Purpose - A growing strand of literature has focused on the returns performance of zero dividend stocks. This paper seeks to provide new evidence on the link between dividend payment and returns history and firms' subsequent stock market performance. Design/methodology/approach - Prior research draws a distinction between those stocks which have never paid dividends and those which formerly paid dividends but subsequently ceased, implicitly using the latter as a measure of financial distress. The analysis is expanded beyond the role of payment history, and also the importance of earnings and past returns in the performance of UK zero-dividend stocks is considered. Findings - In contrast with the prior US evidence, it was found that payment history is not a significant determinant of returns, while past returns play a far greater role. However, this explanatory power seems to be diminishing over time. Research limitations/implications - The paper extends the existing knowledge about the behaviour of UK zero dividend stocks and establishes a benchmark for future research in the area. Practical implications - The prevalence of zero dividend stocks has grown over time. In the US, the proportion of listed companies paying dividends fell from 66.5 per cent in 1978 to only 20.8 per cent in 1999. In the UK, the proportion of companies omitting dividends rose to 25.2 per cent in 1999, from previous recession-year high points of 16.1 per cent in 1982 and 17.9 per cent in 1992. This paper provides new evidence of practical value to investors in this class of stocks. Originality/value - The paper presents original research in the area of dividend policy and stock returns, which should be of interest to both academic and practitioner audiences.