This study examines intertemporal variation in the associations of accounting data with subsequent firm returns. A number of accounting research studies pool data indiscriminately across time and firms. Previous research has disclosed the nature and effects of cross-sectional dependencies in pooled data. On the other hand, intertemporal dependencies associated with real macroeconomic phenomena have not been widely researched. The objective of this study was to provide evidence as to whether accounting data's associations with subsequent firm returns systematically vary across recession-associated and expansion-associated valuation change periods. Eighty-two accounting ratios were examined for evidence of systematic variation in association across business cycle-associated valuation events. Analyses are conducted, using both simple and multiple regression. Business cycle effects on the predictive accuracy of regression models were also examined. / Ph. D.