The value premium and methodological biases: evidence from the UK equity market

We look for the existence of a value premium in the UK equity market for the period of 1987-2002. Previous studies are subject to four methodological biases (1) survivorship bias, (2) look-ahead bias, (3) a downward bias in post-formation growth stock returns caused by excluding recently listed growth stocks from the data, and (4) an upward bias in post-formation value stock returns caused by computing long-term returns by cumulating monthly returns (Conrad and Kaul, 1993). We eliminate the first three biases by using a new survivorship bias and look-ahead bias free dataset, which contains a complete history of all UK stocks that were fully listed at any time during 1987-2002. We eliminate the fourth bias by computing post-formation holding period returns, as recommended in Conrad and Kaul (1993). Our results indicate that the value effect is far smaller than is reported in previous studies, and is neither economically nor statistically significant.