BOOTSTRAPPING THE NUMBER OF FACTORS IN THE ARBITRAGE PRICING THEORY

The number of factors in the APT are re-examined through a new methodology called the bootstrap, which provides a nonparametric alternative to the chi-square test used in prior research. Results suggest that the number of statistically significant factors does not increase when the number of firms increases. Moreover, only the first factor is consistently significant across sample sizes of thirty, sixty, and ninety firms.