When a Safe Harbor Fund Isn’t

In the book “When Genius Failed,” Roger Lowenstein’s account of the notorious collapse of Long-Term Capital Management, the principals of that fund call the market forces that led to their demise “the extreme event.” The extreme event to which they refer was the simultaneous crashing in the late summer of 1998 of the firm’s seemingly diversified portfolio of investments-Russia and other emerging markets, directional trades in developed countries, equity pair trades, yield curve arbitrage, junk bond arbitrage, M&A arbitrage, stocks and last but certainly not least, swaps and equity volatility trades. The firm’s “diversification” turned out to be not so much hedging strategies but individual bets that all went wrong at once, wiping out $4.5 billion in capital in a matter of weeks. As Lowenstein explained, “The Long-Term episode proved that eggs in separate baskets can break simultaneously.”

The Long-Term Capital story, while spectacular in its scale (it nearly brought down Wall Street), is far from anomalous. Hedge funds-large and small-seem to blow up fairly often. The strategies employed don’t have to be as esoteric, complex or far-flung as LTCM’s in order for a collapse to occur. But there are two characteristics that all such disasters have in common: A strategy that isn’t flexible enough to withstand the latest one-of-a-kind market tsunami, and performance-based incentive fees that may have led to a bad bet being pressed. On October 18, 2002, a hedge fund specializing in mortgage-backed and asset-backed securities announced that it was shutting down due to excessive losses. Summit, N.J.-based Beacon Hill Asset Management told investors in a letter that its Bristol fund and its Safe Harbor fund were going to be liquidated following losses of 54% during the month of September. The magnitude of the loss has been estimated at $400 million on a fund that started the month at $700 million. The announcement corrected an earlier communiqué to investors, in which management had pegged losses at just 25%.

“As a result of extraordinary fixed-income market conditions last month and particularly in the final days of September,” the company explained, “Beacon Hill Asset Management’s Bristol Fund realized a loss. This situation was primarily the result of unprecedented, accelerated mortgage prepayments triggered by historically low interest rates, while US Treasury securities prices rose in reaction to heightened global market and political uncertainty.” Implicit in this explanation is that the portfolio was poorly positioned for the foregoing market conditions.

Inarguably, September was a challenging month: The headline event of the month was the high levels of mortgage refinancing. The Mortgage Bankers Association Refinancing Index, which is reported weekly, is setting records, with the index now surging past the heretofore unprecedented 6000 level. Prompting the surge in refis were the twin motivations of low interest rates and rising home values. The yield on 10-year treasuries fell from 4.143% to 3.596%. Similarly, the yield on 2-year treasuries fell from 2.129% to 1.687%, while 30-year conventional mortgage rates dipped from 5.85% to 5.67%.

Chart 1: Refis soar when yields plummet

On balance the economic news during September was flat to bearish. In the latest report, industrial production fell 0.3% in August from July, the first decline since December. According to the Institute of Supply Management, manufacturing activity contracted in September for the first time in seven months. The housing market, which had been on a torrid pace, showed signs of cooling off, with housing starts falling 2.2% in August, the third straight monthly decline. The Federal Reserve released its August survey of national business conditions, known as the “beige book”, in September, and it painted a picture of a sluggish economy. “Economic activity has slowed in recent weeks,” the Fed said. “Manufacturing activity was sluggish. Most districts reported little or no employment gain in July and August. The retail-sales picture was mixed for the nation as a whole.” The continued economic sluggishness and weak equity markets drove investors into the Treasury market.

Going into September, one can only speculate on what the managers of Beacon Hill thought. But if LTCM is a guide, it is likely that they trusted in their models and waited for a reversal in interest rates. But, like LTCM before them, they found that their models didn’t work. (At least not in time: the irony is that the backup in rates the last two weeks would have undoubtedly helped them.) We don’t expect that Beacon Hill is alone in this particular failing. These same factors caused Fannie Mae to lose control of its duration gap. It is likely that Fannie Mae’s efforts to get its duration gap back to within its acceptable limits, while good for Fannie Mae, exacerbated the trend towards lower rates and hurt Beacon Hill and others.

No one but the Beacon Hill managers knows exactly what was in the portfolio leading up to September, but news accounts and market speculation has it that the Beacon Hill funds were positioned to do very well in the event interest rates reversed course and rose, but poorly if rates continued to fall. For example, the Beacon Hill portfolios could have owned significant amounts of interest-only (IO) strips of mortgage-backed securities (IOs perform well when prepayments slow down, i.e., when rates rise), or it could have shorted Treasuries, or it could have owned securities at a high premium (which would have to be amortized quickly in the event of faster-than-expected prepayment speeds), or all of the above. Whatever it was, all of their eggs cracked at once. Beacon Hill apparently did hold some inverse floaters which presumably were not correlated with the rest of their portfolio, and it wasn’t highly leveraged-reportedly no more than 3 or 4 to 1, although mortgage derivatives such as IOs are structurally leveraged-but liquidity most likely dried up as the margin clerks took over.

The lesson of Beacon Hill, like the lesson of LTCM before it, is that a strategy that is geared to one particular outcome will do wonderfully well as long as that outcome obtains. It will destroy capital if- or should we say when- the opposite outcome occurs. The factors that led to LTCM’s collapse were uniquely damaging, as were the conditions that hurt Beacon Hill’s investors, but the point is that there will always be something, a set of market conditions that spell ruin for someone. At the end of the day, perhaps Beacon Hill would have been “right” 9 times out of 10 (or, like LTCM, perhaps 99 times out of 100). As if any of us need to be reminded of this lesson in the wake of the collapse of our stock market bubble, investors need to always know what their fund’s Achilles Heel is, and whether it causes the fund to break or just bend.

The corollary to this lesson is that the prudent investor will seek out strategies that are geared to do well in a wide variety of scenarios, where being “wrong” doesn’t mean being “busted.” Leading into this September, the kind of portfolio that is managed by the folks at Annaly and FIDAC got more defensive as to interest rate direction, the way a driver eases off the gas going into a curve: shortening asset duration, lowering leverage, staying in liquid asset classes. Without a performance fee to worry about, portfolio managers at Annaly and FIDAC are more interested in the long-term best interests of our investors, preferring to keep them rich rather than to make them rich or, much worse, make them poor.

October 23, 2002
By Jeremy Diamond, Executive Vice President
Annaly Capital Management/FIDAC