IN recent years a number of academic studies in industrial economics and company finance have measured the profitability of companies before and after merger. Sometimes their aim has been to assess the gain from merger to particular groups of wealth-holders, such as the acquirer's equity owners; and at other times the exercise has been directed at assessing the impact of merger on efficiency, often as part of a discussion of whether government should foster or inhibit merger. This paper is concerned with the second type of exercise and aims to describe some of the pitfalls which can be encountered by a researcher who is trying to draw inferences for efficiency change from the observed post-merger change in common measures of profitability.