The economics of elections

Abstract

The interconnectedness of the election cycle and the economy has been of interest to economists and political scientists alike for many years, yet little empirical analysis of this association has occurred in the Australian context. The economic literature suggests that there exists a bi-directional relationship between economic variables and elections. First, economic variables are deemed to affect government election outcomes due to the importance of the economy on the welfare of individuals within it. Secondly, elections are hypothesised to affect economic instrument and target variables. Given that voters place considerable weight on certain economic variables when making voting decisions, governments may wish to try and use their policy instruments to favourably manipulate target economic variables around election periods. This thesis provides a thorough empirical analysis of this bi-directional relationship in the Australian political context. Unlike previous Australian studies, the analysis is extended from the Federal level to the States within Australia. This allows for an assessment of fiscal-federal issues related to elections within the Australian economy. With respect to the effect of economic variables on election outcomes, several key results are found in the analysis undertaken. First, two previously unassessed variables (the Term and Opposition variables) are found to have significant effects on election outcomes. Such findings have substantial implications on currently held views within the literature. In particular, the significance of the Term variable calls into question the interpretation of the Honeymoon variable suggested by Cameron and Crosby (2000). Secondly, Federal voters appear to be more concerned with inflation and unemployment rather than economic growth. However, when voting in State elections constituents consider the State levels of economic growth when formulating their voting decisions and not State specific levels of inflation and unemployment. When analysing the effects of elections on economic variables, no such differences between Federal and State results are found. In general, no systematic fluctuations of economic variables are found around election periods. Furthermore, no evidence is found for permanent, partisan differences in economic variables when ideologically differing governments are in power. However, temporary changes in economic variables are found when there is a change in government, suggesting that unexpected changes in government give rise to temporary real changes in economic variables. In addition, the conditional variance of the official cash rate is found to be significantly lower in the four quarters prior to, and following elections, suggesting that central bank decisions are affected by elections.