Resumé

The transition from totalitarian to democratic rule in the Baltic states raises the question of conditions for regime legitimacy and stability. The article focuses on the level of confidence people have in institutions after change of regime. The confidence in political and social institutions is at a surprisingly low level. The data suggest that people have more confidence in institutions producing symbols than they have in policy-making and implementing institutions. There is also evidence to show that leaders are more trusted than institutions as such. A survey of the potential background variables shows that people with higher education have lower confidence in institutions than the less well educated. But low confidence in institutions does not necessarily spell gloomy prospects for procedural democratic development, as long as the elites do have some popular support and the capacity for consensual integration.

Introduction

The purpose of this report is to investigate the level of confidence in institutions of the new democratic states of Estonia, Latvia and Lithuania. Effective democratic government requires popular support in order to perform vital functions. Trust in basic political and social institutions is regarded as a prerequisite of legitimate democracy. Accordingly, a low level of confidence may constitute a threat to the stability of political and social institutions.¹ Experience shows that confidence in political institutions correlates with viable democracy (Almond & Verba 1963; Dahl 1971; Inglehart 1990). The prospects of institutional survival in the new democracies after the breakdown of the Soviet Union will be greater if they are supported by appropriate public attitudes (Finifter & Mickiewicz 1992).

Trust in institutions is of special concern in newly established democratic states recently transformed from totalitarian regimes. It is sometimes taken for granted that the introduction of democracy and market economy will result in positive sentiments towards the new institutions among the people at large. In this study the opposite is found: people in Estonia, Latvia and

Lithuania have little confidence in political, social and economic institutions in comparison to Western countries. The level of confidence is particularly low among highly educated people or what may be referred to as "the social elite". After having discussed and operationalized the concept of "institution", we will approach confidence from three perspectives:

How is confidence related to type of institution?

What influence can an elitist political culture have on people's
confidence?

What kind of relationship exists between social background and
confidence?

Confidence is expected to vary with type of institution, depending on whether the institution has a symbolic function, does policy performance or is associated with its leader. Symbolic institutions not directly affecting the material life of the people will be compared with policy performing institutions. If confidence in institutions is related to institutional output, we will expect popular discontent with living conditions to be transformed into distrust of the political institutions responsible for policy outputs. Confidence may also vary with the age of the institution and its character as a public or private institution.

Leaders play an extremely important role in new democracies. Though lacking democratic experience, elites from the former regime suddenly became involved in a system accountable to popular criticism. Here we are asking if the massive lack of confidence in institutions is reflected in a comparable distrust of the leaders of the institutions.

Popular support of institutions is of special importance in newly established democracies. We will ask how confidence in institutions is related to social background and cleavages. Educational stratification and ethnic cleavages in particular are expected to influence confidence. The well educated may be better informed than others and as such particularly critical of the institutions. Yet another possibility is that they feel more committed than others to democracy and therefore end up with a high level of confidence. The high proportion of non-indigenous persons living in the Baltic states and the revival of national and ethnic sentiments throughout the region render national integration difficult. We expect the confidence level among Russians and other non-indigenous groups to be on the low side.

Level of trust is relative. The Baltic states will therefore be compared with two established democracies: Norway and France. Norway is in many respects similar to the other Scandinavian countries. It is characterized by an egalitarian political culture with a high degree of equality between social groups and close interaction between state and society. Several observers have maintained that the development of the Scandinavian welfare state is

the result of strong social democratic movements promoting political participation from below. France is chosen as a point of reference because of its strongly elitist tradition, its large income differences between social groups, the closed political process and the well-known system of selective recruitment of political and administrative elites (Smith 1972; Hayward 1982).

Egalitarianism and elitism are expected to have opposite impacts on popular support for political and social institutions. To the extent that confidence in institutions is higher in egalitarian societies than in elitedominated systems, we would be inclined to expect Norway to come out on top, followed by France and the three Baltic countries, in that order.

The article relies on data from Baltic surveys in November-December
1992 and the EVSSG Value Study in France and Norway as of 1990.2

Institutions as "Symbols", "Policy Performers" and "Leaders"

The data do not contribute to simple conclusions about underlying causes behind high versus low levels of confidence in institutions. In all likelihood confidence stems from a combination of sources - values connected with symbols, policy performance and/or the special qualities of the leaders - which make themselves felt with varying magnitude depending on the social and political context. In some cases we will expect one element to be predominant, e.g. the symbolic aspect of the Church. Trust in parliament will be related to all three aspects. In times of economic crisis, policy performance is expected to be of paramount importance. In the Baltic countries which do not have much by way of democratic institutional traditions the leaders are likely to be held personally responsible for the way in which the institutions work. Easton (1975) and Gabriel (1992) assert that confidence may come from institutions as such, irrespective of policy outputs and elite persons. In the Estonian paradigm confidence is understood as a fundamental and rather diffuse attitude. Confidence in institutions occurs when institutions or basic processes are taking care of the fundamental interests of the people. In this sense, trust has more to do with general confidence in political symbols than with policy products or elite persons.

Presupposing that people are rational and act in response to the contribution by institutions to their well-being, one would expect the outcome of institutional activities to determine the pattern of attitudes among people. According to Barry (1970) it is the policy performance of democratic institutions over a period of time that gives legitimacy to the political system. Therefore, as Smith (1972, 9) asserts: "value-related explanations take on a subordinate role" for legitimacy and support.

If trust arises from the ability of institutions to produce positive policy outputs, confidence will depend upon how the outputs from institutions meet the expectations of the general public. Low confidence will not stem from distrust in institutions as such, but from poor performance. Economic recession will reduce belief in institutions, but it will not necessarily lead to a desire to change them.

In the elite-oriented political cultures prevalent in the post-communist states, it is also possible for trust in leader persons to make up for institutional distrust. The prominent position of elite persons in a transitional regime without intermediating organizations makes trust in institutional elites into a particularly important stability factor. For newly erected institutions where experience with policy performance is low, for example elected parliaments, the elite persons will not only be representatives of institutions, but are likely to be regarded as the institution itself.

Belief in democratic institutions depends on symbolic attractiveness, the general performance of the system and the ability of the leaders to bring about consensus when confronted with the interests of the various social groups. The crucial question is how different strata of the population react to institutions, and how dependent the authorities are on support from different groups for political stability.

One central question is which group will be in the position to destabilize or overthrow the system by withdrawing its support. In this article the focus is put upon the role of the social elites. It is assumed that trust in political institutions and leaders from this group are of special importance for political stability.

De-Entangling the Concept of "Institution"

Institutions can be classified according to different criteria. Here we discuss
some basic traits of institutions and how differences in properties may
influence confidence.

The distinction between public and private or governmental and nongovernmental is often used in the study of Western countries (Rose 1984). The potential for support for new private institutions resulting from a market economy will probably be large until the positive symbolic content is emptied by harsh realities. Listhaug & Wiberg (1992) propose an empirically reductionist approach and differentiate between order institutions (such as police and army) and others. Here we emphasize that several additional dimensions are relevant to understanding differences between institutions and countries. Additional dimensions are of special importance in transitional regimes, where "public-private" and "order" are put in other contexts than in established democracies.

It may be argued that the "freshness" of institutions and the quality of policy performance in transitional regimes are of special relevance for people's confidence. The mix of old and new institutions constitutes a problem of particular relevance for transitional regimes. The reason for and the driving force leading to the change of regimes in the Baltic states were the establishment of liberal rights, democratic institutions and a market economy. On the other hand, transition of regime means that several former institutions and their personnel have to be preserved in order to carry on vital functions. It is therefore useful to distinguish between institutions carried over from the former regime and new institutions. Old institutions are likely to be associated with the malfunctioning of the former regime; and we therefore hypothesize that new institutions will enjoy higher confidence than the old ones.

Establishing democratic states means not only creating new institutions, like elected parliaments and political parties; it also implies the continuation of old ones, such as the state bureaucracy, and the granting of autonomy to institutions which were previously dependent on the state, like the Church and trade unions. New institutions lack tradition, while old institutions have an ideological heritage to overcome, and we would therefore expect the new institutions to inspire more confidence than institutions representing a continuation of the past. For example, state bureaucracy will have less support than the new political parties. This hypothesis leans on the assumption that support for institutions springs from perceived differences in properties of the institutions under the past and current regimes.

Here we will propose a distinction between policy-making and implementing institutions. Implementing institutions, like the bureaucracy and the police, are to a large extent preserved from the old regime. Policy-making institutions will be new and a byproduct of the transition from one regime to another. The actual scope of transformation will be a function of a number of factors such as uniqueness, change of personnel and change in output. Policy-makers are responsible for decisions made and dependent on regular support because of re-election. Policy-implementers are more in direct touch with the public, but are not responsible for the consequences of their decisions as long as they are in accordance with the instructions and laws given by the policy-makers. We would be inclined to assume that implementing institutions enjoy less confidence than policy-making institutions on the ground that they represent continuity rather than system change. Institutional products are expected to influence attitudes. The Church - which is a producer of diffuse and symbolic goods - is supposed to influence attitudes in quite other ways than institutions producing more specific and tangible goods that affect people's lives in a very direct way, e.g. level of pensions. A third distinction is drawn with respect to institutions producing positional goods, like schools and universities.

It may be hypothesized that in times of transition and economic crisis, institutions like the bureaucracy and market, which are expected to produce specific benefits, will be more open to distrust than institutions like the Church, the national army and the educational institutions which produce less tangible, more long-term and symbolic benefits.

Institutions producing support and demands for the decision-makers constitute yet another category. Political parties, movements and trade unions belong to this category. Some have deep roots into the past, like the trade unions, but the vast majority of the political parties are new.

Opinion-producing institutions like the press perform more symbolic and
diffuse functions and will therefore gain more support than institutions
producing more tangible goods.

The legal system is somewhat ambiguous. Previously, the courts were nothing but the prolonged arm of the power-holders; nowadays the courts are supposed to be independent of the state. The level of trust will have to do with the integrity and neutrality of the individual judges and with the effectiveness of the legal system in coping with crimes.

Empirical Results³

As for the Baltic states, we would expect people to have more confidence in:

new institutions rather than old.

private rather than public institutions

institutions producing symbolic and diffuse outputs rather than specific
outputs

Since differences between ethnic groups for most institutions are marginal, this information is not included in the table. However, regarding confidence in national forces and the press, indigenous people in all three countries have a higher trust level than others, and the opposite is the case for confidence in Russian forces.

Of all the institutions in the Baltic countries, the Church stands out as the most trusted. This may be seen as a byproduct of fact that the Church was thrust into a position of prominence by the most recent independence movement. It has strong ties with the earlier period of national independence and produces diffuse and symbolic outputs which satisfy basic emotional needs at a time when there is a lack of authority. Contrary to what might be expected, the confidence in the Church is no higher in Catholic Lithuania than in predominantly Protestant Estonia and Latvia. The early and close relationship between the Catholic Lithuanian Church and the anti-system opposition against communism (Vardys 1987) have not resulted in higher popularity ratings than the scores obtained by the Church in predominantly

Lutheran Estonia and Latvia with their weak ties to nationalism during the Soviet period. In Norway and France confidence in the Church is not very high. In Norway the Church is in fact among the institutions enjoying the least confidence. France is more similar to the Baltic countries.

The educational system has a high score in all Baltic states. It produces positional, rather than direct, goods which may add up to long-term benefits for the individual citizens. The educational systems in Norway and France have considerably higher scores than the educational systems of the three Baltic countries. But the Norwegians have more confidence in their educational system than the French in theirs.

The national armed forces of the Baltic states constitute important symbols of national independence, unity and historical continuity. They can also be said to represent symbols satisfying national feeling. Not surprisingly, the Baltic peoples have less confidence in the Russian army, which had a considerable amount of troops in Estonia and Latvia at the time of the survey. However, 14-15 percent of the populations, including the substantial Russian speaking minorities in these two countries, say they trust the Russian armed forces. The Baltic forces are still much more of a symbol than a real force. The Norwegian armed forces have the best trust rating of the five armed forces covered by this study, while the French armed forces land on a level not much different from that of the three Baltic countries.

The press gets an average score of 45 percent with Estonia leading the way. The press is mainly privately owned and has a high degree of credibility among people, also in comparison to Western standards. The main point here is that its products are diffuse and affect the lives of the citizens only indirectly. The mobilization of national sentiments through the press during the process of liberalization may go a long way towards accounting for the high level of support.

As for decision-making institutions, the parliament and government have an average of only 28 percent confidence. These are new institutions, in the sense that they are democratically elected. Despite the fact that they constitute the very core in a democratic state, confidence in these institutions is low. This finding may be interpreted as a byproduct of the fact that these institutions either make specific decisions affecting people's daily lives in a direct way or cannot make any decision at all. In either case, it is the parliament and the government which receive the criticism for most of what goes wrong.

The level of confidence is very low in the Baltic countries compared with the trust ratings for parliaments in France (48 percent) and Norway (59 percent). At first glance, this stands out as a rather surprising finding, considering the quick transition from a planned to a market economy and from totalitarianism to liberal democracy. On the other hand, the figures are

not that surprising when one considers the economic crisis and the difficult
decisions politicians have to make.

The position of France between the Baltic and Norway supports our
contention that elitism reduces the legitimacy of parliament. Egalitarian
values tend to support trust between deputies and electorate.

The police and courts are "order" institutions. The police has a confidence score of 27 and the courts a meagre 22 percent in the Baltic countries. In Norway it is 88 percent for the police and 75 percent for the legal system. With scores between 67 percent and 50 percent, France is somewhere inbetween Norway and the Baltic countries. The extremely poor scores in the Baltic states may be a byproduct of the fact that the police and court system are old institutions, associated with the former regime. Their institutional products are also poor; they are having difficulties coping with the soaring crime rate. It seems easier to trust "order-institutions" in established democracies, and to a considerably larger extent in egalitarian than in elitist states.

The main functions of the civil service are state continuity and implementation of policy decisions taken in parliament and government. With an average score of 24 percent, the Baltic countries rank very low on confidence. This is also the case for local government. The low score obviously has to do, first, with the poor performance of public service production and, secondly, with the heritage from the past regime and with organizational continuity, combined with preservation of old style bureaucratic routines and behaviour. In Norway and France, bureaucracy receives one of the lowest confidence ratings, but it is considerably higher than in the Baltic. The relatively low level in Norway may be seen as a byproduct of the traditional scepticism towards a remote bureaucracy seemingly beyond popular control. The somewhat higher rate for France is probably related to the traditional position of the French administrative system accepted as an integral part of a "dirigiste" state machinery.

Business companies are new institutions in the Baltic countries, representing the very idea of transition from plan to a market. Yet, with an average of only 25 percent, the level of confidence in business companies is surprisingly low. This indicates that novelty and privateness of institution are no guarantee of trust. One interpretation is that the majority of the people have not materially benefited from the market economy and are therefore critical. A further explanation is that low confidence in business comes from involvement in black economy activities. In Norway and France the level of confidence in business is much higher. The lower level in Norway is probably part and parcel of an egalitarian ideology rooted in market scepticism.

For political parties/movements we have figures only for the Baltic
countries. The extremely low score of an average of 11 percent confidence is

surprising, since establishing new parties is the very essence of introducing a
democratic regime.

To sum up, it is possible to rank three main groups of institutions
according to confidence:

(1) The new, non-political institutions, not directly involved in influencing policy decisions, which have diffuse and symbolic outputs, have the highest level of trust. These are the Church, the national armed forces and the press. The education system enjoys high confidence among people in general, because it produces opportunities, rather than specific outputs. As we shall return to later, one puzzling fact is that there is a tendency for people with higher education to have greater distrust in the education system than those with of low educational level. So far there is some evidence to support the hypothesis that institutions in the Baltic states with diffuse and symbolic functions will gain more support than institutions producing more specific and tangible outputs. The Baltic states are more similar to Francethan to Norway. Lack of egalitarian democratic traditions has a negative effect on confidence. The higher level of confidence in institutions and leaders in Estonia and Lithuania than in Latvia, and less negative attitudes among the well educated, may be related to a larger potential for consensus-building among the elites and prospects for a more stable political development in Estonia and Lithuania. However, the presence of a huge and dominating neighbouring Russia, combined with a large Russian minority inside Latvia, is probably conducive to national cohesion despite the distrust in institutions.

(2) The lower confidence in policy-making, implementing and order institutions (parliament, government, civil service and local government, courts and police) can be understood as a reaction against outputs affecting the quality of life in a rather direct way. If the attitudes are interpreted as a rejection of these institutions as such, there is a deep crisis of political legitimacy in the Baltic states. The lack of confidence in business companies is remarkable, since the market economy was supposed to be a positive alternative to centralized planning. It can be explained, however, by the poor performance of the market until now. Growing income inequalities, "raw" capitalism and Mafia influence have discredited the market economy.

(3) The extremely low score for political parties and movements is a cause for concern. One precondition for stable democracy is the existence of intermediating and legitimate institutions between the people and the rulers. Explanations are found, first, in the lack of a multiparty tradition, secondly, in the tendency of the voters to hold parties responsible for the

economic difficulties, and, thirdly, the rejection of party politics as a
remnant of the former one-party regime.

Trust in Leaders

Several researchers have stressed the elite character of the defunct Soviet system (Farmer 1992), the circulation and selection of communist elites (Harasymiw 1984), the integration of national elites in the Soviet system (Levits 1987) and the importance of indigenous and competent elite groups in the recent political development of the Baltic states (Lieven 1993; Steen 1994a, 1995).

The question is whether attitudes to institutions as such differ from the attitudes to institutional elites. As shown in the preceding section, people's confidence in institutions producing concrete outputs is low. If people react negatively to such institutions, we would expect the leader persons responsible for such concrete decisions to be particularly exposed to criticism and distrust.

Trust in leaders tends to be higher than trust in institutions. But the differences are considerably larger in Estonia and Lithuania than they are in Latvia. The difference in trust between leaders and institutions is at its highest for parliament and government. There is no support for the hypothesis that leader persons are more exposed to distrust than institutions. When leaders of decision-making bodies like parliament and government are trusted, absolutely and relatively to institutions, it is an

interesting indication of the important role of elite persons in postcommunist societies. Leader persons, even if they are exposed to unpopular decisions and criticized for mismanagement, seem to generate more confidence than abstract institutions.

The low confidence in institutions in Latvia runs parallel to the low trust in leaders. The question of why this is the case for Latvia and not for the other countries may be related to the fact that general elections in compliance with new national election laws had not been arranged at the time when the survey was carried out. But the low score in Latvia may also be the result of a more elite-critical culture than in the other two countries, regardless of the development in the electoral arena. One indication of this is the very low score also for leaders of institutions not directly involved in politics, such as the civil service and business companies.

There may also be one historic explanation: elite cooperation and the recruitment of national elites during the last thirty years took a different turn in Latvia than in Estonia and Lithuania. The Moscow-initiated purge of Latvian national communist leaders in 1959-62 heavily influenced the later development of elites.⁴ It became more difficult to establish a national communist leadership than in the other two Baltic states. The fact that the Latvian leaders were ruled from Moscow to a greater extent than in the other two countries, may be an explanation for the attitudes of the Latvians being more negative.

A low level of confidence in institutions, combined with a relatively high trust in leader persons, indicates a person-oriented political culture. The lower confidence in institutions than in France and especially than in Norway also supports such an interpretation. Egalitarians will be more critical to elite-persons than to institutions. A tentative conclusion is that in an elitist political culture, especially one without democratic traditions, people will have more trust in leaders than in institutions.

Social Background and Confidence: The Importance of Education

How is social stratification among the general population influencing levels of trust in the Baltic countries. The following background variables are examined: ideological preference, residence, age, gender, income level, ethnic group and education.

Several suggestions have been proposed about the effects of these background variables upon confidence in institutions in Western societies. It is not easy to identify any major trends. Listhaug & Wiberg (1992) found that age has a positive effect on confidence, due to internalization of established values over time. Left-right attitudes seem to have an impact on

confidence, with left more critical and right most confident. As to gender, women have more confidence in the Church than men. Education is in general weakly, but positively related to confidence. Confidence in parliament is to a greater extent related to educational level than for other institutions, suggesting "that the most politically informed and involved develop an allegiance to democratic institutions"⁵ (Listhaug & Wiberg 1992, 12).

In his investigation of alienation from political and social institutions in Russia, the Ukraine and Lithuania in 1990 Miller (1993) found that younger people, and in particular people with higher education regardless of age, had low trust in the regime. Concluding that this weakens the hypothesis of the strong socialization effects of education, as formulated by Abrahamson & Inglehart (1986), Miller proposes that the access to critical information makes it possible for the well educated to envision alternative forms of government and therefore helps explain their alienation from the Soviet system.

If the well educated were the most critical of established institutions prior to independence in the Baltic countries, as Miller suggests, we would assume that independence from the Soviet system, the introduction of democracy and market economy would have a considerable positive impact on confidence in institutions, especially among the well educated looking forward to a systemic change.

The total explanatory effect of the model is low for all institutions, except for attitudes towards the Russian forces (0.38). The purpose here, however, is to investigate effects of single variables. In Table 111, the effects of the independent variables are compared for the Baltic states as a group. Findings from the single countries are commented here. The complete tables from all countries are published in Steen (1994b).⁶' 7

One hypothesis about political preference is that because right-wingoriented people will be more sympathetic to change of regime, they will have more trust in new institutions than left-oriented persons.⁸ For the three states, right-oriented people tend to be more positive to national forces, government, parliament, police and political parties. And they are more negative than left-oriented persons towards trade unions and Russian forces. This is what should be expected, and is also in accordance with findings from Western countries. It is more surprising that left-right attitudes do not affect attitudes to business companies.

In Estonia, right-wing political attitudes affect confidence in political institutions positively and significantly. Effects on institutions like the national forces and police are clear in Latvia, while there is no significant effect on political institutions. This is to some extent the case in Lithuania, too. Lithuania is the only country with some effect on attitudes towards business (beta 0.09*).

Residence has a weaker effect than expected. The only significant effect in the Baltic countries is that people living outside cities and towns have more confidence in local government. They are also more sceptical to business companies, but this relationship is not significant. As for the single countries, there is an effect in Latvia on attitudes of rural people towards the Church (b 0.36 and beta 0.16*). Only for Lithuania does the variable have any significant effect on several institutions. In this country it has a positive effect on the confidence of the rural population in education, government, local government, police and the legal system, and a negative impact (b -0.25 and beta -0.14*) on the confidence in political parties.

Age is a central variable in theories of political socialization. If the socialization hypothesis is applicable to the Baltic countries, we would expect older people to have more confidence in old institutions and to be more sceptical towards new ones. In the Baltic states, age has a positive influence on attitudes towards the police (beta 0.10*), the education system (beta 0.12*), the Church (beta 0.15*) and the national forces (beta 0.22*), but it has a negative impact on the attitudes towards business companies (beta -0.25*). Older people trust more traditional institutions, which support order and religious values. Older people are also more sympathetic to the trade unions, which may be an effect of a longer period in working life. The dual transition to democracy and market economy hits elderly people more than others. This may explain the low confidence in business companies. But scepticism may also be related to new institutions to which elderly people cannot adapt as easily as the younger age cohorts. The greater trust of young people in business is probably related to another life situation, with opportunities of better paid jobs in private companies. There is also a gap between the generations concerning trust in the educational system, which does not enjoy the trust of young people to the same extent as that of old people. Confidence in the national forces is especially strong among older people in Latvia (b 0.19 and beta 0.32*). Here they also have more confidence in government and parliament than in the other two countries.

Concerning gender, women live up to our expectations and report more trust in the Church than men (beta -0.09*). The Church seems to influence the attitudes more in Catholic Lithuania (b -0.33 and beta -0.18*), than in the two predominantly Protestant countries. Women are also more confident in the court system (beta -0.07*). This holds particularly true in Estonia (b -0.20 and beta -0.13*). For other institutions there is only a larger effect on attitudes to business companies in Latvia. Here men have most confidence with b 0.30 and beta 0.17*.

In the three Baltic countries, the effect of income is rather modest for most
institutions, except for attitudes towards business companies. Surprisingly
enough, income is positively related to confidence in local government, civil

service, trade unions and political parties. The effects on attitudes towards business are b 0.19 beta 0.17* in Estonia and b 0.19 and beta 0.13* in Latvia; in Lithuania income has no significant effect. The ranking of business attitudes corresponds to the more developed market economy in Estonia, which has resulted in large income differences and trends towards a new middle-class stratum.

In their survey carried out during the final stage of the Soviet Union, Finifter & Mickiewicz (1992) found that ethnicity had an impact on attitudes towards change. But the largest gap was not between Baits and Russians, but between Central Asians and the others. However, in that study the Baits included Estonians only. Owing to the Russian minorities in the Baltic before the Second World War, and the continuous influx of Russian settlers and occupiers after the war, ethnicity has become a major cleavage, especially in the two northern-most Baltic states. We would expect the attitudes towards national questions and symbols such as the armed forces to be influenced by ethnic background. Furthermore, we would expect the ethnic nationals to have a more positive attitude to new political institutions than other ethnic groups.

In the three Baltic states the effect of ethnicity on attitudes towards Russian forces is substantial (beta -53*). For other institutions the effects are more modest, but significant: the press (beta 0.16*), national forces (beta 0.11*) and the parliament (beta 0.08*). The ethnic cleavage is much more clear-cut in relation to institutions related to the former regime, like the Russian army, than to the new institutions established after independence. Even if the general confidence in institutions is low, ethnic affinity does not play an important role in influencing attitudes towards the new institutions. Ethnicity is most of all related to oppressive institutions. The data indicate that even if Russians tend to be sympathetic to Russian forces in the Baltic states, naturally leading to great concern among indigenous elites, this does not mean deeper ethnic splits concerning confidence in other institutions.

As expected, ethnicity has the largest negative effect on attitudes towards Russian forces where the Russian population is largest, in Estonia (b -0.99 and beta -0.55*) and in Latvia (b -0.94 and beta -0.53*). For Lithuania the relationship is also clear but lower with b -0.70 and beta -0.40*. Furthermore, in Estonia and Latvia indigenous people have less confidence in the Church, than in Lithuania. Protestantism seems to nurture less enthusiasm for the Church than Catholicism. As to political institutions, in Latvia there is a positive effect of ethnicity on confidence in government (b 0.17 and beta 0.11*) and in parliament (b 0.28 and beta 0.18*). In the other two countries there is no significant effect on political institutions. However, there is one exception: ethnicity positively affects attitudes towards political parties in Lithuania.

If educational level is taken as an indicator of "political competence", two opposite hypotheses are relevant: the well educated will be especially critical because they are better informed and know the weaknesses of the system, or, the well educated, who are likely to have more than others by way of political insight, will be particularly concerned about democracy and will express loyalty in institutions. This variable produces the most consistent and most astounding effect. Education is systematically negatively related to all kinds of institutions in all three Baltic countries. The more highly educated people are, the less confidence they have in institutions. The impact though not very high is significant for institutions like the Church (beta -0.07*), the press (beta -0.13*), national forces (beta -0.08*), police (beta -0.07*), civil service (beta -0.12*), courts (beta -0.11*) and trade unions (beta -0.08*). For the political institutions the effect is also negative, but not significant. Parliament has beta -0.05, government beta -0.01 and political parties beta -0.07. In their study of Western countries, Listhaug and Wiberg (1992) report a positive beta 0.07 as the effect of education on confidence in parliament. Effects on other political institutions are also positive, but small. This may be interpreted as beneficial to political legitimacy and stability in Western countries.

In Lithuania the negative effect of education on confidence in the educational system is higher compared to the other three countries (b -0.12 and beta -0.10*). The negative effect on confidence in parliament is relatively higher in Latvia (b 0.09) and in Lithuania (b -0.09 and beta -0.07). The effects are not significant, but lend support to the contention that the well educated tend to be less confident in political institutions than people with a lower education. When using another technique suitable for analysis of the proportion with little or no confidence in institutions, we find that the average negative effect of high education on institutions is considerably larger in Latvia than elsewhere: 19.7 percentage points, compared to 11.6 in Estonia and 13.2 in Lithuania.

The findings for the Baltic countries are consistent with Miller's (1993) documentations of alienation from the system among well-educated people in the former Soviet Union in the late Gorbachev era. According to our data, the distrust was also carried over into the current post-communist era. Miller's suggestion that critical attitudes, brought about by the access to freely flowing uncensored information, may apply to the communist societies but not to the post-communist period. Prior to the transition, people with high education had more access to information about alternative forms of institutions than the less educated. Information is now open to all groups, and is not restricted to the well educated as it used to be. The critical attitudes prevailing among the well educated must be a function of factors other than information.

An alternative explanation is that the critical attitudes of the well educated may be a byproduct of elite relations. The more highly educated have a special concern for and interest in institutional leader positions. They are often dependent on and close to leaders. In small countries, such as the Baltic states, the group of leaders is rather limited and leaders tend to be the expression of the institution itself. At a time when institutions and leaders are becoming interchangeable, the well educated may feel they have a special competence to criticize the institutional leaders because they know them well. If this is correct, we would expect well-educated people to be especially critical of leaders.

Multivariate analysis shows that for the three Baltic states the level of education has no significant effect on trust in leaders. For the single countries it has only one significant negative effect, that on trust in leaders of civil services in Latvia (-11*). Since the level of education does not influence trust in leaders, the institutional scepticism of the well-educated must be related to the symbolic and performance functions of institutions rather than to mistrust in leader persons.

Confidence and Consequences for Stable Democratic Development

Political support may be behavioural as well as attitudinal (Gabriel 1992). Voting preferences, election turnout and party membership are indicators of behavioural political support. Attitudes may reveal a more extended and more refined range of preferences than behaviour makes possible. On the other hand, attitudes are not necessarily consistent with behaviour. Attitudinal support may be associated with a wide range of institutions and situations. Institutional symbols will be rather vague. Some policy outputs may be specific in consequences, while others are rather diffuse. Leaders are specific persons, but it is difficult to know if they gain support because of some indefinable charisma or as a result of specific actions or decisions.

Confronted with data, Easton's distinction between specific and diffuse objects of support is problematic. The main problem is how to classify the different institutions which operate under varying conditions according to such abstract categories. This paper sets out to introduce additional institutional dimensions, in order to come closer to the complex empirical reality of new political regimes.

The main finding is the overall low level of confidence in institutions among the populations of the Baltic states. It is particularly low in Latvia. This rather surprising revelation for new democracies was analysed from the perspectives of institutional type, elite culture and social background.

The findings can be summarized in some main points:

(1) There is more distrust in institutions producing policy decisions and
outputs than in institutions with a more symbolic function, and leaders
are more trusted than institutions as such.

(2) Patterns of confidence in the Baltic have more in common with France
than with Norway, and Latvia has a generally lower level of confidence
than Estonia and Lithuania.

(3) The group with the lowest confidence in institutions comprises people with higher education. No other independent variable has such a systematically negative effect on confidence in institutions. There is no negative effect of education on trust in leader persons.

When negative attitudes to political, and market institutions are widespread, one would expect the leaders to be exposed to criticism for having condoned economic mismanagement on a scale that dragged the people into misery. Alienation was a direct consequence of the elitist political culture that prevailed under communism. The weak support for social and political institutions in combination with a more distinct support for elite persons in the new Baltic democracies may be accounted for in terms of a deeply rooted elite culture in these societies. Lack of confidence in institutions as such seems to some degree to be compensated for by trust in leaders. The rather surprising finding that high education means less confidence in most institutions, while we do not find such a negative effect on trust in leaders, may indicate a supportive elite culture in post-communist countries among the higher educated stratum.

It is worth dwelling on what the chances are of developing legitimate and stable institutions in a political climate where people do not trust the institutions as such, but rely more on the leaders. Weak institutional performance provides a fertile ground for elite persons to appeal directly to the public. When intermediating institutions like party system and interest organizations are not sufficiently developed, the elites will have to shoulder even more specific functions. In this respect democratic development will be closely intertwined with the preferences and choices of leaders.

The question of democratic development and stability may be seen from two different perspectives. In a leader-oriented culture where charisma, personal styles and pragmatic politics are more important than long-term political programmes including efficient implementation, political stability will be particularly dependent on the democratic commitment, consensusbuilding and integrating capacity of the elite persons. "Consensually integrated elites" are not necessarily a trusted elite, but vital to democratic stability (Higley & Burton 1989). Political stability will become more a consequence of the ability of the elites to create coalitions or foster strong

leadership than a consequence of popular support. Democratic development is, according to this view, as formulated by Schumpeter (1942), dependent on the ability of the elites to establish the rules of the political game as a method for competing for the vote.

Another perspective is that, in the long run, institutions need general popular support in order to survive. But people do not identify with institutions in the abstract. As Linz (1990) maintains, newly democratized regimes can gain the allegiance of the people only if institutions produce substantive policy changes that are beneficial to solving the daily problems of the people. One conclusion is that lack of confidence in institutions, and in particular among the well educated, constitutes a reason for concern when democracy is defined in the egalitarian tradition as general and active support for the system. Low popular confidence does not, however, exclude democratic stability when elites are competing for leadership under procedures accepted as legitimate.

---

NOTES

1. Since "trust" means belief in the honesty and reliability of someone, the survey used this term about attitudes to leader persons. "Confidence" as a more general concept was used in the survey about attitudes to institutions. The concepts have parallels to the distinction between "specific" and "diffuse" objects of support, as proposed by Easton. In the text of this article "confidence" and "trust" are used interchangeably. The interpretation of the data should, however, take these distinctions into consideration.

2. Data from the Baltic states were provided from the following cooperating survey companies in the Baltic: Estonia (Emor), Latvia (Latvian Social Research Centre), Lithuania (Lithuanian Social Research Centre). Data from France and Norway were kindly supplied by the Norwegian Social Science Data Service.

3. The following question was put to the respondents: "Please look at this card and tell me, for each item (institution) listed, how much confidence you have in them, is it a great deal, quite a lot, not very much, or none at all?" In order to simplify the analysis the categories "a great deal" and "quite a lot" of confidence are put together.

4. Cfr. I. Apine: "Have there been national communists in Latvia?" Manuscript. Undated.

5. Quoted with the permission of the authors.

6. Beta is used to compare effects of independent variables within one country and for the Baltic together. B is used to compare the effect of the independent variables between the countries. Two-tailed test, * significant at the 0.05 level.

7. Coding: Confidence in institutions is coded with four values, with 4 indicating "A great deal" and 1 "None at all". Trust in leaders is coded with four values, with 4 denoting "Total trust" and 1 "Total distrust". Left-right attitudes 1-10 scale: 1 = exleft (1,2 on left-right scale; 2 = modleft (3,4 on 1-r scale); 3 = median (5,6 on 1-r scale); 4 = modright (7,8 on 1-r scale); 5 = exright (9,10 on 1-r scale). Residence: Cities and towns =1; rural =2. Age (age group): 1= -24; 2= 25-34; 3= 35-44; 4 = 45-54; 5 = 55-64; 6 = 65-74; 7 = 75 +. Sex: Male = 1; female = 0. Income: 1 = low; 2 = medium; 3 = high. Education (edlevel): 1 = low; 2 = medium; 3 = high. Ethnic group: 1 = ethnic national; 0 = others (non-ethnic).

---

7. Coding: Confidence in institutions is coded with four values, with 4 indicating "A great deal" and 1 "None at all". Trust in leaders is coded with four values, with 4 denoting "Total trust" and 1 "Total distrust". Left-right attitudes 1-10 scale: 1 = exleft (1,2 on left-right scale; 2 = modleft (3,4 on 1-r scale); 3 = median (5,6 on 1-r scale); 4 = modright (7,8 on 1-r scale); 5 = exright (9,10 on 1-r scale). Residence: Cities and towns =1; rural =2. Age (age group): 1= -24; 2= 25-34; 3= 35-44; 4 = 45-54; 5 = 55-64; 6 = 65-74; 7 = 75 +. Sex: Male = 1; female = 0. Income: 1 = low; 2 = medium; 3 = high. Education (edlevel): 1 = low; 2 = medium; 3 = high. Ethnic group: 1 = ethnic national; 0 = others (non-ethnic).

8. Self-identification on the left-right scale in post-communist countries is related with special problems. One may be unwillingness to disclose one's political preferences, another may be the meaning of the scale to people without multiparty traditions (cfr. Alisauskiene Bajaruniene, undated).

ACKNOWLEDGEMENTS

This survey was supported with funding from the Norwegian Ministry of Foreign Affairs and
the Norwegian Research Council. I acknowledge comments on an earlier draft from Tor
Bjørklund and Ola Listhaug.

REFERENCES

Abrahamson, P. R. & Inglehart, R. 1986. "Generational Replacement and Value Change in Six
West European Societies", American Journal of Political Science 30, 1-25.

Alisauskiene, R. & Bajaruniene, R. (undated). "Peculiarities of the Lithuanian Political
Consciousness 1990-92, and Lithuanian General Elections 1992", Paper. Vilnius: Baltic
Surveys.

Almond, G. & Verba, S. 1963. The Civic Culture: Political Attitudes and Democracy in Five
Nations. Princeton N. J.: Princeton University Press.

Barry, B. 1970. Sociologists, Economists and Democracy. London: Collier-Macmillan.

Dolbeare, D. & Edelman, M. J. 1974. Policies, Power and Change. Lexington; Heath and
Company.

Easton, D. 1975. "A Re-assessment of the Concept of Political Support", British Journal of
Political Science 5, 435-457.

Farmer, K. C. 1992. The Soviet Administrative Elite. Praeger. N.Y.

Finifter, A. W. & Mickiewicz, E. 1992. "Redefining the Political System of the USSR: Mass
Support for Political Change", American Political Science Review 86, 857-874.

Gabriel, O. W. 1992. "Values, Political Trust and Efficacy", Paper prepared for the Fifth
Meeting of the Subgroup "Impact of Values", European Science Foundation Scientific
Program on Beliefs in Government, Colchester, Essex, April 1992.

Harasymiw, B. 1984. Political Elite Recruitment in the Soviet Union. London: Macmillan
Press.

Hayward, J. 1982. "Mobilizing Private Interests in the Service of Public Ambitions: The
Salient Element in the French Policy Style?", in Richardson, J, Policy Styles in Western
Europe. London: George Allen & Unwin.

Higley, J. & Burton, M. G. 1989. "The Elite Variable in Democratic Transitions and
Breakdowns", American Sociological Review 54, (February) 17-32.

Inglehart, R. 1990. Culture Shift in Advanced Industrial Society. Princeton. New Jersey.

Levits, E. 1987. "National Elites and their Political Function Within the Soviet System: The
Latvian Elite", Journal of Baltic Studies XVIII, 176-190.

Lieven, A. 1993. The Baltic Revolution; Estonia, Latvia, Lithuania and the Path to
Independence. New Haven, Mass.: Yale University Press.

Linz, J. 1990. "Transitions to Democracy", The Washington Quarterly, Summer Issue.

Listhaug, O. & Wiberg, M. 1992. "Confidence in Political and Private Institutions", Paper.
University of Trondheim.

Miller, A. H. 1993. "In Search of Regime Legitimacy", in Miller, A. H. et al., eds, Public
Opinion and Regime Change. Oxford: Westview Press.

Rose, R. 1984. Understanding Big Government. London: Sage.

Schumpeter, J. A. 1942. Capitalism, Socialism and Democracy. New York: Harper.

Steen, A. 1994a. Recirculation and Expulsion: The New Elites in the Baltic States. Working
Paper 09/94. Department of Political Science. University of Oslo.

Steen, A. 1994b. Confidence in Institutions in Post-Communist Societies: The Case of the
Baltic States. Working Paper 12/94. Department of Political Science. University of Oslo.

Steen, A. 1995. Consolidation and Competence. The Politics of Recruiting Political Elites in
the Baltic States. Working Paper 03/95. Department of Political Science. University of Oslo.

Vardys, V. S. 1987. "The Role of the Churches in the Maintenance of Regional and National
Identity in the Baltic Republics", Journal of Baltic Studies XVIII, 287-300.