Hedge funds have received considerable attention from academic, political and media circles since the crisis. But what effect does hedge fund activity exactly have on asset markets?
In this article, I write about a recent working paper that studies how a measure of hedge fund illiquidity can forecast asset returns. Is it possible to infer something about asset returns by monitoring hedge funds activities at large? Read on and find out.
There is one type of trading that hedge funds excel at and that is high-frequency trading (HFT). This basically means that hedge funds identify and execute trades in a matter of microseconds. HFT practically means that it is hedge funds that provide liquidity to many asset markets. The fact that hedge funds trade continuously makes sure markets do not get stuck in a state of illiquidity. It, however, also means that hedge funds themselves might become illiquid.
If a hedge fund is to continuously keep trading, it will have a large portion of its cash tied up in the markets. The problem with this is that hedge funds invest other people’s money, who can pull that money out anytime. So hedge funds must be wary of being too illiquid.
Since the hedge funds provide liquidity to the markets, they act as quasi-market makers. But if a hedge fund is too illiquid (and thus it wishes to raise its liquidity) it will be less willing to fulfill this role. Traders, however, need liquidity, so they will bid the price up until the hedge fund is willing to enter the market again.
This claim actually stands a rigorous empirical test. Kruttli, Patton and Ramadorai (2013) construct a surprisingly simple measure of hedge fund illiquidity and find that it is a very good predictor of asset returns in the next month. It outperforms all alternative predictors the authors try; and it performs well in out-of-sample predictions as well. The evolution of hedge fund illiquidity over time can be seen in the figure below.
This measure of illiquidity is based on the fact that hedge funds with higher illiquidity generally exhibit more persistent returns over time. I.e. their returns do not vary as much. Thus if one just looks at a hedge fund’s returns over time and calculates the 1-period autocorrelation coefficient, one can have a very good estimate of the illiquidity of the fund in question. The authors do just that for a lot of hedge funds. They use monthly returns, and they average hedge fund illiquidity over funds for each month in their sample. They then test how a one-month lagged value of this measure can predict returns of international equities, corporate bonds, and currencies.
As mentioned above the results are generally positive, highly significant and tend to outperform other predictors such as dividend yield or lagged returns. Illiquidity works best as a predictor with international equities, and then with corporate bonds. But even with the notoriously hard-to-predict currencies, it is significant in 6 out of 9 cases. Pretty neat.