Ever since academics noticed how some types of hedge funds (you know who you are) had a significant correlation to equity indices, the hunt has been on to find what else drives hedge fund returns. As each new alternative beta is discovered, the remaining magical “alpha” produced by managers shrinks further and further. Alpha, to borrow from a Victorian phrase about the nature of God, “is in the gaps.”
Some of these exotic risk factors are more investable than others. Even when an alternative beta is investable, it may require a lot more human intervention than, say, establishing and managing an index ETF. “Illiquidity risk” is one factor that would arguable fall into this category. Illiquidity risk exists in so many different forms that it’s impossible to agree on just one measure. Several contributors to this website have offered heuristics in an attempt to account for longer lock-ups (“Research puts price on hedge fund illiquidity premium“), as well as simple metrics that may help represent a generic liquidity component in financial markets (“Illiquidity premium that fueled endowment returns falls back to 2005 levels“).
Recently, we wondered if the effect of the so-called “illiquidity premium” was on display in the recent returns of hedge fund-like products including hedge fund replication funds and hedged mutual funds. As you can see from the chart below, it appears as though the less liquid and un-investable HFRI index beats not only the HFRX – as it has for a long time now – but also the average of Morningstar’s long/short mutual fund category. Unfortunately, the Morningstar category only goes back to October 2007 and, hence, is responsible for the rather weird start date on our chart below:
As you can see, the HFRI just regained its high water mark while more liquid alternatives, i.e. HFRX and Long/Short mutual funds, seem to have underperformed. The HFRI also looks to have accomplished this with a volatility no worse that these other options.
Of course, like all investable hedge fund indices, the HFRX is only an approximation of its non-investable cousin. Therefore, expect a lot of unavoidable tracking error to be likely. But having experience with a hedge fund that was part of both an investable and non-investable index, we can say that a significant portion of the tracking error is the result of measures taken to meet the liquidity parameters set by the managed account platform upon which the investable index is built. So we’d submit that liquidity-provision may not directly lead to lower returns, but that it always seems to be near the scene of the crime…
Regular readers may recall the chart below from a paper by Amir Khandani and Andrew Lo at MIT. The duo used return autocorrelation as an indication of asset illiquidity in hedge funds and mutual funds. They found that the funds with the highest correlation (illiquidity) also produced the highest returns. The chart below, created from data in their paper, shows the annual excess return across different autocorrelation buckets. Notice the funds with the highest illiquidity risk also produce the highest excess returns.
This study by Ronnie Sadka at Boston College makes a similar conclusion: funds exposed to high illiquidity risk, a.k.a. “liquidity beta,” produce higher monthly alphas (ignore the red dots in the following chart unless you’re really into stats).
Sadka even concludes that you if went with long hedge funds with a high liquidity risk, i.e. risk that the fund will tank when market-wide liquidity dries up, and short funds with the lowest liquidity risk, you’d make money.
Unfortunately, none of this tells us if the delta between non-investable indices and their more liquid open-ended or investable cousins is a result of their lower liquidity. But we think it’s got to be a contributor. In fact, we might even say that illiquidity may be a multiplier of manager skill since it adds a dimension of inefficiency for managers to exploit. Constraints on liquidity are no different from other investment constraints, including the much maligned “long-only constraint.” Researchers do say that removing that constraint increases the Transfer Coefficient and thus a fund’s Information Ratio – check out the Law of Active Management as articulated by Roger Clarke, Harindra de Silva and Stephen Thorley and our related post.