Praising the leader: personalist legitimation strategies and the deterioration of executive constraints

ABSTRACT
 In the face of current democratic backsliding and autocratization processes, research has rediscovered issues of autocratic legitimation. However, the question of whether rulers’ personalist rhetoric to bolster their legitimacy is followed by congruent political action remains underspecified. Using new expert-coded measures for 164 countries from the Varieties of Democracy project, we examine the political rhetoric–action link using using fixed effects models. The results confirm that shifts towards personalist legitimacy claims are no cheap talk but oftentimes important warning signals for a substantial deterioration of democratic quality, manifested in weaker judicial and legislative oversight of the executive branch. However, in contrast to much current concern, we show that liberal democracies seem to largely escape the negative repercussions of government discourses that increasingly stress the uniqueness of the ruler.


Introduction
Before he regained power in 2010, Hungarian prime minister Viktor Orbán made his legitimation strategies abundantly clear: Not only did he announce an "electoral revolution" 1 but also "tapped into longstanding popular resentment against the political elite" 2 and stressed how he as the leader would overturn the country's existing "corrupt" liberal order. These claims were a precursor of the things to come. Orbán and his Fidesz government changed the Constitution, most notably limiting the separation of powers; filled the Constitutional Court with government supporters; politicized the state-owned broadcaster; co-opted private media; and changed electoral districts to solidify their electoral advantage. 3 The consequences have been dramatic: Hungary, once heralded as a poster child for successful democratic transition in Eastern Europe, has regressed, becoming an electoral-authoritarian regime in 2019 -the first nondemocratic member of the European Union. 4 The public legitimation strategies of Orbán that focused on his personal qualities as a guardian of the people's will and on "religious conservatism and nationalism" 5 were an important discursive step and indeed a warning signal for the ensuing reduction of judicial and legislative checks on the executive. To what extent is the Hungarian case typical? Investigating a global sample of 164 countries, we take governments' legitimation strategies seriously and ask: Are shifting claims to legitimacy that increasingly stress the uniqueness of a political ruler a precursor to substantial action, namely the deterioration of judicial and legal constraints on the executive?
All governmentsdemocratic and autocratic alikemake claims that provide justification for their right to rule. Accordingly, research on the legitimacy of both democracies 6 and autocracies 7 has a long tradition. While the "third wave" of democratization pushed questions on (autocratic) legitimation strategies into the background for some time, the current scholarly focus on different facets of authoritarian rule has brought issues of political legitimacy back in with a vengeance. 8 Important triggers for this rediscovery of questions of authoritarian legitimation have been creeping processes of autocratization and democratic backsliding 9 as well as the success of elected populist leaders that claim to represent the will of the peoplesuch as the United States' Donald Trump, Brazil's Jair Bolsonaro, and, of course, Hungary's Viktor Orbánacross the globe. 10 Making claims that stress their superior qualities as political leaders are defining features of their rule. 11 The collection of new, expert-coded data for almost all the world's countries from 1900 to 2020 by the Varieties of Democracy (V-Dem) project 12 allows us for the first time to examine long-standing questions about the political relevance of regimes' claims to legitimacy in both democracies and autocracies in a systematic cross-national investigation. Thus, we assess the rhetoric-action link and evaluate to what extent a country's institutional quality of democracy affects this link.
We focus on governmental claims to legitimacy and investigate shifts towards legitimation strategies that put the person of the ruler centre stage which is a key component of illiberal reasoning. 13 Accordingly, in this initial analysis, we examine whether an increasing use of personalist claims can signal growing executive aggrandizement. Using established fixed effect models as the main modelling technique, we assess how within-country shifts signify ensuing reductions of executive constraints.
Our findings show that legitimation shifts indeed are an important discursive step towards substantive action, in particular for a subset of political regimes: In authoritarian regimes, which by definition have fewer constraints on the highest executive power, an increase in personalist rhetoric regularly serves as a precursor to a further decline in institutional constraints, while liberal, established democracies are largely spared from its detrimental negative repercussions, at least on the institutional level. This implies that personalist rhetoric might be less harmful for judicial and legislative constraints in liberal democracies than feared in the current intense autocratization debate. 14 With this analysis the article adds to the empirical study of the relationship between personalist government rhetoric and subsequent action and thereby contributes to the discussion about legitimation strategies as key means of political rule.
The article proceeds as follows. We first present our theoretical framework and introduce the concept of legitimation and personalist legitimation strategies. We then present our empirical strategy and discuss the findings. We conclude by presenting considerations for future research on claims to legitimacy as precursors to substantial action.
Theoretical considerations: the importance of shifting legitimation strategies Political legitimacy is a key pillar on which the power of any regime restsin conjunction with the provision of public goods, repression, and co-optation. 15 According to Geddes 16 , "even very coercive regimes cannot survive without some support." Citizens see legitimate authorities as "entitled to be obeyed" and "voluntarily defer" to them. 17 Compared to repression and co-optation, the existence of legitimacy is therefore a particularly (cost-)effective means of rule.
While there are insightful normative conceptions of legitimacy, we employ an analytical, Weberian understanding of it. 18 Conceptually, legitimacy consists of two dimensions: the claims invoked by rulers to justify their rule and their reception by the ruled (Legitimitätsglauben). In this study, we focus on the first aspect: the claims to legitimacy 19 a government makes vis-à-vis its citizens and members of the political elite regarding why it is endowed with the right to rule. 20 These claims can be either purposeful manipulations or genuinely held beliefs among the ruling elite; they can reflect or be totally detached from empirical reality.
In essence, governmental legitimation discourses are concerted efforts to satisfy "the need for justification." 21 According to Weyland, who refers to Goldstein and Keohane 22 , these ideational factors can shape causal beliefs, which in turn "influence actors' calculations about the benefits and costs that different feasible options are likely to yield." 23 More fundamentally, they affect principled beliefs of what is right and wrong. Hence, regime legitimation strategies fulfil three core political functions. First and foremost, they may strengthen the bonds between the regime and citizens and among members of the ruling elite. 24 Second, the claims to legitimacy set the discursive space within which individual actors and organized groups can or cannot express alternative points of view, criticize the government, and voice dissent. 25 Third, they provide the narrative framework within which policy outcomes and political actions are evaluated. In sum, legitimation strategies are the key modes of how governments justify their rule and thereby strengthen their grip on power. 26 The increasing importance of personalist legitimation strategies Political legitimacy may rest on different bases. As Weber found, "the type of obedience, the type of administrative staff developed to guarantee it, the mode of exercising authority […] all depend on the type of legitimacy claimed." 27 Most fundamentally, he introduced three ideal-types of legitimate rule: rational-legal, traditional, and charismatic authority. 28 Our analysis in this article builds on the ideal-type of charismatic rule and its implications. "Charismatic rule," as coined by Weber, stresses that a ruler's personal traits and extraordinary leadership skills can make a political order legitimate in the eyes of their followers. This discursive strategy stresses "the exceptional sanctity, heroism or exemplary character of an individual person, and of the normative patterns or order revealed or ordained by him." 29 In extreme cases the ruler is idealized as the father or mother of the nation, being exceptionally heroic, caring, or wise. 30 Thus, conceptually, while personalism in this context merely is a discursive focus on a leader's extraordinary capacities, the Weberian ideal-type of charisma not only hinges on leaders' portrayal of themselves as indispensable but also focusses on the reception of these strategies by their supporters. Leaders using charismatic legitimation strategies seek to create "linkages with followers that are unmediated, asymmetrical, and deeply emotional in nature." 31 As outlined previously, in this article we focus on the first aspect, the use of charisma as legitimation strategy, while leaving its reception by followers to future research.
As found by Weyland and others, the discursive focus on the person of the leader also constitutes a central element of the political-strategic approach to populism. 32 Classic works in this tradition, particularly in the Latin American context, have repeatedly stressed how central the person of the ruler is for populist legitimation strategies. While populist leaders portray themselves as embodying the "true" will of the masses vis-à-vis a "rotten" elite, they downplay the importance of institutional safeguards such as parliaments and the judiciary that restrain their room for manoeuvre. 33 Furthermore, the discursive attempt to establish a strong direct connection to their followers and to construct "us vs. them" narratives tends to side-line the importance of established formal rules and checks and balances. A distinct but related strand of literature has demonstrated the adverse effects of regime personalization for a host of democratic outcomes. 34 Thus, personalism is a key component of both charismatic rule and populist discursive strategies that inherently downplay the importance of institutional safeguards against executive aggrandisement.
Discursively focusing on the extraordinary capacities and centrality of the ruler may pave the way for silencing counterforces and weakening institutional safeguards. 35 It thereby allows a disregard for "formal equality (before the law) and formal freedom (from arbitrary treatment) in the form of civil and political liberties." 36 This legitimation strategy runs counter to the fundamental tenets of liberal democracy, namely "that the power of the majority must be limited and restrained, the sanctity of individual rights and the principle of the division of powers [preserved]." 37 Though we expect personalist claims to stand out in the cases of interest, governments regularly invoke different claims to legitimacy in parallel to strengthen their rule: Legitimation strategies are not exclusive but always multidimensional. 38 They can even be contradictory. Most notably, in contrast to personalist claims, rational-legal legitimation strategies include a government's claims to have the "right" procedures in place. These legal norms and regulations can be laid out in the Constitution and/or other laws codifying access to power (e.g. elections) and the exercising of power (e.g. rule of law). 39 These claims may diverge widely from the institutional reality. Zimbabwe's current president Mnangagwa has not only repeatedly stressed his role as supreme leader of the nation but has also simultaneously made (completely unrealistic) claims about preserving electoral integrity and procedural fairness. 40 Additionally, Lipset 41 , Easton 42 , and more recent research 43 have underlined the relevance of narratives that laud the performance of a regime in satisfying public demands, be it the provision of security or delivering socioeconomic development. Authors have recurrently stressed how important this performance legitimacy is for authoritarian regimes, particularly for those that do not have strong ideological foundations. 44 Accordingly, we include both performance-based and ideological claims in our analyses. Figure 1 lays out the relationship between the different legitimation strategies, meaning the extent to which they are invoked in parallel.

Legitimation strategies and regime type
It is important to note that there is no direct overlap between claims to legitimacy and a particular political regime type. Figure 2 shows the distribution of leader-centred claims across the autocracy-democracy divide. Autocrats most often focus on the uniqueness of the ruler to inflate their appeal among both the population and the political elite. 45 Still, also in democracies we seeto varying degreesruler-based appeals by leading politicians that transcend the routinized application of impersonal rules and laws. In South Africa, for instance, the first democratically elected president Nelson Mandela had an extraordinarily strong personal appeal. 46 When he held office, ruler-based claims almost doubled in strength and were the most pronounced legitimation strategy, rated between 0.73 and 0.76 (on a 0-1 scale) by the V-Dem project. Similarly, the discursive focus on rational-legal legitimation strategies varies substantially in democracies.
As institutional constraints are already weaker in autocratic regimes than in liberal democracies, it should be generally easier for authoritarian rulers to shift their legitimation strategies and in the following further limit the effectiveness of institutional constraints on executive action. In this sense: Strong democracy may function as a  The lines under the boxplot display the measurement-occurrences. Democracy combines the Regimes of the World's categories "electoral democracy" and "liberal democracy", Autocracy combines "electoral autocracy" and "closed autocracy". solid barrier against its own demise. Indeed, two-thirds of cases of gradual autocratization have occurred in nondemocratic regimes. 47 However, in order to comprehensively assess the implications of an increased governmental reliance on discursive strategies that focus on the person of the leader it is important to cut across the democracyautocracy divide. New cross-sectional and longitudinal data on legitimation strategies by the V-Dem project now provide the basis for such a comparative assessment. 48

Hypotheses
In the early years of the new millennium, the euphoria over the global third wave of democratization had already given way to rising concerns about democracy's "recession." 49 According to data from the V-Dem project, 92 countries are now autocratic, while 2.6 billion people (34% of the world's population) currently live in autocratizing states. 50 Shifting legitimation strategies form an important part of these autocratization processes, also in earlier autocratization periods. 51 A government that increasingly stresses the great achievements and personal capacities of the leader might signal its intention to further strengthen its hold on power. In this way, shifts towards personalist claims discursively can pave the way for rulers' suppression of institutional checks on executive action. In other words, portraying him-/herself as savoir of the nation and embodiment of the people's will allows to justify the increased centralization of power to fulfil his/her political mission. 52 For example, in Russia amidst decreasing economic performance 53 the personal factor has massively gained in importance since President Vladimir Putin's second stint as president began in 2012. In 2011 already, Russia's former deputy prime minister and close advisor Vladislav Surkov declared "that Putin was 'sent by God' to save Russia." 54 Pictures showing the former KGB spy shirtless as a muscular leader who goes fishing and horse riding were meant to contribute to a growing personality cult. As affirmed by Mazepus et al., Putin was increasingly portrayed as the guardian of order, stability and well-being "who understood and represented the needs of ordinary Russians." 55 This shift helped to pave the way for the increasing centralization of power to the Kremlin. 56 A case in point are the constitutional changes in Russia in 2020, which have allowed Putin to further expand his executive dominance while facing almost no institutional or societal resistance. 57 Since Russia's invasion of Ukraine in February 2022, the Putin government further tightened the restrictions on potential opposition while institutional and legal oversight is almost not existent anymore.
Similarly, since 1994 Belarusian president Alexander Lukashenko has increasingly relied on personalist claims, stressing his role as "father of the nation" while simultaneously shifting the discursive focus away from the electoral process and the rule of law. 58 This increasing discursive reliance on the person of the ruler stress the centrality of the man/woman at the top and makes the dismantling of institutional safeguards appear more acceptable in the eyes of the population, and serve also to raise the discursive costs of voicing dissent. 59 In contrast to the comparatively low number of highly-publicised coup d'états where insurgents topple office-holders, elected rulers themselves constitute the biggest threat to democracy. 60 They attack particular provisions that stand in their way and aim to "sell" the erosion of democratic rights to the public. 61 This is also relevant for historic and extreme cases such as Adolf Hitler who came to power in Germany in 1933 via electoral means before he and his followers swiftly destroyed all remnants of the Weimar Republic. As "autocratizers" often pursue a majoritarian understanding of their rule, they downplay institutional safeguards against their executive dominance. 62 Institutional checks, legislative and judicial alike, provide the main guardrails regarding executive power. It is this dimension of democracy that is regularly the first to come under attack. 63 Following these findings, the observable implication for autocratizers is to stress the state's prospects under their leadership. This they do while limiting competing institutions'the judiciary and parliamentcontrol over their power. Ensuing political actions, justified by these personalist claims, can range from ignoring court orders, replacing regime-unfriendly judges, banning opposition parties, co-opting or curtailing the media, or outright altering Constitutions. These claims not only legitimize the goals of the political leader but also serve to delegitimize those oppositional actors and institutions who aim to prevent or hinder changes in this direction. 64 Consequently, we will test the following general hypothesis: H1: Regimes' increasing reliance on personalist claims to legitimacy is likely to signal a subsequent deterioration of judicial and legislative constraints on the executive.
As outlined previously, we expect this effect to be contingent on a country's existing level of democracy. Institutionalized liberal democracies should have an advantage in fending off personalist attacks on their institutional guardrails. Accordingly, Hypothesis 2 proposes: H2: Regimes' increasing reliance on personalist claims to legitimacy has a particularly detrimental effect on judicial and legislative constraints in regimes with lower levels of democracy.

Data and research design
This study utilizes cross-sectional time-series analysis to investigate how shifting claims to legitimacy affected democratic deterioration in 164 countries between 1900 and 2015. 65 In all models, country-years for which all variables are available are included, beginning with 1900 or a country's year of independence. To cover the longest timeframe possible, we employ parsimonious models that focus mainly on the variables of interest. 66 We utilize fixed-effects regression models, as these estimate within-unit effects and account for unobserved unit heterogeneity by absorbing the country-specific, timeconstant effects. Fixed-effects models are well documented and have several benefits compared to other modelling techniques 67 , most notably that they avoid distorted results due to omitted variable bias. We add year fixed effects to account for temporal changes that may affect all included countries in a similar manner (e.g. economic shocks, international disputes, or wars). Lastly, the addition of a set of controls tests for alternative explanations.

Dependent variableexecutive constraints
We use two dimensions to test our hypothesis that the increased reliance on personalist claims to legitimacy precedes a deterioration of executive oversight mechanisms: Judicial constraints and legislative constraints on the executive. Both are captured by V-Dems aggregate measures by the same name. The judicial constraints variable relates to the degree that the executive respects the constitution, its compliance with courts' decisions and higher and lower courts' independence from the executive. Legislative constraints combine four subcomponents which measure whether the legislative questions and controls executive decision-making and whether opposition parties can "exercise oversight and investigatory functions against the wishes of the governing party or coalition." 68 Thus, these measures capture both the potential power and actual performance of courts (judicial institutions) and parliaments (legislative institutions). If, for example, the executive uses its power to limit the courts' independence or simply does not comply with judicial decisions, the individual item-coding would downgrade this index, which we standardize to a mean of zero and a standard deviation of one. If parliament has the right to question the executive, but rarely executes its oversight function due to increased repression or co-optation, we will also see a decline of this variable. Arguably, the new, comprehensive V-Dem project provides more reliable and valid data on democratic quality than other datasets. 69 Our main analyses, thus, rely on V-Dem's regime coding. Additionally, we run robustness tests using Polity IV's "Executive Constraints" as an alternative dependent variable 70 to ensure that the results are robust across different specifications (see Appendix Table A4 and Figure A4).

Independent variablenew data on claims to legitimacy
Issues of legitimation are contested and notoriously difficult to analyse empirically. Directly collecting information on these government strategiesfor instance, through elite surveys or representative pollsis often impossible, particularly in autocratic contexts. 71 Andrews-Lee conducted survey experiments on the revival of charismatic rule in Argentina and Venezuela. 72 Unfortunately, such experiments are inconvenient for large-N cross-sectional time-series analyses. The same still holds for the content analysis of ruler speeches and government propaganda. 73 While their internal validity stands outthis study aims to produce generalizable findings spanning many countries across many decades and thus relies on expert-survey data as a second-best option. In doing so we follow other researchers who have used similar data to examine questions of legitimacy and legitimation. 74 For our measures of legitimation strategies, we use the legitimation items that were assessed in the most recent round of the V-Dem project. 75 Country experts rated the strength of governments' claims to legitimacy on a five-point scale (0-4) annually for 179 countries for the period from 1900 to 2020. These ratings were converted into an interval latent variable, which we standardize to a mean of zero and a standard deviation of one, to make the results comparable. On average, 4.5 expert coders rated the legitimation strategies for all country-years. 76 The expert question on personalist claims to legitimacy ("person of the leader") reads as follows: "To what extent is the Chief Executive portrayed as being endowed with extraordinary personal characteristics and/or leadership skills (e.g. as father or mother of the nation, exceptionally heroic, moral, pious, or wise, or any other extraordinary attribute valued by the society)?" 77 There is clearly a challenge in coding historical data by means of expert surveys. To mitigate this to the best possible extent, the V-Dem project works with a Bayesian item-response-theory model that estimates country-year point estimates based on the experts' coding. 78 This allows using patterns of coder agreement and disagreement to estimate variations in reliability and systematic differences in thresholds between ordinal-response categories. 79 We hypothesize that increasing claims based on the person of the ruler are first steps and thereby early warning signs for deteriorating democratic quality. Thus, we focus on this variablepersonalist legitimation strategiesin our interaction-analysis and in the discussion of the results.

Intervening variable / moderator
To test whether the hypothesized effect is detectable and significant in different regimes we include a moderating variable that captures different levels of (procedural) democracy. Using V-Dem's Electoral Democracy measure (Polyarchy), we rely on a measure that is devoid of our dependent variables' conceptsjudicial and legislative constraintsbut allows us to differentiate between regimes that allow universal, free and fair elections and respect freedom of speech and association on the one hand and those regimes that do not on the other. 80 We rely on this continuous measure for our main analyses, but oftentimes use a categorical distinction of regimes in the text to ease interpretation. 81

Control variables
Going by existing research, the strength of judicial and legislative constraints in a country might also be influenced by other factors, which we include as control variables. First, as outlined, governments regularly invoke different claims to legitimacy. Legitimation strategies other than personalist claims might also signal government intentions to either strengthen or weaken executive constraints. In contrast to personalist claims, rational-legal claims to legitimacy may have a positive precursor-effect by shifting the narrative towards impersonal procedures. We therefore add the V-Dem's rational-legal claims variable as a control. Additionally, we add the performance-based claims variable, but expect its effect to be ambiguous at bestas all regimes rely on performance-based claims to some degree. 82 Furthermore, the V-Dem project assesses whether regimes invoke ideological claims to legitimacy. However, these summarize multiple ideological orientationsconservatism, socialism, nationalism, among othersin a single variable and are, unfortunately, too unspecific to allow for a precise interpretation. 83 Nevertheless, we control for these ideological claims as well.
Second, regime maturity is captured by a "regime duration" variable, as well as by its quadratic and cubic terms. This helps to account for potential nonlinear effects over the life span of a given regime. 84 We do this by coding changes in V-Dem's "v2reginfo" variable 85 as the beginning of a new regime and then counting the number of years until its demise.
Third, we know from Przeworski et al. 86 that economic performance makes democracies more durable, while Tanneberg, Stefes, and Merkel 87 have shown that economic hardship reduces autocratic-regime longevity, thus, potentially influencing our dependent variables as well. We account for this by controlling for GDP growth. Countries' overall prosperity is measured by adding the log of their gross domestic product to the model. Besides negative GDP growthwhich serve as an indicator for economic crises we also control for political crises and natural disasters. All of them create ample opportunities for personalistic rulers to amass power in order to swiftly "solve" a country's predicament. 88 The former is captured by dummies for periods of democratization and democratic breakdown that regularly make the judicial and legislative system more fragile. We control for natural disasters by coding all country-years as 1, where any type of natural disaster occurred as identified by Ritchie and Roser 89 based on the EMDAT International Disaster Database. 90 Fourth, in the final model, year fixed effects are included and to control for any temporal effects that influenced all countries in a similar matter (major wars, economic crises, natural disasters).

Empirical results
The following section summarizes the results of the multivariate models and relates these findings to exemplary cases, which help to understand the underlying dynamics. The interaction analyses add important nuance and show how institutional limits can curb "would-be autocratizers" ambitions. Table 1 summarizes the results of the fixed-effects models for the continuous dependent variables "judicial constraints on the executive" and "legislative constraints on the executive." Model 1 contains only the leadership-based claims to legitimacy, polyarchy, as well as their interaction; Model 2 adds the alternative claims to legitimacy and economic prosperity. Model 3 additionally accounts for GDP growth, for periods of democratization or democratic breakdown and natural disasters, while model 4 adds year fixed effects and the linear, quadratic, and cubic terms for regime maturity. All time-varying independent variables are lagged by one year to account for the temporal ordering that we expect (see Appendix Table A5 for different lags of up to 5 years as robustness checks). Models 5-8 repeat the same analysis for the second dependent variable, legislative constraints.
The results clearly support the hypotheses. For judicial constraints, a one-standard deviation increase in leadership-based claims from one year to the next is related to a reduction of executive constraints by about 0.12 pointsor one eighth of a standard deviation of the dependent variable. As outlined previously, both variables are scaled to a mean of zero and a standard deviation of one. As rulers may formulate their legitimation strategies independently of constitutional reality, claims can be altered more flexibly and to a greater magnitude than institutional structure. Accordingly, even a seemingly small reduction of the dependent variable "judicial constraints" is an important warning sign and can point to gradual changes in this area. By controlling for regime transition periods, we account for times when institutions change rapidly. This finding lends support to H1. That is, changes towards legitimation strategies focused on the person of the ruler signal an erosion of the regime's judicial constraints in the following year, thereby providing the ruler more leeway. 91 By including the interaction with V-Dem's electoral democracy measure, we test whether this observed effect is contingent on the baseline level of electoral democracy present in a given country and period. Figure 3 visualizes the marginal effect. 92 We conclude that only in the most democratic countries judicial oversight mechanisms on averagewithstand such a personalist attack on its judicial oversight mechanisms. This includes almost all long-standing democracies, which lie more than one standard deviation above the long-term global average of V-Dem's Polyarchy measure, such as most EU member states, or other prominent democracies like Table 1. Fixed-effects models for judicial constraints (1)(2)(3)(4) and legislative constraints (5)(6)(7)(8).  Cluster-robust standard errors in parentheses; Significance levels: * 0.05, ** 0.01, *** 0.001; All time-varying variables lagged by one year.

DEMOCRATIZATION
Costa Rica, post-2000 Chile, or Uruguay. These findings of our interaction analysis lend support to H2. To put these findings into perspective, President Alberto Fujimori's self-coup in Peru, at that time an electoral democracy, serves as a drastic example. Fujimori's claims to legitimacy prior to and after his first election in June 1990 stressed his personal appeal as an outsider to the political system who fought the traditional political elite in Peru and the "establishment institutions they controlled." 93 By doing so, he portrayed himself as the saviour of the people who could not accept any compromises (Personalist Claims Δ 1990-1993 = 0.11; all reported changes are on a scale of 0-1). This set the path for demolishing executive constraints in the 1992 autogolpe. Fujimori dissolved parliament, suspended the Constitution, and replaced High Court judges with loyal supporters (Judicial Constraints Δ 1990-1993 = −0.37), and turned Peru's polity into a highly centralized competitive authoritarian regime until he finally fled the country at the end of 2000. 94 In a similar vein, in the USA Donald Trump stressed his irreplaceability as a leader and personal appeal (Personalist Claims Δ 2016-2017 = 0.32) as the savior of the ordinary people who would make America "great again." He subsequently filled open Supreme Court positions with judges that were said to favour his conservative political agenda and recurrently attacked opponents and institutions of executive control. However, the country's courts largely withstood Trump's attacks, andin contrast to V-Dem's liberal democracy variablethe judicial constraints on the executive variable only slightly weakened during his term (Judicial Constraints Δ 2016-2019 = −0.015). The results for the second dependent variablelegislative constraintsalign with the results of the first analysis (Table 1, model 5-8). That is, we also find a negative effect of personalist claims to legitimacy on executive constraintsspeaking in favour of H1. The results are significant at the 99% confidence level and slightly exceed those of the judicial constraints in magnitude. The interaction with polyarchy ( Figure 4) reveals that this effect is also not significant in regimes that score approximately one standard deviation and above in the global average of electoral democracy levels. However, in this case we find a highly significant interaction, meaning that country-years with above average levels of V-Dem's Polyarchy measure experience a significantly weaker decrease of their legislative constraints following an increase in personalist rhetoric.
For example, in Argentinaclassified as an electoral democracy for the last 40 years after Néstor Kirchner was first elected President in 2003, there was a huge increase in personalist claims (Personalist Claims Δ 2002-2003. The analysis, using V-Dem data, shows that in the following years legislative constraints of executive power slightly weakened (Legislative Constraints Δ 2003-2008 = −0.06). After his wife Cristina Fernández de Kirchner was elected president at the end of 2007, personalist legitimation strategies remained high and legislative constraints reduced further. In a country with significantly lower Polyarchy levels, we see a much stronger effect. In Russia, classified as an electoral autocracy since 1993, under President Putin we see increasing personalist claims in 2000, 2004(Personalist Claims Δ 2000) that were all followed by reductions of legislative constraints (Legislative Constraints Δ 2000-2020 = −0.51). These examples demonstrate how leaders with personalist aspirations have a much harder time removing their countries' legislative constraints once a certain level of democratic quality has been surpassed. This might be due to the higher visibility of legislative actions compared to the judicial and the number of actors and parties involved in the parliamentary branch. Attempting to alter legislative constraints in countries with democracy levels above this threshold, thus, seems to be less successful.

Control variables and robustness checks
As for the controls, we find that legitimation strategies that increasingly focus on rational-legal claims work in the opposite direction of personalist claims, meaning they are antecedents to strengthened institutional checks and balances and thereby bolster democratic health. Performance based claims only show a significant influence in the full model on legislative constraints (Model 8). That is, governments that shift their discursive focus towards performance to legitimate their rule, may also attack legislative constraints. As mentioned, ideological claims are difficult to interpret, but generally also favour subsequent executive aggrandizement. For the remaining control variables, the only notable influence on judicial and legislative constraints is the economic performance of a country, whichin the case of judicial constraintsis negative in Models 2 and 3 and turns sign, once we control for regime maturity and add year fixed-effects. During political transition periods, executive constraints behave as expected: They increase following democratization events and decrease during democratic breakdowns. Disasters show no significant influence on our dependent variables.
We run three additional models to probe the robustness of our findings. First, using Polity IV's Executive Constraints variable 95 , which combines judicial and legislative guardrails, we are able to reproduce the findings of the original models (see Appendix  Table A4). Second, we worked with time lags other than one year. All results hold for two-, three-and five-year lags (see Appendix Table A5). Finally, we find that the effects are not driven by a specific time period. Figure A5 and Table A6 in the Appendix summarize the findings.

Conclusion
This article has provided first insights into the consequences of shifting regime legitimation strategies for the strength of judicial and legislative constraints of executive actionsan aspect that existing research could hardly comprehensively examine so far. Our findings clearly demonstrate that changing governmental legitimation strategies that increasingly focus on the person of the leader are important signals for the following expansion of executive powers.
With new comprehensive data from the V-Dem project we have for the first time been able to systematically evaluate the political repercussions of shifts in legitimation. However, there are some limitations to this initial assessment. We have not directly tested how legitimation strategies are utilized as part of the "menu of manipulation," 96 but rather whether or not changing claims signal a subsequent deterioration of judicial and legislative constraints on executive action. The potential causal pathways that follow the increased reliance on leadership-based claims appear as diverse as the leaders that initiate them. Some denounce oppositional and civil society groups as terrorists or foreign agents and declare them illegal. They may also cut down the oversight powers of parliament. Others might follow a more long-term agenda and replace judges after their term-limit is reached with their supporters. The specificities of this rhetoric-action link warrant further analysis in future qualitative and quantitative work on the topic.
However, our initial macroanalysis already adds important nuance. Most notably, the institutional repercussions of an increased discursive focus on the grandeur of the ruler are particularly grave in regimes that are not fully democratic. In contrast to much current scholarly concern, long-standing, liberal democracies with their stronger institutions seem comparatively safe from these executive aggrandizements, at least in the short term. 97 They are able to resist rulers' potential attacks on their judicial and legislative constraints.
Our analysis adds to the rediscovery of questions of political legitimacy and the growing literature on (autocratic) legitimation strategies that has emerged in response to current autocratization processes. 98 It demonstrates that investigating official legitimation narratives enriches comparative politics and comparative authoritarianism approaches that have long had an institutional bent to them.
The key theoretical implication is that personalist discursive strategies form an intricate part of, and facilitate, executive action seeking to remove institutional constraints. Systematically analysing shifting claims to legitimacy and their acceptance should therefore be central to the current scholarly drive to track processes of autocratization.
Notes 20. Dukalskis and Gerschewski, "What Autocracies Say"; Kailitz and Stockemer, "Regime Legitimation, Elite Cohesion." 21. Koschorke,Fact and Fiction,. Goldstein and Keohane, Ideas and Foreign Policy. 23. Weyland, "Autocratic Diffusion and Cooperation," 1240. 24. LeBas, From Protest to Parties. Naturally, this depends on whether citizens actually believe these claims. As Scott reminds us in Domination and the Arts of Resistance., citizens do not necessarily "buy" those narratives but may use both overt and covert forms of resistance to official claims to power. 25. Alagappa, Political Legitimacy in Southeast Asia, 4. 26. Lipset, "Some Social Requisites of Democracy"; Beetham, The Legitimation of Power; Levi, Sacks, and Tyler, "Conceptualizing Legitimacy, Measuring Legitimating Beliefs." 27. Weber,Economy and Society,215. 30