German banks experienced a merger wave throughout the 1990s. However, the success of bank mergers remains a continuous matter of debate. In this paper we suggest a taxonomy as how to evaluate post-merger performance on the basis of cost efficiency (CE). We categorise mergers a success that fulfill simultaneously two criteria. First, merged institutes must exhibit CE levels above the average of non-merging banks. Second, banks must exhibit CE changes between merger and evaluation year above efficiency changes of non-merging banks. We employ this taxonomy to characterise (successful) mergers in terms of various key-performance and structural indicators and investigate the implications for three important policy issues. Our main conclusions are twofold. First, approximately every second merger is a success. Second, the margin of success is narrow, as the CE differential between merging and non-merging banks is one percentage point.