Linear regression models (a.k.a. factor models) have a number of emerging applications in the hedge fund industry. One of the most often-cited here and elsewhere is hedge fund replication (see related posts). But as we discovered recently, regression-based models can also be used to estimate the daily returns occurring between monthly hedge fund reporting cycles (see related post). In addition, MIT’s Andrew Lo has proposed several other applications of linear factors models to address situations such as transitioning between managers and portfolio rebalancing for risk management purposes (see related post).
Now Lo has teamed up with Alexander Healy of Alpha Simplex Group (the company with which Lo is closely affiliated) and proposed yet another application of this truly alpha-centric approach to portfolio management: dealing with redemption gates.
The two suggest that when hedge fund investors are confronted with redemption gates, they can essentially remove their economic exposure to many of the underlying hedge fund betas in much the same way an executive can monetize un-vested stock options. By basically shorting the basket of betas that make up the returns of lock-up hedge fund allocations, investors can reduce volatility dramatically and in some cases, even increase returns (i.e., if the alternative betas in question temporarily deliver negative risk premia).
Drawing on a knack for colourful metaphors, Lo says this is not unlike the strategy taken by the drugs often prescribed to those with high blood pressure:
“We have borrowed the term “beta-blocker” from the pharmaceutical industry where it refers to a class of drugs used to treat hypertension and heart-attack patients by blocking so-called “beta receptors” in the heart and kidneys. Given recent market conditions, blocking financial betas may yield similar salutary effects.”
The following chart from the paper shows the evolution of these estimated betas of a sample of 47 long/short funds (click image if you can’t read the text).
(Linear regression enthusiasts might notice of the 2007 dominance of the relatively puny, yet resource-centric Canadian equity market and the more recent dominance of the Euro, Pound, and Japanese Yen.)
A negative position in these betas produced returns represented by the green line in the chart below from the paper:
As you might expect, Lo’s “beta-blocker” mitigated upside performance surprises and reduced the frequency of high (24-month trailing) standard deviations and large draw downs.
But interestingly the addition of the custom-designed “beta blocker” portfolio actually led to higher overall returns in “turbulent” markets. The authors say this was due to the temporarily negative returns from alternative beta sources that generally deliver a positive risk premium (making short positions in those factors a positive contributor to returns).
So if beta blockers work in certain market regimes but not others, then can you use them to time the market? Sort of, say Healy and Lo. While their dynamic beta blocker approach is not technically market timing, it is what they call “volatility timing.”
So unlike the beta blockers taken by those who suffer from high blood pressure, these ones need not be taken for the rest of your life.