by Dr Yuan Li, Research Associate, Cambridge Centre for Finance and Cambridge Endowment for Research in Finance
![Yuan Li](https://www.jbs.cam.ac.uk/wp-content/uploads/2020/05/li-yuan-229x205-1.jpg)
In 2013, the Nobel committee split the economic prize to Eugene Fama – the pioneer of efficient market hypothesis (EMH) and Robert Shiller – the critic of EMH. This decision indicated that the Nobel committee agreed with both Fama and Shiller. Was the committee right? The answer is yes, according to my findings from a recent research project.
Fama explains EMH as “the simple statement that security prices fully reflect all available information”. The empirical implication of this hypothesis is that except beta (the measurement of a firm’s systematic risk), no other publicly available information can be used to predict stock returns. However, the finance literature has found that many easily available firm characteristics, such as market capitalisation, book-to-market ratio, etc, are related to future stock returns. They are the so-called anomalies. Does the discovery of anomalies reject the EMH? Not necessarily. Because no one knows what a firm’s beta should be, and those firm characteristics can simply be proxies for beta. This is known as the joint hypothesis problem. We can say nothing about EMH unless we know what the correct asset pricing model is. Sadly, we do not know what the correct asset pricing model is.
In this project, I get around the joint hypothesis problem. I assume that a firm’s stock return is composed of two parts: risk-induced return and mispricing-induced return. Because of the joint hypothesis problem, we do not know what the risk-induced return is. However, we can estimate the mispricing-induced return (if there is any) using the forecasts issued by financial analysts. Analysts’ earnings forecasts represent investors’ expectations. More importantly, we know the actual earnings of a firm, and hence we can calculate the errors in analysts’ forecasts, which represent investors’ errors-in-expectations. We can then estimate the returns generated by investors’ errors-in-expectations, that is, the mispricing-induced return. If the market is perfectly efficient, the mispricing-induced return should be zero. I calculate the fraction of an anomaly explained by mispricing as the ratio of mispricing-induced return over the observed return. The fraction of an anomaly explained by risk is thus one minus the above ratio.
I examine 195 anomalies. On average, the fraction explained by mispricing is 17.51 per cent, suggesting that the major fraction of anomalies is not anomalous at all. This result may be disappointing to EMH critics, who seem to think that the stock market is extremely inefficient, and it is very easy to profit from anomalies. However, the good news to EMH critics is that the fraction explained by mispricing varies widely across different anomalies. In particular, the momentum anomalies are almost completely explained by mispricing. Hence, trading on momentum anomalies is likely to generate abnormal returns. In contrast, the high returns from the value strategies are almost entirely compensations for bearing high risk.