Using Canonical Correlation to Identify Arbitrage Pricing Theory Factors

This paper presents a new research design to test the efficacy of the Arbitrage Pricing Theory of Ross [1976], similar to that applied by Christofi, Christofi and Philippatos [1993]. In particular, we use a combination of factor analysis and canonical correlation to test the underlying relationships between APT factors developed using factor analysis and unanticipated changes in five macro-economic variables that have been shown to be related to stock returns. The results of this paper indicate that the first factor of industry returns is strongly related to the S&P 500 while the remaining four factors are highly correlated with the term structure of interest rates, the rate of inflation, the default premium, and the industrial production, respectively.