Ehime University Institutional Repository

IKEDA, Fumi

－ 27 －
Abstract.
Why are the Democrats not likely to promote the anti-nuclear power policy to attract voters? 
This paper argues that the electoral clout of labor unions are one of the reasons why the Demo-
crats are not able to push the anti-nuclear power policy through examining voter mobilization of 
the Electric Power Related Industry Worker’s Union. The economic geography of power plants 
influences voter mobilization of the labor union. In particular, more voters are mobilized in the 
locations of nuclear power plants than in other types of energy power plants. 
1. Introduction
On March 11, 2011, the Great East Japan Earthquake struck Japan and the nuclear accident 
that unfolded thereafter would become the world’s second most serious after the 1986 Chor-
nobyl accident (Nakamura and Kikuchi 2011; Kushida 2013). The accident occurred during 
the administration of Naoto Kan of the Democratic Party of Japan (DPJ) and criticism of his 
handling of the disaster played an important role in the defeat of the DPJ in the following elec-
tion in 2012 (Carlson and Reed, 2018). Since that election, the LDP-Komeito coalition govern-
ment led by Shinzo Abe has pursued a policy of restarting the nuclear power plants after each 
meets strict safety standards. 
 The Japanese response to the Fukushima nuclear disaster stands in stark contrast to the 
German response. German voters reacted to the Fukushima nuclear accident as strongly as did 
Japanese voters. However, the Christian Democrats quickly abandoned its long-standing pro-
nuclear stance, shutting down seven of Germany’s seventeen nuclear power plants and announc-
ing an ultimate phase-out by 2022 (Meyer and Schoen 2017: 425). The Fukushima nuclear 
The Influences of the Economic Geography on Elections: 
The Case of the Electric Power Industry in Japan
Fumi Ikeda　
Fumi Ikeda
－ 28 －
accident had a greater impact on German politics and policy than it had on Japan. The Fuku-
shima nuclear disaster forced the German Christian Democrats to abandon their pro-nuclear 
stance, but the Japanese LDP was not forced to do the same. 
 The difference between the German and Japanese responses cannot be attributed to differ-
ences in public opinion. Opinion polls in Japan also show widespread opposition to nuclear 
power, but the coalition has not lost votes by pursuing this unpopular policy. Aldrich (2008: 
135-137) documents the declining public support for nuclear power plants in the 1980s and 
1990s. Furthermore, this trend accelerated after the Great East Japan Earthquake as table 1 
shows.
Table 1 Public Opinion on Nuclear Energy after Fukushima (%)
Question 1 Question 2
Year Month Favour Oppose Favour Oppose
2011 June 37 42 74 14
　 July 34 46 77 12
　 August 　 　 68 20
　 September 34 45 77 13
　 October 34 48 　 　
　 December 30 57 77 16
2012 February 　 　 66 23
　 March 　 　 70 17
　 July 　 　 67 21
　 August 33 50 　 　
　 November 34 50 　 　
2013 January 　 　 75 16
2015 August 　 　 73 22
Q1: Do you favour or oppose the use of nuclear energy? 
Q2: Do you favour or oppose phasing nuclear reactors out until none are left?
Source: Asahi Shimbun, various years
A majority opposed the use of nuclear energy at all and over two-thirds consistently favoured 
phasing nuclear reactors out until none remain. Yet nuclear power has yet to become an issue 
that impacts Japanese election outcomes.  
 However, the Democrats have not exploited the government’s vulnerability on the issue of 
nuclear power. This paper argues that the key to understanding this failure is the electoral clout 
The Influences of the Economic Geography on Elections: The Case of the Electric Power Industry in Japan
－ 29 －
of the labor unions representing workers in the electric power industry. To campaign against 
nuclear energy promises to win some unknown number of votes from unorganized voters but 
would lose a significant number of organized votes from the labor union of the electric power 
industry. In particular, among the electric power industry, workers in the nuclear power industry 
are mobilized more than in other types of electric power industry such as the thermal and hydro-
power. The economic geography of the nuclear power plants influences voter mobilization. 
2. The Democrats and the Nuclear Power Policy
The Democratic Party of Japan (DPJ) was founded in 1996. Most of its original members came 
from the Japan Socialist Party (JSP) and the new party Sakigake. The JSP had been the primary 
opposition to the LDP from the end of World War II until 1996. It was supported by the Sōhyō 
labor federation led by public-sector unions. Though the history of both the JSP and Sōhyō is 
complex, by 1996 they both opposed nuclear power. Sakigake, a small new party formed in 1993 
from defectors from the LDP, was the closest Japan has yet come to having a green party and 
opposed nuclear power. The Democrats thus started with a clear stance against nuclear power. 
 The Democrat’s position on nuclear power is more popular than the LDP’s position. In the 
2012 general election, 75 percent of respondents reported that they would take nuclear energy 
policy into account when voting, 28 percent said it would weigh heavily in their decision, and 
only 11 percent favored the continued use of nuclear power (Asahi shimbun 10 December 
2012). Even more Horiuchi, Smith, and Yamamoto (2018: 11) find that the strict position 
toward nuclear energy taken by opposition such as ‘strict guidelines for restart reactors’ or 
‘non-restart reactors’ by opposition parties including the DPJ is more popular than the less strict 
position offered by the Liberal Democratic Party (LDP). 
 It would thus appear that nuclear power meets all of the criteria necessary to affect election 
outcomes but the evidence also indicates that the issue has not benefited the Democrats or hurt 
the LDP. In the 2012 general election, Asahi’s exit polls show that, in the PR tier, the LDP won 
the 16 percent of those voters who wanted to end dependence upon nuclear energy immediately, 
the same percentage as the Democrats and the Communists, the latter being the party most 
clearly opposed to nuclear energy. The LDP also won 28 percent of those who wanted to move 
toward zero, more than the Democrats at 18 percent (Asahi exit poll of December 17, 2012, 
reported in Endo, Pekkanen, and Reed 2013: 61). 
 Then, why have the Democrats failed to attract the voters of those who opposed nuclear 
energy by campaigning on the issue to inform anti-nuclear policy?  One of the reasons is their 
Fumi Ikeda
－ 30 －
support organizations and labor unions. Though the Democrats started with a clear anti-nuclear 
power stance, the merger of parties supported by private sector labor unions, Dōmei labor feder-
ation, made it for the Democrats difficult to take a clear stance against the anti-nuclear power 
policy. Different from public sector unions, private sector unions supported the nuclear power 
policy. Considering maintaining a solid political party support base, taking the anti-nuclear 
power policy was not the best choice for the DPJ.
3. Estimating the Electoral Clout 
The Federation of Electric Power Related Industry Worker’s Union of Japan (EPWU, Denryoku 
Sōren) is one of the private sector unions and has been the influential supporter of the Demo-
crats. EPWU is formed by workers in the electric industry and promotes the use of nuclear 
energy (EPWU website). 
 To estimate the electoral clout of EPWU, I use the data of the open-list proportional repre-
sentation (OLPR) system to elect the upper tire of the House of Councillors (HoC). This 
system, and some of the practices that have evolved to take advantage of it, make it possible to 
estimate the number of votes mobilized by some unions for each municipality nationwide. 
 The election takes place in a nationwide district, so all voters are faced with precisely the 
same choice set making the returns equivalent for every municipality in the nation. Voters cast 
a single vote. They may cast it either for a party or for a candidate. All candidates are on a party 
list. The number of seats allocated to a party is the sum of the party vote and the vote for all of 
the candidates on its list. Seats won by the party are then allocated to candidates in the order of 
the number of personal votes they received. A candidate then needs to get around 100,000 voters 
to write her name on the ballot, neither anyone else’s name nor the name of a party. Between 
2001 and 2016, the minimum number of personal votes a candidate needed to win a seat for the 
Democrats ranged from 67,612 to 152,121 votes. 
 Some candidates such as celebrity candidates can obtain the necessary number of personal 
votes based on name recognition. Most, however, require organizational support. Some candi-
dates rely on their organizations, usually those built for campaigning in other elections, and 
others draw on several different organizations. Many of the most successful, however, depend 
almost completely on support from a single nationwide organization. These organizations 
“sponsor” candidates and take full responsibility for electing them (Klein and Reed 2014). 
 When a candidate depends upon an organization, the organization’s primary task is 
to campaign among its members, repeating the candidate’s name and some simple message 
The Influences of the Economic Geography on Elections: The Case of the Electric Power Industry in Japan
－ 31 －
indicating that she is “one of us”. The organization may also do more, contribute money or 
provide campaign staff, but the key is communicating the candidate’s name and convincing their 
members to go the trouble of voting for that candidate instead of simply writing the name of the 
party. The vote for the sponsored candidate is thus a good measure of the organization’s capacity 
to mobilize voters. 
 Using the vote of the sponsored candidate as a measure of an organization’s electoral clout 
has several attractive characteristics. First, because the PR election is conducted in a single 
nationwide district, we obtain equivalent estimates for every municipality in Japan. Second, 
because Councillors serve six-year terms but elections are held every three years, successful 
candidates cannot run in the subsequent election and the organization’s candidate often changes. 
If the vote remains stable even when the candidate changes, it represents an organizational vote, 
neither a personal nor a party vote. Organizational votes are important in all Japanese elections 
but are difficult to measure in any other elections. The OLPR and the practice of sponsoring 
candidates render the organizational vote visible and available for analysis. 
4. The Economic Geography and the Electoral Clout of the EPWU  
EPWU has sponsored candidates and made them elected in the OLPR of the HoC from 2001 to 
2019. By using the OLPR data from 2001 to 2019, I examine where the EPWU has been power-
ful and where it has been weak. 
 The dependent variable is the percentage of eligible votes obtained by the EPWU’s spon-
sored candidates in the OLPR of the HoC from 2001 to 2019. The unit of analysis is munici-
palities. The HoC elections data was collected from Asahi Shimbun (Sangiin Senkyo no Subete 
2001-2019). The independent variable is the locations of the nuclear power plants. I code 
dummy variables. If a municipality contains a nuclear power plant, the dummy variable is coded 
as 1. Municipalities without a nuclear power plant are coded as 0. 
 Six control variables are included. First, I include dummy variables of the other three types 
of electric power plants: thermal, solar, and hydropower plants. Municipalities with these power 
plants are coded as 1 while municipalities without them are coded as 0. Locations of these power 
plants potentially influence the voter mobilization capacities. 
 The percentage of workers in the electric power industry in each municipality is another 
control variable. The number of employees in municipalities might influence voter mobiliza-
tion. It expects that more votes are cast when more workers are in municipalities. The third 
control variable is area size. Area size influence voter mobilization by labor unions (Anzia, 
Fumi Ikeda
－ 32 －
2011).
  Finally, I included the population rate in densely inhabited districts (DIDs). DID popula-
tion ratio shows whether municipalities are rural or urban. It expects that the demographic trends 
influence voter mobilization. In urban areas, the percentage of independent voters tends to be 
higher. Consequently, it might not be easy to mobilize the same amounts of voters constantly. 
In contrast, in rural areas, the percentage of independent voters tends to be lower than in urban 
areas while the number of core voters tends to be larger. These data are collected from the 
National Census (kokusei-chōsa).
 Table 2 reports the coefficient estimates for the ordinary least squares (OLS) regression. 
Models 3 and 4 include dummy variables in 2013, 2016, and 2019 which are years of the HoC 
elections held after the Great East Japan Earthquake in 2011. Fixed effects of years of each elec-
tion are included in models 5 and 6. 
 There are two significant findings from models 1 to 6. First, there is a statistically positive 
coefficient between the percentage of votes for the EPWU and locations of all types of electric 
power plants. This result shows that the economic geography of power plants influences elec-
toral outcomes. Second, from models 1 to 6, the coefficient estimates provide an evidence that 
locations of nuclear power plants are more influential in voter mobilizations than other types of 
power plants. These two findings are consistent from models 1 to 6. 
 In addition to these core findings, the result shows that the number of employees in the 
energy industry sector has a statistically positive coefficient with the number of votes. This 
implies that more voters are mobilized when more workers in the electric power industry are in 
municipalities. Moreover, this result shows that area size does not have a statistically significant 
coefficient on the number of votes. Area size does not influence the amount of voter mobiliza-
tion. Finally, DID population ratio has a statistically negative coefficient on the number of votes. 
Votes are mobilized more in rural areas with lower DID population ratio than in urban areas with 
higher DID population ratio.  
The Influences of the Economic Geography on Elections: The Case of the Electric Power Industry in Japan
－ 33 －
Table 2 The Effects of Locations of Power Plants on Voter Mobilization
(1) (2) (3) (4) (5) (6)
Nuclear power plant 3.554*** 2.760*** 3.554*** 2.752*** 3.554*** 2.751***
(0.0708) (0.0763) (0.0701) (0.0755) (0.0693) (0.0746)
Thermal power plant 0.316*** 0.251*** 0.316*** 0.251*** 0.316*** 0.251***
(0.0345) (0.0336) (0.0341) (0.0333) (0.0338) (0.0329)
Hydropower plant 0.211*** 0.161*** 0.211*** 0.162*** 0.211*** 0.164***
(0.0287) (0.0280) (0.0284) (0.0277) (0.0281) (0.0274)
Solar power plant 0.603*** 0.606*** 0.603*** 0.607*** 0.603*** 0.608***
(0.0804) (0.0782) (0.0796) (0.0774) (0.0787) (0.0764)
Workers in 
the electric power 
industry (%)
0.228*** 0.229*** 0.230***
(0.00839) (0.00830) (0.00820)
Area size 0.00626 0.00626 0.00626
(0.00768) (0.00760) (0.00751)
DID population ratio -0.0355*** -0.0334*** -0.0285***
(0.00775) (0.00767) (0.00758)
Constant -0.0867*** -0.0694*** -0.00305 0.0143 -0.0868*** -0.0695***
(0.00859) (0.00838) (0.0109) (0.0106) (0.00841) (0.00819)
Observations 13,219 13,202 13,219 13,202 13,219 13,202
Dummy 
(2013, 2016, 2019) No No Yes Yes No No
Year-Fixed Effects No No No No Yes Yes
R-squared 0.173 0.217 0.190 0.234 0.180 0.225
Standard errors in parentheses *** p<0.01
 
Fumi Ikeda
－ 34 －
5. Conclusion 
Why are the Democrats not likely to promote the anti-nuclear policy and attract voters who 
are against nuclear power? This study argues that labor unions, support organizations of the 
Democrats, is on of reasons. The electoral clout of lavor unions is influenced by the economic 
geography. Locations of power plants influence on voter mobilization by EPWU. In particular, 
EPWU can mobilize more votes in the locations of nuclear power plants than in the locations of 
other types of power plants. Thus, the economic geography of the electric industry, in particular 
the locations of nuclear power plants, affects electoral outcomes. 
 This study has several limitations. First, this study focuses on EPWU and does not examine 
the electoral clout of other labor unions promoting the nuclear power policy. For instance, labor 
unions formed by workers in automobile and electronics industry support the nuclear power 
policy. To examine the influences of the nuclear power industry on electoral outcomes, the 
future study needs to examine these unions. 
 Another limitation is that this study does not reveal mechanisms of effects of the economic 
geography on voter mobilization. This study does not explain why more voters are mobilized in 
the locations of the nuclear power plants than in the locations of other types of electoral power 
plants. It expects that future studies examine mechanisms.  
The Influences of the Economic Geography on Elections: The Case of the Electric Power Industry in Japan
－ 35 －
References
Books and Journals
Aldrich, Daniel P. 2008. Site Fights: Divisive Facilities and Civil Society in Japan and the West (Ithaca: Cornell 
University Press).
Anzia, Sarah F. 2011. “Election timing and the electoral influence of interest groups” Journal of Politics 73(2): 
412–427.
Carlson, Matthew M, Steven R. Reed. 2018. Political Corruption and Scandal in Japan (Cornell University Press).
Endo, Masahisa, Robert Pekkenan and Steven R. Reed. 2013. “The LDP’s Path Back to Power” in Robert Pekkanen, 
Steven R. Reed, and Ethan Scheiner. Eds. Japan Decides 2012 (Palgrave Macmillan). 
Horiuchi, Yusaku, Daniel M Smith, and Teppei Yamamoto. 2018. “Measuring Voters’ Multidimensional Policy Prefer-
ences with Conjoint Analysis: Application to Japan’s 2014 Election” Political Analysis 26(2):190-209.
Klein, Axel, and Steven R. Reed. 2014. “Religious groups in Japanese electoral politics” in George Ehrhardt, Axel 
Klein, Levi McLaughlin, and Steven R. Reed. Eds. Kōmeitō: politics and religion in Japan (Institute of East Asian 
Studies)
Kushida, Kenji E. 2013. “Japan’s Fukushima Nuclear Disaster: Narrative, Analysis, and Recommendations” Shoren-
stein APARC Working Papers.
Meyer, Marco and Harald Schoen. 2017. “Avoiding Vote Loss by Changing Policy Positions: The Fukushima Disaster, 
Party Responses, and the German Electorate” Party Politics 23(4): 424-436.
Nakamura, Akira, and Masao Kikuchi. 2011. “What We Know, and What We Have Not Yet Learned: Triple Disasters 
and the Fukushima Nuclear Fiasco in Japan” Public Administration Review 71(6): 893–99.
Newspaper
Asahi Shimbun, various years 
Website
The Federation of Electric Power Related Industry Worker’s Union of Japan 
　http://www. http://www.denryokusoren.or.jp/ (access date: 1 June 2022)


English

The Influences of the Economic Geography on Elections: The Case of the Electric Power Industry in Japan

富田, 英司

坪田, 康

河村, 泰之

Japanese

愛大Minecraft 事業の効果に関する初期報告* -発達支持的生徒指導への示唆-

間々田, 理彦

サム, リアン

カンボジアの高度経済成長下における農業経営の実態と課題に関する一考察 -2010年代半ばを事例として-

佐藤, 康行

SATO, Yasuyuki

Neighborhood Association Comparison in Same Elementary School District: Case Study of the Shiojiri District, Ueda City

田口, 絹枝

TAGUCHI, Kinue

花と源氏宮: 狭衣物語における桜、山吹、藤

三輪, 咲愛

MIWA, Sakura

『桜姫全伝曙草紙』の桜姫: 物語における役割とイメージ

柳, 光子

アレクサンドル・デュマによる自作の翻案『銃士たちの青春』…

朝井, 志歩

馬毛島でのFCLP施設建設計画の経緯と問題点

笹沼, 朋子

パワー・ハラスメント防止義務における「人格の否定」について

梶原, 克彦

奈良岡, 聰智

第一次世界大戦と在独日本人の抑留問題(六)

Not a member yet