Bank efficiency measures, M&A decision and heterogeneity

The empirical literature has obtained mixed results regarding the probability for more efficient banks to be bidders in merger and acquisitions (M&A) operations. From an econometric point of view, this might be attributed to an inaccurate control of unobserved bank heterogeneity that can bias parameter estimation severely. In this paper, we adequately control for unobserved heterogeneity through a finite mixture, random parameters logistic model, and we estimate the probability for a bank to be a bidder in an M&A depending on its ex-ante efficiency, therefore avoiding any parametric assumption on the distribution of the random effect. This leads to a likelihood function defined as the integral of the kernel density with respect to the mixing density, which has no analytical solution. For this reason, we approximate the integral with a finite sum of kernel densities, each one characterized by a different set of model parameters. We then obtain a set of non-overlapping clusters with matching values of ex-ante efficiency, and assign each bank to a cluster based on the estimated posterior probability of it being in that cluster. Moreover, in our analysis we use two different sets of measures of bank efficiency, obtained using parametric as well as semi-parametric techniques. Our results are based on a sample of 612 banks, from 34 countries, between 1991 and 2006. They show that, considering unobserved heterogeneity, cost efficiency has a major impact on the probability for a bank to bid in a cross-border M&A, but no effect in the case of domestic M&A.