Adapting to inefficient markets
A new approach to equity investing is drawing attention from some of the more sophisticated Japanese pension funds, as the fall-out of the credit crisis drives them to reassess their portfolios. Although the Efficient Market Hypothesis (EMH) has been the main fundamental premise for building institutional portfolios for the best part of this decade, the credit crisis has triggered a reassessment of the collective wisdom of markets and their ability to price securities correctly. Previously, anyone arguing against the EMH put themselves outside the academic mainstream. But empirical research continued on features such as ‘the amazing January effect’, the consistent outperformance of value-stocks and the lack of predictive power of beta; the presumed single driver of individual stock returns according to the EMH-based Capital Asset Pricing Model (CAPM). These arguments are now finding a new hearing from investment practitioners at several large pension funds in Japan.
Their core question is simple: do high risk stocks indeed over time produce the high returns as predicted by finance theory, or not ? Since the early 1990s several academics (an example is Haugen and Baker (1991 in The Journal of Portfolio Management, called “the efficient market inefficiency of capitalisation weighted stock portfolios”) have argued that the empirical evidence shows that in fact low risk stocks provide very decent returns without the large swings of the broader equity markets. In other words, one can improve the risk/return profile of a portfolio by reducing the absolute risk of the portfolio, without giving up any return. This finding goes against the main premise of the EMH, that an index covering the broad market such as the MSCI World is efficient, in that no other portfolio can improve on its return profile without increasing risk or improve on its risk profile without giving up return. It turns the EMH literally on its head.
These results are now being taken seriously, to the degree that money is actually being allocated on this basis. Firms such as Analytic Investor, Acadian Asset Management and Alfred Berg Global Alpha have developed “minimum variance stock portfolio” strategies out of this concept of investing.
It is ironic that the boom-bust cycle we have just witnessed has been so under-exploited by active managers. In fact, most of them have massively underperformed the market. The logical conclusion seems to be “okay markets may be vastly mis-pricing risk, but it still is impossible to beat a passive index” or in a more popular quote from Keynes: “markets can stay irrational longer than you can remain solvent”. Positioning one’s portfolio to lower risk stocks is one strategy that would have led to out-performance during the market sell-off and the research cited above suggests that over the long run the same is true, but there are periods when the market is seeking risk and positioning yourself away from risk will hurt performance against the index. The same goes for other strategies that use predetermined factors in order to beat the index.
Until the tech bubble developed in earnest in 1998, low PBR (typical “value” stocks) had been a slam-dunk strategy for 25 years until for a few odd years investors started to prefer high PBR stocks. With the bursting of the tech-bubble, low PBR came back with a vengeance and by 2005 many quant managers could show fantastic 5 year track records by betting on low PBR. In late 2007 however these very quant managers faced a crisis - with many quantitative managers underperforming by large margins and some going out of business. Their problem was a fast and fundamental rotation in the market place. In 2007-2008, the value bets that worked so well during the past 5 years reversed strongly and growth factors dominated to an extreme degree. Non-adaptive quantitative models, using only a limited number of factors, were caught totally unprepared by this change. The models suggested markets were not priced “correctly”, but at the same time kept the search for value opportunities to profit from this mis-pricing; in vain.
It is this occasional but persistent change in investor preferences that drives the phenomenon that markets might not price “correctly” but still offer few opportunities to profit from it. Unless off course one can constantly “keep the pulse” of the market and adapt the portfolio to position it for what the market is paying for.
It is this unbiased adaptiveness that now gets a chance as the market efficiency paradigm is breaking down. Professor Andrew Lo of MIT proposes to replace the EMH with the “adaptive market hypothesis”. According to Lo, markets are adaptive structures in a state of continuous change. His point is that market efficiency is context-dependent and dynamic and, as in nature, success depends on the ability to adapt to an ever-changing environment. As the market moves through phases of exuberance and despair opportunities continue to emerge for the agile and adaptive to take advantage of.
Nardin Baker, of the Haugen and Baker paper has actually moved from academia to the money management industry and for the past 15 years has managed portfolios at Alfred Berg based on dynamically adapting the portfolio to market preferences. The Alfred Berg model uses a very large factor set (60) compared to most other managers and the model is flexible and dynamic, allowing factor pay-offs to adapt to changing market conditions. The track record of the model provides ample ammunition for the EMH skeptics. As always, the proof of the pudding is in the eating!