How to improve SME performance using iterative random forest in the empirical analysis of institutional complementaritty

Abstract

Empirically investigating the workings of institutional complementarity in organisations has been a challenge in the social sciences domain for a long time. This paper examines data from the World Management Survey (WMS) using a new machine learning method termed as iterative random forest (iRF), which is used in the field of biostatistics. An empirical study of complementarity was conducted in small and medium-sized enterprises using WMS data. The effects of 18 management quality indicators on profitability, growth and viability were examined using machine learning methods (i.e. random forest [RF] and iRF). The analysis revealed the relative importance of whether high performers are properly rewarded, poor performers are reassigned and retrained and the criteria for high and low performance are well established. Furthermore, the study results revealed that the ability to set short-term goals based on a long-term perspective is complementary to many other indicators. These findings are consistent with the findings of a survey study that examined many empirical studies on the workings of institutional complementarity. This indicates that iRF is a credible and promising method for empirical research on institutional complementarity.