Hedge Fund Replication: Nuts!

In the year since Lo and Hasanhodzic floated the idea of hedge fund replication, a.k.a. hedge fund cloning, there has been plenty of discussion debunking the concept.  But a lot of it gets bogged down in statistical minutiae.

What astonishes me is that major players keep on rolling out replication products — Goldman Sachs’ Absolute Return Tracker, Merrill Lynch’s Factor Index, Deutsche Bank’s Absolute Return Beta Index, JP Morgan….  This is either naive or sarcastic, as a simple explanation should show:

Clones claim that they can produce the same returns as hedge funds by taking positions in tradable indices.  Well, not exactly … they actually aim to produce the same returns as benchmarks of hedge funds, which are notoriously problematic for reasons detailed elsewhere.

Why is this a stupid exercise?

1. To the extent that hedge fund returns can be replicated with some linear combination of indices, the hedge funds are not doing their job of producing alpha, which is defined as producing returns that are uncorrelated with tradable indices.  Granted, there are plenty of “hedge funds” out there that are selling a lot of beta.  Pity the investor paying 2%-and-20% for beta, but don’t try to mimic him.

2. The replicators are explicitly trying to match a hedge fund benchmark — which people only want because it represents alpha — by constructing a basket of beta to match it.  Mathematically we know that with a big enough “basis” of beta instruments we can construct a basket that will match any return history.  But, having fit this basket to the history, will it match the returns going forward?  Again, if it does then it means that the target benchmark has no alpha component.  If it doesn’t, then you failed to capture any alpha.  Either way the clone investor loses — all he gets is beta!  (And insofar as hedge funds are adding any value by timing beta, that timing is being lagged by the cloner!)

3. Why clone the returns of a suspect hedge fund benchmark — indeed, one that represents returns after the hedge funds’ notoriously high fees?  Wouldn’t you rather replicate the returns of hedge funds before fees?  How about an ideal hedge fund index that is completely uncorrelated with beta?  Oops, that’s mathematically impossible … no matter: What investors really want is high returns with low risk.  So why not clone a return series that exactly matches LIBOR + 3%?  And if we can do that, how about absolute returns of 100%?

No matter how you slice it, if hedge fund benchmarks are replicable with liquid securities then both the benchmarks and the replications are useless as sources of alpha.

8 thoughts on “Hedge Fund Replication: Nuts!

  1. WSJ covers these products this weekend.

    Firms offering clones generally acknowledge that the products can’t deliver the returns of the best hedge-fund managers.

    Some firms make more ambitious claims than others. Merrill Lynch says its Factor Model clone should deliver the full, after-fee returns of the average hedge fund, as tracked by the HFRI Fund Weighted Composite Index. But Goldman Sachs’s Absolute Return Tracker “is not really expected to track hedge-fund returns that closely,” a spokesman says. Rather, the product’s performance over time should “broadly resemble” hedge-fund returns, he says.

    Some experts doubt even that goal. A recent study of cloning by researchers at France’s EDHEC Business School showed that most clones generally did a poor job of mimicking the returns of hedge funds.

    A possible reason: Hedge funds are adjusting their portfolios all the time, whereas factor-model clones are based on previous years’ hedge-fund returns, says Lionel Martellini, a co-author of the study.

    Andrew Lo, a finance professor at Massachusetts Institute of Technology whose research helped to spark clone products, says he has found that the clones will capture only about 40% of hedge-fund returns. One reason: Some hedge funds are especially hard to mimic, such as those that try to profit from events such as corporate bankruptcies, because they often hold hard-to-trade securities that are off-limits to clones.

    And some clones may not provide the diversification that investors seek in hedge funds, according to recent research by Harry Kat, professor of risk management at Cass Business School at London’s City University. The clone products offered by the big Wall Street firms generally behaved much like the S&P 500 index in recent years, he says.

    Mr. Kat markets a competing approach to cloning that doesn’t aim to match returns but instead tries to mimic typical hedge-fund characteristics, such as low correlation to broader markets.

  2. Another way to look at this: Alpha is, in a very real sense, a proprietary product. Just like software or hardware that may be built by any other company. Cloning a hedge fund using a historical regression on financial indices is something like trying to clone the appreciation of an individual stock like Google. To the degree that Google does anything special, nobody would believe it could be replicated by trading large baskets of other stocks, bonds, or commodities.

  3. Harry Kat, proponent of “weak replication” (i.e., replicating statistical characteristics of alternatives) explains why “strong replication” (i.e., replicating uncorrelated and above-average risk-adjusted returns) will not work:

    The HFRI Composite is a basket of over 2000 hedge funds, following a large variety of strategies. Since it is extremely diversified, the peculiarities of the various strategies will largely diversify away. The result is an index with little hedge fund-like properties left, mainly driven by equity and credit risk; just like any typical traditional portfolio.

    So here is the problem. Although the HFRI Composite is advertised as a hedge fund index and calculated from hedge fund data, its returns don’t have many hedge fund-like properties at all. Replicating its returns therefore hardly qualifies as true “hedge fund replication” as we have defined it.

    From an investment point of view, the value of these hedge fund replication products is also doubtful. They exhibit a strong correlation with the stock market, which severely limits their attraction as portfolio diversifiers. Investors that buy into these products are basically selling off equity and credit exposure to finance the investment and then buying more or less the same exposure back in a different form. By itself this wouldn’t be so bad, were it not that the new product probably costs 3 or 4 times as much as the old. This also explains why the main players are so eager to issue these products. Pour some hedge fund replication and alternative beta sauce over something old and boring and suddenly it sells again like new.

  4. Kinda late in finding this article…I am doing some research on hedge fund replication. I have found a lot of articles discussing these tools but cannot seem to find the actual tools themselves. Is that a coincidence? Did these projects just fall off the map? Or do my researching skills suck. Any help is much appreciated.

  5. I agree, this idea will never survive past a core set of institutional clients who have this plugged into an overall portfolio management plan that includes dozens of different types of investments…

    – R

  6. Paige – I am in the same boat that you were in 3 months ago!

    Did you have any luck? Can u give me any pointers?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s