Economic Performance and Political Trust: The impact of the financial crisis on European citizens

The bankruptcy of the investment bank, Lehman Brothers, in the fall of 2008 marks the beginning of the ongoing financial crisis. Due to increased financial globalization, a crisis that originally started with the collapse of the US housing market soon spread to the rest of the world, Europe being no exception. The crisis resulted in negative economic growth, high levels of unemployment, and weak macroeconomic prospects for a number of European countries. Following the crisis the frustration amongst European citizens is rising. As the European countries have responded to the crisis with austerity policies, thousands of people have hit the streets and demonstrated against their governments. As illustrated in the opening quote, Greece is one of the countries that have been hardest hit.

The crisis has caused millions of people to lose their jobs, and it has set the world economy back years (Krugman 2009). In this article we set out to disclose both the relationship between economic performance and political trust, as well as financial crisis’ effect on levels of trust. Several studies on how economic performance affects levels of political trust have focused on economic performance in general. However, others have stressed that economic performance might have a different effect in a crisis, compared to a non-crisis. Our contribution is to test how an economic crisis, in this case the current financial crisis, affects levels of political trust.

What makes economic performance important to levels of trust? And why are high levels of trust important? People trust governments more if they have shown a capacity to generate economic growth, create jobs, provide access to social services and perform in a transparent manner (Fiorina 1978; Mackuen, Erikson, and Stimson 1992). If there is doubt that the authorities are capable of managing the national economy and respond to the challenges of an economic crisis, it may create mistrust. Because political trust is essential for the stability of democracy and a basic prerequisite for the legitimacy of those entrusted with political power, it is important that there is not a decline in the level of trust (Hardin 1998; Levi 1998). High political trust signals that the institutions function properly and effectively, and helps to keep democracy alive (Listhaug and Ringdal 2008).

While several studies have been concerned with the nature of and variation on individuals’ political trust (e.g. Levi 1998; Miller and Listhaug 1999; Norris 1999; Uslaner 2002; Dalton 2004; Newton 2008), we will focus on the political trust shared by the population of a given country. The data employed are from the European Social Survey (2004, 2006, 2008, and 2010). Between the collection of the 2008 and 2010 rounds, the financial crisis had made a global impact, which makes the use of 2010 data compared to the earlier rounds advantageous for our purpose. It should however be mentioned that much previous research finds that political trust has been decreasing for some time (Dalton 2004). Rounds 2004 and 2005 are included to reduce the possibility that the evidence found in our analysis only is a continuation or exacerbation of an earlier trend. The method employed is a multilevel regression analysis with three levels; individual, country-year, and country. To measure economic performance we use the variables GDP GROWTH RATE, GDP PER CAPITA, and GROSS DEBT IN PERCENT OF GDP.

Political trust is essential for the stability of democracy, and a basic prerequisite for the legitimacy of those entrusted with political power (Hardin 1998; Levi 1998). The concept of political trust is multidimensional, and different terms are often used to describe the same phenomenon. Political trust can be divided into three forms: thick, interpersonal, and systemic or institutional trust (Levi 1998; Newton 2001). In this study we are concerned with the third form of trust, because this part of the phenomenon is connected to political institutions. Another way to classify political trust is according to Norris (1999) in a hierarchy that ranges from specific to general. Norris’ fivefold conceptualization draws a line between political actors, regime institutions, regime performance, regime principles, and political community. Political community is the most diffuse level while political actors is the most concrete. Using this classification we have focused on the two most specific levels of trust; regime institutions (TRUST IN POLITICAL PARTIES and TRUST IN COUNTRIES PARLIAMENT) and political actors (TRUST IN POLITICIANS). Sztompka (1999) views trust as a strategy to deal with an uncertain an unpredictable future. If one has a functioning set of institutions the danger for the individual is reduced, and trust level are high. According to Hardin (2009) conceptions of distrust is not simply the opposite of conceptions of trust. However, distrust can, just as trust, be viewed as a rational response to protect individuals against harm. Distrust in government has led to the establishment of checks and balances to limit government power. Further, Hardin (2001: 495) describes both trust and distrust as a three-part relation ‘A distrusts B with respect to X. A might trust B on many matters, but not on others’. Levi (1998) argues that citizens are more likely to comply with norms if they perceive the government as credible. Thus, in order to produce and achieve outcomes it is important that trust is present. A decline in confidence can have negative consequences for democracy and the political institutions. High political trust signals that the institutions function properly and effectively (Listhaug and Ringdal 2008), and it also facilitates social and economic exchange and reduces transaction costs in markets. Trust reduced the need for control and supervision, which saves money for the government as well as for firms and other actors in the private sector. This implies that countries with high levels of trust will have an advantage in attracting investments, trade, and tourism.²

Politicians often speak of a crisis in democracy due to constantly declining trust in government. However, it is not certain whether empirical evidence support this conclusion (Norris 1999). According to Newton (2001: 205) an abrupt or steady decline of trust from the long-term trends should be regarded as worrying. Dalton and Wattenberg (2000) find that the trust in government and political institutions has been falling in advanced industrialized democracies since the late 1960s. They also show that in industrialized democracies trust in political parties is eroding. Related to this, public confidence in parliaments has similarly decreased in the last decade (Dalton 2004).

However, research that rejects the hypothesis of a universal decline of trust in the public sector also exists (e.g. Van de Walle, Van Roosbroek, and Bouckaert 2008). These authors argue that there is little evidence of an overall long-term decline in trust in government, and that there are fluctuations rather than a stable trend. McAllister's (1999) research suggests that there are few consistent trends in popular support for the political community, but high and perhaps even growing support for democratic values. But he also indicates that there is a declining support for regime institutions and political leaders.

Table 1 shows the mean values of political trust in the sample countries in 2004, 2006, 2008, and 2010. Political trust is a scale created by collapsing the three measures TRUST IN COUNTRIES PARLIAMENT, TRUST IN POLITICAL PARTIES, and TRUST IN POLITICIANS, each ranging from 0–10. This variable includes both regime institutions and political actors.

The table shows that 15 out of the 25 countries in the sample have experienced decreased trust between 2008 and 2010. Eight countries are experienced increased trust, while two countries have the same levels of trust. Out of the 15 countries that experienced a decrease in trust, it is only Cyprus, Denmark, Spain, Finland, Greece, Croatia, Ireland, and Slovenia we can argue have had an abrupt decrease after the financial crisis. In Belgium, a small decline in trust has been evident for each year the survey has been conducted. However, the decline from the before and after measures is not severe enough for it to be argued that it stems from the financial crisis. In Germany, France, Portugal, Russia, Slovakia, and the UK, trust has both decreased and increased through each survey-year.

Figure 3 illustrates trends in mean political trust in the sample countries when categorized by subgroups. The levels of trust in the Mediterranean countries drops severely between 2008 and 2010, a fall that it is possible to argue is an effect of the crisis.

Several authors relate trust to good performance of government. The performance approach has two main parts: the first deals with macro-performance, like unemployment, economic growth, and inflation (Brown and Coulter 1983; Kornberg and Clarke 1994; Anderson 1995; Zussman 1997; Miller and Listhaug 1999); while the second deals with micro-performance, that is, government service delivery (Norén 2000; Rose and Pettersen 2000). If there are variations in trust across countries, this may be due to variations in unemployment rates, the stability of governments etc. (Listhaug and Wiberg 1995). However, Blind (2006) argues that symptoms do not explain the causes of declining trust, and many different factors may be behind a decline. Public fear that governments are incapable of dealing with previous current fiscal and financial challenges, in addition to period of low economic growth, could be causes of a decline (Blind 2006; Mansbridge 1997; Newton and Norris 2000). It can also be argued that the current crisis is an effect of long-term structural incapacities of states in Southern and Eastern Europe, and that the crisis has just captured this underlying effect. Putnam (1993) viewed trust as the explanatory factor predicting the condition of politics and the economy. With regard to why the north of Europe had higher levels of trust than the south was explained by the absence of wars and turmoil in the former. Papakostas (2012) has investigated the reverse causal relationship, that is, what is the reason for different trust levels. He highlights the importance of a well-functioning public sphere and the strength of the state as factors influencing trust as a dependent variable. This argument is supported by several others who state that high social trust is associated with a strong state with stable democracy, little corruption and a low degree of economic inequality (Inglehart 1999; Rothstein 2005).

In order to answer our research question, we will focus on macro-performance and its impact on political trust. What governments are able to accomplish is important for political trust. This is especially true for economic performance (Listhaug and Ringdal 2008; Nannestad and Paldam 1994). People trust governments more if they have shown a capacity to generate economic growth, create jobs, provide access to social services, and perform in a transparent manner (Fiorina 1978; Mackuen et al. 1992). When the government works well, citizens have confidence in those who govern, whether this is institutions, people, or political parties. In a recession, however, the opposite will occur, and support for the government will be affected in a negative way (Nannestad and Paldam 1994; Miller and Listhaug 1999; Uslaner 2002; Listhaug and Ringdal 2008; Roth, Nowak-Lehman, and Otter 2011). According to Fiorina (1978), Mackuen et al. (1992), and Roth et al. (2011), economic growth and access to work and social services makes citizens rely more on the government. If there is any doubt that the authorities are capable of managing the national economy, or respond to possible challenges in the event of an economic crisis, it may create mistrust. An argument against this is that, if trust has in fact declined, it has done so over a period of economic growth. However, political trust is influenced by many factors, and it may be the case that if growth had not increased, the levels of growth would have been even lower. Nye (1997) argues that citizens’ doubts regarding their national economy and governments’ ability to respond to these challenges could create even more distrust now, compared to earlier, as the economy is becoming more and more globalized. According to Listhaug and Wiberg (1995) a lower gross national income will lower confidence in political institutions. From these arguments we can deduct the following hypotheses:

Several researchers find that increased unemployment correlates with reduced political trust (e.g. Earle 2009; Roth et al. 2011). Research on the effects of increased unemployment from earlier periods also finds that many countries that have experienced an increase in unemployment levels also have experienced reduced trust in politicians, political parties, and political institutions (Listhaug and Wiberg 1995). Veiga and Veiga (2004) performed a case study on Portugal, and found a strong effect of unemployment on a government's popularity. Another study of the UK by Sanders (2000) shows that voters decide on the basis of government results regarding unemployment and inflation, as well as on expectations about the economic future. From this we can assume that in countries where the unemployment rate is high, levels of political trust are low. This would reflect citizens’ dissatisfaction with government's capability to manage the national economy, and respond for example to an economic crisis. From this reasoning our third hypothesis appears:

In a recent study on trust in political institutions Roth et al. (2011) investigates the relationship between debt and trust in political institutions. They looked at 27 European countries with data from both before and after the financial crisis, and concluded that an increase in GDP debt reduces trust. Increased government debt often decreases spending, and when states tighten budgets and the banks become less willing to give out loans, the economic activity goes down. This happens in a situation where economic growth is already low and unemployment high. Therefore, it is reason to believe that for countries where the debt has increased, the citizen's trust will be reduced. Hence, the following hypothesis is deducted:

An alternative hypothesis could, as previously mentioned, be that the financial crisis captures states’ long-term structural incapacities, creating the stagnation that leads to a decline in trust.

The European experience of the 2007–2008 financial crisis was experienced differently from one country to another. To be able to see if economic performance affects levels of trust, we need to control for the actual economic situation by employing macroeconomic indicators. We argue that by looking at the actual situation in the economy, and not only individuals’ perceptions of the economic situation, we are able to get a greater understanding of how economic performance influences political trust.

GDP is an indicator for total value added in a country, and also provides us with an expression of gross income from domestic production activity (World Bank 2013). From Table 2 we see that GDP per capita is lower in 13 out of 25 countries in 2009 compared to 2007.

GDP growth is the annual percentage growth rate of GDP per capita based on constant local currency (World Bank 2013). Table 3 shows that all countries in our sample had a lower growth in 2009 compared to in 2007. Out of the 25 countries, 22 of them have gone from having a positive to a negative growth rate. Some of the countries experienced a more severe decrease than others, but the overall trend is alarming.

According to Krugman (2012), current unemployment has reached levels that would have seemed inconceivable before the crisis. Low unemployment indicates that there is pressure in the labor market. Wages increase, which may lead to price pressure, high production and generally good times. High unemployment indicates the opposite. Trends in the unemployment rate (World Bank 2013) reflect the state of the economy, but the effect does not manifest itself immediately. From Table 4 we read that out of the 25 countries in our sample 21 experienced increased unemployment.

Because the financial crisis has culminated into a sovereign debt crisis from 2010 and onwards in Europe, it is important to control for the effect of this as well (Roth et al. 2011). We have therefore chosen to include gross debt in percentage of GDP (IMF 2013) in the analysis. Twenty of the 25 countries in our sample have experienced an increase in general government gross debt (see Table 5).

The four measures presented can be viewed as proxies for the extent to which each country have experienced crisis. For example, the Norwegian economy has reacted differently compared to Spain's economy. As a result, the level of political trust may be affected differently from one country to another. The economic measures are thus included in our analysis to capture this cross-national variation of the effect of the crisis.

In order to test our hypotheses we perform a multilevel analysis of pooled data from four rounds of the European Social Survey (ESS) covering 25 countries (Jowell and The Central Co-ordinating Team 2005).³ The data for the economic indicators are gathered from World Bank (2013). The use of multilevel modeling is advantageous for this analysis as the data are of a hierarchical structure, as individuals can be considered a subset of country-years, which are considered a subset of countries. When using multilevel analysis we are able to account for variance in a dependent variable measured at the lowest level, by taking the higher levels into consideration (Steenbergen and Jones 2002: 219). In our models the respondents constitute level-1, the country-years level-2, and countries level-3.

We make a distinction between micro and macro explanatory variables. Several scholars (Weatherford, 1987; Lockerbie, 1993; Anderson and Guillory, 1997; Taylor, 1997; Miller and Listhaug 1999) have demonstrated the individual-level relationship between perceptions of government economic performance and levels of political support. According to Huseby (1999) it has been harder to demonstrate a direct relationship between indicators of real economy like unemployment and inflation, and perceptions of economic performance and political trust on an aggregated level. However, to measure the effect of the financial crisis on political trust, we see it necessary to focus mainly on real economic performance using unemployment, GDP per capita, GDP debt, and GDP growth. To avoid Hauser's contextual fallacy,⁴ we have also controlled for individual-level indicators.

The dependent variable POLITICAL TRUST (0–30) is a scale made up by collapsing three measures: TRUST IN COUNTRY'S PARLIAMENT, TRUST IN POLITICIANS, and TRUST IN POLITICAL PARTIES.⁵ A weakness with this operationalization is that it does not include the public administration. These regulatory agencies can also be perceived to have played a part in the economic crisis. People are able to differentiate between the different types of institutions, and Johnson et al. (1998) state that undisciplined bureaucracies creates unfavorable conditions with regard to trust. However, the present dependent variable is the best available in the WVS data. At the individual level we have included the measures POLITICAL INTEREST (1–4), VOTED LAST NATIONAL ELECTION (0–1), UNEMPLOYED (0–1), FEELING ABOUT HOUSEHOLD INCOME (1–4) where high values indicate a positive view, WOMEN (0–1), and AGE. At the country-year level we have included four variables, all of which are economic indicators on the state of economy in a country. These are gathered from the World Bank (2013) and the International Monetary Fund (2013). We have lagged these indicators one year as we argue that it takes some time for the effects to materialize into a change in individual-level trust, and coded them so they show the change from two years ago (in accordance with our sample), and they are as follows GDP PER CAPITA, GDP GROWTH, GDP GROSS DEBT, and UNEMPLOYMENT.

where β₁, β₂ … β_n is the slope for level-1 and level-2 variables X_1ijk, X_2ijk, X_nijk

In Models 1 through 4 we add the macro measures one by one. The last model is the most complex one, and includes all the individual-level variables and all the macroeconomic measures together. We see from Table 6 models 1 through 4 that an increase in GDP GROWTH is associated with higher trust levels, and an increase in UNEMPLOYMENT and GDP GROSS DEBT leads to less political trust. There is no effect for change in GDP PER CAPITA.

Two of the measures are significant in the final model: change in growth and change in debt. This indicates that in countries where growth has decreased between 2008 and 2010, citizen's trust has also decreased. This is in line with our second hypothesis which stated that a decrease in GDP growth rate will reduce citizens’ political trust. The table also shows that an increase in debt is linked with a decrease in trust, thus rendering support to our fourth hypothesis: an increase in GDP gross debt leads to reduced political trust.

The effect of a change in GDP PER CAPITA seen in Model 2, indicates a positive effect on trust. The effect has the same direction in Model 5. However, since the effect is not statistically significant in either model, we cannot confirm our first hypothesis: A decrease in GDP per capita will result in reduced political trust. A change in UNEMPLOYMENT indicates that a decrease is associated with increased trust. The result is significant in Model 3. However, when controlling for the other economic measures the effect is no longer significant. In sum, we can only say that there is some support for our third hypothesis: An increase in unemployment reduced citizen's political trust.

It is a possibility that the reason why changes in GDP and unemployment are not significant is because the data used to measure the effect is to lose in time to the crisis. It is a known phenomenon that unemployment moves counter-cyclical with GDP and is a lagging indicator. Yet, since UNEMPLOYMENT is significant when standing as the sole macro indicator (see Model 3), another possibility may be that the effect of unemployment to some extent is shown through the other economic indicators. To summarize, by including measures on economic performance, political trust is better explained compared to when they are not included. Also, from this it is evident that the financial crisis has had an impact on citizen's confidence in political institutions.

All the level-1 variables are significant in our models. First, high political interest generates high trust levels, and second, citizens who participate politically, by voting in national elections, have higher levels of political trust. The results of UNEMPLOYED indicates that citizens without work, but who are looking for a job, have less trust in political institutions compared to people who are either employed, unemployed, and not looking for a job, permanently sick, or disabled, under education, community, or military service. The demographic variables education, gender, and age also seem to influence the degree of political trust. Increased years of education increase trust, and women have higher levels of trust compared to men. FEELING ABOUT HOUSEHOLD INCOME NOWADAYS has a positive correlation with POLITICAL TRUST, indicating that respondents that are living comfortably on present income have higher levels of political trust compared to respondents feeling that it is very difficult living on present income. The effect of AGE in Model 5 is a decrease up until 70.5 years, and then it starts increasing again.

The time control variables 2004, 2006, and 2008 measures levels of trust in these years compared to 2010. It shows the unexplained variance in political trust when controlling for the other micro- and macro-variables. The effect for 2004 indicates in each of the models, that the trust level is higher in 2004 compared to 2010. The same effect manifests itself between 2006 and 2010 as well. For the comparison between 2008 and 2010 the effect varies when controlling for different measures. In Model 2, where none of the economic measures are included, as well as Model 4 and Model 6, the effect is positive, indicating that levels of trust are higher in 2008. However, when controlling for growth level in Model 3 and unemployment level in Model 5, the effect indicates that trust-levels are lower in 2008 compared to 2010. In the final model the dummy variable 2008, indicate that levels of trust are lower in 2008 compared to 2010. Meaning that the unexplained variance in Y, the variance not explained in the final model, gives increased trust from 2008 to 2010. The effect of the 2008 dummy variable fluctuates around zero and is not significant.

Our findings suggest that economic performance, which is influenced by the financial crisis, does affect political trust. When the economy is doing well, levels of political trust are higher than when the economy is weak. From the tables presented in this article, it is evident that many European countries have experienced a negative shift in the economy. From our regression models we see that GDP growth rate, which is often used to assess the performance of the economy, does affect political trust. We can therefore conclude that citizens in countries that have experienced negative or reduced growth in GDP, as a consequence of the crisis, will have lower levels of confidence.

When an economy is affected by a crisis, economic output declines. One would expect that reduced GDP per capita, would influence the levels of political trust. However, we found no significant relationship in our models. When it comes to unemployment, the results are not clear. When controlling for the change in unemployment as the only economic indicator in the model, the results indicate that the financial crisis has had an effect on political trust. However, when controlling for the other economic measures the effect of unemployment is weaker and no longer significant. When it comes to the levels of debt in percentage of GDP the results indicate that countries with higher debt have lower levels of trust. Still, this finding is not statistically significant.

Blind (2006) argues that symptoms do not explain the causes of declining trust, and that many different factors may be behind such a decline. One factor behind a decline might very well be a concern that the government is incapable of dealing with current fiscal and financial challenges. It is a possibility that people are not just upset with the economic outcomes, but also in the way European policy-makers have handled the crisis. Another aspect is that the economic crisis has hit some countries harder than others. Around 10% of the variance on political trust is between countries. According to Roth (2010), the levels of trust have been reduced more in Spain and Portugal than in France and Germany. As seen in Figure 3, the Mediterranean countries have felt the severity of the crisis to a larger extent than for example the Scandinavian countries.

In this study we have chosen to focus on certain measures meant to capture the impact of the financial crisis. Yet, it is important to stress that also political and economic systems could influence states’ capacity in controlling the crisis. Future research would be well advised to pick up on this thread.

Trust links ordinary citizens to institutions that are intended to represent them (Bianco 1994), and thereby enhancing both the legitimacy and the effectiveness of democratic government (Gamson 1968; Braithwaite and Levi 1998; Hetherington 1998). It is therefore an essential part of a democratic system. The findings presented in this article suggest that the financial crisis has in fact reduced citizen's trust in political institutions, at least in those countries where the crisis has made most impact. By employing a multilevel analysis we have been able to investigate several European countries.

In sum, this study contributes to the understanding of the relationship between economic performance and political trust, as well as how citizen's trust is affected by an economic crisis. A severe economic crisis must be taken seriously by policy makers, as it can contribute to reducing citizen's confidence in and support of political institution. It is their responsibility as elected representatives to lead the economy out of the recession. If they do not handle the economic situation in an appropriate manner, this can provide a foundation for higher discontent. At present, several years after the start of the crisis, many Europeans are still unemployed or underemployed. According to Krugman (2012), at recent rates of reaction we will not be back to normal levels of employment until around 2020.

No potential conflict of interest was reported by the authors.

Veronica Fagerland Kroknes (b. 1989) is working as an analyst and advisor at the Norwegian Tax Administration. She received her masters degree in Political Science in 2013 from Norwegian University of Science and Technology (NTNU).

Tor Georg Jakobsen (b. 1980) is Professor of Political Science at Trondheim Business School, Norwegian University of Science and Technology (NTNU). He received his Ph.D in 2011 from the NTNU. Jakobsen has authored and co-authored articles in, among others, European Sociological Review, Work, Employment and Society, and Social Indicators Research.

Lisa-Marie Grønning (b.1986) is working as a case-worker and advisor at the Norwegian Labour and Welfare Administration. She received her Master's Degree in Political Science in 2013 from the Norwegian University of Science and Technology (NTNU).