Playing with Fire

SPECIAL REPORT FINANCIAL INNOVATION February 25th 2012

Playing with fire

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

SPECIAL REPORT FINANCIAL INNOVATION

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. Paul CONTENTS Volcker, a former chairman of America’s Federal Reserve, who emerged 4 How innovation happens from the 2007-08 nancial crisis with his reputation intact, once said that The ferment of nance none of the nancial inventions of the past 25 years matches up to the ATM. Paul Krugman, a Nobel prize-winning economist-cum-polemicist, 5 Retail investors has written that it is hard to think of any big recent nancial break- The little guy throughs that have aided society. Joseph Stiglitz, another Nobel laureate, 6 Exchange-traded funds argued in a 2010 online debate hosted by The Economist that most inno- From vanilla to rocky road vation in the run-up to the crisis was not directed at enhancing the 9 High-frequency trading ability of the nancial sector to The fast and the furious perform its social functions. 11 Financial infrastructure Most of these critics have Of plumbing and promises market-based innovation in their sights. There is an enormous 12 Small rms amount of innovation going on On the side of the angels in other areas, such as retail pay- 13 Collateral ments, that has the potential to Safety rst change the way people carry and spend money. But the debate and hence this special reportfo- Glossary cuses mainly on wholesale pro- Finance has a genius for obscuring simple ducts and techniques, both be- ideas with technical jargon. Some of the cause they are less obviously terms used in this special report are useful than retail innovations briefly explained here and because they were more Collateralised-debt obligation: a security heavily implicated in the nan- whose payments flow to investors in order cial crisis: think of those evil cred- of their position in the payment hierarchy. it-default swaps (CDSs), collater- Credit-default swap: a type of insurance alised-debt obligations (CDOs) policy that pays out when an issuer and so on. defaults. This debate sometimes re- Exchange-traded fund: a basket of volves around a simple question: securities that tracks an index and is listed ACKNOWLEDGMENTS is nancial innovation good or bad? But quantifying the benets of inno- on an exchange. High-frequency trading: an ultra-fast As well as the people mentioned in vation is almost impossible. And like most things, it depends. Are credit algorithmic trading strategy. this special report the author would cards bad? Or mortgages? Is nance as a whole? It is true that some instru- like to thank the following for their mentsfor example, highly leveraged onesare inherently more danger- Liquidity swap: a transaction to swap help: Francisco Arcilla, Amar Bhidé, illiquid assets for liquid ones. Mike Bodson, Peter Carroll, Dominic ous than others. But even innovations that are directed to unimpeach- ably good ends often bear substantial resemblances to those that are Longevity swap: a transaction to lay off the Casserley, Leo Civitillo, Guy Cough- risk of unexpected changes in life lan, Onur Erzan, Geert Glas, Will now vilied. expectancy. Goetzmann, Chi Hum, Ross Levine, For a demonstration, look at Peterborough. The cathedral city in Emmanuel Naim, Jim Overdahl, Repo: a (typically) short-term funding Robert Pickel, Craig Pirrong, Michael England’s Cambridgeshire is known for its railway station and an under- transaction in which the borrower sells a Poulos, Andrew Reid, Pretty Sagoo, achieving football club nicknamed the Posh. But it is also the site of a security for cash and agrees to buy it back Nick Shellard, Susan Smith, Conrad nancial experiment that its backers hope will have big ramications for at a later date. Voldstad, David Weild, the sta of the way public services are funded. Social-impact bond: an instrument to the Observatory for Responsible Innovation at the École des Mines de Peterborough is where the proceeds of the world’s rst social-im- finance social programmes that links Paris and a number of people who pact bond are being spent. This instrument is not really a bond at all but investors’ returns to the outcome of wished to remain anonymous. behaves more like equity. In September 2010 an organisation called So- those programmes. cial Finance raised £5m ($7.8m) from 17 investors, both individuals and Swap-data repository: an entity designed A list of sources is at to collect data on derivative transactions Economist.com/specialreports charities. The money is being used to pay for a programme to help prevent ex-prisoners in Peterborough from reoending. Reconviction rates and make the information available to An audio interview with regulators. the author is at among the prisoners recruited to the scheme will be measured against a VIX: a measure of the expected volatility Economist.com/audiovideo/ national database of prisoners with a similar prole, and investors will of the S&P 500, known as the “fear index”. specialreports get payouts from the Ministry of Justice if the Peterborough cohort does 1

The Economist February 25th 2012 1 SPECIAL REPORT FINANCIAL INNOVATION

2 better than the rest. If all goes well, the rst payouts will be made in 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. Linking payouts to outcomes is attractive to governments keen to hus- band scarce resources. And if service providers like the people running the Peterborough prisoner-rehabilitation scheme can get a lump sum up front, they can plan ahead without bearing any nancial risk. There is talk of introducing social-impact bonds in Australia, Canada and the United States. Here, surely, is a nancial innovation that even the industry’s critics would agree is worth trying. Yet in fundamental ways an ostensibly good instrument like a social-impact bond is not so dierent from its despised cousins. First, at its root the social- impact bond is about creating a set of cashows to suit the needs of the sponsor, the provider and the investor. True, the investors in the Peterborough scheme may be more willing than the average individual or pension fund to sacrice nancial returns for social benets. But as Franklin Allen of the Wharton School at the University of Pennsylvania and Glenn Yago of the Milken In- stitute, a think-tank, argue in their useful book, Financing the Future, the thread that runs through much wholesale nancial innovation is the creation of new capital structures that align the interests of lots of dierent parties. Second, the social-impact bond is based on the concept of be all that much. risk transfer, in this case from the government to nancial inves- Whether protecting a retirement pot or signalling problems tors who will get paid only if the scheme is successful. Risk trans- with a government’s debt burden, nance can be socially use- fer is also one of the big ideas behind securitisation, the bundling ful (to use a phrase popularised by Adair Turner, the outgoing of the cashows from mortgages and other types of debt on chairman of Britain’s Financial Services Authority) without be- lenders’ books into a single security that can be sold to capital- ing obviously social. Lord Turner himself acknowledged that in markets investors. The credit-default swap is an even simpler a speech he gave in London in 2009: It is in the nature of markets risk-transfer instrument: you pay someone else an insurance that there are some things which are indirectly socially useful premium to take on the risk that a borrower will default. but which in the short term will look to the external world like Third, even at this early stage the social-impact bond is pure speculation. grappling with the diculties of measurement and standardisa- Many people point to interest-rate swaps, which are used to tion. An obvious example is the need to create dened sets of bet on and hedge against future changes in interest rates, as an ex- measurements in order to work out what triggers a payoutin ample of a huge, well-functioning and useful innovation of the this case, the comparison between the Peterborough prisoners modern nancial era. But there are more contentious examples, and a control group of other prisoners in a national database. too. Even the mention of sovereign credit-default swaps, which Across nance, standardisationaround contracts, reporting, oer insurance against a government default, makes many Euro- performance measures and the likeis what enables buyers and peans choke. There are some specic problems with these in- sellers to come together quickly and new markets to take o. struments, particularly when banks sell protection on their own governments: that means a bank will be hit by losses on its hold- Neither angels nor demons ings of domestic government bonds at the same time as it has to For all the similarities, there are two big dierences be- pay out on its CDS contracts. But in general a sovereign CDS has a tween the social-impact bond and other, less lauded nancial in- useful signalling function in an area tilted heavily in favour of struments. The rst is that the new tool has been designed explic- governments (which do not generally have to post collateral and itly for a social purpose. But ask a pensioner how much money can bully domestic buyers into investing). he wants to put into prisoner rehabilitation, and it isn’t likely to The second dierence is that social-impact bonds are still in 1

A selective history of financial innovation 1602 c.3000 1024 Establishment 1667 1774 Banking c.640 First state-backed of first stock First insurance First mutual fund 1924 developed First coins paper money exchange company invented in First modern in Mesopotamia produced in Lydia introduced in China in Amsterdam in London the Netherlands mutual fund founded

1668 First 1780 central bank First inflation-linked founded bonds issued in BC AD in Sweden Massachusetts

3000 2000 1000 500 0 500 1000 1500 1600 1700 1800 1900 1925 Source: “Financing the Future” by Franklin Allen and Glenn Yago, Wharton School Publishing 2010; The Economist

2 The Economist February 25th 2012 SPECIAL REPORT FINANCIAL INNOVATION

2 their infancy, whereas other crisis-era innovations were directly cation of portfolios as banks shed concentrations of risks and in- involved in a gigantic nancial crisis. There are questions to an- vestors buy exposures that suit them. Securitisation is a good swer about their culpability. A few products from that period do thing. If everything was on banks’ balance-sheets there wouldn’t look inherently awed. Only the bravest are prepared to defend be enough credit, says a senior American regulator. the more exotic mortgage products that sprouted at the height of Rather than asking whether innovations are born bad, the America’s housing bubble as lenders found ever more creative more useful question is whether there is something that makes ways to bring unaordable houses within reach. Finance profes- them likely to sour over time. sionals almost blush to recall an instrument called the constant- proportion debt obligation, a 2006 invention of ABN AMRO that Greed is bad added leverage when it took losses in order to make up the short- There is an easy answer: people. When bubbles froth, fall. The end of the structured investment vehicle (SIV), an o- greedy folk use innovations inappropriatelyto take on expo- balance-sheet instrument invented to game capital rules, is not sures that they should not, to manufacture risk rather than trans- much lamented. And the complexity of the CDO-squared has fer it, to add complexity in order to plump up margins rather than been widely condemned. solve problems. But in those circumstances old-fashioned - But even now it is hard to nd fault with the concept, as op- nance goes mad, too: for every securitisation stued with sub- posed to the practical application, of many of the most demo- prime loans in America, there was a stinking property loan sit- nised products. The much-criticised CDO, which pools and ting on the balance-sheet of an Irish bank or a Spanish caja. tranches income from various securities, is really just a capital Du credit analysis is always the cause of the problem, says Si- structure in miniature. Risk-bearing equity tranches take the rst mon Gleeson of Cliord Chance, a law rm. hit when things go wrong, and more risk-averse investors are This argument has a lot of power. When greed takes hold, more protected from losses. (Euro-zone leaders like the idea nance in all its forms is undone. Yet blaming the worst out- enough to have copied it with their plans for special-purpose incomes of nancial innovation on human frailty is hardly help- vestment vehicles for peripheral countries’ sovereign debt.) The ful. This special report will point to the features of nancial inno- real problem with the CDOs that blew up was that they were vations that can turn them into troublemakers over time and stued full of subprime loans but treated by banks, ratings agen- show how these can be managed better. cies and investors as though they were gold-plated. In simple terms, nance lacks an o button. First, the in- As for securitisation and credit-default swaps, it would be dustry has a habit of experimenting ceaselessly as it seeks to blinkered to argue they have no problems. Securitisation risks build on existing techniques and products to create new ones giving banks an incentive to loosen their underwriting stan- (what Robert Merton, an economist, termed the innovation spi- dards in the expectation that someone else will pick up the ral). Innovations in nanceunlike, say, a drug that has gone pieces. CDS protection may similarly blunt the incentives for through a rigorous approval process before coming to market lenders to be careful when they extend credit; and there is a spe- are continually mutating. Second, there is a strong desire to stan- dardise products so that markets can deepen, which often acceler- When bubbles froth, innovations are used ates the rate of adoption beyond the capacity of the back oce inappropriatelyto take on exposures that and the regulators to keep up. should not have been, to manufacture risk As innovations become more and more successful, they rather than transfer it, to add complexity start to become systemically sig- nicant. In nance, that is auto- cic problem with the way that the risk in these contracts can matically worrying, because the consequences of any failure suddenly materialise in the event of a default. can ripple so widely and unpredictably. In a 2011 paper for the But the basic ideas behind both these two blockbuster in- National Bureau of Economic Research, Josh Lerner of Harvard novations are sound. India, with a far more conservative nan- Business School and Peter Tufano of Said Business School also cial system than America, allowed its rst CDS deals to be done argue that in a typical S-curve pattern, in which the earliest in December, recognising that the instrument will help attract adopters of an innovation are the most knowledgeable, a widely creditors and build its domestic bond market. Similarly, securiti- adopted product is more likely to have lots of users with an inad- sationwhich worked well for decadesallows banks to free up equate grasp of the product’s risks. And that can be a big problem capital, enabling them to extend more credit, and helps diversi- when things turn out to be less safe than expected. 7

1946 1967 1970 1983 1994 Birth of 1955 First ATM Securitisation Grameen Bank, microfinance First venture-capital First leveraged machine installed of US mortgages pioneer, becomes credit-default industry buy-out deal in London begins independent bank swap transaction 1949 1973 1989 First hedge fund Black-Scholes First exchange-traded 1997 formed by Alfred options-pricing fund launched First catastrophe- Winslow Jones formula published in Canada bond transaction

1940 1950 1960 1970 1980 1990 2000 2010

The Economist February 25th 2012 3 SPECIAL REPORT FINANCIAL INNOVATION

How innovation happens ooad a particular risk and there is no generic solution to hand, it is the strats’ job to use derivatives to create one. That’s what we have to do when clients call, says Mr Chavez. We can’t tell The ferment of finance them ‘no thanks’. At times like this the calls come thick and fast. When yields are low and uncertainty is high, there is strong demand for investment products that are structured to take more risk but cap potential losses. Take dened-contribution pension plans, Moving from ideas to products to markets where individuals rather than companies now bear the risk that their pension pot will fall short of their retirement needs. They WHAT WITH THE world’s economic woes and the indus- can take out an annuity, but that is unappealing when interest try’s regulatory overhaul, this may not seem like the most rates are ultra-low. So investment banks are seeing more appetite propitious time for nancial innovation. Yet the current environ- for derivatives-based solutions that will guarantee minimum ment is perfect for stimulating nance’s creative juices. Regula- payments and allow for a capped amount of gains. tion has always sparked bouts of inventiveness as practitioners Hedging is another big area of interest for clients. Before seek both to get around new rules (as they did with the introduc- the crisis, hedging was viewed like a tax, says Stéphane Matta- tion of the SIV in the late 1980s) and to benet from them (for ex- tia, Société Générale’s global head of equity ow engineering ample, rule changes in the 1970s that allowed pension funds to and advisory. Now it is becoming an art, just as important as as- put more money into high-risk assets). set allocation. Mr Mattatia recalls a client who approached the Clients have big problems to solve, too. Tempting though it bank in April 2010 nursing some prescient worries about French is to imagine bankers cooking up wild schemes in their Wall CDSs. The bank constructed a hedge based on the euro falling Street lairs, innovation is often triggered by a client coming to a and gold rising, a scenario that the nancial models discounted bank with a specic headache. Sometimes those headaches because gold tended to be negatively correlated with the dollar have a ready-made cure, but often they need bespoke treatment. and positively correlated with the euro. That lowered the cost of Martin Chavez is co-head of a team of strats (nancial engi- the hedge and raised returns to the investor. neers) at Goldman Sachs which acts as Goldman’s in-house re- SocGen has also constructed an o-the-shelf answer to the search-and-development division. If a client wants to take on or risk of a market meltdown: an exchange-traded fund for institutional investors which is based on the VIX, a measure of stockmarket volatility (see chart 1, next page). The fund invests in VIX futures contracts, shifting from cheaper long-dated ones into pricier short-dated ones when the VIX moving average reaches a certain threshold. The idea is to keep the fund’s overall costs down but allow investors to benet from sudden 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 derivatives products a year, outlines a three- stage process for innovation: investiga- tion, creation and validation. The investi- gation phase establishes whether there is a market need for a product. In the creation phase contract specications are developed and the product is tested exter- nally on a small group of early adopters. The validation phase involves feedback from a wider group of people within the CME, such as sales and compliance. If someone at an investment bank has a bright idea for a new product, it typically goes to a new-business committee, where product controllers look at issues like pricing and valuation, risk managers consider how it can be hedged, and legal and compliance folk worry about the ne print. If a product seems to have the potential to tarnish a bank’s brand, it usually gets passed to a reputational-risk committee. At many banks these processes have 1

4 The Economist February 25th 2012 SPECIAL REPORT FINANCIAL INNOVATION

be treated like a child, is the withering verdict of one regulator. Fear, indexed 1 When an innovation is rst launched, it is by denition not 1,600 80 systemic. It might not work out well for its particular buyers and S&P volatility index (VIX) sellers but it will not blow up the economy. The worry for regu- 1,400 70 S&P 500 lators and taxpayers starts when products and processes achieve 1,200 60 a scale where their aws can have far-reaching consequences. 1,000 50 Most innovations take time to reach that point. With mem- 800 40 ories still fresh of a boom when lots of new ideas quickly took 600 30 root, it may seem as though all that is needed to make the money 400 20 ow is an investment banker with a brainwave. But for every market that booms there are others that sputter. 200 10 The classic example is the market for property derivatives. 0 0 Property is the world’s biggest asset class, obvious ground for in- 2007 08 09 10 11 12 Source: Thomson Reuters struments that allow property owners to hedge against price falls and to gain exposure to assets that they cannot buy for themselves. But it is also a very lumpy, illiquid, idiosyncratic as- 2 been tightened up a lot since the crisis. The new product-ap- set, which makes it hard to construct an instrument that matches proval processis extremely conservative, indeed too conserva- the performance of specic properties. tive, says a London-based banker. Enthusiasts have answers to this problem of basis risk, The crisis has highlighted one specic area of diculty: the risk that a hedge will not precisely match movements in the judging the sophistication of a client (see box). At least one big price of the underlying asset. Paul Ogden, one of the founders of bank now steers clear of selling derivatives to municipalities be- InProp Capital, a fund manager that oers investors exposure to cause it feels it cannot be sure that bureaucrats will understand a British commercial-property index, says that derivatives what they are buying. Even supposedly expert investors may not should not be used to trade the risk associated with particular know what they are getting into. A German Landesbank should buildings but the correlated market risk, which drags prices 1

The little guy

What do small investors need more: choice or protection? THE BIGGEST QUESTION for supervisors is some extent. That throws you back to a beneciaries). The FSA’s concern is much how to protect retail investors from - dierent model where you try to ensure that more for the investor. It reckons the product nancial wizardry. A lawyer-turned-regulator products are reasonable and developed with is being sold hard on the basis of actuarial in America recalls how the paperwork for the interests of the client in mind. assumptions that are too complex for a auction-rate securities, a form of debt for That sounds unobjectionable in the- retail buyer to understand. which the interest rates were reset in rolling ory, even to many practitioners. We’re not The problem is that micromanaging auctions, made it clear that the instrument like Canute, telling product regulation to go products can easily mean restricting choice. could be hard to sell if auctions failed. away, says Timothy Hailes, a lawyer at Imagine a nancial instrument that exposes When people complained, I thought it was JPMorgan who chairs the Joint Associations retail investors to rst loss in the event of a ridiculous because this was on page one of Committee, a grouping of industry associa- problem, swings around wildly from day to the document. But it turned out that the tions with an interest in structured pro- day and delivers virtually no returns to salesmen had told them that this was risk- ducts. But there is uncertainty over what Western buyers for a decade. Would equities free, and a liquid alternative to Treasuries. the new approach will mean in practice. have got through the regulators? In Europe in particular regulators are Late last year, for example, the FSA now more inclined to intervene in specic came as close as it can with its current products. Britain’s Financial Services Au- powers to issuing its rst product ban, on thority (FSA) is setting the pace. In Novem- the marketing of life-settlement in- ber the regulator issued guidance on pro- vestments to retail investors. Life-settle- duct development for retail structured ment products are the sort of thing that gets products, a sign of its willingness to get reputational-risk committees to meet, if stuck in at the point when instruments are only because they are also known as death rst launched. bonds. The idea, well-established in Amer- Philosophically, the edice of world ica, is that elderly people can sell their regulation was built on two building life-insurance policies to investors who will blocks, says Martin Wheatley, the man who keep paying the premiums and collect the will head the new Financial Conduct Au- payout when the policyholder dies. That thority when Britain’s supervisory structure may sound chilling, although the same idea is revamped. One is the idea of the rational underpins the sale of annuities; and using a man and the other is that there is a reason- life-insurance policy to generate cash can able level of conduct in the sales channel. be a boon for some pensioners (for ex- Both of those building blocks broke down to ample, those who are very ill or have no

The Economist February 25th 2012 5 SPECIAL REPORT FINANCIAL INNOVATION

Exchange-traded funds Good news for some 2 British life expectancy at birth*, years Forecast year: From vanilla to rocky 86 2010 84 2002 82 road 1991 80 78 1971 1981 76 The Darwinian evolution of exchange-traded funds 74 IF STANDARDISATION IS one way to make nancial inno- 72 Actual vations catch on, mutation is another. By experimenting fu- 70 riously with new products, the industry can hit on iterations that 1966 70 75 80 85 90 95 2000 05 10 15 20 25 30 work. And once something has taken hold, a host of incremental Source: Office for National Statistics *Average of male and female cohorts changes follow as rms compete for custom. Any good idea is immediately copied and propagated like a virus, says Robert Li- tan of the Kauman Foundation. If it’s a bad virus, then you 2 everywhere up or down. Robert Shiller, a Yale professor who is have a pandemic. behind a set of housing futures traded on the CME, reckons that The industry does little patenting. On the retail side there is advances in information technology will eventually allow accu- some attempt to protect new technologies, but for capital-mar- rate local indices to be constructed. But progress has been glacial. kets businesses this is not a priority. In a global ranking of rms The birth pangs of another market, in longevity risk, also assigned patents in America in 2011, drawn up by IFI CLAIMS demonstrate that getting innovations widely adopted is not Patent Services, the rst nancial rm in the list was American straightforward. The need for a product is clear: people are living Expressin joint 259th place. ever longer (see chart 2), which creates risk for institutions such One reason may be that in the capital markets ideas require as corporate pension schemes and annuity providers that will liquidity to take o: the more institutions that imitate an instru- have to provide retirement incomes for longer than expected. ment, the deeper liquidity is likely to be. Another is that copying The idea is to parcel this longevity risk out to someone who is is easy. A nancial product is about as conceptual as you can prepared to bear the cost if it materialises. get, says Wilson Ervin, a senior adviser at Credit Suisse. You just need paper and ink. A patent has to be published after a The lives of others year or so, enabling rivals to design around it. In any case, mone- Basis risk is again a big stumbling-block. A pension fund tising an idea immediately opens the door to copying because of wants to transfer its own specic risk so that it gets a payout if its disclosure requirements under securities laws. Mr Ervin reckons scheme members live longer than expected. But that requires that rms have a window of three to four months before rival lots of actuarial expertise on the part of the buyer. Investors products appear. It does not help that clients shop around to see want something tradable, where they don’t have to make actuar- if they can get the same sort of thing cheaper from another rm. ial judgments and can have standardised products, says David There are cultural and technical barriers to using patents, Howell, the chief executive of Pacic Life Re, a reinsurer. That too, according to Karen Hagberg of Morrison and Foerster, a law means an index measuring rates of mortality improvement that rm. The targets for enforcement would be other nancial insti- can act as a common reference point for buyers. tutions and they are hesitant to sue each other. As a result, in the But there is also a need for a market-maker to bridge the gap past courts have been asked to decide relatively few patent-litiga- between these two communitiesa reinsurer, insurer or invest- tion cases in the nancial-services industry. Without an existing ment bank that has the skills to enter into bespoke transactions body of law to provide some insight into how courts would de- with sellers and is willing to take the basis risk of an index-based cide relevant issues in the future, such as what constitutes an in- transaction to push it out to capital-markets investors. It is the fringement and how to calculate damages, it is riskier to call in same process of intermediation as exists in the more established the lawyers. Instead, products are left to spawn and mutate at 1 market for catastrophe bonds, issued by insurers and reinsurers and designed to pay out when a natural disaster strikes . Longevity risk, however, poses an additional problem. Here to stay 3 Whereas a catastrophe can occur in an instant, longevity risk Exchange-traded funds, assets $trn takes decades to unfold. That creates another mismatch because bond investors typically want to lock up their money for just ve 1.4 or ten years, which may be too short to act as an eective hedge Equity Fixed income 1.2 against people living longer. Alison Martin of Swiss Re, another Commodity reinsurer, says her rm is trying to increase the duration of lon- 1.0 gevity-related instruments. She adds: The point of innovation is to test dierent structures and durations and see what works. 0.8 There is something reassuring about these painstaking ef- 0.6 forts to create a market, but there are risks too. Standardisation is a precondition for growth, but it allows basis risk to build. Dur- 0.4 ing the nancial crisis some intermediaries retained too much of 0.2 this residual risk, and others transferred it to buyers that could not handle it. Our business is ultimately the management of ba- 0 sis risk, says Goldman Sachs’s Mr Chavez. Some do it better 2000 01 02 03 04 05 06 07 08 09 10 11 than others. No one did it well enough ahead of the crisis. 7 Source: BlackRock

6 The Economist February 25th 2012 SPECIAL REPORT FINANCIAL INNOVATION

gy stocks, that focus on emerging-market local-government debt, and many, many more. We’ll stop creating ETFs when you stop having ideas, boasts the website of iShares, BlackRock’s ETF arm. Other exchange-traded products allow investors to gain exposure to commodities, from gold to palladium. In August 2011the SPDR Gold Trust briey wrestled the crown for the world’s largest ETF from a fund tracking the S&P 500. Such vibrancy looks like a victory for the investor over the fund manager. Ye t ETFs have lost some of their lustre recently as the debate about their potential risks has become more vocal. That debate started in spring 2011as both the IMF and the Financial Stability Board (FSB), a global watchdog, voiced concerns about the instrument’s hidden trapdoors. It was given further impetus by erce disagree- ments within the industry about what kinds of products should bear the label ETF. And it demonstrates just how hard it is to control the development of a nancial innovation. Roughly speaking, there are two worries about ETFs. One is that they have become remarkably successful, and the other is that they are opaque. Their suc- cess is something that regulators fret about more than the industry does: in - nance, anything systemic is a potential threat. The absolute size of the ETF market is relatively modest compared with estimated global assets under the control of fund managers of $100 trillion or more. 2 their own pace, and not always in a healthy way. But the institutions at its heartin particular, investment banks Few instruments so encapsulate the mutability of nancial such as Deutsche Bank and Société Généraleare huge, which products as the exchange-traded fund, or ETF. No one has a bad makes people worry about how an ETF crisis might play out word to say about the concept. Invented in the late 1980s by the more broadly. And the rapid take-up of ETFs in itself is enough to Toronto Stock Exchange, the ETF is in essence a cheap mutual cause concern. If something is growing fast over a period of sev- fund: a basket of securities that tracks an index, is wrapped in a eral years and attracting a broader set of players and new en- fund structure and is listed on an exchange so that investors can trants, that is an alarm bell, says Nigel Jenkinson of the FSB. trade in and out whenever they want. They are cheap and tax-ef- cient, and they allow retail investors access to diversied port- Inscrutable folios of assets that had previously been the sole preserve of in- If ETFs had remained the plain vanilla products they stitutional investors. If you were inventing the mutual-fund were in their earliest incarnations, the regulators would have industry today, it would look like this, says Salim Ramji, a con- found little to get alarmed about. What troubles them is the in- sultant at McKinsey. strument’s opacity. The industry is still dominated by physical Even so, the product did not gain momentum immediately. products, with fund managers going out and buying each indi- According to Dan Draper, who runs Credit Suisse’s ETF business, vidual constituent of the index they are tracking. But there is a the spark to spectacular growth came only after the dotcom bust slice of the ETF market, making up a little over 10% of global as- had underscored the importance of diversication. The rise of sets under management and concentrated in Europe, that is fee-based nancial advice also pushed more investors towards synthetic. This means that the returns generated by the fund ETFs. In recent years the instrument has built up huge momen- come from a swap with a counterparty, often an aliate of the tum (see chart 3, previous page), and the 2007-08 nancial crisis fund manager. Instead of using investors’ cash to buy the under- did nothing to slow it. The number of ETFs on the market in lying securities, the money is put into a basket of collateral America has mushroomed from 343 in 2006 to 1,098 in Decem- whose returns are swapped with a counterparty for the returns ber 2011. A recent McKinsey report forecasts that global assets un- of the index being targeted. der management in exchange-traded products, a broader uni- That spooks plenty of people because it exposes investors verse of listed portfolio-tracking products, could grow from to counterparty risk. If the swap counterparty defaults, that about $1.5 trillion in 2010 to $3.1trillion-$4.7 trillion in 2015. leaves the investor holding the contents of the collateral basket Finance’s infectious creativity is again on full display. You as security. This collateral may not bear even a passing resem- can buy ETFs that are sharia-compliant, that invest in clean-ener- blance to the assets that the ETF is ostensibly tracking. And some 1

The Economist February 25th 2012 7 SPECIAL REPORT FINANCIAL INNOVATION

2 regulators worry that banks purposely choose synthetic struc- struments, in BlackRock’s proposed taxonomy), is irreproach- tures so that they can dump their illiquid assets into the collateral ably clear about the risks of compounding. Michael Sapir, its basket and get funding that they otherwise could not. boss, believes that his customers are knowledgeable and use the The fear that some investors may not understand what products appropriately to manage risk or pursue investment op- they are getting themselves into extends to other products. Ex- portunities. change-traded notes, or ETNs, may sound like ETFs, but they are That said, it is dicult to argue against a call for more tran- essentially unsecured debt instruments issued by banks. Anoth- sparency or against the model of innovation governance exem- er set of ETFs that are souped up by leverage and seek to make plied by the debate on ETFs. Regulators congratulate them- daily returns are particularly exposed to an eect called com- selves on having made the industry more introspective and pounding. Imagine a 10% rise on a $100 investment on day one, point out that disclosure practices have improved immeasurably followed by a 10% fall on day two: the value of the investment since they started waving red ags in the spring of last year. Most will end up being $99, not $100, as many people intuitively as- people in the industry seem to think that the end result of this de- sume. Now add in leverage designed to double the movements bate will be some kind of product classication. Among custom- of the investment, so each day sees a 20% swing: the result will ers the wind seems already to be blowing in favour of physical be an investment worth just $96. This compounding eect is also products. Client demand is changing, says Mr Draper at Credit at work in inverse ETFs, which are designed to make money Suisse, which converted four swaps-based funds to physical when markets fall. The risks inherent in leveraged and inverse ones last year. products manifest themselves most in times of volatility. In October BlackRock put itself on the side of the angels by Perpetuum mobile issuing a paper calling for better disclosure around derivatives- It is in the nature of nance that experimentation never backed products, so that investors are clear about the identity of stops, however. So it is with ETFs. The pressure to innovate will counterparties, the composition of collateral and so on, and for intensify as competition increases. The McKinsey report reckons the use of multiple counterparties rather than a single swap deal- that the number of ETF managers in America has grown tenfold er. It also outlined proposals for a formal classication of ex- in the past decade. That guarantees the industry will keep push- change-traded products so that only some instruments can be ing forward with new products. called ETFs. I fear that an exchange- Mr Wiedman believes that there is still lots of room for traded product will break down one of growth in physical ETFs, not just in equities but in xed income, these days and the worry is that it will poi- too. Products could also move into ever more exotic areas in or- son the entire sector, says Mark Wied- der to deliver higher returns, which may yet shove the pendu- man, the head of iShares. lum back in the other direction and require the use of more de- BlackRock’s boss, Larry Fink, has rivatives to replicate the desired exposures. Intriguingly, there is sounded warnings based on his own ex- plenty of talk of active ETFs that would combine elements of dis- perience helping to pioneer mortgage- cretionary stock selection and the tracking of a benchmark. No backed securities. I do believe we have one is sure how this would work, not least because it would re- some responsibility for making sure that quire managers to reveal their strategies to marketmakers, but the market does not morph itself, the same way when I started in the mortgage market 35 years ago, watching a great market Like the Red Queen in Alice Through the Looking-Glass, morph into a monster, he told a confer- the regulators will always have to keep running just to ence in November. Rivals claim that BlackRock’s ap- stand still proach is self-serving: it is one of three dominant providers in America and oers a range of products this is one of those rare areas where patents have been led. made up almost exclusively of physical ETFs. Providers of syn- All of these possibilities will require continuing vigilance thetic ETFs argue that there is plenty of counterparty risk in phys- on the part of nancial watchdogs. Like the Red Queen in Alice ical funds, too, because the funds’ securities are routinely being Through the Looking-Glass, the regulators will always have to lent out to other investors in return for collateral. Derivatives keep running just to stand still. The FSB paper points out that have long been a feature of the ETF market in Europe. They are al- branching out from equities into other asset classes means mov- lowed by the EU’s UCITS fund-management directives, which ing into markets where liquidity is thinner, for example. If an ETF means that the synthetic products are tightly regulated. For ex- is active, presumably sometimes investors do not know what is ample, the rules require the overnight market value of the collat- in their portfolio as managers make discretionary bets: that hard- eral to be at least 90% of the value of the securities. And although ly sounds transparent. Fiercer competition will also encourage much of the debate focuses on the retail investor, this product is providers to make more money from lending securities, which heavily used by institutional investors too. means that even investors in physical products could end up ex- There are also perfectly good reasons to use derivatives posed to rising levels of counterparty risk. when it is hard to run a fund as a physical structure, perhaps be- And all forms of growth will increase the weight of ETFs in cause of restrictions on investing directly in an asset such as com- determining stock prices, a prospect that worries people like Mr modities. The row over synthetic versus physical ETFs is not Litan. He argues that less liquid, smaller stocks already get buet- helping regulators to have a clear view, says Patrice Berge-Vin- ed by wider movements in index-tracking ETFs of which they cent, a French regulator who is preparing a report on the product are constituents, and that buying and selling bundles of stocks for IOSCO, a global body of securities supervisors. It depends leads to excessive correlations between them. Whatever the on each individual ETF as to how dangerous things are. merits of this argument, it opens the door to others: about the ef- Providers of leveraged products, meanwhile, point out that ciency of modern markets, the consequences of passive invest- leveraged mutual funds have existed for more than 15 years. ing, and in particular the role of the most turbocharged nancial ProShares, a provider of leveraged funds (or exchange-traded in- innovation of all: high-frequency trading. 7

8 The Economist February 25th 2012 SPECIAL REPORT FINANCIAL INNOVATION

High-frequency trading ber of contracts, but because of an error in the way the rogue algorithm had been written, this, too, failed to spot a problem. To complete the catalogue of errors, the rm then allegedly The fast and the breached another CME rule when an employee used a col- league’s trading ID to put on positions that would oset its un- furious wanted exposures. This incident was unusual for ending in a ne, but in other respects it was not that uncommon. The ash crash of May 6th High-frequency trading seems scary, but what does 2010, when American equity markets nosedived by almost 10% in the course of a few nerve-shredding minutes, grabbed popu- the evidence show? lar attention. Although it was not directly triggered by high-fre- ON FEBRUARY 3RD 2010, at 1.26.28 pm, an automated trad- quency traders, the ocial reports suggested they helped fuel ing system operated by a high-frequency trader (HFT) the uncontrolled selling. But there are miniature versions of such called Innium Capital Management malfunctioned. Over the ash crashes happening all the time, says John Bates, the chief next three seconds it entered 6,767 individual orders to buy light technology ocer of Progress Software, a software rm. sweet crude oil futures on the New York Mercantile Exchange Often they result from algorithms interacting with each (NYMEX), which is run by the Chicago Mercantile Exchange other and forming an innite loop. For a simplied example, take (CME). Enough of those orders were lled to send the market jolt- two algorithms that are both programmed always to outbid the ing upwards. best oer in the market, so they will go on outbidding each other. A NYMEX business-conduct panel investigated what hap- Usually such loops are spotted and stopped, sometimes man- pened that day. In November 2011it published a list of Innium’s ually and sometimes automatically, without many people notic- alleged risk-management failures and ned the rm $350,000. ing. But the fact that they happen at all feeds the perception that Innium itself neither admits nor denies any violation of the ex- today’s equity markets have turned into something more akin to change’s rules. It takes the same line on a $500,000 ne it was science ction than a device for the ecient allocation of capital. given at the same time for alleged transgressions on the CME it- How, the critics ask, can anyone assess the fundamentals of a self in 2009. company in a fraction of a second? How can lumbering institu- Those alleged failures pull back the curtain on some of the tional investors, let alone the little guy, hope to compete with the safeguards that are meant to protect traders, exchanges and mar- HFTs? And what is to stop a new set of algorithms from unleash- kets from erratic ultra-fast algorithms. The NYMEX panel found ing havoc? that Innium had nished writing the algorithm only the day before it introduced it to the market, and had tested it for only a Science v friction couple of hours in a simulated trading environment to see how it Such questions have gradually drawn the high-frequency would perform. The rm’s normal testing processes take six to traders out into the open. Until recently they saw little need to eight weeks. When the algorithm started its frenetic buying engage with the wider world. HFTs do not have clients but oper- spree, the measures designed to shut it down automatically did ate with their own capital. Proprietary algorithms provide a not work. One was supposed to turn the system o if a maxi- competitive edge, so secrecy is engrained in the culture. But as mum order size was breached, but because the machine was regulators, politicians and the media focus ever more closely on placing lots of small orders rather than a single big one the shut- their activities, the traders have formed groups on both sides of down was not triggered. The other measure was meant to pre- the Atlantic to represent their interests. vent Innium from selling or buying more than a certain num- Many are frustrated by what they perceive as an unfair on- 1

The Economist February 25th 2012 9 SPECIAL REPORT FINANCIAL INNOVATION

eeting trading advantage. But the legitimate explanation for it is Invest, don’t protest 4 that marketmakers cannot aord to be static in case the market S&P 500 median bid-ask spread, $ cents moves against them, and that in an ever-faster market HFTs have to be quicker to adjust prices. 3.0 We have got to get away from the idea that speed equals 2.5 danger, says Richard Gorelick, the G in a Texan HFT called RGM. Professional traders trade continuously and are exposed 2.0 to market movements all the time, so being able to adjust that ex- 1.5 posure quickly gives them condence to quote better prices. That is why the idea of imposing minimum resting times for 1.0 quotes before they can be cancelled would almost certainly lead 0.5 HFTs to widen bid-ask spreads, increasing costs to investors. In their 2011 NBER paper Mssrs Lerner and Tufano argued 0 2003 04 05 06 07 08 09 10 11 that it is virtually impossible to quantify the social impact of a - Source: Knight Capital Group nancial innovation because nance involves so many external- ities, often unintended ones. For example, it would be almost impossible to measure the aggregate costs and benets of a 2 slaught. The gap between reality and rhetoric is bigger for this fundamental innovation like a bank. Instead, they reckoned, a industry than for any other I have seen, says Remco Lenterman, thought experimentimagining what the world would look like the chairman of the European Principal Traders Association and without a particular innovationmight help. the managing director of IMC, an electronic marketmaking rm. A world without HFTs is both easy and very dicult to Plenty of academics support the HFTs’ arguments. On the other imagine. Easy, because the old world of specialist marketmakers side are some big institutional investors who accuse HFTs of and oor trading existed only a few years ago, so people remem- front-running their orders, and high-prole critics like Bart Chil- ber it well. There is little obvious enthusiasm for returning to that ton, a member of America’s Commodities Futures Trading Com- model. Not only were transaction costs higher but the same ar- mission, who has punningly dubbed HFTs cheetah traders. guments about unfair advantages were being put forward in dif- To sift through the arguments on both sides is to confront a ferent forms. Now the complaints are about the milliseconds basic problem with any nancial innovation: the diculty of measuring its benets. For one thing, there are questions of To sift through the arguments on both sides is to denition. HFTs are not the only people confront a basic problem with any nancial innovation: using algorithms to trade: institutional investors use them to break large orders into the di culty of measuring its benets small parcels so that markets do not move against them as they execute the order. And although high-fre- HFTs gain over ordinary investors by putting their servers right quency trading always involves very brief holding periods and next to the exchanges’ data centres; then they were about the mo- very active trading, it breaks down into lots of dierent strat- nopolistic privileges of the specialists and the advantages of be- egies. Some HFTs are momentum traders, riding the wave of a ing on the oor. Institutional investors may complain about be- particular trend. Others arbitrage price dierences. Others still ing forced into dark pools (o-exchange venues where they are marketmakers providing liquidity to buyers and sellers. can deal anonymously) to avoid HFTs, but these pools existed Another problem is that there are not enough good data. before HFTs and were set up in part to avoid being scalped by The ercest debates centre on the role of HFTs as marketmakers. brokers or oor traders. The evidence tends to favour the HFTs, which can point to a solid But imagining a world without high-frequency trading is body of academic research that shows they increase liquidity, as also hard. That is because the HFTs are what Larry Tabb of TABB measured by tighter bid-ask spreads (see chart 4). HFTs also point Group, a research rm, describes as an outcrop of the market to testimony delivered to the Securities and Exchange Commis- structure. They are a natural outcome in a world in which trading 1 sion in 2010 by George Sauter of Vanguard, a big fund manager, who concluded that high-frequency traders provide liquidity and ‘knit’ together our increasingly fragmented marketplace, re- Rise of the machines 5 sulting in tighter spreads that benet all investors. Algorithmic trading, % of total trading But others say that the increase in trading activity brought about by HFTs, in Europe at least, means that fund managers 70 70 Equities have to pay additional costs for storage and electronic reporting 60 60 in order to comply with best-execution requirements. It is hard to F’CAST HFT 50 50 disentangle the eects of s on transaction costs from other Futures changes, such as competition among exchanges. A bigger pro- US 40 40 blem, says Paul Squires, the head of trading for AXA Investment Options Managers, is that increased liquidity can be illusory. You can 30 30 Foreign press the button to buy Vodafone, say, and have it executed in a exchange 20 20 second but in that period 75% of the liquidity has disappeared Europe Asia Fixed and the price has moved. 10 10 It is certainly true that HFTs are constantly sending and can- Income celling orders. Some of that activity may be tied to a manipula- 0 0 tive technique called quote-stung, in which a ood of orders 200405 06 07 08 09 10 11 * 12 13 14 2004 06 08 10 Source: Aite Group *Estimate and cancellations causes congestion on networks and thereby a

10 The Economist February 25th 2012 SPECIAL REPORT FINANCIAL INNOVATION

2 is automated, and in which there is competition between lots of Financial infrastructure dierent exchanges and a need for someone speedily to knit together the prices they oer. The real question is whether humans make worse mistakes when they write algorithms or Of plumbing and when they trade, says Terrence Hendershott of the Haas School of Business at the University of California, Berkeley. promises In practical terms, trading history is highly unlikely to be re- versed. Regulators in developed countries have no evidence that radical change is needed, nor any appetite for it. Developing The back oce moves centre stage countries, which have become the standard-bearers of sensible nancial regulation, are racing towards high-frequency trading FAILURE IN THE nancial crisis had many fathers. There as they seek to improve liquidity (see chart 5, previous page). were failures of regulators, bankers, shareholders, borrow- Meanwhile the industry itself pushes inexorably forward. ers and economists. But two in particular were closely tied to the That certainly entails greater speed: the industry used to think in way nancial innovations work. One was a failure of plumb- terms of milliseconds (it takes you 300-400 of these to blink) but ingthe infrastructure of the markets and the back oces of - is now fast moving to microseconds, or millionths of a second. It nancial rms. The other was a failure of the imagination. also means smarter algorithms. People have gone from trading Infrastructure often lags when an innovation takes o. Re- in open-outcry pits to trading via screens to programming algo- member how markets move over time from being customised to rithms. The next stage, says Mr Bates of Progress Software, will be becoming more standardised. When products are standardised self-learning systems, in which sentient algorithms mine the and demand is high, nance’s manufacturing, sales and distribu- capital markets, spotting correlations that are too complex for tion channels can pump out a vast supply. The frightening humans to see and suggesting trading ideas as a result. Humans growth of over-the-counter derivatives markets attests to that will still be needed to validate these ideas, he says reassuringly. (see chart 6). But the prospect of even faster markets raises the problem The boring detailsof trade conrmations and settlements, posed by the Innium case and by ash crashes large and small: of collateral requests and contractual negotiationtend to get the threat from HFTs to the stability of markets. Software has a left behind in the rush to make money. Lawyers recall how nasty habit of developing bugs. Algorithms behave unpredict- boom-time contracts for commercial mortgage-backed securi- ably once they are out of a testing environment and into the mar- ties would be hurried through, often by junior associates with ket proper. Even leaving aside the potential for deliberate market little knowledge of the product. The far bigger market for resi- abuse, traders will sometimes get things wrong. dential mortgage-backed securities was even sloppier at record- keeping, causing uncertainty about which lenders had formal ti- Plugging the holes tle to properties when the housing bubble burst. So real-time risk-management safeguards have to be put in The back oce is attached to the front oce by a bungee place that work at dierent levelsof the HFTs, the prime bro- cord, and depending on how fast the front oce is running, the kers, the exchanges and the regulatory agencies. Some of these cord gets stretched, says Mark Beeston, who runs a portfolio of safeguards already exist. The exchanges have limits on orders post-trade businesses for ICAP, an interdealer broker. These above a certain size, for instance, and on the number of orders businesses include Traiana, the post-trade risk manager for HFTs that can come in. But there are always holes to be lled. encountered in the previous section, as well as a number of The May 2010 ash crash revealed a glaring problem with rms devoted to a market that truly took o before the crisis: that the structure of American equity markets, for instance: that cir- for credit-default swaps (CDS). Indeed, many people think that a cuit-breakers which automatically pause trading if there are viol- regulatory intervention to try to sort out a horrible paperwork ent price swings kicked in only once the entire market reached a mess among CDS buyers and sellers prevented the crisis from certain threshold. The regulators have since introduced single- becoming even worse. stock circuit-breakers for any stock that moves up or down by By 2005 the infrastructure of the CDS market was 10% or more in a ve-minute period. Most observers are con- swamped. Harmonisation of CDS contracts meant that, in the dent that this would stop another ash crash in its tracks. jargon, they were easy to novate, or transfer to another party. However, a focus on equity markets may distract attention Trading volumes soared. A particular worry was an enormous 1 from 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 of All you can eat 6 high-frequency trading in the vast foreign-exchange markets. It Derivatives, notional amounts outstanding, $trn worried about the capacity of the prime brokers, through which HFTs gain access to credit, to keep pace with their clients. If they 800 do not, the prime brokers themselves are exposed to the pos- sibility of HFTs rapidly accumulating risky positions. Exchange-traded options Exchange-traded futures 600 Again, the industry is aware of this risk. Traiana, a post- Foreign exchange trade processing rm, launched a software program last year that Interest rate † aggregates clients’ positions across lots of dierent venues in real CDS* OTC 400 time and activates a kill switch that stops clients trading once Other OTC pre-dened limits are breached. But this is a voluntary initiative, 200 not one that is required by regulators. Other markets remain un- protected. Regulators should not be afraid to act rmly to dene and enforce standards for market surveillance and trading con- 0 trols across venues and asset classes. And getting the infrastruc- 1998 2000 02 04 06 08 10 11 ture right is important in other areas too. 7 Source: Bank for International Settlements *Credit-default swap †Over the counter

The Economist February 25th 2012 11 SPECIAL REPORT FINANCIAL INNOVATION

2 backlog of unconrmed trades, which in the event of a default would 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 headquarters and knocked heads together. The industry began moving towards electronic conrmations and data repositories to record trades, and the backlogs started to reduce. Everyone knew that it was a paper mess but no one wanted to sign up to a single standard, recalls one participant in that meeting. Regulatory intervention will clean up markets incredibly well. By the time the 2008 tornado hit, the CDS market functioned pretty eciently, at least in an operational sense. The process for settling CDS claims after Lehman Brothers went bust, in particular, passed o more smoothly than expected. But there is only so much room for congratulation. In one crucial 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 against contract are posting collateral every day as prices uctuate. Un- default. Although the 2005 intervention did a lot of good, it did less a default comes totally out of the blue, that can cushion the not deal properly with bespoke, over-the-counter (OTC) trades, impact of a sudden credit event. AIG, notoriously, did not post the bit of the market where AIG was operating. enough collateral until it was suddenly asked for unaordable CDS contracts oer protection against an event (ie, default) amounts, and seemed to be banking on not having to pay out. rather than a move in prices, as happens in markets such as inter- Belatedly, nancial reform is dealing with these infrastruc- est-rate swaps. That means obligations can be triggered sudden- ture aws, and not just for credit derivatives. Many more swap ly, a phenomenon known as jump-to-default risk. The best transactions are heading for clearing houses, which stand be- way to protect against this risk is to ensure that the parties to the tween buyers and sellers and collect collateral centrally. Records 1

On the side of the angels

New ways of lending to small businesses IF THERE IS one bit of nance where people called Crowdcube, which uses the idea of rates are competitive and rms get hold of agree on the need for more innovation, it is crowdfunding to enable lots of investors the money within about two weeks. Samir in lending to small business. Policymakers to buy up small stakes in start-up rms. Desai, a co-founder, dismisses the argument are desperate to get more credit owing to Picking winners among entrepreneurs that lending to small rms requires bankers this vital engine of economic growth. Banks is notoriously dicult. Venture capitalists’ to make personal credit assessments: SMEs claim that lending is muted because de- answer is intensive screening by a small want low costs, a quick process and a trans- mand is subdued, but that is not the only team of dedicated investors, followed by parent fee structure, not a relationship. problem. Small and medium-sized en- hands-on involvement in the business. The Lenders are encouraged to spread their risk terprises (SMEs) are harder credit risks to Crowdcube model, which is due to come to among at least 20 borrowers. assess than large ones, so they attract America if crowdfunding legislation passes, There is another way of speeding up higher capital charges and are often the depends on the ability of thousands of the underwriting process: taking a bet not rst to lose their funding in a downturn. members to ferret out the best ideas. The on an SME itself but on its debtors. The A host of new rms have sprung up general public cannot match the expertise Receivables Exchange, launched in New with solutions. Some are seeking to ll the and commitment of dedicated angel Orleans in 2007, enables investors to buy up gap left by the banks, rather than overhaul investors if a rm gets funded, admits invoices, or fractions of them, owed to small the way that lending is done. Shawbrook is a Darren Westlake, Crowdcube’s founder. But businesses. The idea is similar to factoring, new, specialised lender to small rms in it helps to have lots of investors acting as whereby rms sell o all their invoices at a Britain, where the dominance of a handful advocates for a start-up rm. discount to improve their cashow. But the of big banks makes the choice for SMEs The same peer-to-peer model lies idea behind The Receivables Exchangeand particularly limited. Owen Woodley, Shaw- behind Funding Circle, another British MarketInvoice, a British equivalentis to brook’s boss, says that it can get its credit start-up which launched in 2010 to facilitate break receivables down into small, tradable analysis done faster than the established lending to small businesses for periods of units so that buyers can make judgments on institutions. one to three years. Businesses go through individual debtors and diversify their hold- Other rms are trying to reinvent an underwriting process before they get on ings. We provide electron-level transpa- small-business nancing by providing to the company’s website, where lenders, rency, says Nic Perkin, a co-founder of The virtual marketplaces where investors and predominantly individuals, can bid on the Receivables Exchange. Transactions are SMEs can come together. In the world of rate at which they are prepared to lend. The somewhat less minuscule, approaching a equity capital the pacesetter is a British rm average loan is £40,000 ($63,000), the rate of $1billion a year.

12 The Economist February 25th 2012 SPECIAL REPORT FINANCIAL INNOVATION

2 of all cleared and OTC trades across all the and buy them enthusiastically. When those risks materialise, big classes of derivatives will be collected there is a destabilising ight to safety. The standard argument and stored in so-called swap-data reposi- for nancial innovation is that there are gains from trade, but that tories, so the chances of anyone building model crumbles if you suppose that people do not fully under- up a big undisclosed position should be stand the risks, says Mr Shleifer. His main bugbear is the money- sharply reduced. Much of this is well un- market mutual fund, which oers instant liquidity and promises der way. We are very close to a more com- no risk to investors’ principal. It saw a massive run when one of prehensive view in terms of systems en- them broke the buck in 2008. abling regulators to see the trading This analysis rings true of much of nance: people are lia- universe in intra-day updates, says Peter ble to forget about the risks of products that have already blown Axilrod of the Depositary Trust and Clear- up as well as misjudge those that have been newly created. The ing Corporation (DTCC), the world’s larg- euro zone’s debt crisis has shown that risks even in long-estab- est data repository. lished instruments like government bonds can be underestimat- But some argue that having a bigger ed. But innovations are particularly susceptible to the problem haystack of data makes it harder to nd of self-delusion. If they go wrong early enough, they are unlikely the really important needles. Others want to get o the ground. But once they reach a sucient scale with- a lot more information. No national regu- out a big blow-up, nobody believes that they might be awed. 7 lator can see all of the nancial system: an American regulator can see the CDS exposures of American banks to French banks, for instance, but is not Collateral allowed to see the counterparty risks of the French lenders in turn. And there is critical work to be done to gather data on what collateral is being posted as security on trades and to pull togeth- Safety first er an aggregate picture of the exposures being taken across derivatives classes. Even so, things are far better than they were. We will be in a position in ten years where we have an amazing derivatives infrastructure, says Edmund Parker, a derivatives partner at Mayer Innovative ways of making lenders feel more secure Brown, a law rm. In the meantime the attention of regulators is broadening out to other bits of nancial scaolding. There is talk PUT TOGETHER ALL these aspects of nancial innova- of taking a closer look at the way the securities-lending business tionexperimentation, standardisation, infrastructure works, for instance. An American task force has been at work for gaps and the illusion of safetyand one area of the post-crisis - over two years looking at an obscure but vital area of short-term nancial landscape ashes red as a potential source of problems: nancing called the tri-party repo market. Margining and tech- collateralisation, or giving a lender an asset as security in case a nical policy and back-oce monitoring of positions against col- borrower defaults. lateral are unsexy but it is the stu to be focused on, says Mo- Demand for collateral, at least in areas like OTC derivatives, hamed Norat of the IMF. is growing. In Europe in particular, bank creditors are pushing hard towards secured funding to protect themselves in the event The strangeness of comfort of trouble. A structural shift to collateralised funding is going This is particularly true when it comes to innovations that on, and these forces are very strong, says Imène Rahmouni- pledge to transfer or reduce risk. Many of the instruments and Rousseau of the FSB. Moreover, perceptions of what counts as techniques that were most lauded before the crisis were de- good collateral have changed. Before the crisis residential mort- signed to package risk and shift it away from people who did not gage-backed securities of all kinds were being widely used in want it towards those who did. More transparency might have repo transactions (a form of short-term funding); now the em- made it clear that risk was simply being concentrated some- phasis is on nding highly rated government bonds, itself a where else, or was not really leaving banks’ balance-sheets at all. shrinking universe. A recent IMF paper estimated that a decline This weakness in infrastructure compounded a behaviour- in the amount of pledged collateral, the sort that can be reused al one. Finance is at its most dangerous when it is perceived to be in other transactions, has re- safe. One element in the nancial crisis was a failure to under- duced the overall availabil- stand the risks inherent in various products until it was too late. Squeezed 7 ity of collateral by $4 tril- The resulting fragility is described in a 2010 paper on nancial in- Pledged bank collateral, $trn lion-5 trillion since pre-- novation by Nicola Gennaioli of Pompeu Fabra University, An- Lehman days (see chart 7). drei Shleifer of Harvard University and Robert Vishny of the Europe & Nomura US This collateral squeeze University of Chicago. The authors suggest that many episodes 6 is fuelling all sorts of innova- of nancial innovation (including the excessive issuance of 5 tion. One which has already mortgage-backed securities) start in the same way: with inves- attracted the attention of tors’ demand for a set of cashows that do not carry much risk. 4 regulators is the liquidity There is only a limited number of existing instruments of that 3 swap, whereby holders of sort at a reasonable price. So the nance industry meets the de- liquid assets, such as insur- mand by creating new instruments that are designed to be as 2 ers and pension funds with safe as existing ones (think of those AAA ratings on diversied 1 government bonds, lend bundles of loans). them to banks to use as col- 0 Investors do not necessarily think through all the risks em- 2007 08 09 10 lateral in their secured fund- bedded in these new instruments (for example, that a national Source: “Velocity of Pledged Collateral”, ing. In return, banks lend housing bust would render the tranching within CDOs useless) by Manmohan Singh, IMF, November 2011 them less liquid assets and 1

The Economist February 25th 2012 13 SPECIAL REPORT FINANCIAL INNOVATION

2 pay a fee. Such deals are not entirely new but there are concerns and complacency. For unse- that they could quickly grow in scale, binding the insurance and cured creditors such as bond- Offer to readers banking sectors more closely together. Britain’s Financial Ser- holders, the concern is a pro- Reprints of this special report are available. A minimum order of five copies is required. vices Authority issued some disapproving guidance about the blem known as encum- Please contact: Jill Kaletha at Foster Printing practice last summer and has reportedly blocked a number of brance. It means that the more Tel +00(1) 219 879 9144 planned swaps since. of a bank’s balance-sheet is tied e-mail: [email protected] If one path is blocked, bankers will try to nd others. Chris- up as collateral, the fewer assets Corporate offer topher Georgiou of Ashurst, a law rm, says that banks are creat- will be available to them in the Corporate orders of 100 copies or more are ing guaranteed repackaging structures in which assets are event of a default. Again, data available. We also offer a customisation pledged as collateral and guaranteed by the bank. are in short supply. Fitch, a rat- service. Please contact us to discuss your American regulators are also tracking attempts to launch a ings agency, pointed out in De- requirements. new type of funding instrument called collateralised commer- cember that European banks Tel +44 (0)20 7576 8148 e-mail: [email protected] cial paper, which some think has the potential to replace asset- are both increasingly reliant on For more information on how to order special backed commercial paper as a big source of nancing. It works secured funding and reluctant reports, reprints or any copyright queries by giving investors a claim over collateral used in certain repo to disclose which assets have you may have, please contact: agreements without breaching new rules that prevent money- been pledged as collateral. For The Rights and Syndication Department market funds from investing directly in such repo transactions secured creditors, the risk is that 26 Red Lion Square themselves. The market has not yet caught re, but supervisors they will feel safer than they London WC1R 4HQ are struck by how quickly the industry has started innovating in really are: valuation method- Tel +44 (0)20 7576 8148 Fax +44 (0)20 7576 8492 this area. I am amazed that these things were developed before ologies may not be right, for ex- e-mail: [email protected] the rules were even in place, says one. ample, or assets may not be www.economist.com/rights Some banks see an opportunity to be collateral intermedi- properly segregated. aries. A clearing house may be wary of taking corporate bonds As in other areas of nan- Future special reports Nuclear energy March 10th directly from a pension fund as collateral, say, but a bank can take cial innovation, the best way of Cuba March 24th the bonds and then issue its own letter of credit to the clearing keeping the collateral problem Manufacturing April 21st house. In an echo of the triparty repo market, triparty collateral under control is one of intrusive International banking April 21st management is also becoming more popular. As Rajen Shah, vigilance. That does not mean head of collateral management at Citigroup, explains, under this banning everything in sight. Fi- Previous special reports and a list of forthcoming ones can be found online: system assets used as collateral sit with a third party. If brokers nance has a very good record of economist.com/specialreports want to change the assets they pledge during the day, they can do solving big problems, from en- so more quickly and cheaply via the agent abling people to realise the val- than by managing collateral bilaterally. ue of future income through products like mortgages to protect- Within banks, too, collateral man- ing borrowers from the risk of interest-rate uctuations. It is agement has moved from the back of the tempting to choke o creativity in the aftermath of the crisis, but back oce to a much more prominent po- there is no obvious way of sifting the good innovations from the sition. Institutions are investing in sys- bad at the outset, and far less chance for the industry to mitigate tems that provide them with a central problemssuch as the impact of rising longevity, or ineciency view of what collateral is being held by in public spending, or the lack of credit for millions of poor peo- dierent desks and allow them to move it ple in emerging marketsif it cannot experiment. around eciently, to specify the order in The problems come along after the experimentation phase, which collateral is put to work during the when products have taken root and are growing wildly. The evi- day and to model the correlation risks em- bedded in dierent types of collateral (ie, if a counterparty defaults, whether its col- The market does a brilliant job of nurturing and rening lateral is likely to sour, too). Collateral de- instruments that people want. It is less good at partments were the orphaned, redheaded stepchildren of the organisation, says planning for when things go wrong one executive. Now they are seen as the most interesting entrepreneurial and commercial opportunity. dence of this special report suggests that the market does a bril- In many ways, the collateral craze exemplies the chal- liant job of nurturing and rening instruments that people want. lenges of judging nancial innovation. The idea of using collater- It is less good at planning for when things go wrong. This is al to provide added security for the creditor is a sound one: it where regulators can play a benign role, signalling concerns ear- should reduce the riskiness of lending and the chances of a run. ly and loudly when activities reach a certain scale (as has hap- Such reasoning is reected in the lower capital charges that pened with ETFs and liquidity swaps) and setting standards for banks incur for secured lending. Much can be done to improve data and reporting before the infrastructure starts to creak (as the eciency of collateral-management systems, and a rethink is happened with CDSs in 2005). needed on what counts as a safe asset. Regulators are far from perfect. They proved incapable of But the whirring of nanciers’ minds also spells trouble. As spotting danger before the crisis, and will doubtless make simi- collateral ows more eciently through the system and ways lar mistakes in the future. And nance will still have many old- are found to transform and enhance it, the regulators have a big fashioned ways to blow itself up. That is why it is so important to job to keep up. Working out which collateral belongs to whom, have lots of capital in the system to head o a calamity. But if in- whether it is fairly valued, and what would happen at a systemic novation can be encouraged without being given licence to run level if there was a big margin call on a particular asset class are riot, there will be a better chance of the industry doing good in- all huge tasks for the future. stead of harm. 7 There are also the familiar problems of interconnectedness

14 The Economist February 25th 2012