Yves Nosbusch





Une question ? Une remarque ?
Contactez-nous !


Imprimer l'article
Imprimez l'article



Stock market predictability: An assessment of the evidence


following the 2013 Prize in Economic Sciences in Memory of Alfred Nobel



The predictability of asset prices and therefore of returns on different types of investments has been a central question of finance for centuries. The laureates of the 2013 Prize in Economic Sciences in Memory of Alfred Nobel, Eugene F. Fama and Lars Peter Hansen of the University of Chicago and Robert J. Shiller of Yale University, have deeply advanced our understanding of this issue which is at the heart of professional investment management. So what better occasion could there be to take stock of what we have learned from the intensive research efforts on this topic over the last few decades. The discussion here will focus on the stock market but there is a large amount of related evidence for other asset classes, particularly bonds and currencies.



stock marketFirst, let's take a step back. Should we expect predictability in asset returns? The concept of efficient markets was introduced by Eugene Fama in a celebrated 1970 article. The basic idea is that at any point in time, prices should fully reflect all available information and that there should therefore be no predictability of excess returns on different assets, at least in the short term. Fama distinguishes between three types of efficiency: weak form, semi-strong form and strong form, depending on the type of information which is available – whether it is past prices, public or private information. In practice, research has focused on the first two forms, partly because securities laws do not allow trading on inside information.



The principal way to test this hypothesis has been through so-called event studies. There was an explosion of these studies in the 1970s. Broadly speaking, the evidence gathered during that decade was supportive of the Efficient Markets Hypothesis: for a host of corporate events, the cumulative abnormal (i.e. adjusted for risk) returns after the announcement day were found not to be statistically different from zero. The famous quote by Michael Jensen from 1978 “I believe there is no other proposition in economics which has more solid empirical evidence supporting it than the Efficient Market Hypothesis.” thus largely reflected the prevailing spirit at the time.



In light of this, Robert Shiller's work in the late 1970s and early 1980s seems all the more remarkable. His 1981 paper entitled “Do stock prices move too much to be justified by subsequent changes in dividends?” which continues to be his most cited article to this day, was truly visionary. The paper answered the question posed in the title in the affirmative. This finding is of course interesting in its own right but more importantly, it has profound implications for the long-run predictability of the stock market which formed the basis of a research program which Robert Shiller embarked on with John Campbell in the 1980s and which led to a number of highly influential co-authored articles published in the late 1980s. In particular, these papers examined the predictive power of aggregate valuation ratios like the price-dividend or price-earnings ratios for long-run aggregate stock returns. A by-product of this research was the popularization of the so-called cyclically adjusted price earnings ratio (CAPE). This measure, which scales aggregate stock prices by a 10-year trailing average of aggregate earnings, has become a widely followed indicator of the valuation of the overall stock market.



In parallel, Lars Hansen developed a set of econometric techniques which can be used to test Robert Shiller's claims from the early 1980s in more general settings. These extremely powerful techniques, particularly the so-called Generalized Method of Moments, have since found numerous other applications and continue to this day to form the backbone of research in empirical finance.



While Robert Shiller's work discussed in the previous paragraphs has been mostly to do with returns on the aggregate stock market, a separate strand of the literature has focused on the cross section of stock returns, i.e., the relative performance of different groups of stocks. Here again Eugene Fama has made numerous pathbreaking contributions. In fact it would probably be fair to say that he has - together with a number of co-authors – to a large extent shaped the field of cross-sectional empirical asset pricing as we know it today. His perhaps best-known contribution in this area is the development of the so-called Fama-French factors. The idea of his work with Kenneth French in the early 1990s is to sort portfolios of stocks according to their size and book to market values and to calculate excess returns on portfolios of small stocks over big stocks (giving the SMB factor) as well as excess returns on portfolios of stocks with high book to market values over those with low book to market values (giving the HML factor). It turns out that these two factors add a lot of explanatory power to a simple factor based on the excess return on the market (the pricing factor implied by the Capital Asset Pricing Model). This finding relates to the well-documented size and value effects. The size effect refers to the observation that historically small stocks have tended to outperform large stocks (on a risk-adjusted basis) whereas the value effect refers to the observation that historically value stocks have tended to outperform growth stocks (again on a risk-adjusted basis). Another important empirical regularity documented by authors like Narasimhan Jegadeesh and Sheridan Titman or Mark Carhart is the so-called momentum effect. This refers to the observation that historically stocks which have outperformed in the recent past have tended to continue to do so in the near future and vice versa for underperformers. In a 2012 paper, Eugene Fama and Robert Litterman state that “Of all the potential embarrassments to market efficiency, momentum is the primary one.


In summary, the evidence accumulated since the early 1980s points towards predictability, both at the level of the aggregate stock market as well as in the cross section of individual stocks.



What should we make of this evidence? Broadly speaking, the possible explanations of predictability fall into three categories. A first possibility is that it is just the result of data mining. In other words, if a large number of analysts look at a given set of data long enough, they will end up detecting patterns which just happen to be a fluke of the particular sample at hand. If this is the case, one would expect these patterns to disappear over time. A second set of explanations, sometimes referred to as behavioural, rely on some form of irrationality on the part of investors. To put it simply, the idea is that some investors make mistakes which can distort asset prices and move them away from fundamental value for a period of time. Finally, there are rational risk-based explanations. Here the idea is that investment strategies with predictably higher average returns simply expose investors to higher risk so that the excess return earned on average is simply compensation for taking on more risk.



The debate on which of these explanations is most important has not been settled by the literature with both the behavioural and rational camps still adding new evidence and risk models supporting their respective cases on a regular basis.



One thing however is clear. The research of this year's Nobel laureates has profoundly changed the way not just academics think about predictability of asset prices but also how investment professionals make assessments in the markets every day. For instance, in the area of long-short equity strategies, the Fama-French three-factor model (augmented by Carhart's momentum factor) has become a key benchmark for the performance evaluation of hedge fund managers. And of course the cyclically adjusted price earnings ratio popularized by Robert Shiller is one of the most widely followed indicators of aggregate stock market valuations.




Article completed on December 4 2013 by Yves Nosbusch, Chief Economist of the Bank

Published in the Agefi in December 2013