The determinants of mergers and acquisitions in banking

This paper investigates the determinants associated with the likelihood of a bank becoming involved in a merger or an acquisition. Using a multinomial logistic regression and a Cox regression with time-dependent covariates, we investigate the determinants of being a target or an acquirer from a sample of 777 deals involving EU acquirers and 312 global targets over the period of 1991 to 2006. Both the multinomial logistic and Cox regressions identify the same determinants associated with becoming acquirers or targets. A higher likelihood of becoming an acquirer exists for larger banks with a history of high growth, greater cost X-efficiency, and lower capitalization. In contrast, banks are more likely to be targets if they have lower free cash flows, are less efficient, are relatively illiquid, and are under-capitalized. But, the predictive power of the two regressions is different as the multinomial logistic regression outperforms the Cox regression when predicting the likelihood of becoming an acquirer.