Designing Democracies:

Institutional Arrangements and System Support.  

Pippa Norris,  John F. Kennedy School of Government Harvard University Cambridge MA 02138 T: 617 495 1475 F: 617 495 8696 E: Pippa_Norris@Harvard.edu  

For the John F. Kennedy School of Government Workshop on Confidence in Democratic Institutions: America in Comparative Perspective, 25-27th August 1997 Sheraton Hotel, Washington DC.

Synopsis

Ever since The Civic Culture (Almond and Verba 1963) we have developed a substantial literature on how citizens' support for the political regime varies cross-nationally. New work has recently reexamined trends in public opinion towards democratic governance in the United States (Nye, Zelikow, and King 1997), and Western Europe (Klingemann and Fuchs 1995), while renewed interest in the conditions of democratic consolidation has generated new research on these issues in Central and Eastern Europe (Mishler and Rose; Mishler and Rose 1995; Mishler and Rose 1994; White, Rose, and McAllister 1997; Miller, Hesli, and Reisinger 1996) as well as in Latin America (Booth and Seligson 1994; Lagos). We are now better equipped than ever before with survey data to map system support in emerging, transitional and established democracies yet, it can be argued, we have made less progress in understanding the causes of crossnational variations in public opinion. Theories have lagged behind data.

This paper focuses on explanations of 'system support', meaning positive attitudes towards the political regime, the constitutional order and its structures, at a diffuse level, rather than confidence in specific authorities or satisfaction with particular government policies (Easton 1975). Existing explanations of crossnational variations in system support are usually based on cultural theories of value change in modern society or political economy theories of government performance. In contrast, this paper argues that these approaches need to be supplemented by accounts of system support which emphasise the institutional context of citizens.

The thesis examined by this paper is that constitutional designs may be expected to structure the pattern of winners and losers from government, and therefore public support for the general regime. We therefore examine a series of issues: Do citizens living in nations with well-established political rights and civil liberties have more confidence in the political system, as we would expect, than those who have lived under interrupted democracies? Are people more supportive of the regime in nations with presidential or parliamentary executives, federal or unitary states, majoritarian or proportional electoral systems, and two-party or fragmented multiparty systems?

To examine these questions we draw on data from the World Values Surveys tapping system support in 43 nations in 1990-93. The results of the analysis confirm that institutions matter: winners express more confidence in the regime than losers and some institutional arrangements are more successful in generating system support than others. The conclusion considers the implications of this analysis for democratic theory, for constitutional debates, and for institutional engineering particularly in newer democracies.

Designing Democracies: Institutional Arrangements and System Support.  

Since the early 1960s a well-established literature has documented public opinion towards political regimes in Western nations, and more recently in emerging democracies. Survey after survey has reported substantial cross-national variations in the level of system support. In some established democracies cynicism about the political system is endemic, widespread and persistent, notably in Japan (Pharr 1997) and Italy (Morlino and Tarchi 1996; Morlino and Montero 1995), as well in certain transitional democracies such as the Ukraine, Bulgaria, and Mexico (Mishler and Rose; Lagos). There are some problems of strict comparability of survey items but nevertheless the available evidence since the early 1970s, in the Eurobarometer and other surveys, indicates that in contrast people seem consistently more trusting of politicians and civic institutions, more confident about government, and more satisfied with democracy in other West and North European nations including Denmark, Norway, Finland, West Germany, and the Netherlands (Klingemann and Fuchs 1995; Kaase and Newton 1995; Holmberg 1997; Budge and Newton 1997; Davis 1997; Dalton 1997). Data in the World Values Surveys, analyzed later in this paper, confirm these variations (see Figures 2 and 3).

Moreover these striking cross-national differences are not just in absolute levels of system support, but also in trends over time. America has shifted in recent decades from the more trusting towards the more cynical category. Work has explored the reasons for the growing lack of faith in government in the United States (Nye et al. 1997; Miller and Borrelli 1991), declining confidence in American institutions (Lipset and Schneider 1987; Hibbing and Theiss-Morse 1995), and the phenomenon of the 'angry' voter (Tolchin 1996; Craig 1993). Sweden also seems to have become less trusting (Holmberg 1997). In contrast, the five-volume Beliefs in Government project found that in Western Europe the pattern of satisfaction with democracy showed trendless fluctuations over time, rather than any secular or steady decline (Klingemann and Fuchs 1995; Kaase and Newton 1995). After weighing a wide range of evidence Budge and Newton concluded that popular accounts of a widespread crisis of democracy lacked convincing empirical support in Western Europe.

"It is possible to draw out a few generalizations about mass political attitudes in Western Europe in the post-war period. First, there is little evidence to support the various theories of crisis, contradiction, and catastrophe. There are few signs of a general decline in trust, confidence in public institutions, political interest, or faith in democracy; nor is there much evidence of an increase in apathy, alienation or revolutionary feelings. Second, Western Europe is characterised by random fluctuations rather than general patterns of change." (Budge and Newton 1997).

Explanatory Frameworks

Why are there these persistent and well-established differences in system support between countries? As discussed by Nye, Zelikow and King (1997), we can identify at least three separate schools of thought focussing on the role of cultural values, government performance, and political institutions.

Theories of Cultural Values

Following in the footsteps of Almond and Verba (1963), one strand of research has emphasised how cross-national differences towards the regime are related to deep-rooted social and political values in each country. The imprint of founding ideas may persist in the chromosomes of each culture and leave its mark on the dominant public philosophy towards government. An extensive literature has explored these issues. Theories of post-materialism, and more recently post-modernization, provide the most comprehensive account of political culture, suggesting that a deep-rooted process of value change is gradually transforming citizens' relationship towards government (Inglehart 1977; Inglehart 1990; Inglehart 1997; Abramson and Inglehart 1995; Ester, Halman, and deMoor 1993; Dalton 1996). In this view, core values transmitted through the socialisation process are shaped by family, friends and formative experiences in early youth. These accounts are most powerful when explaining persistent and longterm differences between nations, secular trends over time, and diffuse support such as feelings of national identity. Nevertheless the focus on the socialisation process means we have to turn elsewhere when accounting for short-term fluctuations in attitudes towards the regime.

Theories of Government Performance

Alternatively, theories of political economy commonly focus on how system support relates to public evaluations of government performance, particularly concerning the economy. This perspective notes that confidence in government and political trust are not stable phenomena. Fluctuations in support, it is suggested, reflect the public's overall evaluation of the performance of political leaders and, more generally, the ability of the administration to handle the economy (Citrin 1974; Weatherford 1984; Weatherford 1991). Where successive governments have succeeded in meeting public expectations of peace and prosperity, and more generally fulfilling their campaign promises, this generates diffuse support towards the political regime in general. Particularly in parliamentary democracies, when we tap 'confidence in government', there is ambiguity about whether we are measuring attitudes towards 'the' government (the party in power) or attitudes towards the governmental system. A series of studies have focussed on the relationship between aggregate levels of economic growth, unemployment or inflation, as well as individual-level retrospective and prospective evaluations of the economy, and confidence in governance in the United States, Canada, and Western Europe (Clarke, Dutt, and Kornberg 1993; Kornberg and Clarke 1992; Weatherford 1984; Weatherford 1992; Weatherford 1991; Weil 1989).

If 'performance' is defined fairly narrowly as purely economic then this perspective seems to provide a poor fit, prima facia, for trends in many countries. Both Italy and Japan have experienced rapid economic growth in the post-war era, although political cynicism seems pervasive and widespread in both these countries (Pharr 1997; Morlino and Tarchi 1996). Moreover in America confidence in government declined throughout the 1960s despite a relatively prosperous economy during this period (Lawrence 1997). This theory becomes more convincing if non-economic aspects of government performance are included in the equation, including critical events, for example Watergate and the failure of foreign policy in Vietnam, but then it becomes harder to define any independent and objective measure of 'performance' (Bok 1997). If this account is correct, then we need to explain why the record of some governments seem consistently more successful than others.

Theories of Political Institutions

Lastly, one of the most plausible and yet neglected explanatory frameworks, examined in this paper, seeks to connect the attitudes of citizens with their constitutional context. In this regard, we return to the idea, first suggested by David Easton, that diffuse system support relates to our accumulated experience:

"Members do not come to identify with basic political objects only because they have learned to do so... from others - a critical aspect of socialization processes. If they did, diffuse support would have entirely the appearance of a non-rational phenomenon. Rather, on the basis of their own experiences, members may also adjudge the worth of supporting these objects for their own sake. Such attachment may be a product of spill-over effects from evaluations of a series of outputs and of performance over a long period of time. Even though the orientations derive from responses to particular outputs initially, they become in time disassociated from performance. They become transformed into generalised attitudes towards the authorities or other political objects." (Easton 1975).

But what experience counts? The argument presented in this paper is that the pattern of winners and losers from the political system is structured by the constitutional arrangements, meaning the core institutions of state and the rules of the game, both written and unwritten. Some win, some lose. Over a long period of time this experience can be expected to shape our general orientations towards the political regime. At the simplest level, if we feel that the rules of the game allow the representatives we endorse to be elected to power, we are more likely to feel that the political system is responsive to our needs so that we can trust government. If we feel that the representatives we prefer persistently lose, over successive elections, we are more likely to feel that our voice is excluded from the decision-making process, producing dissatisfaction with the political system. Where constitutional arrangements succeed in channelling popular demands into government outcomes, then we would expect this to be reflected in confidence in the political process. As Weatherford suggests, the methodological challenge for empirical researchers is to ground individual-level findings in theories about the polity (Weatherford 1992).

What evidence is there that institutions matter for system support? An extensive literature has explored the consequences of alternative constitutional designs, for example, the influence of presidentialism or parliamentarianism on executive stability (Powell 1982; Lijphart and Waisman 1996), and the impact of majoritarian or proportional electoral systems on the representation of political minorities, party fragmentation and voter participation (Lijphart 1994; Sartori 1994; Norris 1997). Yet few have looked systematically at the relationship between constitutional arrangements and public support.

The most thorough analysis has recently been conducted by Anderson and Guillory, who compared satisfaction with democracy among consensual and majoritarian political systems in Western Europe (Anderson and Guillory 1997). The study hypothesised that (i) system support is consistently influenced by whether people are among the winners or losers in electoral contests, defined by whether the party they endorsed was returned to government, and (ii) that this process is mediated by the type of democracy. Based on Lijphart's work nations can be classified into majoritarian (Westminster) or consensual (consociational) democracies (Lijphart 1984). Based on theories of responsible party government, during their period of office majoritarian systems empower the winners to impose their preferences over the losers. Accountability and effective government are valued more highly by these systems than representation of all viewpoints (Norris 1997). Anderson and Guillory found that in majoritarian democracies winners consistently express far higher satisfaction with the system and confidence in democracy than losers. In contrast consensual democracies place a higher value on the inclusion of all political minorities in decision-making, giving less emphasis to accountability and the rotation of parties in power. Anderson and Guillory found that these systems produced a narrower gap in democratic satisfaction between winners and losers (see Figure 1). Crossnational differences in system support could therefore be explained as the product of the type of democracy and the distribution of winners and losers in each system. Majoritarian system with few winners could be expected to be characterised by low aggregate levels of system support, while consensual democracies with many winners could be expected to display high satisfaction with democracy. It should be noted that this theory only takes account of the proportion of winners and losers (their distribution), not how much they win or lose (their strength).

[Figure 1 about here]

Anderson and Guillory provided important theoretical insights into system support and an innovative research design linking individual and system-level characteristics. By focussing only on Western Europe, however, the institutional comparison was necessarily limited to established parliamentary democracies rather than other types of political systems. Moreover the analysis is of limited value for issues of constitutional design in emerging and transitional democracies, since it is not clear what specific institutions within majoritarian or consensual systems influenced levels of public support: whether the electoral system played the most important role, for example, or whether the division of executive and legislative powers was also significant. Equally important, recent work has thrown doubt on Lijphart's classification of democracies into the consociational and majoritarian categories. Comparisons have found that in practice there is as much institutional variation within, as between, these ideal-types (Lane and Ersson 1996; McRae 1997).

We can therefore build on this foundation, extending the approach by examining a wider range of institutional arrangements and a broader span of countries than previous studies. On this basis we can examine a series of related issues: Do systems with a wide range of political rights and civil liberties generate high levels of democratic satisfaction? Do multiparty or two-party systems produce greater system support? Do presidential or parliamentary system induce more public endorsement? Are citizens more trusting of government under federal or unitary systems? Do majoritarian or proportional electoral systems produce greater satisfaction with democracy? Our overall logic in exploring these issues, following Anderson and Guillory, is to test whether consensual institutions which maximise the number of 'winners' are more likely to produce high levels of system support, whereas winner-take-all majoritarian arrangements are more likely to produce less system support.

Data, Concepts and Methods

Selection of Countries

To examine these issues this study draws on the World Values Surveys which tapped public opinion in twenty-four nations in 1981-84 and forty-three nations in 1990-91. For our multivariate analysis we focus on a subset of major democracies, defined as countries with a population of 3 million or greater, and with a Gastil Political Rights score of 4 or less at the time of the survey, based on the Freedom House monitor (McColm 1991)(LeDuc, Niemi and Norris 1996 Table 1.1). This included in total 41,791 respondents living in the Americas (the United States, Argentina, Canada, Mexico, Brazil and Chile), Central and Eastern Europe (Hungary and Czechoslovakia), Asia (Japan and India), Scandinavia (Denmark, Norway, Sweden, and Finland), and Western Europe (France, Britain, West Germany, Italy, Netherlands, Belgium, Spain, Ireland, Switzerland, Portugal, Austria and Turkey). This universe included a wide variety of political systems, -- including presidential and parliamentary democracies, federal and unitary states, and significant variations in party and electoral systems, -- as well as differences between established and emerging democracies, and between developed and developing economies.

Measure of System Support

As our dependent variable we developed a System Support Scale (ranging from 6 to 24 points) based on summing confidence in five major political and civic institutions including parliament, the civil service, the legal system, the police, and the army . These items do not refer directly to the government, particular parties in office, political authorities, or policy outputs, but rather to broad attitudes towards the major institutions of state, which makes them suitable measures for diffuse regime support. Given the range of political systems under comparison, these items are also more suitable than more value-laden questions relating to 'satisfaction with democracy'. Responses to these five items were highly intercorrelated (see Appendix Table A), forming a reliable scale (Cronbach's Alpha .75), suggesting that they reflected a general orientation towards the political system rather than specific attitudes towards each separate institution.

In examining the effects of institutional arrangements we can control for (i) differences in political culture, using the standard (four-item) measure of post-materialism; (ii) differences in economic development, as measured by per capita GNP in 1990; and, (iii) the usual social background variables (age, income, education, and gender) which have often been found to be associated with variations in political attitudes. The 'winners' and 'losers' within each country were defined according to the voting intentions of respondents, and whether they supported a party which formed part of the government, drawing on information about the composition of cabinets from the Political Data Yearbook 1991 (Koole and Mair 1991). The major institutional contrasts within democratic political systems, as discussed below, were categorised based on Comparing Democracies (LeDuc, Niemi, and Norris 1996) and Political Systems of the World (Derbyshire and Derbyshire 1996).

Political Rights and Civil Liberties

Based on our earlier arguments, we can develop some testable propositions. First, within democracies we would expect (H#1) the existence of well-established political rights and civil liberties should be associated with high levels of public confidence in the regime. We can examine this using the Gastil Index published by Freedom House which monitors countries worldwide on an annual basis (McColm 1991). We used the monitor for the year prior to the survey in each country. Political rights are measured using a seven-point scale including such items as whether elections are free and fair, whether people have the right to organise competing parties, and whether minority groups have reasonable participation in the decision-making process. Civil liberties are gauged in a similar way monitoring, for example, whether there is a free and independent media, whether there is protection from political terror, and whether there is freedom of religion. The specific components of this scale, and their weighting, is open to debate (Beetham 1994), but this is generally accepted as a standard indicator comparing political rights and civil liberties.

Presidential v. Parliamentary Systems

The most important constitutional arrangements which vary between nations include executive-legislative relationship, party systems, electoral systems, and the federal or unitary centralisation of state power. Within democracies there is a major distinction between parliamentary or presidential systems. Democracies can be categorised based on Lijphart's (1992) three-fold definition: (i) In presidential systems the head of government has a fixed term of office, in parliamentary systems the head of government is dependent upon the confidence of the legislature; (ii) presidents are elected (directly or by electoral college) whereas prime ministers are selected by the legislature; and (iii) presidents have one-person, non-collegial executives whereas parliamentary systems have collective executives.

The merits of parliamentary and presidential systems have been widely debated (Lijphart 1992). Presidentialism with a weak legislature is said to have the advantages of executive stability, greater popular control, and more limited government, but the disadvantages of executive-legislative deadlock, temporal rigidity and winner-take-all government (Linz 1990). Parliamentarianism is said to reverse the pros and cons. Therefore our hypothesis, based on these arguments and our prior assumptions, would be to expect (H#2) that parliamentary systems, where all parties continue to have a stake in the policymaking process, should generate greater system support than winner-take-all presidential systems. Most West European democracies, as well as countries which were former British colonies, are parliamentary systems. In contrast many nations influenced by the United States adopted presidentialism, including the Spanish and Portuguese colonies in Latin America, the Philippines and South Korea. Worldwide there are around 55 parliamentary and 78 presidential executives, plus 12 semi-presidential (or dual) executives (Derbyshire and Derbyshire 1996). The democracies under comparison in this paper fall roughly equally into the parliamentary and presidential categories, based on the criteria of direct election to the executive, although some semi-presidential systems like France and Finland are more difficult to classify (see LeDuc, Niemi and Norris 1996, 283-284).

Party Systems

Parties remain the most important mediating institutions between citizens and the state, the focus of many previous studies of system support. The existence of minor 'protest parties', especially those on the right, has been found to act as a channel for disaffected voters (Miller and Listhaug 1990), although another study found that party fragmentation or polarisation was associated with lower levels of democratic support overall(Weil 1989). King suggests that dealignment among the American electorate, combined with greater ideological polarisation among the Republicans and Democrats in Congress, may have weakened the linkages between voters and their representatives in the United States (Nye et al. 1997).

To analyse this more systematically we need to compare party systems, meaning the durable structure of party competition in government and in the electorate. Early attempts at classification by Duverger distinguished between two-party systems and multiparty systems based on the number of parties (Duverger 1969). In an influential study, Sartori took this a step further, by taking account of two principle dimensions: the strength of parties, conventionally measured by votes in the electorate and seats in Parliament, and the position of parties across the ideological spectrum (Sartori 1976-). This yielded four principle types of party system: predominant one-party systems where one party consistently wins a majority of parliamentary seats; two-party systems, characterised by few parties and a small ideological distance; moderate pluralism characterised by multiparties and a small ideological distance; and polarised pluralism characterised by extreme multi-partyism and a large ideological distance. To compare a diverse range of polities we need to count the number of parties, although the practical problem is how to measure those of unequal size, particularly very small ones (do the Greens and Libertarians, for example, mean America has a multiparty system?). This paper adopts the Laakso and Taagpera index of the 'effective' number of parties in parliament (ENPP) (Laakso and Taagepera 1979; Lijphart 1994), estimated based on seat shares in the lower house (derived from Leduc, Niemi and Norris, 1996, Table 1.4).

Building on theories of responsible party government we would predict that certain systems are more likely to strengthen the link between citizens' preferences and government policies. In particular, in countries with predominant one-party government and fragmented oppositions it is extremely difficult for citizens to use elections as an opportunity to 'kick the rascals out', if dissatisfied with the performance of officeholders. This pattern certainly seems to fit some of the anecdotal evidence about Japan and Italy, both countries which have experienced considerable economic growth yet with predominant one-party systems throughout most of the post-war period. The net effect may have been to make successive governments unaccountable and unresponsive to public opinion, producing disillusionment with the political system in general. Fragmented party systems are characterised by ideological polarization, weak and unstable coalition governments, and bidding war, whereas systems with a few broadly based and centrist parties are better placed to aggregate interests into broad social and ideological coalitions (Diamond, Linz, and Lipset 1995). In contrast, we expect two-party and moderate multiparty systems to function more effectively as a mechanism for translating electoral choices into government policy. On this basis we would expect (H#3) countries with fragmented multiparty systems should have low levels of system support.

Federal v. Unitary States

Democracies can also be classified into unitary and federal states. In unitary systems the state is one and indivisible, meaning that the central government exercises authority over the population directly, whereas in federal systems the central government shares authority with other states within its own territory. Lane and Ersson's comparison classifies most countries within the new Europe as unitary, with the exceptions of Germany, Switzerland, Austria, and Belgium, with perhaps a semi-federal model with decentralised regions in Spain and Italy (Lane and Ersson 1996). Other federal states within our comparison include India, Brazil, Mexico and Canada, as well as the United States. Worldwide only 24 out of 192 states have a federal structure (Derbyshire and Derbyshire 1996), although many more have adopted some form of regional decentralization. Federal states are common in deeply divided societies, characterised by a historical legacy of strong ethnic and/or religious cleavages, and also in countries with a large land mass, like Canada, Australia and Brazil. In terms of confidence in government, advocates like Elazar argue that federalism manages to simultaneously accommodate the needs of different regions, and different groups in the electorate, whereas unitary states allow less flexibility and produce more losers from the system (Elazar 1997; Elazar 1997). To test this thesis, based on our previous assumptions, we would expect (H#4) that federal systems should produce higher levels of system support than unitary states.

Electoral Systems

The electoral system can also be expected to play a major role linking citizens and the state. Ever since the seminal early studies (Duverger 1969; Rae 1971), a flourishing literature has classified the main types of electoral systems and sought to analyse their consequences (Lijphart 1994; Lijphart and Grofman 1984; LeDuc et al. 1996; Bogdanor and Butler 1983; Taagepera and Shugart 1989; Farrell 1997; Cox 1997; Norris 1997). Systems vary according to district magnitude, ballot structures, effective thresholds, malapportionment, assembly size, and open/closed lists, but the most important variations concern electoral formula.

Electoral formula determine how votes are counted to allocate seats. There are four main types: majoritarian formulas (including plurality, second ballot, and alternative voting systems); semi-proportional systems (such as the single transferable vote, the cumulative vote, and the limited vote); proportional representation (including open and closed party lists using largest remainders and highest averages formula); and, mixed systems (like the Additional Member System combining majoritarian and proportional elements). We will therefore compare electoral formula used for the lower house of the national legislature in each country (Leduc, Niemi and Norris 1996 Table 1.2). We can also measure the degree of proportionality of electoral systems as an alternative indicator of electoral system effects. Based on our assumptions, we would expect that (H#5)majoritarian systems should produce less system support than proportional systems. Mixed and semi-proportional systems could be expected to fall somewhere between these poles. To summarise, the core propositions tested within this paper are given in Table 1. [Table 1 about here]

Results and Findings

First, we can examine the distribution of system support at aggregate level among different democracies. As shown in Figure 2, as expected we confirmed considerable cross-national variations, with citizens in Mexico, Italy and Czechoslovakia (which was going through a significant upheaval at the time of the survey) displaying least confidence in their political systems. Japan also displayed widespread cynicism, confirming other observations (Pharr 1997), as did Spain and Portugal, both newer democracies which had been through the transition process during the 1970s. At the other end of the spectrum, citizens proved most confident of their regime in many Anglo-American democracies, including India, Ireland, Britain, the United States and Canada. Confidence in government may have declined over time within America, but even so compared with many other nations system support remains relatively high. Citizens were also relatively positive in Norway and Denmark, As noted elsewhere (Holmberg 1997), as well as in Chile.

[Figure 2 about here]

To analyze trends over time we can compare system support in seventeen countries included in both the 1991-84 and 1990-93 WVS surveys. With only two points of comparison, we cannot monitor secular trends, since both measurements could be at a particular point in a cycle, but we can monitor the direction of differences over time. The results are quite striking and consistent: system support declined slightly in every country under comparison although usually by only a modest amount. France, the Netherlands and West Germany remained fairly stable while countries characterised by relatively low support, including Japan, Italy and Mexico, drifted only marginally downwards. Only Finland, Argentina and South Korea fell by more than one point in support during this period, and the shifts over time were rarely sufficient to change the rank order of countries on the scale.

[Figure 3 about here]

To see how this pattern relates to our main hypotheses we can compare system support in 1990-93 among the winners and losers within each nation. As shown in Figure 4 there is a clear and consistent pattern: winners who backed the party or parties in government consistently displayed greater satisfaction with the political system in every country except one (France). This nicely replicates and confirms the earlier findings of Anderson and Guillory (1997), based on comparing democratic satisfaction in the Eurobarometer with a more limited range of countries. Moreover the gap between winners and losers was particularly marked in three countries which at the time of the second survey (1990-93) had been characterised by long periods of one-party predominant government and fragmented oppositions, notably Mexico (the PRI), Japan (the LDP) and Italy (the Christian Democrats)(Pempel 1990). In such systems, the losers have been excluded from power over successive elections, and we might expect them to lack faith in the system. Other countries with a relatively large gap between winners and losers included West Germany and Britain.

[Figure 4 about here]

To analyse whether this pattern varied at the individual level in a multivariate model we can use ordinary least squared regression analysis, with the system support scale as the dependent variable. The first model entered the general indicators of political rights and civil liberties, and whether respondents were winners and losers from the system. The second model added in the main institutional variables, while the third model also included the controls in terms of post-material values, levels of economic development, and the social background of respondents.

[Table 2 about here]

The results (see Table 2) confirm that system support in democracies was significantly related to the range of civil liberties in each country: citizens express more confidence in the system of government where, according to the Freedom House classification, there are widespread opportunities for civic participation and protection of human rights. But, contrary to expectations, political rights as measured by the Gastil Index were negatively associated with system support. The reasons for the divergence between findings remains a puzzle. Since some of the newer democracies were experiencing a period of regime change at the time of the survey we need to monitor this relationship in later surveys, and these preliminary results need further analysis. The multivariate model found that whether respondents were classified as winners (who endorsed the governing party or parties) or losers (who backed one of the opposition parties) was also significantly related to their system support, confirming our earlier observations.

In the second model, after controlling for these factors, all the institutional variables proved significant and in the expected direction, with the exception of federalism. As predicted, system support proved to be stronger in parliamentary systems and in electoral systems with proportional representation, which allowed more parties to have a voice in the decision-making processes of government. Nevertheless although moderate multipartyism increased this tendency, the results suggest that fragmented multiparty systems produced more dissatisfaction with the regime:the greater the number of effective parliamentary parties the less the system support. In terms of our initial hypothesis, the one contrary findings was that system support proved to be greater in unitary rather than federal states. Perhaps this may be because federalism is not necessarily an appropriate measure of decentralization per se (Lane and Ersson 1996), nor do we take account of the distribution of ethnic and regional cleavages within the nation-state. Possibly we need better indicators of regional and local governments to examine whether a broader distribution of powers produces greater system satisfaction. Alternatively, unitary states may increase accountability and responsible party government.

Lastly, in the third model these relationships were confirmed even after controlling for differences in levels of economic development and post-material values. Indeed the institutional variables proved to be equally, or more strongly, related to system support than either of these factors. In terms of the social background of respondents, older respondents expressed more confidence in the system, as previous studies have commonly found, gender was very weakly related to system support, while the influence of income and education proved insignificant. It should be stressed at this point that this remains only a preliminary analysis of the results and we need some caution in replying upon these initial findings. In further research it is intended to extend the number of countries and examine a wider range of institutional variations, as well as to replicate the findings in the light of alternative survey data. Nevertheless, using a different dataset and measure of system support, the preliminary results do seem to neatly replicate the main theoretical principles of Anderson and Guillory (1997): winners express more confidence in the system than losers, and (most but not all) consensual institutions designed to maximise the number of winners seem to produce more system support than (most but not all) majoritarian institutions. Based on this initial analysis of the evidence the central thesis that institutions have the capacity to influence attitudes towards the regime does seem a fruitful avenue of research which is worth exploring further. The bottom line is that we need to bring the constitutional arrangements back into studies of system support.

Conclusions and Implications

The study started from the premise that regime support reflects the general constitutional arrangements which structure behavior and attitudes towards governance into stable and predictable patterns. We found that citizens who live in democracies with a strong tradition of civil liberties express considerable confidence in their political system. In contrast, people are more skeptical about government, for good reason, in states which lack such conditions, for example where there is widespread electoral corruption and intimidation, where opposition publications are censored or banned, and where rights to protest are suppressed. This finding may seem obvious, but, perhaps surprisingly, it has rarely been examined systematically in comparative studies, perhaps because these have usually focussed on Western democracies where such rights are taken for granted. Nevertheless the disparities in the initial results, in the difference between the impact of political rights and civil liberties, does need further exploration before we have full confidence in the reliability of this analysis.

Second, we found that evaluations of the political regime reflect our experience of whether we are winners or losers over successive elections, defined by whether the party we endorse is returned to government. Democratic institutions represent intermediary mechanisms linking public preferences with the performance of the state. In this regard institutions are not neutral. Instead the arrangements - whether for majoritarian or proportional electoral systems, federal or unitary states, and presidential or parliamentary executives - consistently rule some groups into, and some groups out of, the decision-making process (Powell 1989). We confirmed that people who voted for the governing party or parties are more likely to believe that the political system is responsive to their needs. In contrast, losers in elections are less satisfied with the way democracy works. Cynicism is highest in countries which produce many persistent losers over successive elections, for example, systems governed by one-party predominant parties facing a fragmented opposition. Lastly, we can conclude that certain institutional arrangements are significantly related to higher system support, notably parliamentary executives, proportional electoral systems, and unitary states. Strikingly, none of these features are part of the constitutional structure of the United States.

So what? The implications of these findings may be significant for at least two distinct reasons. First, theoretically this question throws light on the general issue of whether institutions matter (Weaver and Rockman 1993). One of the basic assumptions of 'new institutionalism' is that individual behaviour is constrained and channelled by its rule-based institutional context. If so, the research traditions seeking to explain system support based on long-standing cultural values, or the political economy of government performance, need to be supplemented by taking account of the mediating influence of political institutions.

Equally importantly, if system support is systematically and consistently related to institutional design this may have significant implications for public policy. The constitutional order and the basic rules of the game, like electoral systems, are usually regarded as fairly stable features of the political landscape. Yet the wave of constitution-building following the explosion of new democracies during the 'third wave', combined with new pressures on institutional reform in many established democracies, has encouraged greater interest in the choices involved in 'constitutional engineering' (Sartori 1994) and 'institutional design'(Lijphart and Waisman 1996). Constitutional debates - whether about electoral systems, devolution or executive-legislative relations - have risen on the policy agenda to become first-order issues in countries as diverse as South Africa, New Zealand, Russia, Canada, Mexico, Italy, Japan, and the United Kingdom (Laponce and Saint-Jacques 1997). If institutional arrangements influence system support and public attitudes towards the regime, as we commonly assume, this may have important consequences for constitutional debates. The challenge for further research is to consider these issues in terms of the democratization process. Some nations seem to have managed the successful transitions towards consolidated democracies, including South Africa, the Czech Republic, and Argentina, generating a reservoir of system support among the public to tide the political regime through good times and bad, but in contrast other fledgling democracies like Nigeria have reverted back towards authoritarian rule (Linz and Stepan 1978; Morlino and Montero 1995). If institutional designs can help to strengthen public support for the regime this may provide significant lessons for the process of democratization.

Table 1: Summary of Core Hypotheses

______________________________________________________________

 

Table 2 Models of System Support

_________________________________________________________________________________________________________

_________________________________________________________________________________________________________

Note: OLS regression coefficients with system support as the dependent

variable.

Source: World Values Survey 1990-93.

    Reference List

Abramson, Paul R. and Ronald Inglehart. 1995. Value Change in Global Perspective. Ann Arbor, Mich: University of Michigan Press.

Almond, Gabriel A. and Sidney. Verba. 1963. The Civic Culture; Political Attitudes and Democracy in Five Nations. Princeton, N.J: Princeton University Press.

Anderson, Christopher J. and Christine A. Guillory. 1997. "Political Institutions and Satisfaction With Democracy." American Political Science Review 91(1):66-81.

Beetham, David. 1994. Defining and Measuring Democracy. London: Sage.

Bogdanor, Vernon and David Butler. 1983. Democracy and Elections : Electoral Systems and Their Political Consequences. Cambridge [Cambridgeshire] ; New York: Cambridge University Press.

Bok, Derek. 1997. "Measuring the Performance of Government." Why People Don't Trust Government, eds Joseph S. Nye, Philip Zelikow, and David King. Cambridge, MA: Harvard University Press.

Booth, John A. and Mitchell A. Seligson. 1994. "Paths to Democracy and the Political Culture of Costa Rica, Mexico and Nicaragua." Pp. 99-130 in Political Culture & Democracy in Developing Countries, Larry Diamond. Boulder, CO: Lynne Reinner.

Budge, Ian and Kenneth Newton. 1997. The Politics of the New Europe. Harlow, Essex: Addison Wesley Longman.

Citrin, Jack. 1974. "Comment: The Political Relevance of Trust in Government." American Political Science Review 68:973-88.

Clarke, Harold D., Mitish Dutt, and Allan Kornberg. 1993. "The Political Economy of Attitudes Toward Polity and Society in Western European Democracies." The Journal of Politics 55(4):998-1021.

Cox, Gary W. 1997. Making Votes Count. NY: Cambridge University Press.

Craig, Stephen C. 1993. The Malevolent Leaders: Popular Discontent in America. Boulder, CO.: Westview Press.

Dalton, Russell J. 1996. Citizen Politics: Public Opinion and Political Parties in Advanced Industrialised Democracies. Chatham NJ: Chatham House.

Dalton, Russell J. 1997. "Citizens and Democracy: Political Support in Advanced Industrial Democracies." (Washington, DC., ).

Davis, James A. 1997. Workshop on Confidence in Democratic Institutions: America in Comparative Perspective (Washington DC, ).

Derbyshire, J. D. and Ian Derbyshire. 1996. Political Systems of the World. New York: St. Martin's Press.

Diamond, Larry J., Juan J. Linz, and Seymour M. Lipset. 1995. Politics in Developing Countries : Comparing Experiences With Democracy. 2nd ed. ed. Boulder: L. Rienner Publishers.

Duverger, Maurice. 1969. Partis Politiques. [3d ed.] Translated by Barbara and Robert North. With a foreword by D. W. Brogan. ed. [London]: Methuen.

Easton, David. 1975. "A Reassessment of the Concept of Political Support." British Journal of Political Science 5:435-57.

Elazar, Daniel J. 1997. "Contrasting Unitary and Federal Systems." International Political Science Review 18(3):237-51.

Ester, Peter, Loek Halman, and Ruud deMoor. 1993. The Individualizing Society. Tilburg: Tilburg University Press.

Farrell, David. 1997. Comparing Electoral Systems. London: Prentice Hall.

Hibbing, John R. and Elizabeth Theiss-Morse. 1995. Congress As Public Enemy. New York: Cambridge University Press.

Holmberg, Soren. 1997. "Down and Down We Go: Political Trust in Sweden." Conference on Confidence in Democratic Institutions (Washington, DC, ).

Inglehart, Ronald. 1990. Culture Shift in Advanced Industrial Society. Princeton, N.J: Princeton University Press.

———. 1997. Modernization and Postmodernization: Cultural, Economic and Political Change in 43 Societies. Princeton, NJ: Princeton University Press.

———. 1977. The Silent Revolution : Changing Values and Political Styles Among Western Publics. Princeton, N.J: Princeton University Press.

Kaase, Max and Kenneth Newton. 1995. Beliefs in Government. Oxford ; New York: Oxford University Press.

Klingemann, Hans-Dieter and Dieter. Fuchs. 1995. Citizens and the State. New York: Oxford University Press.

Koole, Ruud and Peter Mair. 1991. European Journal of Political Research. Political Data Yearbook ed. Vol. 21(4). Dordrecht: Kluwer Academic Publishers.

Kornberg, Allan and Harold D. Clarke. 1992. Citizens and Community : Political Support in a Representative Democracy. Cambridge [England] ; New York: Cambridge University Press.

Laakso, Markuu and Rein Taagepera. 1979. ""Effective" Number of Parties: A Measure With Application to West Europe." Comparative Political Studies 12:3-27.

Lagos, Marta. "The Latinobarometro: Media and Political Attitudes in South America." American Political Science Association Annual Meeting (San Francisco, ).

Lane, Jan-Erik and Svante O. Ersson. 1996. European Politics. London: Sage.

Laponce, Jean and Bernard Saint-Jacques. 1997. Contrasting Political Institutions - International Political Science Review. Vol. 18(3)Sage.

Lawrence, Robert Z. 1997. "Is It Really the Economy, Stupid?" Pp. 107-28 in Why People Don't Trust Government, eds Joseph S. Nye, Philip Zelikow, and David King. Cambridge, MA: Harvard University Press.

LeDuc, Lawrence, Richard G. Niemi, and Pippa. Norris. 1996. Comparing Democracies : Elections and Voting in Global Perspectives . Thousand Oaks, Calif: Sage Publications.

Lijphart, Arend. 1994. "Democracies: Forms, Performance and Constitutional Engineering." European Journal of Political Research 25:1-17.

———. 1984. Democracies ; Patterns of Majoritarian and Consensus Government in Twenty-One Countries. New Haven: Yale University Press.

———. 1994. Electoral Systems and Party Systems : a Study of Twenty-Seven Democracies, 1945-1990. Oxford ; New York: Oxford University Press.

———, ed. 1992. Parliamentary Versus Presidential Government. Oxford: Oxford University Press.

Lijphart, Arend and Bernard. Grofman. 1984. Choosing an Electoral System : Issues and Alternatives. New York: Praeger.

Lijphart, Arend and Carlos H. Waisman. 1996. Institutional Design in New Democracies : Eastern Europe and Latin America. Boulder, Colo: Westview Press.

Linz, Juan J. 1990. "The Perils of Presidentialism." Journal of Democracy 1(1).

Linz, Juan J. and Alfred C. Stepan. 1978. The Breakdown of Democratic Regimes. Baltimore: John Hopkins University Press.

Lipset, Seymour M. and William Schneider. 1987. The Confidence Gap : Business, Labor, and Government in the Public Mind. Rev. ed., John Hopkins paperbacks ed. ed. Baltimore: Johns Hopkins University Press.

McColm, R. B. 1991. "The Survey 1991." Freedom Review 22(1).

McRae, Kenneth D. 1997. "Contrasting Styles of Democratic Decision-Making: Adversarial Versus Consensual Politics." International Political Science Review 18(3).

Miller, Arthur H. and Stephen A. Borrelli. 1991. "Confidence in Government During the 1980s." American Politics Quarterly 19(2):147-73.

Miller, Arthur H. and Ola Listhaug. 1990. "Political Parties and Confidence in Government: A Comparison of Norway, Sweden and the United States." British Journal of Political Science 29:357-86.

Miller, Arthur, Vicki Hesli, and William Reisinger. 1996. "Understanding Political Change in Post-Soviet Sociaties." American Political Science Review 90(1):153-66.

Mishler, William and Richard Rose. "Five Years After the Fall: The Trajectory and Dynamics of Support for Democracy in Post-Communist Europe." Conference on Confidence in Democratic Institutions (Washington DC, ).

———. 1994. "Support for Parliaments and Regimes in the Transition Toward Democracy." Legislative Studies Quarterly XIX:5-32.

———. 1995. "Trajectories of Fear and Hope: Support for Democracy in Post-Communist Europe." Comparative Political Studies 28:553-81.

Morlino, Leonardo and Jose R. Montero. 1995. "Legitimacy and Democracy in Southern Europe." The Politics of Democratic Consolidation, Eds Richard Gunther, P. N. Diamandouros, and Hans-Jurgen Puhle. Baltimore, MA: The Johns Hopkins Press.

Morlino, Leonardo and Marco Tarchi. 1996. "The Dissatisfied Society: The Roots of Political Change in Italy." European Journal of Political Research 30(1):41-63.

Norris, Pippa. 1997. "Choosing Electoral Systems: Proportional, Majoritarian and Mixed Systems." International Political Science Review 18(3).

Nye, Joseph S. Jr., Philip D. Zelikow, and David C. King. 1997. Why Americans Mistrust Government. Cambridge, MA.: Harvard University Press.

Pempel, T. J. 1990. Uncommon Democracies: The One Party Dominant Regimes. Ithaca, NY: Cornell University Press.

Pharr, Susan. 1997. "Confidence in Government: Japan." Why Americans Mistrust Government, eds Joseph S. Nye, Philop D. Zelikow, and David C. King. Cambridge, MA.: Harvard University Press.

Powell, Jr. G. B. 1989. "Constitutional Design and Citizen Electoral Control." Journal of Theoretical Politics 1(2):107-30.

———. 1982. Contemporary Democracies. Cambridge, MA.: Harvard University Press.

Rae, Douglas W. 1971. The Political Consequences of Electoral Laws. Rev. ed. ed. New Haven: Yale University Press.

Sartori, Giovanni. 1994. Comparative Constitutional Engineering. London: Macmillan.

———. 1976-. Parties and Party Systems : a Framework for Analysis. Cambridge [Eng.] ; New York: Cambridge University Press.

Taagepera, Rein and Matthew S. Shugart. 1989. Seats and Votes : the Effects and Determinants of Electoral Systems. New Haven: Yale University Press.

Tolchin, Susan J. 1996. The Angry American: How Voter Rage Is Changing the Nation. Boulder, CO: Westview Press.

Weatherford, M. S. 1992. "Measuring Political Legitimacy." American Political Science Review 86:149-66.

Weatherford, Stephen. 1984. "Economic Stagflation and Public Support for the Political System." British Journal of Political Science 14:187-205.

———. 1991. "Mapping the Ties That Bind: Legitimacy, Representation and Alienation." Western Political Quarterly 44:251-76.

Weaver, R. K. and Bert A. Rockman, Eds. 1993. Do Institutions Matter. Washington, D.C.: The Brookings Institution.

Weil, Frederick. 1989. "The Sources and Structure of Legitimation in Western Democracies." American Sociological Review 54:682-706.

White, Stephen, Richard Rose, and Ian McAllister. 1997. How Russia Votes. Chatham, NJ: Chatham House Publishers.

Contact: Pippa Norris, John F. Kennedy School of Government, Harvard University, Cambridge, MA 02138.

Pippa_Norris@Harvard.edu Fax: (617) 495 8696. Tel: (617) 495 1475. www.pippanorris.com 05/31/2001