Purpose This paper aims to use the institutional and information-processing perspectives to explore their association with between internationalization and the Penrose effect phenomenon for business groups (BGs). Design/methodology/approach The authors use ordinary least squares regression models to test arguments about data pertaining to 101 Taiwanese BGs’ foreign direct investments. Findings The results indicate that greater levels of depth and scope in the process of internationalization during one period may negatively affect rates of growth in the following period. The results further demonstrate that institutional distance moderates the relationship. Research limitations/implications Using the perspective of information-processing demands, the authors provide alternate explanations regarding the relationship between the process of internationalization (depth, scope and rhythm) and the Penrose effect. Originality/value Owners and managers should focus on both the depth and the scope of internationalization. BGs are likely to incur high dynamic adjustment costs, which then limit the rate of BGs’ growth. Managers should balance international market uncertainty with current managerial resources when determining how deeply and broadly to expand internationally and where to enter. In addition, as recent major panel studies suggest, management capabilities and practices can improve significantly, which has a positive effect on firm growth and performance. This does require the careful development and acquisition of the managerial resources needed for internationalization.

This study applies resampling data envelopment analyses (DEA) and includes two estimation models: Triangular and Past–Present Models, to explore the operating efficiency of 37 commercial banks in Taiwan from 2014 to 2016. The empirical results are as follows: 1. No matter which models were used, the most inefficient banks from 2014 to 2016 were Bank of PANHSIN. 2. Resample model can more accurately estimate the relative efficiency of the banks than the slacks‐based measure (SBM) model. 3. Based on the Resample model, the empirical results of the SBM model show that the overestimation of the efficiency ranking is more serious.