came under strongupward pressure. These developments forced theBank of Japan to pump massive liquidity into themarket.

To restore financial system stability and in order toavert a run on banks, the Japanese government decidedto introduce further public funds and gave a blanketguarantee for deposits and other liabilities of financialinstitutions for a period of 5 years. From 1992 to 2002,180 deposit-taking institutions were dissolved under thedeposit insurance system (Figure 5).The DIC providedfinancial assistance for the assuming institutions ofabout 19 trillion yen.4 More then 10 trillion yen werepaid by the taxpayers and the rest was funded by pri-vate financial institutions through the deposit insurancesystem. Furthermore, the government injected about12 trillion yen into financial institutions by purchasingpreferred or common stocks and extending subordinat-ed loans (DICJ 2006).All issues related to capital injec-tions were handled by the newly created FinancialCrisis Management Committee (FSA). The volume ofNPLs continued to rise and reached its peak of 43 tril-lion yen or 8.4 percent of the total credit volume of allbanks not until the end of March 2002 (Figure 6).

The aftermath of the crisis

Japanese banks had only limited incentives to removetheir NPLs from the balance sheet. The Bank of Japanmaintained its loose monetary policy, which reducedthe opportunity costs of holding bad loans, and sales of

0

10 000

20 000

30 000

40 000

1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000

1 000

4 000

7 000

10 000

13 000

16 000

19 000

22 000

25 000

1993 1995 1997 1999 2001 2003 2005 2007

PRICE INDICES

Source: Dow Jones and Reuters; Ministry of Land (Japan); Office of Federal Housing

Enterprise Oversights and Standard & Poor's (US).

Stock markets

LHS: Nikkei Index 225 (1 978-2001)

Real estate markets

1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000

-20

-10

0

10

20

1993 1995 1997 1999 2001 2003 2005 2007Annual rate in %

RHS: Dow Jones

industrial index

(1993 -2008)

US

Case Shiller house

price index (1992-2008)

Japan

average land

price index

(1978-2001)

US

OFHEO house price index

(1993-2008)

years

years

years

years

Figure 4

0

10

20

30

40

50

60

1992 1994 1996 1998 2000 2002 2004 2006

0

1

2

3

4

5

6

FINANCIAL ASSISTANCE

Source: DICJ Annual report 2006; Financial Services Agency.

JPY Trillion

RHS: Volume of Grants

0

10

20

30

40

50

1999 2000 2001 2002 2003 2004 2005 2006 2007

JPY Trillion

LHS: Number of Cases

number

VOLUME OF NON PERFORMING LOANS

(NPLS)

Figure 5

4 The average annual nominal GDP of Japan between 1995 and1998 was about 500 trillion yen.


Focus

bad loans were only possible at deeply discountedprices.Thus, in order to fulfil short-term profit expecta-tions the banks used different accounting methods tocover the total need for write-downs. To accelerate theprocess of identifying the bad assets and the disposalsthereof, the FSA conducted several rounds of specialinspections of major banks and later of smaller andregional institutes from 2001 onwards.Additionally, thegovernment installed the so-called Resolution andCollection Cooperation (RCC) to assume failed creditcooperations and to purchase NPLs from failed finan-cial institutions as well as from solvent operating banks,helping them to clean up their balance sheets.

The fall of asset prices implicateda decline of the equity ratios ofJapanese banks, which tightenedcredit conditions for smaller aswell as for large companies tomeet the minimal capital require-ments agreed upon in the Basel Iaccord5 (Figure 6). After theirgrowth rates had already fallen

sharply since the increase of the interest rates in 1989,the volumes of total credit declined from the end of1998 onward. The credit conditions were not gradual-ly untightened until mid-1999.

Comparison

A comparison of the Japanese financial crisis with thecurrent financial and banking crisis shows a numberof similarities (Reinhart and Rogoff 2008). In bothcases, very low interest rates promoted a huge run-upin prices on the real-estate and stock markets.

As in the current situation, where declining lend-ing standards in the mortgages markets, anincrease in loan incentives and a long-term trendof rising housing prices had encouraged US pri-vate households with a low degree of creditwor-thiness to assume difficult mortgages and run intodebt, the starting point of the Japanese financialcrisis was the liberalization and deregulation ofthe real-estate and mortgage market. In bothcases, these developments came along with aninsufficient adjustment of the financial supervi-sion to the new structure of the financial and real-estate markets.

The delayed tightening of monetary conditions leadin both cases to a strong adjustment of asset pricesresulting in striking contractions in wealth and equi-ty capital followed by increases in risk spreads andfinally a liquidity crisis. In contrast to the current cri-sis, where the securitization of debts has lead to aninternationalization of the problem, the Japanesefinancial crisis was regionally bound to Japan. Asduring the Japanese crisis, banks in the United States

-30

-20

-10

0

10

20

30

40

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

CREDIT CONDITIONS IN JAPAN AND THE US

Source: Bank of Japan; Seniour Loan Officer Opinion Survey.

TANKAN, lending attitude of financial institutionsdiffusion index

large companies

C&I loans

-40

-20

0

20

40

60

80

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

in % of banks who tightened their standards for C&I loans

small companies

small

medium and large

Japan

US

Figure 6

5 Basel I is the round of deliberations bycentral bankers from around the world,and in 1988, the Basel Committee (BCBS)in Basel, Switzerland, published a set ofminimal capital requirements for banks.This is also known as the 1988 BaselAccord and was enforced by law in theGroup of Ten (G-10) countries in 1992,with Japanese banks permitted an extend-ed transition period.

0 2 4 6 8 10 12 14 16 18

Japanese Banking Crisis

Nordic Banking Crisis

1990s

Savings&Loans Crisis

1990s

Current Crisis - Realizedlosses

Current Crisis - Worst

Case Estimation

Source: Oxford Economics (OEF 2008).

LOSSES IN THE BANKING SECTOR

in % of nominal GDP

%

Figure 7


Focus

are currently tightening credit conditions in responseto the decline of their equity ratios.

In Japan, the total amount spent in dealing with theNPL problem grew to 86 trillion yen, roughly 17 per-cent of nominal GDP (Figure 7). Estimates for theultimate scale of aggregate losses in the banking sec-tor caused by the current crisis range from one totwo trillion US dollars. The latter figure would rep-resent some 8 percent of US and European nominalGDP. According to this measurement, the Japanesecrisis was by far the most severe of the listed crisis,with relative losses exceeding the worst case scenar-ios of today’s turmoil by far.

References

DICJ (2006), Annual Report 2006,http://www.dic.go.jp/english/e_annual/e_annual_fy2006.pdf(accessed 28 October 2008).

Hoshi, T. (2002), “The Convoy System for Insolvent Banks: How ItOriginally Worked and Why It Failed in the 1990s”, Japan and theWorld Economy 14, 155–180.

Nakaso, H. (2001), “The Financial Crisis in Japan during the 1990s:How the Bank of Japan Responded and the Lessons Learnt”,BIS Papers 6.

OEF (2008), After Paulson, Where Next for World Growth?,http://www.oef.com/OE_FA_Display_Frm.asp?Pg=IntMacSpec&Txt=International%20Macroeconomics(accessed 28 October 2008).

Reinhart C. and K. S. Rogoff (2008), Is the 2007 U.S. Sub-Prime Fi-nancial Crisis So Different? An International Historical Com-parison, NBER Working Papers 13761.

Tamaki, N. (2008), “Bank Regulation in Japan”, CESifo DICEReport 6(3), 9–13.


Focus

WHAT HAPPENED TO KOREA

TEN YEARS AGO?

CHANG WOON NAM*

What began with the panic triggered by the failure ofthe Thai financial system led to the rapid economicdownturn of some of the globe’s most dynamiceconomies in 1997 and 1998. Especially for SouthKorea the consequences of financial and economiccrisis and the intervention of the IMF in overcomingthe accompanying problems were extremely painful.These problems included, for example,

• a large number of bankruptcies of industrial firmsand private banks,

• the increasing pressure on industrial firms tocarry out rapid restructuring and specialisation ina limited number of competitive areas includingthe so-called “big deal” of exchanging businessactivities among large conglomerates,

• the massive dismissal of workers and the subse-quent insecurity in businesses and society,

• the increase of production costs caused by thewon devaluation and the change of relative pricesof imported goods such as semiconductors andnatural resources like petroleum,

• the drastic decrease in domestic households’demand and firms’ investment spending caused byincome reduction and high interest rates, etc.

In fact, South Korea had a real economy crisis thatwas rapidly triggered by a financial shock.According to the standard economic literature, afinancial crisis is generally defined as a situation inwhich a significant group of financial institutionshave liabilities exceeding the market value of theirassets, leading to runs and other portfolio shifts,collapse of some financial firms and governmentintervention. Additionally, it is also quite oftencharacterised as a non-linear disruption to finan-cial markets in which adverse selection and moralhazard problems become much worse, so that

financial markets are unable to efficiently channelfunds to those who have the most productiveinvestment opportunities (see also Hahm andMishkin 2000; Pettis 2001). According to Shin andHahm (1998), the South Korean case not onlydemonstrates all the aspects of a financial crisisunder the first definition but also fulfils the criteriaposited by the latter definition.1 In addition, thistype of financial crisis was accompanied by the cur-rency crisis (i.e. strong depreciation of the won) inthe fourth quarter of 1997.2 Their negative impactscaused the total economic break-down in SouthKorea. This abrupt crash was also partly triggeredby the pressure of the IMF that insisted on adomestic austerity package and on fundamentalstructural reforms in return for bail-out funds (seealso Weisbrot 2007).3

*Ifo Institute for Economic Research.

1 The facts that support their arguments include: (a) the ratio of com-mercial banks’ non-performing loans to total loans increased rapid-ly in 1997, (b) although a domestic depositors’ run was absent, banksand merchant banking operations in South Korea experienced alarge scale foreign creditors’ run at the end of the same year that ledto the sharp reduction in short-term external debts of these financialinstitutions (by more than 40 percent) within a short period of time,(c) fifteen merchant banking corporations were suspended inDecember 1997 and thirteen of them were closed later, while thegovernment intervened in two insolvent banks at the end of the sameyear, and (d) the pattern of rising interest rates and declining stockprices in the midst of an increase in number of incidents of default(the so-called non-linear disruptions in domestic financial markets)had already been in place early in September 1997, which, however,became more evident two months later. For example, the KOSPIstock price index only amounted to 390.3 in December compared to494.1 in November (Shin and Hahm 1998).2 On average, the won was devalued from ca. 844 won/US$ in 1996to 1,850 won/US$ in 1997.3 Policy errors made by the South Korean and other Asian govern-ments and the IMF during the first stage of panic appear to havecontributed to the deepening of economic distress (Feldstein 1998).“The IMF roared into Asia and promised to supply $17 billion toBangkok, $40 billion to Jakarta, and $57 billion to Seoul. [Inreturn], it demanded austerity budgets, high interest rates, and salesof local businesses to foreign bargain-hunters. It claimed that thesemeasures would restore economic health to the “Asian tigers” ...”(Johnson 1998, 659). According to the IMF, “high [real] interestrates would encourage domestic capital to stay at home and foreignlenders to resume lending, which would boost the currency. Thecurrency boost would both make it easier for domestic firms torepay their foreign debts and check the danger of competitive[devaluation]. ... In practice, [however] the increase in real interestrate combined with [tax increases, cuts in government expenditureand other restrictive measures which also led to compression of pri-vate consumption] only depressed firms’ cash flow and raised theirfixed-payment obligations, tipping more and more into insolvency,accelerating the [capital] outflows and reducing inflows. … TheIMF [appeared to forget] that private cash flows are cyclical ...When a whole economy is [in trouble], foreign capital will notreturn whatever the interest rate. ... Requiring a sharp rise in bankcapital adequacy standards in the midst of the crisis caused a cut incredit, a rise in non-performing loans, and further bankruptcies. ...The IMF also required [Asian] governments to guarantee the for-eign debts of local firms and banks. Protected from default, foreigncreditors hung back on rescheduling or rolling over the debt. Thisworsened the hard currency squeeze on local debtors, pushingthem to buy foreign exchange to cover their increased dollar needsand adding to the exchange rate collapse” (Wade 1998, 700–701).

Following the big shocks of the crisis in SouthKorea and other Asian countries, the followingstandard questions were raised at the end of 1990s:After the lessons learned from the previous LatinAmerican economic crisis, how could this happenagain and especially in East Asia? How could suchrapidly growing economies experience a suddendepression, one after the other? Why had the crisisnot been foreseen? And, of course, what could bedone to prevent such crises in the future? Historyrepeats itself. Today Americans and Europeans areasking themselves similar questions. How couldsuch a large scale financial crisis happen to theworld’s most advanced economies just ten yearsafter the Asian crisis? This article revisits theKorean economic crisis at the end of 1990s andidentifies some specific financial and real sectorproblems to provide a better understanding of theorigins of the crisis and its economic consequencesfor the victims.4

Principal causes of the Korean economic crisis

According to Bergsten (1997) and Krugman(1998), the Asian crisis was initiated in the 1980s.The rapid appreciation of the Japanese yenagainst the US dollar in 1985 led to a relativeincrease in production costs in Japan, which trig-gered the relocation of production facilities ofJapanese firms to South Korea and Taiwan. Inorder to counteract its negative impact on domes-tic economic development, Japanese authoritiescarried out an expansionary monetary policy,which, in turn, increased asset prices in Japan andstimulated, due to the low domestic interest rates,capital outflow to South Korea and Taiwan. At theend of the 1980s, these Asian economies wereforced to appreciate their currencies and toaggressively expand the money supply. As hadbeen the case in Japan, the implementation ofthese political measures in Korea and Taiwan cre-ated asset price bubbles at home and large capitalflight to Indonesia and Thailand. On the otherhand, the sharp devaluation of the Chinese yuan

(by about 40 percent in 1994) and the Japaneseyen (by over 25 percent in the period of 1995–96)put tremendous competitive pressure on the otherAsian countries, eroding their trade positions andproducing large trade deficits in South Korea andthroughout East Asia.

Furthermore, it is also quite often argued that therapid creation of bubbles in expanding productioncapacities and investment activities of (large) indus-trial firms (not only in the domestic country but alsoabroad) were, to a larger extent, financially (in manycases also politically) backed by the government (theso-called moral hazard problem).5 In this context, itwas suggested that since the mid-1990s SouthKorean industrial firms have faced strong challengesin the Asian and world market, particularly fromChinese, Thai and Indonesian competitors.Furthermore, labour costs also drastically increased,which again made Korean products more expensivein the world market. In order to maintain their com-petitive position, South Korean industrial firms wereforced to quickly carry out structural changes includ-ing large-scale investments in modern capital stockand to intensify their R&D activities (Bergsten 1997;Kwon 2008). As a consequence, South Koreanimports of modern investment goods have increasedrapidly since the mid-1990s. In addition, the Koreanchaebols also attempted a relocation of their pro-duction facilities to foreign countries.Although largeSouth Korean industrial firms lacked their ownfinancial means and suffered from the serious asset-liability mismatches,6 as Krugman (1998) argued,they could well finance their (in some cases not ade-quately examined, risky and/or less profitable) long-term modernisation and internationalisation pro-jects – thanks to government support – by takingshort-term loans from domestic banks which playedsolely the role of a “cash box” in the country.7

To be sure, corporate sectors with high levels of debtare generally vulnerable to shocks that cause a fall incash flow or an increase in fixed payment obligations


Focus

4 This article is largely based on Nam (2005).

5 This view is also reflected in the IMF programme for Korea, whichdemanded that the government should take measures to weakenthe large, diversified, family-owned chaebols. The proposed mea-sures included the banning of mutual payment guarantees amongthe member firms of the same chaebol, the demand for the publi-cation of consolidated balance sheets for the whole chaebol (ratherthan for individual firms), and the strengthening of minority share-holder rights through stricter disclosure requirements (Chang et al.1998). The implemented big-deal business activity exchange pro-gramme among large conglomerates mentioned above was the con-sequence of these measures.6 According to the Bank of Korea, the debt-equity ratio of Koreanmanufacturing corporations amounted to approximately 340 per-cent between 1973 and 1996.7 Against this type of argument, Chang et al. (1998) suggested: (a)the low corporate profitability in Korea is mainly due to high inter-est payments rather than to inefficiency, (b) the investment that thechaebols had made in the building-up to the crisis were mainly inindustries with stable returns (like those leading export activitieslike petrochemicals, iron and steel, cars, electronics and shipbuild-ing), rather than high-risk high-return industries, which thoseinvestors operating under moral hazard will be inclined to choose,and (c) there has been no instance in last two decades whereKorean government bailed out a failing chaebol. In the period of1990–96 three (Hanyang, Yoowon and Woosung) of 30 conglomer-ates went bankrupt in addition to the six others (Kia, Hanbo,Sammi, Haitai, Jinro and Halla) in 1997.


Focus

– systemic shocks such as a fall in aggregate demand,a rise in interest rates, or a devaluation of the cur-rency when part of the debt is foreign (Wade 1998).In this context, it should also be noted that in SouthKorea there was asymmetry in the deregulationschedules regarding the foreign currency-denomi-nated lending and borrowing by financial institutionsin the country. In spite of the fact that in 1993 theKorean government expanded the positive list ofusage for which financial institutions may provideforeign currency-denominated loans in the scope ofthe financial liberalisation programme, the quantita-tive restriction on long-term borrowing was main-tained, while short-term borrowing (by banks) wasfreely allowed (see also Amess and Demetriades2001; Bratsiotis and Robinson 2005). As a conse-quence, the short-term foreign debts of the SouthKorean banks grew rapidly in the first half of the1990s to meet the increasing investment demand ofthe corporate sector. For these reasons, the creditincrease largely exceeded the growth of availabledomestic resources in the financial sector, and theextra credit supply of banks was provided by thesteady and strong expansion of external debts andforeign currency-denominated assets (Shin andHahm 1998; Chang et al. 1998).8

“Such practices may be sustainable as long as therate of economic growth remains high, as financialinstitutions compensate for the low effective rateof return on loans to these preferred borrowerswith high rates of return on other loans. However,if the rate of growth slows, the bad loan drag beginsto inhibit the ability of banks and non-bank lend-ing institutions to supply credit to the economy.This is what transpired in a number of Asianeconomies [in 1997], as some key export pricesdeclined (computer chips, for example)” (Noland1998, 2). Furthermore such reductions in theexport revenue of Asian countries rapidly createdthe expectation of declining corporate profits andequity prices. As a consequence, domestic and for-eign investors started to transfer money abroad insearch of higher returns. This, in turn, had a signif-icant effect on the collapse of the domestic assetmarket and the exchange rate depreciation inAsian countries.

In other words, apart from the large deficits in thebalance of payments and the insufficient foreign cur-rency reserves, the weakness of the real sectorreflected by the decline in economic and exportgrowth, additionally makes South Korea and otherAsian victims more vulnerable to speculativeattacks, panics and external financial shocks leadingto sudden crises, as is also shown by the monetaryand financial crises in Europe and Latin America in1992–93 and 1994–95, respectively (see alsoCartapanis et al. 2002).

In the beginning of the 1990s, South Korea experi-enced a sharp decline of the real GDP growth ratefrom 9.1 percent (1991) to 5.1 percent (1992).However, this did not cause a crisis, in contrast tothe more recent economic slowdown from 8.9 per-cent in 1995 to 7.1 percent in 1996. Even in 1997 thecountry was able to achieve a GDP growth rate of5.5 percent, although its economy had already start-ed to collapse that autumn. Under the additionalconsideration of the quarterly business climateindex for the South Korean manufacturing sector,provided by the Bank of Korea, one can easily seethat the rapid fall of the real GDP growth rate inthe period of 1991–92 was accompanied by adecline in this leading index.9 In spite of theincrease in the real GDP growth rate from 8.6 per-cent to 8.9 percent in the period 1994 to 1995, how-ever, the business climate index had graduallydecreased since the 3rd quarter of 1994 from avalue of ca. 120 falling sharply to below 40 in the 4thquarter of 1997. This fact clearly shows that SouthKorean manufacturing firms had actually anticipat-ed the decline in their economic activity since 1994,although the GDP growth rate was very high and itsreduction was rather modest in the mid-1990s.Following the idea of Noland (1998), one can assertthat in mid-1994 South Korean and foreigninvestors started to move money abroad in searchof higher returns, and this process continued until1997. This fact is also reflected in the developmentof the South Korean capital account deficits in theperiod between 1993 and 1996. Measured in termsof the 1990 won/US$ exchange rate, the deficitamounted to approximately 536 million US dollarsin 1993, which was reduced by 55 million US dollarsin 1994. In both subsequent years, however, thedeficits increased rapidly reaching 527 and 704 mil-lion US dollars in 1995 and 1996, respectively.

8 For example, in South Korea foreign currency-denominated assetsof banks relative to nominal GDP went up to approximately29 percent in 1996 from ca. 20 percent in 1992, while the ratio oftotal (short- and long-term) external debts to GNP rose to around22 percent from 14 percent in the same period of time and reachedaround 120 billion US dollars in September 1997 (Shin and Hahm1998; Chang et al., 1998)). The total sum of short-term foreign debtreached approximately 95 billion US dollars in the mid-December1997.

9 The quarterly business climate index is calculated on the basis ofbusiness surveys conducted among South Korean manufacturingfirms. Its value of the first quarter of 1991 is set 100.

More immediately, the sudden widening of the tradebalance gap in 1996 made South Korea less immuneto the financial crisis (Krugman 1998; Shin andHahm 1998; Wade 1998). Expressed in 1990exchange rates, the trade (goods and services) deficitincreased from 8 to 25 billion US dollars between1995 and 1996.10 Yet, it should be emphasised thatthis large gap was not created by the reduction intotal exports but by the country’s import growthbeing much faster than its export growth in the sameyear. Since computer chips were the leading exportsof South Korea in the 1990s, the decline in semicon-ductor exports in 1996 created the expectation ofdeclining corporate profits and equity prices, asNoland (1998) and Wade (1998) also suggested.11

Partly thanks to the gradual export increase of itemslike cars, woven fabrics and ships, South Korea wasable to achieve an export growth of 13 percent in1996, although that rate was clearly lower than the28 percent of 1995. As mentioned above, importgrowth was more dynamic in these years: its ratereached 28 and 23 percent in 1995 and 1996, respec-tively. In particular, imports of petroleum andsophisticated semiconductors showed an above-average growth trend: the percentage share ofimports of the former component reached around10 percent of South Korea’s total imports in 1996,while that of the latter amounted to ca. 8 percent. Inother words, the dependence of South Korean man-ufacturing activities on foreign-made semiconduc-tors and petroleum imports had dramaticallyincreased within two years. In combination with theeffects of the decrease in semiconductor exports oncapital flight, South Korean import performancemade the country’s economy more vulnerable to thefinancial crisis. Especially imports of items like semi-conductors and petroleum, which were required forthe manufacturing activities in South Korea, becameobviously more expensive when the won was strong-ly devalued at the end of 1997. This was also one ofthe most decisive reasons why the domestic produc-tion system started to collapse.

Conclusions

The Asian crisis, which began with the panic trig-gered by the failure of the Thai financial system,

rapidly turned out to be a real sector crisis ratherthan a financial and currency crisis. The abruptbreakdown was particularly evident and serious forthe South Korean economy. The list of major causesfor the crisis in South Korea is long.

The financial factors that led to the crisis generallyinclude: the serious asset-liability mismatches ofdomestic firms, the over-reliance on foreign debtspartly caused by the government’s badly designedfinancial liberalisation schedule, the poorly-devel-oped and little transparent domestic financial sys-tem, the insufficient foreign currency reserves, themoral hazard resulting from the close relationsbetween governments and large firms as well as fromthe so-called cross-firm debt guarantee within a con-glomerate.

In addition, some macro- and meso-economic condi-tions and weaknesses have not only made theKorean economy vulnerable to external shock butalso contributed to the sudden capital flight byinvestors and the subsequent economic collapse.Major real sector problems were, for example, the(expectation of a) decline in the economic growthrate from 1995 and the slow growth of export rev-enues in 1996 (due mainly to the world-widedecrease in price of the computer chips – Korea’smost important export item in the 1990s), the sharpincrease in the trade deficit in 1995–96 caused byimport growth of petroleum and sophisticated semi-conductors, the intensive modernisation efforts ofKorean firms, and competitive pressure caused bythe increase in labour costs and recent devaluationof yuan and yen, to name a few.

Furthermore, policy errors made by Korean govern-ments and the IMF during the early stages of panicappear to have deepened the economic distress. Forexample, the increase in real interest rates combinedwith tax increases, cuts in government spending, etc.which were demanded by the IMF in return for bail-out funds depressed Korean firms’ cash flow andraised their fixed-payment obligations, forcing themmore strongly into insolvency.

The unstable development of important Asian cur-rencies, including the Korean won, the transmissionof bubbles from one country to another and therelated less-disciplined monetary policy practice ofindividual Asian countries since the mid-1980s werethe initiators of the Asian crisis. In this context it wasargued that in Asia a more intensive co-ordination in


Focus

10 Measured in terms of 1990 exchange rates, Korea’s total exportsincreased from ca. 135 to 153 billion US dollars between 1995 and1996, while its total imports experienced a stronger growth from ca.143 to 178 billion US dollars in the same period of time.11 As a consequence, the share of semiconductor export revenuedecreased from 16 to 13 percent of total South Korean exportsbetween 1995 and 1996.


Focus

the fields of monetary and exchange rate policy-making, debt negotiations and expansions would benecessary to prevent and to overcome the crisis(Eichengreen and Bayoumi 1996). Although someeconomists and politicians supported the idea ofestablishing a type of Asian monetary union (with acommon currency and central bank), its realisationhas remained fruitless due to significant economicdisparities, structural differences and different finan-cial systems in Asian countries (Wade, 1998).Moreover, Asia lacks the kind of political solidarityand institutional processes that have played animportant role in the economic integration of theEU since the 1950s.

References

Amess, K. and P. Demetriades (2001), Financial Liberalisation andthe South Korean Financial Crisis: Some Qualitative Evidence,Discussion Papers in Economics 01/3, University of Leicester.

Bergsten, C.F. (1997), The Asian Monetary Crisis: ProposedRemedies, Statement before the Committee on Banking andFinancial Services, US House of Representatives, November 13,Washington DC.

Bratsiotis, G. and W: Robinson (2005), “Currency Composition ofDebt, Risk Premia and the 1997 Korean Crisis”, EconomicModelling 22, 459–471.

Cartapanis, A., V. Dropsy and S. Mametz (2002), “The AsianCurrency Crises: Vulnerability, Contagion, or Unsustainability”,Review of International Economics 10, 79–91.

Chang, H. J., H. J. Park and C. G. Yoo (1998), “Interpreting theKorean Crisis: Financial Liberalisation, Industrial Policy andCorporate Governance”, Cambridge Journal of Economics 22,735–746.

Eichengreen, B. and T. Bayoumi (1996), Is Asia an OptimumCurrency Area? Can It Become One? Regional, Global andHistorical Perspectives on Asian Monetary Relations, Center forInternational and Development Economic Research WorkingPaper C96/081, University of California Berkeley.

Feldstein, M. (1998), “Refocusing the IMF”, Foreign Affairs 77/2,20–38.

Hahm, J. H. and F. S. Mishkin (2000), “The Korean Financial Crisis:An Asymmetric Information Perspective”, Emerging MarketsReview 1, 21–52.

Johnson, C. (1998), “Economic Crisis in East Asia: The Clash ofCapitalisms”, Cambridge Journal of Economics 22, 653–661.

Krugman, P. (1998), What Happened to Asia?, January 1988,http://web.mit.edu/krugman/www/disinter.html.

Kwon, E. (2008), The Korean Financial Crisis a Decade Later:Restructuring the Chaebols,http://www.allacademic.com/meta/p250582_index.utml.

Nam, C. W. (2005), Major Causes of the Korean and AsianEconomic Crises, in: Gangopadhyay, P. and M. Chatterji (eds.),Economic Globalization in Asia, Aldershot: Ashgate PublishingLimited, 89–107.

Noland, M. (1998), The Financial Crisis in Asia, Statement beforethe House International Relations Committee Subcommittees onAsia and Pacific Affairs, and International Economic Policy andTrade, February 3, Washington DC.

Pettis, M. (2001), The Volatility Machine: Emerging Economies andthe Threat of Financial Collapse, Oxford: Oxford University Press.

Shin, I. and J. H. Hahm (1998), The Korean Crisis – Causes andResolution, Paper presented at the East-West Center/Korea

Development Institute Conference on the Korean Crisis, August 8,Honolulu.

Wade, R. (1998), “From “Miracle” to “Cronysm”: Explaining theGreat Asian Slump”, Cambridge Journal of Economics 22, 693–706.

Weisbrot, M (2007), Ten Years After:The Lasting Impact of the AsianFinancial Crisis, Center for Economic and Policy Research,www.cepr.net/documents/publications/asia_crisis_2007_08.pdf.


Spotlight

HOW TO DEFINE A

RECESSION?

KLAUS ABBERGER AND

WOLFGANG NIERHAUS*

The famous “R-word” is back again. Not only econ-omists are discussing whether the world economy,regions or specific countries are in or close to reces-sion. Also the public media and ordinary people areconcerned about a recession. But the question mustbe asked: what actually constitutes a recession? Thisis ultimately a matter of definition, which unfortu-nately is not answered homogeneously in the schol-arly literature and also not in the practice of interna-tional organisations or statistical offices.

Business-cycle analysts generally use the concept ofa recession to refer to weak economic phases, ofwhich duration, depth and diffusion exceed the usualbounds; thus one speaks of the “three key dimen-sions of a recession, known as the ‘three Ds’: dura-tion, depth and diffusion” (Fiedler 1990, 131).1 Thismeans that in a recession economic activity decreas-es substantially, the decline affects wide portions ofthe economy and it has some permanence. In theUnited States the Business Cycle Dating Committeeat the National Bureau of Economic Research(NBER) is responsible for the dating of the US busi-ness cycle.2 The NBER describes the concept as fol-lows: “a recession is a significant decline in econom-ic activity spread across the economy, lasting morethan a few months, normally visible in real GDP, realincome, employment, industrial production, andwholesale-retail sales”. However, even this NBERspecification shows that the term is not narrowlydefined and contains some leeway for determiningwhether an economy is in recession or not. Since ajudgment based on these criteria requires the analy-sis of extensive data; the NBER publishes its official

recession dating for the US economy with consider-able lags of often more than one year.3

For this reason another definition has become popu-lar among business-cycle observers in the UnitedStates, and increasingly also in other countries, whichwas first published by Julius Shiskin in The New York

Times in 1974.4 According to Shiskin a recession isdefined as “a decline in the seasonally and calendaradjusted real gross domestic product (GDP) in atleast two successive quarters”. Since GDP is themost comprehensive indicator of economic activityand the critical period for a recession amounts to atleast six months, with this rule of thumb the two cri-teria of diffusion and duration are roughly taken intoaccount. What is the NBER’s position on this defini-tion? It adds: “most of the recessions identified byour procedures do consist of two or more quarters ofdeclining real GDP, but not all of them. Our proce-dure differs from the two-quarter rule in a numberof ways”.5

The main disadvantage of the Shiskin rule is that therate of change of the seasonally and calendar adjust-ed GDP go through erratic fluctuations due to sub-sequent data revisions, whereby a minus can easilybecome a plus. For the recession of the US economyfrom March until November 2001 officially regis-tered by the NBER dating,6 the results of the Bureauof Economic Analysis up to July 2002, for example,showed a minus in real GDP only for the third quar-ter, and subsequently for the first three quarters of2001 in succession. According to present calculations(GDP: first quarter 2008 final) the pluses and minus-es in 2001 alternate from quarter to quarter so thatthe rule of thumb does not signal a recession for theUnited States in 2001 (see Figure 1).

Applying this criterion to Germany, there would bemore recessions than have been normally counted.

* Ifo Institute for Economic Research.1 Fiedler, E. R. (1990), The Future Lies Ahead, in: Klein, P. A. (ed.),Analyzing Modern Business Cycles, Essays Honoring Geoffrey H.Moore, Armonk, NY: M. E. Sharpe.2 http://www.nber.org/cycles/recessions.html.

3 For Germany the Bundesbank – the only official German institu-tion to do so – has published, in its supplement, “SaisonbereinigteWirtschaftszahlen”, a chronology of economic activity up to 1997measured in terms of the local minimums and maximums of thetrend-adjusted industrial production in Germany.4 Achuthan, L. and Banerji, A. (2008), The Risk of RedefiningRecession, CNNMoney.com, May 6.5 http://www.nber.org/cycles/recessions.html.6 http://www.nber.org/cycles/july2003.html.


Spotlight

There is general agreement that recessions occurredin 1974/75, 1980/82 and 1992/93. The rule of thumb,however, has identified more recessions than these,which is ultimately due to the fact that in Germanythe trend growth rate of total economic output islower than in the United States. In the currentdecade alone, GDP, seasonally and calendar adjustedusing the BV4.1 and based on the current data status(National Accounts, 2nd quarter 2008), has fallentwice in at least two successive quarters: from thefourth quarter of 2002 to the first quarter of 2003and in the second and third quarter of 2004. Sincereal GDP declined only slightly, these were notrecessions but only pronounced phases of economicweakness.

Accordingly, for there to be a recession real GDP –using the rule of thumb – has to have clearly fallenfor at least two quarters. In order to adequatelydetermine the position of the economy in the busi-ness cycle, also the utilisation ofproduction capacities shouldalways be taken into considera-tion. Business cycles, accordingto modern business-cycle theory,are to be understood as varia-tions in the utilisation degree ofaggregate production potential(growth cycles). Every cycleconsists of one upswing and onedownturn phase, with the indi-vidual phases being linked byupper and lower turning points.In this delimitation downturnphases include both phases withabsolutely falling production

activity as well as phases withincreasing production activitythat is lower than average(decreasing capacity utilisation)– measured in terms of thepotential rate. When, after astrong upswing stage with highercapacity utilisation than theaverage, real GDP, seasonallyand calendar adjusted, falls intwo successive quarters but theutilisation rate remains abovethe normal level, one should notspeak of a recession. A reces-sion, according to our definition,is only the case when with clear-ly slowing output, seasonally andcalendar adjusted (e.g. in at least

two successive quarters), the aggregate utilisationrate is also simultaneously clearly below its long-term average.7 Using the survey-based Ifo capacityutilisation in manufacturing as proxy for the aggre-gate utilisation rate, this broader recession criterionwas fulfilled in the above-listed recessions in Ger-many: 1974/75, 1980/82 and 1992/93 (see Figure 2).8

For the world economy as a whole, however, the def-inition of recession that we have discussed cannot bereasonably applied because the combined real GDPof the world has increased continuously in the last

94

96

98

100

102

104

I II III IV I II III IV I II III IV

-4

0

4

8

12

16

REAL GROSS DOMESTIC PRODUCT IN THE US, 2000–2002seasonally adjusted

Growth rate in %

percent change from previous period, annualized

Source: BEA; NBER; Ifo calculations.

Index, first quarter 2001 = 100

2000

Annual figures

percent change over previous year

Shaded area represents the most recent dated US recession (from March to November 2001)

3.7%

1.6%

2001 2002

0.8%

Figure 1

72.5

75.0

77.5

80.0

82.5

85.0

87.5

90.0

92.5

95.0

1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008

Shaded areas represent periods of recession

Seasonally adjusted by ASAII

Smoothed by weighted local polynomial regression (a)

Source: Ifo Business Survey.

in %

(a) Abberger, K. and W,. Nierhaus (2008), Die ifo Kapazitätsauslastung - Ein koinzidenter Indikator der

deutschen Industriekonjunktur, ifo Schnelldienst 61(16), 15-23.

IFO CAPACITY UTILISATION IN MANUFACTURING AND RECESSIONS

IN GERMANY

Average 1978–2008

Figure 2

7 For the situation of the world economy and the German economyin autumn 2001, see ifo Schnelldienst 20/2001, 18.8 Results from Ifo business tendency surveys in manufacturing.Respondents are asked to assess the capacity utilization rate oftheir machines and equipment on a categorical scale. For a detailedanalysis of the time series and its behaviour, see Abberger, K. andW. Nierhaus (2008), Die ifo Kapazitätsauslastung – ein gleichlau-fender Indikator der deutschen Industriekonjunktur, ifo Schnell-dienst 61(16), 15–23.


Spotlight

four decades, on average, which is due to the highertrend growth in the emerging markets. Hence, thereare no generally accepted rules for a definition of aworld recession. According to the InternationalMonetary Fund (IMF), there have been three com-prehensive growth recessions since 1970–1975, 1982,and 1991 – in which global economic growth re-mained below the two percent level. Adjusted forthe increase in the world population, the rate ofchange of real GDP was even negative in 1982 and1991.All in all, the IMF in identifying recessions nowfocuses – similar to NBER – on a number of month-ly indicators, for example, world industrial produc-tion or world trade volume.9 On the other hand, ref-erence is also made to a definitive annual worldgrowth threshold (+ 3 percent), below which theIMF sees a global recession.10 Using this new IMFrule, in the past three decades the years 1980 to 1983,1990 to 1993, 1998 as well as 2001 and 2002 were ator below this threshold. For 2009 the IMF predicts+ 2,2 percent growth. So it considers the world econ-omy in a global recession (see Figure 3).

0

1

2

3

4

5

6

7

8

1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009

Shaded areas represent periods of global recession

World output, percent change over previous year

Source: IMF.

WORLD OUTPUT AND GLOBAL RECESSIONSin %

IMF annual world growth threshold for global recession

2008 and 2009: Forecasts

Figure 3

9 International Monetary Fund (2002), World Economic Outlook,April, 10–12.10 International Monetary Fund (2008), World Economic Outlook,October, 43.


Trends

FINANCIAL CONDITIONS

IN THE EURO AREA

The annual rate of growth of M3 decreased to 8.6% in September 2008,from 8.8% in August. The three-month average of the annual growth rateof M3 over the period from July to September 2008 declined to 8.9%, from9.2% in the period June to August 2008.

In August 2008 the monetary conditions index continued its generaldecline that had started in late 2001, signalling greater monetary tight-ening. This is the result of rising real short-term interest rates and a ris-ing real effective exchange rate of the euro.

In the three-month period from August to October 2008 short-terminterest rates rose. The three-month EURIBOR rate increased from anaverage 4.97% in August to 5.11% in October. Yet, the ten-year bondyields decreased from 4.50% in August to 4.42% in October. In the sameperiod of time the yield spread also fell from – 0.47% (August) to– 0.69% (September).

The German stock index DAX declined in October 2008, averaging4,988 points compared to 6,422 points in August and 5,831 points inSeptember. The Euro STOXX also fell from 3,346 in August to2,672 in October. The Dow Jones International also declined, averag-ing 9,177 points in October compared to 11,531 points in August and11,114 points in September.

According to the second Eurostat estimates, euro area (EU15) GDPdeclined by 0.2% and EU27 GDP remained unchanged in the secondquarter of 2008, compared to the previous quarter. In the first quarter of2008 the growth rate had amounted to + 0.7% for the euro area and+ 0.6% for the EU27. Compared to the second quarter of 2007, i.e. yearover year, seasonally adjusted GDP rose by 1.4% in the euro area and by1.7% in the EU27.

In October 2008, the EU Economic Sentiment Indicator (ESI) fell by 7.4 points in the EU27 and decreased by 7.1 points in the euro area, to77.5 and 80.4 respectively. In both regions, the ESI recorded its largestmonth-on-month decline ever to hit its lowest level since 1993. All coun-tries reported a fall in their ESI. Confidence deteriorated most marked-ly in the Netherlands (– 11.3), followed by France (– 6.5), Italy (– 6.1),the UK (– 5.7), Poland (– 5.1), Germany (– 4.8) and Spain (– 3.5).

* The industrial confidence indicator is an average of responses (balances) to thequestions on production expectations, order-books and stocks (the latter with in-verted sign).** New consumer confidence indicators, calculated as an arithmetic average of thefollowing questions: financial and general economic situation (over the next12 months), unemployment expectations (over the next 12 months) and savings(over the next 12 months). Seasonally adjusted data.

In October 2008, the industrial confidence indicator fell in both the EU27and the euro area. The decline was more marked in the EU27 than in theeuro area. The indicator has been on a downward path since its peak inMay 2007 and currently stands below its long-term average in both areas.In parallel the consumer confidence indicator also saw a sharp downwardcorrection in October and currently stands below its long-term average inboth areas.

Managers’ assessment of order books deteriorated from – 13.5 in Augustto – 26.8 in October 2008. In July the indicator had reached – 12.5.Capacity utilisation declined to 80.9 in the fourth quarter of 2008 from82.8 in the previous quarter.

EU SURVEY RESULTS


Trends


Trends

The exchange rate of the euro against the US dollar averaged 1.33 $/€ inOctober 2008, a decrease from 1.44 $/€ in September. (In August the ratehad amounted to 1.50 $/€.)

The Ifo indicator of the economic climate in the euro area (EU15) hasworsened again in the fourth quarter of 2008 for the fifth time in succes-sion, falling to its lowest level in five years. Its decline is the sole result ofless positive assessments of the current economic situation as well asclearly more pessimistic expectations for the coming six months.

Euro area (EU15) unemployment (seasonally adjusted) amounted to7.5% in September 2008, unchanged from August. It was 7.3% in Sep-tember 2007. EU27 unemployment stood at 7.0% in September 2008,compared to 6.9% in August. The rate was also 7.0% in September 2007.Among the EU Member States the lowest rates were registered in theNetherlands (2.5%) and Denmark (2.9%). Unemployment rates werehighest in Spain (11.9%) and Slovakia (10.0%).

Euro area annual inflation (HICP) was 3.2% in October 2008, comparedto 3.6% in September. This is quite an increase from a year earlier, whenthe rate had been 2.6%. The EU27 annual inflation rate reached 4.2% inSeptember, down from 4.3% in August. A year earlier the rate hadamounted to 2.2%. An EU-wide HICP comparison shows that in Sep-tember 2008 the lowest annual rates were observed in the Netherlands(2.8%), Germany (3.0%), Ireland and Portugal (both 3.2%), and thehighest rates in Latvia (14.7%), Bulgaria (11.4%) and Lithuania(11.3%). Year-on-year EU15 core inflation (excluding energy andunprocessed foods) fell to 2.5% in September from 2.6% in August.

EURO AREA INDICATORS

Online information services of the CESifo Group, Munich

The Ifo Newsletter is a free service of the Ifo Institute and is sent by e-mail every month. It in-

forms you (in German) about new research results, important publications, selected events, per-

sonal news, upcoming dates and many more items from the Ifo Institute.

If you wish to subscribe to the Ifo Newsletter, please e-mail us at: [email protected].

CESifo publishes about 20 working papers monthly with research results of its worldwide aca-

demic network. The CESifo Newsletter presents selected working papers (in English) in an easily

understandable style with the goal of making its research output accessible to a broader public.

If you wish to subscribe to the CESifo Newsletter, please e-mail us at: [email protected].

If you wish to receive our current press releases, please e-mail us at: [email protected].

You can also request these services by fax:

Ifo Institute for Economic Research, fax: (089) 9224-1267

Please include me in your mailing list for:

� Ifo Newsletter � CESifo Newsletter � Ifo Press Releases

Name: ………..........................................................................................................

Institution: .....................................................................................................................

Street: .....................................................................................................................

City: .....................................................................................................................

Telephone: ...................................................................................................................

Fax: .....................................................................................................................

E-mail: …………………..............................................................................................

THE GREAT WELSH LEEK BUBBLE OF 1876

RAY REES*

The leek is not only a splendid vegetable, but also the national emblem of my home country, Wales. TheEnglish have the rose, the Scots the thistle, and the Irish the shamrock, all worn on the lapels of theircoats on their respective national days. But the great advantage of the leek over these other rather frag-ile emblems is that it not only has much more of an impact, both on the eye and on the nose, when it isworn on your lapel, but it can also be eaten afterward. Nothing is wasted.

You probably have not heard of the Great Welsh Leek Bubble of 1876, though it deserves to stand upthere with the Dutch Tulip Mania, the English South Sea Bubble, the French Mississippi Bubble, andmore recent events in the United States, from Dotcom to Subprime, as a great case study in human fol-ly and greed. It began in the little town of Llareggub. A local gardener named John P Morgan proudlyannounced one day that he had cultivated a wonderful new variety of leek, with a stem of a purer white,the leaves a deeper green, and a flavour more subtle but powerful, than any other leek in existence.Moreover it could be worn for 3 days without losing its flavour. Demand for the new leek, christened the“Prince of Wales”, soon outstripped supply, the price rocketed to unheard of heights for such an ordi-nary vegetable, and soon John Morgan, who was now a rich man, had sold all of his stock, both of leeksand of seed. This caused something of a panic, because St David’s Day, the national day of Wales, wasonly a couple of months away. Everyone had to have a leek to wear on that day and people did not wantto be seen wearing anything but a Prince of Wales leek. Then a local insurance agent named Arthur IGee, or just A I Gee for short, hit upon the clever idea of selling Prince of Wales leeks forward. For animmediate payment, he would issue a contract guaranteeing the delivery of a bundle of 10 Prince ofWales leeks in the week before St David’s Day. He of course knew very little about growing leeks, butassumed that all the seed that John Morgan had sold would be producing a bumper harvest by the timeSt David’s Day came around. At the same time he contracted with all the growers who had bought JohnMorgan’s seed, giving him an exclusive right to buy their crops at the currently prevailing price, thus cor-nering the market.

The Forward Leek Contracts (FLCs) issued by Arthur Gee started to be traded and indeed sold likehot cakes, soon fetching huge prices. It became clear that a classic bubble was under way. People from asfar away as London were buying FLCs not because they wanted to wear leeks on St David’s Day, but be-cause they expected the price to go on rising, so they could resell their FLCs at a profit. So desperatewere the people of Llareggub to profit from the price spiral that they sold their cows, pigs and sheep (toArthur Gee) in order to put the money into FLCs. For a fee of 5 percent, he would also accept, as col-lateral for loans to buy more FLCs, already existing holdings of FLCs, which he immediately resold onthe market. The price of an FLC soon reached the same level as the price of a small cottage or largepigshed.

In order to stabilize the market, which was becoming a little nervous, Arthur Gee offered insuranceagainst the loss of value of FLC holdings, at a premium equal to 10 percent of their insured value. Manypeople bought this and felt reassured. When a local journalist asked Arthur whether he had the reservesto meet potential insurance claims, he carefully explained the sophisticated system of Credit DefaultSwaps (CDSs) that he had arranged with some big bankers in London. Since he explained all this to thereporter as he was cutting the impressive hedge of beech trees in front of his house, the reporter decid-ed to call them “Hedge Funds”. He also explained that the rating agency he had hired at large expenseto check his solvency had given him a triple A+ rating.

Then an ugly rumour started to circulate. The journalist had calculated that the total number of Princeof Wales leeks required to cover all the FLCs in circulation was about ten times as large as the numberthat could possibly be grown, in the time available, from the seed that John Morgan had sold. The onlyway that it would be possible to meet the contracts to supply all the bundles of leeks would be to mix in-ferior leeks in with the Prince of Wales leeks in a ratio of 9:1, thus creating what was known in the tradeas a Toxic Leek Asset (TLA). This precipitated what was subsequently called the Sub-Prince Crisis. Theprice of FLCs plummeted, and did not stop until it reached the level of the cost of 9 ordinary leeks and1 Prince of Wales leek. Of course, all the holders of FLCs went bankrupt. Arthur Gee disappeared. JohnP Morgan, a much saddened man, gave up growing leeks, emigrated to America and opened a merchantbank. But his leek lives on, and is still worn, and eaten, to this very day.

* Ludwig Maximilian University of Munich.

a joint initiative of ludwig-maximilians-universität and ... · tim to the crisis, unscrupulously...

Documents