february 25th 2012 - the economistmedia.economist.com/sites/default/files/... · 2 the...

February 25th 2012

S P E C I A L R E P O R T

F I N A N C I A L I N N O V AT I O N

Playing with fire

SRFininnov1.indd 1 14/02/2012 15:09

http://bit.ly/yu6rmG

http://bit.ly/x29s2k

Glossary

Finance has a genius for obscuring simple ideas with technical jargon. Some of the terms used in this special report are briefly explained here

Collateralised-debt obligation: a security whose payments flow to investors in order of their position in the payment hierarchy.

Credit-default swap: a type of insurance policy that pays out when an issuer defaults.

Exchange-traded fund: a basket of securities that tracks an index and is listed on an exchange.

High-frequency trading: an ultra-fast algorithmic trading strategy.

Liquidity swap: a transaction to swap illiquid assets for liquid ones.

Longevity swap: a transaction to lay off the risk of unexpected changes in life expectancy.

Repo: a (typically) short-term funding transaction in which the borrower sells a security for cash and agrees to buy it back at a later date.

Social-impact bond: an instrument to finance social programmes that links investors’ returns to the outcome of those programmes.

Swap-data repository: an entity designed to collect data on derivative transactions and make the information available to regulators.

VIX: a measure of the expected volatility of the S&P 500, known as the “fear index”.

1

Playing with fire

Financial innovation can do a lot of good, says Andrew Palmer. It is

its tendency to excess that must be curbed

FINANCIAL INNOVATION HAS a dreadful image these days. PaulVolcker, a former chairman of America’s Federal Reserve, who emergedfrom the 2007-08 �nancial crisis with his reputation intact, once said thatnone of the �nancial inventions of the past 25 years matches up to theATM. Paul Krugman, a Nobel prize-winning economist-cum-polemicist,has written that it is hard to think of any big recent �nancial break-throughs that have aided society. Joseph Stiglitz, another Nobel laureate,argued in a 2010 online debate hosted by The Economist that most inno-

vation in the run-up to the crisis�was not directed at enhancing theability of the �nancial sector toperform its social functions�.

Most of these critics havemarket-based innovation in theirsights. There is an enormousamount of innovation going onin other areas, such as retail pay-ments, that has the potential tochange the way people carry andspend money. But the debate�and hence this special report�fo-cuses mainly on wholesale pro-ducts and techniques, both be-cause they are less obviouslyuseful than retail innovationsand because they were moreheavily implicated in the �nan-cial crisis: think of those evil cred-it-default swaps (CDSs), collater-alised-debt obligations (CDOs)and so on.

This debate sometimes re-volves around a simple question:

is �nancial innovation good or bad? But quantifying the bene�ts of inno-vation is almost impossible. And like most things, it depends. Are creditcards bad? Or mortgages? Is �nance as a whole? It is true that some instru-ments�for example, highly leveraged ones�are inherently more danger-ous than others. But even innovations that are directed to unimpeach-ably �good� ends often bear substantial resemblances to those that arenow vili�ed.

For a demonstration, look at Peterborough. The cathedral city inEngland’s Cambridgeshire is known for its railway station and an under-achieving football club nicknamed �the Posh�. But it is also the site of a�nancial experiment that its backers hope will have big rami�cations forthe way public services are funded.

Peterborough is where the proceeds of the world’s �rst �social-im-pact bond� are being spent. This instrument is not really a bond at all butbehaves more like equity. In September 2010 an organisation called So-cial Finance raised £5m ($7.8m) from 17 investors, both individuals andcharities. The money is being used to pay for a programme to help pre-vent ex-prisoners in Peterborough from reo�ending. Reconviction ratesamong the prisoners recruited to the scheme will be measured against anational database of prisoners with a similar pro�le, and investors willget payouts from the Ministry of Justice if the Peterborough cohort does

A C K N O W L E D G M E N T S

CONTENT S

As well as the people mentioned inthis special report the author wouldlike to thank the following for theirhelp: Francisco Arcilla, Amar Bhidé,Mike Bodson, Peter Carroll, DominicCasserley, Leo Civitillo, Guy Cough-lan, Onur Erzan, Geert Glas, WillGoetzmann, Chi Hum, Ross Levine,Emmanuel Naim, Jim Overdahl,Robert Pickel, Craig Pirrong, MichaelPoulos, Andrew Reid, Pretty Sagoo,Nick Shellard, Susan Smith, ConradVoldstad, David Weild, the sta� ofthe Observatory for ResponsibleInnovation at the École des Mines deParis and a number of people whowished to remain anonymous.

A list of sources is atEconomist.com/specialreports

An audio interview with the author is atEconomist.com/audiovideo/specialreports

4 How innovation happensThe ferment of �nance

5 Retail investorsThe little guy

6 Exchange-traded fundsFrom vanilla to rocky road

9 High-frequency tradingThe fast and the furious

11 Financial infrastructureOf plumbing and promises

12 Small �rmsOn the side of the angels

13 CollateralSafety �rst

SPECIAL REPORT

FINANCIAL INNOVATION

The Economist February 25th 2012 1

2 The Economist February 25th 2012

FINANCIAL INNOVATIONSPECIAL REPORT

2

1

better than the rest. If all goes well, the �rst payouts will be madein 2013.

The scheme is getting lots of attention, and not just in Brit-ain. A mixture of social and �nancial returns is central to a bur-geoning asset class known as �impact investing�. Linkingpayouts to outcomes is attractive to governments keen to hus-band scarce resources. And if service providers like the peoplerunning the Peterborough prisoner-rehabilitation scheme canget a lump sum up front, they can plan ahead without bearingany �nancial risk. There is talk of introducing social-impactbonds in Australia, Canada and the United States.

Here, surely, is a �nancial innovation that even the indus-try’s critics would agree is worth trying. Yet in fundamental waysan ostensibly �good� instrument like a social-impact bond is notso di�erent from its despised cousins. First, at its root the social-impact bond is about creating a set of cash�ows to suit the needsof the sponsor, the provider and the investor. True, the investorsin the Peterborough scheme may be more willing than the aver-age individual or pension fund to sacri�ce �nancial returns forsocial bene�ts. But as Franklin Allen of the Wharton School atthe University of Pennsylvania and Glenn Yago of the Milken In-stitute, a think-tank, argue in their useful book, �Financing theFuture�, the thread that runs through much wholesale �nancialinnovation is the creation of new capital structures that align theinterests of lots of di�erent parties.

Second, the social-impact bond is based on the concept ofrisk transfer, in this case from the government to �nancial inves-tors who will get paid only if the scheme is successful. Risk trans-fer is also one of the big ideas behind securitisation, the bundlingof the cash�ows from mortgages and other types of debt onlenders’ books into a single security that can be sold to capital-markets investors. The credit-default swap is an even simplerrisk-transfer instrument: you pay someone else an insurancepremium to take on the risk that a borrower will default.

Third, even at this early stage the social-impact bond isgrappling with the di�culties of measurement and standardisa-tion. An obvious example is the need to create de�ned sets ofmeasurements in order to work out what triggers a payout�inthis case, the comparison between the Peterborough prisonersand a control group of other prisoners in a national database.Across �nance, standardisation�around contracts, reporting,performance measures and the like�is what enables buyers andsellers to come together quickly and new markets to take o�.

Neither angels nor demons

For all the similarities, there are two big di�erences be-tween the social-impact bond and other, less lauded �nancial in-struments. The �rst is that the new tool has been designed explic-itly for a social purpose. But ask a pensioner how much moneyhe wants to put into prisoner rehabilitation, and it isn’t likely to

be all that much. Whether protecting a retirement pot or signalling problems

with a government’s debt burden, �nance can be �socially use-ful� (to use a phrase popularised by Adair Turner, the outgoingchairman of Britain’s Financial Services Authority) without be-ing obviously social. Lord Turner himself acknowledged that ina speech he gave in London in 2009: �It is in the nature of marketsthat there are some things which are indirectly socially usefulbut which in the short term will look to the external world likepure speculation.�

Many people point to interest-rate swaps, which are used tobet on and hedge against future changes in interest rates, as an ex-ample of a huge, well-functioning and useful innovation of themodern �nancial era. But there are more contentious examples,too. Even the mention of sovereign credit-default swaps, whicho�er insurance against a government default, makes many Euro-peans choke. There are some speci�c problems with these in-struments, particularly when banks sell protection on their owngovernments: that means a bank will be hit by losses on its hold-ings of domestic government bonds at the same time as it has topay out on its CDS contracts. But in general a sovereign CDS has auseful signalling function in an area tilted heavily in favour ofgovernments (which do not generally have to post collateral andcan bully domestic buyers into investing).

The second di�erence is that social-impact bonds are still in

0 500 1000 1500 1600 1700 1800 190010002000 5003000

BC AD

c.3000Bankingdevelopedin Mesopotamia

c.640First coins

produced in Lydia

1024First state-backed

paper moneyintroduced in China

1602Establishment

of first stockexchange

in Amsterdam

1774First mutual fundinvented inthe Netherlands

1667First insurancecompanyin London

1780First inflation-linkedbonds issued inMassachusetts

1924First modern

mutual fund founded

1668 Firstcentral bankfoundedin Sweden

Source: “Financing the Future” by Franklin Allen and Glenn Yago, Wharton School Publishing 2010; The Economist

A selective history of financial innovation

1925

1950 1960 1970 1980 1990 2000 2010

1946Birth of

venture-capitalindustry

1967First ATM

machine installedin London

1973Black-Scholesoptions-pricingformula published

1983Grameen Bank, microfinance

pioneer, becomesindependent bank

1994First

credit-defaultswap transaction

1955First leveragedbuy-out deal

1949First hedge fundformed by AlfredWinslow Jones

1970Securitisationof US mortgagesbegins

1989First exchange-traded

fund launchedin Canada

1997First catastrophe-bond transaction

1940

their infancy, whereas other crisis-era innovations were directlyinvolved in a gigantic �nancial crisis. There are questions to an-swer about their culpability. A few products from that period dolook inherently �awed. Only the bravest are prepared to defendthe more exotic mortgage products that sprouted at the height ofAmerica’s housing bubble as lenders found ever more creativeways to bring una�ordable houses within reach. Finance profes-sionals almost blush to recall an instrument called the constant-proportion debt obligation, a 2006 invention of ABN AMRO thatadded leverage when it took losses in order to make up the short-fall. The end of the structured investment vehicle (SIV), an o�-balance-sheet instrument invented to game capital rules, is notmuch lamented. And the complexity of the �CDO-squared� hasbeen widely condemned.

But even now it is hard to �nd fault with the concept, as op-posed to the practical application, of many of the most demo-nised products. The much-criticised CDO, which pools andtranches income from various securities, is really just a capitalstructure in miniature. Risk-bearing equity tranches take the �rsthit when things go wrong, and more risk-averse investors aremore protected from losses. (Euro-zone leaders like the ideaenough to have copied it with their plans for special-purpose in-vestment vehicles for peripheral countries’ sovereign debt.) Thereal problem with the CDOs that blew up was that they werestu�ed full of subprime loans but treated by banks, ratings agen-cies and investors as though they were gold-plated.

As for securitisation and credit-default swaps, it would beblinkered to argue they have no problems. Securitisation risksgiving banks an incentive to loosen their underwriting stan-dards in the expectation that someone else will pick up thepieces. CDS protection may similarly blunt the incentives forlenders to be careful when they extend credit; and there is a spe-

ci�c problem with the way that the risk in these contracts cansuddenly materialise in the event of a default.

But the basic ideas behind both these two blockbuster in-novations are sound. India, with a far more conservative �nan-cial system than America, allowed its �rst CDS deals to be donein December, recognising that the instrument will help attractcreditors and build its domestic bond market. Similarly, securiti-sation�which worked well for decades�allows banks to free upcapital, enabling them to extend more credit, and helps diversi�-

cation of portfolios as banks shed concentrations of risks and in-vestors buy exposures that suit them. �Securitisation is a goodthing. If everything was on banks’ balance-sheets there wouldn’tbe enough credit,� says a senior American regulator.

Rather than asking whether innovations are born bad, themore useful question is whether there is something that makesthem likely to sour over time.

Greed is bad

There is an easy answer: people. When bubbles froth,greedy folk use innovations inappropriately�to take on expo-sures that they should not, to manufacture risk rather than trans-fer it, to add complexity in order to plump up margins rather thansolve problems. But in those circumstances old-fashioned �-nance goes mad, too: for every securitisation stu�ed with sub-prime loans in America, there was a stinking property loan sit-ting on the balance-sheet of an Irish bank or a Spanish caja.�Du� credit analysis is always the cause of the problem,� says Si-mon Gleeson of Cli�ord Chance, a law �rm.

This argument has a lot of power. When greed takes hold,�nance in all its forms is undone. Yet blaming the worst out-comes of �nancial innovation on human frailty is hardly help-ful. This special report will point to the features of �nancial inno-vations that can turn them into troublemakers over time andshow how these can be managed better.

In simple terms, �nance lacks an �o�� button. First, the in-dustry has a habit of experimenting ceaselessly as it seeks tobuild on existing techniques and products to create new ones(what Robert Merton, an economist, termed the �innovation spi-ral�). Innovations in �nance�unlike, say, a drug that has gonethrough a rigorous approval process before coming to market�are continually mutating. Second, there is a strong desire to stan-

dardise products so that marketscan deepen, which often acceler-ates the rate of adoption beyondthe capacity of the back o�ceand the regulators to keep up.

As innovations becomemore and more successful, theystart to become systemically sig-ni�cant. In �nance, that is auto-

matically worrying, because the consequences of any failurecan ripple so widely and unpredictably. In a 2011 paper for theNational Bureau of Economic Research, Josh Lerner of HarvardBusiness School and Peter Tufano of Said Business School alsoargue that in a typical �S-curve� pattern, in which the earliestadopters of an innovation are the most knowledgeable, a widelyadopted product is more likely to have lots of users with an inad-equate grasp of the product’s risks. And that can be a big problemwhen things turn out to be less safe than expected. 7

When bubbles froth, innovations are usedinappropriately�to take on exposures thatshould not have been, to manufacture riskrather than transfer it, to add complexity


SPECIAL REPORTFINANCIAL INNOVATION

2


1

o�oad a particular risk and there is no generic solution to hand,it is the strats’ job to use derivatives to create one. �That’s whatwe have to do when clients call,� says Mr Chavez. �We can’t tellthem ‘no thanks’.�

At times like this the calls come thick and fast. When yieldsare low and uncertainty is high, there is strong demand for in-vestment products that are structured to take more risk but cappotential losses. Take de�ned-contribution pension plans,where individuals rather than companies now bear the risk thattheir pension pot will fall short of their retirement needs. Theycan take out an annuity, but that is unappealing when interestrates are ultra-low. So investment banks are seeing more appetitefor derivatives-based solutions that will guarantee minimumpayments and allow for a capped amount of gains.

Hedging is another big area of interest for clients. �Beforethe crisis, hedging was viewed like a tax,� says Stéphane Matta-tia, Société Générale’s global head of equity �ow engineeringand advisory. �Now it is becoming an art, just as important as as-set allocation.� Mr Mattatia recalls a client who approached thebank in April 2010 nursing some prescient worries about FrenchCDSs. The bank constructed a hedge based on the euro fallingand gold rising, a scenario that the �nancial models discountedbecause gold tended to be negatively correlated with the dollarand positively correlated with the euro. That lowered the cost ofthe hedge and raised returns to the investor.

SocGen has also constructed an o�-the-shelf answer to therisk of a market meltdown: an exchange-traded fund for institu-

tional investors which is based on theVIX, a measure of stockmarket volatility(see chart 1, next page). The fund invests inVIX futures contracts, shifting fromcheaper long-dated ones into priciershort-dated ones when the VIX movingaverage reaches a certain threshold. Theidea is to keep the fund’s overall costsdown but allow investors to bene�t fromsudden spikes in volatility.

Managing the lightbulb moment

Such irrepressible inventiveness de-mands careful management. Julie Wink-ler of the Chicago Mercantile Exchange(CME), which launches over 400 new de-rivatives products a year, outlines a three-stage process for innovation: investiga-tion, creation and validation. The investi-gation phase establishes whether there isa market need for a product. In the cre-ation phase contract speci�cations are de-veloped and the product is tested exter-nally on a small group of early adopters.The validation phase involves feedbackfrom a wider group of people within theCME, such as sales and compliance.

If someone at an investment bankhas a bright idea for a new product, it typi-cally goes to a new-business committee,where product controllers look at issueslike pricing and valuation, risk managersconsider how it can be hedged, and legaland compliance folk worry about the �neprint. If a product seems to have the po-tential to tarnish a bank’s brand, it usuallygets passed to a reputational-risk commit-tee. At many banks these processes have

How innovation happens

The ferment of finance

Moving from ideas to products to markets


WHAT WITH THE world’s economic woes and the indus-try’s regulatory overhaul, this may not seem like the most

propitious time for �nancial innovation. Yet the current environ-ment is perfect for stimulating �nance’s creative juices. Regula-tion has always sparked bouts of inventiveness as practitionersseek both to get around new rules (as they did with the introduc-tion of the SIV in the late 1980s) and to bene�t from them (for ex-ample, rule changes in the 1970s that allowed pension funds toput more money into high-risk assets).

Clients have big problems to solve, too. Tempting though itis to imagine bankers cooking up wild schemes in their WallStreet lairs, innovation is often triggered by a client coming to abank with a speci�c headache. Sometimes those headacheshave a ready-made cure, but often they need bespoke treatment.Martin Chavez is co-head of a team of �strats� (�nancial engi-neers) at Goldman Sachs which acts as Goldman’s in-house re-search-and-development division. If a client wants to take on or



2

1

been tightened up a lot since the crisis. �The new product-ap-proval process�is extremely conservative, indeed too conserva-tive,� says a London-based banker.

The crisis has highlighted one speci�c area of di�culty:judging the sophistication of a client (see box). At least one bigbank now steers clear of selling derivatives to municipalities be-cause it feels it cannot be sure that bureaucrats will understandwhat they are buying. Even supposedly expert investors may notknow what they are getting into. �A German Landesbank should

be treated like a child,� is the withering verdict of one regulator. When an innovation is �rst launched, it is by de�nition not

systemic. It might not work out well for its particular buyers andsellers but it will not blow up the economy. The worry for regu-lators and taxpayers starts when products and processes achievea scale where their �aws can have far-reaching consequences.

Most innovations take time to reach that point. With mem-ories still fresh of a boom when lots of new ideas quickly tookroot, it may seem as though all that is needed to make the money�ow is an investment banker with a brainwave. But for everymarket that booms there are others that sputter.

The classic example is the market for property derivatives.Property is the world’s biggest asset class, obvious ground for in-struments that allow property owners to hedge against pricefalls and to gain exposure to assets that they cannot buy forthemselves. But it is also a very lumpy, illiquid, idiosyncratic as-set, which makes it hard to construct an instrument that matchesthe performance of speci�c properties.

Enthusiasts have answers to this problem of �basis risk�,the risk that a hedge will not precisely match movements in theprice of the underlying asset. Paul Ogden, one of the founders ofInProp Capital, a fund manager that o�ers investors exposure toa British commercial-property index, says that derivativesshould not be used to trade the risk associated with particularbuildings but the �correlated market risk�, which drags prices

1Fear, indexed

Source: Thomson Reuters

2007 08 09 10 11 120

200

400

600

800

1,000

1,200

1,400

1,600

0

10

20

30

40

50

60

70

80

S&P 500

S&P volatility index (VIX)

THE BIGGEST QUESTION for supervisors ishow to protect retail investors from �-nancial wizardry. A lawyer-turned-regulatorin America recalls how the paperwork forauction-rate securities, a form of debt forwhich the interest rates were reset in rollingauctions, made it clear that the instrumentcould be hard to sell if auctions failed.�When people complained, I thought it wasridiculous because this was on page one ofthe document. But it turned out that thesalesmen had told them that this was risk-free, and a liquid alternative to Treasuries.�

In Europe in particular regulators arenow more inclined to intervene in speci�cproducts. Britain’s Financial Services Au-thority (FSA) is setting the pace. In Novem-ber the regulator issued guidance on pro-duct development for retail structuredproducts, a sign of its willingness to getstuck in at the point when instruments are�rst launched.

�Philosophically, the edi�ce of worldregulation was built on two buildingblocks,� says Martin Wheatley, the man whowill head the new Financial Conduct Au-thority when Britain’s supervisory structureis revamped. �One is the idea of the rationalman and the other is that there is a reason-able level of conduct in the sales channel.Both of those building blocks broke down to

some extent. That throws you back to adi�erent model where you try to ensure thatproducts are reasonable and developed withthe interests of the client in mind.�

That sounds unobjectionable in the-ory, even to many practitioners. �We’re notlike Canute, telling product regulation to goaway,� says Timothy Hailes, a lawyer atJPMorgan who chairs the Joint AssociationsCommittee, a grouping of industry associa-tions with an interest in structured pro-ducts. But there is uncertainty over whatthe new approach will mean in practice.

Late last year, for example, the FSA

came as close as it can with its currentpowers to issuing its �rst product ban, onthe marketing of �life-settlement� in-vestments to retail investors. Life-settle-ment products are the sort of thing that getsreputational-risk committees to meet, ifonly because they are also known as �deathbonds�. The idea, well-established in Amer-ica, is that elderly people can sell theirlife-insurance policies to investors who willkeep paying the premiums and collect thepayout when the policyholder dies. Thatmay sound chilling, although the same ideaunderpins the sale of annuities; and using alife-insurance policy to generate cash canbe a boon for some pensioners (for ex-ample, those who are very ill or have no

The little guy

What do small investors need more: choice or protection?

bene�ciaries). The FSA’s concern is muchmore for the investor. It reckons the productis being sold hard on the basis of actuarialassumptions that are too complex for aretail buyer to understand.

The problem is that micromanagingproducts can easily mean restricting choice.Imagine a �nancial instrument that exposesretail investors to �rst loss in the event of aproblem, swings around wildly from day today and delivers virtually no returns toWestern buyers for a decade. Would equitieshave got through the regulators?



2

1

IF STANDARDISATION IS one way to make �nancial inno-vations catch on, mutation is another. By experimenting fu-

riously with new products, the industry can hit on iterations thatwork. And once something has taken hold, a host of incrementalchanges follow as �rms compete for custom. �Any good idea isimmediately copied and propagated like a virus,� says Robert Li-tan of the Kau�man Foundation. �If it’s a bad virus, then youhave a pandemic.�

The industry does little patenting. On the retail side there issome attempt to protect new technologies, but for capital-mar-kets businesses this is not a priority. In a global ranking of �rmsassigned patents in America in 2011, drawn up by IFI CLAIMS

Patent Services, the �rst �nancial �rm in the list was AmericanExpress�in joint 259th place.

One reason may be that in the capital markets ideas requireliquidity to take o�: the more institutions that imitate an instru-ment, the deeper liquidity is likely to be. Another is that copyingis easy. �A �nancial product is about as conceptual as you canget,� says Wilson Ervin, a senior adviser at Credit Suisse. �Youjust need paper and ink.� A patent has to be published after ayear or so, enabling rivals to design around it. In any case, mone-tising an idea immediately opens the door to copying because ofdisclosure requirements under securities laws. Mr Ervin reckonsthat �rms have a window of three to four months before rivalproducts appear. It does not help that clients shop around to seeif they can get the same sort of thing cheaper from another �rm.

There are cultural and technical barriers to using patents,too, according to Karen Hagberg of Morrison and Foerster, a law�rm. The targets for enforcement would be other �nancial insti-tutions and they are hesitant to sue each other. As a result, in thepast courts have been asked to decide relatively few patent-litiga-tion cases in the �nancial-services industry. Without an existingbody of law to provide some insight into how courts would de-cide relevant issues in the future, such as what constitutes an in-fringement and how to calculate damages, it is riskier to call inthe lawyers. Instead, products are left to spawn and mutate at

Exchange-traded funds

From vanilla to rockyroad

The Darwinian evolution of exchange-traded funds

3Here to stay

Source: BlackRock

Exchange-traded funds, assets $trn

0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

2000 01 02 03 04 05 06 07 08 09 10 11

Equity Fixed income Commodity

everywhere up or down. Robert Shiller, a Yale professor who isbehind a set of housing futures traded on the CME, reckons thatadvances in information technology will eventually allow accu-rate local indices to be constructed. But progress has been glacial.

The birth pangs of another market, in longevity risk, alsodemonstrate that getting innovations widely adopted is notstraightforward. The need for a product is clear: people are livingever longer (see chart 2), which creates risk for institutions suchas corporate pension schemes and annuity providers that willhave to provide retirement incomes for longer than expected.The idea is to parcel this longevity risk out to someone who isprepared to bear the cost if it materialises.

The lives of others

Basis risk is again a big stumbling-block. A pension fundwants to transfer its own speci�c risk so that it gets a payout if itsscheme members live longer than expected. But that requireslots of actuarial expertise on the part of the buyer. �Investorswant something tradable, where they don’t have to make actuar-ial judgments and can have standardised products,� says DavidHowell, the chief executive of Paci�c Life Re, a reinsurer. Thatmeans an index measuring rates of mortality improvement thatcan act as a common reference point for buyers.

But there is also a need for a market-maker to bridge the gapbetween these two communities�a reinsurer, insurer or invest-ment bank that has the skills to enter into bespoke transactionswith sellers and is willing to take the basis risk of an index-basedtransaction to push it out to capital-markets investors. It is thesame process of intermediation as exists in the more establishedmarket for catastrophe bonds, issued by insurers and reinsurersand designed to pay out when a natural disaster strikes .

Longevity risk, however, poses an additional problem.Whereas a catastrophe can occur in an instant, longevity risktakes decades to unfold. That creates another mismatch becausebond investors typically want to lock up their money for just �veor ten years, which may be too short to act as an e�ective hedgeagainst people living longer. Alison Martin of Swiss Re, anotherreinsurer, says her �rm is trying to increase the duration of lon-gevity-related instruments. She adds: �The point of innovation isto test di�erent structures and durations and see what works.�

There is something reassuring about these painstaking ef-forts to create a market, but there are risks too. Standardisation isa precondition for growth, but it allows basis risk to build. Dur-ing the �nancial crisis some intermediaries retained too much ofthis residual risk, and others transferred it to buyers that couldnot handle it. �Our business is ultimately the management of ba-sis risk,� says Goldman Sachs’s Mr Chavez. Some do it betterthan others. No one did it well enough ahead of the crisis. 7

2Good news for some

Source: Office for National Statistics *Average of male and female cohorts

British life expectancy at birth*, yearsForecast year:

1966 70 75 80 85 90 95 2000 05 10 15 20 25 30

Actual 70

72

74

76

78

80

82

84

86

1971 1981

1991

20022010



2

1

their own pace, and not always in a healthy way.Few instruments so encapsulate the mutability of �nancial

products as the exchange-traded fund, or ETF. No one has a badword to say about the concept. Invented in the late 1980s by theToronto Stock Exchange, the ETF is in essence a cheap mutualfund: a basket of securities that tracks an index, is wrapped in afund structure and is listed on an exchange so that investors cantrade in and out whenever they want. They are cheap and tax-ef-�cient, and they allow retail investors access to diversi�ed port-folios of assets that had previously been the sole preserve of in-stitutional investors. �If you were inventing the mutual-fundindustry today, it would look like this,� says Salim Ramji, a con-sultant at McKinsey.

Even so, the product did not gain momentum immediately.According to Dan Draper, who runs Credit Suisse’s ETF business,the spark to spectacular growth came only after the dotcom busthad underscored the importance of diversi�cation. The rise offee-based �nancial advice also pushed more investors towardsETFs. In recent years the instrument has built up huge momen-tum (see chart 3, previous page), and the 2007-08 �nancial crisisdid nothing to slow it. The number of ETFs on the market inAmerica has mushroomed from 343 in 2006 to 1,098 in Decem-ber 2011. A recent McKinsey report forecasts that global assets un-der management in exchange-traded products, a broader uni-verse of listed portfolio-tracking products, could grow fromabout $1.5 trillion in 2010 to $3.1 trillion-$4.7 trillion in 2015.

Finance’s infectious creativity is again on full display. Youcan buy ETFs that are sharia-compliant, that invest in clean-ener-

gy stocks, that focus on emerging-marketlocal-government debt, and many, manymore. �We’ll stop creating ETFs when youstop having ideas,� boasts the website ofiShares, BlackRock’s ETF arm. Other ex-change-traded products allow investorsto gain exposure to commodities, fromgold to palladium. In August 2011the SPDR

Gold Trust brie�y wrestled the crown forthe world’s largest ETF from a fund track-ing the S&P 500.

Such vibrancy looks like a victoryfor the investor over the fund manager.Yet ETFs have lost some of their lustre re-cently as the debate about their potentialrisks has become more vocal. That debatestarted in spring 2011 as both the IMF andthe Financial Stability Board (FSB), a glo-bal watchdog, voiced concerns about theinstrument’s hidden trapdoors. It was giv-en further impetus by �erce disagree-ments within the industry about whatkinds of products should bear the label�ETF�. And it demonstrates just how hardit is to control the development of a �nan-cial innovation.

Roughly speaking, there are twoworries about ETFs. One is that they havebecome remarkably successful, and theother is that they are opaque. Their suc-cess is something that regulators fretabout more than the industry does: in �-nance, anything systemic is a potentialthreat. The absolute size of the ETF marketis relatively modest compared with esti-mated global assets under the control offund managers of $100 trillion or more.

But the institutions at its heart�in particular, investment bankssuch as Deutsche Bank and Société Générale�are huge, whichmakes people worry about how an ETF crisis might play outmore broadly. And the rapid take-up of ETFs in itself is enough tocause concern. �If something is growing fast over a period of sev-eral years and attracting a broader set of players and new en-trants, that is an alarm bell,� says Nigel Jenkinson of the FSB.

Inscrutable

If ETFs had remained the �plain vanilla� products theywere in their earliest incarnations, the regulators would havefound little to get alarmed about. What troubles them is the in-strument’s opacity. The industry is still dominated by physicalproducts, with fund managers going out and buying each indi-vidual constituent of the index they are tracking. But there is aslice of the ETF market, making up a little over 10% of global as-sets under management and concentrated in Europe, that is�synthetic�. This means that the returns generated by the fundcome from a swap with a counterparty, often an a�liate of thefund manager. Instead of using investors’ cash to buy the under-lying securities, the money is put into a basket of collateralwhose returns are swapped with a counterparty for the returnsof the index being targeted.

That spooks plenty of people because it exposes investorsto counterparty risk. If the swap counterparty defaults, thatleaves the investor holding the contents of the collateral basketas security. This collateral may not bear even a passing resem-blance to the assets that the ETF is ostensibly tracking. And some



2 regulators worry that banks purposely choose synthetic struc-tures so that they can dump their illiquid assets into the collateralbasket and get funding that they otherwise could not.

The fear that some investors may not understand whatthey are getting themselves into extends to other products. Ex-change-traded notes, or ETNs, may sound like ETFs, but they areessentially unsecured debt instruments issued by banks. Anoth-er set of ETFs that are souped up by leverage and seek to makedaily returns are particularly exposed to an e�ect called �com-pounding�. Imagine a 10% rise on a $100 investment on day one,followed by a 10% fall on day two: the value of the investmentwill end up being $99, not $100, as many people intuitively as-sume. Now add in leverage designed to double the movementsof the investment, so each day sees a 20% swing: the result willbe an investment worth just $96. This compounding e�ect is alsoat work in �inverse� ETFs, which are designed to make moneywhen markets fall. The risks inherent in leveraged and inverseproducts manifest themselves most in times of volatility.

In October BlackRock put itself on the side of the angels byissuing a paper calling for better disclosure around derivatives-backed products, so that investors are clear about the identity ofcounterparties, the composition of collateral and so on, and forthe use of multiple counterparties rather than a single swap deal-er. It also outlined proposals for a formal classi�cation of ex-change-traded products so that only some instruments can becalled ETFs. �I fear that an exchange-traded product will break down one ofthese days and the worry is that it will poi-son the entire sector,� says Mark Wied-man, the head of iShares.

BlackRock’s boss, Larry Fink, hassounded warnings based on his own ex-perience helping to pioneer mortgage-backed securities. �I do believe we havesome responsibility for making sure thatthe market does not morph itself, the sameway when I started in the mortgage mar-ket 35 years ago, watching a great marketmorph into a monster,� he told a confer-ence in November.

Rivals claim that BlackRock’s ap-proach is self-serving: it is one of threedominant providers in America and o�ers a range of productsmade up almost exclusively of physical ETFs. Providers of syn-thetic ETFs argue that there is plenty of counterparty risk in phys-ical funds, too, because the funds’ securities are routinely beinglent out to other investors in return for collateral. Derivativeshave long been a feature of the ETF market in Europe. They are al-lowed by the EU’s UCITS fund-management directives, whichmeans that the synthetic products are tightly regulated. For ex-ample, the rules require the overnight market value of the collat-eral to be at least 90% of the value of the securities. And althoughmuch of the debate focuses on the retail investor, this product isheavily used by institutional investors too.

There are also perfectly good reasons to use derivativeswhen it is hard to run a fund as a physical structure, perhaps be-cause of restrictions on investing directly in an asset such as com-modities. �The row over synthetic versus physical ETFs is nothelping regulators to have a clear view,� says Patrice Berge-Vin-cent, a French regulator who is preparing a report on the productfor IOSCO, a global body of securities supervisors. �It dependson each individual ETF as to how dangerous things are.�

Providers of leveraged products, meanwhile, point out thatleveraged mutual funds have existed for more than 15 years.ProShares, a provider of leveraged funds (or exchange-traded in-

struments, in BlackRock’s proposed taxonomy), is irreproach-ably clear about the risks of compounding. Michael Sapir, itsboss, believes that his customers are knowledgeable and use theproducts appropriately to manage risk or pursue investment op-portunities.

That said, it is di�cult to argue against a call for more tran-sparency or against the model of innovation governance exem-pli�ed by the debate on ETFs. Regulators congratulate them-selves on having made the industry more introspective andpoint out that disclosure practices have improved immeasurablysince they started waving red �ags in the spring of last year. Mostpeople in the industry seem to think that the end result of this de-bate will be some kind of product classi�cation. Among custom-ers the wind seems already to be blowing in favour of physicalproducts. �Client demand is changing,� says Mr Draper at CreditSuisse, which converted four swaps-based funds to physicalones last year.

Perpetuum mobile

It is in the nature of �nance that experimentation neverstops, however. So it is with ETFs. The pressure to innovate willintensify as competition increases. The McKinsey report reckonsthat the number of ETF managers in America has grown tenfoldin the past decade. That guarantees the industry will keep push-ing forward with new products.

Mr Wiedman believes that there is still lots of room forgrowth in physical ETFs, not just in equities but in �xed income,too. Products could also move into ever more exotic areas in or-der to deliver higher returns, which may yet shove the pendu-lum back in the other direction and require the use of more de-rivatives to replicate the desired exposures. Intriguingly, there isplenty of talk of active ETFs that would combine elements of dis-cretionary stock selection and the tracking of a benchmark. Noone is sure how this would work, not least because it would re-quire managers to reveal their strategies to marketmakers, but

this is one of those rare areas where patents have been �led.All of these possibilities will require continuing vigilance

on the part of �nancial watchdogs. Like the Red Queen in AliceThrough the Looking-Glass, the regulators will always have tokeep running just to stand still. The FSB paper points out thatbranching out from equities into other asset classes means mov-ing into markets where liquidity is thinner, for example. If an ETF

is active, presumably sometimes investors do not know what isin their portfolio as managers make discretionary bets: that hard-ly sounds transparent. Fiercer competition will also encourageproviders to make more money from lending securities, whichmeans that even investors in physical products could end up ex-posed to rising levels of counterparty risk.

And all forms of growth will increase the weight of ETFs indetermining stock prices, a prospect that worries people like MrLitan. He argues that less liquid, smaller stocks already get bu�et-ed by wider movements in index-tracking ETFs of which theyare constituents, and that buying and selling bundles of stocksleads to excessive correlations between them. Whatever themerits of this argument, it opens the door to others: about the ef-�ciency of modern markets, the consequences of passive invest-ing, and in particular the role of the most turbocharged �nancialinnovation of all: high-frequency trading. 7

Like the Red Queen in Alice Through the Looking-Glass,the regulators will always have to keep running just tostand still

High-frequency trading

The fast and thefurious

High-frequency trading seems scary, but what does

the evidence show?

ON FEBRUARY 3RD 2010, at 1.26.28 pm, an automated trad-ing system operated by a high-frequency trader (HFT)

called In�nium Capital Management malfunctioned. Over thenext three seconds it entered 6,767 individual orders to buy lightsweet crude oil futures on the New York Mercantile Exchange(NYMEX), which is run by the Chicago Mercantile Exchange(CME). Enough of those orders were �lled to send the market jolt-ing upwards.

A NYMEX business-conduct panel investigated what hap-pened that day. In November 2011it published a list of In�nium’salleged risk-management failures and �ned the �rm $350,000.In�nium itself neither admits nor denies any violation of the ex-change’s rules. It takes the same line on a $500,000 �ne it wasgiven at the same time for alleged transgressions on the CME it-self in 2009.

Those alleged failures pull back the curtain on some of thesafeguards that are meant to protect traders, exchanges and mar-kets from erratic ultra-fast algorithms. The NYMEX panel foundthat In�nium had �nished writing the algorithm only the daybefore it introduced it to the market, and had tested it for only acouple of hours in a simulated trading environment to see how itwould perform. The �rm’s normal testing processes take six toeight weeks. When the algorithm started its frenetic buyingspree, the measures designed to shut it down automatically didnot work. One was supposed to turn the system o� if a maxi-mum order size was breached, but because the machine wasplacing lots of small orders rather than a single big one the shut-down was not triggered. The other measure was meant to pre-vent In�nium from selling or buying more than a certain num-

ber of contracts, but because of an error in the way the rogue al-gorithm had been written, this, too, failed to spot a problem. Tocomplete the catalogue of errors, the �rm then allegedlybreached another CME rule when an employee used a col-league’s trading ID to put on positions that would o�set its un-wanted exposures.

This incident was unusual for ending in a �ne, but in otherrespects it was not that uncommon. The ��ash crash� of May 6th2010, when American equity markets nosedived by almost 10%in the course of a few nerve-shredding minutes, grabbed popu-lar attention. Although it was not directly triggered by high-fre-quency traders, the o�cial reports suggested they helped fuelthe uncontrolled selling. But there are miniature versions of such�ash crashes happening all the time, says John Bates, the chieftechnology o�cer of Progress Software, a software �rm.

Often they result from algorithms interacting with eachother and forming an in�nite loop. For a simpli�ed example, taketwo algorithms that are both programmed always to outbid thebest o�er in the market, so they will go on outbidding each other.Usually such loops are spotted and stopped, sometimes man-ually and sometimes automatically, without many people notic-ing. But the fact that they happen at all feeds the perception thattoday’s equity markets have turned into something more akin toscience �ction than a device for the e�cient allocation of capital.How, the critics ask, can anyone assess the fundamentals of acompany in a fraction of a second? How can lumbering institu-tional investors, let alone the little guy, hope to compete with theHFTs? And what is to stop a new set of algorithms from unleash-ing havoc?

Science v friction

Such questions have gradually drawn the high-frequencytraders out into the open. Until recently they saw little need toengage with the wider world. HFTs do not have clients but oper-ate with their own capital. Proprietary algorithms provide acompetitive edge, so secrecy is engrained in the culture. But asregulators, politicians and the media focus ever more closely ontheir activities, the traders have formed groups on both sides ofthe Atlantic to represent their interests.

Many are frustrated by what they perceive as an unfair on-



1



2

1

slaught. �The gap between reality and rhetoric is bigger for thisindustry than for any other I have seen,� says Remco Lenterman,the chairman of the European Principal Traders Association andthe managing director of IMC, an electronic marketmaking �rm.Plenty of academics support the HFTs’ arguments. On the otherside are some big institutional investors who accuse HFTs offront-running their orders, and high-pro�le critics like Bart Chil-ton, a member of America’s Commodities Futures Trading Com-mission, who has punningly dubbed HFTs �cheetah traders�.

To sift through the arguments on both sides is to confront abasic problem with any �nancial innova-tion: the di�culty of measuring its bene-�ts. For one thing, there are questions ofde�nition. HFTs are not the only peopleusing algorithms to trade: institutional in-vestors use them to break large orders intosmall parcels so that markets do not moveagainst them as they execute the order. And although high-fre-quency trading always involves very brief holding periods andvery active trading, it breaks down into lots of di�erent strat-egies. Some HFTs are momentum traders, riding the wave of aparticular trend. Others arbitrage price di�erences. Others stillare marketmakers providing liquidity to buyers and sellers.

Another problem is that there are not enough good data.The �ercest debates centre on the role of HFTs as marketmakers.The evidence tends to favour the HFTs, which can point to a solidbody of academic research that shows they increase liquidity, asmeasured by tighter bid-ask spreads (see chart 4). HFTs also pointto testimony delivered to the Securities and Exchange Commis-sion in 2010 by George Sauter of Vanguard, a big fund manager,who concluded that �high-frequency traders provide liquidityand ‘knit’ together our increasingly fragmented marketplace, re-sulting in tighter spreads that bene�t all investors.�

But others say that the increase in trading activity broughtabout by HFTs, in Europe at least, means that fund managershave to pay additional costs for storage and electronic reportingin order to comply with best-execution requirements. It is hard todisentangle the e�ects of HFTs on transaction costs from otherchanges, such as competition among exchanges. A bigger pro-blem, says Paul Squires, the head of trading for AXA InvestmentManagers, is that increased liquidity can be illusory. �You canpress the button to buy Vodafone, say, and have it executed in asecond but in that period 75% of the liquidity has disappearedand the price has moved.�

It is certainly true that HFTs are constantly sending and can-celling orders. Some of that activity may be tied to a manipula-tive technique called �quote-stu�ng�, in which a �ood of ordersand cancellations causes congestion on networks and thereby a

�eeting trading advantage. But the legitimate explanation for it isthat marketmakers cannot a�ord to be static in case the marketmoves against them, and that in an ever-faster market HFTs haveto be quicker to adjust prices.

�We have got to get away from the idea that speed equalsdanger,� says Richard Gorelick, the �G� in a Texan HFT calledRGM. �Professional traders trade continuously and are exposedto market movements all the time, so being able to adjust that ex-posure quickly gives them con�dence to quote better prices.�That is why the idea of imposing minimum resting times forquotes before they can be cancelled would almost certainly leadHFTs to widen bid-ask spreads, increasing costs to investors.

In their 2011 NBER paper Mssrs Lerner and Tufano arguedthat it is virtually impossible to quantify the social impact of a �-nancial innovation because �nance involves so many external-ities, often unintended ones. For example, it would be almost im-possible to measure the aggregate costs and bene�ts of afundamental innovation like a bank. Instead, they reckoned, athought experiment�imagining what the world would look likewithout a particular innovation�might help.

A world without HFTs is both easy and very di�cult toimagine. Easy, because the old world of specialist marketmakersand �oor trading existed only a few years ago, so people remem-ber it well. There is little obvious enthusiasm for returning to thatmodel. Not only were transaction costs higher but the same ar-guments about unfair advantages were being put forward in dif-ferent forms. Now the complaints are about the milliseconds

HFTs gain over ordinary investors by putting their servers rightnext to the exchanges’ data centres; then they were about the mo-nopolistic privileges of the specialists and the advantages of be-ing on the �oor. Institutional investors may complain about be-ing forced into �dark pools� (o�-exchange venues where theycan deal anonymously) to avoid HFTs, but these pools existedbefore HFTs and were set up in part to avoid being scalped bybrokers or �oor traders.

But imagining a world without high-frequency trading isalso hard. That is because the HFTs are what Larry Tabb of TABB

Group, a research �rm, describes as an �outcrop� of the marketstructure. They are a natural outcome in a world in which trading

4Invest, don’t protest

Source: Knight Capital Group

S&P 500 median bid-ask spread, $ cents

2003 04 05 06 07 08 09 10 110

0.5

1.0

1.5

2.0

2.5

3.0

05 07 09 11 13

5Rise of the machines

Source: Aite Group *Estimate

Algorithmic trading, % of total trading

F’CAST

0

10

20

30

40

50

60

70

0

10

20

30

40

50

60

70

2004 06 08 10 * 12 14

Equities

Futures

Foreignexchange

Options

FixedIncome

2004 06 08 10

US

Europe Asia

To sift through the arguments on both sides is toconfront a basic problem with any �nancial innovation:the di� culty of measuring its bene�ts



2

1

FAILURE IN THE �nancial crisis had many fathers. Therewere failures of regulators, bankers, shareholders, borrow-

ers and economists. But two in particular were closely tied to theway �nancial innovations work. One was a failure of plumb-ing�the infrastructure of the markets and the back o�ces of �-nancial �rms. The other was a failure of the imagination.

Infrastructure often lags when an innovation takes o�. Re-member how markets move over time from being customised tobecoming more standardised. When products are standardisedand demand is high, �nance’s manufacturing, sales and distribu-tion channels can pump out a vast supply. The frighteninggrowth of over-the-counter derivatives markets attests to that(see chart 6).

The boring details�of trade con�rmations and settlements,of collateral requests and contractual negotiation�tend to getleft behind in the rush to make money. Lawyers recall howboom-time contracts for commercial mortgage-backed securi-ties would be hurried through, often by junior associates withlittle knowledge of the product. The far bigger market for resi-dential mortgage-backed securities was even sloppier at record-keeping, causing uncertainty about which lenders had formal ti-tle to properties when the housing bubble burst.

�The back o�ce is attached to the front o�ce by a bungeecord, and depending on how fast the front o�ce is running, thecord gets stretched,� says Mark Beeston, who runs a portfolio ofpost-trade businesses for ICAP, an interdealer broker. Thesebusinesses include Traiana, the post-trade risk manager for HFTsencountered in the previous section, as well as a number of�rms devoted to a market that truly took o� before the crisis: thatfor credit-default swaps (CDS). Indeed, many people think that aregulatory intervention to try to sort out a horrible paperworkmess among CDS buyers and sellers prevented the crisis frombecoming even worse.

By 2005 the infrastructure of the CDS market wasswamped. Harmonisation of CDS contracts meant that, in thejargon, they were easy to �novate�, or transfer to another party.Trading volumes soared. A particular worry was an enormous

Financial infrastructure

Of plumbing andpromises

The back o�ce moves centre stage

6All you can eat

Source: Bank for International Settlements *Credit-default swap †Over the counter

Derivatives, notional amounts outstanding, $trn

1998 2000 02 04 06 08 10 110

200

400

600

800

Exchange-traded optionsExchange-traded futures

Foreign exchangeInterest rateCDS*Other OTC

OTC†

is automated, and in which there is competition between lots ofdi�erent exchanges and a need for someone speedily to knit to-gether the prices they o�er. �The real question is whether hu-mans make worse mistakes when they write algorithms orwhen they trade,� says Terrence Hendershott of the Haas Schoolof Business at the University of California, Berkeley.

In practical terms, trading history is highly unlikely to be re-versed. Regulators in developed countries have no evidence thatradical change is needed, nor any appetite for it. Developingcountries, which have become the standard-bearers of sensible�nancial regulation, are racing towards high-frequency tradingas they seek to improve liquidity (see chart 5, previous page).

Meanwhile the industry itself pushes inexorably forward.That certainly entails greater speed: the industry used to think interms of milliseconds (it takes you 300-400 of these to blink) butis now fast moving to microseconds, or millionths of a second. Italso means smarter algorithms. People have gone from tradingin open-outcry pits to trading via screens to programming algo-rithms. The next stage, says Mr Bates of Progress Software, will beself-learning systems, in which sentient algorithms mine thecapital markets, spotting correlations that are too complex forhumans to see and suggesting trading ideas as a result. Humanswill still be needed to validate these ideas, he says reassuringly.

But the prospect of even faster markets raises the problemposed by the In�nium case and by �ash crashes large and small:the threat from HFTs to the stability of markets. Software has anasty habit of developing bugs. Algorithms behave unpredict-ably once they are out of a testing environment and into the mar-ket proper. Even leaving aside the potential for deliberate marketabuse, traders will sometimes get things wrong.

Plugging the holes

So real-time risk-management safeguards have to be put inplace that work at di�erent levels�of the HFTs, the prime bro-kers, the exchanges and the regulatory agencies. Some of thesesafeguards already exist. The exchanges have limits on ordersabove a certain size, for instance, and on the number of ordersthat can come in. But there are always holes to be �lled.

The May 2010 �ash crash revealed a glaring problem withthe structure of American equity markets, for instance: that cir-cuit-breakers which automatically pause trading if there are viol-ent price swings kicked in only once the entire market reached acertain threshold. The regulators have since introduced �single-stock circuit-breakers� for any stock that moves up or down by10% or more in a �ve-minute period. Most observers are con�-dent that this would stop another �ash crash in its tracks.

However, a focus on equity markets may distract attentionfrom other asset classes where HFTs are present and growing.The Bank for International Settlements, a club of central bankers,last September issued a useful fact-�nding report on the role ofhigh-frequency trading in the vast foreign-exchange markets. Itworried about the capacity of the prime brokers, through whichHFTs gain access to credit, to keep pace with their clients. If theydo not, the prime brokers themselves are exposed to the pos-sibility of HFTs rapidly accumulating risky positions.

Again, the industry is aware of this risk. Traiana, a post-trade processing �rm, launched a software program last year thataggregates clients’ positions across lots of di�erent venues in realtime and activates a �kill switch� that stops clients trading oncepre-de�ned limits are breached. But this is a voluntary initiative,not one that is required by regulators. Other markets remain un-protected. Regulators should not be afraid to act �rmly to de�neand enforce standards for market surveillance and trading con-trols across venues and asset classes. And getting the infrastruc-ture right is important in other areas too. 7



2

1

backlog of uncon�rmed trades, which in the event of a defaultwould have left buyers of protection without proof of purchase.

The prospect of systemic legal uncertainty prompted ac-tion. In September 2005 Tim Geithner, now the treasury secre-tary but then the head of the New York Federal Reserve Bank,summoned the broker-dealers to his Wall Street headquartersand knocked heads together. The industry began moving to-wards electronic con�rmations and data repositories to recordtrades, and the backlogs started to reduce. �Everyone knew thatit was a paper mess but no one wanted to sign up to a single stan-dard,� recalls one participant in that meeting. �Regulatory inter-vention will clean up markets incredibly well.� By the time the2008 tornado hit, the CDS market functioned pretty e�ciently, atleast in an operational sense. The process for settling CDS claimsafter Lehman Brothers went bust, in particular, passed o� moresmoothly than expected.

But there is only so much room for congratulation. In onecrucial respect the infrastructure for CDSs failed horribly: the un-noticed accumulation by AIG, an insurer-turned-derivatives-trader, of huge potential liabilities as it �ogged protection againstdefault. Although the 2005 intervention did a lot of good, it didnot deal properly with bespoke, over-the-counter (OTC) trades,the bit of the market where AIG was operating.

CDS contracts o�er protection against an event (ie, default)rather than a move in prices, as happens in markets such as inter-est-rate swaps. That means obligations can be triggered sudden-ly, a phenomenon known as �jump-to-default risk�. The bestway to protect against this risk is to ensure that the parties to the

contract are posting collateral every day as prices �uctuate. Un-less a default comes totally out of the blue, that can cushion theimpact of a sudden credit event. AIG, notoriously, did not postenough collateral until it was suddenly asked for una�ordableamounts, and seemed to be banking on not having to pay out.

Belatedly, �nancial reform is dealing with these infrastruc-ture �aws, and not just for credit derivatives. Many more swaptransactions are heading for clearing houses, which stand be-tween buyers and sellers and collect collateral centrally. Records

IF THERE IS one bit of �nance where peopleagree on the need for more innovation, it isin lending to small business. Policymakersare desperate to get more credit �owing tothis vital engine of economic growth. Banksclaim that lending is muted because de-mand is subdued, but that is not the onlyproblem. Small and medium-sized en-terprises (SMEs) are harder credit risks toassess than large ones, so they attracthigher capital charges and are often the�rst to lose their funding in a downturn.

A host of new �rms have sprung upwith solutions. Some are seeking to �ll thegap left by the banks, rather than overhaulthe way that lending is done. Shawbrook is anew, specialised lender to small �rms inBritain, where the dominance of a handfulof big banks makes the choice for SMEsparticularly limited. Owen Woodley, Shaw-brook’s boss, says that it can get its creditanalysis done faster than the establishedinstitutions.

Other �rms are trying to reinventsmall-business �nancing by providingvirtual marketplaces where investors andSMEs can come together. In the world ofequity capital the pacesetter is a British �rm

called Crowdcube, which uses the idea of�crowdfunding� to enable lots of investorsto buy up small stakes in start-up �rms.

Picking winners among entrepreneursis notoriously di�cult. Venture capitalists’answer is intensive screening by a smallteam of dedicated investors, followed byhands-on involvement in the business. TheCrowdcube model, which is due to come toAmerica if crowdfunding legislation passes,depends on the ability of thousands ofmembers to ferret out the best ideas. Thegeneral public cannot match the expertiseand commitment of dedicated �angel�investors if a �rm gets funded, admitsDarren Westlake, Crowdcube’s founder. Butit helps to have lots of investors acting asadvocates for a start-up �rm.

The same peer-to-peer model liesbehind Funding Circle, another Britishstart-up which launched in 2010 to facilitatelending to small businesses for periods ofone to three years. Businesses go throughan underwriting process before they get onto the company’s website, where lenders,predominantly individuals, can bid on therate at which they are prepared to lend. Theaverage loan is £40,000 ($63,000), the

rates are competitive and �rms get hold ofthe money within about two weeks. SamirDesai, a co-founder, dismisses the argumentthat lending to small �rms requires bankersto make personal credit assessments: �SMEswant low costs, a quick process and a trans-parent fee structure, not a relationship.�Lenders are encouraged to spread their riskamong at least 20 borrowers.

There is another way of speeding upthe underwriting process: taking a bet noton an SME itself but on its debtors. TheReceivables Exchange, launched in NewOrleans in 2007, enables investors to buy upinvoices, or fractions of them, owed to smallbusinesses. The idea is similar to factoring,whereby �rms sell o� all their invoices at adiscount to improve their cash�ow. But theidea behind The Receivables Exchange�andMarketInvoice, a British equivalent�is tobreak receivables down into small, tradableunits so that buyers can make judgments onindividual debtors and diversify their hold-ings. �We provide electron-level transpa-rency,� says Nic Perkin, a co-founder of TheReceivables Exchange. Transactions aresomewhat less minuscule, approaching arate of $1billion a year.

On the side of the angels

New ways of lending to small businesses



1

PUT TOGETHER ALL these aspects of �nancial innova-tion�experimentation, standardisation, infrastructure

gaps and the illusion of safety�and one area of the post-crisis �-nancial landscape �ashes red as a potential source of problems:collateralisation, or giving a lender an asset as security in case aborrower defaults.

Demand for collateral, at least in areas like OTC derivatives,is growing. In Europe in particular, bank creditors are pushinghard towards secured funding to protect themselves in the eventof trouble. �A structural shift to collateralised funding is goingon, and these forces are very strong,� says Imène Rahmouni-Rousseau of the FSB. Moreover, perceptions of what counts asgood collateral have changed. Before the crisis residential mort-gage-backed securities of all kinds were being widely used inrepo transactions (a form of short-term funding); now the em-phasis is on �nding highly rated government bonds, itself ashrinking universe. A recent IMF paper estimated that a declinein the amount of �pledged collateral�, the sort that can be reused

in other transactions, has re-duced the overall availabil-ity of collateral by $4 tril-lion-5 trillion since pre--Lehman days (see chart 7).

This collateral squeezeis fuelling all sorts of innova-tion. One which has alreadyattracted the attention ofregulators is the �liquidityswap�, whereby holders ofliquid assets, such as insur-ers and pension funds withgovernment bonds, lendthem to banks to use as col-lateral in their secured fund-ing. In return, banks lendthem less liquid assets and

Collateral

Safety first

Innovative ways of making lenders feel more secure

7Squeezed

Source: “Velocity of Pledged Collateral”,by Manmohan Singh, IMF, November 2011

Pledged bank collateral, $trn

0

1

2

3

4

5

6

2007 08 09 10

Europe & Nomura US

of all cleared and OTC trades across all thebig classes of derivatives will be collectedand stored in so-called �swap-data reposi-tories�, so the chances of anyone buildingup a big undisclosed position should besharply reduced. Much of this is well un-der way. �We are very close to a more com-prehensive view in terms of systems en-abling regulators to see the tradinguniverse in intra-day updates,� says PeterAxilrod of the Depositary Trust and Clear-ing Corporation (DTCC), the world’s larg-est data repository.

But some argue that having a biggerhaystack of data makes it harder to �ndthe really important needles. Others wanta lot more information. No national regu-lator can see all of the �nancial system: anAmerican regulator can see the CDS expo-

sures of American banks to French banks, for instance, but is notallowed to see the counterparty risks of the French lenders inturn. And there is critical work to be done to gather data on whatcollateral is being posted as security on trades and to pull togeth-er an aggregate picture of the exposures being taken across deriv-atives classes.

Even so, things are far better than they were. �We will be ina position in ten years where we have an amazing derivatives in-frastructure,� says Edmund Parker, a derivatives partner at MayerBrown, a law �rm. In the meantime the attention of regulators isbroadening out to other bits of �nancial sca�olding. There is talkof taking a closer look at the way the securities-lending businessworks, for instance. An American task force has been at work forover two years looking at an obscure but vital area of short-term�nancing called the tri-party repo market. �Margining and tech-nical policy and back-o�ce monitoring of positions against col-lateral are unsexy but it is the stu� to be focused on,� says Mo-hamed Norat of the IMF.

The strangeness of comfort

This is particularly true when it comes to innovations thatpledge to transfer or reduce risk. Many of the instruments andtechniques that were most lauded before the crisis were de-signed to package risk and shift it away from people who did notwant it towards those who did. More transparency might havemade it clear that risk was simply being concentrated some-where else, or was not really leaving banks’ balance-sheets at all.

This weakness in infrastructure compounded a behaviour-al one. Finance is at its most dangerous when it is perceived to besafe. One element in the �nancial crisis was a failure to under-stand the risks inherent in various products until it was too late.The resulting fragility is described in a 2010 paper on �nancial in-novation by Nicola Gennaioli of Pompeu Fabra University, An-drei Shleifer of Harvard University and Robert Vishny of theUniversity of Chicago. The authors suggest that many episodesof �nancial innovation (including the excessive issuance ofmortgage-backed securities) start in the same way: with inves-tors’ demand for a set of cash�ows that do not carry much risk.There is only a limited number of existing instruments of thatsort at a reasonable price. So the �nance industry meets the de-mand by creating new instruments that are designed to be assafe as existing ones (think of those AAA ratings on diversi�edbundles of loans).

Investors do not necessarily think through all the risks em-bedded in these new instruments (for example, that a nationalhousing bust would render the tranching within CDOs useless)

and buy them enthusiastically. When those risks materialise,there is a destabilising �ight to safety. �The standard argumentfor �nancial innovation is that there are gains from trade, but thatmodel crumbles if you suppose that people do not fully under-stand the risks,� says Mr Shleifer. His main bugbear is the money-market mutual fund, which o�ers instant liquidity and promisesno risk to investors’ principal. It saw a massive run when one ofthem �broke the buck� in 2008.

This analysis rings true of much of �nance: people are lia-ble to forget about the risks of products that have already blownup as well as misjudge those that have been newly created. Theeuro zone’s debt crisis has shown that risks even in long-estab-lished instruments like government bonds can be underestimat-ed. But innovations are particularly susceptible to the problemof self-delusion. If they go wrong early enough, they are unlikelyto get o� the ground. But once they reach a su�cient scale with-out a big blow-up, nobody believes that they might be �awed. 7

2

and complacency. For unse-cured creditors such as bond-holders, the concern is a pro-blem known as �encum-brance�. It means that the moreof a bank’s balance-sheet is tiedup as collateral, the fewer assetswill be available to them in theevent of a default. Again, dataare in short supply. Fitch, a rat-ings agency, pointed out in De-cember that European banksare both increasingly reliant onsecured funding and reluctantto disclose which assets havebeen pledged as collateral. Forsecured creditors, the risk is thatthey will feel safer than theyreally are: valuation method-ologies may not be right, for ex-ample, or assets may not beproperly segregated.

As in other areas of �nan-cial innovation, the best way ofkeeping the collateral problemunder control is one of intrusivevigilance. That does not meanbanning everything in sight. Fi-nance has a very good record ofsolving big problems, from en-abling people to realise the val-ue of future income through products like mortgages to protect-ing borrowers from the risk of interest-rate �uctuations. It istempting to choke o� creativity in the aftermath of the crisis, butthere is no obvious way of sifting the good innovations from thebad at the outset, and far less chance for the industry to mitigateproblems�such as the impact of rising longevity, or ine�ciencyin public spending, or the lack of credit for millions of poor peo-ple in emerging markets�if it cannot experiment.

The problems come along after the experimentation phase,when products have taken root and are growing wildly. The evi-

dence of this special report suggests that the market does a bril-liant job of nurturing and re�ning instruments that people want.It is less good at planning for when things go wrong. This iswhere regulators can play a benign role, signalling concerns ear-ly and loudly when activities reach a certain scale (as has hap-pened with ETFs and liquidity swaps) and setting standards fordata and reporting before the infrastructure starts to creak (ashappened with CDSs in 2005).

Regulators are far from perfect. They proved incapable ofspotting danger before the crisis, and will doubtless make simi-lar mistakes in the future. And �nance will still have many old-fashioned ways to blow itself up. That is why it is so important tohave lots of capital in the system to head o� a calamity. But if in-novation can be encouraged without being given licence to runriot, there will be a better chance of the industry doing good in-stead of harm. 7

2

Offer to readers Reprints of this special report are available. A minimum order of five copies is required.Please contact: Jill Kaletha at Foster Printing Tel +00(1) 219 879 9144e-mail: [email protected]

Corporate offerCorporate orders of 100 copies or more are available. We also offer a customisation service. Please contact us to discuss your requirements.Tel +44 (0)20 7576 8148e-mail: [email protected]

For more information on how to order special reports, reprints or any copyright queries you may have, please contact:

The Rights and Syndication Department26 Red Lion SquareLondon WC1R 4HQTel +44 (0)20 7576 8148Fax +44 (0)20 7576 8492e-mail: [email protected]/rights

Future special reports

Previous special reports and a list of forthcoming ones can be found online: economist.com/specialreports

pay a fee. Such deals are not entirely new but there are concernsthat they could quickly grow in scale, binding the insurance andbanking sectors more closely together. Britain’s Financial Ser-vices Authority issued some disapproving guidance about thepractice last summer and has reportedly blocked a number ofplanned swaps since.

If one path is blocked, bankers will try to �nd others. Chris-topher Georgiou of Ashurst, a law �rm, says that banks are creat-ing guaranteed repackaging structures in which assets arepledged as collateral and guaranteed by the bank.

American regulators are also tracking attempts to launch anew type of funding instrument called �collateralised commer-cial paper�, which some think has the potential to replace asset-backed commercial paper as a big source of �nancing. It worksby giving investors a claim over collateral used in certain repoagreements without breaching new rules that prevent money-market funds from investing directly in such repo transactionsthemselves. The market has not yet caught �re, but supervisorsare struck by how quickly the industry has started innovating inthis area. �I am amazed that these things were developed beforethe rules were even in place,� says one.

Some banks see an opportunity to be collateral intermedi-aries. A clearing house may be wary of taking corporate bondsdirectly from a pension fund as collateral, say, but a bank can takethe bonds and then issue its own letter of credit to the clearinghouse. In an echo of the triparty repo market, triparty collateralmanagement is also becoming more popular. As Rajen Shah,head of collateral management at Citigroup, explains, under thissystem assets used as collateral sit with a third party. If brokerswant to change the assets they pledge during the day, they can doso more quickly and cheaply via the agentthan by managing collateral bilaterally.

Within banks, too, collateral man-agement has moved from the back of theback o�ce to a much more prominent po-sition. Institutions are investing in sys-tems that provide them with a centralview of what collateral is being held bydi�erent desks and allow them to move itaround e�ciently, to specify the order inwhich collateral is put to work during theday and to model the correlation risks em-bedded in di�erent types of collateral (ie,if a counterparty defaults, whether its col-lateral is likely to sour, too). �Collateral de-partments were the orphaned, redheadedstepchildren of the organisation,� saysone executive. �Now they are seen as themost interesting entrepreneurial and commercial opportunity.�

In many ways, the collateral craze exempli�es the chal-lenges of judging �nancial innovation. The idea of using collater-al to provide added security for the creditor is a sound one: itshould reduce the riskiness of lending and the chances of a run.Such reasoning is re�ected in the lower capital charges thatbanks incur for secured lending. Much can be done to improvethe e�ciency of collateral-management systems, and a rethink isneeded on what counts as a safe asset.

But the whirring of �nanciers’ minds also spells trouble. Ascollateral �ows more e�ciently through the system and waysare found to transform and enhance it, the regulators have a bigjob to keep up. Working out which collateral belongs to whom,whether it is fairly valued, and what would happen at a systemiclevel if there was a big margin call on a particular asset class areall huge tasks for the future.

There are also the familiar problems of interconnectedness

The market does a brilliant job of nurturing and re�ninginstruments that people want. It is less good atplanning for when things go wrong

Nuclear energy March 10th Cuba March 24thManufacturing April 21stInternational banking April 21st

SPECIAL REPORT


FINANCIAL INNOVATION

february 25th 2012 - the economistmedia.economist.com/sites/default/files/... · 2 the...

Documents