In this analysis, I investigate the causes of early elections in four parliamentary democracies across the world: Great Britain, Australia, Japan, and New Zealand. While I consider a number of explanations for the decisions to hold early elections, I find most theoretical and statistical support for Smith's (2003; 2004) informational thesis. He maintains that governments look to future economic conditions when making their timing decision. This approach, however, also leaves open the possibility that other, non-economic factors can explain why prime ministers call elections earlier than is necessary. I argue that the degree of disproportionality, the measured gap between a party's vote share and seat share, is a key attribute to explain the early election decision. When prime ministers weigh their decision to dissolve government, they cannot assess the effect of changes in their support in the population as accurately when a high degree of disproportionality is present. Using survival analysis, I find some support for a comprehensive attributes and events approach. New Zealand proves an exception; governments tend to fail sooner when high levels of disproportionality are present. This appears to be a result of particular factors related to disproportionality as a political issue, leading to electoral reform in 1996.

In the models analyzed in this paper, there exists an incumbent politician with one objective, two choices, and voters who remember the past differently. The politician's primary goal is to get reelected, which is done by maximizing the number of votes on the day of election. The politician can increase their chances of reelection if they influence the state of the economy over time and ensure the economy is in its 'best' state on the days leading up to the election. In conducting this research, I wanted to study how different rates of memory decay influences the choices the politician makes during the course of their term. Also, I wanted to explore how long a politician would wait to have an election if that were a choice they could make. I found that voters who remember more of the past place a greater constraint on the incumbent leading to moderate fluctuations in the economy and frequent elections.