Resumé

Many modern democracies have experienced a decrease in citizen support for government in recent decades. This article examines attitudes toward public policy as a plausible theoretical explanation for this phenomenon. The connection between public policy and support for the political regime has received considerable academic attention in the United States. Yet very little comparative work has examined whether citizens' policy preferences are related to a decline in diffuse support across different political systems. This article offers a clearer, more concise theoretical specification of the hypothesized relationship between public evaluations of policy outputs and support for the political regime. After specifying the theoretical concerns more succinctly, the article analyzes data from Norway, Sweden and the United States for the quarter century from the late 1960s to the early 19905. The analysis reveals that shifts in evaluations of foreign policy and race-related policies help explain change in political trust for all three countries despite differences in the political systems. Moral issues, such as abortion, however, have no impact on political trust in any of the countries.

Theoretical Concerns

The nexus between public policy and support for a political regime has received varied treatment in political science literature. Some theories of political alienation have dismissed that public policy has any influence on system support (Citrin 1974; Citrin & Green 1986; Tyler, Rasinski & McGraw 1985). Easton (1965) also argues that disagreements over policy outcomes could undermine specific support for incumbents, but not affect diffuse support for a political regime. Yet others have treated policy dissatisfaction and political alienation as nearly synonymous, for example Almond & Verba (1965) who explain the linkage by differentiating between input and output alienation. The "input" element refers to a personally perceived inability to influence government agencies or affect the inputs to government policy making, while "output" alienation means that people

---

* Arthur H. Miller, lowa Social Science Institute, the University of lowa, 123 N. Linn St.. 130 Brewery Square, lowa City, lowa 52242-1409. USA and Ola Listhaug, Department of Sociology and Political Science, The Norwegian University of Science and Technology. N-7034 Trondheim, Norway.

perceive government policies as unresponsive to their preferences (Almond & Verba 1965, 51-2, 63-4). The problem with this approach is that the very definition of alienation incorporates policy-related discontent, and thus policy preferences cannot be treated as a separate cause of alienation.

Although other research suggests a correlation between attitudes on policy issues and a variety of diffuse support indicators, such as efficacy, trust and government responsiveness, it also suffers from limitations. In general, this work often confounds real events, such as assassinations or government scandals, with policy outputs (Converse 1972), or it fails to elaborate the process by which policy-related attitudes become linked with alienation from the political system (Lipset & Schneider 1983). In addition, this previous research usually focused on the United States and was limited to single-time point, cross sectional analysis (Miller 1974; House & Mason 1975; Miller & Borrelli 1991), which raised questions about the generalizability of the findings. The purpose of this article is to address some of these limitations by systematically elaborating the theoretical and empirical relationship between public evaluations of policy outputs and support for a political regime.

Previous literature suggests that all modern democracies have experienced a rather alarming loss of trust in government or diffuse support for the political regime in recent decades (see e.g., Crozier et al. 1975; Lipset & Schneider 1983; Lipset 1994). Much of this work, however, is rather speculative, lacks truly comparative evidence, or focuses on a rather short time frame. This article first operationally defines diffuse support in terms of political trust, then presents the trend in trust for three Western democracies and continues with an empirical examination of attitudes toward public policy as a plausible theoretical explanation for the growth in alienation¹ in these countries over a quarter of a century.

Comparative Trends in Political Support

The loss of diffuse political support as a psychological feeling of divorcement from government has been conceptualized as a multi-faceted phenomenon (see e.g., Finifter 1970; Balch 1974; Weatherford 1992). In general, this conceptual work suggests that a feeling of alienation from government is a reflection of the belief that government is either incapable of or unwilling to respond to citizen demands and hence that the political system is perceived as unresponsive and not to be trusted. Given that responsiveness and trust in government are fundamental principles of democracy, monitoring citizen trust in government is an important means of determining the vitality of democratic political systems.

There are few empirical comparisons of citizen trust in government across
different countries, especially over any length of time. Cross-cultural comparative

research is often curtailed by the absence of comparable empirical evidence from different countries, and available data on trust and efficacy are no exception; however, a number of directly comparable items are available in American, Norwegian and Swedish national election surveys,² and both theoretically and practically, these three countries offer a rich set of comparisons. As social democracies with multiparty, parliamentary political systems, Norway and Sweden provide interesting comparisons with the United States which has a very different type of political system (two-party, presidential, with a balance of power between President and Congress). Yet, because Norway and Sweden are very similar, the three countries combined offer comparisons of both different and similar systems at the same time. Basically, if structural or cultural differences are the primary factors influencing political trust, then we should expect the analysis to display similar results in Norway and Sweden, but the results in both countries should vary systematically in comparison with the United States. Therefore, this combination provides the types of comparisons called for by comparative theorists (for an overview of these concerns, see Rogowski 1993).

Of course, the relationship between public evaluations of policy outputs and political trust could be examined at a single point in time and within a single country. But if theory development is the major concern, then generalizability of the theory across time and place is critical. Although a comparative cross-cultural investigation of trust would shed light on how different political systems function, our principal concern is to determine whether the form of the relationship between public policy preferences and trust in government shifts across time and political systems. Such an investigation not only addresses the generalizability of the empirical relationship but also provides clues that may help explain change in political trust. We do not claim that popular preferences on public issues could provide a full accounting of citizen alienation from government, neither cross-sectionally nor longitudinally. After all, a general explanation of complex attitudes like political alienation would require an examination of several plausible explanatory factors, which is beyond the scope of this article. Here our goal is more modest. We are mainly interested in testing the hypothesized relationship between policy preferences and political trust across societies and time.

A comparison of Norway, Sweden and the United States also makes sense from a practical perspective. The empirical evidence regarding citizen evaluations of government for these specific countries spans the quarter century from the mid-1960s to the early 19905, thus extending over a much broader time period than similar data that reflect the level of political support in other countries. Survey indicators of diffuse regime support have generally been drawn from a standardized battery of trust or external efficacy items. The trust items typically ask if political leaders and government can be trusted

to do what is right, and if they are honest, fair and competent. The external efficacy items generally ask if elected officials stay in touch with the voters, care what people think, and are interested in opinions or just seeking votes. Measures of external efficacy and trust are usually highly correlated and exhibit the same trends across time, thus supporting the assumption that they tap the same underlying and more abstract concept of diffuse regime support (Weatherford 1992).³

As illustrated in Figure 1, trust showed a particularly abrupt decline in all three countries between the mid-1960s and the mid-19705. After that, support recovered in Norway, but continued to decline in Sweden and the United States. In the latter two countries, the downward trend was temporarily halted - between 1980 and 1984 in the United States and between 1979 and 1982 in Sweden - but resumed. Between 1985 and 1989, Norway also witnessed a marked decline in trust.

The similarities and differences in the trend for each country over the entire period for which data are available suggest that common factors influenced trust early on, but that somehow Norway escaped from the more enduring deterioration in support which has afflicted both Swedish and American politics. Between 1964 and 1973, certain world events and social problems affected all three countries. These included the Vietnam War, growing stagflation, increasing criticism of the welfare system, concerns about taxation, and environmental issues. At the same time, there were issues that on the surface seem unique to each country, such as Watergate in the United

States, the European Community (EC) vote in Norway, and the nuclear power
controversy in Sweden, which could potentially help account for the trends in
each country.

An examination of all the alternative explanations for political trust in each country would be enlightening and would help us explain the different levels in trust across these countries. Undoubtedly there are cultural factors and issues specific to each country which partially account for the across-country differences that affect trust. Differences in the party structures, the stronger acceptance of the welfare state ideology in Norway, the degree of polarization or homogeneity in each society, the relative incidence of government scandals and a number of other elements also contribute to these differences (for more on these, see Miller & Listhaug 1990; Risse-Kappen 1991). Such an undertaking, however, goes well beyond the scope of this article. As Dahl (1966, 353) has suggested, a systematic explanation for the variation in political allegiance across cultures goes beyond any available data. In addition, while the available data span more than a quarter of a century, the limited number of time points for the surveys precludes a good deal of sophisticated time series analysis that might be applied to explain the overall trends in political trust. Our goal is more modest, namely to determine if certain general policy-related factors are associated with political support in each country, and if the form of that relationship remains the same across time and political systems - the product of which would provide indirect evidence as to whether or not policy-related preferences help explain the trends in political trust across time and place.

Theoretical Hypotheses

There is nothing intrinsic in the support for a particular policy alternative, or even a pattern of policy preferences, that provides a theoretical expectation linking policy position with distrust. Nevertheless, empirical research from the 1970s (Miller 1974; House & Mason 1975) suggests a number of possible hypotheses that describe the relationship between policy preferences and political trust.

An initial hypothesis observed that in a properly functioning democracy the majority opinion would be enacted into policy and the dissenting minority would exhibit the greatest distrust of government because their alternative was not adopted (Miller 1974, 955). The interpretation here is that distrust results from a series of unsuccessful attempts to influence policy outcomes, hence the feeling that the system is unresponsive.

However, situations arose that contradicted this initial prediction (ibid..
963). When widespread polarization on an issue existed, the possibility arose
that the majority could favor a more extreme position than that endorsed by

the government. Under such circumstances, the majority could be more alienated than the minority. Furthermore, the situation of issue polarization could give rise to a curvilinear relationship between policy preference and alienation when the government pursues a centrist policy. In that situation, those who prefer either an extreme liberal or an extreme conservative policy position would be most discrepant with the government decision, thus most distrusting. It is understandable that individuals with the most crystallized policy attitudes - presumably those who favor the extreme alternatives - should potentially be the most alienated. Therefore, a linear relationship should be expected only when the government makes a policy decision that is clearly liberal or clearly conservative.

Relevant Issues

The goal here is not to catalog all potential issues that might influence attitudes toward government and political authority. Even if this was our goal, data limitations would prevent us from doing so. Rather, we will focus on a subset of issues that are relevant for potentially understanding long-term changes in trust for the three countries, namely the form of the relationship, and the question of whether change in policy decisions and policy preferences help us understand the dynamics of alienation in all three countries.

To meet these modest goals, only three issue areas are examined: foreign policy, race relations, and moral issues. The specific policy questions relevant to the political context of a given country are varied. Nevertheless, some types of policy questions are comparable across societies and should allow us to determine if policy preferences affect political trust the same way in different cultural settings.⁴ All countries deal with the issue of foreign policy and relations with other countries. For major military powers like the United States, this may involve concerns about defense spending or military activity in other countries. For Scandinavia, this issue may involve debates over aid to developing countries. Similarly, problems of race relations in the United States are very much a reflection of its diverse cultural heritage and the historical struggle of minorities to escape subjection. In the more homogeneous societies of Norway and Sweden, race relations have on occasion been framed in the context of protecting the status quo from encroachment by immigrants who attempt to claim welfare state benefits. Finally, each country, particularly the United States, has had continuing concerns in recent times with moral issues such as family values, gay life styles and abortion rights. We discuss these issues in turn, and then draw some general conclusions from the analysis regarding the link between policy preference and feelings of political alienation.

Foreign Policy and Alienation

Some of the earliest writing on the growth of alienation in the United States focused on public reactions to foreign policy and the Vietnam War (Converse 1972; Miller 1974; House & Mason 1975; Bachman & Jennings 1975; Markus 1979). Although Vietnam is generally remembered as a deeply divisive issue, it is not widely known that public attitudes toward Vietnam involvement took some time to crystallize. The US military presence in Vietnam steadily increased during the early 19605, but as Table 1 reveals it was not until 1968 that the percentage of citizens with no opinion about Vietnam dropped as low as 10 percent (see numbers in parentheses). As late as 1964, more than one in every three Americans lacked a preferred policy on Vietnam. Realizing that a significant proportion of the citizenry was yet developing an attitude on Vietnam in the late 1960s is important for understanding the link between preferences on Vietnam policy and alienation.

From 1964 to 1970, support for withdrawal from Vietnam shifted dramatically. The percentage favoring withdrawal from Vietnam increased, while the percentage preferring escalation remained relatively stable from 1968 to 1970. These data suggest that while support for withdrawal increased, there was no concurrent reduction in the polarization of attitudes. Most of the increased preference for withdrawal from Vietnam, therefore, came less from converting "hawks" to "doves" than from the uninformed taking a dovish position as the Vietnam issue became more salient.

A model that relates policy dissent with alienation would predict doves to be more cynical than hawks, since the US policy at the time was to wage war in Vietnam. Such a relationship would help explain the increase in cynicism between 1964 and 1970, the same period during which the ranks of the doves increased. Table 1 reveals that this prediction was largely correct. At each point in time, the doves exhibited relatively lower trust PDI values than the hawks or the moderates. Moreover, the data suggest that as the uninformed sided with the doves, they also became more distrusting of government. As US policy moved gradually toward withdrawal, however, even the hawks became increasingly discontented. In fact, the difference in cynicism between the hawks and doves narrowed over the years as the hawks became alienated at a faster rate than the doves.

The most persistent difference in Table 1, however, is not between those supporting withdrawal and those preferring escalation. Rather, the largest difference is between these two groups and those who favored continued involvement but in an effort to end the fighting. Respondents who selected this position consistently displayed more trust than hawks and doves. Hence, a curvilinear rather than a linear relationship best describes the association between Vietnam policy preference and political trust. One interpretation of this curvilinear pattern is that both hawks and doves had become increasingly

dissatisfied with the war, but for opposite reasons. The doves were inherently opposed to US involvement from the start and became increasingly dissatisfied as the conflict wore on, especially after Richard Nixon failed to quickly end the war effort as promised. In contrast, the hawks became increasingly dissatisfied with Vietnam policy even as a stronger stand was pursued, because in their opinion the escalation was not sufficient to win the war.

The shifts in attitudes toward Vietnam policy certainly help explain change in trust during the 1960s and 19705. However, since even those preferring a middle position became increasingly unsupportive across time (see Table 1), other issues or causal factors had to be operating as well, a point we return to later.

The Vietnam war officially ended in 1975, which excludes the issue as an explanation for long-term decline in political trust. Yet, in the face of changing international events, one facet of the debate endures. Military spending has remained one of the most controversial and hotly debated issues in US foreign policy. Given this continuing controversy, conflict over military spending and US involvement abroad could potentially have contributed to the long-term decay in political trust.

In 1972, with the Vietnam war drawing to a close, 35 percent of Americans favored cutting military spending (see Table 2). By 1980, however, with Americans held hostage in Iran, only 9 percent favored military reductions, while 61 percent favored increased military spending. The Reagan administration incorporated the public demand for increased defense spending into its 1980 election mandate and dramatically raised government outlays for the Defense Department. Popular support for this military buildup did not last

very long, however. By 1984, the proportion of those who favored military reductions had again increased to 28 percent, and support for more military spending had declined to 31 percent (see Table 2). Then, with the end of the cold war, the American public shifted dramatically toward a position of decreased military spending (supported by 42 percent in 1992).

Across time, those in favor of decreased military spending were generally more distrusting of government than those who favored increases. In each year, however, there is some curvilinearity to the relationship. Those positioned on both extremes express relatively more discontent with government (see Table 2). Miller & Borrelli (1991) have previously demonstrated that change in satisfaction with defense policy was significantly correlated with shifts in political trust during the 1980s. The increased military spending during Reagan's first term in office helped increase government support. But Reagan's continued spending despite growing budget deficits, alienated liberals without fully satisfying conservatives. The net result, by the end of Reagan's second term, was renewed political cynicism among both hawks and doves.

The Scandinavian foreign policy experience during the 1970s and 1980s, while less tumultuous than the American, nevertheless involved a number of controversial issues. The Vietnam issue played a minor role in Scandinavian politics in the 1960s and early 19705, although it had some influence on the radicalization of younger people. Empirical evidence on Scandinavian public opinion toward the Vietnam war, however, is extremely limited. Other foreign relations issues have had a more direct impact, as partially evidenced by the availability of empirical data.

We begin with the NATO issue in Norway, where the evidence reveals an irregular relationship between issue position and trust (see Table 3). In the period 1977-1985, the small minority that wanted Norway to leave NATO was consistently less trusting than those who supported membership in the alliance. In 1973 and 1989, however, there was no difference in trust levels between the two sides. The shift in opinion on the issue was marginal, although there was a drift toward increased support for NATO from 88 percent in 1973 to 92 percent in 1989. The tiny change in the aggregate distribution of opinion about NATO virtually precludes that it is a major explanation for changes in trust over time.

Similar arguments, on the other hand, would not apply to the controversial EC issue only a year before the first time point in our series. In September 1972, Norway held a referendum in which EC membership was narrowly defeated after a long and heated debate, which would lead to the expectation that those in favor of membership should be most cynical. There was a tendency in that direction in the 1973 data, but between 1977 and 1989 the

evidence reveals virtually the same level of trust regardless of issue position. While the EC question is given much credit for causing the political upheavals of the 19705, the issue was not strongly related to trust at that time. This is consistent with the prevailing observation that the losers in the 1972 referendum quickly adapted to the outcome and that the opposing political elites did not actively promote their cause for many years to come. By 1993, however, with a new European membership vote on the horizon for 1994 and the government taking a pro-EU position, the relationship with trust changed. Between 1989 and 1993, those in favor of membership displayed increased trust, whereas virtually no change occurred among those who opposed membership (see Table 3). This differential change helps to explain the slight overall rise in trust for Norway between 1989 and 1993.

Because there was no early vote on EC membership in Sweden, the expected relationship between EC attitudes and trust would be the opposite of that initially expected for Norway: we would expect Swedes opposed to EC membership to be more distrusting because most political leaders and the citizen elite have favored a close link with Europe. This position was apparently also true of most citizens as a majority of them voted in favor of EU membership in the 1994 referendum.

At the end of the 1960s and early 19705, however, the Swedish population showed little interest in the EC issue. This was especially evident in 1968 and 1973, when a large percentage of respondents were unable to give an opinion for or against membership. Nonetheless, in each year with available data, except for 1988, those opposed to EC membership were, as hypothesized, more distrusting of government (see Table 4). As in Norway, the shift in ECrelated attitudes between 1988 and 1991 partly explains the Swedish change in trust. Between 1988 and 1991, the proportion of Swedes favoring EC membership increased to 54 percent, and for this segment of the population trust remained constant. But among those who either opposed EC membership or wanted to maintain the status quo (labeled pro/con), distrust grew dramatically, which partially accounts for the continued downward spiral of trust.

While Vietnam, NATO and even the EC issue have had intermittent effects on Scandinavian public opinion, the question of aid to underdeveloped countries has been both more enduring and divisive. This issue emerged in the 1950s as the governments of Norway and Sweden initiated programs to transfer capital and expertise to third world countries. Public opinion on these matters has been quite supportive of government policy (note that in Tables 3 and 4 large majorities favor foreign aid). Moreover, the issue was related to cynicism in both countries, with those holding negative views on foreign aid as the most cynical.⁵ Nevertheless, an examination of the joint trends is instructive.

From 1968 to 1973, Swedish opposition to foreign aid grew from 13 percent to 17 percent, while distrust remained fairly constant (PDIs moved from -25 to -28). From 1982 to 1991, Swedish anti-foreign aid sentiments grew only gradually, but the continued increases in distrust were far more evident among those opposed to foreign aid (see Table 4). In short, a continued drift away from support for other countries offers a partial explanation for the long-term decline in political trust for Sweden (the time series correlation between support for foreign aid and trust is .88).

In Norway, the change in public opinion on aid to underdeveloped countries similarly helps explain the trend for trust in that country. From the 1950s to the early 19705, support for foreign aid increased (Ringdal 1977). Then between 1973 and 1977, counterbalancing changes occurred (opposition to foreign aid decreased from 14 percent to 5 percent, but at the same time support for the present policy on foreign aid fell from 43 percent to 36 percent), which makes it difficult to determine if overall opinion became more positive or negative (see Table 3). From 1981 to 1985, public opinion became clearly more supportive of foreign aid, whereas negative sentiments increased considerably from 1985 to 1989, resulting in twice as much support for cuts as for increased foreign aid (26 percent vs. 13 percent). As in Sweden, Norwegians opposed to foreign aid were generally more cynical than those

who supported foreign aid, and the change in attitudes toward foreign aid does correlate with the dynamics of political trust (time series correlation of .62). Yet, more Norwegians than Swedes remain committed to foreign aid, which may partly explains the higher level of political trust in Norway.

Race Relations

Governments in democratic systems are regularly called upon to protect the rights of minorities and eliminate discrimination through legal means. However, members of majority or dominant groups often fail to perceive that discrimination against minorities exists and thus view these government actions as unnecessary or as providing special favors to undeserving minorities (Edsall & Edsall 1992). Because of the government's normative role in balancing the interests of majority and minority elements in society, it is possible that neither set of interests are fully met. This certainly appears to be the case with US race relations.

Considerable evidence on attitudes toward blacks show marked shifts in a positive direction across time (Schuman, Steeh & Bobo 1985). For example, the average rating of civil rights leaders (using a scale from 0 as most negative to 100 as most positive and 50 as neutral) increased from 41 in 1972 to 57 in 1992. On the surface this, coupled with strong government involvement, might suggest that those favoring an activist position on civil rights would be the least alienated, while those opposed to government involvement would be most discontented. Nevertheless, this hypothesized correlation is not supported by the data.

The relationship between racial attitudes and trust was examined by employing a two-item index of opinions on the changing situation of blacks in America and perceptions of the speed with which civil rights for minorities were changing. Figure 2 demonstrates that those who wanted improvements in civil rights to occur at a faster pace as well as those who felt that change in this policy area was already going too fast were more distrusting than those who were satisfied with the current situation. Prior to 1976, those who demanded rapid change in the civil rights area (labeled "left" in Figure 2) were consistently more alienated than those who favored a slower pace (labeled "right"). As of 1976, however, those who favored a more conservative policy became as alienated as the liberals, even during Reagan's presidency despite his conservative civil rights record.

What accounts for the curvilinear form of this relationship and the shift among conservatives after 1976? Previous research on racism in America has argued that during the 1970s and 1980s a new form of racism emerged (Sears 1988; Kinder & Sears 1981; McConahay 1986). This "modern racism" is subtle and lacks conspicuous displays of discrimination. In contrast to traditional

racism, it makes no direct mention of minorities as inherently inferior
or socially undesirable and thus deserving of legal segregation or regulation.

In fact, modern racism is a conservative ideology that extends well beyond the issues of civil rights. The new racism emphasizes traditional values such as loyalty to the larger community and patriotism, family and parental responsibility, religion and rules, work ethic and social stability. People who express this orientation are opposed to special treatment for minorities. They are likely to believe that discrimination is a past phenomenon, and are very likely to attribute success or failure in life to personal responsibility. They fear a growth in drug use, crime and number of welfare recipients in the country. On the other hand, there is the liberal-oriented ideology of those who champion the rights of individuals and promote programs to compensate for past discrimination against certain minority groups.

The data suggest that, despite real improvements in racial attitudes, both liberals and conservatives have shown a good deal of impatience in recent years concerning civil rights. Blacks perceived considerable change in their situation as of 1964 when more than half said their social position had changed "a lot." By 1992, however, only 17 percent of blacks felt that their situation was improving, and nearly half of the respondents believed that the pace of civil rights was too slow. There were similar signs of growing impatience among whites. In 1992, the percentage who saw "a lot" of change was lower than it had been a quarter of a century earlier, despite a slight increase between 1964 and 1972. Yet, nearly a third of all whites believed that blacks were pushing too fast for improvements in civil rights, and these individuals displayed the greatest propensity toward alienation.

When viewed across time, aggregate shifts in attitudes on civil rights were strongly correlated with the trend in political trust. The aggregate correlation between the two variables using all eleven data points for which the measures were available during the period of 1964-1992 was .90 for whites and .73 for blacks. Consequently, racial attitudes in America appear to have played a role in the long-term decline in political trust.

A similar impact of racial attitudes on alienation can also be found in Norway and Sweden, although the form of the relationship differs from that observed for the United States. While the historical ethnic homogeneity of Scandinavia might suggest an absence of racial conflict in Norway and Sweden, the issue of race relations has become more and more salient as the immigration of colored workers and refugees into Norway and Sweden has increased.⁶

Immigration has fueled racial antagonism in both countries with the general public displaying more negative attitudes toward immigration than was reflected by government policy and the political elite. Anti-immigration sentiments have been linked to popular perceptions that immigrants raise the crime rate, create civil unrest, and receive more generous welfare benefits than non-immigrants.

Attitude toward immigration policy was measured in both countries with an agree/disagree statement asking if immigrants should receive increased governmental economic support for protection of their culture.⁷ This item was included in all Swedish studies from 1979 onwards and in the Norwegian studies of 1985, 1989 and 1993. The Swedish public was almost evenly divided on economic support for immigrants in 1979, but the preponderance of public opinion moved strongly against supporting immigrants in successive years (see Table 5). Similarly, in Norway, the opinion balance jumped from a point differential of 10 in 1985 to a difference of 53 percentage points in 1989 and moderating to 15 points in 1993. These considerable shifts concur with the overall trend toward cynicism for both countries. The aggregate time series correlation between attitude toward immigration and trust was .67 in Norway and .84 Sweden. Despite the relatively limited number of time points with available data, the magnitude of these correlations suggests that opinion change on the immigration issue contributed significantly to the development and change in political cynicism. Moreover, the cross-section correlations are fairly strong and linear in each year, with those holding negative attitudes toward immigrants as the most cynical (see Table 5).

Moral Issues

Moral issues, including family values, tolerance of gay lifestyles and abortion, have played a major role in the politics of all three countries in recent decades, particularly in the United States. Unfortunately, the only moral issue with available indicators over time is abortion. Nevertheless, abortion has been a salient and controversial issue in all three countries, especially since the early to mid-1970s when laws protecting freedom of choice were first introduced. With the Supreme Court upholding Roe v. Wade in 1973, the United States became the first of the three countries to give women a legal right to abortion. Sweden followed shortly after, adopting free choice in 1974, while Norway did not legalize free choice until 1977.⁸

Generally, attitudes on abortion are similarly linked to cynicism in all three countries - those strongly opposed to abortion and government policy were consistently the most alienated (see Table 6). The alienation of pro-lifers reflects their support of a position that, despite the rhetoric of many politicians, has not been strongly supported by any government, whether the incumbent leaders were of the left, center or the right. The relationship between abortion and cynicism in Norway is particularly fascinating because it provides insight into the dynamics of how issues and distrust become linked. In 1973, prior to the adoption of a specific policy on abortion, there was no relationship between abortion preferences and cynicism. After the full implementation of free choice in 1977, Norwegian public opinion moved to

accept the new law (the pro-choice percentage rose from 28 percent in 1973 to 40 percent in 1977 and 60 percent by 1981). At the same time, distrust increased for the group that was most strongly opposed to the new policy (the PDI dropped from 35 in 1973 to 21 in 1981 and 6 in 1993), a case where the outcome of a particular issue controversy, in the form of newly implemented public policy, alienated the losing side. Nevertheless, in this specific case, the process does not contribute much to explaining trends in trust over time for Norway and the US, because the size of the alienated group was quite small and remained relatively stable across time (roughly 10 percent in the United States and 2 percent in Norway). Moreover, as Table 6 reveals, the relationship across time between abortion attitudes in Sweden and the United States has decreased. Thus, although the abortion issue has been very divisive in the United States, it does not appear to have contributed much to the change in political trust among the three countries.

Multivariate Analysis

The bivariate, longitudinal analysis presented above is very instructive. Nevertheless, the observed correlations between issue preference and distrust may be spurious, simply reflecting demographic or partisan differences associated with policy preferences. Furthermore, the bivariate analysis does

not allow us to sort out the independent effects of the various issue areas. Since a citizen's policy preference orientation may be consistently liberal or conservative from issue to issue, it is necessary to employ multivariate analysis to gauge the relative importance of each issue.

However, the curvilinearity observed for the relationship between trust and several of the issue measures for the United States raises questions about what type of multivariate analysis is appropriate. Monotonic relationships were consistently found for Norway and Sweden, but not the United States. Indeed, when a regression analysis was computed for the United States with all variables coded in a linear fashion from low to high, military spending was the only issue that approached significance. On the other hand, repeating the same analysis with each issue measure recoded into a pair of dummy variables - indicating either a more liberal or more conservative preference relative to the center - produced more meaningful results and a significant increase in the explained variance. Given this outcome, the discussion of the multivariate analysis for the United States will focus only on the dummy variable regression while those for Norway and Sweden will not.

The results of the multivariate analyses, in which we have entered the available issues for each year along with demographic and partisan control variables, summarize and confirm the earlier bivariate findings. In Norway, aid to underdeveloped countries is the best overall predictor of political cynicism (see Table 7). With the exception of 1985, the beta hovers around .20 for this variable. The somewhat weaker impact in 1985 can be attributed to the fact that the non-socialist parties were in power at that election, which possibly reduced some of the policy dissatisfaction among those who displayed negative attitudes toward foreign aid. The impact of economic support for immigrants did not reach statistical significance in 1985, but in 1989 and 1993 the item was clearly related to alienation. This change should be read in conjunction with the sharp increase in immigration to Norway between 1985 and 1989, and the increased policy polarization on the immigration issue during that period. The NATO issue is the only example where a leftist policy preference was linked to cynicism, as those who wanted Norway to withdraw from NATO were more cynical at all time points with the exception of 1973. The consistent alienation of the small anti-NATO minority reflects the political seclusion of a group whose policy position throughout the post-war period has been contrary to what all governments - left, center or right - during that time have pursued. The regression models corroborate the conclusion that the EC issue, at least for the five national elections following the referendum, did not contribute to political distrust in the Norwegian electorate. Contrary to the pattern of the EC question, policy position on abortion had a weak, but statistically significant, impact on trust in the years after the liberalization of the law as shown by the cynical position of anti-abortion advocates (but note that 1985 and 1993 are exceptions).

The results of the multivariate analysis for Sweden are very similar to those found for Norway. Aid to underdeveloped countries and for immigration are the strongest policy predictors of political cynicism (see Table 8). In the years when both questions were asked, foreign aid was more strongly linked to distrust than immigration. In 1982 and 1985, the betas are about twice as large for foreign aid as for immigration. In 1988, however, the relative weight of immigration policy approaches that of foreign aid, suggesting that immigration had become more salient in Sweden. Of the remaining foreign policy issues, Vietnam gained significance only in 1973, and when discussions of an upcoming EC vote for Sweden started to heat up, the EC question became statistically significant in 1991. Moral issues, however, were not related to long-term distrust in Sweden.

For the United States, the analysis clearly demonstrates that civil rights issues played an important and continuous role in determining citizen attitudes toward government (see Table 9). In the early 19705, those who felt civil rights were progressing too slowly were particularly alienated, but by the late 19705, an apparent backlash effect emerged as the coefficients for those on the right equaled and often surpassed the magnitude of those on the left. In general, those who opposed US foreign involvement were also consistently more cynical (see coefficients for the involvement item in Table 9). Yet, this general orientation did not overshadow the independent effect that Vietnam had in alienating those who demanded an end to the war. Likewise, dissatisfaction with military spending had an enduring effect on trust in government (with the exception of 1980), but again primarily in alienating those on the left. Across the years, hawks were never significantly more alienated than the remainder of the population (see Table 9). Given that the US government has generally pursued a policy of active foreign involvement and regular increases in military spending, it is not too surprising that the staunchest supporters of these policies were not significantly more discontented than the rest of the population. As for moral issues, position on abortion was significantly related to trust only in 1976 when both pro-choice and pro-life advocates were more alienated.

In short, the regression analysis demonstrates considerable similarities and yet some differences across the three countries. The regressions reveal a limited impact of moral issues on trust in all three countries, and those opposed to foreign involvement, directly or through aid, tend to be more alienated regardless of country. When we move to racial issues, however, some differences emerge. It is in this area that the results for the US are most nonmonotonic. Discontent among those who favored a conservative race relations policy had an increasing impact on trust toward the end of the 1980s in all three countries. Yet, in the US, unlike Scandinavia, civil rights liberals also felt that government actions were less than satisfying. Setting aside differences in the form of the relationship (linear versus non-linear), the

evidence demonstrates the powerful impact that racial cleavages can have on
citizen trust in government regardless of the type of political system.

Conclusion

The evidence and analysis presented in this article reconfirm the findings of previous research on the United States that hypothesized a relationship between policy dissatisfaction and political distrust. The results demonstrate that when discontent with the policies pursued by the government endures across time, trust decreases and the potential for alienation increases. Under these circumstances, evaluations of the content of decisions as opposed to evaluations of procedures that lead to policies influence the development of support or alienation (Lipset 1994).

In this particular case, we have documented two broad issue areas - foreign policy and race relations - with long-term effects on changing levels of political trust in all three countries examined. Although Norway and Sweden are much smaller countries than the United States, evaluations of government involvement in foreign affairs still played a significant role in how citizens perceived the trustworthiness of their government. These foreign policy assessments may very well have reflected moral judgments of how citizens felt their government ought to participate in the global political arena. But toward the end of the 1980s and early 19905, these evaluations also represented a growing feeling that there were real problems at home that took precedence over foreign affairs. Future research should address some of these other domestic issues, particularly the preferred level of spending on welfare programs.

For Scandinavia, race relations is a relatively new problem area which could become even more divisive, especially during periods of economic recession, as the number of immigrants increases. In the United States, the Los Angeles riots of May 1992 were a vivid reminder of the tremendous difference in quality of life between whites and blacks, and yet many people believe that discrimination is a thing of the past. In such a divided society it is difficult for the government to shape policies that will satisfy the competing sides.

That moral issues did not have a major impact on declining trust in government coincides with previous findings (Miller & Listhaug 1993). New social issues such as abortion have had little effect on shifting attitudes toward the political system. While these issues have certainly been controversial, especially in the United States, they are either too new to influence diffuse support, or they have not dissatisfied large enough segments of the population to significantly affect overall political alienation.

Another major finding is the different forms of relationship between policy
preference and trust in Scandinavia and the United States. The nonmonotonic

relation for the US versus the linear association for Norway and Sweden emphasizes the impact of different types of social and political systems on the form of this relationship. The greater diversity of cleavages in the United States, the constitutional division of powers, the two broad-based parties and the panoply of interest groups produce a system bent on managing conflict through centrist policies. In such a society, when issue polarization exists, enduring policy discontent is likely to alienate those who seek change as well as those who wish to return to a more traditional approach. On the other hånd, Norway and Sweden are smaller, more homogeneous societies that have been politically dominated by social democratic parties for much of the post-war period. The task of these governments, given the corporatist system in Scandinavia, has been consensus management rather than conflict resolution. Such a system is more likely to create linear correlations between policy orientation and alienation in which those who prefer a liberal policy are most supportive of the government and the system.

But although the relationship between policy preferences and alienation is similar in the three countries, there may be different explanations for why the association develops. One explanation involves a self-interest interpretation, i.e., the policy is perceived as helping others (groups or individuals) and hurting me (House & Mason 1975,125). Others have argued that the outcome does not necessarily have to affect the individual directly to be perceived as undesirable (Tyler 1984; Rasinski 1987). If an individual perceives a series of government decisions as unfair, he/she may eventually believe that the rules of the political game and the avenues of institutional decision making are inherently biased and incapable of fair decisions. Another interpretation of the correlation between policy dissatisfaction and alienation arises from the belief that government is simply inefficient and thus ineffective (Sundquist 1980). According to this view, government is simply not capable of producing outputs that will effectively solve problems. Eventually, government itself is perceived as untrustworthy.

Each of these interpretations specifies an alternative theory that may account for the observed correlation between policy preferences and political trust. An empirical examination of these alternative theories is, however, beyond the scope of this article, Nevertheless, investigating these alternatives should be the goal of future research agendas because the relative importance of each of these interpretations holds both theoretical and practical political implications for the development of democratic political systems.

---

NOTES

1. In this article we use "alienation" and "cynicism" synonymously as opposites of "trust."

2. The Norwegian surveys were carried out by Henry Valen and Bernt Aardal of the Institute for Social Research, Oslo, the Swedish studies by Soren Holmberg and colleagues at the Department of Political Science, University of Gothenburg, and the American surveys by the Center for Political Studies, University of Michigan. The data were made available by the Norwegian Social Science Data Services, the Swedish Social Science Data Service, and the Inter-University Consortium for Political and Social Research.

3. See Appendix for a description of the specific measures of diffuse support used in the analysis reported here. For convenience, in the remainder of the article the indices formed with these items will be referred to as measuring levels of public trust in the political regime.

4. Our choice of both the broader issue areas and the more specific policy questions used in the analysis reflects a consideration of three sources of information suggesting what major issues of the time might have influenced public attitudes toward government. These sources included: (1) a review of political science and sociology literature linking policy preferences and trust in government; (2) a review of the headlines from major newspapers during the historical period covered in the analysis (in the US): and (3) the array of issue questions included in the surveys from which the empirical evidence was drawn. These three sources of information revealed that issues of foreign policy, race relations and moral principles were important in all three countries during the quarter century under investigation. Other specific issues were also important during this period, but they were either not relevant to all three countries or were simply not available in all the surveys.

5. For Norway it is interesting that those who favored increased aid were as trusting as those who favored the status quo. This can be explained by the faet that for much of the period government policy was to increase aid, thereby making the policy satisfaction relatively equal for both categories.

6. Hernes & Knudsen (1990, 47) report two recent major immigration waves in Norway and Sweden. The first wave occurred in the early 1970s and the second started in the mid-1980s. The first immigration wave primarily reflected economic growth. and numerous guest workers came to Scandinavia. The second wave was fueled by growing political problems in Eastern Europe, Africa and Asia.

7. The 1993 question asked in Norway was slightly different from the one used in 1985 and 1989 when the question referred to economic support for immigrants to protect their culture. The 1993 item referred to economic support so immigrants could teach their own language. Also, the 1976 Swedish question referred to stopping immigration rather than support for immigrants, so it was reversed to correspond to the other items.

8. Lcgalization of abortion in Norway actually occurred in two steps. In 1975, the procedure for requesting an abortion was liberalized, and in 1977 full legal rights were extended.

ACKNOWLEDGMENTS

This research was partially funded by a grant from The Norwegian Marshall Fund. We wish to
thank Chia-Hsing Lu, Andy Peebler, Svein Åge Relling and Roar Håskjold for research
assistance, Gwyn Erb for editing the manuscript, and Peggy Swails for secretarial assistance.

REFERENCES

Almond, G. & Verba, S. 1965. Civic Culture. Boston, MA: Little Brown.

Bachman, J. G. & Jennings, M. K. 1975. "The Impact of Vietnam on Trust in Government."
Journal of Social Issues 31, 141-55.

Batch, G. I. 1974. "Multiple Indicators in Survey Research: The Concept 'Sense of Political
Efficacy,'" Political Methodology 1, I^3.

Citrin, J. 1974. "The Political Relevance of Trust in Government." Amena™ Political Science
Review 68, 973-88.

Citrin, J. & Green, D. 1986. "Presidential Leadership and Trust in Government." British
Journal of Political Science 16, 431-53.

Converse, P. E. 1972. "Change in the American Electorate," in Campbell, A. & Converse, P. E.
eds., The Human Meaning of Social Change. New York, NY: Russell Sage.

Crozier, M. J., Huntington, S. P. & Watanuki, J. 1975. The Crisis of Democracy. New York:
New York University Press.

Dahl, R., ed. 1966. Political Oppositions in Western Democracies. New Haven, CT: Yale
University Press.

Easton, D. 1965. A System's Analysis of Political Life. New York, NY: Wiley.

Hernes, G. & Knudsen, K. 1990. Svart på hvitt. Osco: FAFO.

House, J. S. & Mason, W. 1975. "Political Alienation in America, 1952-1968," American
Sociological Review 40, 123-47.

Kinder, D. R. & Sears, D. O. 1981. "Predjudice and Politics: Symbolic Racism Versus Racial
Threats to the Good Life," Journal of Personality and Social Psychology 40, 414-31.

Lipset, S. M. 1994. "The Social Requisites of Democracy Revisited," American Sociological
Review 59, 1-22.

Lipset, S. M. & Schneider, W. 1983. The Confidence Gap: Business, Labor and Government in
the Public Mind. New York, NY: Free Press.

Markus, G. B. 1979. "The Political Environment and the Dynamics of Public Attitudes: A
Panel Study," American Journal of Political Science 23, 338-59.

McConahay, J. B. 1986. "Modern Racism, Ambivalence and the Modern Racism Scale," in
Dovidio, J. F. & Baertnev, S. L., ed., Prejudice, Discrimination, and Racism. Orlando, FL:
Academic Press.

Miller, A. H. 1974. "Political Issues and Trust in Government: 1964-1970," American
Political Science Review 68, 951-72.

Miller, A. H. & Borrelli, S. A. 1991. "Confidence in Government During the 1980s," American
Politics Quarterly 19, 147-73.

Miller, A. H. & Listhaug, O. 1990. "Political Parties and Confidence in Government: A
Comparison of Norway, Sweden and The United States," British Journal of Political
Science 20, 357-86.

Miller, A. H. & Listhaug, O. 1993. "Ideology and Political Alienation," Scandinavian Political
Studies 16, 167-92.

Rasinski, K. A. 1987. "What's Fair is Fair or is it? Value Differences Underlying Public Views
About Social Justice," Journal of Personality and Social Psychology 53, 201-11.

Ringdal, K. 1977. "Folkemeininga on utviklingshjelp," Tiddskrift for samfiinnsforskning 18,
47-75.

Risse-Kappen, T. 1991. "Public Opinion, Domestic Structure, and Foreign Policy in Liberal
Democracies," World Politics 43, 479-512.

Rogowski, R. 1993. "Comparative Politics" in Finifter, A. W., ed., Political Science: The State
of the Discipline 11. Washington, D.C.: American Political Science Association.

Schuman, H., Steeh, C. & Bobo, L. 1985. Racial Attitudes in America: Trends and
Interpretations. Cambridge, MA: Harvard University Press.

Sears, D. O. 1988. "Symbolic Racism," in Katz, P. A. & Taylor, D., eds., Toward the
Elimination of Racism: Profiles in Controversy. New York, NY: Plenum.

Sundquist, J. L. 1980. "The Crisis of Competence in Our National Government," Political
Science Quarterly 95, 183-208.

Tyler, T. R. 1984. "Justice in the Political Arena," in Folger, R., ed., In the Sense of Injustice.
New York, NY: Plenum.

Tyler, T. R., Rasinski, K. A. & McGraw, K. M. 1985. "The Influence of Perceived Injustice on
the Endorsement of Political Leaders," Journal of Applied Social Psychology 15, 700-25.

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