Crowded Hedge Fund Trades: Rules Changed in 2007

Earlier this week, we told you about a recent European Central Bank report containing some research on correlations between hedge funds.  Since position-level analysis is very difficult, researchers used the average cross-correlation between hedge funds following similar investment strategies.  A high average correlation was interpreted as meaning that hedge funds had likely pursued the same underlying trades.  

Position-level data may indeed be next to impossible, but not quite.  In the United States, hedge funds are required to file a quarterly form called a “13f” that contains their common shares, convertible preferred shares and convertible bond holdings (long positions only).  While the data is somewhat stale (managers have 45 days after quarter-end to submit the report), it provides some interesting insight into so-called “crowded” hedge fund trades.

In a report issued to clients on August 27, 2007, Thomson Financial analyzed the largest US holdings of the top 25 biggest hedge funds in the United States.  Collectively, this group managed slightly over US$400 billion.  According to Thomson, over 3% of these assets were invested in Sears Holdings (although they acknowledge that most of that may be ESL).  The list of 25 also includes names such as Autozone, CVS, Google, Alcoa and RIM.

But does the proportion of the top 30 hedge funds’ investments indicate a trade is “crowded”.  Perhaps the number of top 20 hedge funds in the name is a better indication?  On this grounds, CVS, Wellpoint, and Qualcomm would be the most crowded. 

Examining hedge fund ownership as a percentage of market cap might be a more accurate measure how popular a company is amongst hedgies.  Goldman Sachs has created an index it calls “The Most Concentrated Positions” (Bloomberg: GSTHHFHI) is an index of the stocks with the highest percentage of hedge fund ownership (by 700 hedge funds with $936 billion under management in aggregate as of October 1, 2007). 

In a November 21 research note, Goldman said that a strategy of buying the top 20 S&P 500 stocks in which hedge funds have the highest percentage ownership has outperformed the index by over 17% on an annualized basis.  While they underperform in market pullbacks, they usually rebounded soon after.  According the the research note, this was because these names were oversold due to “ownership characteristics” not “fundamental reasons”.  (This list includes names like Goodyear, Alcan, and Virgin Media).  So score one for the “smart money”, right?

Maybe not.  The firm also found that these stocks broke with tradition and didn’t actually rebound after the July/August sell-off.  In fact, they continued to underperform the S&P 500 in September.  Although Goldman said this strategy “may continue to outperform”, they sent the rocket scientists back to the drawing board. 

The result was Goldman’s “Very Important Positions [VIP]” list (Bloomberg: GSTHHVIP).  This is a list of the stocks the firm feels “matter most” to fundamentally-driven hedge funds. (Score one for the fundamental managers over the quants).  Goldman says it now prefers stocks owned by fundamentally-driven hedge funds to stocks that simply have a high proportion of hedge fund ownership.

The VIP list contains stocks that make the most number of “top 10″ appearances in the portfolios of hedge funds (ranked by position-size).  To isolate ”fundamentally-driven” funds, Goldman only surveyed funds with between 10 and 200 positions.  This removed quant funds and primate equity-like funds from the mix.

Contrary to the “Most Concentrated” basket, the “VIP list” outperformed the S&P 500 during the fall recovery and ensuing correction (although it seems to have underperformed since November 21.) 

 

So while hedge funds only own 1% of behemoth General Electric, the company makes a top 10 appearance in 24 of the funds examined (ranking it #8 on the VIP list).  Other top ranked VIPs included: Google, Cisco, Alcan and Corning. 

Goldman’s research seems to corroborate the ECB’s (and NY Fed’s) hypothesis that crowded trades and the resulting fire sale of positions were behind August’s hedge fund meltdown.  It also suggests that hedge funds did not dive back into their favorite positions afterwards - leaving those stocks even more vulnerable to November’s market slide. 

Comments

[Malkiel]

While mutual fund ownership is declining, it must still be significant, and these guys are also a herd mentality--so how to cross-reference hedge fund and mutual fund buying? It would be interesting to know if they synergize or whether they buy completely different stuff at different times...