The past few weeks have seen a surge of inflows into US equity mutual funds, following many years in which investors have preferred allocating money to bonds rather than stocks. The week ended January 9 saw the fourth largest weekly cash flow into equity mutual funds since 1992, and large investment companies such as BlackRock have spoken of a sea change in the opinion of small investors towards equities. Some analysts see this as the start of a great rotation from bonds into stocks, thus reversing the pattern of the last decade.
Others, however, point out that cash inflows from small investors tend to be contrarian indicators, since they are often driven by recent market behaviour, rather than by fundamental valuation, which is what actually determines market returns in the long run.
An interesting academic paper has recently appeared on this topic, written by two behavioural economists from Harvard, Robin Greenwood and Andre Shleifer. The paper, which is well summarised here in Free Exchange in The Economist, discusses the signals that can be derived from investor sentiment and flow data, and then contrasts these results with some standard predictions from the theory of finance. Given their results, some unexplained puzzles remain.
Investor Expectations and Future Returns
It will come as no surprise to market practitioners that Greenwood and Shleifer find that many surveys of investor sentiment are highly correlated with each other, and that they track actual flows of money into equity mutual funds fairly accurately. In other words, when investors are feeling bullish, they allocate more money into stocks!
What is more interesting is that Greenwood and Shleifer show that investors tend to be trend followers when forming their opinions, so they chase rising stock prices and vice versa. One consequence of this is that these investors tend to buy high and sell low, which scarcely seems rational. Furthermore, while it would be nice to think that institutional investors act more rationally than small investors, data from Robert Shiller suggest that the characteristics of the two classes of investors are in fact very similar.
Greenwood and Shleifer then show that investor sentiment is, in fact, a contrarian indicator, with bullish sentiment predicting abnormally low stock market returns over one and, especially, three years ahead. This is worrying, since at present sentiment is bullish and equity inflows are high.
It is also contrary to the predictions of the standard theory of finance, based on rational expectations. In his 2010 Presidential Lecture to the American Finance Association, John Cochrane says that it is extremely well established that asset market returns will turn out to be high when market valuations are attractive (ie when discount rates are high, dividend yields are high and price-earnings ratios are correspondingly low).
In other words, future returns will turn out to be high when they are expected to be high. That occurs when perceived risk is high, so attractive expected returns are needed to compensate for this. This pattern of behaviour is required for investors to fulfill rationality criteria, but it seems to have been directly contradicted by the results of Greenwood and Schleifer, which is why their work will not make welcome reading in Chicago.
If the empirical results of the Greenwood and Shleifer are valid, then investors act systematically in an irrational manner, because they allocate more money to equities when stocks are expensive. This appears to be what they are doing at present, since the price-earnings ratio on US stocks is actually well above its very long-term historical average (though not above its 20 year average). It is not very appealing to believe that investors are systematically irrational, so what is going on?
Time Horizons and Momentum Models
In my view, the explanation boils down to the fact that investors use different methods to derive their asset price expectations over different time horizons. Over long horizons of three years or more, most investors do seem to form their expectations from rational models, in which attractive valuations today (ie low p/e ratios) lead to the expectation of high stockmarket returns over the medium term. Given enough time, that method works.
However, very few professional investors have the luxury of being able to sit and wait for these long-term models to justify themselves, so shorter-term expectations are formed by trading models in which the recent behaviour of asset prices plays a very large role. In other words, price momentum dominates expectations in the shorter term.
The finance literature often describes momentum trading as irrational or even speculative. It certainly seems to contradict the tenets of perfect markets, which hold that past information (such as previous price patterns) should not be able systematically to predict the future.
Yet there is in fact an enormous amount of evidence that momentum trading systems produce attractive risk-adjusted returns over long periods of time. For example, an influential paper, which reports very attractive Sharpe Ratios for momentum trading systems in all asset classes since the mid 1980s. I understand that a recent research paper for clients by AQR Capital Management shows that similar impressive results apply over more than a century [1].
Momentum trading systems are in widespread use, formally or informally, throughout the financial markets, and of course they appear in their purest form in CTA or managed futures hedge funds. (A disclaimer: we use momentum models along with many other approaches at Fulcrum, so I am not an unbiased observer.) These models did extremely well in the financial melt-down in 2008, but most of them have struggled to make high returns in the risk on/risk off environment of 2011-12. Nevertheless, their pedigree over many decades suggests that they should not be discarded too easily.
Conclusions
My conclusions from all this? Recent investor flows into equities, and improving sentiment, are driven mainly by price momentum, which works over short horizons, but says nothing about a longer-term rotation from bonds to stocks. In judging medium-term returns, we have nothing better to rely on than fundamental valuation, which also fits better with the theory of finance. At present, valuation indicates that US stocks (though not European stocks) are fairly expensive compared with their own history, while bonds are extremely expensive. Based on valuation, US stocks should therefore out-perform bonds in the medium term, but overall real returns on both assets in the US may be fairly low.
—————————————————————————————–
Footnotes
[1] It is debatable whether the success of momentum systems demonstrates that financial markets are imperfect. One interpretation, which I discussed in this blog in 2010, is that momentum-driven portfolios result in higher returns than other portfolios because they also involve higher risks.
