Illiquid Assets: Dimensions of Risk

Exploding spreads in structured mortgage products, which have driven sharp drops in asset values and collateral calls have shined a bright light on the risks posed by illiquid assets. Bear Stearns is now being held up as the latest victim of flying too close to the Sun, loading leverage on top of complex, illiquid assets that, in fact, had values far lower than those reflected on its books in the wake of a liquidation. Today's article in the Wall Street Journal made some good points that plainly chronicle some of the issues facing those investing in illiquid assets:

Figuring out the risk profile of illiquid assets -- and funds that invest in them -- can be tricky. Typical methods for assessing risk rely on measuring volatility -- the choppier returns are, the riskier the investment. But because illiquid assets don't trade regularly, marking to market -- or using recent sales prices to determine an asset's value -- may not be possible. In these cases, a fund manager may instead use a mathematical model to value an asset, a practice called marking to model.

Such models tend to smooth returns, making an asset look much less risky, says Massachusetts Institute of Technology finance professor Andrew Lo, who is also a principal in AlphaSimplex Group LLC, an asset-management company that runs a hedge fund.

Using broker-dealer quotes for illiquid assets can also damp volatility because they are often based on an average of bid and offer prices rather than actual sales prices. What's more, price quotes can vary widely from one dealer to the next.

********************

Mr. Lo has found that returns for illiquid assets and funds that invest in them tend to have little variation from one month to the next. Paradoxically, it is this smoothness of returns that show how illiquid, and risky, a position might be.

Andrew Lo is a smart guy. He has done seminal work in the area of analyzing the properties of hedge fund returns. His statements are both accurate and logical and, not surprisingly, mirror those of someone who has written about the risks of running an illiquid asset book - me. From Broadening the Multi-Strategy Mandate 07/21/2006:

For instance, if you can't arrive at a fair value for an asset, a "thumb in the air" method is used like book value minus an illiquidity haircut. This approach dear friends, is not very scientific and is extremely prone to error. Beyond error, once an approach is used for a particular asset that approach tends to be used again and again and again. This has two potentially adverse effects upon disclosure and compensation: (1) compensation may be overstated because NAV might be much lower that the level reflected in the NAV calculation; and (2) volatility of returns will be understated since this asset, whose value invariably is moving up and down but whose value is difficult to measure, is being valued in the same way quarter after quarter after quarter. This is called autocorrelation. This skews the analysis of fund returns and makes a fair comparision across different firms and different strategies nearly impossible. Does any of this sound good?

Bottom line: if you are running an illiquid asset book, especially one that is leveraged, you'd better have some of the following working for you:

1. Immense liquidity to support margin calls in an eroding market;
2. A diversified portfolio so this one book doesn't drag down your entire firm (a la Amaranth);
3. Valid, conservative market data points on the bid side to back up "mark to model" valuations; and
4. A long time horizon.

And you don't need to be running a complex, leveraged mortgage derivatives book for this to be the case, which is why I kind of chuckle when all this emphasis is being placed on Bear Stearns. Risks of this sort abound across a wide array of portfolios, many of which are anything but exotic. How about the following:

1. A large short position where the borrow is tight and the market is rallying;
2. A large long position in an thinly-traded stock;
3. A highly concentrated portfolio of small- and mid-cap stocks;
4. A leveraged book of correlated large-cap longs in a declining market.

Are these "exotic" risks? No. Can they be marked-to-market and not just marked-to-model? Yes. Can you get competitive dealer bids? Yes. Is leverage the killer in each situation? No. Is illiquidity the killer in each situation? No. The nature of the risk differs based upon the situation. Sometimes it is liquidity-based. Other times it is leverage-based. Sometimes it is both. So let's not throw stones at Bear Stearns and make like they are another LTCM. Because they're not. Funds take risks. Stuff happens. Disclosure of these risks needs to be extremely clear, accurate and detailed. Investors need to know what they're getting, and valuations need to be done in a rigorous, transparent manner. If this is done and the s**t hits the fan, so be it. It is the by-product of playing the game. Hedge fund returns, as Mr. Lo and others have discovered, are generally negatively skewed, e.g., akin to selling options, collecting premium, generating superior returns for a period and then BANG - getting hit hard, and sometimes getting carried out. Hedge fund investment is not a business for the meek and certainly not one for the stupid. If investors thought they were getting a free lunch by investing in leveraged "high quality" paper and collecting juicy returns, shame on them. Because in the markets, you seldom get something for nothing.   

Comments

[DrFox]

Hi Roger,

Very nice article!

It is still deep for me, but I am taking it by the pieces.

Anyway I liked very much the concept of "exotic" illiquid portfolios.

Regards from Brazil!

[Motts McGregor]

Telegraph.co.uk: "Banks set to call in a swathe of loans":

http://www.telegraph.co.uk/money/main.jhtml?xml=/money/2007/06/26/cnusecon126.xml&CMP=ILC-mostviewedbox

-Motts

[Yaser Anwar]

One thing most investors fail to realize is that the risks facing hedge funds investing are nonlinear and more complex than those in traditional asset classes. Because of the dynamic nature of hedge-fund investment strategies, and the impact of fund on leverage and performance, HF risk models require more sophisticated analytics and users.

Due to the phase-locking phenomenon, hedge fund exposure to various market factors changes drastically (sometimes even changing signs from positive to negative exposures) depending on the state of the market.

A good # of hedge funds concentrate on hedge fund performance relative to a certain benchmark (BSC to ABX Index). An important question relates on the risk profile of HFs in terms of dynamic risk exposure, specifically, phase-locking exposure to fat-tail events and the switch in volatility of hedge fund risk factor.

In a normal state of the market, the exposure to risk factors could be very low but as soon as the market risk factor captured by the S&P500 (or that certain benchmark- ABX in BSC's case) moves downwards characterized by negative returns and high volatility, the exposure of hedge fund indexes to the S&P500 (ABX for BSC) and especially to other risk factors changes significantly, resulting in phase-locking.

[Motts McGregor]

Roger, thanks for that and welcome back. May I ask for your comment on the following interpretation as I am not a hedge fund / big company guy by trade....

In many of the recent articles on the Bear Stearns funds and the withdrawing of investors, seizing of collateral by lenders, and ultimately the infusing of capital by the parent co., it appears that (and this is pure tea-leaf-reading) ALL MARKET PARTICIPANTS have gone to great lengths to PREVENT a market price from being struck on these CDO instruments. I think of an auction by ML or JPM that was cancelled at the very last minute when they reached a deal with Bear Stearns? Hmmm....

Presumably this is justified in folks' own minds on the basis of "it would be bad for the market" but it seems to me it has the effect of avoiding a new mark-to-market price, which everyone realizes they would then have to use to value their OWN similarly rated and structured CDO tranches. And repricing the whole asset class down would be bad for "everyone" and so a day of reckoning is unnaturally postponed.

This seems to give the lie to the STATED rationale of the mark-to-model approach -- "oh we just can't get a reliable market bid so we use this", and reveals a much more self-serving one -- "it's a heck of a lot better for us to keep this thing stable and keep the fees coming in off an artificially inflated asset base, and it's better for all the other guys on the Street too."

Coming right out and saying "we believe the value of these instruments is more accurate coming off the model" obviously won't fly as it is not compliant with fair-value GAAP and other reporting standards, let alone common sense... but it just makes doing everything in one's power to prevent a substantive and highly informative after-market trade from occurring all the more important as a way to keep the mark-to-model game going a while longer.

Conspiratorily yours,
Motts