This thesis analyses the Monopolies and Mergers Commission's investigations of referred bids over the period 1974-84. The purposes of undertaking this study were as follows. First, to discuss the practice of merger policy and the framework within which it has been set. Second, to compare this with the theoretical and empirical debates which have been developing in relation to mergers and merger policy. Third, to analyse the extent to which a predictive model could be constructed so that significant elements of the public interest may be identified. It was shown that it is official policy to give mergers, in general, the benefit of the doubt. This is inspite of the fact that most of the evidence involving merger activity has tended to be critical of their impact. In addition, it also appears that market discipline does not drive merger activity. In terms of interpreting the Monopolies and Mergers Commission's reports, probit analysis was unable to identify elements of the public interest which appeared to be consistently important to the Commission. In addition, univariate and multivariate analysis showed that it was virtually impossible to distinguish between raiders and targets. Thus, as it stands, merger policy does not provide meaningful, or clear, signals to either potential raiders or targets. Current policy deems that, under certain circumstances, mergers are undesirable. The problem is that the circumstances are unknown. Thus published guidelines are essential given the need for a merger policy. Further, we conclude that a stronger policy is necessary. In particular it should be incumbent upon firms to show expected benefits rather than a lack of detriments.