The method of electing the head of state in a parliamentary system is a critical constitutional matter. A popular argument made is that allowing direct presidential elections strengthens democratic practices. Another argument posits that multiplying the number of political contests may fatigue voters and decrease their participation levels. This article considers electoral turnout in a global sample of parliamentary democracies with a nonhereditary head of state from 1945 to 2006 and finds that direct presidential elections decrease turnout in parliamentary elections by about 7 percentage points. This effect is stronger than that of most existing explanations of turnout.
Keywords: voter turnout; direct presidential election; voter fatigue; semipresidential system; parliamentary system
The number of parliamentary republics in the world is constantly growing. One commonly recurring political debate when setting up or reforming these systems is over whether or not the public should directly elect the head of state. Fierce policy debates over presidential selection procedures have recently taken place in a diverse range of countries, including Austria, the Czech Republic, Estonia, Germany, Hungary, Ireland, Italy, Moldova, Slovakia, and Turkey. Of all constitutional issues deliberated during the regime transitions in Eastern Europe, the question of whether or not to directly elect presidents received the most attention in a majority of countries. This issue is also critical in advanced democracies: Australia, for example, remains a constitutional monarchy after republicans in favor of direct elections caused the 1999 referendum proposing parliamentary selection of the president to fail.
Several issues are raised in policy discussions on the advantages and consequences of holding popular elections for heads of state. A common argument among those debates is that direct elections should be favored over indirect ones because the former strengthen democratic practices and increase political involvement.1 Lindberg (2006), for example, proposed that simply holding elections helps countries democratize. One could argue that people will be more involved in politics if they are able to elect their head of state and that this will increase their trust and participation in the political system. A presidential election may be an event that triggers general interest in politics, as the debates between candidates include discussions of broad political issues.
On the other hand, however, multiplying the number of political contests may overwhelm and fatigue voters, creating incentives to neglect one's democratic duty and involvement in the political process altogether. Franklin (2003) has demonstrated that introducing direct elections to the European Parliament significantly suppresses participation in national elections. The cases of Switzerland and the United States are commonly cited as nations with multiple and frequent elections but exceptionally low participation among advanced democracies (Norris 2004).
Neither the studies on presidents in parliamentary systems, which are generally few in number, nor those on electoral turnout have empirically analyzed the relationship between turnout and the selection mechanism of the head of state in parliamentary systems. Aggregate-level studies of turnout tend to consider a broad cross-national sample of countries (Geys 2006), which does not allow studying the intricacies characteristic to certain regime types. Answering this question, however, not only has clear practical implications to constitution designers but also would contribute both to our understanding of the functioning of and differences between parliamentary systems and to uncovering reasons for why electoral turnout differs across countries. This article considers electoral turnout in a global sample of parliamentary democracies with a nonhereditary head of state from 1945 to 2006 and finds that direct presidential elections decrease turnout in parliamentary elections by about 7 percentage points. …