Contrarian investment strategies in a European context

During recent years contrarian investment strategies have gained a lot of attention in the academic literature. A contrarian investor acts in a way opposite to what the majority is doing at the time. His trading strategy is premised on an assumption of a negative serial correlation in market prices. A contrarian believes that investors put too much weight on recent news and too little weight on historical information. One of the contrarian strategies calls for buying stocks with low prices relative to value measures such as earnings, cash flows, book values and dividend yields. Such strategies are generally referred to as value strategies. These are the opposite of glamour strategies which call for buying stocks with high prices relative to these measures. Until now contrarian investment strategies have mainly been studied for the United States and Japan. In this paper we have studied value strategies for four European countries, i.e. France, Germany, the Netherlands and the United Kingdom. An important problem in constructing portfolios of stocks of firms from different countries and/or different branches of industry is that firms from the same branch of industry and/or the same country tend to have clustered relative value rankings. Because of this, return differences between different value portfolios may be attributed to differences in country and/or industry performance. In this paper we have presented a methodology to correct for these biases by sorting stocks by relative ratios. In our empirical study we have constructed five portfolios for four relative value variables, i.e. the E/P ratio, the GF/P ratio, the B/M ratio and the dividend yield (Yld), using data from June 1982 to June 1993. We have found that the hedged returns for the value portfolios outperformed the hedged returns for the glamour portfolios. This difference turned out to be especially remarkable for the GF/P ratio (20.8%). In a regression analysis, in which all four value variables were taken into account, as well as a correction for the size effect, we find a difference of 11.8% for the GF/P ratio. We have also shown that this result cannot be explained by risk differences alone. Reasons for this are that the value strategies did not lead to an underperformance in bad years and that the differences in standard deviations of the stock could only explain a small part of the return differences. These results confirm the results earlier found by Chan et al. (1991) for Japan and by Lakonishok et al. (1994) for the United States. The explanations that can be given for the outperformance are likely to be the same as the explanations given by Shefrin and Statman (1995) and Lakonishok et al. (1994) for the outperformance of contrarian investment strategies in the United States. Shefrin and Statman (1995) demonstrate that good companies are companies with low B/M ratios. They argue that, according to the behavioural asset pricing theory (see Shefrin and Statman, 1994), noise traders make cognitive errors that lead to the belief that good stocks are stocks of good companies. Therefore, these investors prefer glamour strategies over value strategies. Information traders, such as money managers of investment or pension funds, do not nullify this effect through arbitrage. The reason for this is that they also have a preference for glamour strategies, because their clients are more forgiving of losses on stocks of good companies than of losses on stocks of bad companies (see Shefrin and Statman, 1995). Lakonishok et al. (1994) present a similar explanation. They argue that institutional investors may prefer glamour stocks because 'they appear to be "prudent" investments, and hence are easy to justify to sponsors'. This last argument may be especially relevant for one of the countries in our study, i.e. the Netherlands. In this country a great deal of savings is in the hands of very large pension funds. These pension funds have difficulties in persuading supervisors to give them permission to increase their investments in shares. Therefore these pension funds will be motivated to show that they are 'prudent' investors. Lakonishok et al. (1994) mention that they will focus their future research on the question whether there is a relation between the inferior performance of pension funds relative to the market and the outperformance of value strategies in relation to glamour strategies. We argue that such an investigation may be particularly interesting for a country such as the Netherlands, where large investments are in the hands of pension funds, which have difficulties in convincing supervisors to let them increase their investments in stocks.