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

This paper investigates institutional complementarity in small and medium-sized enterprises (SMEs) using World Management Survey (WMS) data and a novel machine learning method: iterative random forest (iRF). The study examines the effects of 18 management quality indicators on profitability, growth, and viability. iRF analysis revealed the importance of rewarding high performers, reassigning/retraining poor performers, and establishing clear performance criteria. The ability to set short-term goals aligned with long-term perspectives showed strong complementarity with other indicators, consistent with previous research. The findings suggest iRF as a promising tool for analyzing institutional complementarity.

Atsushi Sannabe