A special report on international banking May 16th 2009

Rebuilding the banks

BANKING.indd 1BANKING.indd 1 5/5/09 14:35:065/5/09 14:35:06

The Economist May 16th 2009 A special report on international banking 1

A tamer banking industry is already emerging from the debris of theold, failed one, says Andrew Palmer

hearing about your annual bonus. Now itmeans getting �red. America’s �nancial­services �rms have shed almost half a mil­lion jobs since the peak in December 2006,more than half of them in the past sevenmonths. Many have gone for good.

The pain is nowhere near over. Thecredit crunch has been a series of multiplecrises, starting with subprime mortgagesin America and progressively sweepingthrough asset classes and geographies.There are now some glimmers of opti­mism in the investment­banking world,where trading books have already beenmarked down ferociously and credit expo­sures to the real economy are more limited.But most banks are hunkering down formore misery, as defaults among consum­ers and companies spiral. In its latest Glo­bal Financial Stability Report, the IMF esti­mates that the total bill for �nancialinstitutions will come to $4.1 trillion.

With so much red ink still to be spilled,it may seem premature to ask, as this spe­cial report does, what the future of bank­ing looks like. For most industries, failureon this scale would mean destruction,after all. Banks, notoriously, are di�erent.The most seismic event of the crisis to date,the bankruptcy of Lehman Brothers lastSeptember, demonstrated the costs of let­ting a big �nancial institution collapse.Trust evaporated and credit dried up. �Oc­tober was the most uncomfortable mo­ment in my career,� recalls Gordon Nixon,

Rebuilding the banks

BANKING is the industry that failed.Banks are meant to allocate capital to

businesses and consumers e�ciently; in­stead, they ladled credit to anyone whowanted it. Banks are supposed to makemoney by skilfully managing the risk oftransforming short­term debt into long­term loans; instead, they were undone byit. They are supposed to expedite the �owof credit through economies; instead, theyended up blocking it.

The costs of this failure are massive.Frantic e�orts by governments to savetheir �nancial systems and buoy theireconomies will do long­term damage topublic �nances. The IMF reckons that aver­age government debt for the richer G20

countries will exceed 100% of GDP in 2014,up from 70% in 2000 and just 40% in 1980.

Despite public rage over bank bail­outs,the industry has also comprehensivelyfailed its owners. The scale of wealth de­struction for shareholders has beenbreathtaking. The total market capitalisa­tion of the industry fell by more than halfin 2008, erasing all the gains it had madesince 2003 (see chart 1, next page).

Employees have scarcely done better.The popular perception of bankers asPorsche­driving sociopaths obscures thefact that many of the industry’s sta� aremodestly paid and sit in branches, infor­mation­technology departments and call­centres. Job losses in the industry havebeen savage. �Being done� used to refer to

More articles about banking are at

Economist.com/banking

An audio interview with the author is at

Economist.com/audiovideo

A list of sources is at

Economist.com/specialreports

Exit rightThe contract between society and banks willget stricter. Page 3

Don’t blame CanadaA country that got things right. Page 5

From asset to liabilityThe shifting shape of bank balance­sheets.Page 6

Too big to swallowThe future of securitisation is the industry’smost pressing question. Page 9

Opportunity gently knocksWho will gain from the crisis? Page 11

The revolution withinThe way banks manage risk will change�aswill the way they reward managers. Page 12

Back at the branchMore Swedish lessons for the banking in­dustry. Page 14

From great to goodBanks will still make money, just less of it.Page 16

Also in this section

AcknowledgmentsApart from the people quoted in this report, the author isgrateful to the following people for their help: ViralAcharya, David Aldrich, Lucian Bebchuk, MarkusBrunnermeier, Simon Gleeson, John Grout, Colm Kelleher,Naguib Kheraj, Mark Richards, Til Schuermann, LarryTabb, Michael Tory, Rick Watson and numerous people atAllen & Overy, the Bank for International Settlements, theBoston Consulting Group, Deloitte, Oliver Wyman andPricewaterhouseCoopers.

1

2 A special report on international banking The Economist May 16th 2009

2 the boss of Royal Bank of Canada (RBC).�There was a possibility that the entire glo­bal banking system could go under.�

Concerted actions by governmentssince then, �rst in the form of capital injec­tions and liability guarantees, and more re­cently via schemes to buy or guaranteeloans, have signalled their determinationto stabilise and clean up their big banks.

Politics notwithstanding, the commit­ment of governments to defend theirbanking systems removes the existentialthreat to the biggest institutions (or, moreprecisely, transfers it to sovereign borrow­ers). Bank bosses have learnt not to pro­nounce too con�dently about the future. Ifthe IMF’s loss predictions turn out to be ac­curate, there is still too little capital in thesystem. But most think that the chance ofanother Lehman­style blow­up has beengreatly reduced.

There is still great uncertainty about thenature and extent of the support that gov­ernments will end up o�ering to theirbanks. But governments are now deeplyembedded in banking systems. They areguaranteeing far more retail deposits thanbefore the crisis. They are guaranteeing theissuance of new debt. They own preferredshares in many banks, common equity inothers and stand ready to inject capital inothers still. Banks that have not taken ascrap of government money still bene�tfrom their stabilising presence. �We all ex­ist at the largesse of the government rightnow,� says a bank boss.

The types of losses that banks now facehave also changed. The huge writedownson trading­book assets that de�ned the�rst phase of the crisis were horribly un­predictable. The complexity of structured�nance made it di�cult to know howlosses would cascade down the ladder ofinvestors in securitised assets. The patchycredit histories of subprime and low­docu­mentation borrowers made it hard to mod­el default rates accurately. And mark­to­market accounting meant that banks werevaluing illiquid assets at prices which re­�ected a lack of buyers as much as under­lying credit quality (accounting­standardsbodies have since been bullied into allow­ing bankers to exercise more judgment inhow they classify and value such assets).

Although the losses that banks face intheir loan books are ugly, they should bemore predictable. Shocks are still likely: forinstance, the size of the bubble and scale ofthe bust may overturn historic relation­ships such as that between unemploy­ment rates and credit­card losses. Butlosses on loans can be recognised in the ac­

counts more slowly. And the assets that arenow under scrutiny may be much biggerthan their subprime predecessors but theyare also better understood. �The scale ofthe recession is unprecedented but it ismore familiar terrain,� says John Varley,the chief executive of Barclays.

The forgotten artWith government backing assured and im­pending losses somewhat more predict­able, the big banks are slowly starting to lifttheir heads from the �oor. Meetings withinvestors have been dominated for thepast 18 months by discussions aboutbanks’ balance­sheets and, in particular,the amount of capital that banks had.�This is my �rst experience of the quarter­ly­earnings game where no one has caredabout earnings,� says Bob Kelly, the boss ofBank of New York Mellon.

That is changing. Even the biggest vic­tims of the crisis expect to return to pro�t­ability this year. Galling as it may be to con­template the returns that will once againaccrue to banks, the rest of us badly needthem to make money. Just as the prospectof continuing losses is what has stoppedprivate capital from entering the system,the prospect of future pro�ts is what willlure investors back in to replace govern­ments. Pro�tability is also critical to theability of banks to cover future losses with­out calling on further government cash.The situation is �uid but analysts at Bar­clays Capital reckoned in March that cum­ulative pre­tax and pre­provision incomeat the top 20 American banks for this year,2010 and 2011 will be $575 billion, justenough to cover their estimates of losses inthat period of $415 billion­$560 billion.

Pro�ts need to be sustainable, of course.They may be the �rst line of defenceagainst trouble but they disappeared all

too quickly during this crisis, wiped out bywritedowns and by the implosion of busi­ness models. �The discounted future pro�tstreams of �nancial institutions went fromquite something to almost nothing in aninstant,� says Andy Haldane, head of �­nancial stability at the Bank of England.

Banks recognise this as much as regula­tors do. There is a striking degree of conver­gence between the thrust of planned regu­latory reforms and the new strategicthinking of many institutions. Greater re­silience is a shared objective. Banks are re­ducing their dependence on wholesalefunding and increasing their reliance on�stickier� deposits. They are reducing theamount of risk they take, which means re­ducing their proprietary trading and con­centrating more on clients and activitiesthat consume less capital. They are rapidlyshrinking their balance­sheets. �The bank­ing industry got it so wrong and destroyedso much value that it is di�cult to sit infront of investors and say we are going tocarry on as before,� says Richard Ramsden,an analyst at Goldman Sachs.

The future looks di�erent to di�erenttypes of banks. For smaller ones that falloutside the comforting embrace of thestate or have less diversi�ed loan portfoli­os, the outlook is bleaker. American re­gional banks and Spanish savings banks,or cajas, are among those coming under in­creasing pressure as commercial­propertyportfolios su�er. Mike Poulos of OliverWyman, a consultancy, expects the num­ber of banks in America, currently some8,000 or so, to drop by 2,000 or more as aresult of the crisis.

Banks in many emerging markets willsu�er as the economic climate deterioratesbut they need to deleverage less. There isalso less need for regulatory change. TheAsian banks kept their exposure to cross­border funding �ows under control, for ex­ample, unlike their peers in eastern Eu­rope. The scale of structural change thatthese institutions face is relatively limited.

But for those banks at the heart of thecrisis, the household names of Western �­nance, the landscape is di�erent. Their fu­ture is secure enough for them to be able toplan beyond survival. Their failures havebeen big enough for them to know thateverything they do, from the way theymanage their balance­sheets to the waythey pay their managers, has to change.But in seeking to work out what the newnormality will be for banks, the �rst ques­tion to ask is how quickly and on whatterms governments will disentangle them­selves from the industry. 7

H12007

H22007

-3.2

1Worth less

Source: Boston Consulting Group *Year end

Global banking industry, total market capitalisation$trn

0

2

4

6

8

2007* Q12008

Q22008

Q32008

Q42008

2008*

-13.2

-10.0-12.7

-29.9

% change on previous quarter

The Economist May 16th 2009 A special report on international banking 3

1

NOTHING highlights the scale of bank­ing’s upheaval better than the inter­

vention of governments. An industry thatembodied the free market turns out to bepathetically dependent on the state for itssurvival. In some cases, the civil servantsare o�cially in charge. The taxpayer is al­ready the majority owner of Royal Bank ofScotland (RBS) and Lloyds Banking Groupin Britain. The German government ispoised to take control of Hypo Real Estate.American taxpayers are set to own thelargest single stake in Citigroup. In manymore cases, o�cials exercise control with­out formal representation, imposing paylimits and lending targets. The governmentis the industry’s largest shareholder andthe guarantor of its liabilities.

Yet the magnitude of this shift can easi­ly be overstated. Governments routinelystep in to rescue banks at times of systemicdistress, observes Claudio Borio of theBank for International Settlements. Ratingagencies have long assessed banks’ credit­worthiness in part on the likelihood ofgovernment support should they get intotrouble. Their judgment, as everyoneknows, is not always right. Moody’s waspilloried in early 2007 for awarding gold­plated AAA ratings to the big Icelandicbanks on the false premise that the au­thorities in Reykjavik could a�ord to res­cue them. But the assumption that govern­ments will try to help a big bank in crisis isnothing new.

This contract to intervene was �rst le­gally recognised after the Depression,when the Glass­Steagall act of 1933 createdthe Federal Deposit Insurance Corporation(FDIC). Since then, similar deposit­guaran­tee schemes have been created around theworld to help persuade savers, who areotherwise unsecured creditors of theirbank, not to remove their money if it getsinto trouble. Indeed, some advocates offree markets argue that this guarantee ofcompensation helped to cause the currentcrisis, by reducing the incentives for depos­itors to look closely at their banks.

Whatever the merits of that argument,it is whistling in the wind to suggest thatthe state should withdraw from its com­mitment to support banks in times of trou­ble. �The body cannot survive without

blood,� says Bo Lundgren, one of the archi­tects of Sweden’s vaunted bank­rescuepackage of the early 1990s, �and the econ­omy cannot survive without banks.� Butnow that this commitment has been calledon so dramatically, three questions arise.The �rst is how long the state will remainso explicitly involved in the industry. Thesecond is what immediate distortions thatinvolvement creates. And the third is whatadditional charges governments will levyon the industry in future for providingbanks with such a huge safety net today.

The answer to the �rst question will bemeasured in years. Take Sweden’s bankbail­out. It was more successful than any­one had expected but it still took four yearsfor the liability guarantees to be lifted.Nordbanken, the seed of today’s Nordea,was fully nationalised in 1992 and partlyre�oated three years later but the Swedishstate remains its largest shareholder.

The Swedish policymakers’ task wasalso less daunting. The bad assets in theirbanks were more homogenous and easierto value than those currently cloggingthings up. The Swedes intervened at theend of a recession, so banks quickly bene­�ted from the recovery. Governments to­day have had to step in earlier in the eco­nomic cycle, implying a longer period ofengagement for two reasons.

First, while loan losses continue to raisedoubts about banks’ solvency, the pres­ence of governments will be necessary toreassure creditors and counterparties. InAmerica the healthiest banks are increas­

ingly vocal about their desire to repaymoney from the Troubled Asset Relief Pro­gramme (TARP). But many bankers also re­cognise that they should not be too hastyin their bid for freedom. We are not goingto pay o� TARP money until we are certainwe don’t need it, says a bank boss.

Regulators may not allow relativelystrong banks to buy out the governmentearly in any case, for fear of a further lurchdownward in the economy and of leavingstraggling institutions vulnerable to attackfrom short­sellers. �The idea of TARP re­payment is a nonsense,� steams a WallStreet executive. �It all has to be paid backat the same time.�

Finding a way outEven if it is feasible to replace governmentequity fairly quickly, most believe that itwill take far longer for governments to exittheir debt guarantees. Banks have lots ofbubble­era debt to re�nance this year andnext. The coming torrent of governmentborrowing may make it harder for banks toattract private funding. And the more gov­ernment­backed bank debt is issued, thegreater the risk of creating another re�­nancing problem when state guaranteesexpire.

The second reason why governmentsneed to stay engaged is to counter thebanks’ usual instincts during slowdowns.The obvious thing for banks to do in a re­cession, let alone one in which trust incounterparties has been shattered and acredit bubble is de�ating, is lend less (seechart 2). Governments are urging banks tolend more to prop up the economy, eventhough in the long term they will wantthem to be more cautious lenders.

The political imperative for govern­ments to try to make a return on their in­vestments complicates matters further.Banks will have to look relatively risk­proof before they can be passed back intoprivate ownership at a pro�t. All of whichsuggests that governments have to negoti­ate a prolonged transition before they willexit all of their investments in banks or re­move their liability guarantees.

The longer governments stay involved,the more they will distort competition.Normally, private �rms moan about hav­

Exit right

The contract between society and banks will get stricter

2Drying up

Sources: Federal Reserve;Goldman Sachs; The Economist

*Excluding Washington Mutualand money-market assets

US commercial-bank credit*3-month moving average, % change:

J A S O N D2007

J F M A M J J A S O N D08

J F M09

5

0

5

10

15

20

+

–

on previous 3 months, annual rate

on previous year

4 A special report on international banking The Economist May 16th 2009

2

1

ing to compete with state­backed rivals butin this case government backing is likely tochange from a boon into a handicap. Bankbosses in America who welcomed the ini­tial injection of TARP capital have becomeprogressively less enthusiastic about it.Those who have stayed outside the gov­ernment net, in terms of equity participa­tion at least, are revelling in their indepen­dence. �There is some tactical advantage togovernment money but it is deeply politi­cised,� says the boss of a big bank whichhas not taken state cash.

Compensation is the obvious example.The top 25 employees at banks that havetaken TARP money face tight regulation oftheir incentive­based pay until the govern­ment has been paid back. Prior bonusesare also at risk from punitive tax proposals.That may be sustainable for a while, saysanother boss: �We can say for a year or twothat ‘we value you, you’re a leader and if

you stay with us we will make it up toyou’.� But eventually competition fromunfettered rivals will tell.

Freedom to act on the internationalstage is particularly prized by institutionsthat have not taken government cash. Tax­payers have little interest in seeing theirmoney used to �nance activities in othercountries: they want it used for lending athome. The dismantling of RBS’s global em­pire is the most conspicuous example ofthis type of �nancial nationalism, but pres­sure to lend domestically is universal.With many competitors gone, impaired orunder the cosh of government masters,banks that have been able to keep operat­ing normally in global markets are alreadygrabbing new wholesale business. Capi­tal­raising is easier for independent bankstoo because shareholders and politicianshave di�erent priorities. �Investing capitalwhere government is involved on a con­

tinuing basis is di�cult because of con­cerns over restrictions on marketing andcompensation expenses,� says Gary Parrof Lazard, an investment bank.

There are also disadvantages to havinggovernment­owned rivals. The obviousone is unfair competition. Northern Rock,a British bank which was nationalised inearly 2008 and was originally told to shutits doors to new borrowers and shrink itsbook, abruptly changed course in Febru­ary. It now aims to lend an extra £5 billion($7.6 billion) in mortgages in 2009, and upto an additional £9 billion in 2010. If gov­ernment­owned banks were to underpricerisk for a long period of time in order tomeet lending targets, everyone would feelunder pressure to respond.

The shadow of the stateLet’s be foolhardy and assume the best.Economies start to recover relatively rapid­ly. Governments are able to plot a relative­ly fast exit from their equity investments.And a revival in funding markets allowsfor a smooth exit from debt guarantees (ashappened in Sweden). Even so, the crisiswill leave a lasting mark on the terms oftrade between banks and the taxpayerswho periodically come to their rescue.�Banks have to have a licence to operate,which is granted by a common under­standing of what is right and fair,� saysHans Dalborg, the chairman of Nordea.

Some elements of this new contract be­tween banks and society are already clear.Amendments to bank­capital regimes arecertain, although regulators clearly do notwant to squeeze banks to raise more capi­tal until credit shortages have eased. Thereis now impressive momentum behind theidea of a leverage ratio, a measure that putsa �xed ceiling on the total amount of assetsthat a bank can hold relative to its capital.Some countries, including America, al­ready have such a system, and others arefast coming around. The Swiss are intro­ducing just such a ratio for their two big­gest banks, which will be phased in by2013, to sit alongside the international �Ba­sel 2� capital rules.

Basel 2 takes a di�erent approach tocapital, charging banks on the basis of howrisky their assets are. These �risk weights�will also become far more punitive. Ask su­pervisors about the biggest �aws in theprevious regulatory framework and manywill point to the meagre capital chargesthat banks faced in their trading books,which were based on disastrously optimis­tic assumptions about the liquidity, riskpro�le and price stability of assets such as

The Economist May 16th 2009 A special report on international banking 5

2

1

mortgage­backed securities. These chargesare going to be driven higher.

The liquidity of banks’ balance­sheetswill also be regulated more intensively.Britain’s Financial Services Authority(FSA) has already issued proposed guide­lines on liquidity which will require banksto hold a greater cushion of liquid assets,mainly in the form of government bonds.The proposals have attracted plenty of crit­icism but they are indicative of what iscoming: a more robust approach to liquid­ity in general and, in the wake of the Leh­man bankruptcy and the collapse of theIcelandic banks, greater e�orts by nationalregulators to safeguard the local opera­tions of foreign banks from the risk of theirparents getting into trouble.

If the regulation of balance­sheets is setto become more prescriptive, other thingswill be designed to increase levels of un­certainty. Take the stance of Britain’s newlyscary FSA. Its previous philosophy meantthat the regulator focused primarily on the

management controls and systems thatbanks had in place. That passive approachwill be replaced by a more intrusive andcapricious regime, which questions the de­cisions of individual institutions.

Uncertain timesWidespread enthusiasm for a more �mac­roprudential� approach to regulation�inwhich regulators think harder about thestability of the system in addition to thehealth of individual institutions�also im­plies a higher level of uncertainty for exec­utives. Banks that may be doing a good jobcould still �nd themselves subject to high­er charges if systemic risks are rising.Countercyclical rules requiring banks tobeef up capital in good times and run itdown in bad times may well rely on thediscretion of authorities.

What of the two big structural ques­tions that now dog industry regulators�whether to separate out �utility� retailbanks from �casino� investment banks;

and what to do about those banks that aretoo big to fail? Both problems have gotmore acute because of the crisis. Dealssuch as the takeovers of Bear Stearns byJPMorgan Chase, and of Merrill Lynch byBank of America, have further blurred theboundaries between retail and investmentbanks, not sharpened them. Combina­tions like those of Wells Fargo and Wacho­via, Lloyds TSB and HBOS, Commerzbankand Dresdner Bank have bloated the big­gest institutions, not slimmed them. Andthe trend towards concentration of depos­its among America’s top banks has acceler­ated as a result of these deals, for example(see chart 4 on next page).

Yet despite some talk about the need fora new Glass­Steagall act to separate retailand investment banking, and for highercapital charges based on size, the idea ofbreaking up institutions does not havegreat momentum. No business model hascome through the crisis unscathed and sizeis manifestly not the only attribute that

�IT IS the only time I feel like royalty,�says the boss of a big Canadian bank,

describing the reception he now gets inAmerica. He is not the only one basking inacclaim. All of Canada’s main banks werepro�table in the quarter ending January31st, when market conditions were at theirworst. None has needed government in­vestment. The country’s �nancial systemhas been praised by Barack Obama.

Trouble is, some di�erences betweenthe two countries are culturally ingrained.�The United States has an inherently high­er risk appetite,� says a banker familiarwith both sides of the fence. It is hard to�nd pre­crisis equivalents in America ofthe decision by Toronto­Dominion (TD) toexit its structured products business in2005, or the 20­30% band that RBC im­poses on the share of earnings that its cap­ital­markets business can contribute.

Structural di�erences matter too. TheCanadian system is an oligopoly of �vedominant banks. That dampens pricecompetition: independent brokers origi­nate less than one­third of the mortgagesin Canada, compared with up to 70% inAmerica during the bubble. It also makes

it easier for Canadian banks to pull backwhen things are getting too risky.

Having a few banks that are clearly toobig to fail has led to more stringent super­vision, including imposing a maximumleverage ratio and a single regulatory re­gime for commercial and investmentbankers. Laxer and more fragmented capi­tal regimes allowed the balance­sheets ofbanks elsewhere to balloon (see chart 3).

Perhaps the most striking divergencebetween Canada and America is in theirregulation of mortgages. Interest paid onhome loans is tax­deductible in America,encouraging people to borrow more; notso in Canada. American mortgages arenon­recourse in many states, making itharder for lenders to pursue defaultingborrowers; not in most of Canada. (Thenagain, Britain is like Canada in these re­spects but still has soaring defaults).

Canadians taking out mortgages witha loan­to­value ratio over 80% must alsotake out insurance on them from a federalagency called the Canada Mortgage andHousing Corporation (CMHC). The banksinsure the rest of their portfolios with theCMHC, which keeps them honest by ap­plying strict standards to the mortgagesthey guarantee. Taking out insurance alsobrings the risk weighting that regulatorsapply to these mortgages down to zero,which means that the banks derive nocapital advantage from funding themthrough securitisation. Some argue thatFreddie Mac and Fannie Mae, America’shousing­�nance giants, should likewiseguarantee mortgages but not buy them.

A country that got things rightDon’t blame Canada

3Bank on Canada

Sources: Thomson Datastream; The Economist;Royal Bank of Canada; Royal Bank of Scotland

Banks’ assets, 1997=100, C$ terms

1997 99 2001 03 05 07 08

1000

500

1,000

1,500

2,000

2,500

Royal Bank of Scotland

Royal Bankof Canada

6 A special report on international banking The Economist May 16th 2009

2

1

makes a bank too important to fail. Standalone investment banks have

failed and, as Lehman vividly demon­strated, were too central to the architectureof global �nance to disappear smoothly.Pure retail banks have imploded too. In­vestment bankers archly observe thatjudgments on which bit of the business isthe casino ought to be withheld until theend of the credit cycle. The woes of Citi­group put paid to the myth of the inde­structible universal bank, even as the suc­cess of Canada’s banks (see box onprevious page) showed that a system of afew domestic giants can work.

Any radical regulatory surgery wouldalso require governments to mark outsome very arti�cial boundaries. Take thedistinction that some make between de­posit­taking institutions, which should beprotected, and wholesale­funded entities,which should not. With so much whole­sale funding coming ultimately from indi­vidual investors in the form of pensionand mutual funds, that distinction is blur­rier than it �rst looks.

There are similar problems with de�n­ing the borders between acceptable and

unacceptable activities. Peter Sands, theboss of Standard Chartered, an emerging­market leader, argues that there areswathes of the industry doing blamelessbut critical things like cash managementand trade �nance for companies that falloutside the de�nition of narrow banking.

What is more, any form of lending en­tails risk. The extension of credit to a smallbusiness is one of the riskiest things a bankcan do, but it wins taxpayers’ unequivocal

support. Credit­default swaps are vili�ed,by contrast, but they serve a valuable func­tion. �We will buy credit protection but notsell it, buy catastrophe risk [protection] butnot sell it,� says the boss of a bank that hasnegotiated the crisis successfully. Fine, butthat implies it is useful for someone to beselling these kinds of instruments. Propri­etary trading is harder to defend when it issheltered by a government guarantee butany bank that acts as a marketmaker be­tween buyers and sellers will end up tak­ing some form of proprietary risk.

Faced with this untidy set of choices, asensible philosophy would not makehard­and­fast judgments about what busi­nesses belong together. Quality of man­agement, not business models, is better atexplaining the di�erence in performancebetween banks. The right approach con­ceptually is a dynamic regulatory regimethat looks sceptically at the boardroomsand strategies of �nancial institutions andis capable of intervening e�ectively whenneed arises. In any case, systemic changesto institutions’ balance­sheets will have asubstantial impact on the types of busi­nesses banks become. 7

4Winners take more

Source: Goldman Sachs *February

Top 3 US banks’ domestic deposits, % of total

1994 96 98 2000 02 04 06 09*0

5

10

15

20

25

30

35

THE dirty secret of the golden age of �­nance was that it was obscenely easy to

make money. The supply of credit wasseemingly inexhaustible, so banks couldfund their expansion at will. Demand wasequally insatiable, providing those infa­mously complex structured products witha stream of ready buyers. The years ofplenty disastrously skewed risk models, al­lowing banks to run with lower capital onthe assumption that past performancewas, contrary to the industry’s standardadvice, a guide to future returns. And thetheory that risk had been dispersed be­cause of securitisation added to the falsesense of security.

The result for many banks was a strat­egy of expanding their balance­sheets bywriting more and more loans and holdingever more securities. With risk low, liquid­ity ubiquitous and many institutions un­der �re for appearing to be overcapitalised,there seemed to be little cost to growth. �Ifwe could have an in�nite balance­sheet fora penny return, we were going to take it,�

says a Wall Street veteran. Things are somewhat di�erent now. If

boardroom discussion in the past decaderevolved around the asset side of the bal­ance­sheet, the next decade will see man­agers focusing on the liabilities side�theamount and quality of capital they hold toprotect against losses, and the durationand sources of their funding. Banks will gofrom being unconstrained by their bal­ance­sheets to being caged by them.

Start with capital, which has suddenlybecome the industry’s scarce resource.That is particularly true today as the pros­pect of further losses continues to unnerveprivate investors. But it will remain trueafter the immediate crisis has eased. Theamount of capital that banks have to holdwill go up, and not just because their regu­lators want them to have a bigger bu�eragainst losses.

The risk weightings on assets are risingas the e�ects of the downturn feed throughinto banks’ risk models, forcing them to setaside more capital. Bondholders want a

greater cushion beneath them in the capi­tal structure to protect them against losses.Shareholders too are belatedly happy totrade higher returns on equity for a re­duced chance of being wiped out. Sobanks with more equity capital are nowvalued more highly by the market. Be­tween 2000 and 2007 there was no corre­lation between equity­capital ratios andthe total return on banks’ shares, says MrVarley of Barclays. �Now the correlation ismeaningful.�

The amount of capital banks hold is notthe only thing under scrutiny. They alsoneed to have the right kind. Their capital isa mix of common equity, which is �rst inline when losses strike, and various otherinstruments, often hybrids of equity anddebt. A gradual pre­crisis increase in theproportion of this sort of capital has accel­erated rapidly in recent months, as com­mon equity has been eaten up by lossesand governments have largely �lled thegap with preferred shares, which helps toavoid nationalisation.

From asset to liability

The shifting shape of bank balance­sheets

The Economist May 16th 2009 A special report on international banking 7

2

1

The curious e�ect of this changing capi­tal mix has been to bolster banks’ overallcapital bases while disturbing share­holders, who see common equity as theonly dependable bulwark against losses.Thus banks with less of it have been pun­ished by the markets. �It turns out that hy­brids don’t have much loss­absorbing ca­pacity and are not much use in a period ofstress,� concludes Mr Ramsden of Gold­man Sachs. That realisation helps to ex­plain why American commercial banks,despite appearing well capitalised com­pared with their European peers ahead ofthe crisis, have still had a capital problem.It also helps to explain the banks’ rush tobuy back hybrid debt at discounted prices:they can book the gains as pro�ts and usethese to beef up capital where it counts.

Holders of hybrid instruments, as wellas of other forms of junior debt, have beengiven their own reasons to re�ect. The Brit­ish government’s decision in February toamend the terms of subordinated debt is­sued by Bradford & Bingley, a nationalisedbank, spooked European markets, for ex­ample. Bondholders were locked in whenDeutsche Bank decided in December notto redeem a ¤1billion ($1.3 billion) subordi­nated bond at its �rst opportunity. JohnRaymond of CreditSights, a research �rm,says investors used to like buying debtlower down banks’ capital structure be­cause they thought its higher yield over­compensated for a marginal increase inrisk. Now their thinking has changed.

Senior debtholders, who rank �rst inthe hierarchy of unsecured creditors if abank is liquidated, have less to worryabout. In general regulators have stuck tothe standard script of bank bail­outs, inwhich shareholders take the pain andbondholders are protected. And a biggerequity cushion should help to reduce thecost of debt by counteracting fears that

debtholders are too exposed to losses. Bank debt of all kinds will nevertheless

be perceived as more risky after this crisis.Investors will not soon forget WashingtonMutual’s failure last September, when as­sets and deposits of the Seattle­based thriftwere transferred to JPMorgan Chase but itscreditors were left high and dry. Even incountries that do not formally prioritisedepositors over other creditors, as Ameri­ca does, the political necessity of reimburs­ing taxpayers before anyone else has be­come crystal clear. Thanks to Iceland’scrisis the creditworthiness of banks willalso be far more closely tied to the credit­worthiness of the countries in which theyare headquartered.

The primary e�ect of all these changeswill be to make it more expensive to ex­pand the balance­sheet. Scared bankshareholders will now demand a higherrisk premium, as will debtholders. Compe­tition for capital and safer forms of debtwill be greater, as investors demand for­tress­like balance­sheets. And equity willgo less far in a world where banks are moreconstrained in the amount of assets theycan support with each unit of capital.

All this in turn will lead banks to thinkharder about where they deploy capital.Executives will ask tough questions aboutactivities that absorb lots of capital buthave lower returns now that leverage islower, the risks are clearer and cost of fund­ing is higher. �Some banks’ balance­sheetscould be expanded inde�nitely in thepast,� says Paul Calello, the boss of CreditSuisse’s investment bank. �Now that capi­tal is more scarce the banks have to be evenmore e�cient in their balance­sheet andcapital usage to maximise pro�tability.�Dedicated proprietary­trading desks,where a group of traders put the bank’sown capital at risk, look much less attrac­tive in this changed environment, for ex­

ample. The advantages of running suchdesks have largely gone, says Bill Wintersof JPMorgan Chase’s investment bank.

Finer judgments about the liquidity ofassets will also come into play. When mar­kets are less liquid, assets stay on the bal­ance­sheets for longer. That exposes insti­tutions to greater risk and ties up capitalthat could be better deployed elsewhere.Credit Suisse is planning to continue to op­erate in American residential mortgage­backed securities (RMBS), for example,where markets are deeper and more liq­uid, but exit the sludgier European RMBS

market, where the bank is forced to holdassets for longer.

Businesses that throw o� plenty ofearnings without absorbing much capitalor running great risks are naturally in de­mand. Take the advisory businesses of in­vestment banking, an area in which plentyof boutiques make a decent living withouthaving a balance­sheet at all. Or custodybusinesses, where banks look after the as­sets of other �nancial institutions. Or assetmanagement, where someone else’s mon­ey is at risk. The two remaining indepen­dent investment banks, Goldman Sachsand Morgan Stanley, have both signalledgreater emphasis on less capital­intensivebusinesses.

Debt dilemmasCapital is not the only bit of the balance­sheet that will get a lot more attention infuture. Executives (and regulators) willalso focus more on the funding pro�le oftheir businesses in light of the crisis, as thecosts of borrowing rise, lenders demandgreater security and keener awareness ofliquidity risk informs behaviour.

They will pay greater attention, for ex­ample, to whether assets can be used ascollateral for further borrowing. Huw vanSteenis, an analyst at Morgan Stanley, says

8 A special report on international banking The Economist May 16th 2009

2 banks will divide activities between thosegenerating collateral that can be placedwith central banks (high­quality mort­gages, for example) or is otherwise decentenough to be used as security (shares, say),and those that do not throw o� any collat­eral at all and therefore consume unse­cured funding. The cost and scarcity of thistype of funding means that these business­es�equity underwriting is an example�will command higher margins.

Above all, banks will have a deeperawareness of funding risk�their ability toroll over debts as they come due. Institu­tions that keenly exploited the pricing dif­ferences between long­term assets andshort­term liabilities paid heavily when li­quidity dried up and they were unable tore�nance fast­maturing debts or sell the as­sets that they held.

To be more precise, the weakness re­vealed by this crisis has been in short­termwholesale funding, which rolls over quick­ly but does not have the government­backed guarantees that help to keep retaildepositors quiescent. This type of fundinghas been at the heart of the crisis.

Many subprime mortgage­backed secu­rities were held in o�­balance­sheet vehi­cles that funded themselves by issuingshort­dated, asset­backed commercial pa­per to money­market funds and other in­vestors. When those funds suddenlystopped buying paper, the banks’ liquiditylines to these vehicles abruptly came intoforce. Similarly, the amount of funding thatinvestment banks were doing throughovernight repo agreements surged be­tween 2004 and 2007; they were rollingover one­quarter of their balance­sheetsevery day prior to the crisis, making themvulnerable to a sudden loss of con�dence.

Short­term wholesale funding alsohelped to sink Northern Rock, one of theearliest victims of the crisis. The bank’sfailure in September 2007 is indelibly asso­ciated with images of Britain’s �rst retailbank run since 1866 and is often blamed onits enthusiastic use of securitisation to ex­pand its mortgage book. Yet a 2008 paperby Hyun Song Shin of Princeton Universi­ty, dissecting the bank’s implosion, sug­gests that neither the run nor securitisationwas the principal culprit.

The retail run on the bank came in mid­September, when news broke that theBank of England was providing it withemergency support. Yet Northern Rockhad been experiencing funding problemssince mid­August. The retail run came afterthe bank had already been destabilised bywholesale­funding problems. Nor can se­

curitisation really be blamed for the with­drawal in funding. Northern Rock’s securi­tisation vehicle issued relatively long­termnotes to investors, so it did not face thethreat of massive redemptions from thisparticular quarter. The �rst and most da­maging run on the bank took place in itsother short­ and medium­term wholesaleliabilities (see chart 5).

Faced with this stress fracture in theirfunding structures, banks have two obvi­ous ways to respond. The �rst is to length­en the maturity of the wholesale debtsthat they have. Some institutions have lessfar to go than others: the unsecured debt ofGoldman Sachs already boasts an averagematurity of eight years, for instance. Butwhen markets get back to normal, fundingmaturities are likely to rise.

There are limits to issuance of longer­dated liabilities. A big rise in the propor­tion of long­term bond funding across theindustry is bound to be costly, especiallysince banks will have to compete to attractinterest from bond investors who already

have exposure to many of these institu­tions anyway. Securitisation markets arebadly damaged (see next section).

The second option, and the more im­portant shift, is for banks to increase theirdependence on more stable deposits. Themost dramatic volte­face has been that ofthe surviving investment banks, GoldmanSachs and Morgan Stanley, which becamebank holding companies in September,making it easier for them to take deposits.

The upheaval in funding pro�les is ar­guably greater for retail banks, however.Just as capital ratios are now strongly corre­lated with share prices, so too are loan­to­deposit ratios (LDRs). Among emergingEuropean countries, where cross­bordercapital �ows were critical and have nowdried up, those with higher LDRs (ie, fewerdeposits) have seen their banks’ shareprices dive the most (see chart 6).

Retail­bank bosses who had �nancedloan growth by tapping wholesale sourcesof funding are now targeting lower LDRs.Before it was snapped up by Lloyds TSB,HBOS, another stricken British lender, wastrying to dispose of businesses that weredependent on wholesale funds and there­fore increased the group’s LDR. Manybanks have now imposed limits on assetgrowth, requiring that loans do not expandmore quickly than deposits.

A perennial industry debate, on themerits of bank branches versus other cus­tomer channels, has been given fresh pi­quancy by this need to gather in deposits�and the branch is likely to emerge strength­ened. Studies consistently suggest thatbranch networks with a strong local pres­ence are the most e�ective way to win de­posits. Mr Shin’s analysis of Northern Rockcontains some fascinating detail on the re­tail­deposit run. Despite those snakingqueues, customers with branch­based ac­counts were stickier than many others.Online account­holders �ed in roughlysimilar proportions to branch customers.Holders of postal accounts and o�shoreaccounts were �ightier.

In truth, banks need to have a diverseset of funding sources and maturities,whether wholesale or retail. Relying on de­posits alone still entails risk. Deposits canbe withdrawn at a moment’s notice, afterall, and government guarantees do notstop depositors from discriminating be­tween institutions: companies and indi­viduals alike want to avoid the hassle ofhaving to retrieve deposits from a failedbank. There will be a bigger question toconsider as well. Are there enough depos­its to go round? 7

6LDRs of the pack

Source: Citigroup

Emerging European countries’ banks

Loans as % of deposits, 2007

Ban

k sh

ares

, 20

08

, % fa

ll on

pre

viou

s ye

ar

0

20

40

60

80

100

0 40 80 120 160 200 240

Latvia

Baltics

Russia

HungaryRomania

Bulgaria

PolandTurkey

Czech Republic

5Wholesale disaster

Source: Hyun Song Shin, Princeton University

Northern Rock’s liabilities, £bn

0

20

40

60

80

100

120

June 2007 December 2007

Loan fromBank of England

Securitisednotes

Covered bonds

Retail

Wholesale

ONE of the canards of the credit crisis,trotted out regularly by politicians

and pundits, is that banks have stoppedlending. It is a charge that bankers vehe­mently reject and the data largely backthem up. It is true that overall �ows of cred­it have fallen steeply. Yet analysis by OliverWyman, a consultancy, suggests that netlending by American banks, for example,has contracted by amounts that are broad­ly in line with previous recessions, whendemand for credit naturally diminishesand lending standards inevitably tighten.Indeed the worry of some observers, giv­en that easy credit got us into this mess, isthat banks are still lending too much.

The really precipitous contraction incredit has come from non­bank lenders�the array of money­market funds, hedgefunds, former investment banks, ex­change­traded funds and the like that issometimes called the �shadow bankingsystem�. These capital­market lenders areespecially important in America�bankshave supplied only 20% of total net lend­ing in the country since 1993 (see chart 7,left­hand side)�but they play an increas­ingly important role elsewhere too.

In particular, non­bank lenders havebeen buyers of securitised products, loansthat are bundled together into securitiesand sold on to investors. An estimated $8.7trillion of assets worldwide are funded bysecuritisation. More than half of the creditcards and student loans originated inAmerica in 2007 were securitised. ManyEuropean banks used securitisation tofund the expansion of their loan books inthe boom (see chart 7, right­hand side).

There is a stylised model of what ismeant to happen when the shadow bank­ing system contracts, in which banks act as�lenders of second­to­last resort�. Borrow­ers who can no longer get money from cap­ital markets can call instead on contingentfunding commitments made by the banks.And banks can fund their expanded assetbase because at the same time deposits areattracted into the banks by the comfort ofdeposit insurance. A 2005 paper from EvanGatev and Philip Strahan of Boston Col­lege and Til Schuermann of the Federal Re­serve Bank of New York showed how this�ight to traditional banking operated

when the Long­Term Capital Managementhedge fund failed in 1998.

Some of these things happened thistime too. Liquidity lines from banks to o�­balance­sheet entities such as conduitsand structured investment vehicles (SIVs)were activated as securitisation marketsevaporated. Bank executives report heavyloan demand as a result of the collapse innon­bank credit. Some savings �owed intobanks too. The problem is that the amountof money needed from the banks this timearound is so vast. Oliver Wyman calcu­lates that in the �rst three quarters of 2008,lending via capital markets in Americashrank (on an annualised basis) by $950billion. In contrast, banks’ total net lending

in 2007 was just $850 billion. Supposing for a moment that the banks

actually wanted to take on the credit riskassociated with these assets, the sums sim­ply do not add up for two reasons. First,taking securitised assets back on to bankbalance­sheets implies extra demand forcapital that would be very hard for banksto meet in benign circumstances, let alonethese ones.

Second, plenty of banks depend on se­curitisation for a big chunk of their ownfunding, so they would have to replace thissource of �nance with deposits. In mostmature markets, savings penetration is al­ready relatively deep, so there are limitedoptions for driving deposits higher still.There is greater capacity to increase depos­its in emerging markets, where there ismore cash under the mattress (see chart 8,next page), but doing so takes years. �Youcannot be disintermediated over ten yearsand then reintermediated in a month,�says Mr Nixon of RBC. Hence the near­uni­versal agreement that securitisation needsto be revived.

Resuscitation proceduresIf only it were that simple. The intellectualcase for securitisation certainly remainsstrong, and not just because without it, de­leveraging will be even more painful.Banks that have concentrations of risk intheir portfolios can reduce them by sellingassets to other investors. Those investorswho cannot extend credit directly to indi­viduals or small businesses can get expo­

Too big to swallow

The future of securitisation is the industry’s most pressing question

7Out of the shadows

Sources: Oliver Wyman; Citigroup *Q3

US banks’ net lending, % of total credit European bank system, loans as % of deposits, Dec 2008

1952 60 70 80 90 2000 08*20

0

20

40

60

AVERAGE 1952-1992=39%

AVERAGE 1993-2008=20%

80

+

–100

110

120

130

140

Britain France Euroarea

Spain Italy Germany

The Economist May 16th 2009 A special report on international banking 9

1

10 A special report on international banking The Economist May 16th 2009

2

1

sure to these assets via securitisation. �Wedidn’t come out of the internet bubble andsay that we should give up on venture cap­ital,� says a regulator.

Optimists point out that some of theworst excesses of the market have alreadygone. Ludicrously complex securitised pro­ducts, the CDO­squareds and ­cubeds,have gone forever. Greater emphasis onthe quality of borrowers will mean thatrisk should become more predictable.�The problem comes when you start secur­itising things for which you cannot com­pute the odds of default,� says StephenCecchetti, chief economist at the Bank forInternational Settlements. Even if thosepredicted default rates are high, the riskcan be mitigated by techniques such asovercollateralisation, where there is an ex­cess of loans to cover losses.

There is also an emerging consensus onhow to �x securitisation’s biggest �aw, themoral hazard which meant that origina­tors had less incentive to care about thequality of the business they wrote becausethey thought the risks were someone else’sproblem. By making issuers take the �rstloss on any defaults in the securitised poolof assets (and stipulating that they cannothedge that exposure away), regulators willgive them a clear incentive to think aboutasset quality.

This goal of aligning the interests of is­suers and investors also explains o�cialenthusiasm for covered bonds, a type ofsecured­funding instrument in whichcreditors have recourse to both assets andthe issuing bank. By keeping all the assetson the balance­sheet, however, a surge incovered bonds would still require banks to�nd a lot of additional capital. That cost ismore manageable if banks keep some ex­posure but sell most of the securities toother investors who have no recourse. As­sume a risk­weighting of 20% on a portfo­lio of high­quality mortgages, calculates Ja­mie Dimon, JPMorgan Chase’s boss, andretaining a 10% slice of a $50 billion pool ofmortgages would imply a capital charge of$80m. �That’s doable,� he says.

There is broad agreement on how a re­vived securitisation market would work(high­quality assets, simple products,some retained risk on the part of the issu­er). But big worries remain. First, regulatorsmay impose higher capital charges onbanks for the contingent risks they run as aresult of securitisation. Banks were not justundercharged for the formal liquidity linesthey o�ered to conduits and SIVs; theywere also undercharged for reputationalrisk, the informal obligation to reabsorb

troubled o�­balance­sheet assets to helptheir clients. That reputational exposurewill surely attract a more explicit cost in fu­ture. Coming changes to FAS140, an Ameri­can accounting rule for o�­balance­sheetassets, will also mean that banks can nolonger claim capital relief by securitisingassets through special­purpose vehicles.

Second, many buyers of securitisedproducts are also likely to be more con­strained in future. Leveraged investors,such as some hedge funds, are going to �ndit harder to gear up, making the returns onsecuritised products less attractive.

Banks themselves, also important buy­ers of securitised products, will have lessroom for manoeuvre too. Matt King, an an­alyst at Citigroup, believes that the surge insecuritisation during the bubble can partlybe explained by a massive mismatch be­tween the regulatory regimes of Americanand European banks. Those Americanbanks whose regulator imposed a leverageratio had an incentive to move assets o�

their balance­sheets. European bankswhich operated only under a risk­weight­ed capital regime were able to buy thosevery same assets because they attracted alow capital charge. With risk weightings onthe rise, and leverage ratios all the rage, thecapacity of European banks to purchasethese assets is shrinking.

Money­market funds, which investedheavily in securitised products, will alsobe more constrained. One of the most un­nerving moments of the crisis was themassive out�ow of cash from these fundsafter the announcement by one of themlast September that it had �broken thebuck�, meaning that its net asset value hadfallen below $1 a share and investors weregoing to get less back than they had put in.With $3.4 trillion of assets under manage­ment, allowing a run on money­marketfunds was unthinkable. The Americangovernment stabilised the market with atemporary guarantee that investors wouldnot lose money.

The issue is what kind of quid pro quomoney­market funds will now face. Thereis a particular focus on their break­the­buck commitment, which means that theymimic a bank by engaging in maturitytransformation while promising share­holders that they can get all their moneyback whenever they want it. A choice islooming for the industry�either to keepthis commitment and submit to greaterregulatory oversight, potentially includingcapital charges, or to drop it and makeshareholders understand the risk.

Neither outcome is great for securitis­ers. If money­market funds keep the break­the­buck promise, they are likely to moveinto more liquid asset classes than securi­tised products. If they abandon it, they willdemand even higher yields on securitisedassets or even greater amounts of creditenhancement, which inevitably meanshigher borrowing costs for issuers. (On the�ip side, if funds produce lower yields ormore risk in future, that could lead inves­tors to keep more of their money in banks).

Even for long­term investors�think ofpension funds and insurers with long­dat­ed liabilities of their own�likely levels ofdemand for securitisation are horriblymurky. Rating agencies are going to be farmore wary of giving AAA ratings for struc­tured products. Since many of these inves­tors have to put their money into top­ratedproducts, that implies a smaller market.

When AAA ratings are awarded, inves­tors will in any case derive less comfortfrom them. That is partly because of thehigh­pro�le failures of rating agencies and

8Banking on the unbanked

Source: Oliver Wyman

Savings and investments as % of GDP, 2007

0 25 50 75 100 125 150

Switzerland

Japan

Belgium

South Korea

Italy

Canada

United States

China

Ireland

Austria

Germany

Britain

Australia

Netherlands

Greece

Spain

Denmark

India

Finland

Sweden

France

Poland

Norway

Indonesia

Russia

Brazil

Turkey

Mexico

Savings

Investments

The Economist May 16th 2009 A special report on international banking 11

2 partly because investors are rethinkingtheir assumptions about the supposed di­versi�cation bene�ts of securitised pro­ducts. A large portfolio of securities clearlyo�ers greater protection against idiosyn­cratic risk�the chance that a particular bor­rower will get into trouble�than buying asingle­name corporate bond, say. But as apaper by Joshua Coval and Erik Sta�ord ofHarvard Business School and Jakob Jurekof Princeton University argues, a diversi­�ed portfolio o�ers far less protectionagainst systemic risk such as a general eco­nomic downturn. The chance of losses onsecuritised products increases as the econ­omy worsens; for single­borrower bonds,�rm­speci�c factors are more importantthan the economic climate. Growingawareness of this disproportionate expo­sure to systemic risk may reduce investors’appetite for securitised products.

The uncertainties do not end there.

Government intervention in America andelsewhere to ease homeowners’ repay­ment di�culties will shake investor con�­dence in future income streams. The pros­pect of court­ordered reductions inmortgage principal�or �cramdowns��isparticularly alarming. According to AnnaPinedo of Morrison & Foerster, a law �rm,there is also fogginess around the tax statusof securitisation trusts, the entities intowhich securitised assets are placed. For taxpurposes, they are structured as �pass­through� entities, meaning that the servic­ing �rms that administer mortgage pay­ments have little scope to modify the termsof loans if borrowers get into di�culty.With servicers now given greater leewayto intervene, questions about how far theycan go without compromising trusts’ taxstatus hang over the industry.

How these various uncertainties re­solve themselves will not be known for

years but two assertions look pretty safe.The �rst is that the market for securitisa­tion will shrink substantially. Borrowersare scaling back, buyers are thinner on theground, risk aversion is up and banks arein any case under pressure to improvetheir loan­to­deposit ratios. The second isthat the extent of banks’ continued deleve­raging depends to a large extent on thescale of that drop.

The wild card, of course, is the degree oflong­term support that governments arewilling to provide to buttress the market,whether through guarantees, loan pro­grammes for investors or future incarna­tions of government­sponsored enter­prises such as Freddie Mac and FannieMae. Whisper it softly, but one of the last­ing e�ects of this crisis could end up beinginstitutionalised guarantees for buyers ofsecuritised assets to sit alongside guaran­tees for retail depositors. 7

DESTRUCTIVE? Absolutely. But willthe �nancial crisis also be creative?

When incumbents disappear and estab­lished business models no longer work,that is usually good news for up­and­com­ers. The massive disruption in bankinghas members of the industry’s fringe rub­bing their hands. They include:

Advisory boutiques. �Like gnats� ishow an executive at a big investmentbank describes boutiques. Without �­nancing capacity, a global presence or bigcapital­markets businesses, they lack the�repower of bigger rivals. But the crisishas nevertheless increased their capacityto irritate the giants. Clients’ faith in theadvice of the industry’s big names hasbeen badly dented by their conspicuousinability to manage their own businesses.Many banks have damaged client rela­tionships more directly, by skimping oncredit as they slim their balance­sheets.Con�icts of interest for large banks arealso more common now that their rankshave thinned. And boutiques have lots ofhigh­quality job­hunters to choose from.

Peer­to­peer lending platforms.These websites, through which saverspool money and lend to borrowers, havealso been boosted by the crisis. Derisory

interest rates are encouraging savers toseek better returns elsewhere. Zopa, a Brit­ish website that pioneered the concept,says the number of lenders joining it hassoared. For borrowers spurned by theirbanks, low­cost and unleveraged sociallenders are an attractive alternative.Zopa’s boss, Giles Andrews, says new en­trants like his should gain from how thecrisis has undermined customers’ faith inbanks’ solidity and intensi�ed theirdoubts about whether the banks havecustomers’ best interests at heart.

Islamic �nance. This was boomingbefore the crisis, thanks to oil­fuelled li­quidity in the Gulf, rising devoutnessamong Muslims and a fast­developingmarket infrastructure. But its emphasis onrisk­sharing and prohibition of specula­tion has a fresh resonance given the fail­ures of Western �nance. Its backers stressthe ethical side of sharia­compliant �­nance. However, the Middle East is su�er­ing its own economic headwinds and theindustry’s fundamental problems, includ­ing an over­reliance on short­term fund­ing, have yet to be solved.

Supermarkets. They see the crisis asan opportunity to push further into �nan­cial services. Their costs of acquiring cus­

tomers are low, because they already havemillions of shoppers passing throughtheir stores. Their brands are trusted. Andthose who have seen how retailers workwith banks in joint ventures consistentlynote how much more focused grocers areon the customer’s needs. �Retailers think�rst about the customer, banks about thepro�t,� says an executive. Britain’s Tescoannounced an ambitious expansion of itsbanking activities in March.

Just how capable non­banks are of tak­ing big chunks of the market is unclear.The downturn is hitting most institutions,retailers included. Regulators will alsohave a big say. The rules may have beentweaked to make it more attractive forprivate­equity �rms to invest in Americanbanks, for example, but Douglas Landy ofAllen & Overy, a law �rm, expects con­tinuing hostility to the idea of non­banksowning banks. And serious questionshover about whether it makes sense to en­courage more competition in banking.�Anything that smacks of loosening regu­latory standards is going to be politicallyhard,� says Andrew Schwedel of Bain, aconsultancy. There are great opportunitieslying among the debris of the banking in­dustry but reaching them may be tricky.

Who will gain from the crisis?Opportunity gently knocks

12 A special report on international banking The Economist May 16th 2009

1

THE changes to the environment inwhich banks operate�tougher regula­

tion, higher capital requirements andscarcer funding�will have a dramatic im­pact on the way that banks are managed.But banks are also re�ecting hard on somefundamental internal questions, such ashow to manage risk, compensation andgrowth itself. Too many bosses and share­holders accepted years of double­digit re­turns without probing the sources and sus­tainability of those pro�ts. �No one wasasking the ‘Columbo’ questions,� saysToos Daruvala of McKinsey, a consultancy.

The most basic of these questions, par­ticularly for banks with large wholesaleoperations, is what kind of businessesthey want to be. The bubble was character­ised by a game of copycat, in which banksstrove to match the returns of their mostpro�table rivals by piling headlong into as­set classes where they were lagging, irre­spective of the risks. �The securities indus­try was based on revenue, not onrisk­adjusted returns,� says a bank boss.

Consultants armed with league tablesand presentations full of �gap analysis� in­creased the pressure on sluggards to catchup. Mr Winters of JPMorgan Chase recallshow executives at the bank worried aboutits underperformance in �xed­incomemarkets. �We used to beat ourselves todeath about it and wonder ‘what aren’t wegetting right?’ Now we know.� For the fore­seeable future, managers will think harderabout where they have a competitive ad­vantage over rivals, not where they don’t.

Besides working out what they aregood at, banks must decide how much riskthey want to take. Helped along by theratcheting­up of capital charges in tradingbooks and other planned regulatorychanges, a sweeping shift in risk appetite isalready under way. There are obviouslydistinctions between �rms: GoldmanSachs has maintained a stronger bias to­wards risk exposure than Morgan Stanley,for example. But in general proprietaryrisk­taking is being scaled back drastically.Risk capital will reside outside the bankingsystem, in hedge funds and private­equity�rms, much more than before.

The likes of Deutsche Bank, UBS andCredit Suisse have all unveiled strategies to

cut their proprietary activities in illiquidmarkets and focus on high­volume ��ow�businesses: for example, helping clients tomanage exchange­rate and interest­raterisk. That means leaving some money­making opportunities on the table, a mostunbubble­like thing to do. �We could haveheld on to certain assets and made moneynow but we cannot have this kind of riskirrespective of future potential,� says JosefAckermann, the boss of Deutsche Bank.

Fireproo�ngBanks are also taking measures to ensurethat a poor year in more volatile business­es cannot overwhelm a decent year insteadier ones. And they are reviewing theappropriate mix of earnings between divi­sions, given the capital­intensity and riskpro�le of some activities. The �rewalls be­tween businesses are being forti�ed, too,so that managers have a clearer idea of thestandalone pro�tability of each division.

UBS was especially guilty of underpric­ing its internal funding, letting its invest­ment bank take advantage of the bank’scheap overall cost of funds without payingan appropriate premium for the risks itwas taking. The Swiss bank has reorgan­ised itself to ensure that businesses aremore autonomous and are funded at mar­ket rates. Such changes arguably havemore impact than any regulatory reforms.�The real revolution will be within thebusinesses,� says Charles Roxburgh of

McKinsey, �as managers see real detail onwho is making money and how.�

The mechanics of risk management arealso in upheaval. Articulating how muchrisk to take or deciding how much tocharge internally for a certain activity isless clear now that many banks’ risk mod­els have proved unreliable. (The impres­sion of additional uncertainty is itselfpartly illusory: the clarity models provid­ed during the bubble was misleading.)

In truth, the crisis will make modelsmore useful. They will be using data from awhole economic cycle rather than lookingmyopically at a period of exceptionallyhigh returns. The improved risk pro�le ofbanks’ borrowers also means they willhave better data to work with. Method­ological improvements will capture the re­lationships between institutions�the ef­fect on its peers of Lehman Brothers goingbust, say�as well as their independent riskpro�les, which are commonly assessed bya measure called �value at risk� (VAR). To­bias Adrian of the Federal Reserve Bank ofNew York and Markus Brunnermeier ofPrinceton University have proposed ameasure called CoVAR, or �conditionalvalue at risk�, which tries to capture therisk of loss in a portfolio due to other insti­tutions being in trouble. Taking account ofsuch spillover e�ects greatly increasessome banks’ value at risk (see chart 9).

Despite such improvements, risk man­agers are well aware of the need to beef uptheir qualitative controls too. Stress tests,designed to think through how institu­tions cope with periods of pressure, willbecome more important to boards as theyseek to de�ne institutions’ risk appetite.They will also become more important toshareholders. Bank of New York Mellonhas started to include �gures in its earningsstatements showing what could happen toits capital under various scenarios.

Stress tests will also become more de­manding. Take the assumptions abouthow long liquidity can disappear for. Mea­sures such as VAR seek to capture the ef­fects of a single explosive event within arelatively short period. This crisis, saysKoos Timmermans, chief risk o�cer ofING, a Dutch bank, has been �more likeslow death by torture�. Peter Neu of the

The revolution within

The way banks manage risk�including how they reward managers for taking it�will change greatly

9Even more at risk

Sources: Bank ofEngland; Bloomberg

*The return on the bank’s share price relativeto the risk-free return has a 10% chance of

suffering a fall at least this great

Selected British banks, post-2007 crisis, %*

Banks (names withheld)

1 2 3 4 5 6 7 8 90

5

10

15

20

Value-at-risk (VAR)

VAR conditional onother institutions beingin distress (CoVAR)

The Economist May 16th 2009 A special report on international banking 13

2

1

Boston Consulting Group says stress testsmust also become more �coherent�. Toomany banks de�ned stress events in isola­tion�asking what kind of losses theymight sustain in the event of, say, a 20%stockmarket fall without asking what sortsof changes in the economic climate wouldprompt a fall that big.

Even Goldman Sachs, widely regardedas the best manager of risk in the industry,did not foresee quite how bad things couldget. The bank’s most demanding pre­crisisstress test�known as the �wow�, or worstof the worst, test�took the most negativeevents to have happened in each marketsince 1998 and assumed that they got 30%worse and all happened at the same time.That still wasn’t pessimistic enough.

Banks must revisit their assumptionsabout how e�ective their defences areagainst multiple risks. The crisis will livelong in the collective memory for showingthat all markets can become illiquid and allrisks are correlated, removing many of thebene�ts of diversi�cation. �The fourthquarter of last year was remarkable forshowing how fragile the system has actu­ally turned out to be,� says Wilson Ervin,chief risk o�cer of Credit Suisse.

The inadequacy of speci�c hedges,something known as �basis risk�, alsocame as a shock to many. A corporate bondand a cash­collateralised credit­defaultswap written on the same company oughtto o�set each other�if the company lookslikely to default, the bond will fall and theswap rise. In late 2008 the system­wideevaporation of liquidity meant that bankscould lose money on both.

A degree of calm has returned to themarkets since then, reversing some of thelosses banks su�ered from basis risk. Theamount of counterparty risk in the systemwill be reduced greatly by central clearing­houses for credit­default swaps. But con�­dence in hedges and market liquidity as away of mitigating risk has been badlydamaged. In response, banks will use asimpler set of palliatives. They will takegreater account of their gross as well as netexposures. They will charge more for tak­ing on risk on clients’ behalf. And to the ex­tent that they continue to package and sellsecuritised assets to investors, they will re­duce the amount of inventory they hold.

A game of pay senseAll of these aspects of risk management,from models to hedges, are important. Butanother risk­related question�bankers’pay�has dominated the public debate onthe industry’s failures. Pay has been the

touchstone issue of the �nancial crisis, vil­i�ed both as the incentive that drove bank­ers to take foolish risks as well as the mostinequitable feature of an industry thatmakes obscene pro�ts in the good timesand comes crawling to the taxpayer whenit gets into trouble. From the bonuses paidto executives at AIG, a monumentallyfailed insurer, to the expensive tastes ofJohn Thain, a former head of MerrillLynch, and the huge pension granted to SirFred Goodwin, a former boss of RBS, payhas captured the public’s attention, farmore than the banks’ many other failings.

Managers admit privately that thingsgot way out of line. �It was better to be anemployee than a shareholder,� says abank’s chief executive. The traditional ar­gument against changing pay structureshas been that no institution could moveunilaterally without competitors poachingits best people. Now, no bank can fail to al­ter its compensation policy without hav­ing its executives publicly humiliated bypoliticians and the news media, andfrowned upon by regulators.

The broad thrust of the coming changeson pay is clear. Banks will tie compensa­tion more closely to performance andspread rewards over longer periods. Itshould be said that neither idea is foreignto the industry. Bonus pools based on pro­�ts (though not revenues, an indefensiblepractice) may be seen as a problem nowbut are clearly more closely tied to perfor­mance than a �xed base salary. Awards ofshares were common within the industrybefore the crisis and caused employees,those of Lehman Brothers included, to suf­fer vast losses when share prices dropped.What the industry as a whole did not dowell enough was to design pay so that itbetter re�ected long­term risk.

According to a survey of industry prac­tices published by the Institute of Interna­tional Finance (IIF) in March, many banksstill fail to use risk­adjusted measures ei­

ther to calculate the size of their bonuspool or to allocate it. That will change (seechart 10). Economic­capital models, whichcalculate the use of capital based on as­sumptions about expected losses, will bemore widely used to set bankers’ pay in fu­ture. The bonus/malus structure intro­duced by UBS in 2008, whereby a cash por­tion of a bonus award is held back at theend of a �nancial year and reduced if tar­gets are not met in subsequent years, willalso become more common as institutionsseek to track and reward the performanceof senior managers over time.

Some banks will be more sophisticatedstill. With costs and capital under so muchpressure, the incentive for executives toidentify those who add genuine value to abank has rocketed. A few banks already tryto adjust, when calculating bonuses, forfranchise value�the advantage derived byemployees from the bank’s brand value,league­table positions and other institu­tional strengths. An industry veteran saysthat more managers of big banks willcome to realise that they do not need topay twice over for the same bit of business,�rst by building a global infrastructure andthen by rewarding an investment banker.�They would get one in �ve calls for bigprojects anyway,� he says.

Other ideas in the vanguard of design­ing pay structures include �S­curves�,which pay less below a certain thresholdof pro�t so as not to reward employees formarket conditions and franchise value, butalso pay out less above a certain threshold,to discourage excessive risk­taking. Thesetypes of thinking are likely to becomemore prevalent.

Many of these changes are welcome,with two caveats. First, no system can befoolproof. Risk­adjusted measures of com­pensation work only if risk is being mea­sured properly, for example, and the indus­try has proved how unsafe an assumptionthat is. And attempts to control pay in onearea tend to in�ate it in another. As bonus­es fall, pressure on banks to increase basicpay is already rising. That pressure willgrow as the industry recovers and compe­tition for the best sta� increases. �At somepoint in the next few years, the industry isgoing to have an absolutely stellar year,�says a pay consultant who predicts that�rms with clawback policies will have too�er more in upfront pay to attract recruits.The second caveat is that some employeesreally are worth lots of money. Asked todefend levels of pay prior to the crisis,many in the industry would reach for theanalogy of �lm or sport, two other indus­

10Risk-free returns

Source: Institute of International Finance, survey

Banks’ use of risk adjustment in bonus calculations% of total respondents

0 10 20 30 40

Not used, plansto implement

Used in generating andallocating bonus pool

Used in generating orallocating bonus pool

Not used

14 A special report on international banking The Economist May 16th 2009

2

1

tries where talented individuals are criticalto success and are richly rewarded as a re­sult. The trouble with this defence is that itwas not just the big­name stars who gotreally rich in �nancial services; the extrasdid too. Lower pro�ts and more sensitivepay structures will mean that most jobs arerepriced across the industry but the bestpeople will still be the subject of frenziedcompetition and will still command hugesums. That may be distasteful to many out­siders but if pay structures better re�ect in­formation about the risks such star bank­ers are taking and if their pay levels do notin�ate the compensation of everyonearound them, it ought to be defended.

The biggest upheavals in pay and in riskmanagement will be in wholesale bank­ing. The assumptions that underpin theway retail banks manage risks and payhave withstood the crisis better. There arestill lessons to be learned, of course. One

result, for example, will be that lenders de­mand more data on customers, leadingborrowers to concentrate more of theirbusiness on particular institutions. But thebasics of credit­risk management havebeen reinforced rather than overturned.

There is a problem with this picture,however. Retail banks may have less tochange operationally (their funding pro�leis the obvious exception) yet they still gotinto a ton of trouble. The worst mistakes ofthis crisis were arguably made in relativelysimple areas of retail and commercialbanking�from the concentration of risk inthe corporate­loan book of HBOS to Wa­chovia’s kamikaze acquisition of GoldenWest, a Californian lender stu�ed full ofmortgage­shaped grenades. Complexity isnot much of an excuse here. For manybanks, the crisis re�ects a simpler tale offrenetic asset growth and the inevitableturn of the credit cycle.

And that raises a bigger managementquestion�how institutions can resist thepressure to grow when a boom is in pro­gress. Such pressure comes from all quar­ters: from shareholders who want growth,from analysts who want to see higher re­turns on equity, from sta� who want bo­nuses, from managers who want to keeptheir jobs, and from politicians who wanthigher employment and tax takes. Oneway of getting around this is to operate inmarkets that o�er high growth without re­quiring great risks. �We run a boring busi­ness model in exciting markets,� says MrSands of Standard Chartered, which isheadquartered in London but operates indeveloping countries. �The problem wasthat others were running exciting businessmodels in boring markets.�

Industry bosses agree that saying �no�to opportunity is one of their most impor­tant jobs and among their most di�cult.

IF A bank posts record results during theworst quarter in living memory for �­

nancial markets, it could be a quirk. Whenthe same bank has produced higher­than­average returns on equity compared withits peers for a number of years, it deservesa closer look. And when it has a businessmodel that appears to answer some of themain governance concerns a�icting theindustry, it repays much wider attention.

The bank is Sweden’s Svenska Han­delsbanken, a retail bank with operationsin Scandinavia, Britain and elsewhere.Handelsbanken posted a 39% quarter­on­quarter jump in operating pro�ts in thefourth quarter of 2008. It has gobbled upgreat chunks of market share in depositsand new lending in the past year. Theworst of the economic downturn is yet tocome in Sweden but the bank has goodreason to believe it can navigate stormywaters, since it sailed through the coun­try’s 1990s banking collapse unscathed.

The bank’s managers put its successdown to an extremely decentralised man­agement model, introduced in 1972 after aperiod when Handelsbanken had got intotrouble. Branch managers are the banks’main decision­makers, following what isknown internally as the �church­tower

principle��namely, that you should dobusiness only as far you can see from thelocal church tower. Responsibility for allcredit decisions rests with the branches.No loans can be extended over the headsof branch managers (larger sums also re­quire approval from higher up).

The bank is unimpressed by the idea ofselling loans on to other investors. UlfRiese, the bank’s chief �nancial o�cer,says 30% of credit losses can be traced tothe initial decision to extend credit but70% come from changes in borrowers’ cir­cumstances and the way banks respondto them. Banks need to have deep custom­er relationships to spot and respond tothese changes, he says. If loans do sour,Handelsbanken has no specialist centralworkout team, like those at many otherbanks, to come in and sort out the mess.The job is left to branches, which similarlyhave responsibility for cost management,salary levels and product o�erings. A tierof regional management makes the deci­sions on where to open new branches.

The e�ects of making branches re­sponsible for their own fate run deep. Thebank’s credit culture is consistent through­out the cycle, meaning that it loses marketshare in boom times and wins business in

environments like this one. There are noformal budgets or projections for the yearahead, on the principle that customerneeds, not product targets, should deter­mine growth. Handelsbanken eschewsbonuses too, on the grounds that theywork against long­term relationshipswith customers and employees. If thebank meets its return­on­equity goals,however, a portion of the pro�ts goes intothe bank’s pension scheme, which is itslargest shareholder.

Is Handelsbanken just a Scandinavianoddity or can it teach others something?Its approach works in part because it is se­lective about the types of customers ittakes on. A mass­market bank would �ndit tougher to copy its model and be pro�t­able. Mr Riese reckons that the bank’s ini­tial shift to a decentralised model washelped by the fact that lending growthwas very tightly regulated in Sweden atthat time. Handing full control tobranches would lead to more missteps ina deregulated market. But the bank’s corephilosophy�a focus on customers, notproducts; on pro�tability at the level ofeach operating unit; and on long­term re­lationships, not short­term gains�is clear­ly of its time.

More Swedish lessons for thebanking industryBack at the branch

The Economist May 16th 2009 A special report on international banking 15

2 Those who did sit out some of the boomwere heartily criticised for doing so. EdClark, the boss of Canada’s TD, recalls theheat he got from analysts for exiting thestructured­products business. Ulf Riese ofSvenska Handelsbanken (see box on previ­ous page) remembers the pressure that thebank resisted to join its peers in the Balticlending boom. Mr Timmermans, the riskchief at ING, points to the problem of get­ting out of positions at the right time. �It isrelatively easy to get discipline into theprocess of putting assets on to the books.The problem is when you have held themfor two years and think it may be time too�oad,� he says.

The governance gapThe memory of this most painful of epi­sodes should make it easier for bosses toshake their heads, at least for a few years.Private capital will be more patient andmanagers will be more focused on sustain­able growth rather than short­term returnson equity. Wrong­headed assumptionsabout risk dispersion will be less easilymade. But there is an increasing recogni­tion that the governance of �nancial insti­tutions needs to be reviewed carefully (theBritish authorities have already initiatedjust such an exercise).

One obvious area of scrutiny will bethe quality and composition of bankboards, which were found sorely wantingin many cases. That does not mean that di­rectors should take responsibility for risk

management, a job for bank executives.�Directors do not design aeroplanes forBoeing or make the food for Taco Bell,�says Mr Dimon of JPMorgan Chase.

But it does mean that they can do a bet­ter job of vetting key executive appoint­ments�for example, the rise of ChuckPrince, a lawyer, to head Citigroup and ofAndy Hornby, a youthful former retailer, tolead HBOS should have prompted moresearching questions. It means dedicatingmore time to reviewing the business,which implies a limit to the number of di­rectorships that board members hold. Itmeans separating risk and audit commit­tees. It ought to mean dividing the role ofchairman and chief executive. And itmeans asking more robust questionsaround such things as �key person� risk, inwhich only a few employees really under­stand what is going on in a particular lineof business.

Profound questions are also beingasked about the right model of bank own­ership. Some fondly remember the olddays of private partnerships on Wall Street.But for banks that need lots of money tooperate, that is not an option. �Capital islike heroin,� says an investment banker.�Once you go down the capital­intensiveroute, you cannot go back.� Others pro­mote the merits of mutuals, banks that areowned by their customers. Tony Prestedgeof Nationwide, a British building societythat has come through the crisis relativelywell so far, says that being unlisted, mutu­

als can avoid being obsessed with short­term growth targets and can live with peri­ods of reduced pro�ts. Then again, Nation­wide has spent much of the crisis snappingup other mutuals that have got into trou­ble, so the model is not infallible.

With quality of management beingboth the best defence against bank failureand something that can change with theappointment of a new chief executive or arush of empire­building madness (step for­ward the managers of Bank of Americaand Lloyds TSB), regulators are likely to ad­dress the problem of governance in twodi�erent ways. The �rst will be to cushionthe impact of those bank failures that dooccur by creating better resolution regimesfor large institutions and for non­banks.There are also proposals for banks to buyan option on capital via a kind of disaster­insurance scheme, paying out premiumsto long­term investors in return for dollopsof equity when crisis strikes.

The second direction of policy will beto intervene more forcefully to prevent fail­ures in the �rst place, stepping in whenev­er asset growth accelerates, demanding agreater say in board appointments and ve­toing dodgy acquisitions on the grounds of�nancial stability as well as competitionconcerns. More daring voices are even sug­gesting that there may be a case for an o�­cial presence at board meetings. There is atleast time to get all of these things right. Itwill be a long time until anyone has toworry about the next bubble. 7

16 A special report on international banking The Economist May 16th 2009

1

FUNDING markets are damaged. Bor­rowers have to recover from the biggest

credit bubble in history. Bankers’ reputa­tions are mud. Regulators are not just read­ing riot acts, they are rewriting them. Yetmany industry executives are surprisinglybouncy about the future. Investmentbankers in particular have been soundingbrighter, thanks to a healthy start to theyear. Are banks in denial or do they havegenuine cause for optimism?

The answer is obscured by a couple ofbig unknowns. One is the length anddepth of the recession. A depressing analy­sis by Citigroup looks at what happened tobanks in four previous episodes of ex­treme stress, including the Depression, Ja­pan’s �lost decade� in the 1990s and theSwedish banking crisis of the 1990s. Loanbooks collapsed in all cases (by 50% frompeak to trough in America, 30% in Japanand 25% in Sweden), greatly reducing earn­ings even before credit losses were takeninto account.

Direct comparisons are dangerous.Banks have fewer loans as a percentage oftotal assets nowadays (because they holdmore securities) and they also have thechance to gain business that had been go­ing to the shadow­banking system. But thedynamics that operated in earlier periodsof stress are also present now�falling de­mand, pressure to deleverage to meet newcapital rules and reduce loan­to­deposit ra­tios, and dipping asset values. Europeanbanks look especially leveraged in com­parison with their American counterparts.If things turn out anywhere near as badlyas before, says Simon Samuels of Citi­group, banks’ pre­provision returns have alot further to fall.

Another important unknown is the ex­tent to which globalisation unravels. Thethreat of �nancial nationalism, sparkedinitially by political pressure on lenders tofocus on domestic markets and reinforcedby the likely tightening of rules on liquid­ity and capital for any bank operatingwithin a country’s borders, is arguably thebiggest long­term worry for internationalbanks. (Local banks, by contrast, should�nd it easier to win more business.)

Business volumes are likely to fall inmarkets that have been producing a rising

proportion of revenue at the big banks (seechart 11). Returns will drop if banks have toset aside more capital at the national level,or fund themselves from domestic depos­its. Big customers may take things into theirown hands if the system gets too fragment­ed. �If international banking gets more dif­�cult, multinationals will end up doingthings like cash management themselves,�says Mr Sands of Standard Chartered.

Let us again make some non­apocalyp­tic assumptions: that the business of inter­national banking is less pro�table but sur­vives broadly intact and that the recessionreaches a bottom in the relatively near fu­ture. That still leaves many banks with thetask of �nding a new set of pro�t drivers toreplace the old ones.

The extraordinary returns on equity

that banks enjoyed in recent years (seechart 12) were largely created by leverage,the ability to increase the amount of assetsthey held relative to their equity, and by�asset velocity�, which let banks reuse cap­ital multiple times during the course of ayear as assets were originated and speedi­ly moved o� balance­sheets through se­curitisation. The new emphasis on stabil­ity of capital and funding ensures thatneither source of pro�ts will be readilyavailable to banks in the future. The banks’hope is that they can compensate by in­creasing their unleveraged returns, whichmeans grabbing higher volumes of busi­ness and repricing their products.

They do have some cause for optimism.The structural potential of developingmarkets remains intact. And in maturemarkets, banks’ �nancing and risk­man­agement capabilities are arguably in great­er demand than ever. Lots of companiesstill need to raise capital, for example, asevidenced by the rush of bond issuance inthe �rst two months of the year. The advi­sory business is ticking over too, as wavesof companies seek to restructure debts.

Still hedgingMany expect clients to demand morehedging because of the crisis. �There arecompanies that cannot continue operatingtoday as a result of a failure to hedge,� saysMr Winters at JPMorgan Chase, who alsoreckons that clients will ask for more pre­cise, and therefore expensive, forms of pro­tection given the inadequate performanceof some hedges through recent months. �Ifyou are exposed to real estate in the [Eng­lish] Midlands it is no good being hedgedwith a European property index,� he says.

A heightened awareness of risk will af­fect clients’ relationships with the banksthemselves. Banks are supposed to worryabout borrowers going bust. Now the re­verse is also true. Mergers and acquisitionsmandates often require companies to paybanks a fee even if they are no longer in­volved at the time a deal is done, for in­stance. Some clients now want engage­ment letters for the services of banks tospell out what would happen if the banksfailed in the interim. The bankruptcy ofLehman Brothers gave a harsh lesson to

From great to good

Banks will still make money, just less of it

11Globalisation halted?

Source: Oliver Wyman*Including Japan

†Including eastern Europe

Wholesale bank revenues, % of total

2001 02 03 04 05 06 07 080

20

40

60

80

100

Americas

Europe†, Middle East and Africa

Asia-Pacific*

12Nice while it lasted

Source: Citigroup *Estimate

European banks’ return on equity, %

1995 97 99 2001 03 05 07 08*

0

5

10

15

20

25

The Economist May 16th 2009 A special report on international banking 17

2

1

hedge funds about the dangers of doing allof their borrowing and saving with a sin­gle prime broker. Custody banks are win­ning lots of hedge­fund business as a resultof this. Tri­party collateral management,whereby a third bank acts an intermediarybetween a buyer and seller, is anothergrowth area for custodians. Bank of NewYork Mellon is currently servicing $1.8 tril­lion of tri­party collateral a day, up from$1.2 trillion in 2007.

Trend­watchingChanges in consumer behaviour can alsocreate opportunities for retail banks. Ashift towards saving is one trend to capital­ise on. Retail bankers are already thinkingabout structured savings products that of­fer consumers the chance to start puttingmoney back into shares while protectingtheir principal. Given worries about thestability of the dollar, says David McKay ofRBC, there will also be greater demand forproducts denominated in other currenciessuch as the euro.

More important is the fact that competi­tion has fallen sharply in many markets, ei­ther because banks have disappeared orbecause they are �nancially and politicallyconstrained. The credit environment haschanged from being demand­driven tosupply­constrained, which means thatmarket share is up for grabs and pricingpower has increased markedly. A recent re­port on the future of wholesale bankingfrom Morgan Stanley and Oliver Wymanreckons that bid­o�er spreads have in­creased by anything from 50% to 300%.

�The change in the competitive landscapehas been absolutely brutal but for the win­ners, volumes are up, margins up and mar­ket share up,� says Mr Varley of Barclays.

Survivors of the crisis will also be pro­tected by higher barriers to competition.Regulators are going to be nervier aboutletting new entrants into the �nance indus­try and allowing foreign banks free rein intheir markets. Many of the most importantsources of earnings in the new bankinglandscape, such as cash­management ser­vices and �ow businesses, are gigantic,technology­heavy operations that are di�­cult to replicate. Economies of scale willalso count for more in areas such as depos­it­gathering, risk analysis, cross­selling andwholesale­debt issuance. Although thereis much talk about constraining banks thatare too big to fail, the smallest institutionsare the ones that will su�er most in thischanged environment.

All of these factors help to explain whybanking will continue to be a highly attrac­tive business. But they do not make up forwhat has been lost. Huge swathes of thewholesale industry’s product o�ering (in­cluding some of its most pro�table areas)have disappeared. So have many of itsnewer customers�analysts at MorganStanley reckon that hedge­fund assets fellby around 40% in the second half of 2008alone, and that a further 15­30% of assetswill be redeemed this year. The contribu­tion that prime brokerage, structured creditand private­equity activities made to pro­�ts in wholesale banking rose from ap­proximately 20% in 2000 to around 35% in

2006, according to estimates by Oliver Wy­man. These sources of revenue will noteasily be replaced.

The goal of many retail customers,meanwhile, will be to deleverage. The factthat households, not businesses, have somuch debt to unwind is something thatmarks this episode out from many previ­ous banking crises. According to McKinsey,American consumers have accounted formore than three­quarters of the country’sGDP growth since 2000 and for more thanone­third of worldwide growth in privateconsumption since 1990. Although deleve­raging can also occur through incomegrowth, the immediate response of con­sumers has been to save more, depressingdemand for credit (see chart 13). That is like­ly to continue for the foreseeable future.(The situation in emerging markets is dif­ferent: assets there will probably grow rap­idly again once the economic cycle turns,although the need to reduce loan­to­de­posit ratios will weigh on several easternEuropean markets.)

The ability of retail banks to make mon­ey from those customers who do still needto borrow is also more constrained than itmay appear. The politics of ramping uplending rates to taxpayers is sensitive, tosay the least. As Andy Maguire of the Bos­ton Consulting Group points out, there isalso an adverse­selection problem. Bor­rowers who are applying for credit rightnow are likely to be the ones that are hav­ing trouble getting loans elsewhere. Mov­ing existing customers on to higher­pricedloans prematurely can strain relations.

Nightmare scenarioLow interest rates have steepened theyield curve, the di�erence between short­and long­term rates, but they also makethis a terrible environment for depositmargins, which banks calculate as the dif­

13Hitting the credit limit

Sources: Federal Reserve; Bureau of Economic Analysis

US households’ net new borrowing as % of GDP

1952 60 70 80 90 2000 08

2

0

2

4

6

8

10

12

+

–

18 A special report on international banking The Economist May 16th 2009

2

Previous special reports and a list offorthcoming ones can be found online

Economist.com/specialreports

Future special reportsCountries and regionsTexas July 11th The Arab world July 25thIndonesia September 12th

Business, �nance, economics and ideas Business in America May 30thThe euro area June 13thGlobal greying June 27th

Economist.com/rights

O�er to readersReprints of this special report are available at aprice of £3.50 plus postage and packing. A minimum order of �ve copies is required.

Corporate o�erCustomisation options on corporate orders of 100or more are available. Please contact us to discussyour requirements.

Send all orders to:

The Rights and Syndication Department26 Red Lion SquareLondon WC1R 4HQ

Tel +44 (0)20 7576 8148Fax +44 (0)20 7576 8492e­mail: rights@economist.com

For more information and to order special reportsand reprints online, please visit our website

ference between what they pay for depos­its and what they make by putting them towork in money markets. With interestrates so close to zero, banks are having tocut their lending rates but have no room todrop their deposit rates further. Spreadscompress as a result. �The nightmare sce­nario is a period of extended low interestrates like Japan,� says Mr Clark of TD.

There is another threat to pro�ts. Banksmake money not just from the spreadsthey can command on lending but alsofrom fees. The politicisation of bankingcould easily mean that the fairness of bankfees comes under closer scrutiny. Britain’sO�ce of Fair Trading has already ruledsome bank charges unfair. American law­makers are taking aim at credit­card fees ina proposed law. With voters, ie, consumers,now in charge of the industry, other feessuch as overdraft charges may also fall un­der the spotlight. O�shore banking secrecyis an example of something that did notcause the crisis but has been vigorously tar­geted in its aftermath.

Wealthier clients are also likely to beless inclined to pay fat fees in such busi­nesses as asset management, as fallingmarkets, frauds such as the Bernard Ma­do� scandal and broken promises of abso­lute returns make investors question thevalue they are getting. As the full e�ect ofthe crisis on savings and pensions be­comes clearer, consumer activism is likelyto rise.

A glistering era endsAdd to this picture the drag of continuinglosses from toxic assets and souring loans,and it is clear that as an industry, banks aregoing to �nd it much tougher to make mon­ey than before. Clearly, costs, particularlythose related to pay, will fall as well as rev­enues. But there seems to be broad consen­sus among industry observers that averagereturns on equity through the economiccycle will be in the low­ to mid­teenshenceforth, well down on the 20%­plusachieved before the current crisis.

Another way of looking at the industryis to compare its growth with GDP growth.In emerging markets, the industry shouldstill be able to grow faster than GDP as theuse of �nancial products spreads. In ma­ture markets, with the turbo­boost of lever­age gone and bank balance­sheets still tobe slimmed, a growth rate in line with GDP

is probably as much as can be hoped for.That would still make banking a decentbusiness, comparable to many other in­dustries. And if you look at returns on arisk­adjusted basis, as some converts to the

cause now urge, it may even be a morepro�table one than before. But masters ofthe universe it ain’t.

It is possible to glance at the emerginglandscape of banking and think that notan awful lot is going to change. Aside froma few tweaks to capital here, some tougherrules on liquidity there, and the disappear­ance of a handful of badly­run institu­tions, the same big names dominate the in­dustry. And yes, banks will make lessmoney than before but the industry willstill return decent pro�ts and still pay itspeople well. Their �rst­quarter earningsshowed that they can generate hugeamounts of money in even the most di�­cult times. With so many assets trading atsuch distressed levels, many expect thewholesale side of the industry to recordmassive gains when sentiment properlyturns around.

Regulators themselves wonder wheth­er the measures now being discussed gofar enough. As Mr Borio at the Bank for In­ternational Settlements points out, manyof the ideas around countercyclicality (set­ting aside more capital in good times) andmacroprudential regulation (safeguardingthe stability of the whole banking systemas well as of individual banks) were oven­ready, having been worked on by a coterieof central bankers, academics and regula­tors for a number of years. Calls to disman­tle the biggest institutions and split up uni­versal banks have not got far.

Yet the scale of the change sweepingover banking should not be minimised.Banks will seek to conserve capital, not�nd ways to run it down. They will cuttheir dependence on wholesale funding,and grow more slowly as a result. They

will manage risk, not assume it away. Sta�and lines of businesses will have to showthey add value to a bank, not just increaseits revenues. Regulators will bare theirteeth more, and look away less. And tax­payers, whether explicit owners or implic­it guarantors, will peer at the industry andits leaders with hostility, not admiration.

As dramatic as these changes will be tothose inside the banks, they will be just asstriking for banks’ customers. During thebubble and during the crisis, credit was tid­al. It swept in, buoying everything fromsubprime mortgages to leveraged buy­outs. And then it swept out again, strand­ing everyone from investment­grade com­panies to emerging­market oligarchs. Inthe future, credit will be riverine. It willstream towards more creditworthy bor­rowers. It will follow a more de�nedcourse, constrained by embankments ofcapital, funding and risk management. Its�ow will be more domestic, less global.Above all, it will be scarcer.

Given what has gone before, that mayseem like no bad thing but it will entailcosts. No one knows exactly what the rightbalance of debt and equity is in an econ­omy, but the shrinkage of securitisation inparticular makes it more likely that the pro­cess of deleveraging will overshoot. Cus­tomers, such as new businesses or immi­grants, who lack a credit history but couldwell be terri�c economic bets will �nd ittougher to raise money. Emerging marketsthat need to wean themselves o� cross­border capital will grow more slowly thantheir potential. For borrowers such asthese, the failure of the banks will not bemeasured in periods of a few dramaticmonths. Its legacy will last years. 7