Does Divided Government Control Unilateral Policymaking?

• In Event: The Presidency and Unilateral Action

Abstract

Unilateral action is one of the important policymaking tools in the presidency. Previous studies on the American presidency empirically and theoretically examine the impact of the relationship between the executive and the legislature on unilateral policymaking, but they have a significant omitted variable problem. With the close election regression discontinuity design and the innovative dataset on the chair elections in Japanese municipalities, we deal with this problem and examine the causal relationship between the legislative context and unilateral policymaking.

Japan municipalities are ideal for examining the causation between the two. There are about 800 city governments, and the mayors can issue “senketsu-shobun,” the Japanese version of the Executive Order. In addition, mayoral and city council elections are often held at different times, and chair elections are frequently held. Therefore, we can exploit close election regression continuity design by assuming chair elections as “randomization” of divided government. For example, suppose that a mayor is related to the Liberal Democratic Party (LDP) and that the city council member related to the Democratic Party of Japan and the one related to the LDP are competing for the chair post. When the former wins the election, the city government is divided. On the other hand, when the latter wins, the government is unified.

To examine the causation, we developed an innovative dataset on the chair elections in Japanese municipalities. We collected about 12000 candidates’ data on when the chair elections were held, who got how many votes, and each candidate’s relationship to the national parties. Then, we exploit the close election regression discontinuity design.

We contribute to the literature in two ways. First, we expand the literature to East Asia. Unilateral policymaking is essential in the presidency in the world, but the previous studies examine the cases mainly in the United States. We examine the theory in Japan, which has been clearly ignored. Second, we examine the vital theory with a causal inference approach. The previous studies mainly examine the theory with a dataset on the Executive Orders in the United States federal government. As a result, the studies have a significant omitted variable problem. The studies cannot control the variables because there is only one president in one period in the United States. We re-examine the theory by dealing with the problem with the close election regression discontinuity design.

Author

• Takaharu Saito, Nagoya University of Commerce and Business