Sources of Public Trust in Government: East Asian Evidence. Chong-Min Park Korea University & Jung-Hyun Bae Korea University

Transcription

1 Sources of Public Trust in Government: East Asian Evidence Chong-Min Park Korea University & Jung-Hyun Bae Korea University Prepared for delivery at the IIAS Study Group on Trust and Public Attitudes Workshop on Trust in Public Administration in Seoul, Korea on December 11-12, 2012

2 Public distrust in institutions of government becomes increasingly a global phenomenon (Norris 1999). In trilateral democracies a decline in trust in government has long been regarded as one of the most notable political phenomena (Pharr and Putnam 2000; Dalton 2004). A diversity of public opinion surveys repeatedly show that public confidence in government has declined and remains low. However, such a lack of trust in government is not exclusive to advanced Western democracies. Recent public opinion data reveal that third-wave new democracies in East Asia are no exceptions (Park and Chang 2011). Some of them display alarmingly low levels of public trust. Since trust in government contribute to citizen compliance and the viability of the political system (Levi 1998), low levels of trust in government pose a serious challenge for effective governance, especially in new democracies. There has been so much scholarly effort to understand why citizens come to lose their faith in government, especially in advanced Western democracies (Nye at al. 1997; Pharr and Putnam 2000; Hibbing and Theiss-Morse 2001b). In this paper we attempt to address the same question in the context of East Asia. There exist a variety of scholarly attempts to account for a decline in trust in government. Of them, we focus on government performance-based accounts. There exist three such accounts, each of which emphasizes different dimensions of government performance. The process accounts highlight how the government conducts its business. The output accounts emphasize what the government decides. The outcome accounts stress what the government delivers. By utilizing public opinion data drawn from the Asian Barometer Surveys, we test and compare these three accounts in the context of East Asia. The paper is organized into six sections. The first section reviews three performance-based accounts of trust and confidence in government. The second section develops our analytic framework based on prior theory and research. The third section deals with the data and measures of key concepts. The fourth section examines the levels of trust in government in the sample countries. The fifth section analyzes the influence of each dimension of government performance on trust in government. The final section highlights key findings and explores their implications. Prior Theory and Research Trust in government As King (2000) points out, the study of trust in government should begin with understanding of two concepts: trust or confidence on the one hand and government on the other. The meaning of trust is abstract and complex. Barber (1983) provides two specific meanings of trust. The first is related to the expectation of technically competent role performance. The second concerns expectations of fiduciary obligation and responsibility. In this conception, trust is government is based on popular beliefs about the government s competence and sense of fiduciary responsibility. Similarly, Hardin (2000) argues that trust has two distinct elements, that is, competence to perform what one is trusted to do and motivation to perform. In this conception, trust in government reflects evaluation of competence and moral commitment of public officials. Hence, distrust in government may be viewed as resulting from a gap between expectations and performance (Citrin and Luks 2001). The meaning of government is not simple because the term is used differently in different contexts. However, it is normally regarded as consisting of three sets of institutions, namely the legislature, the executive and the judiciary. In different contexts, the government sometimes 1

3 means just the executive. Hence, trust in government may mean more broadly trust in three branches of government or more narrowly trust in the executive. Since we are interested in the influence of policy performance on trust in government, we primarily focus on two policymaking branches of government, the national government (or the executive) and parliament. Trust in government is one type of orientations to the political system. As an attitude object the political system is a complex concept. Easton (1965) distinguishes between three levels of political system: political community, regime and authorities. The political community refers to a group of persons bound together by a political division of labor. The regime refers to the authority structure and its principles and values. The authorities refer to the present incumbents of authority roles. Prior research on citizen orientations to the political system has built on this conceptual distinction. Following this line of research, we distinguish trust in institutions of government from trust in incumbent authorities, although both are related. The former reflects general affective orientations or diffuse support which should be distinguished from evaluation of incumbent performance or specific support (Easton 1975; Dalton 2004). In this study we consider public trust or confidence to be directed toward institutions of government, not incumbents. Of the three branches of government we focus on two political ones - the executive and the legislature. We use confidence in government and trust in government interchangeably. 1 Performance-based explanations Political scientists have long searched for the determinants of a decline in trust in institutions of democratic government. Among more recent attempts, Putnam and his colleagues (2000) develop a general framework for analyzing the causes of the decline in confidence in government. They argue that confidence in democratic institutions is a function of the information to which citizens are exposed, the criteria by which the public evaluates government and politics, and the actual performance of those institutions. They see a decline in the public trust as reflecting a change in any of these variables. Furthermore, given that perceived government performance matters for trust in government, they offer two possible explanations for citizen evaluations of government performance. One has to do with the capacity of political agents to act on citizens interests and desires and the other the fidelity with which political agents act on citizens interests and desires. Newton and Norris (2000) develop three explanations of confidence in government and assess them empirically. First, social-psychological explanations consider trust a character trait of confidence in government. Second, social and cultural explanations hold that individual life situations and experiences, especially participation in voluntary associations, generate trust and norms of reciprocity, which in turn helps create effective political institutions in which people come to have confidence. Lastly, institutional performance explanations view government performance as the key to public confidence in government. They find little evidence for the first and the second explanations. In contrast, they provide substantial support for institutional explanations which link government performance to confidence in institutions. Alesina and Wacziarg (2000) propose three hypotheses concerning sources of distrust in government based on an economic model of voter dissatisfaction. The policy quality hypothesis 1 Hardin (2000) argues that Because there is no evidence that people actually trust government in the ways they do trust some persons, and because it is very difficult to see how they sensibly could, it makes sense analytically to speak of declining confidence in government rather than declining trust. 2

4 posits that citizen dissatisfaction originates from the government s incompetence and the lack of honesty among elected officials. The heterogeneity hypothesis posits that citizen dissatisfaction originates from greater polarization of preferences. The representation hypothesis holds that citizen dissatisfaction originates from the deviation of policies from the electorate s median. The first suggests that citizen distrust in government reflects poor policy outcomes as well as a lack of good governance (della Porta 2000). The last two hypotheses suggest that citizen distrust in government reflects the policy disagreement between citizens and the party in power. In accounting for a decline in trust in government in the United States, Orren (1997) identifies two types of influences. One includes two long-term factors that shape people s views of government: the traditional antipathy toward government and authority-challenging postmaterial values. The other includes four immediate, short-term factors: the public s appraisal of government performance, the public s ideological views on certain policies, and the public s assessment of the ethics and integrity of the people and processes of government, and the denunciation of government by political leaders and the press. Of these short-term influences, the first is related to policy outcomes. The second pertains to the policy agreement between citizens and the government. The third has to do with some elements of good governance. The last pertains to a mood of government bashing. Most recently, Norris (2011) develops two performance-based explanations based on a rational choice perspective. The process performance accounts emphasize standards of democratic governance such as accountability, participation, and rule of law as determinants of democratic satisfaction. From this perspective, citizens have confidence in government when government performs well against standards of democratic governance. On the other hand, the policy performance accounts emphasize policy outcomes and outputs as determinants of democratic satisfaction. From this perspective, citizens have confidence in government when government delivers goods and services demanded by citizens. Overall, these performance-based accounts emphasize different dimensions of policy performance - outcome, output or process. From these accounts, citizens come to lose faith in government when government fails to meet standards of good governance, when the government makes policies counter to their preferences or when the government fails to deliver desired outcomes and benefits. Analytic Framework Considering prior theory and research, we propose an individual-level performance-based framework analyzing confidence in policymaking institutions of government across East Asia. Our model combines three performance-based accounts. First, it considers policy outcomes as one source of trust in government. The outcome accounts stress the delivery of desired policy benefits. Economic well-being, national or personal, has been considered one of the most important outcomes. From this account, a decline of trust in government reflects the failure of government to solve economic problems such as employment and inflation. However, desired policy benefits are not necessarily confined to wealth or prosperity. They include other areas of government activity such as domestic security, socio-economic security and equality, and environmental protection (Roller 2005). The government s inability to handle these problems may also undermine public confidence in government. In this study we include three areas of government activity economic well-being, public safety and anti-poverty. They by no means encompass all desired policy outcomes but still constitute major ones. We regard conditions associated with these areas as policy outcomes or effects, although it is not obvious whether 3

5 economic and social conditions are the consequences of government action (Bok 1997). The outcome-based explanation primarily emphasizes technical competent role performance of government. There exists evidence that what matters are not absolute levels of outcomes but relative levels of outcomes (Tyler 2001). For instance, explanations based on adaptation theories emphasize changing expectations rather than historical records as criteria of comparison. Other accounts based on relative deprivation theory link dissatisfaction to violations of the sense of deservedness. Considering different criteria of comparison, we distinguish between current economic assessments from retrospective ones. The former means evaluation of the current economy as compared to changing expectations. In contrast, the latter refers to evaluation of the current economy as compared to past one. Second, our model considers policy outputs as another source of trust in government. 2 The output accounts emphasize the policy agreement between citizens and government. According to this perspective, citizens lose confidence in government when government adopts policies counter to their preferences. In democracies where policy direction can be altered through elections, policy disagreements may not persist. However, citizens come to lose confidence in institutions of government when successive governments run counter to their policy preferences. Orren (1997) maintains that some lost faith because they felt that government was doing too good a job in advancing the wrong mission. Similarly, Tyler and his colleagues (1985) argue that citizen dissatisfaction with government reflects not only the failure of government to solve social and economic problems, but also a lack of policy congruence between citizens and government. In this account, a decline in trust in government reflects a widening policy disagreement between citizens and government. The more polarized citizen policy preferences become, the more there are conflicts between winners and losers. This increased conflict of preferences would result in increased dissatisfaction with government (Alesina and Waczuarg 2000). In this account, political winners are more likely to have trust in government than political losers (Anderson et al. 2005). Hence, excessive partisanship may be one of the sources of a decline in public trust government. As government s policy direction and choices diverge from citizens desired outputs, there would be less public confidence in government. It is an empirical question whether government s actual policies are congruent with the preferences of citizens. Although they may diverge from its promises, citizens who voted for the party in power are expected to agree more with government s actual policies than those who voted for the opposition. Hence, we hypothesize that those voting for the party in power are likely to have trust in government than those voting for the opposition. Third, our model considers policy processes as another source of public confidence in government (Hibbing and Theiss-Morse 2001a; Tyler 2001). As discussed above, the process accounts emphasize the standards of good governance such as impartiality, equality before the law, the rule of law, transparency, fairness and so forth. A lack of such standards in the governing process causes citizens to lose faith in institutions of government. From this perspective, the manner in which government authority is exercised shapes citizen attitudes about institutions of government. No arbitrary rule or the rule of law is one of core universal norms against which the exercise of government authority is evaluated. No arbitrary rule or the rule of law means that all 2 While policy outcomes pertain to policy impacts, policy outputs refer to actual policies. Hence, while outcomes-based accounts look at policy benefits, outputs-based accounts look at policy decisions. 4

6 citizens are equal before the law and that official corruption is detected, punished and controlled, and that public officials are impartial in exercising their authority and treat people with equal concern and respect (Rothstein and Teorell 2008). Although mistrust in government may facilitate corruption, as della Porta (2000) argues, political corruption worsens government performance, thus reducing trust in the government s capacity to address citizens demands. In this study we consider four standards of good governance associated with no arbitrary rule or the rule of law - impartiality, law-abidingness, transparency, and lack of corruption. These standards of governance ensure the impersonal conduct of public affairs and remove the problems of particularism and clientelism in the public sector. When the outcomes accounts emphasize the technically competent role performance of government, the process accounts highlight the integrity and moral commitment of government as a fiduciary. Finally, considering the alleged harmful effects of mass media, we include exposure to media news about politics and government as an alternative source of distrust in government. In brief, our model emphasizes citizen evaluation of government performance as the immediate, short-term source of satisfaction government. The outcomes-based explanation posits that when government fails to meet citizen demands of prosperity, public safety and anti-poverty, citizens consider government incompetent and thus lose trust in government. The outputs-based explanation holds that when government adopts policies counter to voter preferences, those voting for the party in power consider government untrustworthy and thus lose confidence in government. The process-based explanation holds that when government fails to conduct business according to standards of good governance, citizens consider government arbitrary and unethical and thus lose trust in government. Do every dimension of government performance matters for trust in government? Are they equally important? If not, which dimension matters most or least? In the ensuing sections, we compare these performance-based accounts in the context of East Asia. Data and Measures Data The data used here are drawn from the third round of the Asian Barometer Survey (ABS III hereafter) conducted between 2010 and We selected nine of eleven countries surveyed, excepting Cambodia, China and Vietnam: Japan in 2011 (N=1,880), South Korea in 2011 (N=1,207), Mongolia in 2010 (N=1,210), the Philippines in 2010 (N=1,200), Taiwan in 2010 (N=1,592), Thailand in 2010 (N=1,512), Indonesia in 2011 (N=1,550), Malaysia in 2012 (N=1,214), and Singapore in 2010 (N=1,000). Since we use cross-sectional survey data, it remains challenging to disentangle the direction of causality. Although we assume that perceptions of government performance affect trust in government, it is possible that trust in government shapes perceptions of government performance. Dependent variable Our dependent variable is confidence or trust in government. To measure it we selected two questions: one asked respondents how much trust they had in the national government and the other, in parliament. Although government means all three branches of government, we focus on the two because they constitute key national institutions responsible for policymaking. Four response categories were provided: a great deal of trust, quite a lot of trust, not very much trust, and none at all. By combining responses to both questions, we construct 7-point index of trust or confidence in government. 5

7 Explanatory variables We have three clusters of key explanatory variables. The first cluster pertains to variables associated with policy outcomes. We consider economic outcomes as well as other outcomes associated with public safety and anti-poverty. Since our variables measure evaluation of the conditions of policy areas, they do not necessarily capture policy effects. The first two measures capture assessments of the nation s economy. One asked How would you rate the overall economic condition of our country? Five response categories were provided: very good, good, neither good nor bad, bad, and very bad. Responses to the question reflect current assessments. The other asked How would you describe the change in the economic condition of our country over the last few years? Five response categories were provided: much better, a little better, about the same, a little worse, and much worse. Responses to the question reflect retrospective assessments. Although both measures focus on the nation s economy, they use different criteria of comparison. The first involves current expectations whereas the second, past records. The second two measures capture assessments of the respondent s household economy. One asked How do you rate the economic situation of your own family? Five response categories were provided: very good, good, neither good nor bad, bad, and very bad. Responses to the question reflect current assessments. The other asked How would you compare the current economic condition of your family with what it was a few years ago? Five response categories were provided: much better, a little better, about the same, a little worse, and much worse. Responses to the question reflect retrospective assessments. As in the assessments of the national economy, they use different criteria of comparison. The first involves current expectations whereas the second, past records. We have two additional measures associated with other policy outcomes. 3 First, to measure evaluation of public safety conditions we asked How safe is living in this city, town or village? Four response options were offered: very safe, safe, unsafe and very unsafe. Second, to measure evaluation of anti-poverty conditions we asked respondents whether they agree or disagree with the statement People have basic necessities like food, clothes and shelter. The second cluster pertains to variables associated with policy outputs. Lacking better measures of policy agreement, we employed voting for the party in power as a proxy. To measure it we asked respondents which parties or candidates they voted for in the most recent national elections. 4 Those who voted for the winning camp were viewed as agreeing with policy outputs whereas those who voted for the opposition, as disagreeing with them. The third cluster pertains to variables associated with policy processes and includes four variables. We consider four aspects of rule-based governance: law-abidingness, transparency, lack of corruption and impartiality. All of them reflect the rule of law, one of the hallmarks of good governance. They go beyond evaluation of particular decisions or outcomes. They reflect evaluations of how government authority is exercised. First, to measure law-abidingness we asked How often do you think government leaders break the law or abuse their power? Four response categories were offered: always, most of the 3 For a variety of outcome measures covering a wide range of government performance, see Bok (1997) and Roller (2005). 4 In Singapore where this question was not asked, we used a question asking respondents Which party if any do you feel closet to? 6

8 time, sometimes, and rarely. Second, to measure transparency we asked How often do government officials withhold important information from the public view? The same four response options as in law-abidingness were provided. Third, to measure lack of corruption we asked How widespread do you think corruption and bribe-taking are in the national government? Four response options were provided: hardly anyone is involved, not a lot of officials are corrupt, most officials are corrupt, and almost everyone is corrupt. Lastly, to measure impartiality or fairness we asked whether respondents agreed or disagreed with the statement Rich and poor are treated equally by the government. 5 Control variables Although some emphasize the positive role of modern media in promoting civic participation, others point out the corrosive effects of media on politics by arguing that media s negative coverage of politics and government tends to encourage viewers to be cynical of political institutions (Norris 2000). Hence, we include media exposure as a control variable. 6 We introduce four standard demographic controls gender, age, educational attainment, and household income. To measure household income we used subjective income instead of objective income to facilitate cross-national comparison. 7 Levels of Public Trust In this section we briefly examine how much confidence East Asians have in their nation s institutions of government. As discussed earlier, we focus on two core policymaking institutions: the national government and parliament. As shown in Table 1, liberal democracies such as Japan, South Korea and Taiwan displayed low levels of confidence in institutions of government. For instance, for the last decade public trust in the national government ranged, on average, from a low of 18 percent (in Japan) to a high of 36 percent (in Taiwan). Public trust in the legislative branch of government was even lower, ranging, on average, from a low of 11 percent (in South Korea) to a high of 19 percent (in Taiwan). Our electoral democracies may be divided into two groups: in Thailand (for the last decade) and Indonesia (for the last five years), on average, more than half of their publics expressed some degree of trust in both institutions (from a low of 54 percent to a high of 61 percent). By contrast, for the last decade in Mongolia and the Philippines, on average, just half or less than half displayed confidence in the same institutions (from a low of 42 percent to a high of 50 percent). Overall, electoral democracies obviously enjoyed greater public confidence than liberal democracies. 5 Impartiality requires that the government treats people equally on the basis of clear rules. Hence, this measure reflects government impartiality or procedural non-discrimination regardless of social classes. However, it may also mean distributive fairness or social equality. 6 To measure it we asked How often do you follow news about politics and government? Five response categories were provided: everyday, several times a week, one or twice a week, not even once a week, and practically never. 7 To measure it we asked Does the total income of your household allow you to satisfactorily cover your needs? For response categories were provided: Our income covers the needs well, we can save, Our income covers the needs all right, without much difficulty, Our income does not cover the needs, there are difficulties, and Our income does not cover the needs, there are great difficulties. 7

9 Our electoral autocracies or competitive authoritarian regimes apparently fared best in the eyes of their publics. 8 For the last five years a large majority of citizens in Malaysia and Singapore (ranging, on average, from a low of 67 percent to a high of 88 percent) expressed some degree of trust in their nation s policymaking institutions. Notable is that for the period examined most sample countries but Malaysia experienced more or less a decline in public trust in government. Mongolia, one of the new democracies in East Asia, experienced the highest decline: public trust in the national government sharply dropped from 58 percent in 2002 to 35 percent in Similarly, public trust in parliament plummeted from 60 percent in 2002 to 28 percent in Noteworthy is that most of the sharp decline there occurred during the last few years. In contrast, the decline in trust appears to be gradual in Thailand and Indonesia. Although the decline in public trust in liberal democracies was gradual, public trust remained low throughout a decade. (Table 1 about here) Overall, low levels of public trust or its steady decline is not exclusive to advanced democracies. Evidence shows that citizens in East Asian democracies had low levels of confidence in institutions of government. Moreover, citizens in liberal democracies appear to be far more cynical than their counterparts in electoral democracies or even electoral autocracies. Perhaps since government propaganda and media censorship may distort public perceptions, citizens in non-democracies could be less cynical than their counterparts in democracies. Whatever the reasons may be, these findings challenge the thesis that democratic transition and consolidation strengthens public confidence in representative institutions. Perhaps what matters for confidence in government may be not the quality of democracy in general but the quality of government more specifically. Determinants of Public Trust In this section we analyze the influence of three dimensions of government performance on trust in institutions of government and assess three performance-based accounts in the context of East Asia. In this multivariate analysis we employed pooled cross-sectional data drawn from ABS III. Since the levels of public trust seem to vary depending upon regime types, we divide the sample countries into three groups and performed the analysis separately: liberal democracy (Japan, South Korea and Taiwan), electoral democracy (Indonesia, Mongolia, Philippines, and Thailand) and electoral autocracies (Malaysia and Singapore). Notable is that three liberal democracies and two electoral autocracies are more affluent than four electoral democracies. Liberal democracies Let us begin with liberal democracies (Japan, South Korea and Taiwan). As shown in the first column of Table 2, all three dimensions of performance - outcome, output and process - mattered. Notable is that current assessments of economic conditions, whether national or personal, had significant effects. In contrast, retrospective economic assessments had either negligible or no effects. This finding suggests that even if economic well-being today is perceived as better than a few years ago, people may not be satisfied because the current economic conditions remain short of their expectations. In contrast, when people think that the current economic conditions fared better than their expectations, they are more likely to have confidence in government. That 8 For the concept of competitive authoritarianism, see Levitsky and Way (2010). 8

10 national economic conditions mattered more than personal financial situations suggests that political cynicism reflects sociotropic economic judgment more than pocketbook one (Kinder and Kiewiet 1981). Second, the influence of other policy outcomes turned out to be mixed. Those who considered their city, town or village safe were more likely to have trust in government than those who considered it unsafe. In contrast, those who considered the basic needs satisfied were no more likely to have trust in government than those who considered them unsatisfied. In these advanced countries satisfactory government performance on anti-poverty played little role in increasing public confidence in government. Third, in our liberal democracies, the policy agreement had significant effects. Those who voted for the party in power were more likely to have confidence in government than those who voted for the opposition. This finding strengthens the outputs-based accounts that citizens come to lose faith in government when government adopts policies counter to citizen preferences. In our liberal democracies where elections are free and fair and peaceful transfer of power is institutionalized, citizens agreement with government s policy direction and choices encourages confidence in government. Fourth, all four standards of good governance had significant effects. Especially, the influence of impartiality, transparency and lack of corruption seemed pronounced. As discussed above, our measure of impartiality indicates equal treatment, which may reflect distributive as well as procedural fairness. Hence, all its effects may not reflect process effects. In contrast, lawabidingness, transparency and lack of corruption obviously constitute key elements of rule-based governance. The finding strengthens the argument that the rule of law is one of key ingredients in establishing viable and effective state institutions. Fifth, exposure to media news had negative effects. Those who more often followed news about politics and government were less likely to have trust in government. This finding seems to confirm the media malaise thesis. In liberal democracies where media freely cover government scandals, exposure to media news seemed to lower public confidence in institutions of government. Lastly, no standard demographic variables had effects. Men, the young, the educated and the poor were no less likely to have trust in government than women, the old, the uneducated and the affluent. This finding suggests that life circumstance associated with gender, age and socioeconomic status played little role in trust in government. Distrust in government seemed to cut across every segment of the population in these liberal democracies. The entire set of five clusters of variables accounted for only two-fifths (21%) of the variance in trust in government, indicating the limited explanatory power of the performance-based model. (Table 2 about here) Electoral democracies Next, we turn to electoral democracies (Mongolia, Indonesia, the Philippines and Thailand), as presented in the second column of Table 2. In our electoral democracies assessments of the nation s economy had pronounced effects. Notable is that current assessments mattered far more than retrospective assessments. As in liberal democracies, people seemed to adjust their expectations. Hence, even if the nation s economy improves than a few years ago, people s assessments would remain short of their changing expectations. In contrast, assessments of personal economic well-being had negligible or no effects. Second, other policy outcomes had significant effects. Those who viewed their neighborhood 9

11 as safe were more likely to have trust in government than those who viewed it as unsafe. Performance on anti-poverty had significant, albeit weaker, effects. Of policy outcomes prosperity and law and order appear to be important in these less affluent unstable democracies. Third, policy agreement had no effects. Those who voted for the party in power were no more likely to have confidence in government than those who voted for the opposition. Contrary to the outputs accounts, the finding suggests that citizens may not lose confidence in government even if government s actual policies are viewed as incongruent with their policy preferences. Political distrust seemed to cut across parties and ideological orientations. This finding may indicate that in these electoral democracies vote choices do not necessarily reflect policy preferences. Fourth, three of four standards of good governance had significant effects. Those who considered the government law-abiding, uncorrupt and transparent were more likely to have confidence in government. All these elements are associated with law-based governance, strengthening the argument that the rule of law is one of the key ingredients in establishing an effective modern state. Surprising is that impartiality had no significant effects. As discussed above, our impartiality measure reflects not only procedural fairness but also distributive justice. To the extent that people react to the latter, this finding may suggest that in our electoral democracies government performance on social equality played little role in producing confidence in government. Fifth, exposure to media news had no significant effects, although its direction seemed negative. Although the freedom of the press is limited in electoral democracies than liberal ones, this finding challenges the media malaise thesis. This finding suggests that the content of media matters. Lastly, only one of four demographic controls, education, had significant, albeit negative, effects. The more educated were less likely to have confidence in government than the less educated. Yet, the young and the affluent were no less likely to have trust in government than the old and the poor. This finding suggests that cognitive skills and worldviews help people to develop critical and challenging attitudes toward government. In general, distrust in government seemed to cut across most major segments of population. As in liberal democracies, the entire set of five clusters of variables accounted for only twofifths (22%) of the variance in public trust in government, indicating the limited explanatory power of our performance model. Electoral autocracies Finally, we turn to electoral autocracies (Malaysia and Singapore), as reported in the third column of Table 2. As in our electoral democracies, both current and retrospective assessments of the nation s economy had significant effects. As in both types of democracies, current assessments turned out to be more influential than retrospective assessments, strengthening the argument that levels of expectations constantly adjust to experiences. As in our electoral democracies, assessments of personal financial situations had no effects. What mattered to public trust in government in these non-democracies was national economic well-being, not personal economic well-being. Second, the influence of other policy outcomes was mixed. Unlike in both types of democracies, evaluations of public safety conditions had no effects. In contrast, as in less affluent democracies, evaluations of anti-poverty conditions had effects. More notable is that performance on anti-poverty had far greater effects, suggesting that meeting basic needs remains 10

12 a key ingredient in determining public trust in these apparently affluent countries. Third, the policy agreement had notable effects, strengthening the outputs-based explanation. Those who voted for or felt close to the party in power were far more likely to have confidence in government than those who voted for or felt close to the opposition. In these electoral autocracies the playing field remains skewed in favor of incumbents and institutions may not be distinguishable from incumbent leaders. Since policy preferences of the opposition have been excluded from political processes, partisanship may be closely related to trust in institutions of government. Fourth, as in liberal democracies, all four standards of good governance had significant effects. Of them two turned out to be more pronounced. They were impartiality and lack of corruption. Those who consider government treatment equal and public officials uncorrupt were more likely to have confidence in government. To the extent that impartiality reflects fair distribution of outcomes, satisfactory performance on social equality would encourage confidence in institutions of government. Fifth, contrary to the media malaise thesis, exposure to media news had significant, albeit positive, effects. Those who more often followed news about politics and government were slightly more likely to have confidence in government. Perhaps because of media censorship in these electoral autocracies the content of media news may exclude negative news about government and include pro-government propaganda. Hence, more exposure to media news would encourage rather than discourage trust in government. Lastly, only one of the four demographic controls, age, had significant, albeit small, effects. The young were less likely to have trust in government than the old. Trust in government seemed to cut across major segments of the population. Our performance model worked better in non-democracies than democracies. Nonetheless, the entire set of five clusters of variables accounts for one-third (32%) of the variance in public trust in government. Overall, regardless of regime types or levels of socioeconomic development East Asians reacted to the condition of their nation s economy. Moreover, current assessments were more influential than retrospective assessments, strengthening adaptation theories. In contrast, the influence of personal financial conditions seemed inconsistent. A notable influence was found only in liberal democracies where citizens also reacted to current assessments of private economic lives. The influence of other policy outcomes was not robust. Only in democracies, evaluation of public safety contributed to confidence in government. In contrast, only nondemocracies evaluation of anti-poverty had pronounced effects. The outputs-based explanation received mixed result. Most notable is that in electoral autocracies the policy agreement emerged as a strong predictor of trust in government. In contrast, the policy agreement had no independent role in electoral democracies. Regardless of regime types, most procedural elements of good governance contributed to confidence in government. The influence of lack of corruption was more or less strong across East Asia. The influence of transparency and law-abidingness was somewhat weaker but still consistent across the region. Although the influence of impartiality was not found in electoral democracies, it was somewhat stronger in the rest of the region. In brief, government performance mattered for public trust in government across East Asia, strengthening the performance based accounts. Yet, depending upon regime types, three aspects of government performance had varying effects on trust in government. Moreover, that the 11

13 government performance model worked better in electoral autocracies than democracies suggests that the former relies on performance than legitimacy. Summary and Conclusion A lack of trust in government poses a serious challenge for the viability of the political system because trust in government is a key to citizen compliance and effective governance. Hence, understanding sources of trust in government is increasingly important, particularly for new democracies. There has been much scholarly effort to understand sources of trust or distrust in government, especially advanced Western democracies. In contrast, there has been not much such effort in the context of East Asia. This paper attempts to fill the gap. Of a variety of explanations, we focused on policy performance-based explanations. We identified three dimensions of government performance, each of which emphasizes different aspects of policy performance. The first stresses policy outcomes, as measured by assessments of the nation s economy, personal economy, public safety and anti-poverty. The second emphasizes policy outputs (policy agreement), as measured by voting for the party in power. The last emphasizes policy processes, as measured by evaluation of impartiality, law-abidingness, transparency and lack of corruption. We estimated and compared the influence of each factor on trust in policymaking institutions of government. It was found that public trust in government was much lower in democracies than nondemocracies across East Asia. The publics in democracies were far more disillusioned with their political branches of government. This finding suggests that holding free and fair elections does not necessarily lead to greater public confidence in government. Moreover, it was found that trust in government was lower in liberal democracies than electoral democracies, which suggests that more civil liberties and political rights do not necessarily lead to higher public trust. All the findings suggest that not the quality of democracy (or how to constitute government power), but the quality of government (or how to exercise government power) may play a greater role in generating public trust. What matters for trust in government across East Asia? Is it whether government is capable of handling policy problems, whether government adopts policies favored by citizens, or whether government conducts its business in accordance with standards of good governance? A multivariate analysis reveals that policy outcomes, especially national economic performance influenced confidence in government. Perhaps the nation s current economic conditions may not be policy effects. Yet, the nation s economic conditions appeared to be seen as reflecting the competence of government. The finding suggests that when government is perceived as failing to handle effectively critical problems in society, (particularly, but not exclusively, economic problems), citizens come to lose faith in institutions of government. Critical problems facing society may vary. What is important is that government should be perceived as effective and competent in handling these critical problems. The influence of policy agreement appears to be mixed. Only in electoral democracies voting for the party in power played little role. Notable is that it had larger effects in electoral autocracies, suggesting that policy agreement rather than electoral contest itself matters to confidence in government. This finding suggests that when policy disagreements become divisive, there will be more widespread public distrust in government in these non-democracies. The standards of good governance had consistent effects. Regardless of regime types key aspects of rule-based governance had notable influence on trust in government. Arbitrary rule and a weak rule of law rendered people disillusioned with institutions of government. When 12

14 government is perceived as abusing or arbitrarily exercising its power, citizens come to lose faith in institutions of government. Overall, government performance shapes public confidence in government across East Asia. Trust in government largely reflects satisfaction with economic management and the procedural quality of governance. When people believe that the nation s economy is bad, they conclude that government is incompetent and thus untrustworthy. When people think that government is corrupt or law-breaking, they conclude that government is immoral and thus untrustworthy. Not only what government delivers but also how it conducts its business seems to have pronounced influence. When government is perceived as incompetent, self-serving and arbitrary, citizens come to lose faith in government. Perhaps the problems facing governments become too complex to solve. Citizens expectations of governments become increasingly polarized. As a result, policy outcomes and outputs may be limited in restoring confidence in government. In this context, the procedural quality of good governance is more important than ever (Hibbing and Theiss-Morse 2001c). Our finding suggests that one of the hallmarks of the modern state, the rule of law, is a key ingredient in determining public confidence in government. As a final note, we would like to mention the influence of media exposure. Despite some measurement problem, the influence of exposure to media news turned out to be inconsistent. It had negative effects in liberal democracies while positive effects in electoral autocracies. This seemingly contradictory finding may not weaken the thesis that media is responsible for declining public trust in government. In liberal democracies the press often denounces government leaders, whereas in electoral autocracies the press often praises or rarely denounces government leaders. Considering the content of media news, it is little wonder that exposure to media news had negative effects in liberal democracies while positive effects in electoral autocracies. The key is the content of media report. 13

15 References Alesina, Alberto and Romain Wacziarg The Economics of Civic Trust, in Susan J. Pharr and Robert D. Putnam, eds. Disaffected Democracies: What's Troubling the Trilateral Countries? Princeton: Princeton University Press. Anderson, Christopher J., Andre Blais, Shaun Bowler, Todd Donovan, and Ola Listhaug Losers Consent: Elections and Democratic Legitimacy. Oxford: Oxford University Press. Barber, Bernard The Logic and Limits of Trust. New Brunswick: Rutgers University Press. Bok, Derek Measuring the Performance of Government, in Joseph S. Nye, Philip D. Zelikow and David C. King, eds. Why People Don't Trust Government. Cambridge: Harvard University Press. Citrin, Jack and Samantha Luks Political Trust Revisited: Deja Vu All Over Again? in John R. Hibbing and Elizabeth Theiss-Morse, eds. What Is It about Government that Americans Dislike. Cambridge: Cambridge University Press. Dalton, Russell J Democratic Challenges, Democratic Choices: The Erosion of Political Support in Advanced Industrial Democracies. Oxford: Oxford University Press. della Porta, Donatella Social Capital, Beliefs in Government, and Political Corruption, in Susan J. Pharr and Robert D. Putnam, eds. Disaffected Democracies: What's Troubling the Trilateral Countries? Princeton: Princeton University Press. Easton, David A Systems Analysis of Political Life. New York: Wiley. Easton, David A Re-Assessment of the Concept of Political Support, British Journal of Political Science 5: Hardin, Russell The Public Trust, in Susan J. Pharr and Robert D. Putnam, eds. Disaffected Democracies: What's Troubling the Trilateral Countries? Princeton: Princeton University Press. Hibbing, John R. and Elizabeth Theiss-Morse. 2001a. Process Preferences and American Politics: What the People Want Government to Be, American Political Science Review 95: Hibbing, John R. and Elizabeth Theiss-Morse, eds. 2001b. What Is It about Government that Americans Dislike. Cambridge: Cambridge University Press. Hibbing, John R. and Elizabeth Theiss-Morse. 2001c. The Means Is the End, in John R. Hibbing and Elizabeth, eds. What Is It about Government that Americans Dislike. Cambridge: Cambridge University Press. Kinder, Donald R. and Roderick D. Kiewiet Sociotropic Politics: The American Case, British Journal of Political Science 11(2): King, Anthony Distrust of Government: Explaining American Exceptionalism, in Susan J. Pharr and Robert D. Putnam, eds. Disaffected Democracies: What's Troubling the Trilateral Countries? Princeton: Princeton University Press. Levi, Margaret A State of State, in Valerie Braithwaite and Margaret Levi, eds. Trust and Governance. New York: Russell Sage. Levitsky, Steven and Lucan A. Way. (2010). Competitive Authoritarianism: Hybrid Regimes After the Cold War. Cambridge: Cambridge University Press. Newton, Kenneth and Pippa Norris Confidence in Public Institutions: Faith, Culture, or Performance, in Susan J. Pharr and Robert D. Putnam, eds. Disaffected Democracies: What's Troubling the Trilateral Countries? Princeton: Princeton University Press. 14