Trust in Political Institutions. the Nordic Countries Compared With Europe

Embed Size (px)

Citation preview

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    1/24

    Trust in Political Institutions: The Nordic Countries Compared with

    Europe

    Ola Listhaug and Kristen RingdalThe Norwegian University of Science and Technology

    Paper prepared for the Norwegian Political Science Meeting, NTNU, Trondheim,January 3-5, 2007

    Data are from the 2004 European Social Survey made available by TheNorwegian Social Science Data Services

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    2/24

    2

    Abstract

    Using data from the 2004 European Social Survey we study where the Nordiccountries rank on political trust when compared with other countries in Europe.Citizens trust in the major political institutions is important for the state ofdemocracy as well as for the functioning of broader social and economic processes insociety. Our findings confirm previous research showing that the Nordic countriesrank very high in political trust in national institutions. The empirical analysis showsthat the high trust levels in the Nordic countries may be explained as a combination ofcompositional effects and macro characteristics. The compositional effects are causedby favorable scores on the indicators of the countrys performance, especially on theevaluation of the economy, and on indicators of political distance betweengovernment and citizens. The macro effects are caused by the top positions of theNordic countries on the Human Development Index, which again reflects the goodliving conditions in these countries.

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    3/24

    3

    Introduction

    In this paper we study where the Nordic countries rank on political trust whencompared with other countries in Europe. Citizens trust in the major institutions in asociety is important for the state of democracy as well as for the functioning ofbroader social and economic processes in a society. High trust levels signal thatinstitutions are working effectively, thus reducing the chance that non-democraticforms of government will receive support. High trust levels facilitate social andeconomic exchange and reduce transaction costs in markets. Trust reduces the needfor control and supervision, which saves money for government as well as for firmsand other actors in the private sector.

    In the comparative dimension countries with high trust will be at an advantage inattracting investments, trade, and tourism. For these reasons we see political trust as asuccess criterion for societies. Based on what we know from previous research theNordic countries place favorably on political trust measures. While we know that the

    Nordic score high on political trust we know little about what can explain this pattern.

    Is there a peculiar quality to the Nordic countries - a Nordic model (sterud 2005) -or is the success of these countries in creating political trust among citizens, primarilya reflection of the fact that they are small, rich, and homogenous in social andeconomic terms, such that we will find that other countries that share thesecharacteristics also will have high levels of political trust?

    High trust levels are in most cases possible to achieve only when institutions receivesupport from major social and political groups in society. This is an indicator thatsocial integration is working successfully. At a time when cleavages in societies may

    regain their important due to immigration of groups that differ by religion and othercultural aspects from native populations, support for institutions that bind groupstogether will become increasingly important. The recent territorial expansion of theEuropean Union, the establishment of the European Economic Area, and theSchengen treaty area are factors that facilitate geographical mobility, and, indirectly,contributes to the increasing ethnic diversity of European countries.

    Trust in political institutions is part of a wider concept of political trust that builds ona classification by Easton (1965) and is available in a more recent version by Norris(1999b). This conceptualization classifies trust in a hierarchy that goes from specificto general: Political actors (elected political officials and political leaders, regime

    institutions (electoral institutions, order institutions, public service institutions),regime performance (evaluation of how the actual democracy works), regimeprinciples (support for the basic ideas of democracy), and political community(identification with the nation).

    Political trust is not limited to the nation state but is increasingly relevant atsupranational level. In Europe trust in the European Union is crucial as EU is in aprocess of increasing its political authority over the member states, and the success ofthe Union will depend on how strongly the evolving institutions are able to connectwith the identifications and interests of the publics in the member states. The EUmight be successful in economic terms, but a broader cultural unity is lacking. In the

    words of Johan P. Olsen: "The EC/EU has had some success in conceptualizingEuropean as an economic unit, but less so with the idea of political, cultural, and

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    4/24

    4

    social unity in Europe based on a common heritage such as Christendom, Latin, andRoman Law, or common, future projects" (Olsen 1995: 26). The more than ten yearsthat has passed since this remark has increased its relevance due to the increasedpolitical and cultural diversity of the EU resulting from the inclusion of new memberstates from Eastern Europe, and the potential inclusion of Turkey, a major Muslim

    country. It seems obvious that citizens trust in EU institutions is a key measure of thesuccess of the European Union in establishing a more firmly grounded legitimacy. Inthis paper we restrict the analysis of political trust at the EU level to the support forthe European Parliament.

    Our expectations for trust in the EU institutions are different from what we expect fornational institutions. The Nordic countries have had a problematic relationship to theEuropean. Norway has twice rejected membership in national referendums. Iceland isnot a member, and Denmark, member since 1973, has at several occasions been atodds with EU on important policy directions and institutional reforms. Sweden andFinland became members in 1995. Finland has had some success in mobilizing public

    support for EU institutions, while Swedish mass publics have been quite negative tomembership. We expect that trust in the European parliament will reflect the basic EUorientations and experiences with EU membership and show a quite different patternacross the five countries than trust in national political institutions.

    Recent research: Trends and comparisonsRecent research on trust in institutions and other aspects of political trust has beenoverwhelmingly associated with concerns for trendsis trust declining or not? Thelarge Beliefs in Government project of the European Science Foundation (BIG)published its findings in 1995 (but time series data ended c. 1990) and concluded thatthere was no general decline in political trust in Western Europe (Klingemann andFuchs 1995). Pippa Norris (1999a) concluded her follow up on BIGextending thetime series past the mid 1990s and broadening the data to cover non-Westerncountriesthat citizens remained committed to the values of democracy, but hadbecome more critical to the core institutions of democracy. In the most recent work ontrends in political trust Russell Dalton demonstrated that political trust is declining inadvanced industrial countries (Dalton 2004).

    The Third Wave of democratization and the process of globalization havestimulated studies that take a closer look at political trust in a comparativeperspective. Much of the comparative research on political trust uses data from the

    European Values Study/World Values Survey (EVS/WVS). These surveys have beenperformed in more than 80 countries and territories covering all inhabited continentsand constitute the most global dataset that we have of relevant data on trust. Allsurveys have included at least some questions on political trust, and some of thesurvey waves have had political trust and support of democracy as a major topic.Using Values survey data from 1995-1997 Klingemann (1999) found that support fordemocratic ideals and principles are robust in all parts of the world, but the support issomewhat stronger in Western Europe than in the rest of the countries.

    Listhaug and Aardal (2003) performed a detailed analysis of variations of democraticsupport in Europe using EVS data from 1999/2000. On the basis of Freedom House

    indicatorsand building on previous research by Klingemann (1999) - they classifiedEuropean countries in three groupsstable democracies (countries in Western

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    5/24

    5

    Europe), new democracies (most of the post-communist countries in Central- andEastern Europe), and transitional democracies (Croatia, Russia, Ukraine, andBelarus). They compared these countries on five dimensions of political trust: Howthe current system is rated relative to the communist regime, satisfaction with howdemocracy is developing, rejection of non-democratic forms of government, rejection

    of criticism of democracy, and support for democracy as principle. They found that onall dimensions, support levels are higher in stable democracies than in newdemocracies and that the transitional regimes on most of the indicators have thelowest trust readings. The negative findings for new democracies and transitionalregimes remain when they control for an extensive set of variables that are related totrust.

    In an extension of this analysis Anderson et al. (2005) showed that the gap in politicaltrust between winners and losers is larger in new democracies and transitional regimesthan in the stable democracies in Western Europe. While it is quite normal that voterswho support parties that lose elections will become less trusting toward the political

    system, the gap between losers and winners should be kept within a reasonable size.With the definitions and measures that we use it is difficult to decide when the gap isoutside a boundary where the low trust of losers can be a critical factor for thesurvival of the democratic system.

    The study by Listhaug and Aardal (2003) provides some information about therankings of the Nordic countries. This study is based on the European Values Studydata from 1999/2000 and only Denmark, Finland, and Sweden are included. Amongthese countries Denmark has consistent the highest ranking, and Denmark is also oneof the countries with the strongest support for democratic values among all Europeancountries. Sweden, and, especially, Finland, showed a somewhat lower support level.

    One could argue that for all stable democracies the support levels for democraticvalues are so high that variations within this group are not necessarily of importance.For established democracies it might me more meaningful to compare trust levels inthe medium and lower levels of the trust hierarchy like confidence in institutions andtrust in politicians.

    With data from the first two waves of the European Values Study in 1981 and 1990Listhaug and Wiberg (1995) compared confidence in six government institutions forthe countries in Western Europe. Their analysis includes four Nordic countries, but

    data for Iceland and Denmark were available only for 1990. At the first time pointNorway is clearly number one in trust in government, but Sweden and Iceland (datawere collected in 1984) are also in the top half. In 1990 Norway and Iceland are ontop (in that order) and Sweden is about in the middle of the rankings.

    Extending this time series with data from the 1995-1997 World Values Survey,Listhaug (2005) demonstrated that the Nordic countries in the surveyFinland,Norway, and Sweden - remain highly ranked among European countries. Again,Norway is on top, with trust in parliament as especially strong. More recently politicaltrust in Norway has declined, and there are indications that the countrys comparativestatus in trust rankings has suffered as well. Norway is unique among developed

    countries in having an important oil sector which produces a huge surplus ingovernment budgets. The economic surplus is put into an oil fund that is invested

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    6/24

    6

    abroad. The contrast between a growing pile of money that cannot be touched, andcitizens preferences for spending more to solve problems in Norway, can have leadto frustrations that have fuelled the new distrust.

    A weakness of the studies cited above is that few systematic explanations are

    introduced to make sense of comparative patterns. In some cases this might beexplained by the fact that we have data from a small number of countries whichmakes it difficult to test hypotheses about cross-national variations. In this chapter weattempt to improve on previous research by introducing more rigorous measurementof trust differences as well as estimating statistical models that include variables thatare relevant to explain comparative differences in political trust.

    The Model

    The standard model to explain variations in political trust at the individual include twomain categories of independent variables: Political distance and performanceevaluations. We expect that an increasing political distance between government and

    citizens will lead to a decline in trust. Distance is primarily measured as policydistance on the salient issues of the day as developed in the classic article by Miller(1974), but may also include measures of ideological distance or distance by partythrough the division of voters who vote for winning parties and those who are on thelosing side (Miller and Listhaug 1990, 1998, Anderson and Guillory 1997, Andersonet al. 2005).

    Political distance can sometimes result in a curvilinear pattern as for example whenvoters on either side of the political spectrum are more distrusting than voters in thepolitical centre. This can be explained by the fact that most governments are eithercentre-left or centre-right, which can alienate citizens on both sides.Political distance is measured along issues where citizens take different sides orpositions.

    In contrast to this we can have a situation where the electorate agrees on policypositions. In this case policy distance to government becomes irrelevant for politicaltrust. What counts is government performance, how well government is able to fulfilthe goals that citizens agree on. Most of the research on performance tests hypothesesabout performance of the economygood times breed trust in government, bad timeslead to a decline in trust levels (Listhaug 2006). Huseby (2000) has extended theanalysis of performance to the environment and social policy (care for the elderly and

    health policy). These policy areas have relatively high consensus about the goalsatleast in the Nordic countries. Thus performance considerations become dominant andtrust will be affected by evaluations of how good job the government is doing inproviding good care for the elderly and the sick.

    It is likely that factors accounting for variations in trust levels between citizens willalso be relevant in explaining variations among countries. In most of the reportedresearch, the Nordic counties are placed on the high end of political trust distributions.By most accounts the policy distances between citizens and government in thesepolities are small, possibly reflecting an egalitarian social and economic structure andpopulations that are homogenous along cleavages. Moreover, the Nordic countries

    have systems of proportional elections which strengthen political representation of allsocial and political groups, further reducing political divides.

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    7/24

    7

    The Nordic countries are also relatively rich and their economies have performed wellover extensive periods of time. All have welfare states that take care of basic needsbased on universalistic principles. At times the Nordic economies have suffereddownturns, like in Finland in the 1990s and also Sweden has had problems in keepingup economic growth. Needless to say, the active role of government in solving

    problems and providing services for citizens tend to increase expectations that aredifficult to meet. This is obvious in the current oil case of Norway, but may also havea more general relevance.

    In the remaining part of this study, we first present the basic multilevel model anddiscuss the data that we use. In the empirical analysis we attempt to answer tworesearch questions. First, we simply ask where the Nordic countries rank on thevarious dimensions of political trust. Based on previous research and theoreticalarguments, we expect trust in national political institutions to be relatively high whencompared with other European countries. For trust in the European Union thisexpectation will not necessarily be true as the membership issue has been contentious

    in the Nordic countries.

    The second research question attempts to sort out explanations for the cross-nationalvariations in trust. We make a distinction between micro- and macro explanatoryvariables. The set of micro-variables includes measures of policy issues that arerelevant for political distance, performance evaluations, and demographical controls.The results of this analysis will tell us how much the placement of the Nordiccountries can be explained by compositional effects. In addition, we include twomacro variablessize of the country populations (natural log of the size ofpopulation) and country score on the UN Human development index (HDI). Theeffects of these variables will tell us if the Nordic countries are distinct on politicaltrust because they are small or because they are rich and highly developed. As analternative to the detailed macro variables, we will introduce a country classificationmainly based on welfare regimes.

    Methods

    To answer the second research question we employ a multilevel analysis of pooleddata from the ESS 2004 covering 24 countries. The multilevel model will briefly bepresented below, and a more extensive description of the technique is found in theintroductory chapter. The individuals constitute level 1 and the countries constitute

    level 2. Our basic multilevel model with a random intercept may be expressed in thisgeneral equation where the boldfaced letters indicate vectors.

    Yij =0 X+ Z + uojeijThe dependent variable, Y, has two subscripts, i for individuals andj for countries. 0is the weighted average regression constant among the countries. X is a vector of

    individual level explanatory variables. is a vector of their regression coefficients. Zis a vector of country characteristics, or macro-variables and their regressioncoefficients. Uoj captures the deviation from the weighted average regression constantfor countryj, i.e. this is the basis for calculating the between-country variation in Y.Finally, eij represents the individual level error term, i.e. the basis for calculating thewithin-country variation in Y.

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    8/24

    8

    The basic assumptions of the multilevel regression model are defined in theintroductory chapter. The model restrains the X to have the same, averagerelationships or effects across all countries. This assumption may be relaxed by

    allowing the s to vary among the countries and by adding cross-level interactions.

    Political trust indicators

    The ESS 2004 includes seven questions on political trust covering both national andinternational political institutions. The questions were introduced in this way:

    Using this card, please tell me on a score of 0-10 how much you personally

    trust each of the institutions I read out. 0 means you do not trust an institution

    at all, and 10 means you have complete trust. Firstly

    country's parliament (B4) the legal system (B5) the police (B6) in politicians (B7).

    in political parties (B8) in the European Parliament (B9) in the United Nations (B10)

    Our analysis will be based on all questions except the item about the United Nations.The items that we analyse tap trust for the electoral system (B4, B7, B8), trust in thelegal system (B5, B6), and trust in the European Parliament (B9).

    Individual level explanatory variables

    These include a set of demographic characteristics, measures of political distance, aset of indicators of performance evaluations, and a set of questions on political issues.The first group: age in years, age squared, and years of education are included mainlyas controls. Political distance is measured in two ways. The first one is whether or notthe respondent at the last parliamentary election voted for a party now in governmentposition. The second indicator is the left-right scale based on question (B23):

    In politics people sometimes talk of left and right. Using this card, where

    would you place yourself on this scale, where 0 means the left and 10 means

    the right?

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    9/24

    9

    The first political issue is about immigrants. We formed a scale as the mean of thescores on three questions about negative or beneficial consequences of immigration:

    Would you say it is generally bad or good for [country]s economy that people

    come to live here from other countries?

    And, , would you say that [country]s cultural life is generally undermined

    or enriched by people coming to live here from other countries?Is [country] made a worse or a better place to live by people coming to live

    here from other countries?

    The wordings at the ends of the 0 to 10 scales differ, but 0 means that theconsequences are bad and 10 that they are positive. The scaleImm_goodranging from0 to 10 has excellent psychometric properties: It is one-dimensional with high internalconsistency (Cronbachs alpha = 0.84).

    Next, follows three single questions on different issues. A question from thesupplementary self-completion questionnaire is used to make a 01-indicator ofsupport for the environment:He/She strongly believes that people should care for

    nature. Looking after the environment is important to him/her. This question on selfdescription has response categories ranging from 1. Very much like me, to 6. Notlike me at all. The two first categories as joined into 1 and the remaining ones arecoded 0.

    We also include a question on income differences: The government should takemeasures to reduce differences in income levels (B30). The response categories rangefrom 1. Agree strongly, to 5. Disagree strongly. The scoring of the question isreversed in our version so that high values mean support for government action toreduce income differences.

    The final issue included in our analysis is about European unification:Now thinkingabout the European Union, some say European unification should go further. Others

    say it has already gone too far. Using this card, what number on the scale best

    describes your position?[0. Unifications has already gone too far 10. Unification

    should go further]

    We include two categories of performance evaluations: Assessments of personal well-being and life satisfaction, and evaluation of how well the country is doing.Life satisfaction is measured by the following question:All things considered, howsatisfied are you with your life as a whole nowadays? Please answer using this card,

    where 0 means extremely dissatisfied and 10 means extremely satisfied.

    In addition, we include a question on income adequacy (F33):Which of the descriptions on this card comes closest to how you feel about your

    households income nowadays?

    1. Living comfortably on present income.

    2. Coping on present income.

    3. Finding it difficult on present income.

    4. Finding it very difficult on present income.

    We recode this question into two categories, distinguishing between these who cope

    or may live comfortably on present income from the other alternatives.

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    10/24

    10

    We include three variables tapping the extent to which the respondent is satisfied withthe present state of the economy, education, and the health services:On the whole how satisfied are you with the present state of the economy in

    [country]? Still use this card [0=Extremely bad10=Extremely good] (B25).Now, using this card, please say what you think overall about the state of education in

    [country] nowadays? (B28).Still using this card, please say what you think overall about the state of health

    services in [country] nowadays? (B29).

    Country level explanatory variables

    The low number of countries put severe restrictions on the number of country levelvariables it is meaningful to include. The two main dimensions covered by our twoindicators are size of the country, measured by the natural logarithm of populationsize, and the standard of living measured by the Human Development Index (HDI)published by the United Nations since 1990. We also tried out gross national productper capita (GNP/cap.) and the increase in GNP, but their explanatory power warlargely captured by the other two macro-variables. GNP alone was also clearlyinferior to the HDI.

    The macro-variables are frequently strongly correlated. The reason for this is thatcountries tend to cluster, especially towards the ends of the scales. As an example, theNordic countries are small, rich, have long democratic traditions, peaceful, etc, whichreduces the possibilities for

    Another way of dealing with the strong correlations among the macro-variables is to

    replace the macro-variables with a country classification. We use the Ferrera (1996)classification according to welfare regimes, adding an extra category for East-European countries: 1. Bismarckian (Austria, Belgium, France, Germany,Luxembourg, the Netherlands, Switzerland), 2. Nordic (Denmark, Finland, Iceland,Norway, Sweden), 3. Anglo-Saxon (UK, Ireland), 4. Southern (Spain, Greece, andPortugal), 5. Eastern (Czech Republic, Estonia, Hungary, Poland, Slovakia, Slovenia,and Ukraine). We use the Eastern Europe as the reference category in our multilevelanalysis.

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    11/24

    11

    Results

    In this section we present the results of the empirical analysis. We attempt to answerour two research questions. First, we ask where the Nordic countries are located in the

    rankings of political institutions, are these counties still on top of the list as previousresearch has demonstrated? Second, which factors can explain the trust levels? Wepursue factors both at country level and at the individual level.

    Gross country differences in political trust

    In Table 1 we display the country means on the political trust indicators. Thecountries are listed from high to low scores on each of the six questions. The Nordiccountries are printed in bold arte be easy to identify. We may first note that there arelarge differences in political trust between the countries as the mean ranges from 6-7to around 2. The midpoint between no trust and complete trust is 5 (on the 0-10 scale)

    Furthermore, the trust in the legal system is higher than the trust in the electoralsystem. Trust in the European Parliament is weaker than for trust in nationalinstitutions, and the differences between countries are smaller. The high percentage ofDont know -answers to the EU questions reflects the lower saliency of thatinstitution when compared with national institutions.

    Table 1 about here

    It is evident that the political trust in the Nordic countries is high compared to othercountries. For trust in national political institutions the Nordic countries are found inthe top group for all institutions. Denmark is the clear leader and tops the ranking forfour of the five institutions and is number two on the fifth, police, where Finlandcomes is number one. All five countries are placed in the top third on all rankings,

    with a minor exception for trust in politicians, where Sweden is number 9 of the 24nations. Norway, which used to be a leader in political trust, is now in the lower partof the Nordic distribution, giving further support to the thesis that that the big oilfortune that Norway is building in foreign investments leads to frustrated economicexpectations that undermine political trust (Listhaug 2005).

    As excepted we find that trust in the European Parliament shows a quite differentdistribution. The Nordic countries are now scattered, with a non-member country,Iceland, at third place, Finland is number seven, Denmark is at tenth place, Norway isnumber fifteen, and Sweden is next to last among the 24 countries. On average theNordic countries score comparatively much weaker on European trust than onpolitical trust of national political institutions.

    The similarity in the rankings across the five indicators of trust in national politicalinstitutions indicate that the data may be simplified by constructing one or morecomposite scale based on the five indicators. There are two advantages of scalescompared to single indicators: first, the number of dependent variables in our analysiswill be smaller. Second, composite measures have large variability and are morereliable than each single indicator.

    A factor analysis of the five questions on trust in national institutions gave a one-

    dimensional solution in the total sample. However, it is important to validate themeasurement model by looking at each country separately. Doing so, a clear picture

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    12/24

    12

    emerges. Two factors are needed to obtain measurement model with rather similarfactor structure across all the countries. These two factors explain more than 78percent of the variation in the five indicators in all countries. Not surprisingly the firstdimension covers trust in the electoral system and the second dimensions capturestrust in the legal system. We, therefore, constructed a scale for each of the dimensions

    as the means of the scores on each of the relevant indicators. Both scales, TelectandTlegal, have excellent psychometric properties. The scales are one-dimensional andtheir internal consistency measured by Cronbachs alpha are 0.90 and 0.79respectively.

    The mean score for each country on the two scales are graphed in Figure 1. The highcorrelation between the two political trust dimensions at the country level is clearlyvisible in that the points for the countries are found along a narrow band from thelower left to the upper right corner. The actual correlation at the country level is 0.90.The scales are also rather strongly correlated at the individual level (r=0.65).

    The clusters of countries is evident with the Nordic countries plus other smaller richcountries at the top on both dimensions and the four Eastern European countries arefound in the lower left corner, scoring low on both trust in the electoral and in thelegal system. This brings us to our multilevel analysis which has two purposes. First,we attempt explain the observed differences among the countries and, second, toestimate within-country differences as captured by the individual level explanatoryvariables in the model.

    Figure 1 about here

    A multilevel analysis of political trust

    Table 2 presents the variables to be used in our analyses. The dependent variables arethe three measures of trust for the electoral and the legal system plus the singlequestion on trust in the European Parliament. The next panel in the table presents thedemographic controls, followed by our political distance measures: voted for the partyin government, and the left-right scale. To capture, non-linear relationships to politicaltrust, the left-right scale is decomposed into six categories with the middle category asthe reference category in the multilevel analyses. The remaining individual levelexplanatory variables are the performance evaluations indicators and the politicalissue indicators. The last panel of the table describes the two country level

    explanatory variables,HDIandLnpop (population size), and the countryclassification.

    Table 2 about here

    The multilevel analyses of the three political trust measures are presented in Table 3and 4. Table 3 shows the variance components in four models for each of the threedependent variables. Model 0, the null model, has no explanatory variables. Itspurpose is to split of the variance in the dependent variables in the within-countrycomponent (Se) and the between-country component (Su). These components are alsothe basis for calculating the intra-class correlation which shows the proportion of thevariance in the dependent variable that is attributed to between-country variation. Thisproportion is 0.20 and 0.23 for trust in the two national institutions, but only 0.04 for

    trust in the European Parliament.

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    13/24

    13

    The two variance components in model 0 may also be seen as estimates of themaximum amount of variation at the individual level an at the country level ourmodels can explain. The amount of variance at the two levels explained by eachmodel is also found in Table 3. Model 0, the baseline, explains nothing by definition.Model 1 adds the demographic controls, age, age squared and years of education.

    These variables explain only trivial amounts of the variance at the individual as wellas at the country level. In model 2, the remaining individual level explanatoryvariables are added. They include measures of political distance, performanceevaluations and political issues. These variables explain more than 20 percent of thevariation in the three political trust measures.

    The variables also explain most of the between-country variation. They explain about69 percent and 63 percent, respectively, of the between-country variation in trust inthe electoral system and in the legal system. For trust in the legal system, we observeabnormalities as the explained variance actually turns negative. This may happen inmaximum likelihood estimation if the variance component is small at the outset.

    Finally, the individual level variables explain nothing at all of the between-countryvariation in the trust in the European Parliament.

    Table 3 about here

    In model 3, we introduce the two macro-variables,HDIandLnpop. They increase theamount of explained between-country variation to 70 percent for the first and to 75percent for the second national trust measure. The two variables also explain about30% in the between-country variation in our third trust measure. Model 4 is analternative to model 3 in that the two macro-variables are replaced by the countryclassification. For the two trust scales, the country classification does marginallybetter in terms of explanatory power than the macro-variables in model 3. For trust inthe European Parliament, the country classification explains about 39% percent of thebetween-country variance.

    Models 3 and 4 perform about equally well, with an advantage to model 4. Theindividual level variables do, however, have almost identical coefficients in model 2-4. Therefore, the results for model 3 are the basis for Table 4. The results for thecountry classification from model 4 are presented as an additional panel at the bottomof the table.

    Table 4 about here

    Starting from the top, all three types of political trust show a weak curvilinearrelationship to age. The level of trust decreases weakly until people reaches 40 to 50years for thereafter to weakly increase. Thus, at least for the two dimensions of trustin national institutions, the level of support is highest among older people.Men show slightly lower levels of political trust than women. Trust also increaseswith years of education completed, but this effect is rather weak. The predicteddifference between a person with no education and one with 20 years of education isonly about 0.2 points on the trust scales ranging from 0 to 10.

    The first of the political distance measures is whether one voted for a party ingovernment. In model 3 this variable is represented by two categories: Yes, and

    Did not vote, and those who voted for parties in opposition constitute the referencecategory. The results are as expected: those who did not vote show the lowest levels

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    14/24

    14

    of trust, and those who voted for a party in government show the highest level of trust.The differences are statistical significant but they are rather small.

    The left-right scale is represented as a categorical variable with six categories, themiddle one being the reference category. People identifying with the left have lower

    trust in the legal system than others, followed by those who answered Dont know.People identifying with the right side of politics show high trust in the electoralsystem an in the European Parliament, but they do not distinguish themselves in trustfor the legal system. Those who answered Dont know show low levels of trust onall three measures.

    Next follows the issue indicators. Only the immigration issue has any substantialeffect on trust. Citizens who evaluate the effects immigrations as positive have higherpolitical trust than those who evaluate the effects of immigrants as negative. As to beexpected, the question on the EU-issue is rather strongly and positively related to trustin the European Parliament.

    We have five performance variables in the model. The two variables capturing theconditions for the respondents show only minor effects on political trust. The threevariables tapping the performance of the country, satisfaction with the economy, theeducation and the health services in the country, all show relatively strong effects.Since they all are measured in the same 010-scale, their effects are directlycomparable. All signs of the coefficients are positive indicating the high satisfactionwith performance is related to all three dimensions of political trust. The satisfactionwith the economy of the country shows the strongest effect. The maximum effects ofthis performance evaluation indicator, is around 2.5 on the scale from 0 to 10 for trustin the electoral system. Its effect on the two remaining trust measures is smaller butstill substantial. The effects of evaluation of the education and the health system issmaller, their maximum effects are around one point on the 010 trust scales.

    Finally, the two last panels in Table 4 shows the results for the country-level variablesin model 3 and 4. Population size is negatively related to all dimensions of politicaltrust, but the relationships are not statistical significant.The Human Development Index (HDI) is positively related to political trust in theelectoral and the legal system. However, only the latter, rather strong relationship, isstatistical significant. The strongest relationship is with trust in the EuropeanParliament. This relationship is strongly negative and statistical significant.

    The country classification in the last panel, give more detailed results on netdifferences among groups of countries. The reference category for the countryclassification is the East European countries. With one exception, all othercoefficients are positive for trust in the electoral and the legal system. This indicatesthat, also controlling for all our individual level variables, the East European countriesshow the lowest political support.

    The Nordic and the Bismarckian countries score highest on trust in the electoralsystem. The Anglo-Saxon countries score lowest together with the East Europeancountries. The same ranking is found for trust in the legal system.

    For trust in the European Parliament the pattern is different. The South and the EastEuropean countries show the highest trust in this institution. At the other end, we find

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    15/24

    15

    the Nordic and the Anglo-Saxon countries with the lowest trust in the EuropeanParliament.

    Discussion

    The main purpose of this chapter has been to examine country differences in trust in

    political institutions, especially focusing on the Nordic counties. Previous research aswell as the present study has demonstrated that the Nordic countries rank very high inpolitical trust in national institutions, but that their support for European institutionsare weaker. We have tried to explain these patterns by analyzing the most recentsurvey data from ESS by a multilevel statistical model where the individualsconstitute level 1 and countries constitute level 2.

    Initially we found it necessary for both theoretical and statistical reasons todistinguish between three dimensions of political trust: trust in the electoral system,trust in the legal system, and trust in the European Parliament. The two firstdimensions were fairly strongly correlated, especially at the country level. The Nordiccountries and other small rich countries showed the highest level of trust in the twonational political institutions. At the bottom were the East European countries. Thepattern was quite different for trust in the European Parliament. On this dimension,the Nordic countries scored lower, and the South European countries, were found thetop. Also, the overall country differences were smaller for this dimension than fortrust in the two national political institutions.

    How can we explain these gross country differences? The country differences may bedue partly to compositional effects and partly to macro-characteristics of thecountries. Our multilevel analysis clearly shows that the demographic controls cannot

    explain the between-country variation in political trust. The attitudinal explanatoryvariables, on the other hand, especially the evaluation of performance indicators, doexplain the major part of the between-country variation in trust in national politicalinstitutions. These variables are, however, unrelated to the variation in trust in theEuropean Parliament among the countries.

    Adding the country classification in model 4 showed the net differences in politicaltrust between the country categories. After controlling for all individual levelvariables, the between-country differences are smaller than the gross ones, but basicpattern remains with the Nordic countries highest on trust in national politicalinstitutions, and the East European countries at the other end of the scale, showing

    low levels of trust.

    Adding the macro-characteristics of countries in model 3 explained most of theremaining between-country differences in all three dimensions of political trust.Especially the Human Development Index is related to all three dimensions ofpolitical trust. Countries high on the HDI show high trust in the electoral and the legalsystem, but they are less inclined than others to trust the European Parliament.

    These empirical findings show that the high trust levels in the Nordic countries maybe explained as a combination of compositional effects and macro characteristics. Thecompositional effects are caused by the high scores on the indicators of the countrys

    performance, especially evaluation of the economy, and indicators of politicaldistance. The macro effects are caused by the top positions of the Nordic countries on

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    16/24

    16

    the Human Development Index, which again reflects the good living conditions inthese countries.

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    17/24

    17

    ReferencesAnderson, Christopher J., Andr Blais, Shaun Bowler, Todd Donovan, and Ola

    Listhaug. 2005.Losers Consent: Elections and Democratic Legitimacy.Oxford: Oxford University Press.

    Anderson, Christopher J. and Christine A. Guillory. 1997. Political Institutions andSatisfaction with Democracy.American Political Science Review 91: 66-81.

    Dalton, Russell J. 2004.Democratic Challenges, Democratic Choices. Oxford:Oxford University Press.

    Easton, David.1965.A Systems Analysis of Political Life. New York: John Wiley.

    Ferrera, Maurizio.1996. The Southern mode of welfare in Europe. Journal ofEuropean Social Policy 6(1): 17-37.

    Huseby, Beate. 2000. Government Performance and Political Support. Dr.polit.-dissertation in political science. Trondheim: The Norwegian University ofScience and Technology.

    Klingemann, Hans-Dieter. 1999. Mapping Political Support in the 1990s: A GlobalAnalysis. In Norris, Pippa. (ed.): Critical Citizens. Oxford: Oxford UniversityPress.

    Klingemann, Hans-Dieter and Dieter Fuchs (eds.).1995: Citizens and the State.Oxford: Oxford University Press.

    Listhaug, Ola. 2005. Oil Wealth Dissatisfaction and Political Trust in Norway: AResource Curse? West European Politics 28: 834-851.

    Listhaug, Ola. 2006. Political Disaffection and Political Performance: Norway 1957-2001. In Torcal, Mariano and Jose Ramon Montero (eds.): Political

    Disaffection in Contemporary Democracies. Oxford: Routledge.

    Listhaug, Ola and Bernt Aardal. 2003. Support of Democracy in Europe. Paperprepared for the conference on Democracy in the New Europe, Institutdtudes Politiques de Paris de Dijon, November 13-16.

    Listhaug, Ola and Mattti Wiberg. 1995. Confidence in Political and PrivateInstitutions. In Klingemann, Hans- Dieter.and Dieter Fuchs (eds.): Citizensand the State. Oxford: Oxford University Press.

    Miller, Arthur H. 1974. Political Issues and Trust in Government: 1964-1970.American Political Science Review 68: 951-972.

    Miller, Arthur H. and Ola Listhaug. 1990. Political Parties and Confidence inGovernment: A Comparison of Norway, Sweden and the United States.British

    Journal of Political Science 20: 357-386.

    Miller, Arthur H. and Ola Listhaug. 1998. Policy Preferences and Political Trust: AComparison of Norway, Sweden and the United States. Scandinavian PoliticalStudies 21: 161-187.

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    18/24

    18

    Norris, Pippa (ed.). 1999a. Critical Citizens. Oxford: Oxford University Press.

    Norris, Pippa. 1999b. Introduction: The Growth of Critical Citizens? In Norris, P.(ed.): Critical Citizens. Oxford: Oxford University Press.

    Olsen, Johan P. 2002. Reforming European Institutions of Governance.Journal ofCommon Market Studies 40: 581-602.

    sterud, yvind. 2005. Introduction: The Peculiarities of Norway. West EuropeanPolitics 28: 705-720.

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    19/24

    19

    Tables and figures

    3.00 4.00 5.00 6.00 7.00 8.00

    Trus t in legal system, country mean

    1.00

    2.00

    3.00

    4.00

    5.00

    6.00

    7.00

    Trustinelectoralsystem,countrymean

    AT

    BE

    CH

    CZ

    DE

    DK

    EE

    ES

    FI

    FR GB

    GR

    HU

    IE

    IS

    LU

    NLNO

    PL

    PT

    SE

    SI

    SK

    UA

    Figure 1. Trust in electoral and legal system, gross country differences.

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    20/24

    20

    Table 1. Political trust in 2004, country means.A. Trust in country'sparliament B. Trust in politicians

    Denmark 6.286 Denmark 5.589

    Finland 6.009 Luxembourg 5.178

    Iceland 5.924 Iceland 4.971

    Luxembourg 5.762 Finland 4.877

    Switzerland 5.517 Switzerland 4.770

    Norway 5.424 Netherlands 4.690

    Sweden 5.351 Norway 4.244

    Spain 5.089 Belgium 4.240

    Ukraine 4.797 Sweden 4.191

    Austria 4.775 Ireland 3.924

    Ireland 4.713 Ukraine 3.736

    Greece 4.687 Spain 3.685

    Belgium 4.682 Greece 3.595

    Netherlands 4.668 United Kingdom 3.589United Kingdom 4.288 France 3.486

    France 4.269 Estonia 3.315

    Germany 4.214 Austria 3.254

    Estonia 4.191 Germany 3.230

    Slovenia 4.127 Slovenia 3.099

    Portugal 3.719 Czech Republic 2.727

    Hungary 3.634 Hungary 2.682

    Czech Republic 3.186 Slovakia 2.526

    Slovakia 3.051 Portugal 2.058

    Poland 2.405 Poland 1.917

    C. Trust in political parties D. Trust in the legal systemDenmark 5.647 Denmark 7.214

    Finland 4.996 Finland 6.897

    Luxembourg 4.971 Norway 6.352

    Iceland 4.885 Switzerland 6.140

    Netherlands 4.800 Luxembourg 6.137

    Switzerland 4.636 Iceland 6.009

    Sweden 4.398 Austria 5.829

    Norway 4.340 Sweden 5.769

    Belgium 4.286 Germany 5.542

    Ireland 3.973 Netherlands 5.496

    United Kingdom 3.676 Greece 5.382

    Spain 3.668 Ireland 5.205

    Ukraine 3.611 United Kingdom 5.116

    Greece 3.505 Estonia 4.907

    France 3.397 Belgium 4.830

    Austria 3.396 France 4.766

    Slovenia 3.209 Spain 4.717

    Germany 3.184 Hungary 4.429

    Estonia 3.088 Portugal 3.939

    Czech Republic 2.745 Ukraine 3.910

    Hungary 2.711 Slovenia 3.849

    Slovakia 2.663 Czech Republic 3.720

    Portugal 2.085 Slovakia 3.583Poland 1.891 Poland 3.006

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    21/24

    21

    E. Trust in the police F. Trust in the European Parliament

    Finland 7.956 Ireland 5.370

    Denmark 7.936 Greece 5.344

    Iceland 7.279 Iceland 5.288

    Norway 7.131 Hungary 5.223

    Switzerland 6.858 Luxembourg 5.223

    Ireland 6.589 Spain 5.051

    Sweden 6.486 Finland 4.996

    Germany 6.477 Belgium 4.983

    Luxembourg 6.472 Estonia 4.874

    Austria 6.185 Denmark 4.833

    United Kingdom 6.116 Ukraine 4.826

    Greece 6.031 Slovakia 4.738

    Netherlands 5.967 Switzerland 4.609

    Spain 5.906 Netherlands 4.606

    Belgium 5.784 Norway 4.552

    Estonia 5.692 Slovenia 4.534France 5.661 Czech Republic 4.383

    Hungary 5.174 France 4.306

    Portugal 5.056 Poland 4.261

    Slovenia 4.710 Germany 4.183

    Poland 4.576 Portugal 4.037

    Slovakia 4.352 Austria 4.022

    Czech Republic 4.226 Sweden 3.955

    Ukraine 3.299 United Kingdom 3.548

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    22/24

    22

    Table 2. Descriptive statistics for the variables to be analysed, minimum n=29455a

    Variables Description Min. Max. Mean Std.

    Political trust

    telect Trust in political, electoral system, three items 0 10 3.972 2.171

    tlegal Trust in legal system and police 0 10 5.629 2.336eptrust Trust in EU parliament 0 10 4.503 2.392

    Controls

    age Age in years 12 99 44.519 17.436

    age2 Age in years squared 144 9801 2285.981 1654.714

    male Gender, 1=male, 0=female 0 1 0.479 0.500

    eduyrs Years of full-time education completed 0 26 12.075 3.753

    Political distance

    votegov Voted for party in government, 3 categories

    1. Voted for party in government 0 1 0.313 0.464

    2. Did not vote, don't know 0 1 0.374 0.484

    3. Did not vote for party in government 0 1 0.313 0.464

    lrscale5 Left-right scale in six categories1 0-2 Left 0 1 0.091 0.288

    2 3-4 Left 0 1 0.190 0.392

    3 6-7 Right 0 1 0.187 0.390

    4 8-10 right 0 1 0.127 0.333

    5 88 Don't know 0 1 0.088 0.283

    6 5 Middle (reference category) 0 1 0.317 0.465Political issues

    imm_good Scale, immigration good for country 0 10 5.083 2.063

    env_impEnvironment important, 1=important, 0= not soimportant 0 1 0.697 0.460

    incdiff

    Agree government should reduce income

    differences 1 5 3.761 1.064

    euftfEuropean Union: European unification gofurther or gone too far 0 10 5.349 2.648

    Performance evaluations: for person and country

    stflife How satisfied with life as a whole 0 10 7.068 2.173

    comfinc Living comfortably on present family income 0 1 0.771 0.420

    stfecoHow satisfied with present state of economy incountry 0 10 4.797 2.382

    stfedu State of education in country nowadays 0 10 5.617 2.293

    stfhlth State of health services in country nowadays 0 10 5.260 2.516Country level variables (n=24)

    Lnpop Natural logarithm of total population 12.574 18.229 15.961 1.450

    HDI Human development index 0.780 0.960 0.908 0.045Ferrera Country classification

    1 Bismarckian (AT, BE, CH, DE, FR, LU, NL) 0 1

    2 Nordic (DK, FI, IS, NO, SE) 0 13 Anglo-Saxon (IE, GB) 0 1

    4 Southern (ES, GR, PT) 0 1

    5Eastern (CZ, EE, HU, PL, SI, SK, UA)(reference category) 0 1

    a Min.: minimum value, Max.; maximum value, Std.: standard deviation. N, thenumber of individuals varies from 29455 to 31873.Table 3 Variance components from a multilevel regression analysis of political

    trust in Europe.a

    Trust in electoral system Model 0 Model 1 Model 2 Model 3 Model 4

    Se 3.670 3.631 2.706 2.706 2.706

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    23/24

    23

    Su 0.897 0.866 0.282 0.265 0.196

    Explained Se 0.000 0.011 0.262 0.262 0.262

    Explained Su 0.000 0.035 0.686 0.704 0.782

    Intraclass correlation 0.196

    -2 Log likelihhod 132929.5 132588.3 123138.4 123137.0 123130.1

    Trust in legal system

    Se 4.136 4.109 3.282 3.282 3.282

    Su 1.198 1.157 0.444 0.302 0.207

    Explained Se 0.000 0.006 0.207 0.207 0.207

    Explained Su 0.000 0.034 0.629 0.748 0.827

    Intraclass correlation 0.225

    -2 Log likelihhod 138153.8 137943.9 130637.3 130628.5 130619.9

    Trust i European Parliament

    Se 5.296 5.181 4.124 4.124 4.124

    Su 0.235 0.229 0.318 0.165 0.144

    Explained Se 0.000 0.022 0.221 0.221 0.221Explained Su 0.000 0.022 0

    b0.296 0.385

    Intraclass correlation 0.042

    -2 Log likelihhod 135063.7 134404.7 127582.5 127567.51 127564.5

    a Model 0: Only intercept, Model 1: M0+Demographic variables, Model 2:M1+Attitudinal variables, Model 3: M2+ country variables, Model 4: M2+ countryclassification.Se: within country variance, Su: between country variance, Explained: proportion ofthe variance in the null model explained by models 1-4.All variance components are statistical significant at the .01 level.b

    Due to an increase in the between country variance component this value is set tozero.

  • 8/2/2019 Trust in Political Institutions. the Nordic Countries Compared With Europe

    24/24

    Table 4. A multilevel analysis of political trust in Europe, fixed regression

    coefficients from model 3.*Trust inelectoralsystem

    Trust in legalsystem

    Trust in EUparliament

    Variables Description B Sig. B Sig. B Sig.

    Intercept Regression constant -0.248 0.929 -5.508 0.075 11.105 0.000

    Age Age in years -0.025 0.000 -0.006 0.034 -0.053 0.000

    age2 Age in years squared 0.000 0.000 0.000 0.014 0.000 0.000

    Male Gender, 1=male, 0=female -0.086 0.000 -0.104 0.000 -0.178 0.000

    Eduyrs Years of full-time education 0.009 0.001 0.014 0.000 0.010 0.005

    Votegov Voted for government party? 0.000 0.000 0.000

    1 Did not vote, don't know -0.201 0.000 -0.156 0.000 -0.035 0.264

    2 Yes 0.178 0.000 0.091 0.001 0.157 0.000

    3 No 0.000 - 0.000 - 0.000 -

    Lrscale5 Left-right scale 0.000 0.000 0.000

    1 0-2 Left -0.078 0.027 -0.300 0.000 -0.033 0.460

    2 3-4 Left 0.090 0.001 -0.053 0.077 0.144 0.0003 6-7 Right 0.161 0.000 0.042 0.162 0.138 0.000

    4 8-10 right 0.204 0.000 0.027 0.430 0.200 0.000

    5 88 Don't know -0.260 0.000 -0.230 0.000 -0.330 0.000

    6 5 Middle (reference category) 0.000 - 0.000 - 0.000 -

    imm_good Scale, immigration good for country 0.122 0.000 0.109 0.000 0.121 0.000

    env_imp Environment important 0.065 0.002 0.050 0.029 0.096 0.000

    IncdiffGovernment reduce incomedifferences -0.017 0.080 -0.030 0.004 0.001 0.940

    Euftf EU unification go further 0.059 0.000 0.040 0.000 0.204 0.000

    Stflife How satisfied with life as a whole 0.014 0.008 0.095 0.000 0.016 0.014

    Comfinc Living comfortably on family income 0.027 0.288 0.090 0.001 0.054 0.099

    StfecoHow satisfied with economy incountry 0.255 0.000 0.180 0.000 0.193 0.000

    Stfedu State of education in country 0.111 0.000 0.131 0.000 0.126 0.000

    Stfhlth State of health services in country 0.115 0.000 0.123 0.000 0.077 0.000

    Lnpop Natural logarithm of total population -0.048 0.552 -0.009 0.916 -0.068 0.292

    HDI Human development index 2.214 0.376 8.360 0.004 -8.943 0.000

    Ferrerab

    Country classification 0.071 0.001 0.001

    1 Bismarckian 0.574 0.030 0.891 0.002 -0.580 0.013

    2 Nordic 0.686 0.015 1.301 0.000 -0.969 0.000

    3 Anglo-Saxon -0.008 0.983 0.588 0.123 -0.866 0.010

    4 Southern 0.409 0.196 0.808 0.018 0.085 0.752

    5 Eastern (reference category) 0.000 - 0.000 - 0.000 -

    N Number of respondents 31565 31873 29455a

    B: regression coefficient in metric scale, Sig.: the probability value of B based on the t-statistic. The probability values of Votegov, Lrscale5and Ferrera, are based on the Waldstatistic for each set of coefficients.bCoefficients for country classification from model 4.