The Separation of Corporate Ownership and Control : A Reinterpretation of the Evidence of Berle and Means

Berle and Means' classic study of the separation of ownership and control remains authoritative and influential despite having been criticised on various grounds by a number of authors. This paper argues that, firstly, the Berle and Means approach to determining company control implicitly assumes a static framework inappropriate to analysing early twentieth century corporations. Secondly, accepting their approach, their control-type criterion in terms of shareholding concentration is too high and biases their result. Thirdly, the use of the same criterion for all companies fails to recognise the importance of shareholding dispersion and further biases their results. A probabilistic-voting model is described which makes explicit the assumptions behind the concept of factual control. This is applied to the data on 16 companies used by Berle and Means and their classification of them as managerial is shown to be invalid.