We decompose the redistributive effect of direct taxes into vertical, horizontal, and reranking components applying the methods of Urban and Lambert (Public Finance Review, 2008). In the first such application to the UK, and using yearly data covering 1977–2020, we find that redistributive effect increased over the period. However, there is no clear trend in horizontal inequity and this component forms a very small fraction of total redistributive effect by comparison with reranking and especially vertical components. It is also the vertical component that best tracks trends in redistributive effect. We give specific attention to the choice of the bandwidth used to define 'close equals' in terms of pre-tax income. We also show that implausible estimates of the horizontal inequity component arise for some years regardless of bandwidth used.