If you're employed to measure risk on Wall Street, it's obvious there's much less of it being taken now than last summer. There's one hitch: No one can tell you how much less.
Forget about ascertaining if the caution stems from the September 1998 collapse of
Long Term Capital Management or from
Y2K scenarios of widespread havoc, financial and otherwise.
Unless you're a regulator, it's impossible to know exactly what's changed in the last year. In part, that's because much of the risk that's being taken is by minimally regulated hedge funds like LTCM, or on an off-balance-sheet basis by heavily regulated banks and brokerage firms. Those numbers simply don't show up in the public domain.
The anecdotal evidence, however, is plentiful and consistent. In March,
Treasury's Lewis Sachs told
Congress that the government had seen "some tightening of market discipline since last fall" inasmuch as "there is some evidence that banks and other suppliers of credit to highly leveraged financial institutions are demanding more collateral or requiring larger haircuts on their repurchase agreements and derivative transactions."
Yet even if the numbers were readily available, risk is extremely difficult to measure.
Anyone remotely familiar with the LTCM story has heard something about the role of leverage in its demise. Leverage is the bond market's version of buying on margin. (LTCM is primarily a fixed-income hedge fund.)
If you can use $100 million of bonds as collateral to borrow another $100 million and invest it in bonds, you are leveraged 2-to-1. Like margining in stocks, leverage magnifies any gains and deepens any losses. But leverage isn't as risky in bonds as in stocks, because bonds are so much less volatile.
LTCM was leveraged 28-to-1 at the end of 1997, controlling $129 billion of assets with $4.7 billion of capital, according to a
study of its collapse by the Treasury Department, the
Fed, the
SEC and the
Commodity Futures Trading Commission released in April.
By that measure, it was far and away the most highly leveraged large hedge fund reporting to the CFTC. The regulators' principal conclusion was that excessive leverage needs to be prevented, but they also conceded that leverage is a doozy to measure. In Long Term Capital's case, the 28-to-1 ratio "does not include any leverage that LTCM may have been able to assume through derivatives or other off-balance-sheet transactions," Treasury's Gary Gensler told Congress.
And that's not the half of it. A June
industry report on how to prevent another crisis is deadly reading. But it contains a fascinating (really!) analysis of the various ways to measure leverage. Sparing you the details, there are at least six.
Even "net economic leverage," the fourth-most-sophisticated method detailed, would rate as equally leveraged a fund with a single large position in an illiquid emerging-market equity, as one with a like amount of three-month Treasury bills, if both used 25% equity and 75% debt.
The most accurate way to measure leverage is a mouthful: Asset liquidity-adjusted value-at-risk leverage.
The trouble is that the concept at the heart of it -- the value-at-risk part, which estimates potential losses on a position over a specified time period -- has subjective elements. The study released in April blamed flaws in the risk models of LTCM and its creditors for failing to predict how far out of the money the fund's positions might go. Yet, it also acknowledged that the events that triggered the fund's demise were the financial equivalent of a 100-year flood, and it isn't clear how such possibilities should fit into risk models.
Hard evidence that there is less risk now than a year ago is sparse, but clear as far as it goes. (And remember, it isn't clear there ever was a leverage problem extending beyond Long Term Capital.)
Rather than describing the amount of leverage that exists, the evidence describes the ways in which leverage is created. Much of it is created in the over-the-counter "repo" market, where securities are used as collateral to borrow at short-term interest rates. (Repo stands for repurchase. The borrowers actually sell their securities to the lenders and agree to buy them back the next day.)
LTCM was busy in the repo market and had about 75 repo counterparties. Its absence is being felt. Weekly outstanding amounts of repo and reverse repo agreements during the first half of the year are down 6.7% from last year, according to a
Bond Market Association analysis of
New York Fed data.
Repo Subsides
Average daily amount outstanding of repurchase agreements and reverse repurchase agreements, in billions |
 |
| Source: New York Fed |
The other major way leverage created in the bond market is through derivatives. Derivatives volume is as high as ever. But in both repo and derivatives, the amount of leverage that can ultimately be created is a function of collateral requirements.
In repo, for example, you can't borrow the full value of the collateral you post. You can borrow a bit less. The dealer's "haircut" is insurance against the possibility that you'll default on your loan, but it also keeps you from leveraging yourself ad infinitum, if each time you can borrow a little less.
Similarly, in derivatives, you post margin. In a simple example, if you take a position in bond futures at the
Chicago Board of Trade, whether long or short, you have to post initial margin. Then, if the market moves against you, your loss is subtracted from your initial margin and you have to post maintenance to make up the difference.
About 85% of the world's derivatives activity is conducted on an over-the-counter basis, rather than on exchanges. But the basic rules are the same: Margin -- or at least collateral -- is required for derivatives exposure, and the amount of collateral required ultimately constrains the amount of leverage that can be created. Long Term Capital had about 50 OTC derivatives counterparties.
That's still true, though perhaps to a lesser degree, said Linda Sullivan, fixed-income analyst specializing in repo at
IDEAglobal.com. "It does cost hedge funds more money to finance their positions now than it did last year," she said. Haircuts went up and credit limits went down. More recently, Sullivan said, "some of the larger, more liquid funds did see their haircuts lowered a little bit, but the marginal customers did not."
On the derivatives front, CBOT
data show a marked decline in open interest in interest rates futures over the last year.
The OTC derivatives market is even more opaque than the repo market because it is fragmented. But Peter Nerby, securities industry senior analyst at
Moody's Investors Service, says: "It's a reasonable assumption that a lot more margin is being posted to a lot more people on a lot more positions."
