Populism and state capture: Evidence from Latin America

This article argues that populist governments, in comparison to non-populist ones, diminish the effectiveness of the rule of law in preventing state capture by economic elites. (Uncon-strained) populism in power that reduces existing levels of the rule of law generates more state capture than (contained) populism which has no negative effects on the rule of law. The moderating effect we theorize is that the strength of the rule of law on state capture will show decreased effects at controlling state capture when a country is a populist relative to a non-populist. Populist leaders may seek to rid the old corruption networks, but once they aggran-dise their power and weaken the rule of law, they have incentives to establish corrupt linkages with opportunistic economic elites willing to work with them, increasing state capture as a result. To test this theory, the paper employs a random effects model with 6 different specifications across 18 different Latin American countries, throughout the time period 1996-2017.


Introduction
While it is important to know what enables the rise of populist leaders, it is also of critical importance to understand the impact that populism in power has on the quality of democratic governance.Some extol populism's participatory and inclusive traits and argue that it is bound to strengthen democracy (Canovan, 1999;Mény & Surel, 2002;Laclau, 2005).Others are more skeptical and point out that populist leaders tend to erode democracy (Müller, 2016;Galston, 2018;Weyland, 2020).Many populist leaders come to office promising an end to "politics as usual," to "drain the swamp," and these promises are attractive in a context where state capture has become a growing and global concern, and citizens feel alienated from their governing institutions and are fed up with political corruption.Indeed, state capture -private and public actors influencing the rules of the game for their own benefit through illicit mechanisms and private payments (Hellman et al., 2000;David-Barrett, 2021) -has increased exponentially alongside other forms of corruption.The scholarly concern with state capture grew with the fall of the Soviet Union and the advent of neo-liberalism (Rose-Ackerman & Palifka, 2016), and it has been studied in many contexts, including Latin America (Hellman et al., 2000;Karl, 2019;Durand, 2019).Despite the overwhelming literature on populism, very little work has been conducted on the impact of populism on state capture.For instance, two widely read handbooks (Rovira Kaltwasser et al., 2017;de la Torre, 2018) and two comprehensive edited volumes on populism (de la Torre & Arnson, 2013;de la Torre, 2015) have no index entry for state capture.Given that populists usually run for office campaigning against corrupt elites and their undue influence on the state and public policy, this disinterest is surprising.Do populists in power in fact reduce state capture?Or do they increase it?This paper seeks to answer these questions.
The literature suggests two contrasting hypotheses for the consequences of populism in power on state capture.Given their avowed calls for "draining the swamp," one would expect that state capture would be curtailed under populism.This expectation derives from the premise that populists are usually outsiders who come to power on a promise to end old government corruption and collusion with economic elites (Barr, 2009;Carrión, 2015;Curini, 2018;Fieschi & Heywood, 2004;Kubbe & Loli, 2020).Their outsider status and lack of ties with old elites could enable populists to move against them (Geddes, 1994).On the other hand, we know that several countries that have experienced state capture are led by populist leaders (Karl, 2019;Crabtree & Durand, 2017;Durand, 2019;Petkoff, 2011).Even populist chief executives who espoused left-wing and egalitarian rhetoric have ended up establishing mutually beneficial ties with economic elites once they consolidated their power, as Farthing (2019) has shown in the case of Evo Morales in Bolivia, and Petkoff (2011) and Carroll (2013) have illustrated for the case of Hugo Chávez in Venezuela.
This paper seeks to assess these contrasting claims.We argue that to properly analyze the role that populism has on state capture, one must acknowledge that not all populist governments are the same nor do they have consistent effects on democracy and the rule of law (Carrión, 2022;Huber & Schimpf, 2016;Pappas, 2019;Weyland, 2020).This is important because we know that a major driver of state capture is a weak rule of law (Blake & Morris, 2009;Korman, 2022) and the works just cited show that populism in power affects the rule of law to varying degrees.Thus, to examine the impact of populism in power on state capture, one must know how populism interacts with the rule of law.This claim has substantial consequences because it implies that we should not statistically assess the relationship between populism and state capture without linking populism with the rule of law.This means that the best modelling strategy to follow is a multiplicative design, in which the main predictor of interest (populism in power) is interacted (multiplied) with existing levels of the rule of law.The presence or absence of a populist chief executive provides two contrasting scenarios under which the rule of law operates in controlling state capture.Is the impact of the rule of law on state capture similar under populist and non-populist governments?If one accepts the views of the defenders of populism, one should find that the rule of law will be more effective in combatting state capture under conditions of populism in power than under non-populism.
We argue here that the anti-pluralistic and power-seeking behavior of most populist chief executives in Latin America creates governing conditions that reduce the effectiveness of courts and institutions of governmental oversight in fighting state capture.In other words, the rule of law is less effective in fighting state capture under populism than under non-populist governments.We argue below that the mechanisms behind this moderating (weakening) effect is the control of the judiciary and the emergence of network effects under populism.To test this hypothesis, we use a pooled time series cross-sectional design using available data from 18 Latin American countries from 1996 to 2017.We employ a random effects model because it enables us to assess the impact of our independent variables on state capture in a robust manner that minimizes bias to our coefficient estimates yet maximizes the generalizability of our results.Our main determinant of interest is the interaction (multiplication) term of the dummy variable for a populist president in office and the the Varieties of Democracy (V-Dem) rule of law index.
Given that our study, as far as we know, is the first that explicitly connects populism and state capture in a quantitative fashion, it is convenient to restrict it to countries that have ample experience with both, hence our decision to focus on Latin America.We argue that our findings offer a starting point for those who want to replicate them in other regions.The paper is organized as follows.In the first and second sections, we offer a definition of state capture and populism and explain how we measure both in this paper.The third section discusses the mechanisms that explain why populism in power lessens the impact of the rule of law in controlling state capture.After that, we detail our research design, methodology, and results.Finally, we conclude with a reflection on the overall implications of the findings on the role of populism in state capture in Latin America.

What is state capture?
State capture has gained significant attention in recent years due to its potentially negative impact on democratic institutions, the rule of law, and economic growth.State capture refers to individuals or groups of individuals using their positions of power or influence to manipulate the decisions made by the state in their favor.In other words, state capture occurs when private interests can take control of public institutions and decision-making processes, often for their own personal gain (Hellman et al., 2000).Capture economies offer an environment where the rules of the game are skewed in favor of incumbent (domestic) interests.Thus, captor firms turn to illicit mechanisms including bribery and other corrupt means to gain entry (Hellman et al., 2000).Captor firms essentially purchase private property protection a la carte from the state (Hellman et al., 1999).In capture economies, certain business actors remain favored over others, and they are able to benefit from state-linkage privileges such as the awarding of government contracts and licenses to operate in a given market.This scenario offers a hostile environment for enterprises that desire to conduct business in licit manners free from corruption and state capture.At the same time, state capture creates a flexible environment where a given government can punish some old interests and establish new linkages with other private interests.
State capture can occur through various means, including bribery, corruption, cronyism, nepotism, and favouritism.This can involve the use of money, gifts, or other incentives to influence public officials and institutions.In some cases, state capture can also involve the manipulation of policies and regulations to favor specific interests, resulting in a distortion of the market or a reduction in competition.Ultimately, state capture undermines the principles of transparency, accountability, and fairness that are essential to a functioning democratic system (Durand, 2019).A captor environment skews the rules of the game in favor of actors who hold wealth and power relative to those who do not.When corruption becomes a national political issue, voters can fall prey to populist messages that promise to give them back control of their governments and fight corruption.
Measuring state capture is a difficult task because it involves examining activities that tend to be hidden from public view given their unethical or even illegal character.Nevertheless, scholars have devised different strategies to measure the prevalence of state capture.A common strategy is to rely on the perception of corruption indices (such as Transparency International's Corruption Perception Index), which are based on surveys of country experts and businesspeople.A second approach relies on determining the prevalence of certain corrupt practices, such as bribery and embezzlement, using public opinion polls or experts' surveys to measure them.Examples of these measures are the Amer-icasBarometer's Mass perception of corruption among public officials, and the Varieties of democracy's public sector corruption index.A third strategy involves analyzing the concentration of economic power and the degree of competition in a given sector or industry.This can be done by examining data on market concentration, profitability, and the presence of barriers to entry, among other factors (Durand, 2019).
None of these strategies provides a perfect way to measure state capture, especially if one is interested in determining the variation that occurs on a yearly basis.Given their imperfection, the best approach is to use an indicator that combines different strategies for measuring state capture.We believe that the World Bank's Control of corruption index (CofC for short) does a good job in this regard.This index builds on the Business Environment and Enterprise Performance Survey developed by the World Bank in 2000.Later, the World Bank refined the measurement approach, resulting in the creation of the CofC.This index has since been utilized in numerous studies to analyze state capture (Aidt & Dutta, 2008;Bagashka, 2014;Blake & Morris, 2009;Innes, 2014).It measures perceptions of the extent to which public power is exercised for private gain to include both petty and grand forms of corruption, "as well as 'capture' of the state by elites and private interests" (WGI, 2022).CofC's main advantage is that it considers acts of state capture perpetrated by private actors.Moreover, in addition to being multifaceted, the CofC includes a combination of public opinion data as well as experts' opinions on the prevalence of corruption and state capture in different spheres of government.In its construction, it uses over thirty different data sources, both official and non-official (WGI, 2022).Its main drawback is that 1996 is the earliest year for which it is available.

Defining populism
There is a vast scholarship that explains why populism is prevalent and attractive to the mass public in Latin America.While there is no consensus on its root causes, political discontent caused by crisis situations and grievances against elites for their policy failures and enduring social inequalities feature in most explanations (Weyland, 2001;de la Torre & Arnson, 2013;Carrión, 2015;Ruth, 2018).Similarly, the impact of populism on democratic governance in Latin America remains a subject of ongoing debate, with some extoling its democratizing benefits (Collins, 2022) and others warning against the dangers it poses (Weyland, 2020;Carrión, 2022).Some populist governments have had quite negative effects on the rule of law (Weyland, 2020;Carrión 2022).Centralization of power, constitutional reforms to remain in office, interference with the judiciary, and control over or restrictions on opposition organizations have been identified as common strategies employed by some populist leaders to consolidate their political dominance and undermine institutional checks and balances (Levitsky & Loxton, 2019).
There is a vibrant discussion about how to define populism.The "ideational" approach conceptualizes populism as a "thin ideology" that portrays society as divided between a virtuous "people" and a corrupt "elite" (Mudde & Rovira Kaltwasser, 2017).An alternative view considers populism as a political strategy for seeking and exercising power characterized by a personalistic leader who tries to establish direct and unmediated linkages with followers (Weyland, 2001).We embrace the latter approach.This political-strategic approach is a more suitable way to classify populist chief executives because it relies on their actual practices and strategies, as determined by country experts.The ideational approach, by contrast, suffers from "excessive breadth" because it uses discourse analysis to classify them.This approach frequently produces "false positives," as Weyland (2017, p. 53) has shown.It could also produce "false negatives," given that populist leaders such as Álvaro Uribe or the Kirchners score low in the Global Populism Dataset (which uses speeches to determine populism).To that end, our decision of who qualifies as a populist is largely based on the published literature (Houle & Kenny, 2016;Huber & Schimpf, 2016;Ruth-Lovell, Lührmann, & Grahn, 2019;Weyland, 2020).Later in the paper we specify which presidents are treated as populists (Table 1).Populism in power is thus a concept that includes a significant variation in terms of how it affects the rule of law.As noted, some populist governments are able to change the rules of the game to perpetuate themselves in power, whereas others are removed or impeached when they try to aggrandize their power.Some just leave the office as scheduled.Therefore, to examine its effect on state capture, one needs to analyze how populism interacts with the rule of law.

Populism, rule of law, and state capture
State capture and corruption are pervasive phenomena documented in Latin America under all kinds of governments, democratic or not (Durand, 2019, p. 37).In addition, numerous case studies have shed light on specific instances of corruption and state capture under populist leaders (Schedler, 2002(Schedler, , 2006;;Coppedge et al., 2011).While the literature has emphasized the corrupting impact of populism in Latin America, few studies have undertaken rigorous quantitative analyses to assess the extent to which populism in power influences state capture.Does populism in power have a negative or positive impact on the levels of state capture?This research gap motivates our study, which aims to provide a comprehensive statistical analysis of the relationship between populism in power, the rule of law, and state capture in Latin America from the mid-1990s to the present day.
Our study focuses on the environment that populism in power creates for the rule of law to fight state capture.We argue that populist governments, in comparison to non-populist ones, diminish the effectiveness of the rule of law in preventing state capture.While populist leaders in office have an incentive to dislodge existing corrupt practices associated with state capture, they frequently replace the old ties between the state and private actors with new ones.Given their anti-pluralistic predispositions, populists tend to "colonize," "occupy," "seize," or seek "power asymmetry" to secure self-aggrandizement (Müller, 2016;Pappas, 2019;Carrión, 2022).Several mechanisms are used to achieve this power dominance.The first is a constitutional change to grant themselves the ability to appoint those in charge of oversight and judicial institutions (Corrales & Penfold, 2015;Mauceri, 2006;Brewer-Carías, 2010).The second is the weakening and control of civil society organizations to curtail their ability to hold populist chief executives accountable for their power-seeking behaviour (Conaghan, 2017).Finally, the control of the judiciary enables them to use what Weyland (2013) describes as "discriminatory legalism," and Corrales (2015) as "autocratic legalism," for example, the harassment and repression of opponents by employing formally legal means but effectively arbitrary use of state power.Not surprisingly, long-lasting forms of populism in power, such as Venezuela's, can lead to what López Maya (2018) calls a "criminal state."In short, populism in power leads to greater state capture by reducing the impact of the rule of law in halting predatory behaviour, either public or private (David-Barrett, 2021).
Once they have been able to aggrandize their power, populist leaders have the incentive and means to build their collusive networks with opportunistic private interests.Their monopoly of power generates network effects -where an individual enjoys a direct benefit by aligning his or her behaviour with the behaviour of others (Easley & Kleinberg, 2010).This situation provides incentives to private actors to support and strike deals with the populist government.As more private actors become supportive of the regime, the expected payoff of their support increases, providing greater direct benefits and offering impunity for their actions if they remain loyal to the populist leader.Therefore, once populist leaders rid the Ancien Régime of old corruption and economic collusion with elites, those who have effectively cemented power asymmetries offer incentives to economic elites thus creating new corrupt networks.
Indeed, case studies show how populist chief executives who have eroded judicial constraints have no problem developing ties with private interests to buttress their rule.Catherine Conaghan (2006) speaks of politically motivated corruption under Alberto Fujimori and the "greed rings" that surrounded his government.Moreover, in the absence of the fear of punishment, public officials engaged in widespread corruption (Conaghan, 2006).John Crabtree and Francisco Durand (2017, p. 103) write that Fujimori created a state that was "functional to the interests of business."Teodoro Petkoff (2011) writes that a group of business interests associated with the government, which Venezuelans labelled as the "boli-bourgeoise," are one of the pillars that held Chávez's regime together.The use of hydrocarbon rents and the corruption associated with them in the absence of governmental transparency also affected Bolivia's Evo Morales (Farthing, 2019).A significant portion of these rents were transferred to the Indigenous Fund, to support grassroots development projects.However, as Andreucci (2017, p. 174) notes, "private appropriation of funds on the part of campesino leaders and the use of the Indigenous Fund to co-opt indigenous leadership became increasingly more frequent."Therefore, our main expectation is that populism in power will provide an environment that will diminish the effectiveness of the rule of law in combatting state capture.

Research design
In this section, we discuss the data and variables we use in the statistical analysis.
Populism in power.As noted, measuring populism remains a difficult task due to a lack of consensus on definitions.The Global populism database relies on the analysis of speeches, which produces several false positives and negatives.We prefer the strategy followed by Ruth-Lovell, Lührmann and Grahn (2019).They build on the work of Ruth (2018), who classifies populist governments by a "combination of both a qualitative literature review and expert validation" (Ruth-Lovell et al., 2019, p. 15).We contrast their list of populist chief executives with those of other published authors to generate a classification of populist governments in Latin America.In short, we hand-coded populism in dummy variable format to code 1 for all countries and years that were governed by a populist government between 1996 and 2017, and 0 otherwise.& Grahn, 2019;and Weyland, 2020.
Table 1 presents the countries used in this analysis and their populist presidents along with the years they served that we constructed in order to create our populism dummy variable.In some cases, populists were in power before 1996, which is the first year we include in the analysis due to temporal limitations in the dependent variable state capture.This table lists, however, the full range of presidents and the years they were in power for the reader's sake.The actual analysis includes only the relevant populist presidents and years from 1996 through 2017 as detailed in Table 1.All other observations (presidents not listed in Table 1) in the dataset are coded as non-populist.State Capture.We use the CofC index, developed by the World Bank (WGI, 2022), to operationalize and assess state capture.Despite its name, the index evaluates more than minor or major forms of public corruption or the perceptions regarding the extent to which public power is exploited for private gains.This measure also takes into account the level of "capture" by elites and private interests (WGI, 2022), enabling a specific assessment of the impact of state capture committed by the private sector.The index ranges from -2.5 to 2.5, representing the highest to the lowest degree of state capture.To ease interpretation, we transform the original scale by adding 2.5 to each score.We also flip the sign.The resulting scale, which we use here, ranges from 0 to 5, wherein lower values indicate lesser state capture and higher values indicate greater state capture.
Rule of Law.The other major predictor in our models is the extent to which a government respects the rule of law.We measure this attribute by using the V-Dem Rule of Law Index (v2x_rule).According to V-Dem, this index measures "To what extent are laws transparently, independently, predictably, impartially, and equally enforced, and to what extent do the actions of government officials comply with the law" (Coppedge et al. 2022, p. 303) This is a 0-1 continuous index, where 0 represents the absence of the rule of law and 1 represents a country that predictably and reliably adheres to the rule of law.We argue that the effectiveness of the rule of law in preventing state capture is reduced under populist rule.To test this hypothesis, we look at the interaction term (multiplication) between the dummy variable for populism in power and the rule of law index and how it affects levels of state capture.

Control variables
State capture is also influenced by a series of factors in addition to the rule of law and whether populism is in power or not.To account for these possible influences, we include a wide set of control variables that are known determinants of state capture.One key control, which we use in all models, is regime type (Korman, 2022).Our initial base model uses this variable (described below) along with our main predictor of interest.The control variables we use derive from the World Bank, the Database on Political Institutions from the Inter-American Development Bank, the Heritage Foundation, Freedom House, and the V-Dem project.These controls are: Regime type.The Polity10 measure is utilized to assess the level of democracy and autocracy in each country as a control for regime type.Research has highlighted that democratic regimes are generally associated with lower levels of state capture compared to autocratic regimes (Treisman, 2007;Geddes et al., 2018).
Party in power [years].This variable tracks the consecutive years that a governing party remains in power.This variable comes from the Inter-American Development Bank Database on Political Institutions.Evidence suggests that longer tenures of the ruling party correspond to higher levels of state capture (Korman, 2022(Korman, , 2023)).
Property rights.Countries with stronger property rights frameworks exhibit lower levels of state capture, as these frameworks foster stability, incentivize investment, and discourage rent-seeking behaviors (Acemoglu et al., 2001).We utilize the "v2xcl_prpty" measure from the V-Dem dataset to assess property rights (Coppedge et al., 2022).
Log GDPPC.There is a negative relationship between higher GDP per capita and state capture, indicating that countries with higher levels of economic development tend to experience lower levels of state capture (Korman, 2022;Treisman, 2007).We take this variable from the World Bank indicators.
Economic freedom.The Economic Freedom Index, constructed by the Heritage Foundation, serves as a composite measure encompassing various indices related to business freedom and general business-friendliness.Studies have demonstrated that more open economies tend to exhibit lower levels of corruption and state capture (Morris, 2004).
Civil society participation.Civil society organizations play a vital role in promoting transparency, accountability, and good governance.Higher levels of civil society participation are associated with lower levels of state capture (Grzymala-Busse, 2012).We measure this variable by using Civil Society Participation Index (v2x_cspart) from V-Dem.

Methodological approach
The following analysis includes 18 Latin American countries from 1996 through 2017.The temporal scope is determined by data availability.Our study is an unbalanced panel dataset, given the missing data in our dependent variable in the years 1997, 1999, and 2001 for all units involved.The temporal scope ensures wide variation in the governments we examine not only for our key independent variables -the rule of law and populism -but also for our dependent variable -state capture.This is important because some countries in Latin America, such as Chile, Costa Rica, and Uruguay, have controlled state capture to a much better extent than other countries in the region, such as Venezuela and Nicaragua.Overall, across our six models, each estimation contains between 330 and 341 observations while our full specification and most unrestricted model (model 6) has a total of 330 observations.We use six different models, providing different combinations of the control variables.The three main predictors of conceptual interest are present in all models: the dummy variable for whether a populist government is in power or not, the Rule of Law Index from V-Dem, and the interaction term between these two predictors.Since we are using a multiplicative model, the interpretation of the unique effect of populism in power (dummy variable) is constrained: it is the impact of populism in power on state capture when the value of the rule of law index is equal to 0. Given that we have no instances of a score of 0.0 for the Rule of Law Index in the database, the interpretation of this particular coefficient is merely statistical.The coefficient of theoretical interest is the interaction term between populism in power (dummy variable) and the rule of law index.
To estimate our six different model specifications, we employ a random effects model to allow for random intercepts amongst country-level groups.Thus, unlike a fixed effects model, we exploit between country-group variation (Stimson, 1985;Zhu, 2012).This enables us to generalize our results about Latin America while accounting for both observed and unobserved heterogeneities to a much greater extent than a pure pooled OLS model would, while minimizing the bias to our estimates.After we ran both a Breusch-Godfrey test for autocorrelation and a Breusch-Pagan test for heteroskedasticity, we checked the robustness of our analysis by running a second round of estimates with our six same specifications.Testing for autocorrelation is important to ensure that we have properly accounted for the temporal dynamics in the study.Meanwhile, ensuring we have accounted for heteroskedasticity is critical to ensure that our results are not biased by some observations in the data doing a systematically better job at explaining state capture than others.As expected, our base estimates and model employed to estimate our six estimations were confirmed to have both autocorrelation and heteroskedasticity.Therefore, our robustness results present the same six estimations while reporting serial-correlated robust standard errors in order to control and account for both heteroskedasticity and autocorrelation in the model.

Results
The random effects regression results are shown in Table 2. Model 1 is our baseline specification and our most restricted estimation as we only assess our key independent variable (interaction term between populism in power and the rule of law) along with an additional regressor to control for regime type (we also include the individual predictors for populism in power and the rule of law).Model 6 presents our full specification and our most unrestricted estimation and is the core model of this paper.Table 3 presents the results of our random effects regression model with serial-correlated robust standard errors as a robustness check after accounting for both heteroskedasticity and autocorrelation.The same six specifications estimated in Table 2 are also estimated in Table 3.All six specifications in both Table 2 (our base results) and Table 3 (our robustness check results) include our key interaction term populism * rule of law to ensure that our argument holds across a myriad of specifications.Confirming the results of existing literature, the rule of law emerges as a highly statistically significant predictor of state capture.Given that this is a multiplicative model, the coefficient indicates the unique effect of the rule of law on state capture when populism is equal to 0 (that is, when there is no populist president in power).In total, about 72 per cent of our observations in the dummy variable are non-populist years.Our baseline specification shows this coefficient with a negative sign, after controlling for all other predictors, which suggests that when a Latin American country is not ruled by a populist leader, the rule of law depresses the score of state capture.This is true across all six models.The unique effect of populism in power on state capture when the rule of law index is equal to 0 emerges with a negative sign.Given that we have no instances of rule of law index = 0, this result is of little to no conceptual consequence.When populism in power interacts with the rule of law, we see that the sign of the relationship emerges as positive.This means that, at varying levels of the rule of law, populism in power produces more state capture than non-populist governments.This effect is statistically significant at the p < .05level after controlling for all other predictors.We explore the impact of this interaction term in greater detail through Figure 3 below which plots the marginal effects of this interaction term as we discuss Model 6. Model 6 in Table 2 is our main model and full specification, and it stands out as a robust approach to estimating state capture while addressing omitted variable bias.With nine control variables, we are able to uncover the impact of populism in power on state capture at different levels of the rule of law.In general, populism in power provides a more fertile environment for state capture by moderating the impact of the rule of law on state capture.Model 6 exhibits the highest explanatory power, explaining approximately 31 per cent of the variation in state capture according to the adjusted R 2 , surpassing all other specifications.This model offers strong support for our hypothesis.
To better understand the true impact of the rule of law on state capture, we turn to Figure 1.The figure illustrates the substantive effects, plotting the relationship between rule of law (x-axis) and state capture (y-axis), with 95 per cent confidence intervals shown in grey (based on the results of Model 6).As we progress from weaker to stronger levels of the rule of law, a consistent and significant reduction in state capture is observed.At the lower end of the rule of law spectrum (0 value), state capture levels hover around 3.7, reflecting a high degree of state capture in the absence of strong legal foundations.Conversely, at the upper end (1 value), state capture levels decrease to approximately 2.2, holding all other variables at their mean values.The magnitude of the effect of the rule of law on state capture is substantial.We now turn to analyze in more detail our interaction term reported in Model 6 Table 2 (rule of law * populism).The coefficient is positive and highly statistically significant, suggesting that populism offers a more conducive environment for state capture by blunting the effects of the rule of law on state capture.We present a forest plot (Figure 2) of our regression coefficients (Model 6) to visualize the magnitude of our predictors on state capture.We see from Figure 2 that of the three predictors that have a positive impact on state capture, the interaction term rule of law * populism shows the strongest effect.In order to better assess the true effect of this variable, we turn to Figure 3.To assess the true impact of the interaction term on state capture, we now turn to Figure 3, which displays an interaction plot showing the marginal effects of the rule of law on state capture when a country is ruled by a populist (solid line) relative to a non-populist (dotted line) while holding all other variables at their means.This figure is consistent with our argument: populist leaders moderate the impact of rule of law on state capture.At stronger levels of rule of law (from about a value of .60 on the x-axis onward), the rule of law decreases its effectiveness in combating state capture when a country is populist relative to non-populist.

Robustness checks
To ensure the robustness of our results, we also present Table 3, which re-estimates the same six specifications from Table 2.However, in Table 3, we report the regression output with serial-correlated robust standard errors to address the presence of heteroskedasticity and autocorrelation, as confirmed by the Breusch-Pagan and Breusch-Godfrey tests.These robust standard errors account for potential biases and improve the reliability of our estimates.Overall, the results from Table 3 confirm the robustness of our main specification (Model 6 in Table 2).While there are slight variations in the regression output between the two tables, with some regressands losing statistical significance or experiencing lower confidence levels, the core findings remain consistent and robust.Specifically, when analyzing Model 6 in Table 3, which employs serial-correlated robust standard errors, the results align with those presented in Model 6 of Table 2.It is important to note that the robustness check using serial-correlated robust standard errors inherently leads to a loss of precision in our estimates, as the standard errors increase, subsequently affecting the level of statistical significance.This is a common characteristic of robustness checks that aim to ensure unbiased results while maintaining consistency.While the findings in Table 3 may exhibit slightly weakened statistical significance compared to Table 2, it is crucial to emphasize that our results remain consistent and reliable.The main predictor of interest emerges with the hypothesized negative sign, lending further support to our hypothesis.

Discussion and conclusion
Our study seeks to analyze the impact of populism on state capture by exploring how effective the rule of law is in combating state capture under conditions of populism and non-populism.Our hypothesis posited that given the propensity of some populist leaders to establish unconstrained and dominant power asymmetries, populism in power would diminish the effectiveness of the rule of law in controlling state capture in relation to non-populists.Our results not only confirmed this hypothesis but also exhibited robustness, accounting for serial-correlated robust standard errors.Our study contributes to the literature on state capture by explicitly examining the role of populism on state capture, instead of solely focusing on the influence of business power on governments, as previous research has done (Durand, 2019;Crabtree & Durand, 2017).Our findings suggest that the potential of populism in power not only undermines the rule of law but in doing so, also strengthens the very state capture they claim to fight against when running for office.Indeed, several case studies (cited) suggest that populists in power generate their own corrupting networks, thus increasing state capture.Using a quantitative design that encompasses 18 countries from 1996 to 2017, we find that, indeed, populism in power can encourage state capture by reducing the effectiveness of the rule of law.When populist leaders can control the judiciary, the path is available to both captor actors in informal networks and populist leaders to collude.Therefore, for the protection of judicial independence, powerful mechanisms to fight state capture should include robust procedural measures for judicial appointments and adequate resources for the courts and legal system.This approach would minimize delays and attempts to subvert the rule of law.Ultimately, democratic governance that respects the rule of law remains the optimal path to prevent state capture (Whitehead, 2017).Unfortunately, this is a challenging task for Latin America, given the region's historical struggles in upholding the rule of law (Edwards, 2009).
While our findings are robust, this study also reveals areas for further improvement and exploration.First, despite its benefits and that it remains the standard, our dependent variable, state capture, relies heavily on perceptions by both experts and the general public.The need to devise better ways to measure state capture continues to be an important task.Another limitation lies in our use of a dummy variable format to represent populism to capture its presence or absence.While we attempted to account for the role of populism in state capture by interacting populism with the rule of law, a more refined approach would involve constructing an ordinal/interval variable that quantifies the extent to which populist chief executives undermine institutions of accountability.Creating an ordinal variable of populism, akin to the democratic indices constructed by Varieties of democracy, is a task for future research.
We finish by stressing that our main finding, which is that populism in power increases state capture by moderating the effects of the rule of law on it, needs to be replicated in other parts of the world where the rule of law is less subject to the vagaries of who is in power and the legislative majorities they control.Perhaps the impact of populism in power on state capture is less dramatic in regions where the rule of law is not in such flux as it has been in Latin America.As it is, our study offers conclusive evidence that the most fertile terrain to reduce state capture is to have medium to high levels of the rule of law in combination with non-populist governments.

Figure 1 .
Figure 1.Substantive effects rule of law on state capture

Figure 3 .
Figure 3. Marginal effects rule of law on state capture in the presence/absence of populism is a Professor of Latin American Politics at the Department of Political Science and International Relations at the University of Delaware.He was the Founder and Director of the Center for Global and Area Studies.He holds a degree in Sociology from the University of San Marcos (Lima-Peru) and a PhD in Political Science from the University of Pittsburgh, Pennsylvania.

Table 1 .
Populist presidents and years served in office

Table 3 :
Robustness check results in random effects model | Dependent variable is state