The purpose of this study is to explore the complex impact of research and development (R&D) on firm performance, as represented by firm capabilities. Specifically, it focuses on predictive impact analysis and delineates the versatile impact patterns of R&D, contingent on resource conditions and the level of capabilities.
This study presents an innovative three-stage data envelopment (DEA)-Tobit regression (Tobit)-multi-layer perceptron neural network (MLP) approach. The integrated analytic process includes an evaluation of firm capabilities (DEA), explanatory analysis (Tobit) and predictive pattern analysis (MLP).
This study finds that the effect of R&D on firm capabilities is significant, in general, with a quadratic effect (U-shaped). However, the delineated impact analysis reveals that the effect size varies, contingent on firm-specific conditions and the divergent impact patterns showing that one-size-fits-all solutions are not applicable.
The study’s methodological advancement provides industry managers with a meaningful decision-making aid. By understanding these complex impact patterns, managers can take advantage of what-if scenario testing for resource deployment and pursue feasible options best suited to firm-specific conditions.
By proposing a cascading analytic process, this study presents an innovative empirical approach in that it delves into the complex impact mechanism of R&D on capabilities as a holistic performance measure. Unlike prior efforts, this study focuses on capturing asymmetric impact patterns and predicting the interpretable effect size, which holds significant pragmatic value.
