A Functional-Form-Free Test of the Research and Development/Firm Size Relationship
提出一种无需预设函数形式的检验方法，用于分析研发投入与公司规模的关系。研究发现，某些行业中研发强度随公司规模变化，且多为负相关。
Abstract Parametric testing of the relationship between research and development (R&D) and firm size is limited by the hypothesized functional form, predominantly assumed to be linear in the logarithms of both variables. With a transformation of variables, the null hypothesis of constant returns to scale can be tested employing the multiple-rank F-test procedure. The findings of this analytical approach suggest that R&D intensity varies with firm size in some industries but not in others, and where it does vary, there is more evidence that R&D intensity is negatively rather than positively related to firm size. KEY WORDS: NonlinearityRank F testRank regression