This paper develops a theory of political timing. It argues that the timing of those political `events' over which politicians have some discretion is non-random and that politicians attempt to influence the timing of events in such a way as to maximize the political benefits or minimize the political costs for themselves. In relation to `good news', it is argued that politicians will attempt to increase the public attention given to it or the positive impact felt by electors and to reduce public attention, or the impact on welfare, of `bad news'. Consideration of memory research, prospect theory and structural factors in the media suggests ways in which politicians can use their discretion over the timing of events to further these ends. Four specific behavioural hypotheses are advanced, regarding the way discretion will be exercised by politicians, according to whether the `news' is, or is likely to be, good or bad, and the resulting patterns in the timing of events. The behaviours are labelled respectively as packaging, splitting, highlighting, and phasing. Data on the timing of three very different types of phenomenon - British parliamentary by-elections; the timing of announcements of increases in medical prescription charges in the British National Health Service; and the timing and nature of changes in official lending rates in the United Kingdom and the United States - are chosen to illustrate the hypotheses.