Koizumi Carried the Day:
Did the Japanese Election Results
Make People Happy and Unhappy?
Yoshiro Tsutsui (OsakaUniversity)
Miles Kimball (University of Michigan)
Fumio Ohtake (OsakaUniversity)
July 2006
Correspondence:
Yoshiro Tsutsui,
Institute of Social and Economic Research,
OsakaUniversity
6-1 Mihogaoka, Ibaraki,
567-0047 Japan,
Phone:+81-6-6879-8560
Fax:+81-6-6878-2766
e-mail:
1
Abstract
This paper investigates whether Japanese people became happy and unhappy with the general election conducted on September 11, 2005, in which the Prime Minister Koizumi won a land sliding victory. We conducted a large survey just after the election to ask their happiness and supporting parties. Although there are consistent tendencies thatsupporters for ruling parties became happier and supporters for opposition parties became unhappier, the effect was not significant. Considering the results of a previous study that Americans showed significant response to the result of a presidential election, this suggests that Japanese people are rather indifferent to politics.
Keywords: election, happiness, Koizumi Cabinet, survey, Japan
JEL classification number: I31, D72, C42
1
1. Introduction
How much are people interested in politics? Standard economic theory assumes that the people’s utility depends on their own consumption, so that it often disregards the importance of political institutions, although they essentially determine overall economic outcomes. Of course, people admit that political system is an inevitable infrastructure to achieve high economic performances and is very important to improve happiness of people (Frey and Stutzer, 2002).
Nevertheless, many people do not participate to any political activities because political system is public goods which can be free-ridden. Alternatively, political process can be regarded as a battle between two competing opinions instead of provision of public goods.Even in this case, however, people tend to think that they are too tiny and helpless to influence the outcome, so that they do not join to political activities. This is also a kind of a free rider problem within a political group. Actually, voting rate has been declining in many countries, including Japan and the U.S., andindifference to politics is an important social problem in democratic countries.
This paper investigates whether Japanese people became happy and unhappy with the election outcomes.Although democratic countries worry about their nation’s indifference to politics, there are no good ways to measure the degree of interest of nations in election. Asking people of their happiness after an election and comparing the happiness of winners and losers may be a good method to measure their interest in election because people who are interested in victory or defeat in an election should become happy or unhappy with results. Nevertheless,there have been only a few studies,to our knowledge, that investigated happiness of voters (Wilson et al., 2003 and Gilbert et al., 1998). These previous studies asked only 52 and 57 voters respectively, so that it is not convincing that their results are robust and represent a tendency of the United States. Meanwhile, we conducted a large survey in which we obtained about 1400 responses every month that are randomly sampledfrom all over Japan. Thus, this is the first comprehensive study to investigate how much people are concerned with an election.
Our study is also unique from the viewpoint of happiness study. Most of the surveys on happiness ask people of their overall feeling of happiness.[1] Typical question is: Overall, to what degree are you currently feeling happiness? Averaging happiness may make it similar to utility, so that such a question may be suitable for some cases, e.g. as a proxy for utility data. However, this type of question neglects important character of the feeling of happiness, i.e. it fluctuates with news and personal events (Kimball and Willis, 2006).In contrast, this study focuses on how people’s happiness is influenced with macro news, such as election results. In order to investigate how happiness depends on news and personal events, we did a survey on happiness which asks respondents of their current feeling every month in Japan.[2]
In this paper, we analyze the effect of general election conducted in September 11, 2005. This election was rather heated one and the Koizumi Cabinet won overwhelming victory that exceeded predictions by the media. Surprisingly, our study does not find significant change in happiness of winners and losers, suggesting that Japanese people are indifferent to politics.
The rest of this paper is organized as follows. In the next section, we first explain how the electionwas focused and outline of our survey. Then, we show that there is a tendency that supporters of ruling parties became happier and those for opposition parties unhappier, but they are insignificant. In section 3, we examine how the elation and dismay by the election faded away by looking happiness overthesix days after the election. In section 4, we analyze happiness of cabinet supporters. In section 5, we examine whether happiness of voters were affected by win or loose of their candidates. Section 6 are devoted to discussion and conclusions.
2. Did Japanese become happy and unhappy?
2.1. The general election in September 11, 2005
The general election was conducted on Sunday, September 11, 2005,and the Prime Minister Koizumi won a land sliding victory. Liberal Democratic Party (hereafter LDP) carried 296 seats out of 480 seats of the House of Representatives. Summing up the seats carried by New Komeito (Komei) with which LDP concludes alliance, the ruling parties carried over 2/3 (=327/480) of the House of Representatives. The Japanese Diet consists of the House of Representatives and the House of Councilors, and to be approved bills should clear the both. However, even if they are rejected in the House of Councilors after the approval of the House of Representatives, they are approved if it wins over 2/3 votes in the House of Representatives again. Therefore, the victory of over 2/3 seats in the House of Representatives implies thatthe Prime Minister Koizumi can pass any bills that he wants.
To understand why this election result is important, we need to explain the situation under which Koizumi dissolved the House of Representatives on August 8. The alliance of LDP and Komei commanded a majority in both Houses, so that they could have passed any bills in principle. However, the privatization of Japan Postal Service Public Corporation, which was the most important public promise of Koizumi, met strong opposition from many diet persons of LDP. In the consequence, the bill passed the House of Representatives by a hair, but was rejected in the House of Councilors on August 8. The Prime Minister Koizumi dissolved the House of Representatives at once, saying that he wants to ask the nation’s voice on the privatization. Thus, the general election became a kind of national referendum on the policy of the reform of postal services.How this election attracted the nation’s attention was reflected in the high casting rate: 68 % in this time vs. 60% in the last general election. As for the main events reported in news papers for the period from August to October, refer to Appendix.
2.2 Our survey
In this paper, we use the monthly surveys conducted from August 2005 to February 2006, focusing on the impact of the general election conducted on September 11, 2005.
Let us explain the outline of our survey, taking the September survey as an example. The surveys in the other months have quite similar features. 2000 people who were over 20 years old were randomly selected from all over Japan and interviewed.[3] Number of effective responses was 1399; the response rate was 70.0%. The survey was conducted from September 13 to September 20, the number of respondents on each date is shown in Table 1.[4]
Table 1 The number of responses on each day in September survey
total / 13-Sep / 14-Sep / 15-Sep / 16-Sepcount / 1399 / 1 / 149 / 192 / 183
ratio (%) / 100% / 0.1 / 10.7 / 13.7 / 13.1
17-Sep / 18-Sep / 19-Sep / 20-Sep
count / 327 / 353 / 190 / 4
ratio (%) / 23.4 / 25.2 / 13.6 / 0.3
653 respondents (46.7%) were male and 746 (53.3) were female. As for their age, those who were 20-29 years old share 10.4%, 30-39 years old, 18%, 40-49 years old, 17%, 50-59 years old, 22%, over 60 years old, 33%. As for academic background, 15% was grade school, 52%, high school, 32%, university. We also have the data on household income, occupation, residence location, and size of cities, as well as their attitude to the Koizumi Cabinet and supporting party and their view on business condition.[5]
749 respondents (53.5%) supported the Koizumi Cabinet, 416 (29.7%) opposeed to it and 234 (16.7%) answered‘do not know’. Supporting parties of respondents are shown in Table 2. 32% supported LDP and 15%, Democratic Party of Japan (DP), while 42% did not support any party. Thus, how to attract those who did not support specific party was the key to carry the election, and actually about a half of them voted for LDP according to interviews at the exit of polls, leading to the victory of Koizumi.
Table 2 The number of supporters of each party
total / LDP / DP / Komei / Communist / SDPcount / 1399 / 446 / 207 / 61 / 32 / 23
ratio (%) / 100 / 31.9 / 14.8 / 4.4 / 2.3 / 1.6
PNP / NPN / Other parties / No party / do not know
count / 3 / 1 / 3 / 589 / 34
ratio (%) / 0.2 / 0.1 / 0.2 / 42.1 / 2.4
Note: Communist stands for Japanese Communist Party, SDP, Social Democratic Party, PNP, The People’s New Party, NPN, New Party Nippon.
As for happiness, the following question is asked:
Please remember how did you feel in these one week. To what degree were you feeling happiness in the last week? Using a scale from 0-10 where “10” is “very happy”and “0” is “very unhappy”, how do you rate your level of happiness in the last week?
We define a variable,Happiness, for this value.The results of the surveyconducted in Septemberare shown in Table 3.Out of 1399, those who chose 5 was the largest (397; 28.4%), and those who chose 7, 8, or 6 were over 200 (about 15%). The average happiness was 6.3.[6]
Table 3. Happiness in September survey
total / 10 / 9 / 8 / 7 / 6 / 5count / 1399 / 82 / 69 / 242 / 211 / 203 / 397
% / 100 / 5.9 / 4.9 / 17.3 / 15.1 / 14.5 / 28.4
4 / 3 / 2 / 1 / 0 / do not know
count / 57 / 61 / 17 / 13 / 9 / 38
% / 4.1 / 4.4 / 1.2 / 0.9 / 0.6 / 2.7
2.3 Supporters of ruling parties became happier
We want to investigate whether happiness of supporters of ruling parties rose from August to September and whether happiness of supporters ofopposition parties fell from August to September. In this subsection, we calculate the averages of happiness for these groups for each month to compare them. Then, what we should be careful is that happiness changes over months from various reasons, so that we should adjust this variation caused by other events than the election.
As Figure 1 shows, the average happiness of all respondents substantially varied over the period. Especially, we are interested in a change in happiness from August to September and average happiness of all respondents in August is higher than that in September.Thus, if we find a decrease in happiness of supporters of ruling parties in September, it does not necessarily imply that the election results negatively affected the supporters. In order to identify the effect of the election results on happiness of supporters of ruling parties and opposition parties, it is necessary to adjust the change in average happiness between months. To do so, we divide the happiness of supporters of ruling and oppositionparties by the average happiness of all respondents of each month.
Figure 1. Average happiness of all the respondents
Indeed, happiness of supporters of ruling parties in September was lower than August. However, once we adjust the change in average happiness of all respondents, the outcome is reversed. Blue rectangles in Figure 2, which present the thus normalized happiness, reveal that supporters of ruling parties rose in September, then declined until December and rose again in January and February. The rise in September may be due to the victory in the election.
Figure 2. Normalized happiness of supporters of ruling parties,opposition parties, and those who support oppositions but not Koizumi
Note: Normalized happiness is (average happiness of a group)/(average happiness of the whole sample).
We present the normalized happiness of supporters ofopposition parties by red rectangles in Figure 2. They fell in September, and fell more in October, and then went up until January.[7]The result that they fell in September is consistent with the notion that it was caused by the defeat in the election. One may argue, however, thatthe result that they fell more in October than in September is unreasonable, if the result was caused by the defeat in the election.
Our survey also asks whether or not people support the Koizumi Cabinet.According to the results,a large portion (XX %) of supporters of opposition parties,at the same time,supported the Koizumi Cabinet in September.This is not strange because the Prime Minister Koizumi appeals that he is the real reformer of LDP, so that people who are against LDP may, at the same time, support Koizumi. They may feel happy to hear the election results that Koizumi won.
Considering this fact, it may be appropriate that we focus on the happiness of those who supported oppositions and did not support Koizumi Cabinet. Yellow rectangles in Figure 2 present the results, which reveal that the happiness of those people fell in September and rose in October, and then went up until December. This result is consistent with the supposition that people who were against Koizumi and LDP became unhappy by the defeat in the general election in September, 2005.
The ruling party consists of LDP and Komei, while opposition party consists of the DemocraticParty of Japan (DP), Japanese Communist Party (Communist), Social DemocraticParty (SDP), the People’s New Party (PNP), and New Party Nippon (NPN). Let us examinewhether the results on ruling and opposition parties above apply for each party. However, the supporters of Communist, SDP, PNP, and NPN are too few to get reliable statistic for each party, so that we aggregated the supporters of these small parties (SMALL) and looked at LDP, Komei, DP, and SMALL.
Blue rectangles in Figure 3 show the result of LDP. The happiness rose in September. But, the difference from August was only 0.004 and the happiness rose more in October. Thus, the rise of happiness of LDP supporters was small in September.
Red rectangles in Figure 3 show the result of Komei. The happiness rose in September and fell in October. The difference from August is 0.04, which is ten times larger than that of LDP. Thus, the rise of happiness of Komei (smaller ruling party) supporters is more vivid than LDP (larger ruling party) supporters.
Figure 3. Normalized happiness of supporters ofLDP and KOMEI
Figure 4. Normalized happiness of supporters ofDP, SMALL, and SMALL cum anti-cabinet
Blue rectangles in Figure 4 show the result of DP. The happiness fell in September and rose in October. However, the difference from August was only 0.001.
Red rectangles in Figure 4 show the result of SMALL, i.e. Communist, SDP, NPN, and PNP. The happiness fell in September. The difference from August was 0.04, which is forty times larger than that of DP. Thus, the fall of happiness of SMALL’s supporters was more vivid than DP’s supporters. Considering the results of LDP and Komei together, these results suggest that the supporters of smaller parties were more strongly affected to the results of the election.
However, the result that the happiness in October dropped much more than in September seems strange. This may be due to the fact that a considerable part of the supporters ofopposition parties, at the same time, supportedthe Koizumi Cabinet. For example, Among 32 supporters of Communist in September, 19 did not support the Cabinet and ten supported the Cabinet.[8] The average happiness of the former group was 5.2, while that of the latter group was 8.0. Thus the anti-cabinet Communist supporters were muchunhappier in September than pro-cabinet Communist’s supporters.
To get rid of the bias from pro-cabinet supporters ofopposition parties, we show the happiness of supporters of SMALL who were anti-cabinetby yellow rectangles in Figure 4.It fell in September almost by 0.10. While the happiness was lower in October than in September, the spread between September and Octoberwas only the half of the spread between August and September.
The results of Figures2-4 suggest that the supporters of ruling parties relatively felt happier in Septemberthan in August, while those for opposition parties felt less happiness in September than in August. Considering that the survey was conducted a few days after the election, the landslide victory of Koizumi may be the cause of this rise and fall of happiness. The effect was lager for the supporters of smaller parties: for ruling parties, the effect on Komei was larger than LDP, and for the opposition parties, SMALLwas larger than DP.
2.4 Is the change significant?
In this subsection, we will examine whether the rise and the fall of the happiness in September found in the former subsection is significant or not. To test the significance of the change in the happiness of supporters of ruling parties, we estimate the following equation.
(1)
where AUG, SEP, OCT, NOV, DEC, JAN, FEB are dummy variables standing for each month. DRULE is a dummy variable which takes on unity if the respondent is a supporter for ruling parties and 0 otherwise. Subscript i stands for respondents of all the period (i=1,…,8592). The change in happiness of supporters of ruling parties from August to September is measured by coefficient of SEP, .
The estimates of (1) are shown in Table 4, which reveals that average happiness in September is significantly lower (at the 5% level) than August (see SEP).[9]Supporters of ruling parties are significantly happier (at the 5% level) than others (see DRULE). The supporters of ruling parties became happier in September than in August, but this is not significant (p-value is 0.315, see DRULE*SEP). Still, the coefficient is larger and more significant than those of other months, suggesting that the impact of the election might be captured by this coefficient.