Measures of firm performance, earnings changes, and the prediction of stock returns

This study extends the existing research on the use of financial statement variables to predict one-year-ahead earnings changes. A principal component analysis was conducted on 61 financial statement variables in an attempt to describe the dimensionality of the variables and facilitate the development of parsimonious earnings prediction models. This study finds that the 61 variables embody a much richer array of information than suggested by previous research. The variables could not be described by a small number of principal components.