Firm Size and R&D Spending: Testing for Functional Form

One of the most frequently investigated topics in the economics of technological change field is the relationship between firm size and the corresponding level of research and development (R&D) spending. For more than 30 years, researchers have been examining numerous functional forms in an effort to determine whether larger firms spend more on R&D relative to their size than do smaller firms. A double-log regression model has evolved to be the accepted specification. However, the statistical appropriateness of a particular functional form has not been investigated. Tests for functional form are run using an extended Box-Cox model. Results of analyzing 1985 firm-level industrial data for 9 separate industry groupings indicate that the double-log specification is the most appropriate within the class of models represented by the Box-Cox transformation.