Abstract
It is well documented that the public is often poorly informed about the economy. In the domain of fiscal policy, this may make voters susceptible to favour spending, while underestimating its costs (fiscal illusion). While politicians may have more information to judge the need for prudent economic policies, voters may be less inclined to support prudent fiscal policy if they do not believe that these politicians act in their best interest—an idea that in recent decades has become more prevalent. Using a novel dataset from the Netherlands, this paper examines whether strong populist ideas lead to more expansionary fiscal preferences, thereby reinforcing the risk of fiscal illusion. The findings indicate that respondents' populist attitudes significantly predict their fiscal preferences. Additionally, higher literacy and information provision contribute to more prudent fiscal preferences. However, the impact of literacy is conditioned by the level of populist sentiment. Poorly literate respondents show higher support for tax relief only when holding strong populist attitudes, not when expressing lower levels of populist sentiment.
Similar content being viewed by others
1 Introduction
For some decades now, countries across the world have witnessed a surge in the vote share of parties that challenge the political elite. One of the core features that many of these challenger parties of both the left and the right have in common is their populist set of ideas (Rooduijn et al., 2019). This set of ideas holds that there is an antagonistic division between the ordinary people and an ‘evil elite’ which is not acting according to the popular will as it should, and can be present in the rhetoric of political parties, but also in the mind of individuals (Hawkins et al., 2018; Mudde, 2004). Recently the literature on populism has started to investigate the consequences of the prevalence of populism for economic policies (see Afonso, 2015; Otjes & Louwerse, 2015; Stöckl & Rode, 2021).
A widely held premise of economists is that populism is a threat to sound economic policymaking and sound public finances in particular (Andersen et al., 2017; Davidson, 2018; Dornbusch & Edwards, 1991; Guiso et al., 2017). When people believe the political elite to be financially or morally corrupt (i.e., when they hold populist attitudes), they are less likely to support unpopular measures than when they believe that the elite acts according to the people’s interest. As a result, people with populist attitudes will be more likely to conceive public policy as being complex and budgets as non-transparent. This may result in fiscal illusion, i.e. voters appreciating spending programs, but underestimating the (future) costs in terms of taxation or debt (Alesina & Perotti, 1995; Buchanan & Wagner, 1977).
Economists have expressed concerns about the viability of a populist economic agenda. Yet, they have not explored the link between populist sentiments and fiscal preferences. A more rigorous assessment is needed for at least two reasons. First, the empirical support for fiscal illusion—and the deficit bias that goes with it—is not strong, especially in those countries where voters are more sophisticated (Alesina & Passalacqua, 2016; Eslava, 2011). Second, recent research has highlighted various examples of economic policies that have been unresponsive to the concerns of a large share of voters. This goes, most notably, for the disruptive labour market effects of trade with China, which have been found to play a causal role in the electoral success of populist parties (Autor et al., 2020; Colantone & Stanig, 2018, 2019). In a similar vein, populist parties may also cater to voters who hold intrinsically more expansionary preferences than the political elite, e.g. due to a less benign socioeconomic position. In line with this, Piketty (2020) mocks the tendency of mainstream political actors to label parties ‘populist’ merely because they propose policies that they deem too radical.
The aim of this paper is to evaluate the impact of increased prevalence of populist ideas for fiscal preferences of voters. More specifically, we empirically assess whether people with strong populist ideas also report more expansionary fiscal preferences, and to what extent populist attitudes reinforce the risk of fiscal illusion. Our survey is set in the Netherlands, a country with a strong tradition of prudent fiscal policy and a literate population which on average holds relatively conservative fiscal preferences (European Commission, 2010). The survey was performed in 2017, at a time of strong discussions on fiscal policies (see more in Sect. 3). We measure fiscal preferences by asking respondents how they would use the tax windfalls that were foreseen at the time of the survey (September 2017): for debt reduction, tax relief and/or more spending. In turn, we measure the extent to which individuals adhere to populist ideas by a tested index of individuals’ ‘populist attitudes’ (Akkerman et al., 2014, see more below). We assess the role of fiscal illusion by including a measure of the literacy of respondents. In addition, we set up an information experiment in which we treat a random share of respondents with information about public debt dynamics.
This paper makes three main contributions. First, we examine to what extent fiscal preferences can be explained by populist attitudes at the individual level. To our knowledge, this has never been done before. We find that populist attitudes of respondents prove a very relevant predictor of their fiscal preferences. Second, in line with previous research, we find that literacy and information provision—which can alleviate the occurrence of fiscal illusion—contribute to more prudent fiscal preferences. Third, to assess whether populist attitudes reinforce the risk of fiscal illusion, we evaluate to what extent populist sentiment moderates the effect of literacy and information on fiscal preferences. We find that the effect of literacy is conditional on the level of populist sentiment when it comes to support for tax relief. To be precise, poor literacy only spurs support for tax relief when respondents hold relatively strong populist attitudes. Furthermore, as regard to support for more spending, we find that the effect of our information experiment is larger for respondents with stronger populist attitudes, suggesting that information provision can also alleviate fiscal illusion with voters with strong populist attitudes.
This paper proceeds as follows. The next section reviews the literature on populist attitudes and fiscal preferences and presents our hypotheses. Section 3 describes our research design. Section 4 presents the main results, i.e., the relationship between populist attitudes and fiscal preferences, as well as the impact of literacy and information provision. Among other robustness tests, Sect. 5 assesses whether our results are robust to an instrumental variable (IV) estimation in which we instrument populism with pre-crisis trust in national politics and the financial sector. Section 6 assesses to what extent populist attitudes reinforce the effect of literacy and information provision on fiscal preferences. Section 7 concludes. Supplementary material is included in the Online Appendix.
2 Selected literature review and hypotheses
2.1 Attitudes towards the political elite and fiscal preferences
The starting point for our hypothesis that people with strong populist ideas hold more expansionary fiscal preferences is given by the model of Cukierman and Tommasi (1998). It consists of three premises that we deem rather realistic. First, public policy is complex, as outcomes do not only depend on government policies, but also on external circumstances. Second, politicians have a comparative advantage in economic policymaking, as they have more information and access to expert judgement on the state of the world. Last, the model assumes that voters cannot observe the elite’s ideological preferences and whether these are aligned with theirs.Footnote 1
In such a setting, voters will condition their evaluation of economic policy proposals on their judgement of the political elite’s position vis-à-vis topics that matter to them. When people believe that the political elite acts according to the people’s will, they will be more likely to support their plans. Arguably, this mechanism greatly facilitated the job of political and economic elites in the days when there were strong ideological and religious ties between the electorate and the elite (i.e., when party systems in Europe were still ‘frozen’, see Lipset & Rokkan, 1967). Yet, when for whatever reason people believe that the political elite is not acting in the people’s interest as it should, they may be sceptical of the elite’s policy proposals, especially if these policy proposals align with their views of the elite. If voters fear that elites will use public funds for their own means, they may hence favour a lean government (Hayo & Neumeier, 2017; Otjes et al., 2018; Roth et al., 2022). However, the opposite may also be true. When voters are especially worried that the elite is pursuing an agenda that mainly caters to the interests of big business, as in the model of Acemoglu et al. (2013), the elite will have to propose expansionary fiscal policies in order to be credible. Additionally, in order to broaden their vote base, nowadays many populist parties propagate a generous welfare state to appeal to working-class voters that in economic terms typically have leftwing preferences (Afonso, 2015). This line of argument is also supported by a study on the Dutch PVV showing that voting has become slightly more left-wing on the period 2006–2017 (Otjes, 2019). On the basis of this discussion, our first hypothesis will beFootnote 2:
Hypothesis 1
People with strong populist ideas will be more favourable to expansionary fiscal policy.
Importantly, there is also evidence suggesting that the relationship between populist attitudes and fiscal preferences goes in the other direction. Analysing how fiscal consolidation undertaken in the aftermath of the Great Recession affected welfare spending at the regional level, Fetzer (2019) finds that the support for Brexit was especially large in districts where the cuts to welfare spending were largest. Similarly, Guiso et al. (2019) hypothesize that in euro area countries the crisis has spurred frustration over the loss of economic policy space, and find support that the resort to populist parties was stronger in countries in the eurozone than outside. Although these studies focus on aggregate-level processes, and do not examine individual-level populist attitudes, they do warrant for an empirical set-up to correct for endogeneity of populist attitudes.
2.2 Fiscal illusion and the role of literacy and information
There is a large literature in the public choice tradition on ‘fiscal illusion’, which holds that voters appreciate spending programs, but underestimate the (future) costs in terms of taxation or debt (Alesina & Perotti, 1995; Buchanan & Wagner, 1977). This can, for instance, arise when people observe the fruits of public spending, but do not observe the costs when spending is paid for with an increase of public debt. This would cause fiscal policies to be biased towards deficits. While the concept of fiscal illusion is intuitively very powerful and is often taken for granted, scholars have put forward that fiscal illusion is at odds with the fact that voters often actually support politicians with fiscal conservative agendas (Alesina & Perotti, 1995; Eslava, 2011).
The financial sophistication of the public and collective learning by the public and the media to judge fiscal policies have been put forward as mechanisms that can mitigate fiscal illusion. Indeed, empirical studies have highlighted that a large share of the public is ill-aware of economic facts and mechanisms (Blinder & Krueger, 2004; Boeri et al., 2001; Caplan, 2002; van der Cruijsen et al., 2015). When public policy is complex, and budgets are non-transparent, poorly literate voters are less suited to judge economic and fiscal policies. In turn, as voters are more literate and receive information on the public budget, they are less prone to deficit bias.
On the basis of a study on Germany, Hayo and Neumeier (2017) indeed find that the more knowledgeable respondents are, the more they favour prudent fiscal policies. Furthermore, a few studies have used randomized information experiments and have confirmed that exposure to factual information can spur public support for pension reforms (Boeri & Tabellini, 2012), dampen support for raising teachers’ salaries (Lergetporer et al., 2018) or alter overall fiscal preferences (Roth et al., 2022). In particular, this latter study finds that respondents become more supportive of debt reduction, although they do not update their preferences on taxation.
Hypothesis 2
People with poor literacy skills will be more favourable to expansionary fiscal policy.
Hypothesis 3
When people receive information about the intertemporal budget constraint of the government, they will become less favourable to expansionary fiscal policy.
2.3 The moderating effect of populism on the effects of literacy and information
To the best of our knowledge, there is no literature explicitly linking literacy and populist attitudes in relation to fiscal policy preferences. Yet, from the model of Cukierman and Tommasi (1998) presented above we can derive predictions on this link. The model first of all holds that it is more difficult for voters to evaluate economic policy than for politicians due to less information and access to expert judgement. Yet, of course, this does not apply to all voters to the same extent, as some voters are more literate and have gathered more information on economic policymaking than others. Likewise, support for economic policies depends on people’s judgement of the political elite’s position vis-à-vis topics that matter to them. Yet, also here, voters differ in their evaluation of the political elite’s intentions. These two effects can be expected to reinforce one another. When voters are particularly suspicious of the elite’s intentions, especially poorly literate individuals may demand expansionary policies to be convinced that the elite is acting in their interest. Yet, when they are reassured that the elite is acting in the people’s interest, poor literacy may not lead to more demand for expansionary fiscal policies.
Hypothesis 4
The positive effect of poor literacy on support for expansionary fiscal policy is larger, when people hold stronger populist views.
When it comes to information, a similar mechanism may be in place. In line with Hypothesis 3, factual information can be expected to dampen expansionary preferences. Yet, there is also a possibility that people who are sceptical of the motives of the political elite also tend to be sceptical of expert advice and third-party information in general. Indeed, recent research has found that populism is correlated with anti-intellectualism—a measure that includes attitudes towards economists (Merkley, 2020). Hence, when respondents are negative about the elite, they may be inelastic to information provision. Indeed, a study among US respondents finds that information only minimally alters fiscal preferences among respondents with low trust in government (Kuziemko et al., 2015).
Hypothesis 5
The negative effect of information on support for expansionary fiscal policy is more muted, when people hold stronger populist views.
2.4 Controls of fiscal preferences
A large theoretical and empirical literature has highlighted a host of determinants of fiscal preferences that we need to control for in our regression of fiscal preferences.
2.4.1 Gender, age and children
First of all, it is well established that since several decades women in industrialized countries have grown to be more supportive of redistribution and government spending (Alesina & Giuliano, 2011; Alesina & La Ferrara, 2005; Dassonneville, 2021). Furthermore, most empirical evidence suggests that support for redistribution increases with age (Alesina & Giuliano, 2011; Gärtner et al., 2017). People with children may be less supportive of high debt due to intergenerational concerns. Yet, the evidence is not conclusive (Hayo & Neumeier, 2017; Heinemann & Hennighausen, 2012).
2.4.2 Education
On the basis of the data from both the US General Social Survey and the World Values Survey, Alesina and Giuliano (2011) find that less-educated people demand more redistribution, also controlling for income, although less-educated people have also been found to be less favourable to debt reduction, again controlling for income (Stix, 2013).
2.4.3 Economic position
According to the Meltzer-Richard hypothesis, one’s position in the income distribution is the key driver of support for government spending (Meltzer & Richard, 1981), although empirical support for this is more nuanced. Likewise, in the model of Cukierman and Meltzer (1989) people who are more financially constrained are more supportive of borrowing from future generations. Empirical work has confirmed that financially constrained people are more opposed to debt consolidation (Stix, 2013). Furthermore, Alesina and Giuliano (2011) report that people who have experienced an unemployment spell are more supportive of redistribution.
2.4.4 Personality traits
The literature has also highlighted how several personality traits affect fiscal preferences. First of all, and most straightforwardly, more risk averse individuals demand more redistribution (Alesina & La Ferrara, 2005; Gärtner et al., 2017). Second, studies have found that myopic individuals are more tolerant of debt (Hayo & Neumeier, 2017; Stix, 2013). Third, studies have highlighted that people who believe in control over lifetime economic outcomes, rather than luck or fate, attach less value to redistribution (Alesina & Giuliano, 2011; Kouba & Pitlik, 2014). Last, Bakker (2017) inspects the role of the ‘Big Five’ personality traits and found that conscientious individuals are less supportive of redistribution, while agreeable and neurotic individuals are more supportive.
2.4.5 Attitudinal predispositions
Finally, fiscal preferences have been linked to a variety of predispositions held by individuals. Almost by definition, people who place themselves to the left of the political spectrum are keener on redistribution (Alesina & Giuliano, 2011). Furthermore, in the European context, fiscal policy is highly influenced by European rules. Attitudes towards the European Union may therefore drive fiscal preferences, as people who are supportive of the EU may be more supportive of fiscal consolidation efforts to comply with European rules. At the same time, there is also evidence that Euroscepticism has grown as a result of fiscal consolidations implemented during the crisis, which lends support to the view that one’s attitude towards the EU could be an endogenous regressor (Armingeon et al., 2016; Fetzer, 2019; Guiso et al., 2019).
3 Research design and data description
3.1 Context of our study
Our study is set in the Netherlands. With government debt at 43.0 percent of GDP in 2007, its public finance position was relatively strong (CPB, 2019). Yet, government debt rose to 67.8 percent of GDP in 2014, despite expenditure cuts and tax rises that were taken from 2011 onwards. Since then, growth turned positive again and government debt started to fall. In Summer 2017, the economic forecasts for the years to come were better than assumed in the initial budgetary projections. In the run-up of the budget for 2018—that was presented to the public in September 2017—there was discussion on how to allocate the tax windfalls that would result from this, for debt reduction (consistent with budgetary rules) or for tax relief or more spending. This is precisely the question that we asked respondents in our September 2017 survey.
It is furthermore worth mentioning that, compared to other European countries, Dutch households appear to be relatively financially literate (Fornero & Lo Prete, 2019) as well as fiscally conservative (Debrun & Kinda, 2017; van Geest & van Vuuren, 2018). Arguably, this tradition has made Dutch voters one of the most sophisticated audiences for judging fiscal policy. Compared to households in other Northern European countries, Dutch households are relatively debt averse when it comes to their fiscal preferences. In a 2010 Eurobarometer poll, 77 percent of Dutch households agreed that measures to reduce public deficit and debt could not be delayed (European Commission, 2010). This percentage was only 3 percentage points higher than the EU average; yet, the Netherlands had also a relatively low debt level (59 percent of GDP, against an EU average of 80 percent in 2010).
Furthermore, the Netherlands is also a suitable case for studying the influence of populism. First, next to right-wing populists (LPF, PVV, FVD), also left-wing populists (SP) have been successful in this country. Second, it has been demonstrated that populist attitudes can be measured validly in this country (Akkerman et al., 2014). Third, the populist message is relatively widespread in the Netherlands (Rooduijn, 2014). The main surveys used for this study, on populism and fiscal preferences, were held just several months after the Parliamentary elections in which populist parties fared relatively well (together, PVV, SP and FVD obtained 24 percent of the seats).
3.2 Surveys used
We employ various modules of the DNB Household Survey (DHS), conducted by CentERdata at Tilburg University. DHS is a panel dataset that includes approximately 2000 households from which one or more household members may take part. The panel is designed to be representative of the Dutch population and includes questions on demographics, occupational status, education, earnings, wealth, health and psychological concepts, in various modules spread out over the year. A majority of respondents stays in the panel year after year; in case of attrition, new participants with similar characteristics are recruited. Among other things, the panel has been used to study financial literacy and financial behaviour (see, e.g., Guiso et al., 2008; van Rooij et al., 2012; Kramer, 2016).
We use DHS data for the year 2017. In addition to recurrent questions, additional questions can be added to the questionnaire on an ad hoc basis. In a special module of the September 2017 DHS survey, we asked respondents several questions on fiscal policy. The survey was presented to 2773 members of the panel, and completed by 2299 of them (i.e., the response rate was 82.9%). Furthermore, in June 2017, we conducted a special module in the DHS on political attitudes of households (this survey was also used in Rooduijn et al., 2017). This survey was presented to 3035 members of the panel, and completed by 2358 of them (i.e., the response rate was 77.7%). Conduction of two separate surveys for fiscal preferences and populist attitudes avoids the risk of potential priming as well as common source bias (see also Sect. 5). In addition to the 2017 data, for robustness purposes, we also use some information from older surveys. In particular, to instrument populism with two indicators of trust (see more in Sect. 5), we draw from the trust survey that is administered by De Nederlandsche Bank and embedded in the DHS modules of 2006 and further (for more information, see van der Cruijsen et al., 2016). Tables A.1a and A.1b in the Online Appendix contain an overview of all variables used and the respective module.
3.3 Dependent variables
Our dependent variables capture the support of respondents for three margins of fiscal policy. To be precise, we asked respondents to what extent they would prefer to use the tax windfalls that were foreseen at the time for (i) debt reduction, (ii) tax relief, and (iii) increased spending, leaving the other fiscal margins the same (see Table A.1a in the Online Appendix for the exact wording of our survey questions). Respondents could answer all three questions on a 1 (do not agree at all) to 4 (strongly agree) scale.Footnote 3
Given that we are interested in relative preferences for the three margins of fiscal policy, we control for the fact that some respondents overall report more extreme answers than others. To make sure we capture preferences for fiscal policy components relative to one another we divide the scores per item by the the total scores awarded. This adjustment results in three continuous variables, which we again transform to the original scale from 1 to 4. We will use these three variables—‘support for debt reduction’, ‘support for tax relief’ and ‘support for more spending’—as our dependent variables throughout the paper. Correcting for the total scores awarded is not trivial, however, as the total scores awarded is significantly correlated with populism (r = 0.17). Yet, in the robustness section we show that our results are similar without this correction. This also applies to a probit estimation based on a dummy variable for those respondents who (strongly) agree using the tax windfalls for the three dimensions of fiscal policy.
Table 1 below presents summary statistics for all variables used, restricting the sample to respondents for whom we observe their fiscal preferences. From an inspection of the means of our three dependent variables, we can see that respondents were on average most likely to support more spending and, to a lesser extent, debt reduction, while support for tax relief was substantially lower. To test our hypotheses, we interpret support for debt reduction as less expansionary, and support for tax relief and more spending as more expansionary.
3.4 Regressors of main interest
The regressor of our main interest is ‘populism’. While political scientists initially studied the populist ideas of political parties, recently they have started to measure to what extent individuals are prone to populist ideas (Akkerman et al., 2014; Castanho Silva et al., 2020; Hawkins et al., 2012; van Hauwaert & van Kessel, 2018; Wüttke et al., 2020). To be precise, this set of ideas comprises (i) a distinction between the ordinary people and an evil elite, (ii) an antagonistic relationship, and (iii) the premise that politics should follow the general will and respect popular sovereignty (Mudde, 2004).Footnote 4 Following the work of Akkerman et al. (2014), in the June 2017 survey we asked respondents to rank their agreement with six statements on political elites. For instance, respondents were asked to what extent they agreed with the following statement “I would rather be represented by a citizen than by a specialized politician” (see Table A.1a of the Online Appendix for the exact wording of all statements). As in other studies, the scores on these statements are strongly related, allowing us to integrate the scores into a composite index which is internally consistent.Footnote 5 This composite indicator, ‘populism’, is a continuous variable ranging from 1 to 5.
Furthermore, to measure the numerical sophistication of respondents we use an index for probability literacy (‘literacy’).Footnote 6 This index traces to what extent respondents can account for uncertainty, such as an event like job loss (Hudomiet et al., 2018).The index is based on four numerical questions, where respondents have to select the right answer on an answering scale from 0 to 100. We reward each correct answer with 1 point, leaving us with an index running from 0 (all questions answered incorrectly) to 4 (all questions answered correctly).
Last, in order to assess the effect of information provision on fiscal preferences we conducted an information experiment, which was presented to DHS respondents just before the question on fiscal preferences in the September 2017 survey (note that the questions on populist attitudes were asked in a separate survey in June). In the experiment, half of the respondents was given some information about the intertemporal budget constraint of the government, whereas the other half received no message.Footnote 7 To be precise, our message is as follows: ‘When the government spends more money in a year than it receives through taxation, the government runs a budget deficit. To finance this, the government must borrow money. As a result, the total debt of the government (‘government debt’) will increase. The government cannot let government debt rise endlessly. If the government keeps on borrowing, eventually it will have to raise taxes and/or cut expenditures so as to stop government debt from increasing.’ The variable ‘debt experiment’ takes the value of 1 if respondents received the information text, and a value of 0 if they received no text. Table A.3 in the Online Appendix shows the descriptive statistics for the treatment and control group respectively, which are very similar.
3.5 Controls
Due to the embedding of our questions in the DHS survey, we can merge our survey questions with a rich set of controls that have been shown to be relevant in modelling fiscal preferences, as discussed in Sect. 2. Namely we control for age, income, educational level, personality traits and the financial situation of the household (e.g., whether they have a hard time making ends meet).Footnote 8Footnote 9 For a detailed description of all variables, see Table A.1a in the Online Appendix or the DHS codebook.Footnote 10 It furthermore should be noted that we have imputed some values (based on surveys filled in by the same respondent in other years) in order to minimize the loss of observations. For relatively stable variables such as personality traits (Cobb-Clark & Schurer, 2012, 2013; Salamanca, 2018) and income, we have imputed the average value of the available observations over the entire period (note that income enters our regression via categorical dummy variables). For the variable ‘hard to get by’, we used a stricter imputation method, imputing the value of the observation in the previous or next year only, or the average thereof, if both were available.
Moving further in Table 1, ‘right-wing’ is the self-placement of respondents on a left-to-right scale, ranging from 0 (extreme left) to 10 (extreme right). Furthermore, the table shows respondents’ agreement (on a 1 to 5 score) with the statement that the government should take measures to minimize income differentials (‘equality’). Both variables come from the June 2017 survey.
Turning to the end of the table, under the ‘auxiliary variables’, we report the summary statistics for variables that are used for robustness purposes. First of all, pre-crisis trust in national politics and the management of financial institutions refer to the trust that respondents reported in the years 2006, 2007 and 2008 (as respondents stay in the panel for a limited number of years, this leads to a strongly reduced sample size). We will use these as instruments for populism in Sect. 5. Furthermore, in our robustness section we test whether the results hold when we include a regressor that we suspect to be endogenous to fiscal preferences, i.e., respondents’ attitudes towards a strengthening of EU cooperation (‘EU cooperation’). Finally, Table 1 highlights that, due to non-response and the merging process, for some variables there are quite a few missing values. As we shall see, in our baseline regressions we end up with a sample of 1636 observations, which is 71 percent of the full sample. In the robustness section, we test whether the smaller sample selection that we end up with results in different estimates.
4 Main results
We now turn to our baseline results, which allow us to test Hypotheses 1, 2 and 3. Our dependent variables are the scores of respondents on the three options to use tax windfalls that were at the time discussed in the policy debate, namely for debt reduction, tax relief or more spending. As our dependent variables are continuous, we use OLS. In all regressions, we cluster standard errors at the household level to control for the possibility that errors correlate among members of the households, capturing e.g. exposure to the same media and acquaintances that shape populist attitudes and/or policy preferences. We standardize the personality traits and use categorical dummy variables for income and education.
Table 2 reports the results of our main regressions. The three columns refer to the specifications in which the dependent variable is the score based on debt reduction (column 1), tax relief (column 2), spending reduction (column 3), respectively. The first row shows our main result: populist attitudes yield a strongly significant coefficient on all three fiscal preferences. In line with Hypothesis 1,respondents with strong populist attitudes are less debt averse, more inclined to favor tax relief, and more supportive of spending. Importantly, this effect holds when controlling for literacy and information measures. In line with Hypothesis 2, we find that our literacy measure exerts a highly significant effect on debt reduction and tax preferences. Yet, there is no effect on support for more spending. Last, treatment with our educational message causes respondents to report more support for debt reduction, and less support for more spending. In line with the findings of Roth et al. (2022), the effect on tax relief is not statistically significant.
We will not discuss all other results in detail, but mention some that stand out. First of all, support for more spending does not differ among educational groups, yet more education goes hand in hand with more support for debt reduction, and less support for offering tax relief. Second, in contrast to the prediction of the Meltzer-Richard hypothesis, one’s income position does not matter for fiscal preferences. Yet, in line with the model of Cukierman and Meltzer (1989), we do find that respondents who have a hard time getting by are less supportive of reducing public debt (although this result is significant at 10 percent only once we add our literacy measure).Footnote 11 Third, several of the personality traits appear highly significant regressors of fiscal preferences, although their impact differs across our three fiscal policy margins. The results for risk aversion, patience and locus of control are broadly in line with previous research. As to the Big Five personality traits, broadly in line with Bakker (2017), we find that conscientious respondents are significantly more supportive of tax relief, while individuals who score high on agreeableness are less supportive of tax relief, and more supportive of increasing spending. Fourth, there appear to be some interesting nuances when it comes to the differences between right-wing self-placement and support for income levelling. As regard to debt reduction, it is right-wing respondents that stand out with a significantly higher support for debt reduction. When it comes to support for tax relief, it is supporters of income levelling that stand out, reporting significantly lower levels of support for tax relief. Regarding support for more spending, both dimensions are found to be statistically significant regressors.
Table 3 reports the results of a decomposition of the explained variance. First and foremost, it shows that populist attitudes are not only a significant regressor for fiscal preferences, but they are also very important in material terms. When it comes to support for debt reduction and tax relief, a quarter of the explained variation can be attributed to populist attitudes. Its contribution is bigger than many variables that play a central role in the literature on fiscal preferences, such as income and right-wing ideology. The relevance of populist attitudes for support for more spending is more limited, whereas the role of ideology is much bigger here. Second, a similar pattern can be observed with regards to our literacy measure. Whereas literacy accounts for about a fifth of the fit of our models of debt reduction and tax relief, this is only 3 percent for attitudes towards spending. Third, while our information treatment has a significant effect on two out of three fiscal preferences, it can only account for a very small portion of the model fit. Table 3 highlights the significant impact of personality traits on attitudes towards fiscal policy, underscoring the importance of these controls in our analysis. Taken together, they are even the most important predictors of support for more spending.
5 Robustness analysis
5.1 Endogeneity
As mentioned earlier, we have one primary methodological concern regarding the regression presented in Table 2. Our main regressor of interest, populist attitudes, could be endogenous for three main reasons. First, there could be an omitted variable bias, i.e. fiscal preferences and populism are both influenced by another factor, such as feelings of vulnerability. A second concern is that the relation is simultaneous, i.e. there is reverse causation from fiscal preferences to populism. In our case, we cannot exclude the possibility that respondents with more expansionary fiscal preferences have grown populist sentiments, when the government consolidated in the midst of a large recession. Finally, it would be problematic if errors in the measurement of our variables—which by themselves are inevitable in observational research—would not be random. For instance, a respondent’s mood can influence the answers given in the same survey in similar ways. This is particularly a concern when variables are taken from a common source.
In our baseline regression, several elements of our estimation strategy already mitigated some of these sources of endogeneity. For instance, we employ a very rich set of regressors, diminishing the bias resulting from omitted variables. Likewise, our questions come from surveys held at various moments in time, which alleviates concern over common source bias. Most importantly, this holds for our main regressor of interest, populist attitudes, which were measured in June 2017, and our dependent variables (fiscal preferences), which were recorded three months later in September 2017. Nonetheless, there remains a risk of endogeneity due to other omitted variables, the possibility of reverse causation and measurement error that affect the answering in all surveys (e.g. a tendency to social desirability).
A more general solution to control for endogeneity is to employ an instrumental variable approach. To address the possibility of reverse causation, we opt for instrumental variables that were collected with a considerable lag from the time we inquired about respondents' fiscal preferences, ideally before the economic crisis. As the measurement of populist attitudes was first done in June 2017, a lagged variable of populism is not an option. Yet, as our survey was embedded in the rich DHS survey environment there are other lagged variables that we can consider. In particular, since 2006, the DHS survey has included questions about respondents' trust in various entities, including national politics. Additionally, respondents were asked about their faith in the expertise and integrity of financial firms' management (see Table A.1a in the Online Appendix for the exact wording).Footnote 12
Trust in politics is typically understood by political scientists as an evaluation of how well politicians fulfill people’s expectations (van der Meer, 2018) and is a different concept than populism (Geurkink et al., 2020). For instance, according to populist discourse, politics should follow the ‘will of the people’, which is not a necessary condition for people to trust their government. Yet, low trust and populism have in common that they are, at least to a large extent, an expression of discontent with the performance of the political elite. Likewise, respondents’ attitudes to the management of financial institutions are likely to pick up adverse sentiment towards elites. To rule out the possibility that our instruments capture feelings of anger over fiscal consolidation measures taken in the aftermath of the crisis, we take the average of the available observations that were available in 2006 to 2008, before the crisis hit. This limits our sample to 610 observations. Yet, even with this time lag, our instruments could still be endogenous to our dependent variables (fiscal preferences) as there could be unobserved fixed individual characteristics that influence both attitudes towards fiscal policy and political views. While there is a literature linking fiscal preferences to trust in politics (Hayo & Neumeier, 2017; Stix, 2013), we are not aware of any theoretical or empirical linkages between pre-crisis attitudes towards the financial sector and fiscal preferences. Table A.5 in the Online Appendix shows the correlation between fiscal preferences of respondents in 2017 and their trust in national politics and the management of financial firms in the years before the crisis. The analysis reveals a significant correlation between pre-crisis trust in national politics and fiscal preferences in 2017, even after accounting for other relevant variables. However, no significant correlation is found between fiscal preferences and pre-crisis trust in financial sector management. After introducing controls, the partial correlation between the two becomes very low. This finding provides reassurance that at least one of our two instrumental variables is exogenous to the dependent variables, which is a necessary condition for the Sargan over-identification test we will perform later.
Table 4 presents the results of our two-stage-least square (2SLS) regression. While the table focuses on the regression of fiscal preferences (our second stage), it also includes the highlights of the first-stage regression where we instrument populist attitudes with the two instruments, pre-crisis trust in politics and the financial sector management, while including all controls from explanatory regression. The results indicate that our two instrumental variables turn out to be highly significant regressors. The partial R2 of the first stage is 0.159 and the F-statistic 55.1, which largely exceeds the threshold of 10 which is widely used to test the relevance of an instrument. In the regression of fiscal preference, our second stage, we include the regressors of Table 2. We first run the same regressions, but then with our limited sample size (N = 610) so that we can attribute any differences in the estimated coefficient to the estimator and not to differences in the sample. While the point estimates of the OLS and 2SLS differ, the 2SLS results confirm our main finding from our baseline regression, i.e., a significantly positive coefficient for populist attitudes.
Lastly, given that we have multiple instruments, we perform a Sargan test of over-identifying restrictions to assess the exogeneity condition of our instrument set. At the 5 percent significance level, we find no evidence to reject the null hypothesis of validity for all three fiscal preferences. However, we also evaluate the necessity of using an IV estimation as OLS is known to be a more efficient estimator.Footnote 13 Comparing OLS and 2SLS results, the Wu-Hausman test scores formally rejects endogeneity of populist attitudes. Hence, while our IV estimation is valid and the results support that populism is a significant predictor of fiscal preferences, we may just as well rely on our OLS regression.
5.2 Inclusion of attitudes towards the European Union
As highlighted in the literature review, attitudes towards the European Union (EU) may also play a significant role in shaping fiscal preferences. However, it is essential to consider that attitudes towards the EU could be influenced by fiscal preferences, possibly due to frustration over consolidations during the crisis. This makes it an endogenous regressor, highly correlated with populist attitudes (0.42, see Table A.4 in the Online Appendix). Incorporating multiple endogenous covariates could significantly confound our regressions, leading us to exclude attitudes towards the EU from our baseline regression. Instead, we address endogeneity through an IV regression, using pre-crisis trust levels as instruments for populist attitudes. However, it's important to acknowledge that pre-crisis populist attitudes may also capture some effect of (pre-crisis) attitudes towards the EU. In Table A.6 in the Online Appendix we run the same regressions as in Table 2, but now also including attitudes towards EU cooperation. While the coefficient of populist attitudes is somewhat smaller, it remains highly significant and all other results also hold.
5.3 Loss of observations
As noted in Sect. 3, using variables from different DHS modules comes at the cost of a loss of observations. In our baseline regressions we end up with a sample of 1636 observations, which is 71 percent of the full sample of fiscal preferences. To test whether restricting the sample influences our results, we repeat the regressions with a smaller set of controls and hence with larger samples. In Table A.7 in the Online Appendix we repeat the regression of Table 2 for debt reduction (the first of the three fiscal preferences) with a smaller set of regressors for which we have substantially more observations. Comfortably, we find that all coefficients are remarkably similar. The only difference is that in one case (i.e., a dummy for respondents aged 65 and over) the coefficient is only statistically significant in the larger sample, and not in our baseline sample.
5.4 Correction of scaling
As discussed in Sect. 3, we have adjusted the scales of our dependent variables to account for the total number of scores awarded, aiming to filter out the influence of more extreme answering. However, this correction is not straightforward, as the answering behaviour is correlated with populist attitudes. To ensure the robustness of our main results, we conduct tests using the uncorrected scale as well. Table A.8 in the Online Appendix compares the results of Table 2 (full model) with fiscal preferences, when no correction is made (to compare the coefficients, the scale of both the unadjusted scale as the adjusted scale runs from 1 to 4). When we do not correct the scaling, the results for populist attitudes are even stronger. Furthermore, also based on the uncorrected scaling, instead of an OLS regression we also run a probit regression based on a dummy variable taking the value of 1 in case respondents agree or strongly agree with incurring this fiscal policy margin. Results (Table A.9 in the Online Appendix) are largely the same, although the coefficients of populism and the information experiment on support for more spending are no longer significant.
6 Interaction effects
6.1 Moderating effect of populism on effect of literacy
To test Hypotheses 4 and 5, this section extends the regression of Table 2 with interaction effects of populism with literacy and information provision, respectively. Figure 1 graphically reports the results of the interaction analysis of populist attitudes and literacy. The y-axis depicts the marginal effect of literacy on fiscal preferences at different levels of populism (x-axis). As such, the chart reports the combined effect of the coefficients of literacy, and the interaction term of literacy and populist attitudes.Footnote 14 The brown bars display the distribution of populist attitudes.
Moderating effect of populism on the effect of literacy on fiscal preferences. Notes: Panels show at the y-axis the marginal effect of literacy on fiscal preferences (i.e., debt reduction, tax relief and more spending) at various levels of populism (x-axis). Shaded area covers the 90% confidence interval. Brown bars display the distribution of populist attitudes. (Color figure online)
The chart in panel (a) of Fig. 1 shows that the effect of literacy on debt reduction—which on average yielded a highly significant coefficient of 0.20 (see Table 2)—does not vary much with populism, although at very low levels of populism the effect of literacy on support for debt reduction is no longer significant. The effect is stronger when it comes to tax relief (see panel (b) of Fig. 1). On average, we found a highly significant coefficient of − 0.17 of literacy on support for tax relief. Yet, panel (b) shows that the effect of literacy on support for tax relief is not significant at low levels of populism, but is significant at high levels of populist attitudes, lending support to our Hypothesis 4. The effect of literacy on support for spending does not vary significantly with the level of populism (see panel (c) of Fig. 1), which is not surprising as there was also no significant overall effect of literacy on support for more spending in the first place. These results are robust to using the uncorrected scale (results available upon request).
6.2 Moderating effect of populist attitudes on effect of information
Figure 2 reports the results of the models in which we include an interaction term of the information experiment with populist attitudes. The figure shows the predicted level of fiscal preferences (i.e., support for the fiscal policy margin on a scale of 1 to 10), both for people who have been treated with the information experiment (red line) as well as for those in our control group (blue line). In Table 2 we estimated the average effect of the information experiment over the entire sample, yielding a significant coefficient for debt reduction (0.28, i.e., more support debt reduction) and spending (− 0.17, i.e., less support for more spending). Panels (a) and (b) of Fig. 2 show that the effect of the information experiment on support for debt reduction does not vary along respondents’ level of populist attitudes. Panel (c), however, shows that the information experiment reduces support for more spending especially when populist attitudes of respondents are stronger.Footnote 15 Hence, information provision can alleviate fiscal illusion also especially with voters who with strong populist attitudes.Footnote 16 This is in contrast to our Hypothesis 5, and may be explained by the fact that respondents with low populist sentiment report much lower support for increasing spending in the first place, leaving less scope for adjusting their fiscal preferences in response to information provision.
Moderating effect of populism on effect information on fiscal preferences. Notes: At the y-axis is the level of support for our three fiscal policy margins (i.e., debt reduction, tax relief and more spending) at various levels of populism (x-axis) for respondents in the treatment group of our information experiment (red line) and those in our control group (blue line). Shaded area covers the 90% confidence interval. Brown bars display the distribution of populist attitudes. (Color figure online)
All in all, we find some supportive evidence on our hypothesis that providing information about the intertemporal budget constraint of the government leads respondents to have less expansionary fiscal preferences. Moreover, we find that respondents with stronger populist attitudes, who report significantly higher support for more government spending, also reduce their support for more spending more strongly after being exposed to our information treatment. This suggests that information provision can mitigate fiscal illusion among voters with strong populist attitudes.
7 Conclusions
This paper assesses whether populist attitudes lead to more expansionary fiscal preferences, and whether populist attitudes reinforce the risk of fiscal illusion. Our findings show that populist attitudes are indeed a highly significant and materially important predictor of fiscal preferences (a finding that is robust to the use of an instrumental variable (IV) estimation). Individuals with strong populist sentiments are less supportive of debt reduction, and more supportive of tax relief and more spending. Notably, the explanatory power of populist attitudes in predicting debt reduction and tax relief preferences far surpasses that of widely studied covariates in the political economy literature, such as income and left-right ideology. This is a very important finding because it shows that key socioeconomic preferences that have traditionally been associated with classical left-right positions currently appear more closely linked to a new dividing line like populism or anti-establishment politics (see Uscinski et al., 2021). This suggests that the way in which political and economic attitudes are rooted in ideologies could have changed over time. And this, in turn, means that economists and political scientists should carefully rethink and restudy the ideological foundations of attitudes about welfare.
To assess whether populist sentiment reinforces the risk of fiscal illusion, we have also inspected the role of literacy and information, and their interaction with populist attitudes. We find that literacy is a statistically significant and materially relevant predictor of support for debt reduction and tax relief. We take this as support of the occurrence of fiscal illusion. In addition, we find that populist attitudes moderate the effect of literacy on support for tax relief (but not attitudes towards debt and spending). At high levels of populist attitudes, literacy is a significant predictor of support for tax relief, but not at low levels of populist attitudes. Our results hence suggest that populist sentiment reinforces the risk of fiscal illusion that comes with poor literacy. Turning to information, our information experiment confirms that providing information about the intertemporal budget constraint of the government causes respondents to have less expansionary fiscal preferences. We find that respondents with strong populist attitudes, who report significantly higher support for more government spending, also reduce their support for more spending more strongly after being exposed to our information treatment.
Our findings offer various lessons to economic policymakers who may view populism as a threat to sound economic policymaking. First of all, our results corroborate the finding that poor literacy spurs the risk of fiscal illusion which calls for investing in the numerical and fiscal policy sophistication of voters (Fornero, 2015). Furthermore, this risk is even larger when people, for whatever reason, have come to believe that the political elite is not acting in their interest as it should. This means that in the current era, in which the religious and/or ideological ties between voters and the elite have become increasingly loose, investing in knowledge and skills is even more important. Second, our results suggest that information provision can alleviate fiscal illusion especially with people with strong populist sentiment, as they are most prone to a deficit bias in the first place. Of course, a precondition is that information can reach such voters, which may be complicated because of the lower tendency of voters with strong populist attitudes to make use of established news sources (Schulz, 2019).
Yet, the results of our analysis also imply that fiscal illusion is not the full story behind the expansionist ‘populist economic agenda’, as dubbed by economists. For one thing, after controlling for literacy and information, there remains a very large independent effect of populism on fiscal preferences. Our literature review has highlighted some mechanisms that may be at play here. Most prominently, according to the model of Acemoglu et al. (2013), individuals who think the elite is not acting in the people’s interest as it should, may find that the political elite’s agenda caters more to the needs of big business than to the ordinary people. It is not in the scope of this paper to judge whether such a view is correct or not. In either case, as put forward by Piketty (2020), it is dangerous to equate the economic agenda of populist parties as merely short-termist and unsustainable, as it can reinforce the idea that the elite is not responsive to the needs of ordinary voters and can also inhibit debates about fundamental economic policy questions, complex as they may be. And in either case, the elite may need to signal better that it is really acting in the interest of ordinary people, e.g., by investing in universal welfare schemes and shifting taxation more to big firms and wealthy households.
We conclude this paper with some suggestions for further research. First of all, with our dataset we have only been able to analyse fiscal preferences and populist attitudes in a cross-section setting. By instrumenting populist attitudes with pre-crisis trust levels, we ruled out the possibility of reverse causation. Yet, while we showed that one of our instrumental variables was exogenous to fiscal preferences from a statistical point of view, it could be argued that unobserved individual traits may link it endogenously to the dependent variable. This can only be addressed by using longitudinal data, once these become available. Second, the strong empirical relationship between populist attitudes and fiscal preferences warrants more theoretical literature on the mechanisms underlying this relationship.
Notes
The model of Cukierman & Tommasi (1998) continues with the prediction that under certain circumstances, ‘unlikely’ parties will be more successful in implementing substantial policy reform, as their proposal of such policies can be more credible.
Besides preferences for the aggregate stance on debt, taxation and spending, populist attitudes may also come with preferences on the exact composition of taxation and spending. We leave this issue for further research.
Clearly, a score of 1 or 4 on all three dimensions would be inconsistent. This is the reason why we delete those cases. Together, this concerns 31 cases.
It has been demonstrated that the index of populism, also used in this research, measures different underlying attitudes than older concepts such as political trust and external political efficacy, i.e. the extent to which respondents feel they can influence the political process (Geurkink et al., 2020).
Cronbach’s alpha = 0.85. To construct the index, following Akkerman et al. (2014), we take a simple average.
We also considered using two measures for financial literacy available in the DHS, a parsimonious one based on three questions as designed by Lusardi and Mitchell (2014) and an extensive one based on sixteen questions as designed by van Rooij et al. (2012). However, the questions to construct these indices were not asked in 2017. Using the index of another year would result in a loss in the number of observations (by 75% in the case of the extensive measure). The correlation of our probability literacy index with the parsimonious and the extensive financial literacy indicator is 0.33 and 0.43 respectively. When running our baseline regressions with the literacy measures in a smaller sample, our conclusions are the same.
With information experiments there is a risk that respondents try to conform with the researcher’s hypothesis, although such ‘experimenter demand effects’ have been shown to be of only limited quantitative importance in online surveys (de Quidt et al., 2018). We have tried to formulate our experiment as neutral as possible. Yet, it is possible that respondents are somewhat more inclined to support conservative fiscal policies. This could influence our results only to the extent that respondents differ in their tendency to conform to the survey.
The DHS survey presents respondents with a series of statements to measure risk aversion, patience, locus of control and all so-called ‘Big Five’ personality traits (i.e., openness, conscientiousness, extraversion, agreeableness and neuroticism, using Goldberg’s 50-item personality scale). See Table A.2 in the Online Appendix for the exact statements used.
Table A.4 in the Online Appendix displays pairwise correlations of all our regressors. It can be derived that the highest correlation between covariates in our baseline regression is − 0.42, i.e. the correlation between left-right self-placement (rightwing) and support for income levelling (equality). Furthermore, there is a similarly high correlation between populist attitudes and attitudes towards EU cooperation (0.42), which is not included in our baseline regression but in one of our robustness checks.
For the DHS codebook see https://www.dhsdata.nl/site/users/login.
We also included an actual unemployment spell as a regressor; yet the results were not statistically significant. In general, in Table 2 we find that a number of covariates turn out to be statistically insignificant. This may be due to the rather rich specifications in which some factors absorb the effects of others.
This variable has also been used in a recent paper on financial literacy and trust in financial institutions (van der Cruijsen et al., 2021).
Another drawback is that our IV estimate is strictly speaking only applicable to those respondents for which our instruments properly predict the level of populism. In more technical terms, the IV models estimate the ‘local average treatment effect’ (LATE) for those people that positively respond to the IV (‘the compliers’), while the models without IV estimate the ‘average treatment effect’ (Cameron & Trivedi, 2005: 884–885).
Regression results can be found in Table A.10 of the Online Appendix.
Regression results can be found in the Table A.11 of the Online Appendix.
Again, results are robust to using the uncorrected scale. Results available upon request.
References
Acemoglu, D., Egorov, G., & Sonin, K. (2013). A political theory of populism. Quarterly Journal of Economics, 128, 771–805.
Afonso, A. (2015). Choosing whom to betray: Populist right-wing parties, welfare state reforms, and the trade-off between office and votes. European Political Science Review, 7, 271–292.
Akkerman, A., Mudde, C., & Zaslove, A. (2014). How populist are the people? Measuring populist attitudes in voters. Comparative Political Studies, 47, 1324–1353.
Alesina, A., & Giuliano, P. (2011). Preferences for redistribution. In J. Benhabib, A. Bisin, & M. O. Jackson (Eds.), Handbook of social economics (pp. 93–131). Elsevier.
Alesina, A., & La Ferrara, E. (2005). Preferences for redistribution in the land of opportunities. Journal of Public Economics, 89, 897–931.
Alesina, A., & Passalacqua, A. (2016). The political economy of government debt. Handbook of Macroeconomics (Vol. 2, pp. 2599–2651). Elsevier.
Alesina, A., & Perotti, R. (1995). The political economy of budget deficits. IMF Staff Papers, 42, 1–31.
Andersen, T., Bertola, G., Driffill, J., James, H., Sturm, J., Uroševic, B., & Fuest, C. (2017). Economic policy and the rise of populism—It's not so simple. In EEAG Report on the European Economy 2017 (pp. 50–66). CESifo Group Munich.
Armingeon, K., Guthmann, K., & Weisstanner, D. (2016). How the Euro divides the union: The effect of economic adjustment on support for democracy in Europe. Socio-Economic Review, 14, 1–26.
Autor, D., Dorn, D., Hanson, G., & Majlesi, K. (2020). Importing political polarization? The electoral consequences of rising trade exposure. American Economic Review, 110(10), 3139–3183.
Bakker, B. N. (2017). Personality traits, income, and economic ideology. Political Psychology, 38(6), 1025–1041.
Blinder, A. S., & Krueger, A. B. (2004). What does the public know about economic policy, and how does it know it? Brooking Papers on Economic Activity, 35(1), 327–397.
Boeri, T., Börsch-Supan, A., & Tabellini, G. (2001). Would you like to shrink the welfare state? A survey of European citizens. Economic Policy, 16, 8–50.
Boeri, T., & Tabellini, G. (2012). Does information increase political support for pension reform? Public Choice, 150, 327–362.
Buchanan, J. M., & Wagner, R. E. (1977). Democracy in deficit: The political legacy of Lord Keynes. Academic Press.
Cameron, A., & Trivedi, P. (2005). Microeconometrics: Methods and applications. Cambridge University Press.
Caplan, B. (2002). Systematically biased beliefs about economics: Robust evidence of judgmental anomalies from the survey of Americans and economists on the economy. The Economic Journal, 112, 433–458.
Castanho Silva, B., Jungkunz, S., Helbling, M., & Littvay, L. (2020). An empirical comparison of seven populist attitudes scales. Political Research Quarterly, 73(2), 409–424.
Chávez Juárez, F. (2012). SHAPLEY2: Stata module to compute additive decomposition of estimation statistics by regressors or groups of regressors. Statistical Software Components. Boston College Department of Economics.
Cobb-Clark, D. A., & Schurer, S. (2012). The stability of big-five personality traits. Economics Letters, 115, 11–15.
Cobb-Clark, D. A., & Schurer, S. (2013). Two economists’ musings on the stability of locus of control. The Economic Journal, 123, F358-400.
Colantone, I., & Stanig, P. (2018). The trade origins of economic nationalism: Import competition and voting behavior in Western Europe. American Journal of Political Science, 62, 936–953.
Colantone, I., & Stanig, P. (2019). The surge of economic nationalism in Western Europe. Journal of Economic Perspectives, 33, 128–151.
CPB. (2019). Macro Economische Verkenning 2020, Verzamelde bijlagen MEV 2020 met lange tijdreeksen.
Cukierman, A., & Meltzer, A. H. (1989). A political theory of government debt and deficits in a neo-Ricardian framework. American Economic Review, 79, 713–732.
Cukierman, A., & Tommasi, M. (1998). When does it take a Nixon to go to China? American Economic Review, 88, 180–197.
Dassonneville, R. (2021). Change and continuity in the ideological gender gap: A longitudinal analysis of left-right self-placement in OECD countries. European Journal of Political Research, 60, 225–238.
Davidson, S. (2018). Populism and fiscal illusion. Australian Economic Review, 51, 418–425.
Debrun, X., & Kinda, T. (2017). Strengthening post-crisis fiscal credibility: Fiscal councils on the rise—a new dataset. Fiscal Studies, 38, 667–700.
de Quidt, J., Haushofer, J., & Roth, C. (2018). Measuring and bounding experimenter demand. American Economic Review, 108(11), 3266–3302.
Dornbusch, R., & Edwards, S. (1991). The macroeconomics of populism in Latin America. University of Chicago Press.
Eslava, M. (2011). The political economy of fiscal deficits: A survey. Journal of Economic Surveys, 25, 645–673.
European Commission. (2010). Standard Eurobarometer 73, European Commission.
Fetzer, T. (2019). Did austerity cause Brexit? American Economic Review, 109, 3849–3886.
Fornero, E. (2015). Economic-financial literacy and (sustainable) pension reforms: Why the former is a key ingredient for the latter. Bankers, Markets & Investors, 134, 6–16.
Fornero, E., & Lo Prete, A. (2019). Voting in the aftermath of a pension reform: The role of financial literacy. Journal of Pension Economics and Finance, 18(1), 1–30.
Gärtner, M., Mollerstrom, J., & Seim, D. (2017). Individual risk preferences and the demand for redistribution. Journal of Public Economics, 153, 49–55.
Geurkink, B., Zaslove, A., Sluiter, R., & Jacobs, K. (2020). Populist attitudes, political trust, and external political efficacy: Old wine in new bottles? Political Studies, 68, 247–267.
Guiso, L., Sapienza, P., & Zingales, L. (2008). Trusting the stock market. Journal of Finance, 63, 2557–2600.
Guiso, L., Herrera, H., Morelli, M., & Sonno, T. (2017). Demand and supply of populism. Centre for Economic Policy Research London, UK.
Guiso, L., Herrera, H., Morelli, M., & Sonno, T. (2019). Global crises and populism: The role of Eurozone institutions. Economic Policy, 34, 95–139.
Hawkins, K. A., Carlin, R. E., Littvay, L., & Rovira Kaltwasser, C. (Eds.). (2018). The ideational approach to populism: Concept, theory, and analysis. Routledge.
Hawkins, K. A., Riding, S., & Mudde, C. (2012). Measuring populist attitudes. Political Concepts Committee on Concepts and Methods. Working Paper Series, 55, 1–35.
Hayo, B., & Neumeier, F. (2017). Public attitudes toward fiscal consolidation: Evidence from a representative German population survey. Kyklos, 70, 42–69.
Heinemann, F., & Hennighausen, T. (2012). Understanding public debt preferences. FinanzArchiv/Public Finance Analysis, 68, 406–430.
Hudomiet, P., Hurd, M. D., & Rohwedder, S. (2018). Measuring probability numeracy. RAND Working Paper, R-1270.
Kouba, L., & Pitlik, H. (2014). I wanna live my life: Locus of control and support for the welfare state. MENDELU Working Papers in Business and Economics, 46/2014.
Kramer, M. (2016). Financial literacy, confidence and financial advice seeking. Journal of Economic Behavior & Organization, 131, 198–217.
Kuziemko, I., Norton, M., Saez, E., & Stantcheva, S. (2015). How elastic are preferences for redistribution? Evidence from randomized survey experiments. American Economic Review, 105(4), 1478–1508.
Lergetporer, P., Schwerdt, G., Werner, K., West, M., & Woessmann, L. (2018). How information affects support for education spending: Evidence from survey experiments in Germany and the United States. Journal of Public Economics, 167, 138–157.
Lipset, S. M., & Rokkan, S. (1967). Party systems and voter alignments: Cross-national perspectives. The Free Press.
Lusardi, A., & Mitchell, O. S. (2014). The economic importance of financial literacy: Theory and evidence. Journal of Economic Literature, 52, 5–44.
Meltzer, A. H., & Richard, S. F. (1981). A rational theory of the size of government. Journal of Political Economy, 89, 914–927.
Merkley, E. (2020). Anti-intellectualism, populism, and motivated resistance to expert consensus. Public Opinion Quarterly, 84(1), 24–48.
Mudde, C. (2004). The populist zeitgeist. Government and Opposition, 39, 541–563.
Otjes, S., & Louwerse, T. (2015). Populists in parliament: Comparing left-wing and right-wing populism in the Netherlands. Political Studies, 63(1), 60–79.
Otjes, S., Ivaldi, G., Jupskås, A. R., & Mazzoleni, O. (2018). It’s not economic interventionism, stupid! Reassessing the political economy of radical right-wing populist parties. Swiss Political Science Review, 24, 270–290.
Otjes, S. (2019). What is left of the radical right? The economic agenda of the Dutch Freedom Party 2006–2017. Politics of the Low Countries, 1(2), 81–102.
Piketty, T. (2020). Capital and ideology. Harvard University Press.
Rooduijn, M. (2014). The mesmerising message: The diffusion of populism in public debates in Western European media. Political Studies, 62, 726–744.
Rooduijn, M., van der Brug, W., De Lange, S. L., & Parlevliet, J. (2017). Persuasive populism? Estimating the effect of populist messages on political cynicism. Politics and Governance, 5, 136–145.
Rooduijn, M., van Kessel, S., Froio, C., Pirro, A., De Lange, S., Halikiopoulou, D., Lewis, P., Mudde, C., &Taggart, P. (2019). The PopuList: An overview of populist, far right, far left and Eurosceptic Parties in Europe. www.popu-list.org.
Roth, C., Settele, S., & Wohlfart, J. (2022). Beliefs about public debt and the demand for government spending. Journal of Econometrics., 231(1), 165–187.
Salamanca, N. (2018). The dynamic properties of economic preferences. Melbourne Institute Working Paper, No. 4/18.
Schulz, A. (2019). Where populist citizens get the news: An investigation of news audience polarization along populist attitudes in 11 countries. Communication Monographs, 86, 88–111.
Stix, H. (2013). Does the broad public want to consolidate public debt?—The role of fairness and of policy credibility. Kyklos, 66, 102–129.
Stöckl, S., & Rode, M. (2021). The price of populism: Financial market outcomes of populist electoral success. Journal of Economic Behavior & Organization, 189, 51–83.
Uscinski, J. E., Enders, A. M., Seelig, M. I., Klofstad, C. A., Funchion, J. R., Everett, C., Wuchty, S., Premaratne, K., & Murthi, M. N. (2021). American politics in two dimensions: Partisan and ideological identities versus anti-establishment orientations. American Journal of Political Science, 65(4), 877–895.
van der Cruijsen, C., de Haan, J., & Jansen, D. (2016). Trust and financial crisis experiences. Social Indicators Research, 127, 577–600.
van der Cruijsen, C., de Haan, J., & Roerink, R. (2021). Financial knowledge and trust in financial institutions. Journal of Consumer Affairs, 55(2), 680–714.
van der Cruijsen, C., Jansen, D., & de Haan, J. (2015). How much does the public know about the ECB’s monetary policy? Evidence from a survey of Dutch households. International Journal of Central Banking, 11, 169–218.
van der Meer, T. (2018). Economic performance and political trust. In E. M. Uslaner (Ed.), The Oxford handbook of social and political trust (pp. 599–616). Oxford University Press.
van Geest, A. S., & van Vuuren, D. (2018). CPB and the assessment of structural reforms. In K. de Haan & J. Parlevliet (Eds.), Structural reforms. Springer.
van Hauwaert, S. M. V., & van Kessel, S. (2018). Beyond protest and discontent: A cross-national analysis of the effect of populist attitudes and issue positions on populist party support. European Journal of Political Research, 57(1), 68–92.
van Rooij, M., Lusardi, A., & Alessie, R. J. (2012). Financial literacy, retirement planning and household wealth. The Economic Journal, 122, 449–478.
Wüttke, A., Schimpf, C., & Schoen, H. (2020). When the whole is greater than the sum of its parts: On the conceptualization and measurement of populist attitudes and other multidimensional constructs. American Political Science Review, 114(2), 356–374.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The views expressed in this article are the ones of the authors and do not necessarily reflect the position of the Central Planning Bureau (CPB). We are grateful to the editor, the associate editor, an anonymous referee, Roel Beetsma, Brian Burgoon, Thomas Buser, Jakob de Haan, Joop Hartog, Elsa Fornero, Niels Gilbert, Kees Goudswaard, Mauro Mastrogiacomo, Maarten van Rooij and Johannes Wohlfahrt for valuable comments and suggestions.
Supplementary Information
Below is the link to the electronic supplementary material.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Parlevliet, J., Giuliodori, M. & Rooduijn, M. Populist attitudes, fiscal illusion and fiscal preferences: evidence from Dutch households. Public Choice 197, 201–225 (2023). https://doi.org/10.1007/s11127-023-01101-9
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11127-023-01101-9