‘European Price Momentum and Analyst Behavior

Previous studies have found evidence that selecting stocks with positive price momentum is effective in the U.S., European, and emerging stock markets for periods up to a year. The reasons that historical price momentum forecasts the direction and magnitude of stock returns, however, are not clear. Insight into the determinants of price momentum would allow investors to judge whether and how price momentum should play a role in their investment strategies. Studying the European stock markets, we found that positive price momentum is caused by analyst underreaction to new earnings information. We found earnings surprises, expected earnings growth, and earnings revisions to be systematically related to historical price movements. Importantly, the data show that European price momentum is distinct from the widely documented value and size effects. Our findings clarify the benefits of assessing analyst behavior to predict whether momentum investing might work in the next period.

Price momentum refers to a tendency of stock prices to continue on the same path from one period to the next. The existence of such return continuation for periods up to a year has been documented for the U.S., European, and emerging equity markets in studies conducted by both academics and practitioners. The cause (or causes) of price momentum, however, is subject to much controversy. Some argue that it is a manifestation of slow reaction by analysts to earnings-related news. The study reported in this article provides empirical evidence, based on European data, of the validity of this hypothesis.

We used a comprehensive sample of all European stocks that had at least one earnings forecast in the I/B/E/S database spanning the period February 1987 and June 1999. We confirmed that price momentum strategies were profitable in the sample period. This finding is robust to corrections for size effects, value versus growth effects, estimated earnings-growth effects, and the effects of country-specific risk.

To test the underreaction hypothesis, we examined the relationships among earnings surprises, expectations for future earnings growth, earnings revisions, and the returns from momentum investing. The primary finding of this study is that a pessimism bias exists for portfolios composed of strong-price-momentum stocks and an optimism bias exists for portfolios consisting of weak-price-momentum stocks. The optimism bias is definitely more substantial than the pessimism bias; underreaction of analysts to earnings-related news is likely. We based this conclusion on the observation that earnings surprises are positive for strong-momentum stocks and negative for weak-momentum stocks. Moreover, just after an earnings announcement by a company, the expected growth in earnings is more overestimated for weak-momentum stocks than it is for strong-momentum stocks. The revisions were generally downward but were more downward for weak-momentum stocks.

Furthermore, we found that the widely documented value and size anomalies are distinct from the price momentum anomaly. An almost perfectly linear negative relationship exists between price momentum and the book-to-market ratio, whereas if the momentum and the book-to-market effects were the same phenomenon, we would expect a positive relationship. We found a relationship between market capitalization and momentum to be either absent or positive, which is inconsistent with the hypothesis that these two anomalies are the same phenomenon.

When we examined how robust the results are to some changes in sample selection and the method of portfolio construction, we found that the empirical results were not driven by one particular European country and were similar for the first and second halves of the sample period. In addition, the results are robust to alternative currency assumptions for the returns and to alternative methods of portfolio construction.

Based on our findings, price momentum strategies were profitable in the European markets in the period we studied, but these strategies may produce incremental returns only as long as underreaction to earnings information persists. Therefore, this study implies that to monitor the risks of using a price momentum strategy, investors should continuously observe analyst behavior and changes in behavior. Indicators for underreaction are earnings surprises within the price momentum portfolios and autocorrelations in earnings-forecast revisions. Whether price momentum strategies will continue to generate excess returns after a significant change in analyst behavior depends on the existence of other causes for price momentum—which calls for further research.