Crisis,Ideology,and Interventionist Policy Ratchets

Abstract

Proper government reaction to economic crisis has long been a central element of public policy debate and is experiencing a revival after the Great Recession of 2008. Previous studies argue on theoretical and empirical grounds that crises may lead to more interventionist policies, but also cause deregulation and liberalization. This article claims that policy responses will partly depend on the core economic ideology of government, causing ideologically heterogeneous post-crisis strategies. Employing a panel of 69 countries for which salient ideology measures can be constructed, we find that growth crises between 1975 and 2015 caused larger increases in government size and regulatory policy when countries have left-wing governments. We also find some evidence of policy ratchets, meaning that certain crisis policies present a tendency to become permanent, regardless of the ideology of successive governments in power. Rolling back the public sector in size and scope seems to be possible, but our results show that, on average, it does not clearly occur as an ideologically driven reaction to anti-crisis policies.

Keywords

government growth regulation economic crisis government ideology policy ratchets

Introduction

Following the 2008 financial and economic crisis, discussions on the optimal level of government intervention in the economy have become very intense. Some prominent voices, such as Krugman (2012) or Piketty (2014), have argued for much more active government, postulating that current public sector size, or regulatory activity, are insufficient to overcome the long-term effects of the Great Recession. Large increases in government spending were precisely the initial reaction to the 2008 crisis, according to recent findings by Alesina et al. (2015). In an attempt to offset the expansionary increase in spending though, many Organisation for Economic Co-operation and Development (OECD) governments shortly afterwards designed deficit reduction policies that had exactly the opposite objective. As this concrete example highlights, the question of whether crisis-driven increases in government size and regulation become permanent, whether they are purely temporary, or whether perhaps the overall effect is even negative, is highly relevant in economic policy.

Previous theoretical and empirical studies argue that economic crises may lead to more interventionist policies and bigger governments, but also cause deregulation and reductions of public sector size. Notably, Peacock and Wiseman (1961) highlight with their famous displacement effect hypothesis that government not only increases as a consequence of economic growth (Wagner, 1893), but also through political crises, namely, interstate wars. Such events permanently displace the government’s capacity to tax, as spending is never readjusted to prewar levels and new ideas have meanwhile emerged among the electorate on what tolerable tax levels should look like. Similarly, Buchanan and Wagner (1977) argue that in the case of economic crises, the acceptance of Keynesian economic policy tends to relax political constraints on deficit finance, allowing government to spend more than otherwise possible. As such, while Buchanan and Wagner focus on what may be termed the political supply side, Peacock and Wiseman’s theory rests on changes in the electorate’s demand for interventionist policies.

Regarding the long-term effects of an economic crisis, Higgs (1987) believes it to be unlikely that government size will be rolled back to pre-crisis levels, thereby giving every economic crisis a permanent component, denoted as the ratchets effect. On the other hand, several authors maintain that an economic crisis should actually be beneficial to the acceptance of pro-market reforms, including the reduction of public sector size and economic regulation (e.g. Drazen and Grilli, 1993; Krueger, 1993; Rodrik, 1996). In this view, an economic crisis is a politically unique moment in time where deteriorating economic conditions facilitate major policy changes to permit a more liberalized economy with a smaller public sector.

Interestingly, both viewpoints interpret major economic crises as a type of critical juncture, where political reactions will, at least in part, be determined by historically path-dependent processes, but where the decisions taken at the moment of the critical juncture will also affect future events (cf. Soifer, 2012). Due to small differences in initial conditions and the decisions taken to combat the slump, the same crisis can send countries down radically different paths with respect to government intervention. According to Acemoglu et al. (2009), this is why all empirical studies dealing with critical juncture events need to control for constant, potentially historical, factors.

Maybe for this same reason, the empirical evidence from cross-country studies is equally mixed, with some authors finding that crises cause governments to tighten the interventionist grip on the economy, and others concluding that they are moments in which deregulation and less interventionist policy can be implemented. Funashima (2017), for example, finds the displacement effect to play a crucial role in the long-run growth of the public sector, and Baier et al. (2012) conclude that crisis episodes are associated with lower overall economic freedom. In contrast, De Haan et al. (2009) only find this effect to be temporary and that in the long run, public sector size is actually reduced after a banking crisis. Likewise, Pitlik and Wirth (2003) find that substantial crisis episodes facilitate overall economic deregulation, and Pitlik (2008) adds the important notion that this relationship is mediated by political institutions, with democratic governments being more prone to liberalize after a crisis. Recently, Young and Bologna (2016) undertook an exhaustive study of how several different types of crises affect the size and scope of government, concluding that reactions are for the most part very idiosyncratic and cannot be generalized across countries.

Given the previous findings, our article introduces an important innovation that has not been tested jointly in the empirical literature. First, we argue that while economic crises do represent critical junctures where more substantial change is possible, whether or not governments implement interventionist policies will depend heavily on their core economic ideology, where views of government control and the market economy differ greatly across the policy spectrum. On the political demand side, a long tradition in political science on issue ownership documents that specific types of parties own, or clearly represent, subjects that are of particular relevance to certain groups in society (Petrocik, 1996), and that voters align their political choices along mental models where many different positions are aggregated into an ideology (Denzau and North, 1994). Regarding the political supply side, Haidt (2012) further shows that ideology is highly stable over an individual’s lifetime and that it principally functions as a moral belief system, which determines personal views on capitalism and the appropriate role of government in society. Therefore, it should also have a decisive impact on the political answer to a crisis, as those in government will fall back on this belief system to make decisions at a point in time, when reliable information about the economy is only available with a certain time lag. Empirical studies by Potrafke (2010), Pickering and Rockey (2011, 2013), Bjørnskov and Potrafke (2012), and Faccini and Melki (2014) all confirm that government ideology is a potentially important determinant of economic policy. We thus expect ideologically heterogeneous policy reactions to crises, with more left-wing governments creating bigger and more permanent stabilization programs.¹

Second, we argue that Tullock’s (1975) transitional gains trap, as well as democratic political dynamics, may prevent crisis policies from being rolled back when the crisis is over. Tullock (1975) argued that any policy benefiting a particular industry will inevitably create a dynamic, where the short-run transitional losses from renewed deregulation exceed the immediate welfare cost of reduced output and higher prices from regulation. In addition, the industry in question will use intense lobbying activity to prevent the program from being eliminated, despite the fact that exceptional benefits only accrue to the first generation of license holders. Rent-seeking groups may thus be able to maintain regulation for extended periods of time, especially if substantial technological change is absent in a particular market (Coates et al., 2010; Thomas, 2009). Given that this dynamic is extended to a sufficient number of groups, for example, via Keynesian anti-crisis programs, democratic economic policy may very well become dominated by intense lobbying activity that has the principal goal to maintain or expand these programs. In the words of Olson (1982), these lobbying activities lead to institutional sclerosis in which special interests and poor policies reinforce each other.² To the extent that interventionist crisis policies create transitional special interests and political credibility problems, we therefore expect to observe policy ratchets, that is, that temporary crisis policies become quasi permanent by increasing government’s role in the economy.

Employing a panel of 69 countries between 1975 and 2015 for which we can construct an indicator of government’s economic ideology, we find that crises in general cause more interventionist policies when countries have left-wing governments. We track the development in indicators of government size and regulatory interventions by using two subcomponents of the Economic Freedom of the World index developed by Gwartney et al. (2017). Findings also show clear evidence of policy ratchets in the case of government size, albeit not necessarily for overall regulation, indicating that some of the policies chosen during the crisis tend to become permanent. Finally, we discuss evidence suggesting that our main findings are robust and can be interpreted as reasonably causal.

The rest of this article is structured as follows: Section “Data and empirical strategy” describes our data and the estimation strategy. Section “Results” reports the results, while we discuss the findings and conclude in section “Conclusion.”

Data and Empirical Strategy

In this article, we explicitly focus on government growth and changes in its regulatory activity. Potentially, each can act as substitutes for the other, as highlighted by Posner (1971). If one accepts his description of regulation as a form of taxation, any empirical investigation will automatically need to take both aspects into account, because by focusing only on the budgetary actions of government, we would be leaving out an important dimension of government activity that is functionally equivalent and directly related to its overall size (Holcombe, 2005).³

To measure government size and economic regulation, we employ two underlying elements of the Economic Freedom of the World (EFW) index by Gwartney et al. (2017). This index is published annually by the Canadian Fraser Institute, reflecting the degree to which the economic institutions and policies of a country correspond to free market principles. The index is divided into five major areas: (1) Size of government: Expenditure, taxes, and enterprises, (2) Legal structure and security of property rights, (3) Access to sound money, (4) Freedom to trade internationally, and (5) Regulation of credit, labor, and business. In all areas, 0 represents the least free and 10 the most free.

For reasons of better interpretation, we reverse the index such that higher index values imply larger government consumption and more intervention and employ area 1 to measure variations in the size of government, and area 5 to measure changes in governments’ regulatory activity. Both are rather broadly encompassing measures, which assures that we do not overlook any possible relationship by focusing on just one indicator and thus a single policy instrument. Although we would potentially also like to split both areas into their individual components, so as to highlight the exact variations of the underlying variables, missing data partly precludes such an analysis. In addition, depending on which type of welfare state regime a country has, crisis-related increases may register in the different subareas of each category (e.g. government consumption vs transfers and subsidies), where separate policies can ultimately obtain the same aims. These are all best captured with the composite indicators.

The focus in this article is on explaining changes in government size and regulatory activity, not in comparative levels. Holcombe (2005) highlights that explanations for the size of government and its growth are naturally interrelated, but differences are nonetheless important: Government size could hypothetically be the outcome of some collective choice mechanism with a preference-driven optimality, but related theories ultimately fail to explain why government size would keep changing, unless preferences change to a substantial degree. Ultimately, this makes collective choice theories incompatible with explanations that highlight budget maximization and policy ratchets, where there really is no predefined maximum size that government can have (e.g. Holcombe, 1993; Miller and Moe, 1983; Peltzman, 1980). As this article explicitly tests for ideologically motivated budget maximization and crises as critical junctures that facilitate it, we will focus on the explanation of government growth.

We define economic crisis as the count of years with negative annual growth at 5-year intervals, meaning that we really measure growth crises. These are calculated from real GDP per capita (PPP) data, obtained from the Penn World Tables, version 9.0 (Feenstra et al., 2015). Obviously, alternative crisis definitions and measures are also available, some of which we employ as robustness checks in the Appendix (see Table A3). As can be observed in the summary statistics of Table 1, the average number of crisis years per country is around one for each 5-year period. Growth crises are, however, not evenly distributed and are much more concentrated among the developing countries of our dataset, which is also reflected in the relatively high standard deviation of this variable. This explains why the average number of crisis years might be regarded as relatively high, when judged from the perspectives of Western developed countries.

Table 1.

Descriptive Statistics.

Variable	Mean	Std. deviation	Observations
Government size (EFW area 1)	4.140	1.597	627
Regulation (EFW area 5)	3.500	1.333	579
Crisis years	1.053	1.144	608
Light crisis	0.293	0.456	608
Hard crisis	0.285	0.452	608
Log GDP per capita	9.506	0.778	619
Openness	0.629	0.537	619
Log population size	1.991	1.636	619
Presidential democracy	0.269	0.444	625
Civilian autocracy	0.101	0.301	625
Military dictatorship	0.066	0.248	625
Coalition government	0.432	0.496	625
Government ideology	0.196	0.417	609

EFW: Economic Freedom of the World; GDP: gross domestic product.

In addition, we create two variables that distinguish between a hard crisis and a soft crisis operationalized as follows: Hard crises are defined so that the cumulative GDP losses for the corresponding economy have to be larger than 4% of GDP, while all other crises are defined as soft crises. Both variables capture the amount of years that a country spends with the corresponding negative annual growth. Effectively, this also separates short crises from longer crises, as we observe no crises with large GDP losses that last only for 1 year.

As hypothesized, whether or not governments implement more or less interventionist policies during an economic crisis will likely depend on their core ideology. While most ideology measures heavily focus on positions in the liberal-authoritarianism sphere to classify parties, often judging economic policy positions to be of secondary importance, Bjørnskov and Potrafke (2011) construct an indicator that exclusively captures the economic ideology of central governments. Similar to Wagner and Meyer (2016), this measure examines the positions and salience of economic issues as a separate dimension in political party competition, so that the ideological approach to non-economic social issues or immigration can differ substantially from economic ideology. Compared to indicators that mix both spheres, we believe that our measure is far more adequate for the investigations conducted in this article.⁴

In detail, this variable is measured with modern social democrat parties as an anchor around which other parties are placed. All parties in national parliament are distributed on a scale from −1 to 1, with −1 representing completely left-wing (communist or unreformed socialist) parties, −0.5 representing modern socialist parties, 0 representing modern social democratic parties such as the SPD in Germany, 0.5 represents conservative parties, and 1 is set to represent parties with an ideological base in classical liberalism.⁵ This categorization is employed to place all central governments on a −1 to 1 scale, where the seat share in parliament of each party is used as weights when calculating the ideology of coalition governments. To capture government ideology for 5-year intervals, ideology scores are simply averaged over each time period. Albeit discussable for obvious reasons, Figure 1 shows for a selection of three countries that this measure does capture meaningful ideology differences for the period under observation by presenting the annual values and 5-year averages.

Figure 1.

Government Ideology, Three Examples.

Most importantly, we introduce an interaction term between government ideology and our crisis variables, which will be used to test any possibly heterogeneous crisis reactions by governments of differing economic ideology. In accordance with evidence from previous research, we expect the reactions of left-wing administrations to be stronger and more durable across time, when compared to the reactions of right-wing governments.

Furthermore, we introduce a number of additional controls, which previous research suggests are significantly related to government size across countries. These are all summarized in Table 1. Following Wagner’s Law, stating that the size of government grows as a share of GDP with rising average per capita income (Wagner, 1893), we introduce the logarithm of initial real GDP per capita (PPP) to our basic model, which is also taken from the Penn World Tables.

Relatedly, it has been argued by a number of scholars that there is a direct connection between openness to trade and the size of government. Most notably, Rodrik (1998) treats this relationship in his famous compensation hypothesis, which states that more open economies build larger governments as an insurance mechanism against external trade shocks. Despite recent criticism (e.g. Benarroch and Pandey, 2012; Rode and Sáenz de Viteri, 2018; Shelton, 2007), we introduce the total value of all traded goods over GDP to our model, so as to account for the existence of such a possible mechanism. Similarly, recent papers have further found smaller countries to present comparatively larger public sectors, arguing that these countries also trade more by default (i.e. Alesina and Wacziarg, 1998). Therefore, we also include the log of total population into our model. Data for both variables are also from the Penn World Tables.

Another important factor that has been argued to affect the size and scope of government is a country’s political regime (Pitlik, 2008). Here, it is generally accepted knowledge among political economists that under democratic regimes the public sector grows, as a response to voters’ redistributive demands, while non-socialist authoritarian regimes are able to keep public sectors small (e.g. Tavarés and Wacziarg, 2001). This hypothesis is also indirectly supported by Aidt and Jensen (2009), who find that the expansion of the voting franchise significantly raised government taxation and spending in the nineteenth and early twentieth centuries. We therefore introduce two different dummy variables to our model that takes the value of one, if the country is a political regime that is a civilian autocracy or a military dictatorship, respectively. Both variables are from the much-used dataset by Cheibub et al. (2010), which codes democratic and autocratic regimes worldwide since 1946.⁶

Relatedly, Persson and Tabellini (1998) find strong support that the size of government is smaller under presidential regimes, when compared to parliamentary democracy. We further introduce another dummy variable from Cheibub et al. (2010) to our basic model that takes the value of one if a country has a democratic presidential regime. Finally, Persson and Tabellini (1998) and Kontopoulos and Perotti (1999) find that countries with electoral systems with proportional representation and frequent coalition governments have, on average, larger governments. Therefore, we further introduce a common measure of government concentration to our basic model, the Herfindahl-Hirschmann index of the legislature, which is derived from our data on political parties. The common explanation is that political bargaining processes in parliamentary coalitions have a tendency to increase the size of new spending programs.

For the following reasons, we conversely chose not to employ countries’ share of the dependent population, and their degree of fiscal federalism, although they are potentially important controls. First, the share of the dependent population was never statistically significant in any estimation conducted, and generally seems to be more important for explaining government size in levels, rather than its short- to medium-term variation. Second, to our knowledge, there is no institutional index of fiscal federalism available that has a similar coverage to our dataset, and any existing measure would have significantly reduced our number of observations. Notwithstanding, the comparative degree of federalism should be adequately accounted for by our estimation procedure, as it is usually highly stable over time.

Regarding data structure and our estimation procedure, it should be noted that we observe aggregated 5-year changes in government size and regulatory activity as a function of the variables mentioned above. The dataset contains 69 countries, aggregated at 5-year periods between 1975 and 2015 (following the structure of the EFW dataset), which yields up to 535 individual observations. Observing annual data might have been somewhat more desirable, but it is neither available before 2000, nor is the lag structure with crisis obvious in that case. What is relevant for our results is that, if anything, our current data and lag structure should create a downward bias for our findings (see Table A1 of the Appendix for additional testing). Our sample is further restricted by the requirement that all countries must have formal-constitutional political institutions similar to those of Western Europe, which we implement by only including the OECD, OECD-like countries, and Latin America.⁷ Table A7 of the Appendix lists all the countries included in the analysis.

All estimations are conducted using simple ordinary least squares (OLS) fixed effects regression for panel data, introducing a one-period time lag of 5 years between the dependent variables and the independent controls. Standard errors are clustered at the country level. The initial levels of government size and regulation are also introduced as primary control variables to the model, but due to the estimation procedure with country (and time) fixed effects, the dependent variable will only capture changes in both the variables of interest.⁸ The use of country fixed effects also means that we can effectively control for stable institutional environments and the possible presence of path-dependent reactions to the extraordinary event of an economic crisis, as highlighted by Acemoglu et al. (2009). Marginal effects of interacted variables are calculated by the Delta method (Brambor et al., 2006). Our baseline specification models change in the size and scope of government in country i at time t, as follows

\begin{array}{l} {GovSize}_{i t} = β_{0} + β_{1} {GovSize}_{i t - 1} + β_{2} {Crisis}_{i t - 1} + β_{3} {Ideology}_{i t - 1} \\ + β_{4} ({Crisis}_{i t - 1} \times {Ideology}_{i t - 1}) + β_{5} {Controls}_{i t - 1} + {cfe}_{i} + {tfe}_{t} + ε_{i} \end{array}

A potential problem in this setting is that countries with a comparatively larger government (or excessive and inadequate regulation) might also present a larger crisis risk. In fact, this seems to be highly probable, according to some recent contributions (e.g. Bjørnskov, 2016; Taylor, 2009). Still, while the crisis variables may be endogenous to changes in interventionist policies, we argue that the interaction term is unlikely to suffer from endogeneity bias. According to Nizalova and Murtazashvili (2016), when one of the interaction variables is plausibly exogenous, an interaction term and the heterogeneity of an effect can still be interpreted causally. This is exactly the situation in our case, because the ideology of the government is not affected by crises: Our data show that governments often change during an economic crisis, but what is important in this context is that the change in government ideology is not in any way directional, meaning that voters are not any more likely to vote for governments of a specific ideology during an economic downturn.⁹ In addition, there is also no reason to believe that either right- or left-wing governments are more likely to create an economic crisis, due to the fact that crisis-conducive policies are usually of a much more structural nature (Taylor, 2009). Our dataset also shows a very similar distribution of government ideology with and without the occurrence of an economic crisis.

If we interpret government ideology as exogenous to crisis, the interaction term will also produce unbiased coefficients, as long as the direction and extent of omitted variable bias produced by the crisis measure does not directly depend on government ideology (cf. Dreher et al., 2016). As we argue, this is unlikely to be the case and, in addition, all country-specific structural factors that could jointly produce crisis and governments of a certain ideology are effectively controlled for by the country and time fixed effects. Therefore, coefficients on the crisis-ideology interaction variable can reasonably be interpreted as exogenous.¹⁰

Results

Figure 2 shows the evolution of our dependent variables and a measure of crisis prevalence for a sample of countries that have been democratic for at least 20 years and are therefore labeled as “stable democracies” (see Table A7).¹¹ It can be observed that average government size was particularly large in the early 1980s, after the second oil crisis. Since then it has experienced an almost continuous decline, with an absolute low point reached around the year 2005. Probably as a consequence of the crisis episode that started in 2008, the figure shows a renewed increase for the period between 2005 and 2010, after which it again declines to an absolute low in 2015. A similar evolution is visible for average regulation, albeit at a lower relative level. Overall, the graphical evolution gives some support to the hypothesis that government size and regulation increase as a reaction to economic crisis, even though it does not support the idea that (democratic) government continuously grows over time and cannot be rolled back. Naturally, this begs the question of what the general drivers of the overall decline in government size and regulation are. In the figure, rollback episodes appear to be more concentrated in phases of overall economic expansion, which undoubtedly makes this task much easier. Recent research also points to policy diffusion with a potentially prominent role for intergovernmental organizations on the political supply side (Bergh et al., 2017) and possible mentality changes on the political demand side (Pitlik and Rode, 2017). Still, it should be noted that the facts presented in Figure 2 would make an ideologically driven increase in government size or regulation from anti-crisis policies all the more notable, as these would effectively be against the overall long-run time trend during the period under analysis.

Figure 2.

Government Size, Regulation, and Crisis Prevalence Over Time.

Table 2 presents the results of our baseline model, where we look at the effects of a crisis in the preceding period, presenting also the full set of individual control variables. We employ area 1 to capture changes in the size of government, and area 5 to capture changes in overall economic regulatory activity. Columns 1 and 2 show estimates for the entire country sample, while columns 3 and 4 exclude military dictatorships, and columns 5 and 6 exclude all political autocracies from the model.

Table 2.

Overall Crisis Effects.

Sample	All	All	No military	No military	No autocracies	No autocracies
Dependent	Gov. size	Reg.	Gov. size	Reg.	Gov. size	Reg.
	1	2	3	4	5	6
Initial level	0.366*** (0.039)	0.431*** (0.066)	0.369*** (0.037)	0.419*** (0.069)	0.360*** (0.037)	0.403*** (0.074)
Log GDP per capita	0.094 (0.221)	0.381* (0.217)	0.273 (0.356)	0.420* (0.244)	0.480 (0.347)	0.760*** (0.260)
Openness	−0.197 (0.193)	−0.392*** (0.131)	−0.222 (0.206)	−0.397*** (0.132)	−0.229 (0.300)	−0.416** (0.183)
Log population size	−0.205 (0.514)	0.889* (0.532)	−0.170 (0.393)	0.821 (0.548)	0.118 (0.683)	1.078 (0.704)
Presidential democracy	−0.191 (0.483)	0.494* (0.265)	−0.298 (0.725)	0.629 (.447)	–	–
Civilian autocracy	−0.560* (0.543)	0.596** (0.237)	−0.601 (0.692)	0.658** (.265)	–	–
Military dictatorship	−0.420 (0.485)	0.407 (0.249)	–	–	–	–
Coalition government	−0.004 (0.118)	−0.061 (0.117)	0.038 (0.125)	−0.084 (0.121)	0.019 (0.131)	−0.055 (0.109)
Government ideology	−0.377** (0.165)	−0.182 (0.116)	−0.399** (0.178)	−0.192 (0.116)	−0.454** (0.176)	−0.176 (0.121)
Crises years	0.144*** (0.047)	0.073*** (0.021)	0.125** (0.049)	0.078*** (0.022)	0.139*** (0.049)	0.069** (0.024)
Crisis * ideology	−0.148 (0.094)	−0.163** (0.074)	−0.132 (0.099)	−0.189** (0.076)	−0.157 (0.104)	−0.149** (0.073)
Period FE	Yes	Yes	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes	Yes	Yes
Observations	534	502	506	489	472	458
Countries	69	69	69	69	67	67
Within R squared	0.439	0.695	0.434	0.688	0.444	0.672
Between R squared	0.597	0.263	0.764	0.257	0.738	0.104
F statistic	21.88	47.73	26.13	45.90	24.89	39.70
Crisis effect at:
10th percentile	0.217** (0.080)	0.155*** (0.048)	0.191** (0.085)	0.172*** (0.051)	0.217** (0.086)	0.145*** (0.052)
25th percentile	0.144*** (0.047)	0.073*** (0.021)	0.125** (0.049)	0.078*** (0.022)	0.139*** (0.049)	0.069** (0.024)
Median	0.102** (0.042)	0.028 (0.024)	0.088** (0.043)	0.025 (0.024)	0.095** (0.045)	0.028 (0.023)
75th percentile	0.069 (0.049)	−0.008 (0.036)	0.059 (0.050)	−0.017 (0.035)	0.060 (0.053)	−0.005 (0.033)
90th percentile	0.056 (0.055)	−0.023 (0.042)	0.047 (0.056)	−0.034 (0.041)	0.046 (0.059)	−0.019 (0.039)

Standard errors in parentheses.

Marginal effects of interactions are calculated by the Delta method (Brambor et al., 2006).

p values are significant at:

***

p < 0.01.

p < 0.05.

p < 0.10.

Before describing the principal association of interest, a few words apply regarding the other control variables. First, we find that initial levels of government size and economic regulation enter all columns in Table 2 with a positive and statistically highly significant coefficient. This indicates that those countries with already bigger governments and more regulation also have a tendency to further increase both in the following period. Second, we find little indication that GDP per capita levels, total population, presidential republics, or coalition governments bear a significant relation to changes in government size and economic regulation, although some of these result may arguably be influenced by the fact that we are not looking at levels. Third, we find some indications that military dictatorships share a tendency to increase economic regulation, all else being equal. Interestingly, they are usually associated with smaller public sectors, as reviewed in the previous section. These results would indicate that a smaller government is perhaps compensated for by comparatively higher levels of economic regulation. Fourth, the latter finding is further confirmed by the results for civilian autocracies, which also present a tendency to increase economic regulation, but show a parallel inclination to reduce government size in the economy. Finally, we find that more open countries present a significant tendency to reduce economic regulation, while the overall size of the public sector seems unaffected.

Coming to our principle variables of interest, Table 2 shows that governments with a more conservative (i.e. economically classic liberal) ideology reduce public sector size in the economy outside of crises, but that government ideology presents absolutely no relation with changes in economic regulation.¹² In turn, coefficients on the current crisis variable show that more years spent in an economic downturn are associated with larger following increases in both the size of the public sector and the overall regulation. Still, these coefficients should not be interpreted without considering the interaction effects of both variables, which is also our primary focal point in this article. The lower part of Table 2 presents the marginal effect of years spent in an economic crisis, dependent on different levels of government ideology. These results are further represented in Figures 3 and 4, where we graphically show the marginal effects of an economic crisis on government size and regulation, dependent on political ideology.

Figure 3.

Marginal Effect of Crisis on Government Size.

Figure 4.

Marginal Effect of Crisis on Economic Regulation.

The results in column 1 of the table and figures reveal that center and left-wing governments significantly tend to increase government size during a crisis. Coefficients further imply that socialist governments generate larger public sector increases as a reaction to an economic crisis, even though ideological differences per se are not statistically significant and should thus be regarded as indicative for the moment. Conversely, ideological differences are statistically significant in the case of regulation, where left-wing governments present larger regulation increases in column 2, while the results for center and right-wing governments are insignificant. The findings are confirmed when excluding different types of autocracies in columns 3–6, where a comparison of coefficients indicates that the effect seems to be very similar for democratic regimes only.

So far, a main finding from Table 2 is that governments react to a current economic crisis in rather heterogeneous ways: Governments of a center and left-wing political ideology increase the size of government as an immediate crisis reaction, but this increase is more pronounced for governments on the left side of the political spectrum. Similarly, to the extent that the average crisis estimates are unbiased, only socialist governments increase economic regulation to a significant degree, while governments with a classically liberal ideology in general seem to refrain from doing so. In both the cases, coefficients denote that the combined effect is not overly sizable, but neither is it inconsequential.

Given the fact that we find heterogeneous reactions to a current economic crisis raises the question of whether potential ratchet effects are also primarily driven by political ideology, and if left-wing administrations also create more lasting increases in the size and scope of government. Table 3 evaluates this particular question by simply introducing a two-period lagged crisis effect to our basic model to capture policy ratchets. All models presented here further employ the full set of control variables from the previous table, but due to the fact that findings are practically identical, they are simply omitted.

Table 3.

Including Ratchet Effects.

Sample	All	All	No military	No military	No autocracies	No autocracies
Dependent	Gov. size	Reg.	Gov. size	Reg.	Gov. size	Reg.
	1	2	3	4	5	6
Full baseline included
Government ideology	−0.386* (0.216)	−0.249** (0.121)	−0.425* (0.225)	−0.276** (0.128)	−0.469** (0.205)	−0.198 (0.139)
Crises years	0.136*** (0.045)	0.078*** (0.020)	0.119** (0.046)	0.082*** (0.021)	0.138*** (0.047)	0.072*** (0.023)
Crisis * ideology	−0.137 (0.094)	−0.170** (0.075)	−0.125 (0.095)	−0.196** (0.075)	−0.168* (0.099)	−0.157** (0.074)
Lagged crises years (t-2)	−0.051 (0.045)	−0.128*** (0.049)	−0.064 (0.044)	−0.124** (0.052)	−0.066 (0.046)	−0.079 (0.059)
Lagged crisis (t-2) * ideology	−0.009 (0.107)	0.098 (0.107)	0.010 (0.106)	0.099 (0.113)	0.029 (0.108)	0.048 (0.127)
Period FE	Yes	Yes	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes	Yes	Yes
Observations	524	494	496	481	464	450
Countries	69	69	69	69	67	67
Within R squared	0.453	0.707	0.449	0.700	0.459	0.677
Between R squared	0.498	0.299	0.676	0.294	0.864	0.142
F statistic	21.32	61.58	27.10	55.27	26.18	42.37
Crisis effect at:
10th percentile	0.205*** (0.077)	0.163*** (0.049)	0.182** (0.078)	0.180*** (0.051)	0.221*** (0.079)	0.150*** (0.053)
25th percentile	0.136*** (0.045)	0.078*** (0.020)	0.119** (0.046)	0.082*** (0.021)	0.138*** (0.047)	0.072** (0.023)
Median	0.097** (0.042)	0.030 (0.023)	0.084** (0.043)	0.027 (0.022)	0.090** (0.044)	0.028 (0.022)
75th percentile	0.068 (0.051)	−0.007 (0.034)	0.057 (0.051)	−0.016 (0.034)	0.054 (0.054)	−0.006 (0.032)
90th percentile	0.054 (0.056)	−0.023 (0.041)	0.045 (0.056)	−0.034 (0.039)	0.038 (0.059)	−0.021 (0.038)
Lagged crisis effect (t-2) at:
10th percentile	−0.047 (0.090)	−0.176* (0.096)	−0.069 (0.088)	−0.173* (0.101)	−0.081 (0.089)	−0.103 (0.115)
25th percentile	−0.051 (0.045)	−0.128** (0.049)	−0.064 (0.044)	−0.124** (0.052)	−0.066 (0.046)	−0.079 (0.059)
Median	−0.054 (0.034)	−0.100*** (0.033)	−0.061* (0.034)	−0.096*** (0.035)	0.058 (0.036)	−0.065 (0.039)
75th percentile	−0.056 (0.041)	−0.078** (0.037)	−0.058 (0.042)	−0.074* (0.038)	0.052 (0.045)	−0.055 (0.043)
90th percentile	−0.057 (0.048)	−0.069* (0.042)	−0.058 (0.049)	−0.065 (0.044)	0.049 (0.052)	−0.050 (0.049)

Standard errors in parentheses.

Marginal effects of interactions are calculated by the Delta method (Brambor et al., 2006).

p values are significant at:

***

p < 0.01.

p < 0.05.

p < 0.10.

Generally, the estimates in Table 3 confirm results from Table 2 on the heterogeneous effect of ideology in anti-crisis policies: The lower part of Table 3 presents the marginal effect of years spent in an economic crisis, depending on different levels of government ideology. Here, the earlier finding that, as a reaction to an economic crisis, more left-wing governments also generate larger increases in the size and scope of government is confirmed. In addition, marginal effects for economic crises lagged by a period are further presented in the lower part of the table where we find absolutely no significant evidence for heterogeneous policy reversals for the size of government, regardless of the political ideology of the following government in power. In the case of economic regulation though, lagged economic crises present an association with following reductions in economic regulation, which seem to be driven by center and left-wing autocracies. On excluding these regimes in columns 5 and 6, this effect becomes insignificant once again.

The results of Table 3 thus show very little evidence of heterogeneous policy reversals. This also means that previous increases in government size are, at least on average, not rolled back again after the crisis is over, indicating that Higgs (1987) was somewhat correct and that policy ratchets are a general phenomenon in Western democracies. Yet, Figure 2 shows that government size can, and has, been rolled back over the time period under observation, also in stable democracies. These findings are not necessarily incompatible though: As our results show, relative reductions in government size and scope do not occur as a long-term consequence of economic crisis, but perhaps, as indicated above, rather during strong growth phases and as part of a general restructuring of welfare states. The important point is that governments are not scaled back as an ideologically driven reaction to anti-crisis policies, which is the objective of our investigation, but rather as a consequence of more long-run structural developments that are essentially outside the scope of this article.

Finally, Table 4 repeats the main estimates from the previous two tables, but separates economic crises according to their total impact on the economy. This shows that most of our previous findings are driven by hard economic crises. It can be observed in column 1 that practically all governments preside over increases in government size during a hard economic crisis, regardless of their political ideology. Still, coefficients again show that more socialist governments generate much larger increases in the public sector. Similarly, left-wing governments also present comparatively larger increases in regulation as a crisis reaction in column 2, but results for centrist governments also show that they share a significant tendency to increase economic regulation during a hard economic crisis. Interestingly, columns 5 and 6 show that democratic regimes with center and left-wing governments will also significantly react to a light economic crisis, even though coefficients indicate that the magnitudes are much smaller than for a hard economic crisis.

Table 4.

Separating Short and Long Crises.

Sample	All	All	No military	No military	No autocracies	No autocracies
Dependent	Gov. size	Reg.	Gov. size	Reg.	Gov. size	Reg.
	1	2	3	4	5	6
Full baseline included
Government ideology	−0.345** (0.139)	−0.259** (0.102)	−0.353** (0.158)	−0.268** (0.104)	−0.372** (0.154)	−0.237** (0.106)
Light crisis years	0.225* (0.120)	0.142* (0.082)	0.253** (0.116)	0.140* (0.082)	0.315*** (0.115)	0.174* (0.095)
Light * ideology	−0.221 (0.273)	−0.137 (0.169)	−0.231 (0.273)	−0.133 (0.171)	−0.354 (0.248)	0.178 (0.199)
Hard crisis years	0.454*** (0.122)	0.196** (0.075)	0.404*** (0.121)	0.208*** (0.079)	0.462*** (0.116)	0.181** (0.081)
Hard * ideology	−0.474** (0.241)	−0.213 (0.193)	−0.419 (0.256)	−0.276 (0.209)	−0.499** (0.251)	0.139 (0.209)
Period FE	Yes	Yes	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes	Yes	Yes
Observations	535	503	507	490	472	459
Countries	69	69	69	69	67	67
Within R squared	0.442	0.692	0.437	0.685	0.450	0.671
Between R squared	0.573	0.272	0.709	0.269	0.776	0.104
F statistic	19.84	52.04	24.78	50.61	22.94	44.74
Light crisis effect at:
10th percentile	0.330 (0.223)	0.210 (0.153)	0.369 (0.228)	0.207 (0.154)	0.492** (0.213)	0.263 (0.183)
25th percentile	0.225* (0.120)	0.142* (0.082)	0.253** (0.116)	0.140* (0.171)	0.315*** (0.115)	0.174* (0.095)
Median	0.165* (0.086)	0.103* (0.059)	0.188** (0.089)	0.103* (0.059)	0.216** (0.094)	0.124** (0.063)
75th percentile	0.119 (0.090)	0.073 (0.065)	0.138 (0.109)	0.074 (0.065)	0.138 (0.109)	0.085 (0.067)
90th percentile	0.099 (0.101)	0.061 (0.073)	0.116 (0.125)	0.062 (0.074)	0.105 (0.123)	0.069 (0.076)
Hard crisis effect at:
10th percentile	0.691*** (0.207)	0.302** (0.152)	0.614*** (0.221)	0.346** (0.164)	0.712*** (0.211)	0.251 (0.170)
25th percentile	0.454*** (0.122)	0.196** (0.075)	0.404*** (0.121)	0.208*** (0.079)	0.462*** (0.116)	0.181** (0.081)
Median	0.321*** (0.098)	0.136** (0.064)	0.286*** (0.100)	0.131** (0.064)	0.321*** (0.099)	0.142** (0.060)
75th percentile	0.217** (0.104)	0.089 (0.083)	0.194* (0.116)	0.069 (0.085)	0.212* (0.118)	0.112 (0.078)
90th percentile	0.173 (0.112)	0.069 (0.096)	0.154 (0.129)	0.044 (0.099)	0.165 (0.132)	0.098 (0.092)

Standard errors in parentheses.

Marginal effects of interactions are calculated by the Delta method (Brambor et al., 2006).

p values are significant at:

***

p < 0.01.

p < 0.05.

p < 0.10.

Generally speaking, it therefore seems that a more severe economic crisis creates strong incentives for policymakers of all political ideologies to relax constraints on government spending and augment regulatory activities, but these tendencies are substantially stronger for all governments on the left. In addition, democratic politics seems to follow a dynamic that also makes center and left-wing governments react to light economic crises with stabilization programs or other types of expansions.

So as to check the robustness of our results, we present a number of additional tests in the Appendix. These provide good indications that our findings are not influenced by reverse causality or omitted variable bias. In addition, results are robust to different types of crisis, different measurement techniques, and a whole array of subsamples that could theoretically influence the outcome. None of these tests suggest that our findings present any serious methodological problems and, as such, we can be quite confident that they actually do represent crisis reactions by ideologically different governments.

Conclusion

Recent crisis episodes have once again led to heated public debates on how much government should intervene in the economy. In this article, we argue that the degree of interventionist policy pursued by governments during a crisis will crucially depend on its core economic ideology. Findings indicate that crises cause larger increases in government size and regulation when countries have left-wing administrations. We mainly encounter evidence for heterogeneous policy reversals for the area of overall regulation, while the size of government seems to be unaffected. Importantly, more severe crises create stronger incentives for policymakers of almost all ideologies to relax constraints on government spending and augment regulatory activity, but also these tendencies remain stronger for the left.

It is interesting to note that especially countries with a democratic regime seem to correspond to classical political economy and public choice expectations (cf. Tullock, 1975): Contrary to the full sample, the ideology of the incumbent government is also relevant in a light economic crisis, compared to a severe crisis. In addition, there are no ideologically heterogeneous policy reversals with respect to overall regulation in our sample of democracies. It seems as if economic crises act as events that permanently relax the ordinary political constraints, especially on democratic governments, which left-wing administrations are prone to momentarily use to maximize government budgets and introduce tighter regulations on markets. Conversely, neither right-wing nor left-wing administrations reduce the size of government on their next ascent to power, thereby effectively creating policy ratchets. On average, these increases are only partially permanent over the time period under observation, and rolling back the public sector in size and scope seems to be politically possible. Still, our results show that government is, on average, not scaled back as an ideologically driven reaction to anti-crisis policies. Other determinants of a more structural nature must be at work, which could be an interesting topic for future research.

Supplemental Material

PSX807858_supplemental_material – Supplemental material for Crisis, Ideology, and Interventionist Policy Ratchets

Supplemental material, PSX807858_supplemental_material for Crisis, Ideology, and Interventionist Policy Ratchets by Christian Bjørnskov and Martin Rode in Political Studies

Footnotes

Acknowledgements

We thank Sutirtha Bagchi, Niclas Berggren, Sebastián Coll, Enrico Colombatto, Sven-Olov Daunfeldt, Vera Eichenauer, Erich Gundlach, Jerg Gutmann, Randy Holcombe, Henrik Jordahl, Florian Neumeier, Panu Poutvaara, and three anonymous reviewers for comments and suggestions on earlier versions of the paper.

Funding

We gratefully acknowledge financial support from the Institute for Research in Economic and Fiscal Issues. Bjørnskov also gratefully acknowledges financial support from the Jan Wallander and Tom Hedelius Foundation. Naturally, all remaining errors are entirely ours.

Notes

Supplementary Information

Additional supplementary information may be found with the online version of this article.

Appendix: Crisis, Ideology, and Interventionist Policy Ratchets

Table A1: Lagging Crisis.

Table A2: Randomly Assigned Neighbor Crises.

Table A3: Alternative Crisis Indicators.

Table A4: Accounting for Early Elections.

Table A5: Only Stable Democracies.

Table A6: Determinants of Monetary Policy Responses.

Table A7. Countries Included in the Sample (Total 5-Year Observations as Stable Democracies).

Figure A1: Jackknife Estimates, Area 1.

Figure A2. Jackknife Estimates, Area 5.

Author Biographies

Christian Bjørnskov is Professor of Economics at the Department of Economics, Aarhus University, in Aarhus, Denmark, and Affiliated Researcher at the Research Institute of Industrial Economics (IFN) in Stockholm, Sweden. Most of his research focuses on the political economy of institutions.

Martin Rode is Assistant Professor of Economics at the Department of Economics, Universidad de Navarra, in Pamplona, Spain. His research focuses on political economy and public policy.

References

Acemoglu

Johnson

Robinson

, et al. (2009) Reevaluating the Modernization Hypothesis. Journal of Monetary Economics 56: 1043–1058.

Aidt

Jensen

(2009) Tax Structure, Size of Government, and the Extension of the Voting Franchise in Western Europe, 1860–1938. International Tax and Public Finance 16: 362–394.

Alesina

Wacziarg

(1998) Openness, Country Size and Government. Journal of Public Economics 69: 305–321.

Alesina

Barbiero

Favero

, et al. (2015) Austerity in 2009–13. Economic Policy 30: 383–437.

Baier

Clance

Dwyer

(2012) Banking Crises and Economic Freedom. In: Gwartney

Lawson

Hall

(eds) Economic Freedom of the World 2012 Annual Report. Vancouver, BC, Canada: Fraser Institute, pp.201–217.

Beck

Katz

Migozzetti

(2014) Of Nickell Bias and Its Cures: Comment on Gaibulloev, Sandler, and Sul. Political Analysis 22 (2): 274–278.

Benarroch

Pandey

(2012) The Relationship between Trade Openness and Government Size: Does Disaggregating Government Expenditure Matter? Journal of Macroeconomics 34: 239–252.

Bergh

Dackehag

Rode

(2017) Are OECD Policy Recommendations for Public Sector Reform Biased against Welfare States? Evidence from a New Database. European Journal of Political Economy 48: 3–15.

Bjørnskov

(2016) Economic Freedom and Economic Crises. European Journal of Political Economy 45: S11–S23.

10.

Bjørnskov

Potrafke

(2011) Strikes, Ideology and Inequality: Is There Any Connection? Washington, DC: Southern Economic Association.

11.

Bjørnskov

Potrafke

(2012) Political Ideology and Economic Freedom across Canadian Provinces. Eastern Economic Journal 38: 143–166.

12.

Brambor

Clark

Golder

(2006) Understanding Interaction Models: Improving Empirical Analyses. Political Analysis 14: 63–82.

13.

Buchanan

Wagner

(1977) Democracy in Deficit. The Political Legacy of Lord Keynes. New York: Academic Press.

14.

Cheibub

Gandhi

Vreeland

(2010) Democracy and Dictatorship Revisited. Public Choice 143: 67–101.

15.

Coates

Heckelman

Wilson

(2010) The Political Economy of Investment: Sclerotic Effects from Interest Groups. European Journal of Political Economy 26: 208–221.

16.

De Haan

Sturm

Zandberg

(2009) The Impact of Financial and Economic Crises on Economic Freedom. In: Gwartney

Lawson

Hall

(eds) Economic Freedom of the World 2009 Annual Report. Vancouver, BC, Canada: Fraser Institute, pp.25–36.

17.

Denzau

North

(1994) Shared Mental Models: Ideologies and Institutions. Kyklos 47: 3–31.

18.

Drazen

Grilli

(1993) The Benefit of Crises for Economic Reforms. American Economic Review 83: 598–607.

19.

Dreher

Eichenauer

Gehring

(2016) Geopolitics, Aid, and Growth: The Impact of UN Security Council Membership on the Effectiveness of Aid. World Bank Economic Review 32 (1): 268–286.

20.

Faccini

Melki

(2014) Political Ideology and Economic Growth: Evidence from the French Democracy. Economic Inquiry 52: 1408–1426.

21.

Feenstra

Inklaar

Timmer

(2015) The Next Generation of the Penn World Table. American Economic Review 105 (10): 3150–3182.

22.

Funashima

(2017) Wagner’s Law versus Displacement Effect. Applied Economics 49 (7): 619–634.

23.

Gwartney

Lawson

Hall

(2017) Economic Freedom of the World 2017 Annual Report. Vancouver, BC, Canada: Fraser Institute.

24.

Haidt

(2012) The Righteous Mind: Why Good People Are Divided by Politics and Religion. New York: Random House.

25.

Higgs

(1987) Crisis and Leviathan: Critical Episodes in the Growth of American Government. Oxford: Oxford University Press.

26.

Holcombe

(1993) Are There Ratchets in the Growth of Federal Government Spending? Public Finance Quarterly 21: 33–47.

27.

Holcombe

(2005) Government Growth in the Twenty-First Century. Public Choice 124: 95–114.

28.

Kontopoulos

Perotti

(1999) Government Fragmentation and Fiscal Policy Outcomes: Evidence from OECD Countries. In: Poterba

(ed.) Fiscal Institutions and Fiscal Performance. Chicago, IL: University of Chicago Press, pp.81–102.

29.

Krueger

(1993) Virtuous and Vicious Circles in Economic Development. American Economic Review 83: 351–355.

30.

Krugman

(2012) End This Depression Now. New York: W.W. Norton.

31.

LeDuc

Niemi

Norris

(2010) Introduction. In: LeDuc

Niemi

Norris

(eds) Comparing Democracies 3: Elections and Voting in the 21st Century. London: SAGE Publishing, pp.1–10.

32.

Levitsky

Way

(2002) The Rise of Competitive Authoritarianism. Journal of Democracy 13: 51–65.

33.

Miller

Moe

(1983) Bureaucrats, Legislators, and the Size of Government. American Political Science Review 77: 297–322.

34.

Nizalova

Murtazashvili

(2016) Exogenous Treatment and Endogenous Factors: Vanishing of Omitted Variable Bias on the Interaction Term. Journal of Econometric Methods 5: 71–77.

35.

Olson

(1982) The Rise and Decline of Nations: Economic Growth, Stagflation, and Social Rigidities. New Haven: Yale University Press.

36.

Peacock

Wiseman

(1961) The Growth of Public Expenditure in the United Kingdom. London: National Bureau of Economic Research.

37.

Peltzman

(1980) The Growth of Government. Journal of Law and Economics 23: 209–287.

38.

Persson

Tabellini

(1998) The Size and Scope of Government: Comparative Politics with Rational Politicians. Working Paper No. 6848 (December). Cambridge, MA: National Bureau of Economic Research.

39.

Petrocik

(1996) Issue Ownership in Presidential Elections, with a 1980 Case Study. American Journal of Political Science 40: 825–850.

40.

Pickering

Rockey

(2011) Ideology and the Growth of government. Review of Economics and Statistic 93: 907–919.

41.

Pickering

Rockey

(2013) Ideology and the Size of US State Government. Public Choice 156: 443–465.

42.

Piketty

(2014) Capital in the Twenty-First Century. Cambridge, MA: Harvard University Press.

43.

Pitlik

(2008) The Impact of Growth Performance and Political Regime Type on Economic Policy Liberalization. Kyklos 61: 258–278.

44.

Pitlik

Rode

(2017) Individualistic Values, Institutional Trust, and Interventionist Attitudes. Journal of Institutional Economics 13 (3): 575–598.

45.

Pitlik

Wirth

(2003) Does Crises Promote the Extent of Economic Liberalization? An Empirical Test. European Journal of Political Economy 19: 565–581.

46.

Posner

(1971) Taxation by Regulation. Bell Journal of Economics and Management Science 2: 22–50.

47.

Potrafke

(2010) Does Government Ideology Influence Deregulation of Product Markets? Empirical Evidence from OECD Countries. Public Choice 143: 135–155.

48.

Rode

Sáenz de Viteri

(2018) Expressive Attitudes to Compensation: The Case of Globalization. European Journal of Political Economy 54: 42–55.

49.

Rodrik

(1996) Understanding Economic Policy Reform. Journal of Economic Literature 34: 9–41.

50.

Rodrik

(1998) Why Do More Open Economies Have Bigger Governments? Journal of Political Economy 106: 997–1032.

51.

Shelton

(2007) The Size and Composition of Government Expenditure. Journal of Public Economics 91: 2230–2260.

52.

Soifer

(2012) The Causal Logic of Critical Junctures. Comparative Political Studies 45: 1572–1597.

53.

Tavarés

Wacziarg

(2001) How Democracy Affects Growth. European Economic Review 45: 1341–1378.

54.

Taylor

(2009) The Financial Crisis and the Policy Responses: An Empirical Analysis of What Went Wrong. Working Paper No. 14631 (January). Cambridge, MA: National Bureau of Economic Research.

55.

Thomas

(2009) Deregulation Despite Transitional Gains. The Brewers Guild of Cologne 1461. Public Choice 140: 329–340.

56.

Tullock

(1975) The Transitional Gains Trap. The Bell Journal of Economics 6: 671–678.

57.

Wagner

(1893) Grundlegung der Politischen Ökonomie. Teil I: Grundlagen der Volkswirtschaft. Leipzig, DE: C.F. Winter’sche Verlagshandlung.

58.

Wagner

Meyer

(2016) The Radical Right as Niche Parties? The Ideological Landscape of Party Systems in Western Europe, 1980–2014. Political Studies 65 (1): 84–107.

59.

Young

Bologna

(2016) Crisis and Government: Some Empirical Evidence. Contemporary Economic Policy 34: 234–249.

60.

Zakaria

(1997) The Rise of Illiberal Democracy. Foreign Affairs 76: 22–43.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.40 MB