The impact of Fintech on bank development: A meta-analysis investigation

To test whether the results of the empirical literature on bank Fintech are affected by the characteristics of specific research, the paper selects existing high-quality empirical literature to conduct a meta-analysis. It finds that the empirical estimation results conducted across various studies are influenced significantly by factors such as sample interval, estimation methods, measurement indicators for Fintech and bank development, and the inverse of the model count. Specifically, the probability of obtaining significant estimation findings increases with earlier sample start-time and the use of risk or Fintech index data; however, the inverse is true for more models adopted, or the use of dynamic panel estimation methods. Meanwhile, the probability of obtaining significant positive estimation findings increases with the wider sample coverage. Furthermore, estimating methods, and measurement indices for Fintech or bank’s risk, all significantly contribute to the significant negative estimation results. Moreover, the funnel plot asymmetry analysis reveals the existence of publication biases in the sample research; the greater the significance of the empirical estimation results, the higher the probability of publication for the article. Therefore, it is vital to consider the heterogeneity and possible publishing biases among the extant empirical research when examining the impact of Fintech on bank development.