The dangers of derivatives
Derivatives have brought the market down once before. Given the state of investors’ understanding of these complex financial vehicles, they could do so again.
Warren Buffett once referred to derivatives as financial weapons of mass destruction'. Many pundits have gone as far as to warn that these instruments could bring down the financial system. But few people understand the alphabet soup of CDOs, CDSs and ABCDSs that appear regularly in the financial news. So what are derivatives and just how dangerous are they?
What are derivatives?
The first question is easily answered; the second less so. Derivatives have been around for centuries, with crude versions stretching back two millennia or more. They first evolved as a way for farmers and merchants to manage the risk of crop prices moving against them; a farmer who wanted to be certain what price he would get for his grain at harvest could use forwards or futures contracts to lock in a price in advance. In time, speculators became a crucial part of derivatives markets. They had no interest in hedging their own exposure, but instead used derivatives to bet on the prices of agricultural commodities.
The same basic principles apply today, but derivatives have become far more complex than simple punts on pork bellies. There are derivatives that allow people to bet on and hedge against movements in currencies, interest rates, shares, debt, property and more besides. The market is vast; the International Swaps and Derivatives Association (ISDA) says the total notional amount of derivatives outstanding was $236trn at the end of 2005, up 20% on 2004 (although "gross credit exposure" the amount at risk on the contracts at any point is probably less than 5% of that, according to the Bank for International Settlements).
There are several important points to bear in mind with derivatives. The contracts themselves are a zero-sum game for every winner there must be a loser. Many are enormously complex, requiring advanced mathematics to get a handle on how they work. The hedge funds and banks that trade them make extensive use of leverage trading with a very small amount of their own capital and a very large amount of borrowed money to juice up returns. All this can make derivatives extremely risky and there's plenty of evidence that supposedly sophisticated, professional investors often don't fully understand how much risk they're taking on.
The dangers of derivatives: Long-Term Capital Management
The most famous illustration of this was Long-Term Capital Management (LTCM), a hedge fund that has become the posterchild for everyone who fears derivatives. LTCM was founded by John Meriwether, formerly the head of bond trading at Salomon Brothers; the stellar line-up included two Nobel economics laureates. If anyone should have understood derivatives, it was this team.
LTCM's strategy was to uncover small pricing anomalies in the market. It would buy underpriced assets, short-sell similar but overpriced assets, wait for the discrepancy between them to close and then pocket the difference.The discrepancies were small, so it used highly leveraged derivatives to boost returns. At first, all went to plan. Discrepancies were found, LTCM invested, the valuation gaps closed, and LTCM pocketed the profit. Then the Asian crises struck. Investors fled risky assets for higher-quality, safer ones. Stockmarkets became more volatile and credit spreads on riskier bonds widened (a credit spread is the difference between a bond's yield and the yield on a risk-free bond, such as US government bonds). LTCM saw this as a huge opportunity; it bet heavily that things would soon bounce back.
They didn't. In fact, crisis followed crisis, culminating in Russia defaulting on its domestic debt. LTCM took huge losses, multiplied by its use of leverage up to $550m in one day alone. Its brokers demanded cash to cover its liabilities cash the fund didn't have. In September 1998, LTCM was on the edge of collapse and shockwaves ran through the financial world. The fund had such large liabilities in so many markets that some felt that if it was unable to settle its debts, the outcome would threaten the stability of the financial system. The New York Federal Reserve didn't wait to find out; it persuaded 14 banks to inject $3.6bn into LTCM to recapitalise it, making it possible to wind up the fund in an orderly fashion.
How severe would the fallout have been had LTCM been allowed to collapse?
The money at stake was far greater than LTCM's capital of around $4bn; the fund used leverage of 25 times, meaning it controlled positions worth around $100bn. If LTCM had defaulted on these, then other investors wouldn't receive payments they were counting on to offset their other liabilities. They might have gone under as a result, while brokers who had extended loans to LTCM would find themselves nursing heavy losses.
The big question was whether these problems would have cascaded through the markets, with losses being multiplied and amplified by the widespread use of derivatives and leverage. Investors might panic and liquidity dry up. Eventually, one or more of the core elements of the financial system, such as the big banks, might be undermined. Would this have happened? No one knew then or knows now; the situation was so complex that the possible lines of cause and effect were incalculable. All that's certain is that the authorities were worried enough to bail LTCM out.
The growth of credit derivatives
But markets have short memories. There have been no more incidents on the scale of LTCM and the growth of derivatives has continued unchecked, with total notional value quadrupling since 1998. Much of this growth has been in credit derivatives. It's no exaggeration to say these products have revolutionised global finance and added hugely to the world's stock of acronyms.
Credit derivatives allow credit risk to be shared among many more investors than in the past. For example, a bank that issues mortgages would once have had to accept the risk of lots of its mortgagees defaulting. But today it can use credit derivatives to pass that risk onto other investors very often hedge funds.
The same thing happens across a range of lenders and a range of loans. So if there is a surge in defaults, lots of people take a small hit rather than a few suffering hefty losses. At least, that's the theory.
To some, this is a breakthrough in risk management. The risk of a lender suffering a fatal blow from one of their debtors defaulting is much reduced; instead the damage is spread across so many investors in such small quantities that everyone will usually survive. The market becomes like a spider web; quite weak in terms of individual links, but stronger by virtue of its structure.
Could we face another LTCM-type incident?
But others fear the outcome will be quite different. They believe the credit derivatives market is so large and so complex, while risk oversight is so poor, that another system-threatening crisis on the scale of LTCM is certain. Financial institutions say they use sophisticated mathematics to model the risks that they are taking and help them hedge against severe problems. But these models are imperfect; some problems such as the risk that liquidity in your markets will dry up are very hard to model. And supposedly unlikely events such as major movements in asset prices are far more common than theory says they should be. LTCM's models indicated that its collapse was a ten-sigma event'. This is much more unlikely than the chances of someone winning the lottery for three weeks in a row. Maybe they were very unlucky or maybe someone got their sums wrong.
Concerns are heightened by the fact that current levels of derivatives exposure have yet to be tested under tough economic conditions. As James Grant says in Grant's Interest Rate Observer, "the supercharged growth in credit derivatives has taken place in a time of collapsing credit spreads, record-low bank failures" and, until recently, very low interest rates. This low-rate, low-default climate has forced investors to take on more risk in search of decent returns and encouraged them to make unrealistic assessments of how much risk they are really taking.
One recent story suggests there may also be risks outside the models. Much of the modern derivatives trade is over-the-counter rather than on an exchange. Therefore, the only record of the trade is between the two parties involved. But earlier this year the ISDA revealed that one in five credit-derivative deals involve major errors. There is a huge backlog of unconfirmed trades and some traders appear to have been recording their trades on the backs of old envelopes. Former Federal Reserve chairman Alan Greenspan, who regards derivatives as a breakthrough in reducing risk, called the situation "unconscionable" and said that he was "frankly shocked". Financial regulators are now clamping down on recording methods, but the fact that the problem could develop in the first place hasn't increased anyone's faith in the financial institutions.
And although credit derivatives have so far avoided any LTCM-style incidents, that's not to say there haven't been problems. Some investors saw heavy losses in 2001 as the US went into recession and corporate default rates shot up. There was another nasty wobble in 2005 when the credit ratings of Ford and GM were cut to junk. The worry is that if a major credit event happens such as a severe recession driving up corporate bankruptcies the result will be much worse than a wobble. Far from cutting risk, the spider web of derivatives could push it through the financial system, with losses in unexpected places.
Derivatives: assessing the risks
We may have seen an example of how derivatives trading can have unforeseen results during the May stockmarket turmoil; some claim the falls were exacerbated by hedge funds trading in variance swaps (bets on how volatile the markets will be in the future) being forced to buy and sell equities to hedge their positions on these. And this was a minor incident; a major crisis could see large hedge funds collapse as a result of risks they'd overlooked. Even big banks could face problems, as some of these are heavily exposed to derivatives. Any hint that one of these was in trouble could spark panic and cause mass-selling, amplified by further derivatives losses, sending asset markets in general into meltdown. In this scenario, global losses could easily spiral beyond the maximum "gross credit exposure" of the derivatives world. In that sense, playing with derivatives might not be a zero-sum game after all.
It's an apocalyptic picture. Could it happen? It's impossible to say but we know that throughout history investors have consistently underestimated the risks they are taking. A worrying statistic was reported by the European Central Bank in June. Apparently the correlation between different hedge-fund strategies is rising in other words, hedge funds are increasingly placing bets in the same direction as each other which also means they are exposed to the same risks. As the ECB points out, by late 2005, correlation levels had exceeded "those that had prevailed just before the collapse of Long Term Capital Management".
The high level of correlation "raises concerns that a triggering event could lead to highly correlated exits across large parts of the hedge-fund industry".
It's also certain that "a bear market in complex debt instruments will eventually succeed that current resplendent bull market", as Grant puts it. At the very least, the smart money must be on some speculators getting a very nasty shock.
Just what exactly is a credit derivative?
Credit derivatives hinge around two concepts CDOs and CDSs. A CDO is a collaterised debt obligation. A bank collects together a bunch of debt instruments and parcels them together into a freestanding financial instrument or special purpose vehicle the CDO. These debt instruments can be corporate bonds, mortgages, corporate loans pretty much any kind of debt.
Investors can buy into the CDO in several ways: senior notes, which have the lowest rate of interest but are first in the queue to get paid; mezzanine notes, which pay higher interest but come after senior notes; and equity, which gets all the return that doesn't go to the senior and mezzanine notes, but comes bottom of the pile in protection terms. The CDO uses the money it raises from these investors to pay the bank for the debt.
It's easy to see how the bank has benefited from this transaction; it's managed to shift the risk of this debt off its balance sheet and onto the investors in the CDO. But what's in it for the CDO's investors? Well, since the debt the CDO owns is generally higher risk (for example, junk bonds), it pays a reasonable rate of interest, which means the return on thenotes issued by the CDO is good compared to what can be found elsewhere in this low interest-rate, yield-starved investment climate.
As you would expect, higher risk means there is a higher chance of the bonds defaulting. But because the CDO is a pool of debt, if one bond defaults it won't affect the CDO's ability to pay interest to its senior and mezzanine investors the equity investors will have to cover it instead. Of course, if enough bonds default, it will eat into the income available to pay the interest to the mezzanine investors. For taking this risk, the mezzanine receives a higher rate of return than thesenior notes.
The other part of the story is the CDS or credit default swap. This effectively insures against a debtor being unable to pay his debt. One party to the CDS pays the other a premium. If the debtor doesn't default, the premium-payer gets nothing.
But if the debtor fails to pay up, then the premium-payer gets a payout from the CDS seller in compensation. CDSs started as a way for those who own debt to insure themselves against the risk of default. But today, a buyer of the CDS doesn't necessarily own the debt; they are bought and sold as investments in their own right. For example, an investor who expects Ford to default on its debt might buy CDSs on Ford bonds to profit.
But this is by no means the end of it. CDSs and CDOs combine together in something called the synthetic CDO. The demand for CDOs is so great that it is hard to find enough debt to create enough CDOs to meet the demand. This is partly due to the complexity of putting together a CDO in the traditional way; many types of debt require the debtor's consent if ownership of his loan is to be moved from the bank to the CDO.
Thus the synthetic CDO was born a vehicle that doesn't buy real bonds or mortgages, but instead puts its investors' money into safe bonds such as government debt and also sells CDSs on the type of debt it represents. The premiums it receives for selling the CDSs together with the interest onthe government bonds meet the interest payments to its investors. If one of the debts on which it has sold a CDS defaults, then the CDO pays the CDS buyer compensation and the CDO investors see their returns cut just as they would if the CDO really owned the debt that had defaulted.
These are the basic building blocks of credit derivatives. After that, you get lost in a forest of acronyms. An ABCDS is a CDS on asset-backed securities, such as a pool of mortgages. A CDS of a CDO offers protection against a CDO defaulting. CDO2 is a CDO comprised of notes issued by another CDO. A CDO3 you get the idea. By now, you may also be rapidly starting to realise why no one really knows just how risky these things are...