Does citizen participation affect municipal performance? Electoral competition and fiscal performance in Japan

ABSTRACT

Subnational fiscal performance has been receiving a lot of attention from researchers in various disciplines. However, there is very little published on the impact of citizen involvement on fiscal performance. This paper shows that a lack of citizen involvement in the electoral process was associated with fiscal performance in all 807 city-level Japanese municipalities from 2006 to 2012.

KEYWORDS:

• Electoral competition
• fiscal performance
• Japan
• municipality

Subnational fiscal performance has received a great deal of attention from various academic disciplines. In particular, scholars studying fiscal federalism (Oates, 1999) have examined the country-specific situation (Jones, Sanguinetti, & Tommasi, 2000) of fiscal behaviour by national governments. The fiscal performance of subnational governments can be improved by setting goals: for example revenue diversification and revenue balance (Suyderhoud, 1994); fiscal centralization and decentralization (Panizza, 1999); fiscal autonomy (Jacobs, 2003); debt ratio (De Haan & Sturm, 1997); and a balanced budget (Tujula & Wolswijk, 2004). Previous studies have used various approaches to measure fiscal performance based on a country’s or government’s specific goals. Thus, approaches to measuring the fiscal performance of government vary across contexts, reflecting different institutional conditions. In this paper, we use fiscal autonomy and fiscal capability as indicators of the fiscal performance of municipalities.

Various determinants of fiscal autonomy and fiscal capability have been identified. These include socioeconomic factors (Panizza, 1999; Volkerink & De Haan, 2001; Woo, 2003), institutional factors (Alesina & Perotti, 1999), and political factors (Ades & Glaeser, 1994; Carlsen, 1997; Feld, 2002; Kontopoulos & Perotti, 1999; Henisz, 2000; Remmer & Wibbels, 2000). However, there have been few empirical studies published about how voter involvement (public engagement) affects municipalities’ fiscal performance (Andrews, 2012; Geys, Heinemann, & Kalb, 2010). Theories of fiscal decentralization suggest that an increase in the level of fiscal autonomy will promote citizen engagement in public affairs. Low levels of fiscal autonomy might be a reason for low political involvement. However, there is no real evidence to date about whether or not active citizen involvement in politics leads to higher fiscal autonomy and fiscal capability in municipalities. Principal-agent theory concerning political accountability suggests that local government officials and politicians act as agents for citizens. Therefore, citizen participation in local politics must have a significant role in holding municipal officials and politicians accountable for their performance (Schaltegger & Torgler, 2007). Social capital literature also suggests that vibrant civic communities positively affect the performance and responsiveness of municipal governments (Andrews, 2012; Putnam, Leonardi, & Nanetti, 1994).

We examined how citizens’ electoral involvement is correlated with the fiscal performance of municipalities by using a panel data set of Japanese municipalities. Citizen participation in voting is low at the local level in Japan. This is partly due to the fact that there is very little choice in local elections. Especially in rural areas, there are not enough candidates running for mayoral and local council elections for people to actually vote. As a result, ‘walk- over’ rates have been increasing in local elections. For instance, the percentage of city mayors who held their position through an uncontested election was 21.0% in 2011, 22.1% in 2012, 23.1% in 2013, and 25.2% in 2014 (Japan Research Institute for Local Government, 2012–2015). This lack of competition can lead to less responsive and less functional government. Thus, we examined how the lack of citizen involvement in the electoral process is associated with fiscal performance by investigating 807 city-level Japanese municipalities from 2006–2012.

Fiscal autonomy and capability

As explained above, among the various measurements of fiscal performance, we focused on fiscal autonomy and fiscal capability. Fiscal autonomy is generally concerned with the financial independence of municipalities from the impact of upper-level government and other third parties. It can be defined as a percentage of an entity’s own financial resources in total revenue (Mikesell, 2010). Having more of one’s own revenue sources, such as local tax, fees and charges, and investment income, leads to less dependence on other entities. Increasing the financial autonomy of municipal governments is thought to lead to better services (Oates, 1999; Tiebout, 1956) and to reduce costs (Brennan & James, 1980). More recently, cross-national studies of fiscal autonomy have been published (for example Baskaran & Feld, 2013). In particular, research on the determinants of fiscal autonomy has increasingly been attracting the attention of scholars looking at such examples as the unification of Germany (Torgler & Werner, 2005), municipal mergers (Jacobs, 2003), and the progress of economic and monetary integration in Europe (Alesina & Spolaore, 1997).

While fiscal autonomy measurement is only concerned with the revenue side of municipal finance, fiscal capability covers both revenue and expenditure (both the demand and supply sides of public goods). The fiscal capability of subnational governments has been generally measured as the ratio of financial resources to local public needs (Brudney, 1984; Gordon, Auxier, & Iselin, 2016; Zhang & Li, 2016). It measures the extent to which municipal governments maintain the financial ability to meet the demands of local public goods and services from citizens. The concept of fiscal capability covers both the revenue and expenditure sides, so that it has been used similarly to the concept of budget balance to indicate the difference between revenue and spending (Mikesell, 2011; Tujula & Wolswijk, 2004).

Previous studies have examined the factors affecting the fiscal performance of municipalities, for example budgetary, macroeconomic, institutional, and political factors, as well as managerial quality. For instance, mandatory and discretionary grants from the central government, or local tax instruments or budgetary processes can improve fiscal capability (Alesina & Perotti, 1999). Budgetary variables, such as debt level and change in debt ratio and lagged budget balance, are important (Mikesell, 2011). Macroeconomic factors, such as unemployment rate, economic growth, interest rate (Volkerink & De Haan, 2001), and welfare level (Woo, 2003), are also considered to be determinants of fiscal capability. Political factors have been considered as one of the crucial dimensions affecting the fiscal performance of municipalities. In fact, scholarly interest in the topic of political and institutional effects on fiscal performance both at national and local level has been growing (Hagen & Vabo, 2005). For instance, political fragmentation—measured by the number of public agencies—political parties (Remmer & Wibbels, 2000), or political ideology (Carlsen, 1997) have all been considered to be significant factors for fiscal capability. The political budget cycle, which is ‘the possibility of a macroeconomic cycle induced by the political cycle’ (for example Brender & Drazen, 2005, p. 1272) has been much studied, with the conclusion that politicians increase spending and reduce taxation before upcoming elections in order to increase their likelihood of being re-elected. In addition, political ideology is also a factor that affects the composition and size of public spending (Hagen & Vabo, 2005). Another line of study that focuses on managerial capacity suggests that managerial quality, such as education and career background, matters in municipal fiscal performance (Avellaneda, 2009, 2012).

Political involvement and fiscal performance

While a wide range of political factors has been studied as determinants of fiscal performance, few empirical studies have considered whether or not political involvement improves fiscal performance, especially fiscal autonomy and capability (Geys et al., 2010). While theories of fiscal decentralization suggest that an increase in levels of fiscal autonomy promotes citizen engagement in public affairs, there have been few empirical studies examining the effects of citizen involvement on fiscal performance (Andrews, 2012; Geys et al., 2010). As explained above, the principal–agent theory of political accountability suggests that local government officials and politicians act as agents for citizens (for example Han & Demircioglu, 2016; Romzek & Dubnick, 1987), so participation in local politics has a significant role in holding municipal officials and politicians accountable to the public for their activities (Schaltegger & Torgler, 2007). More active citizen involvement in public affairs increases political awareness and interests, which leads to more pressure on local government officials and politicians. However, a low level, or absence, of political involvement reduces citizens’ motivation to actively monitor politicians. Among various forms of political participation, voting is the most common way for citizens to be involved in local politics. Active citizen involvement in the electoral process is believed to improve local government performance (De Janvry, Finan, & Sadoulet, 2008; Geys et al., 2010). In order to provide the services and goods necessary to meet citizens’ requirements (i.e. the amount of expected expenditure), local governments need to raise revenue. To collect as much revenue as possible, it is essential for local governments to improve their capacity to autonomously cover the required revenue rather than simply rely on upper-level governments (Mikesell, 2011).

The effect of political involvement on fiscal performance might also be contingent on other factors such as political and institutional factors (Tujula & Wolswijk, 2004). For instance, according to Kontopoulos and Perotti (1999), fiscal performance is influenced by mayors’ re- election times, political party support, and conservative party support. However, such political contexts are not related to the strength of citizen involvement. Therefore, political factors may moderate the impacts of citizen involvement on fiscal performance. Our study tested the following hypotheses:

H1a: Citizen involvement in the electoral process is associated with higher fiscal autonomy.

H1b: Citizen involvement in the electoral process is associated with higher fiscal capability.

H2a: The effect of citizen involvement on fiscal autonomy is moderated by the political context.

H2b: The effect of citizen involvement on fiscal capability is moderated by the political context.

Case study: Japanese local governments

Our study focused on all city-level municipalities in Japan. We selected Japan for the following reasons:

• Japan had adopted a unitary system, not a federal system, therefore all Japanese municipalities have similar municipal structures, have experienced similar institutional change, and have dealt with similar historical and macroeconomic factors. This homogeneity allowed us to analyse the impact of citizen involvement on performance more precisely.

• The number of walk-over elections, which takes a crucial opportunity for political involvement away from citizens, has been increasing in Japan. Therefore, Japan provides a good laboratory to test how the lack of involvement in electoral process affects municipal performance after an election.

• Data on the variables which interest us is available for all city-level municipalities, which allowed us to conduct a nationwide analysis.

• Japan’s local governments have been understudied.

Japanese subnational governments consist of two units, prefecture (regional units) and municipality (local units). Municipalities are categorized into cities, towns, and villages. Japanese local government consists of the legislative branch and the executive branch. Members of local assemblies, governors, and mayors are directly elected by the public. Local assemblies have voting rights in such matters as budget and ordinances, and can conduct a no- confidence vote to remove mayors (CLAIR, 2013). The main responsibilities of municipalities include social welfare; the establishment and management of nursing homes for the elderly; elementary and middle schools; nursing insurance; national health insurance; urban design; construction and management of municipal roads, bridges, water, and sewerage; collection and disposal of general waste; fire-fighting operations; medical emergency support; and resident registration (MIC, n.d.). Mayors and local assembly members are directly elected by residents for four-year terms. After the Second World War, local elections were scheduled to take place in March, with inaugurations taking place one month later. However, due to mayors’ deaths, impeachments, terminations, and other issues, elections in 2015 were held on a different schedules in almost 70% of localities (MIC, 2017a). Mayors’ duties include enacting regulations, preparing budgets, proposing bills, and appointing or dismissing staff (CLAIR, 2013).

Local governments have two primary sources of income: their own revenue collection, and transfers of funds from the prefecture and the central government. They rely heavily on transfers from the central government. Local revenue collection is from local taxes, fees, rents, donations, and the leasing and sale of property. While cities vary considerably in terms of local revenue collection, in 2012 the total revenue collected locally amounted to 48.95% of the total for all city-level municipalities. The main source of local revenue is tax, which accounted for 34.87% of total revenue. The central and prefectural governments provide the remaining 51.05% of local revenue (MIC, 2018).

Data collection and variable operationalization

Our unit of analysis was the municipality-year. We targeted all city-level municipalities from 2006 to 2012, covering 807 cities. As the mayoral term in Japan is four years, our panel contained data from two mayoral administrations. Towns and villages were excluded because of lack of data. The two main data sets we used were the Zenkoku Shucho Meibo (List of Mayors) collected by the Japan Research Institute for Local Government (2007–2013) and the Regional Statistics Database (MIC, 2018). The first data set contained mayors’ political data (party affiliation, re-election times, vote share, political party support). Other data was obtained from the second data set.

Measuring fiscal performance

Japan records the highest amount of government debt among OECD member countries. The central government’s austerity policy constrains the public finances of local municipalities (Suzuki, 2017). Due to increasing debt, a declining population, and economic conditions, increasing their financial autonomy through revenue expansion has become a primary goal for many local governments in Japan (Japan Research Institute for Local Government, 2010). Major tools for revenue expansion include local tax; fees, charges and allotments from other organizations; property revenue; donations; and accounts brought forward (Japan Research Institute for Local Government, 2011).

An amendment to the Local Autonomy Act in 2006 granted power to localities to obtain revenue by utilizing their own assets such as landed property and movable property (leasing/ sales). Therefore, although most Japanese local governments largely depend on national transfers, they are also seeking to expand their revenue through locally-controlled channels (Japan Research Institute for Local Government, 2011). These efforts can be measured by the following two indicators: the fiscal autonomy index and the financial capability index. Our first dependent variable, fiscal autonomy, is simply a ratio of the independent revenue to the total amount of revenue for a municipality. The main sources of external revenue include local allocation tax from the central government, national government disbursement, and local transfer tax.

We used the ‘financial capability index’ (zaisei ryoku shisu in Japanese) as an indicator for the fiscal capacity of municipalities. It is an index used by the central government to decide whether the central government should distribute local allocation tax to particular municipalities or not. Municipalities whose financial capability index scores more than 1 do not receive any local allocation tax, since the central government considers them to have sufficient financial resources. Those who score below 1 receive local allocation tax. The financial capability index is a ratio of a city’s own standard revenue to their standard fiscal needs. This index shows how much revenue each city collects by itself. Their own standard revenue shows the amount of their own tax revenues with adjustment. Standard fiscal needs are the financial resources necessary for municipalities to provide standard public goods and services. Scores are calculated considering unit cost, measurement unit, and the characteristics of municipalities. A score of 1 means that a municipality can fully cover its expenditures with its own financial resources based on its taxes (MIC, 2017b, 2018). Unlike the fiscal autonomy measurement, the financial capability index only considers tax as a revenue resource. In addition, unlike the fiscal autonomy indicator, the fiscal capability index considers the demand (expenditure) side for public goods, as well as the revenue side.

Independent variable

Our main independent variable was whether or not mayoral elections were walk-overs (i.e. no contest). We created a binary variable, which gives a value of ‘0’ to municipalities with mayors winning their seat through a contested election and ‘1’ to those with mayors whose election was uncontested. Other moderating political variables included re-election times, conservative party support, and the number of political parties that supported the mayor during the elections. Re-election was a continuous variable, representing how many times a mayor was re-elected consecutively. If a mayor was elected three times consecutively, the indicator was 3. We used a centred variable to reduce any skew. Conservative party support was a dummy variable. Category 1 shows that a mayor received support from the conservative parties—the Liberal Democratic Party (LDP) and the Komei Party—during the previous election campaign. Category 0 indicates no support from conservative parties. (The LDP is one of the largest political parties in Japan. The party has been in power almost continuously since 1955. The Komei party, which relies on a Buddhist organization, Soka Gakkai, entered into coalition with the LDP in 1999. See, for example, ; Klein et al., 2018; Krauss & Pekkanen, 2010 for an explanation of Japanese political parties.) The number of political parties supporting a candidate refers to the previous election campaign. We also use a centred variable for this variable.

Control variables

We controlled for other factors that we expected would affect our dependent variables. These were income level, lag of municipal expenditure, population, area, and municipal mergers. Many Japanese municipalities merged after 1999, which we expected to have had a large effect on the subsequent fiscal performance of the merged units (Suzuki & Sakuwa, 2016; Yamada, 2016). Municipal merger is a categorical variable. If a merger did not take place, we scored ‘0’. Municipalities that experienced municipal merger through municipal annexation scored 1. A value of 2 was given to municipalities that were created as a new municipality as a result of municipal merger. Table 1 provides the descriptive statistics for all the variables.

Table 1. Descriptive statistics (807 city-level municipalities).

	Observations	Mean	SD	Minimum	Maximum
Dependent variables
Fiscal autonomy	5691	46.91	14.96	5.11	89.30
Fiscal capability	5357	66.64	26.73	11	192
Independent variables
Walk-over election
(1 = no contest, 0 =  contested)	5630	0.21	0.41	0	1
Political context
Re-election times	5630	1.75	1.01	1	10
Conservative party support
(1 = yes, 0 = no)	5638	0.38	0.48	0	1
Political party support	5638	0.99	1.30	0	6
Control variables
Income per capita	5691	3023.52	607.37	2050.5	11266.5
Lag of municipal expenditure	5691	53,900,000	110,000,000	4,553,889	1,670,000,000
Population	5691	141,652.20	246,467.70	4,110	3,633,130
Area	5691	26634.06	28413.20	510	217,767
Municipal merger
(0 = No merger, 1 =  Annexation, 2 =  New city)	5638	0.89	0.92	0	2

Analysis and results

Our dataset was comprised of panel data of 807 city-level municipalities from 2006 to 2012. Some variables, such as electoral contests, re- election times, party support, and municipal merger, do not vary across time within a municipality. Having time-invariable measures meant we could not report fixed-effect estimations. Consequently, for each of our two dependent variables, we report five different estimations for robustness checks: random- effects with White-Huber standard errors to correct for heteroskedasticity; between effect; population average; linear regression with panel-corrected standard errors; and Arellano- Bond dynamic panel model (Arellano & Bond, 1991). The Arellano-Bond model includes a lag of the dependent variable as an independent variable.

Explaining fiscal autonomy

Table 2 presents the results of the estimations for fiscal autonomy by using both control and political variables without applying moderating effects. Our main variable of interest—walk- over election—shows a positive sign in the between-effects model (p < 0.05) and in the panel-corrected standard error model (p < 0.01). This result is the opposite of what we had expected. Lack of voter involvement as a result of uncontested election is positively associated with higher fiscal autonomy. However, the results from the other models were not statistically significant for the relationship between voter involvement and fiscal autonomy. Table 2 also shows that most variables were statistically significant across all five of the estimation models (p < 0.01). Income and population had positive impacts on fiscal autonomy. Municipal mergers in the form of new municipality creation present negative effects. Area of municipality positively affected fiscal autonomy in all models except the Arellano-Bond model. Lag of municipal expenditure was negatively associated with fiscal autonomy across all models except in the Arellano-Bond model. Conservative party support was positive across two models (random effects and population-averaged models). Political party support was negative and statistically significant in the random effects and the population-averaged models (p < 0.05) and in the panel-corrected standard errors model (p < 0.01).

Table 2. Explaining fiscal autonomy (807 municipalities).

	Random effects model	Between effects model	Population- averaged model	Panel- corrected SE	Arellano-Bond dynamic panel model
Walk-over election
(1 =  no contest, 0 =  contested)	0.01	0.06**	0.00	0.03***	−0.01
	(0.00)	(0.02)	(0.00)	(0.01)	(0.01)
Conservative party support
(1 = yes, 0 = no)	0.02**	−0.02	0.02***	−0.00	0.00
	(0.01)	(0.03)	(0.01)	(0.01)	(0.01)
Re-election (centred)	−0.00	0.01	−0.00*	0.00*	0.00
	(0.00)	(0.01)	(0.00)	(0.00)	(0.00)
Political party support (centred)	−0.01**	−0.02	−0.01**	−0.01***	−0.00
	(0.00)	(0.01)	(0.00)	(0.00)	(0.00)
Income per capita (ln)	1.10***	0.97***	1.10***	0.96***	1.75***
	(0.07)	(0.06)	(0.03)	(0.03)	(0.10)
Lag of municipal expenditure (ln)	−0.40***	−0.65***	−0.39***	−0.61***	0.14*
	(0.03)	(0.03)	(0.01)	(0.03)	(0.08)
Population (ln)	0.42***	0.66***	0.42***	0.63***	2.05***
	(0.04)	(0.03)	(0.01)	(0.03)	(0.43)
Area (ln)	0.02***	0.04***	0.02***	0.04***	−0.16
	(0.01)	(0.01)	(0.01)	(0.00)	(0.25)
Municipal merger (baseline = no merger):
Annexation	0.02*	0.01	0.02	0.01**	−0.09***
	(0.01)	(0.02)	(0.02)	(0.00)	(0.02)
Municipal merger: New municipality	−0.08***	−0.10***	−0.08***	−0.10***	−0.09***
	(0.01)	(0.02)	(0.01)	(0.01)	(0.03)
Lag of DV					0.03
					(0.05)
Constant	−3.10***	−0.58	−3.17***	−0.65**	−34.43***
	(0.66)	(0.53)	(0.31)	(0.30)	(6.35)
Observations	5,604	5,604	5,604	5,604	3,998
R2		0.72		0.68

Note: Robust standard errors in parentheses.

*** p <  0.01, ** p <  0.05, * p <  0.1.

Table 3 reports moderating effects of walk-over elections and political variables by using interaction terms for these variables. We created three interaction terms: walk-over election and re-election times, conservative party support, and political party support. The underlining idea is that the impact of a walk-over election on fiscal autonomy might differ, contingent on other political factors. The results, which take into account moderating variables, were to some degree different from those not including the moderating variables, depending on the selection of statistical models. Table 3 shows that a walk-over election had a positive impact on fiscal autonomy in the random effects and the panel-corrected standard errors models (p < 0.05) and in the population- averaged model (p < 0.01). This means that the lack of citizen involvement in mayoral elections had a positive impacts on municipal performance in terms of fiscal autonomy. Political support was negative and statistically significant (p < 0.1, p < 0.5, p < 0.01) in four models. Conservative party support was positive and statistically significant (p < 0.1, p < 0.5, p < 0.01) in two models.

Table 3. Explaining fiscal autonomy—interaction terms (807 municipalities).

	Random effects model	Between effects model	Population- averaged model	Panel- corrected SE	Arellano-Bond dynamic panel model
Walk-over election	0.02**	0.04	0.02***	0.03**	0.02*
(1 =  no contest, 0 =  contested)	(0.01)	(0.04)	(0.01)	(0.01)	(0.01)
Conservative party support	0.02**	−0.03	0.02***	−0.01	0.01
1 = yes, 0 = no)	(0.01)	(0.03)	(0.01)	(0.01)	(0.01)
Re-election (centred)	−0.00	0.01	−0.00*	0.00	0.00
	(0.00)	(0.01)	(0.00)	(0.00)	(0.00)
Political party support (centred)	−0.01**	−0.02*	−0.01***	−0.02***	−0.00
	(0.00)	(0.01)	(0.00)	(0.00)	(0.00)
Walk-over election x re-election (centred)	0.00	0.02	0.00	0.01*	−0.00
	(0.01)	(0.02)	(0.00)	(0.01)	(0.01)
Walk-over election x conservative party support	−0.04**	0.03	−0.04**	0.01	−0.07**
	(0.02)	(0.09)	(0.02)	(0.03)	(0.03)
Walk-over election x political party support	0.01**	0.02	0.01**	0.01	0.02**
	(0.01)	(0.03)	(0.01)	(0.01)	(0.01)
Income per capita (ln)	1.10***	0.98***	1.10***	0.96***	1.74***
	(0.07)	(0.06)	(0.03)	(0.03)	(0.10)
Lag of municipal expenditure (ln)	−0.40***	−0.65***	−0.39***	−0.61***	0.14*
	(0.04)	(0.03)	(0.01)	(0.03)	(0.08)
Population (ln)	0.42***	0.66***	0.42***	0.62***	2.06***
	(0.04)	(0.03)	(0.02)	(0.03)	(0.43)
Area (ln)	0.02***	0.04***	0.02***	0.04***	−0.16
	(0.01)	(0.01)	(0.01)	(0.00)	(0.25)
Municipal merger: Annexation	0.02*	0.00	0.02	0.01**	−0.09***
	(0.01)	(0.02)	(0.02)	(0.01)	(0.02)
Municipal merger: New municipality	−0.08***	−0.09***	−0.08***	−0.10***	−0.09***
	(0.01)	(0.02)	(0.01)	(0.01)	(0.03)
Lag of DV					0.03
					(0.05)
Constant	−3.09***	−0.66	−3.17***	−0.70**	−34.46***
	(0.66)	(0.53)	(0.31)	(0.29)	(6.34)
Observations	5,604	5,604	5,604	5,604	3,998
R2		0.72		0.68

Note: As table 2.

The interaction term between a walk-over election and conservative party support was negative (p < 0.05). This means that the combination of a walk-over election and conservative party support negatively affects fiscal autonomy. In other words, when support from the conservative party is high, the enhancement of political participation increases fiscal autonomy. We conducted joint F-tests for these interaction terms for a robustness check. The results of a joint F-test for an interaction term between a walk-over election and conservative party support were statistically significant in the population- averaged and the Arellano-Bond models (p < 0.05) and in the random effects model (p < 0.1). The interaction term between no electoral contest and political party support was positive (p < 0.05), suggesting that, if there was only one mayoral candidate with support from more than political party, the municipality was likely to achieve higher fiscal autonomy than those municipalities with an electoral contest and fewer political parties supporting the mayor. The joint F- test results for the interaction between no contest and party support were significant in the random effects and the population- averaged models (p < 0.05) and in the Arellano-Bond model (p <  0.1). These results show the robustness of the estimations.

In sum, municipalities that did not go through a contested election tended to have higher fiscal autonomy than those with competitive elections. Political involvement in the electoral process seems to be negatively associated with fiscal autonomy, yet the relationship is positive under specific Japanese contexts such as the support of conservative party. When a mayor won a competitive election with support from a conservative party, the municipality was likely to achieve higher fiscal autonomy. On the other hand, when there was no electoral contest and many political parties supported a single mayoral candidate, the municipality was likely to achieve higher fiscal autonomy. Therefore, in the Japanese case, the relationship between political participation in mayoral elections and fiscal autonomy is mixed, and is considerably affected by the specific political situation.

Explaining fiscal capability

Table 4 shows the results of the five regression models for fiscal capability, without considering the moderating impacts of specific political situations. Walk-over election was correlated to higher fiscal capability but only in the panel- corrected standard error model (p < 0.01). Conservative party support was positive and statistically significant across four models. Re- election times were statistically significant across all statistical models. However, the results were not consistent. Political party support was negative and statistically significant (p < 0.1, p < 0.5, p < 0.01) in three models. Most of the results on control variables were consistent and statistically significant across four or five models, thus indicating robustness in the findings. Factors such as income per capita, population, and municipal annexation were connected with higher fiscal capability. On the other hand, lag of municipal expenditure and municipal area were negatively associated with fiscal capability.

Table 4. Explaining fiscal capability.

	Random effects model	Between effects model	Population-averaged model	Panel-corrected SE	Arellano-Bond dynamic panel model
Walk-over election (1 = no contest, 0 = contested)	0.00	0.02	0.00	0.02***	−0.00
	(0.00)	(0.02)	(0.00)	(0.00)	(0.00)
Conservative party support (1 = yes, 0 =  no)	0.01***	0.02	0.01***	0.02***	0.01***
	(0.00)	(0.03)	(0.00)	(0.01)	(0.00)
Re-election (centred)	−0.00***	0.02**	−0.00***	0.01***	−0.00***
	(0.00)	(0.01)	(0.00)	(0.00)	(0.00)
Political party support (centred)	−0.00	−0.03**	−0.00*	−0.02***	0.00
	(0.00)	(0.01)	(0.00)	(0.00)	(0.00)
Income per capita (ln)	0.83***	1.50***	0.82***	1.39***	0.87***
	(0.04)	(0.08)	(0.02)	(0.06)	(0.02)
Lag of municipal expenditure (ln)	−0.09***	−0.59***	−0.09***	−0.49***	−0.03***
	(0.02)	(0.04)	(0.01)	(0.04)	(0.01)
Population (ln)	0.29***	0.66***	0.30***	0.58***	0.31***
	(0.02)	(0.04)	(0.01)	(0.04)	(0.04)
Area (ln)	−0.11***	−0.01	−0.11***	−0.03***	−0.15***
	(0.01)	(0.01)	(0.01)	(0.01)	(0.03)
Municipal merger: Annexation	0.08***	0.07***	0.07**	0.07***	0.00
	(0.03)	(0.02)	(0.03)	(0.00)	(0.00)
Municipal merger: New municipality	−0.02	0.02	−0.02	0.01*	0.00
	(0.02)	(0.02)	(0.02)	(0.00)	(0.00)
Lag of DV					0.72***
					(0.01)
Constant	−3.14***	−5.02***	−3.21***	−4.77***	−7.26***
	(0.43)	(0.73)	(0.20)	(0.46)	(0.52)
Observations	5,320	5,320	5,320	5,320	3,761
R2		0.79		0.77
Number of municipalities	784	784	784	784	776

Note: As table 2.

Table 5 presents the estimators for fiscal capability using the interaction terms for specific political contexts. A walk-over election alone did not achieve statistically significant results for the fiscal capability variable, unlike the fiscal autonomy variable. Conservative party support alone was positive and statistically significant (p < 0.5, p < 0.01) across four models. Re-election times again led to inconsistent results. Political party support was negatively associated with fiscal capability in two models (p < 0.5, p < 0.01). The interaction term between walk-over election and conservative party support was positive in the random effects model (p < 0.1) and in the population-averaged model (p < 0.05). However, considering political party support, the interaction term between walk-over election and number of supporting parties was negative in the population-averaged model (p < 0.1). In other words, when the number of supporting parties was high, the enhancement of political participation improves fiscal capability. Despite a few significant results, the joint F-tests for these interaction terms shows that only the estimator for the population-averaged model was statistically significant, showing a lack of robustness. In sum, walk-over elections do not seem to affect the fiscal capability of municipalities.

Table 5. Explaining fiscal capability with interaction terms.

	Random effects model	Between effects model	Population- averaged model	Panel- corrected SE	Arellano-Bond dynamic panel model
Walk-over election (1 =  no contest, 0 =  contested)	−0.00	0.02	−0.00	0.01	0.00
	(0.00)	(0.05)	(0.00)	(0.01)	(0.00)
Conservative party support (1 = yes, 0 =  no)	0.01***	0.02	0.01***	0.01**	0.01***
	(0.00)	(0.03)	(0.00)	(0.01)	(0.00)
Re-election (centred)	−0.00***	0.02	−0.00***	0.01***	−0.00**
	(0.00)	(0.01)	(0.00)	(0.00)	(0.00)
Political party support (centred)	−0.00	−0.03**	−0.00	−0.02***	−0.00
	(0.00)	(0.01)	(0.00)	(0.00)	(0.00)
Walk-over election x re-election (centred)	0.00	0.02	0.00	0.02***	−0.00
	(0.00)	(0.03)	(0.00)	(0.01)	(0.00)
Walk-over election x conservative party support	0.02*	0.01	0.02**	0.02	−0.01
	(0.01)	(0.09)	(0.01)	(0.03)	(0.01)
Walk-over election x political party support	−0.01	0.01	−0.01**	−0.00	0.00
	(0.00)	(0.03)	(0.00)	(0.01)	(0.00)
Income per capita (ln)	0.83***	1.50***	0.82***	1.39***	0.87***
	(0.04)	(0.08)	(0.02)	(0.06)	(0.02)
Lag of municipal expenditure (ln) Population (ln)	−0.09***	−0.59***	−0.09***	−0.49***	−0.03***
	(0.02)	(0.04)	(0.01)	(0.04)	(0.01)
	0.29***	0.66***	0.30***	0.58***	0.31***
	(0.02)	(0.04)	(0.01)	(0.04)	(0.04)
Area (ln)	−0.11***	−0.01	−0.11***	−0.03***	−0.15***
	(0.01)	(0.01)	(0.01)	(0.01)	(0.03)
Municipal merger: Annexation	0.08***	0.07***	0.07**	0.07***	0.00
	(0.03)	(0.02)	(0.03)	(0.00)	(0.00)
Municipal merger: New municipality	−0.02	0.02	−0.02	0.01**	0.00
	(0.02)	(0.02)	(0.02)	(0.00)	(0.00)
Lag of DV					0.72***
					(0.01)
Constant	−3.15***	−5.04***	−3.22***	−4.79***	−7.25***
	(0.43)	(0.74)	(0.20)	(0.46)	(0.52)
Observations	5,320	5,320	5,320	5,320	3,761
R^2		0.79		0.77
Number of municipalities	784	784	784	784	776

Note: As table 2.

Discussions and conclusions

Our study explored whether uncontested mayoral elections affect fiscal performance measured by fiscal autonomy and capability by using a panel data set of 807 city-level Japanese municipalities from 2006 to 2012. Our findings were that a walk-over election is positively correlated to fiscal autonomy, contrary to our hypothesis 1a. We hypothesized that citizen participation through the electoral process increases political awareness and interests, which leads to more pressure on municipal officials and politicians. As a consequence, political participation leads to higher fiscal autonomy. However, we found that the opposite was true for Japanese municipalities. One of the potential explanations is that pressure to satisfy citizen demands for public goods may be lower for walk-over mayors than for those who competed to be elected. Mayors who face limited restraint from citizens because of an uncontested election may be able to pursue unpopular financial options in order to ensure fiscal autonomy. Therefore, these mayors can employ unpopular financial reconstruction measures such as raising fees and charges. A combination of a walk-over election and political party support was also found to be positively related to fiscal autonomy. This also makes sense because mayors are more likely to gain support from local councils, which may allow walk-over mayors to succeed in conducting fiscal reconstruction programmes. A combination of a walk-over election and conservative party support negatively affects fiscal autonomy, which means that political participation enhances fiscal autonomy of local Japanese governments when the conservative party support is a strong supporter. This result needs to be investigated in another study. Nevertheless, the result implies that the relationship between citizen participation and fiscal autonomy can produce mixed results, depending on specific political contexts.

We also found that a walk-over election did not affect the fiscal capability of a municipality. However, conservative party support positively affected financial capability and political party support alone negatively affected it. Future research should investigate the causal mechanism of the positive effects of walk-over elections on fiscal performance. Moreover, survey and case studies of local governments and mayoral quality should be conducted to identify such causal mechanisms.

Impact

Although practitioners and researchers have strong interests in fiscal performance and citizens’ political participation, we have a limited understanding of how the latter affects the former. The evidence in this paper shows that citizen participation does not always lead to better fiscal outcomes at the municipal level. Practitioners should consider various factors that may potentially affect the fiscal performance of municipalities. In particular, practitioners should take into account the political contexts that may affect the participation–fiscal performance relationship.

ORCID

Kohei Suzuki http://orcid.org/0000-0002-5403-4826