A firm's cost of equity capital is an opportunity cost, reflecting what the shareholders can expect to obtain elsewhere for the same level of risk. I t follows that enquiry into the cost of equity capital a t the corporate (prepersonal tax) level must consider the implications of shareholders' individual assessments of their required returns, made on a post-personal tax basis. The difficulty encountered in formulating such a framework is that the relationship between earnings a t the corporate level and a t the post-personal tax level ii dependent on how shareholders receive their return: via dividends less income tax or via capital gains less capital gains tax. One way of avoiding the problem (the US preferred approach) is to assume that investors receive their returns as some combination of dividends and capital gains that are liable to some overall or 'effective' rate of taxation. The relationship between a firm's returns on a net-of-corporate-tax and on a net-of-personal tax basis is therefore established, at least algebraically; and the analysis for the impact of investors' expectations on a firm's cost of capital is able to proceed. However, such an overall or 'effective' rate of taxation fails to exploit the essential interrelationship between a firm's dividends and capital gains. Subsequent analyses thereby fail to provide an insight into the nature of the structural problem faced by firms in their bid to maximise their shareholders' afterpersonal tax returns.In the context of a UK imputation tax system, the problem has been explicitly recognised by Ashton (1989aAshton ( , 1989bAshton ( and 1991. Ashton argues that it may be reasonable empirically to assume that td = ti (where t d , tb represent, respectively, the effective aggregates of investors' marginal tax positions on their dividends and capital gains), so that the problem of how shareholders receive their returnas dividend or capital gainis avoided. Ashton refers to the position as the assumption of the tax irrelevancy of dividend policy.