Black boxes

May 17th 2007
From The Economist print edition

Investment banks' inventions for transforming risk are ingenious, but hard to fathom


ADVANCES in technology and information have always been good for capital markets. Richard Sylla of New York University's Stern School of Business argues that financial globalisation started earlier than previously thought, around 1816. At that time packet sailing ships began to make regular scheduled crossings between New York and British ports, cutting the average time it took for mail—and information about securities prices—to cross the Atlantic. Later that century the telegraphic transatlantic cable, along with steam shipping, heralded the huge growth of global markets between 1870 and 1914.

The computer is the cable of our age, and it has turned investment banks into hothouses of financial innovation. The financial principles are not new: parallels have been drawn between the spectacular growth of credit derivatives in recent years and the introduction of wheat futures in America in the mid-19th century. But thanks to computing power the products have become infinitely more malleable. They are designed to appear complex and impenetrable. But they are not patented, and staff turnover across the investment-banking industry is estimated to be up to 20%, so ideas travel quickly. Margins fall, products go through the inevitable boom-and-bust cycle and the whole thing starts all over again.

The peephole into this mysterious world of debt is a single line in many investment banks' trading accounts known as FICC, for fixed income, currencies and commodities. For the five big Wall Street firms (Goldman Sachs, Morgan Stanley, Merrill Lynch, Lehman Brothers and Bear Stearns) taken together, these revenues have quadrupled since the start of this decade, according to Dominion Bond Rating Service, a rating agency (see chart 3). Back in 2000, when stockmarkets were booming, share-trading generated twice as much revenue as did fixed income. But it has been growing far more slowly, to only $27 billion last year, against $44 billion for fixed income.

From copper to catastrophe
FICC encompasses a broad swathe of assets, from American subprime mortgages to Japanese yen, copper futures to catastrophe insurance, General Motors bonds to Zambian debt. Some of the fastest growth has been in tried-and-tested asset-backed securities such as commercial and residential mortgages, which have soared since 2000 whereas straight company debt issuance has stagnated (see chart 4). But the most profitable area has been the growth of derivative and structured credit products, such as CDSs and CDOs.

These have enabled banks to separate credit risk from interest rates and trade that risk among those who want to hold it and those who don't. This process has freed credit risk from the underlying bonds, leading to an explosion of secondary-market activity. For the banks it has been a goldmine, but for outsiders it is devilishly hard to know how much risk the new instruments entail.

The cornerstone of the new market is the CDS, a form of insurance contract linked to underlying debt that protects the buyer in case of default. The market has almost doubled in size every year for the past five years, reaching $20 trillion in notional amounts outstanding last June, according to the Bank for International Settlements. That makes it far bigger than the underlying debt markets. Some 70% of these products are linked to an individual issuer and are not much more complex than selling a bond short. The fun starts with the other 30%. The mathematicians of Wall Street and London have found ways of bundling indexes of CDSs together and slicing them into tranches, based on riskiness and return. The most toxic tranche at the bottom exposes the holder to the first 3% of losses but also gives him a large portion of the returns. At the top, the risks and returns are much smaller—unless there is a systemic failure.

CDOs grew out of the market for asset-backed securities which took off in the 1970s and encompassed mortgages, credit-card receivables, car loans and even recording royalties. The structured CDO is a more complex variation, using lots more leverage and bundling bonds, loans and CDSs into securities that are sold in tranches. According to the Bond Market Association, an industry body, $489 billion-worth of CDOs were issued last year, twice the level in 2005. One-third were based on high-yield loans and are known as collateralised loan obligations (CLOs). The rest involved mortgage-backed securities, CDSs and even other CDOs (which become supertankers of leverage, known as CDO2 and CDO3).

Investment bankers are fascinated by the possibilities created by structured products, which they think offer almost limitless ways for their clients to manage risks. Some of the more ambitious are working on risk-transfer instruments that deal with weather, freight, emissions, mortality and longevity (see article). The most immediate opportunities, though, may be in asset classes—such as property derivatives—that have already proven successful in America but are still emerging in Europe and barely exist in developing countries.

Anshu Jain at Deutsche Bank says over half of the firm's trading profits last year came from this ill-defined area, which he loosely calls “intellectual capital”. About one-third of the business was done with hedge funds and the other two-thirds with financial institutions needing to match assets and liabilities, such as insurance companies. He believes that this aspect of trading may be less cyclical than other sources of revenue, such as proprietary positions.

Matt King, a credit strategist at Citigroup, says that demand for CDOs has lately been stimulated by the approaching implementation of the Basel 2 capital accord, which encourages banks to swap risky loans on their books for CDO tranches to avoid high capital charges. This has been helped by the increasing willingness of banks to sell loans into the capital markets in order to diversify their portfolios. Huw van Steenis of Morgan Stanley estimates that some 78% of senior secured loans in America have now been sold in this way, compared with 29% in 1995. In Europe 53% are now securitised, up from 12% in 1999, still leaving considerable room for expansion. Alessandro Profumo, boss of UniCredit, a booming bank that has stitched together an investment-banking franchise across Italy, Germany, Austria and central and eastern Europe, says it sells €40 billion-50 billion ($54 billion-68 billion) of loans a year and would eventually like to make it a lot more.

Maria JeevesSuch financial innovation not only spreads the risk for investment banks' clients. It may also make the banks themselves more diversified as they search for the holy grail of finance: unrelated portfolios that do not all melt down at the same time. There is little correlation between credit and government bonds, for example. When credit spreads widen, investors flock to “safe haven” products such as government bonds.

That is the theory, at least. But there are doubts, which may help explain the low multiples on investment banks' earnings. Guy Moszkowski, securities industry analyst at Merrill Lynch, describes the institutions as “black boxes”, revealing little about their overall exposures. A senior banker agrees. “There is a total lack of transparency. We make it impossible for our investors and analysts to understand what is going on.” He thinks that investors would be more impressed than scared if there were more disclosure, but accepts that it is difficult to reveal all unless competitors do the same.

Astronomical leverage
CDOs and CDSs generate other concerns too. One is the leverage embedded in them, which does not appear on banks' balance sheets. A senior European financial regulator draws particular attention to the way banks value deeply subordinated portions of CDOs. “They offer very attractive returns, but could fall off a cliff if something goes wrong,” he says. Lombard Street Research, a firm of economic analysts, calculates that such parts of a CDO might be geared up about nine times, and that the leverage will be many times greater still if a hedge fund has invested in them with borrowed money. The leverage becomes “astronomical”, the firm reckons, when the CDOs are based on other CDOs.

The second concern is the possible need to pay out on CDSs if a wave of defaults were to hit. The attraction for the CDS seller, as with any insurance contract, is receiving a steady stream of payments for a risk it hopes will not occur. If its calculations are wrong, the result can be devastating. The BIS says the best way to gauge the total volume of CDS exposures is by their gross market value, which was $294 billion last June. That is a small fraction of the $20 trillion in notional amounts outstanding, and the whole lot would never blow up at once. But there would still be tidy sums involved if the sellers—including investment banks—ever had to pay up.