Empirical tests of the capital asset pricing model (CAPM) have shown that movements in a single factor market index poorly explain returns on individual securities. Assuming that a possible reason is the existence of pricing anomalies, more recent research has extended the model by adding new factors to explain systematic variations in stock returns and achieved significant results. In particular, securities with high book to market ratios (book equity divided by market equity, BE/ME) appear to command a value premium. One suggestion is that the BE/ME premium compensates for the systematic risk associated with financial distress. It has been argued that industries experience periods of growth and financial distress and it has been reported in the US literature that the BE/ME premium is not stationary. To explore this proposition, this paper examines the behaviour of CAPM and three-factor models when tested using portfolios of different industries through time. The results show that the three-factor model provides a better explanation of industry returns than the CAPM. Loadings on the BE/ME premium are not stationary and appear to vary for the different industries through time. Monthly returns in excess of the one month t-bill tend to be generally higher for the Basic industrial sector through time.