Layout Februari 2016 Does Democracy Make You Happy? Multilevel Analysis of Self-rated Happiness in Indonesia SUJARWOTO Department of Public Administration, Universias Brawijaya, Malang, Indonesia. Email: sujarwoto@ub.ac.id ABSTRACT The linkage between democracy and citizen happiness in developing countries is rarely examined. This study examines the link between democracy and citizen hap- piness in Indonesia, a new emerging democratic country in South East Asia. Data comes from the Indonesian Family Life Survey 2007 (N individual = 29.055; N household = 12.528; N district = 262) and the Indonesian Family Survey East 2012 (N individual = 5.910; N household = 2.546; N district = 55). Results from a three-level ordinary logit model show that democracy as measured by age of direct local democracy is not associated with citizen happiness. Instead of age of direct local democracy, district community social capital and spending public services give benefit for citizen hap- piness. The results are robust against individual and district characteristics related to happiness. The results highlight the importance of promoting community social capital and improving district capacity in delivering public service to improve citizen happiness in Indonesia. Keywords: democracy, happiness, three-level ordinary logit model, Indonesia INTRODUCTION During the last two decades, citizen happiness has become an important development agenda in Western countries, particularly rich countries across Europe and North America (Argyle 2001; OECD 2013). The economic crisis in Europe and United States at the end of 2009 have realised that economic welfare or GDP is not the only ultimate goal of development. The government and lead- ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ Recieved 11 September 2015; Accepted 07 September 2015 http://dx.doi.org/10.18196/jgp.2015.0021 26-49 27 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ Vol. 7 No. 1 February 2016 ers in these countries have realised that citizen happiness and life satisfaction are more importance. Since 2008 therefore, government and policy makers across European countries have promoted hap- piness as an indicator to measure national development progress. France and United Kingdom for example have included happiness in their national wellbeing index. The International development organisation such as OECD has also used this indicator for measur- ing social and development progress in their development program (OECD 2013). Studies to understand citizen happiness and its factors in West- ern Europe and United States have been widely documented. Among them conclude that democracy is an important factor of citizen happiness (Frey & Stutzer, 2000; Argyle 2003; Blanchflower & Oswald 2004; Lane 2009). Frey and Stutzer (2000) found that local autonomy and direct democracy both increase citizen happiness, and they suggested that this positive effect could be attributed to two factors: local citizens’ feeling of being more closely connected to political outcomes, and the benefits of political participation. However, Veenhoven (2000) found that political and private free- doms exert a positive influence on subjective wellbeing only in coun- tries with well-established democracies. In emerging democratic countries, the linkage between democracy and subjective wellbeing or happiness is still little known. This study examines the linkage between democracy and citizen happiness in Indonesia. The country is an emerging democratic country in South East Asia as its authoritarian political system was radically reformed in 1999. Since 1999, free and fair national, cen- tral elections of parliament and president have been introduced across the country. More than forty new political parties partici- pated in this direct election which, coupled with new media open- ness, supported a burgeoning country-wide democracy (National Election Committee 2004). From then onwards, citizens experienced http://dx.doi.org/10.18196/jgp.2016.0021 26-49 28 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ JOURNAL OF GOVERNEMENT & POLITICS increased freedom to make their voice heard through the press, over which the government has continued to relax its control. Moreover, in 2001, Indonesia embraced local autonomy, a move which trans- formed the country’s local government political system. Local au- tonomy has given every district or local government the power to perform the key functions of state, including the provision of health, education, environmental and infrastructure services. Further re- forms in 2005 allowed citizens to elect their own mayor and parlia- ment through direct local elections: by the end of 2006, more than half of all districts had conducted direct elections (The Ministry of Home Affair 2007). All these national and political reforms may affect people’s lives and their happiness across archipelago. This study contributes to existing literature on democracy and happiness in the following ways. First, it combines insight from two strands of earlier research on contextual and individual factors of citizen happiness in developed countries (Argyle 2003; Blanchflower and Oswald 2004; Diener and Biswar-Diener 2008; Graham 2009; Lane 2009). Second, by focusing on Indonesian local democracy, this study provides a contrast to the far more extensive work on happiness and wellbeing that draws on data from developed coun- tries and Western country settings (especially the UK, US and West- ern Europe). Third, prior studies examining the linkage between democracy and citizen happiness used either aggregate data at the individual level (see for example Frey and Stutzer 2002) or ignore the nested structure of data (see for example Bjornskov et al. 2008). It is crucial to distinguish the individual-compositional sources of variation from the place-contextual ones in order to minimize con- founding (Subramanian et al. 2001). The three-level ordinary logit model used in this study is able to account for the clustering of individuals within households and districts by separating their vari- ance in happiness from the household and district variance (Rabe- Hesketh and Skrondal 2012). Using this model is thus most appro- http://dx.doi.org/10.18196/jgp.2016.0021 26-49 29 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ Vol. 7 No. 1 February 2016 priate for examining the effects of individual and contextual sources of happiness on citizen happiness. THEORETICAL FRAMEWORK Social scientists have long been interested in the study of how and why people are happy and unhappy; they do so by identifying factors associated with happiness. The findings of previous studies suggest that the factors contributing to happiness fall into two broad categories: contextual and individual. Contextual sources of happi- ness pertain to factors at the community and state level such as democracy, social capital, better environment, good government and per capita GDP (Argyle 2003; Blanchflower and Oswald 2004; Lane 2009). Individual sources of happiness pertain to the socio-demo- graphic characteristics of individuals and their unique economic circumstances. The individual sources include gender, age, marital status, companionship, social ties, religiosity, health, employment status, education and personal or household income (Argyle 2003; Diener and Biswar-Diener 2008; Graham 2009). Figure 1 shows the interrelationships among contextual and individual sources of happiness. The model allows for an interaction effect between indi- vidual and contextual factors in the production of happiness. FIGURE 1: INTERRELATIONSHIP BETWEEN INDIVIDUAL AND CONTEXTUAL SOURCES OF HAPPINESS http://dx.doi.org/10.18196/jgp.2016.0021 26-49 30 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ JOURNAL OF GOVERNEMENT & POLITICS Contextual sources of happiness have been a consistent topic of interest for sociologists, economists and political scientists. Research shows democracy, social capital, better environment, good govern- ment and GDP to be good predictors of happiness. Frey and Stutzer (2000) found that local autonomy and direct democracy both in- crease happiness, and they suggested that this positive effect could be attributed to two factors: local citizens’ feeling of being more closely connected to political outcomes, and the benefits of politi- cal participation. However, Veenhoven (2000) found that political and private freedoms exert a positive inf luence on subjective wellbeing only in countries with well-established democracies. Rodriguez-Pose and Maslauskaite (2012) found that good govern- ment as measured by its capacity to deliver public services is posi- tively associated with citizen happiness. Bjornskov et al. (2008) found that the higher performance of economic and judicial institutions affects happiness in medium- and high-income countries. Studies have also shown the benefit of community social capital for indi- vidual happiness (Putnam 1995; Bjornskov 2003). Putnam (1995) explained that social capital provides a channel for the personal and social support that increases individual happiness. Mookerjee and Beron (2005) find that religious fractionalization is associated with reduced happiness. Thao et al (2009) argue that ethnic diver- sity may relate with less happiness. Some scholars have also found that higher GDP increases citizen happiness. Corruption has a det- rimental effect on happiness (Bjornskov et al. 2008). Di Tella et al. (2003) and Helliwell et al. (2013), among others, have shown that people who live in countries with higher GDPs have higher levels of happiness. Individual sources of happiness have been identified by most psychologists and economists. Studies have found gender, age, mari- tal status, health, companionship, religiosity, social ties, employment status, education and personal or household income to be predic- http://dx.doi.org/10.18196/jgp.2016.0021 26-49 31 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ Vol. 7 No. 1 February 2016 tors of individual happiness. Women tend to be happier than men (Graham 2009); one somewhat contentious explanation for this is that women tend to have lower levels of aspiration and, thus, a higher level of happiness (Frey and Stutzer 2002). Age is positively associ- ated with happiness, with older people likely to be happier than younger people (Argyle 2001). Blanchflower and Oswald (2008) posited a U-shaped relationship between age and happiness, dem- onstrating that both younger and older people tend to be happier than middle-aged people. Marital status and companionship are consistent predictors of happiness, with married individuals and individuals in a partnership likely to be happier than widowed and divorced individuals (Blanchflower and Oswald 2004); loneliness has the effect of decreasing happiness (Lane 2001). Religiosity and social ties are also sources of individual happiness. Abdel-Khalek (2006) found that people who are religiously devout tend to enjoy not only better mental health but also higher levels of happiness. Helliwell (2006) found that social ties are one of the main factors of subjective wellbeing. Physical health is a stable predictor of happi- ness; a healthy individual is likely to be happier than an unhealthy individual (Graham 2009). Employment is positively associated with happiness, with employed people likely to be happier than those who are unemployed (Clark 2003); unemployment has a detrimen- tal effect on happiness (Clark and Oswald 1994). Prior studies also show that education may contribute to happiness by enabling indi- viduals to better adapt to a changing environment (Clark and Oswald 2004). Household and individual incomes are consistent predic- tors of happiness (Lane 2003; Blanchflower and Oswald 2004). RESEARCH METHOD INDONESIAN FAMILY LIFE SURVEY (IFLS) AND OFFICIAL STATISTICS To examine the linkage between democracy and citizen happi- http://dx.doi.org/10.18196/jgp.2016.0021 26-49 32 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ JOURNAL OF GOVERNEMENT & POLITICS ness in Indonesia, we assembled individual and district level data from various sources. In this analysis, the data possesses a multi- level structure, with individuals nested within households and dis- tricts. Data on individuals and households is taken from IFLS 2007 and IFLS East 2011, while district data comes from Indonesian Vil- lage Potential Census (PODES), Indonesian Population Census, and official statistics. Our analysis used the IFLS 2007 and IFLS East 2012 as the hap- piness question was first asked in both surveys. The sample in the surveys was restricted to respondents aged 15 years and older who gave complete information on the happiness question. The IFLS 2007 sample included 29,055 individuals from 12,528 households living in 262 districts, which corresponds to approximately 99% of the IFLS 2007 sample included in the happiness module. On aver- age, there were 4-5 individuals within each household and 100 house- holds within each district. The IFLS East 2012 sample included 5,910 individuals from 2,546 households living in 55 districts, which cor- responds to approximately 99% of the IFLS 2012 sample included in the happiness module. On average, there were 2-3 individuals within each household and 70-100 households within each district. The IFLS data was linked to a number of other surveys and offi- cial statistical datasets using district codes. First, we linked the IFLS data with the local and national election database of the Indone- sian Ministry of Home Affairs as the data contains information about districts that by 2007 had elected their major and local parlia- ment through direct election. Next, we linked it with the govern- ment Village Potential Statistics (PODES) 2006 and 2011. PODES contains detailed information about the incidence of local conflict and violence as well as the number of community groups within districts, calculating aggregates at village and urban neighbourhood levels to measure their distribution. Third, we linked the IFLS with the ethnic and religious fractionalisation index measured from the http://dx.doi.org/10.18196/jgp.2016.0021 26-49 33 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ Vol. 7 No. 1 February 2016 Indonesian Population Census 2010. Fourth, we linked the IFLS data to district fiscal data. Collected by the Ministry of Finance, this dataset provides information about district spending for public services (Indonesian Ministry of Finance 2008). We use fiscal data from 2006 and 2011 (the year prior to the IFLS survey), as district development spending in the Indonesian budgeting system takes at least one year to take effect. MEASURE OF DEMOCRACY AND CITIZEN HAPPINESS Democracy is measured by age of district direct election. It as- sumes that districts which have older age of local democracy have more mature democracy than district with younger age. Citizen happiness is measured by self-rated happiness (Veenhoven 1984). In the IFLS survey, respondents were asked: “Taken all together, how do you feel these days: ‘very happy’, ‘pretty happy’, ‘not too happy’ or ‘very unhappy’?” This item has been widely used and has been validated during use in previous studies (Frey and Stutzer 2002; Krueger and Schkade 2008; Oswald and Wu 2011). CONTROL VARIABLES Control variables include contextual and individual/household factors of happiness. Contextual factors include district GDP, com- munity social capital, spending for public services, local conflicts and violence, ethnic and religious fractionalisation. Following Putnam (1993), we use the density of community groups active in a district to measure social capital. This provides information about, among others, kelompok pengajian and kelompok kebaktian (religious groups), karang taruna (youth groups), persatuan kematian (funeral groups), and kelompok wanita (women’s groups), all active commu- nity groups found within villages or urban neighbourhoods in In- donesia. We use district spending for public services to measure district capacity to deliver public services. To address whether local http://dx.doi.org/10.18196/jgp.2016.0021 26-49 34 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ JOURNAL OF GOVERNEMENT & POLITICS conflicts and violence affect happiness, we include number of local conflicts and violence in the estimation. Ethnic and religious fractionalisation were measured as one minus the Herfindahl in- dex of ethnic or religious group shared, and reflected the probabil- ity that two randomly selected individuals from population belonged to different groups (Alesina et al. 2003). Individual and household factors of happiness are also included for control variables. Indi- vidual factors of happiness were selected from prior studies. These individual factors include age, gender, per capita household expen- diture, poverty status, years of schooling, employment status, mari- tal status, health, social ties, trust, ethnicity, religiosity and social participation, loneliness, parental status, and migration. Meanwhile, household factors include household expenditure, household loca- tion in an urban or in a rural area and household size. THE THREE-LEVEL ORDINARY LOGIT MODEL A three-level ordinary logit model is used to account for indi- vidual, household and district characteristics on citizen happiness. This model is more appropriate than the use of Ordinary Least Squared (OLS) regression, which aggregates data at the individual level and ignores the nested structure of data; if ignored, this nest- ing of individuals within household and district units can lead to the underestimation of standard errors regarding the effect of house- hold and district characteristics (Snijders and Bosker 1999). One consequence of failing to recognise hierarchical structures is that standard errors of the regression coefficient will be underestimated, leading to an overstatement of statistical significance. Standard er- rors of the coefficient of higher-level predictors will be the most affected by ignoring the grouping as does OLS estimation. Three-level ordinary logit model, or multilevel, analyses combine the regression and the variance component models to account for the nested structure of the data. This model accounts for the clus- http://dx.doi.org/10.18196/jgp.2016.0021 26-49 35 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ Vol. 7 No. 1 February 2016 tering of individuals by separating individual variance in happiness from that of household and district variance in happiness. The to- tal variance is partitioned into the variance between households and districts (ó2 µ ) and the variance within households and districts (ó2 e ). These variances capture the effects of unobserved heterogene- ities, variables that are independent of the covariates in the model (Rabe-Hesketh et al. 2012). We carried out three-level ordinary logit models with generalised linear latent and mixed models commands (GLLAMMs) using Stata 13.0 software. Rabe-Hesketh et al. (2004) explained that GLLAMMs are a class of multilevel latent variable models for (multivariate) responses of mixed type, including con- tinuous responses, counts, duration/survival data, dichotomous, ordered and unordered categorical responses, and rankings. In this analysis, GLLAMM is used with an ordinal logit link as the depen- dent variable (self-rated happiness) is ordinal. For each of the mod- els, the estimated coefficient, standard errors, individual, house- hold and district variances, and log likelihood as an indicator of model fit are reported. All models were estimated using maximum likelihood estimation. FINDINGS First, we describe the district characteristics of the survey. The surveys reflect the contrast in socio economic development between western and eastern Indonesia, particularly concerning GDP, spend- ing for public services, ethnic and religious fractionalisation, social capital and age of democracy. As IFLS East was collected in 2012, the age of local direct democracy in eastern Indonesia is longer. Levels of district GDP and spending for public services in eastern Indonesia are greater than in western Indonesia. Community social capital in the eastern region is substantially greater than in the west- ern region. Ethnic and religious fractionalisation in eastern Indo- nesia is higher than in western Indonesia. Poverty as measured by http://dx.doi.org/10.18196/jgp.2016.0021 26-49 36 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ JOURNAL OF GOVERNEMENT & POLITICS percentage of households with consumption of less than 2 USD per day is fairly similar in both regions; poverty in the eastern re- gion is slightly higher at 68%. The mean self-rated happiness score was 3.00 (SD = 3.55-3.58) for both samples, measured on a 1 (very unhappy) to 4 (very happy) scale. Consistent with the Ipsos 2010 and the 2012 Happy Planet Index results, most respondents in the surveys report happy and very happy (94% for IFLS 2007 and 88% for IFLS East 2012). The detailed of socio-demographic variables in both samples are shown in Table 1. Bivariate ordered logistic regression in Table 1 shows that most variables were statistically significantly associated with happiness at the 5% level. However, age of local democracy is not significantly associated with citizen happiness. Ethnic fractionalisation, commu- nity social capital and spending for public services are strongly re- lated with happiness. Religious fractionalisation and GDP have a weak correlation with happiness. Local conflict and violence are negatively associated with citizen happiness. Only a small number of observations were missing. With the relatively small amount of missing data, it is expected that bias and efficiency loss in the esti- mation of the multivariate models are minimal. Table 2 shows results of the three-level ordered logit model. To examine the degree of variance in citizen happiness at the individual, household, and district levels, we estimated an empty model before regressing happiness on any predictor. The empty model includes only variance components; it is not reported in the tables. It re- vealed that the unexplained individual variance in citizen happi- ness is 0.049, the unexplained household variance in happiness is 0.038, and the unexplained district variance comes to 0.007. Thus, the intra-class correlation (ICC) for individuals within a household is 0.124; for individuals within a district the ICC is 0.101. This means that 12.5% of the variation occurs between households and 10% http://dx.doi.org/10.18196/jgp.2016.0021 26-49 37 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ Vol. 7 No. 1 February 2016 TABLE 1 SAMPLE CHARACTERISTICS AND BIVARIATE ORDERED LOGISTIC ANALYSIS http://dx.doi.org/10.18196/jgp.2016.0021 26-49 38 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ JOURNAL OF GOVERNEMENT & POLITICS occurs between districts; ignoring this can lead to inefficient and biased estimates. Across the models, democracy is not significantly associated with citizen happiness (0.020, *p >5%). The widespread local conflict and violence following democracy and decentralisation in the local level have detrimental effect on citizen well-being. The negative as- sociation of local conflicts and violence on citizen happiness is con- sistent in the models. The negative association between district GDP and citizen happiness indicates the paradox of economic growth in the country archipelago. This finding signals that district economic development does not go with citizen happiness. As districts across Indonesia are rich with community social capital, this community social capital gives benefit for citizen well-being. Individuals who live in district with richer community social capital are happier than those who live in district with poorer community social capital. In- stead of age of direct local democracy, the capacity of district gov- ernments in spending for public services gives benefit for citizen well-being. Across the models, spending for public services is sig- nificantly associated with happiness. Studies also report that higher ethnic fractionalisation across the Indonesian archipelago often leads to ethnic conflicts, which have detrimental effect on citizen well- being (Kaiser and Hofman 2003; World Bank 2006; Baron et al. 2009). The negative association between ethnic fractionalisation and citizen happiness is shown in this study. Most of the individual and contextual factors show consistent results with existing studies. Being married, being healthier and richer, being educated and employed, being religious, and having social ties are source of happiness in the individual and household level in Indonesia. However, happiness is beyond ethnicity and reli- gious affiliation. The findings indicate that basic sources of citizen happiness in the country are the same regardless of ethnic and reli- gious differences. Age is positively associated with happiness. Mar- http://dx.doi.org/10.18196/jgp.2016.0021 26-49 39 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ Vol. 7 No. 1 February 2016 TABLE 2: REGRESSION RESULTS OF THREE-LEVEL ORDERED LOGIT MODEL 40 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ JOURNAL OF GOVERNEMENT & POLITICS ried individuals are happier than divorced and widowed individu- als. Women are happier than men are. Richer household is happier than poorer household. Employed and educated individuals are also happier. Support and good social relationship benefit for happi- ness. Consistent results are shown from the association of social ties, trust and social participation on happiness. In contrast, loneli- ness and childlessness have detrimental effect on happiness. Reli- gious individuals are happier, but religious affiliation is irrelevant. Ethnicity affiliation is also irrelevant. In all models, we found being Muslim and being Javanese are not significantly associated with happiness. The null association of being Javanese is not surprising in that little ethnic discrimination, whether de jure of de facto, has existed in modern, independent Indonesia. Some studies argue that Chinese have been discriminated against Javanese in Indonesia, but considering their large wealth, they are not so much discriminated against as envied (Suryadinata et al. 2003). The significant associa- tion of urban residence and happiness is only shown for eastern Indonesia. The variances at the household and district levels are significant across all specifications. The estimation of these works towards en- suring the remaining estimates is robust against unobserved house- hold and district heterogeneities. Single-level studies, which ignore unobserved heterogeneities at either the household or the district level, may not be as robust. This is worth bearing in mind when comparing these results with those discussed in the current litera- ture. DISCUSSION This paper examines the linkage between democracy and citizen happiness in Indonesia. Indonesia provides an interesting case not only because recent findings show the country to be one of the happiest in the world but also because relatively little research has http://dx.doi.org/10.18196/jgp.2016.0021 26-49 41 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ Vol. 7 No. 1 February 2016 explored to examine the relationship between current local democ- racy development and citizen happiness in the country. The main findings show that democracy as measured by age of local democracy is not associated with citizen happiness. Instead of democracy, community social capital and district capacity in deliv- ering public services are district factors of citizen happiness. The null association of local democracy on citizen happiness may indi- cate ineffectiveness of local democracy in the country. Ineffective- ness of local direct democracy has been noted by some studies. Aspinall and Mietzner (2010) note the influence of political elites, patronage and ‘money politics’ in local elections. Local democracy in Indonesia has also been accompanied by an increasing amount of local conflict, which has not only rendered local democracy less effective (Choi 2004; Nordholt and Van-Klinken 2005), but which is also in itself detrimental to citizen happiness (Sujarwoto and Tampubolon 2014). Under such conditions, the widening partici- pation of political parties and local direct elections may not guaran- tee effective local leadership, which can provide better local policies and services, which improve wellbeing. The pessimism of citizens on effectiveness of local democracy to bring effective local leader- ship was also indicated from the increasing number of non-voters in local elections since political reform from 20% in 2005 to 40% in 2013 (The Ministry of Home Affairs 2014). Although some qualitative studies report such local conflicts were also related to religious tension (Suryadinata et al. 2003; Klinken 2007; Choi 2004), we do not find significant relation between dis- trict religious fractionalisation and citizen happiness. Religious af- filiation is also not significantly associated with happiness. Similar finding were also shown by the studies of Ferriss (2002) and Sohn (2013). They find that the levels of happiness differ little among Muslims, Protestants, Christians, Catholics and Jews. In our study, the effect of religion on happiness is shown in the form of indi- http://dx.doi.org/10.18196/jgp.2016.0021 26-49 42 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ JOURNAL OF GOVERNEMENT & POLITICS vidual religiosity. We find those who do daily prayers are happier than those who do not do daily prayers. Further, the positive asso- ciation between religious fractionalisation and individual religiosity may signal the way individual religious freedom is expressed and associated with citizen happiness, as suggested by some scholars that the majority of Indonesians harbour tolerant and friendly towards other faiths, especially in the context of the wider Muslim world (Pizani 2014). Indonesia is a society in transition, but retains many traditional features. Ethnicity still plays important roles in citizen’s daily life and therefore may influence citizen happiness (Ricklefs 2001; Suryadinata et al 2003). The importance of ethnicity on happiness is shown in this study. In all models, we find that citizen who live in district with higher ethnic fractionalisation are less happy than those who live in district with less ethnic fractionalisation. The negative association of ethnic fractionalisation may relate with increasing conflict and violence in high ethnic fractionalisation districts in Indonesia since 1998. Nordholt and Van-Klinken (2005) for example show that increasing conflict involving ethnic groups since the fall of the authoritarian regime in 1998 mostly occurred in districts with high ethnic fractionalisation such as Maluku, Kalimantan, and Papua. Ethnic fractionalisation in conflict districts in those islands resulted in social tension and reduced social tolerance, which may also make citizen less happy. The benefits of indigenous community social capital across In- donesia archipelago such as kerjabakti (voluntary labor), arisan or binda (rotating credit association), kelompok pengajian and kelompok kebaktian (religious groups), karang taruna (youth groups), persatuan kematian (funeral groups), and kelompok wanita (women’s groups) on various aspects of citizen well-being were documented in prior studies. Miller et al. (2006) for example found that these type of community social capital benefits mental health among adults in http://dx.doi.org/10.18196/jgp.2016.0021 26-49 43 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ Vol. 7 No. 1 February 2016 Indonesia. Beard (2007) shows the positive role of this indigenous tradition in improving community well-being. Sujarwoto and Tampubolon (2013) explain some benefits of community and indi- vidual social capital for well-being, which include support from com- munity and family members, better access to public services and amenities, and sharing knowledge and information. Such channels may link the positive association of social capital and citizen happi- ness in Indonesia. For example, the importance of family and community support in improving well-being was highlighted in the recent BPS 2013 survey, which found that time spent with family and neighbours are among main factors of household happiness in Indonesia (BPS 2014). The association between spending for public services and citizen happiness signal that happiness in the country is increased through better capacity of district government to deliver public services. Using the IFLS 2007, Sujarwoto and Tampubolon (2014) found that fiscal decentralisation in Indonesia is significantly associated with citizen happiness. Bjørnskov et al. (2008) explain mechanisms by which fiscal decentralisation leads to citizen happiness. They find that it encourages local governments to provide citizens with the public goods and services that they need. It entails a shift in political deci- sion-making from central to local government, with the implication that local government has a greater potential to tailor its policies specifically to the demands of local citizens. This improved match- ing of public goods and service delivery to local needs necessarily increases happiness levels of the beneficiaries. Indonesians living in richer districts as indicated by higher GDP are less happy than those living in poorer districts. This negative association contrasts with previous cross-country analyses, which have shown a positive association (Di Tella et al. 2003; Clark and Senik 2011). In the literature of happiness, this phenomenon is often called the growth paradox (Graham 2009). This enigma may http://dx.doi.org/10.18196/jgp.2016.0021 26-49 44 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ JOURNAL OF GOVERNEMENT & POLITICS be explained by the fact that districts with higher GDP tend to expe- rience higher levels of economic growth. Recent studies on Indone- sian economic development indicate that despite high growth rates between 2005 and 2011, economic inequality has deepened (Yusuf et al. 2014). The Asian Development Bank (2012) and Euromonitor (2012) report that Indonesia has experienced the highest increases in income inequality levels in the world in the past few years (rising from a Gini coefficient of 0.33 in 2005 to 0.47 in 2011); districts with the highest GDPs also have the highest levels of economic inequality (World Bank 2008). Based on the National Survey Data 1993-2013, Yusuf et al (2014) show the rise in inequality was pre- dominantly visible in the period after the 1999 crisis, or era of po- litical reform and democratisation. This increase of inequality in richer districts may explain why citizens living in richer districts are less happy. CONCLUSION In Indonesia, the quality of district administration seems to be more important for citizen wellbeing than the provision of local democracy and the freedoms it implies. This study found that citi- zens report being happier when their district authorities prove them- selves more capable of providing better public services and promot- ing community social capital. Providing better policies and services that can ensure adherence to basic norms of equity and fairness, as well as promoting social support and collective culture, are essential to increasing citizen happiness in the country. The importance of the role of districts in improving citizen happiness across archipelago is clearly paramount and should not be underestimated by govern- ment and policy makers. The findings have an important implication both for theory link- ing democracy and citizen happiness and local democracy practice in developing countries. The null association between democracy http://dx.doi.org/10.18196/jgp.2016.0021 26-49 45 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ Vol. 7 No. 1 February 2016 and citizen happiness in Indonesia is in contrast with findings in the Western Europe and United States which show strong relation- ship between democracy and happiness (Frey and Stutzer, 2002; Lane, 2009). This contrasting findings imply that theory linking democracy and citizen happiness in developing countries may have different pathways. In the context of developing countries, local democracy will benefit for citizen happiness if political participa- tion and freedom can be translated into better public services and more cohesive and safer community. From a policy perspective there- fore, politicians and government should notice that the ultimate goal of democracy is not enlarging political participation and free- dom but achieving citizen wellbeing or happiness through provid- ing better policies and services as well as through promoting social support and collective culture. REFERENCES Alesina, A, Devleeschauwer, A, Easterly, W, Kurlat, S, & Wacziarg, R. (2003). Frac- tionalization. Journal of Economic Growth, 8, 155-194. Argyle, M. (2001). The psychology of happiness. East Sussex: Routledge. Aspinall, E. & Mietzner, M. (2010). Problems of democratisation in Indonesia: Elec- tion, institutions, and society. Singapore: ISEAS. Ballas, D. & Tranmer, M. (2010). Happy people or happy places? A multilevel mod- elling approach to the analysis of happiness and wellbeing. International Re- gional Science Review, 1(1),1-33. Barron, P, Kaiser, K, & Pradhan, M. (2009). Understanding Variations in Local Con- flict: Evidence and Implications from Indonesia, World Development, 37(3), 698-713. Beard, V. (2005). Individual determinants of participation in community develop- ment in Indonesia. Environment and Planning C: Government and Policy, 23(1), 21-39. Beard, V. (2007). Household contributions to community development in Indone- sia, World Development, 35(4), 607-625. Bjørnskov, C., Dreher, A., & Fischer, J. (2008). On decentralization and life satisfac- tion. Economics Letters, 99(1),147-151. Bjørnskov, C., Dreher, A., & Fischer, J. (2010). Formal institutions and subjective wellbeing, Revisiting the cross-country evidence. European Journal of Political Economy, 26(1),419- 430. http://dx.doi.org/10.18196/jgp.2016.0021 26-49 46 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ JOURNAL OF GOVERNEMENT & POLITICS Bjørnskov, C. (2003). The happy view: Cross-country evidence on social capital and life satisfaction. Kyklos, 56(1),3-16. Blanchflower D. & Oswald, A. (2004). Wellbeing over time in Britain and the USA. Journal of Public Economics, 88(1),1359-1386. Blanchflower, D. & Oswald, A. Is well-being U-Shaped over the life cycle? Social Science & Medicine, 66(8), 1733-1749. Clark, A. (2003). Unemployment as a social norm: Psychological evidence from panel data. Journal of Labor Economics, 21(2),323-351. Clark, A. & Oswald, A. (1994). Unhappiness and unemployment. The Economic Journal, 104(424),648-659. Clark, A. & Senik, C. (2011). Will GDP growth increase happiness in developing countries? IZA Discussion Paper No. 5595. Diener, E. (1984). Subjective wellbeing. Psychological bulletin, 95, 542-575. Diener, E., Suh, E.M., Lucas, R.E., & Smith, H.L. (1999). Subjective wellbeing. Psy- chological bulletin, 125(2), 276-302. Diener, E. & Biswar-Diener, R. (2008). Happiness: Unlocking the mysteries of psy- chological wealth. Malden: Blackwell Publishing. Diener, E., Sandvik, E., Seidlitz, L., & Diener, M. (1993). The relationship between income and subjective wellbeing: relative of absolute? Social Indicators Research, 28(1),195-23. Di Tella, R. & MacCulloch, R. (2005). Partisan social happiness. Review of Economic Studies, 72(2),367-393. Di Tella, R., MacCulloch, R., & Oswald, A. (2003). The macroeconomics of happi- ness. Review of Economics and Statistics, 85(4),793-809. Easterlin, R. (1974). Does economic growth improve the human alot? In David, P. and Reder, M. (Eds.), Nations and households in economic growth. New York: Academic Press. Euromonitor (2012). Income inequality rising across the globe. Paris: Euromonitor International. Fearon, JD. (2003). Ethnic and cultural diversity by country. Journal of Economic Growth, 8, 195-222. Frey, B. S. & Stutzer, A. (2002). Happiness and Economics: How the economy and institutions affect human wellbeing. Princeton: Princeton University Press. Graham, C. (2009). Happiness around the world. Oxford: Oxford University Press. Grootaert, C. (1999). Social capital, household welfare, and poverty in Indonesia. World Bank Policy Research Working Paper Number 2148. Washington, DC: World Bank. Helliwell, J. (2003). Wellbeing and social capital: Does suicide pose a puzzle? Paper presented at the Conference on Wellbeing and Social Capital, Harvard 7-9 No- vember 2003. Helliwell, J. (2006). Wellbeing, social capital and public policy: What’s new? The http://dx.doi.org/10.18196/jgp.2016.0021 26-49 47 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ Vol. 7 No. 1 February 2016 Economic Journal, 116(510),C34-C45. Ipsos. (2011). Tracking Global Happiness. New York: Ipsos. Jorgensen, D. (2002). Did we lose the war on poverty? In Jorgensen, D. (Eds.), Econometrics volume 3: Economic growth in the information age. Cambridge, Massachusetts: The MIT Press. Kaiser, K. & Hofman, B. (2003). Decentralisation, democratic transition, local gover- nance in Indonesia. World Bank: Jakarta. Krueger, A. & Schkade, D. (2008). The reliability of subjective wellbeing measure. Journal of Public Economics, 92(8),1-16. Lane, R. (2000). The Loss of Happiness in Market Democracies. New Heaven: Yale University Press. Layard, R. (2005). Happiness: Lessons from a new science. New York: The Penguin Press. Lessman, C. (2008). Fiscal decentralization and regional disparity: Evidence from cross-section and panel data. Environment and Planning A., 41(10),2455-2473. Lim, C. & Putnam, R. (2010). Religion, social networks, and life satisfaction. Ameri- can Sociological Review, 75(6),914-933. Litvack, J., Ahmad, J., & Bird, R. (1998). Rethinking Decentralization in Developing Countries. Washington, D.C. : Sector Studies Series, Poverty Reduction and Eco- nomic Management, World Bank. Miller, D., Schiffer, R., Lam, S. & Rosenberg, R. (2006). Social capital and health in Indonesia. World Development, 34(6), 1084-1098. Mookerjee, R and Beron, K. (2005). Gender, religion and happiness. The Journal of Socio-Economics, 34, 674–685 National Economic Foundation (2012). Happy Planet Index 2012. NEF. National Election Committee (2004). Indonesian National and Local Election data- base, National Election Commitee, Jakarta. Nordholt, H. & Van-Klinken, G. (2007). Introduction. In Nordholt, H. and Van- Klinken, G. (Eds.). Renegotiating Boundaries. Local politics in post-Suharto In- donesia. Leiden: KITLV Press. Nordholt, H. & Van-Klinken, G. (2005). New actors, new identities: Post-Suharto etnic violence in Indonesia. In Anwar, D.F., Houvier, H, Smith, G and Tol, R. (Eds.). Violent internal conflicts in Asia Pacific: Histories, political economies and policies, Jakarta: KITLV-Jakarta, Oswald, A. & Wu, S. (2011). Wellbeing across America. The review of economics and statistics, 93(4),1118-1134. Pisani, E. (2014). Indonesia: Exploring improbable nation. WW Norton and Com- pany, New York Rabe-Hesketh, S. & Skrondal, A. (2012). Multilevel and longitudinal modeling us- ing Stata. Texas: Stata Press. Radcliff, B. (2001). Politics, markets and life satisfaction: The political economy of http://dx.doi.org/10.18196/jgp.2016.0021 26-49 48 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ JOURNAL OF GOVERNEMENT & POLITICS human happiness. American Political Science Review, 95(4),939-956. Resosudarmo, B. & Jotzo, F. (2009). Working with nature against poverty: Devel- opment, resources, and the environment in Eastern Indonesia. Singapore: ISEAS. Rice, N. &d Jones, A. (1997). Multilevel models and health economics. Health Eco- nomics, 6(6),561-75. Ricklefs, MC. (2001). A History of Modern Indonesia since C. 1200. Stanford Uni- versity Press. Shiffman, J. (2002). The construction of community participation: village family planning groups and the Indonesian state. Social Science and Medicine, 54(8), 1199-1214. Sohn, Kitae. (2013). Source of happiness in Indonesia. Singapore Economic Re- view, 58(2), 1350014. Snijders, T. & Bosker, R. (1999). Multilevel Analysis: An introduction to basic and advanced multilevel modeling. London:Sage. Strauss, J., Beegle, K., Dwiyanto, A., Herawati, Y., Pattinasarany, D., Satriawan, E., Sikoki, B., Sukamdi, & Witoelar, F. (2004). Indonesian living standards: Before and after the financial crisis. Santa Monica: RAND Corporation. Stock, W.A., Okun, M.A., Haring, M.J., & Witter, R.A. (1982). Reporting reliability: A case study of life satisfaction research. Paper presented at the 90th Annual Convention of the American Psychological Association. Washington, DC. Subramanian, S., Kim, D., & Kawachi, I. (2002). Social trust and self-rated health in US communities: A multilevel analysis. Journal of Urban Health, 79(4),s21-s34. Sujarwoto, S & Tampubolon, G. (2013). Mother’s social capital and child health in Indonesia. Social Science & Medicine, 91, 1-9. Sujarwoto, S & Tampubolon, G. (2014). Decentralisation and citizen happiness: a multilevel analysis of self-rated happiness in Indonesia. Journal of Happiness Studies, 15(2). Suryadinata, L, Arifin, EN, & Ananta, A. (2003). Indonesia’s population: ethnicity and religion in a changing political landscape. ISEAS: Singapore. The Indonesian Ministry of Finance (2008). Sistem Informasi Keuangan Daerah di Indonesia. Jakarta: The Indonesian Ministry of Finance. The Ministry of Home Affair (2007). Indonesian National and Local Election data- base, The Ministry of Home Affair, Jakarta. Thomas, D., Witoelar, F., Frankenberg, E., Sikoki, B., Strauss, J., Sumantri, C. & Suriastini, W. (2012). Cutting the costs of attrition: results from the Indonesia Family Life Survey. Journal of Development Economics, 98(1), 108-123. Tselios, V., Rodriguez-Pose, A., Pike, A., Tomaney, J., & Torrisi, G. (2011). Income inequality, regional development, and decentralization in Western Europe. Mimeo. Veenhoven, A. (1984). Conditions of happiness. Boston and Lancaster: Reidel, Dordrecht. http://dx.doi.org/10.18196/jgp.2016.0021 26-49 49 ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ Vol. 7 No. 1 February 2016 Veenhoven, R. (2000). Freedom and happiness: A comparative study in forty-four nations in the early 1990s. In Diener, E. and Suh, E. (Eds.), Culture and subjec- tive wellbeing. Cambridge MA/London: The MIT Press. Welsch, H. (2008). The social cost of civil conflict: Evidence from survey of happi- ness. Kyklos, 61(2), 320-340. World Bank (2003). Decentralizing Indonesia. Jakarta: World Bank. World Bank (2006). Making services work for the poor: Nine case studies from Indonesia. INDOPOV: World Bank. World Bank (2008). Spending for development: Making the most of Indonesia’s new opportunities. Washington DC: World Bank. Yusuf, AA, Sumnerb, A, & Ruma, IA. (2014). Twenty Years of Expenditure Inequality in Indonesia, 1993–2013, Bulletin of Indonesian Economic Studies, 50,(2), 243– 54. http://dx.doi.org/10.18196/jgp.2016.0021 26-49 A A History of Modern Indonesia Ai Weiwei Amporn Tamronglak athumthani University aulus Wirotomo autonomy B Badan Layanan Umum Bambang Wahyu Sudarmadji Bangkokthonburi Universit Banten Beijing Occupy Bekasi Biswar-Diener Blacksburg Manifesto Perspective Brian Smith Bulletin of Indonesian Economic Studies C Cambridge Camilo Valderrama Canada Censorship Central Java Charoen Pokphand Group Chiangmai University China Chulalongkorn University Colonel Thanom Kittikajorn D democracy Department of Cooperatives, Industry, and Trade Developing Countries Diah Ayu Ardiyanti E Economic Studies Edi Suharto Egypt Electricity Generating Authority of Thailand F Fareastern University Frederick W. Taylor G GDP Gemolong GFMIS GLLAMMs Government Financial Management Information System Growth from Knowledge H H.J. Krysmanski Higher Education Hong Kong I IFLS Indah Surya Wardhani Indonesia Indonesian Village Potential Census International Monetary Fund intra-class correlation ITU J Jakarta Jakarta Metropolitan Region Japan Jasmine Jin-Wook Choi JMR Journal of Business Ethics Journal of Development Economics Journal of Economic Perspectives K Kalipaten Kartasasmita Kasetsart University Khon Kaen University, King Rama V Krirk University L Laohavichien Lippo Karawaci local government Local Power M macroeconomics Mahasarakam University Mahidol University Malaysia Massachusetts Max Weber, Ministry of Communication and Information Ministry of Home Affair monetary power complex Multilevel Mung N National Election Commitee National Election Committee National Reform Council Nawaporn Roengsakul New Public Management New York NGO Nielsen. Richard P North America North Bangkok University, North-Chiangmai University NPM NPO Nurmandi O Occupy Wall Street OECD Office of The Civil Service P Pathumthani University Payakumbuh Pince of Songkla University Pitsanulok University PKL PODES Proposal of Power Structure Reform PSAs Psychological bulletin Public Administration Public Sector Development Public Sector Restructuring Act R Rachbini Rajabhat Universities Ramkhamhaeng University Royal Decree on Good Governance S Sataporn Roengtam Self-rated Happiness Siam Cement Group Square Creative Merchants Sri Lestari Munajati Srinakharinwirot University Street Vendors Sukhothai Thammathirat Syamsir Syarif Makmur T Tangerang Thailand Thammasat University Tiananamen U ublic Administration United States Urban Theories and Urban Life V Van Dijk Van-Klinken, G W Wardoyo Weibo West Sumatra Western countries Who Rules America William Dommhoff William G. Domhoff World Bank Z Zainuddin Muda Z. Monggilo Zengcheng