This paper examines the effect of mergers on the performance of banks. We use a unique and exhaustive panel data set of mergers of Austrian banks covering the period from 1996 to 2002. A probit selection equation is formulated to explain the adoption of a merger strategy. We use various matching techniques to estimate the treatment effects of bank mergers on the banks' performance. The analysis provides evidence in favour of the view that there are longer lasting positive effects on bank performance, especially, in terms of improved cost efficiency. The findings also suggest that pre-merger effects are likely to occur in terms of higher cost efficiency immediately before the establishment of the merger. Finally, smaller banks involved in merger activities are more likely to enjoy cost-efficiency gains earlier than larger banks.