![Sample our Politics & International Relations journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5947/TOC-Subject-Sample-2021-Politics-International-Relations.jpg"})
ABSTRACT
This paper examines whether boundary changes lead to changes in the relationship between voters and politicians. We focus on the wave of municipal mergers in Japan that took place in the 2000s in order to examine this question. Municipalities with small population size, once they merge with their larger neighbours, would have a small number of voters relative to the size of the electorate in the post-merger municipalities. Therefore, municipal politicians in the post-merger municipalities do not have strong electoral incentives to receive support from geographic areas corresponding with pre-merger municipalities of small population size. Using a survey of voters in 89 locations corresponding with pre-merger municipalities, we demonstrate that voters who live in small municipalities that experienced mergers interact with politicians less frequently and have less favourable impressions of politicians than before the mergers. These patterns are not observed in municipalities that did not experience mergers.
Acknowledgement
Earlier versions of this paper and related works were presented at the World Congress for Korean Politics and Society (August 2015), the Midwest Political Science Association annual conferences (April 2016 and April 2017), a research workshop at Hokkaido University (September 2016), and the Japan Public Choice Society annual conference (November 2017). We sincerely appreciate comments from the participants of these workshops and conferences. We are also grateful for comments and suggestions from two anonymous reviewers, Masahisa Endo, Alan Gerber, Masataka Harada, Yusaku Horiuchi, Go Murakami, Steven R. Reed, Frances Rosenbluth, Jun Saito, Hironori Sasada, Seiki Tanaka, and Masahiro Yamada.
Disclosure statement
No potential conflict of interest was reported by the authors.
Supplementary material
Supplemental data for this article can be accessed here.
Notes
1. Throughout the paper, when we say ‘size’ of local government, we mean the population size.
2. See Tavares (Citation2018) and Van Houwelingen (Citation2017) for reviews of empirical works examining the impact of merger. There is also an emerging literature that focuses on proliferations of subnational units (splitting of a political unit and establishment of new ones), which have been observed mainly in developing countries (Grossman, Pierskalla, and Boswell Citation2017; Lewis Citation2017; Pierskalla Citation2016).
3. Jakobsen and Kjaer (Citation2016) show that in Denmark shortly after the 2007 municipal mergers, peripheral areas within merged municipalities had a larger number of representatives per person than in central, more populous areas; the other studies that focus on Czech Republic, Finland, and Japan show that peripheral areas lost representatives and/or experienced a decline in public services.
4. The LAT grant is a transfer from the central to local governments (municipalities and prefectures). The amount allocated to each local government is determined by a formula. The special bond could cover up to 95% of the total project cost. Seventy per cent of the amount to be repaid would be covered by the LAT grant in later years. Thus, a merged municipality could launch a new project, such as the construction of a new city hall, with substantial financial assistance from the central government.
5. This point is consistent with Saarimaa and Tukiainen (Citation2014), who focus on the concurrence of municipal mergers in Finland in 2007 and find that municipalities with small population sizes were induced to merge despite the likely loss of their post-merger political powers because they had been fiscally struggling and wealthier neighbours were willing to merge with them ‘only if they [wealthier neighbors] maintain political control of the resulting coalition (p.112).’ In the Finnish context, Hyytinen, Saarimaa, and Tukiainen (Citation2014) also argue that electoral considerations of incumbent assembly members likely affected their voting decisions over proposed mergers.
6. The exception is cities designated by ordinance (seirei shitei toshi) – such as Osaka, Sapporo, and Yokohama – where electoral districts correspond with wards within the city. For Tokyo’s 23 Special Wards, each ward has its own assembly with the electoral district for the assembly corresponding with the entire ward.
7. Some candidates received far more votes than required to guarantee a seat, which allowed a candidate who received as few as 2491 votes to be elected. The election data are obtained from Takasaki City’s website.
8. Note that our survey was conducted only in 2012. In order to facilitate interpretation, we use expressions associated with changes (‘decreased’, ‘increased’, ‘changed’, etc.) to explain the difference between the respondents’ reported experience today and 10 years ago (e.g., frequency of interactions with politicians decreased from 10 years ago to today). We acknowledge that in measuring respondents’ attitudes and experiences 10 years ago, we rely on what respondents recall about 10 years ago; there might be some problems, such as recall bias, which we briefly discuss in Section 5.6.
9. The population size of 5495 was the first quartile in the municipalities in Japan as of 1999. We picked this cut-off point in order to focus on municipalities with small population size. DID ratio is obtained from the census data and sometimes used in the studies of election and politics in Japan to measure the urban-ness of a geographic unit (e.g., Horiuchi and Saito Citation2003; Reed, Scheiner, and Thies Citation2012).
10. Variables included are population size, population growth rate from 1995 to 2000, area, percentage of the elderly population (65 years old or older), percentage of the young population (15 years old or younger), home ownership rate, percentage of the working-age population employed in the agriculture, forestry, or fishery sector, percentage of the working-age population employed in the construction sector, per capita income, per capita transfer from the central government, fiscal strength index, ordinary balance ratio, and the ruling coalition’s vote shares in the 1990 and 1993 Lower House elections.
11. King and Nielsen (Citation2019) cautions the use of propensity score for causal inference but claims that their results ‘do not necessarily implicate the many other productive uses of propensity scores, such as regression adjustment … [and] stratification … (p.436)’.
12. Due to the limitation in space in the questionnaire, we could not ask questions about mayors.
13. Assembly members have 4-year terms. We asked about the situation in the last 4 years to make sure to include the period during which an election was held. If we ask about the situation in the last 3 or 6 months, for example, some respondents refer to the period when the election was held (thus assembly members were more active), while others do not.
14. We asked the questions about today and 10 years ago back to back, and there might be a concern about the order effect, which refers to the tendency of one’s response to the first question affecting their responses to subsequent questions. However, we believe it is less of a concern here. In fact, our goal is to compare responses between the two questions to construct our dependent variables. In other words, we want the respondents to recall the situation 10 years ago in comparison with today.
15. The variable takes the value between 0 and 1. We used the population data as of 1999. Suppose Municipality A (10,000 residents) merged with B (10,000 residents) and C (30,000 residents). Then, the variable takes the value of 0.20 (10,000 divided by 50,000).
16. We report the analyses for these two sets of dependent variables because the size of the effect is much larger than that of the dependent variables related to municipal politicians’ activities.
Additional information
Funding
Notes on contributors
Kyohei Yamada
Kyohei Yamada is an associate professor in the Graduate School of International Relations at International University of Japan. His research focuses on local government and politics, intergovernmental relations, and Japanese Politics.
Kiichiro Arai
Kiichiro Arai is an associate professor of political science in the Faculty of Policy Studies at Chuo University. His research interests bridge the area of political psychology and political behaviour with a particular focus on political participation, public opinion, and voting behaviour