Countering autocratization: a roadmap for democratic defence

Many democracies around the world face the challenge of democratic recession and autocratization as democratically elected incumbents increasingly show autocratic tendencies. Existing research has mainly focused on the circumstances under which these autocratizing incumbents erode democracy and on the structural factors explaining the resilience of democratic institutions. Much less is known about the actors within those institutions and when they stand up against the autocratizing incumbent to defend democracy. In this article, we present a novel theoretical framework of democratic defence that focuses on the interaction between the incumbent, institutional elites, and citizens. Developing a two-level model of democratic defence, we show how the democratic defender ’ s personal interests, repression by the incumbent, the perceived ambiguity of the autocratic action, and the perceived credibility of the democratic defender interact to a ﬀ ect the occurrence of democratic defence. The resulting framework can guide future research on the role of speci ﬁ c actors in defending democracy. We demonstrate the utility of our framework with illustrative case studies of (attempted) democratic defence in Senegal (2011 – 2012) and Poland (2017 – 2018). An actor-based approach of democratic defence is crucial to understand what actions domestic and international actors can take to prevent (further) democratic recession.


Introduction
Democracy around the world is under threat.Increasingly, democratically elected incumbents seek to undermine democracy, and increasingly citizens appear to be willing to support them in their actions. 1Autocratizing incumbents pack courts, subvert checks and balances, and aggrandize their executive power. 2 They show great creativity in undermining democracy and in justifying their actions.Yet, when democracies are under threat, pro-democratic actors can stand up to defend democracy.Existing research has mainly focussed on the specific actions of autocratizing incumbents and the circumstances under which they succeed, but much less is known about unsuccessful attempts at autocratization and the role of democratic defenders.In particular, we lack a comprehensive theoretical framework to account for the occurrence of democratic defence.
Democratic defence should be an important part of the autocratization literature. 3n the past, domestic attempts to undermine democracy predominantly came from extremist parties that had no real shot at coming into power through democratic means.Capoccia studies how pro-democracy incumbents could use formal provisions and informal political strategies to limit the extremists' chances even further. 4By proclaiming and acting on support for democracy, many incumbents tried to make democracy "the only game in town". 5Since the turn of the century, however, democratically elected incumbents increasingly seek to undermine democracy from within.Democratic recessionthe substantial decline of core requirements of democracy 6 is especially worrisome as it does not only affect hybrid or emerging democracies, but also consolidated democracies. 7As incumbents in many democracies around the world no longer default to defending democracies, the very institutions designed to protect democracy become the target of incremental changes and attacks.We focus our framework on those puzzling cases that were (at least) electoral democracies when democratic recession started, and in which democratic recession was led by incumbents who themselves were elected in free and fair elections.
We employ an actor-centred approach to democratic recession to contribute to this literature and conceive of autocratization and democratic defence as a strategic interaction between elites and citizens.Autocratizing incumbents implement autocratic actions that erode aspects of democracy.Democratic defenders, in turn, aim to uphold the democratic status quo.We argue democratic defence can take place on two levels: by elite democratic defenders and by citizens.This article develops a model of democratic defence that distinguishes between these two levels and conceptualizes the trade-offs democratic defenders face on each level.On the first level of democratic defence, elite democratic defenders face trade-offs between their personal interest and repression by the incumbent.On the second level, citizens face trade-offs between elite democratic defenders' credibility and the autocratic action's ambiguity.We illustrate the value of our framework with an empirical application to the cases of Poland and Senegal.
This article is structured as follows: in the next section we provide a brief review of the state of the art of research on democratic recession and resilience, highlighting why research on democratic defenders is urgently needed.The subsequent section outlines the two-level model of elite and citizen democratic defence.Then, we apply this framework to the cases of Poland and Senegal, providing an empirical illustration of the model.The conclusion provides three avenues for further research on democratic defence.

Beyond democratic resilience: democratic defence
There is ample research on the structural factors that increase the likelihood of autocratization.These explanations focus on institutional, economic, social, and international factors. 8In terms of institutional factors, for example, it is argued that in more majoritarian or presidential systems elites and citizens are more used to powerconcentration in executive offices.This facilitates autocratization when an incumbent with autocratic ambitions is elected. 9In terms of economic factors, research has demonstrated that the weaker a state's economy is (for example higher inequality or lower GDP), the more vulnerable it is to autocratization. 10Economic decline and high economic inequality are associated with lower trust in politicians and political institutions, 11 which can turn into crises allowing more concentration of executive power. 12In terms of social factors, polarization has been found to be a key condition making autocratization more likely.As polarization creates growing distrust between social groups, leaders are incentivized to ignore democratic norms, ensuring more benefits for their in-group. 13Finally, in terms of international factors research on autocratization has found that international normative preferences for democracy often help states remain democratic or democratize: the EU is one of the prime examples. 14n the other hand, linkages to autocratic regimes, specifically China and Russia, increase the likelihood of autocratization. 15owever, despite the extensive body of research on the conditions that facilitate autocratization, less is known about when and why autocratization fails.To better understand cases where autocratization was attempted but failed ("near-misses" 16 ), scholars have turned their focus to explaining democratic resilience.The resilience of democraciesdefined as the ability of a democratic system to prevent and react to anti-democratic challenges and attacks 17 depends on three main pillars: the strength of democratic institutions, the presence of accountability mechanisms, and citizen values.Horizontal accountability mechanisms prevent backsliding when one actor's power limits the power of others: they cannot singlehandedly decide to turn autocratic. 18The judiciary "as the last bulwark against autocracy" 19 is one example.Vertical accountability via elections plays an important role as well, giving oppositions the opportunity to democratically oust an autocratizing incumbent. 20Lastly, a vibrant civil society increases democratic resilience by enabling diagonal accountability. 21hen democracy is truly consolidated and the only game in town, the mere presence of strong accountability mechanisms and institutions deters and prevents an incumbent from challenging democracy.
However, it has become clearer that even when democracy appears to be consolidated, autocratizing incumbents find cracks in the system to further their autocratic agenda.Venezuela, Brazil, Turkey, Hungary, and the USA all show how more-or-less consolidated democracies can still be challenged. 22This could be explained by assuming that the democratic institutions in these cases were not sufficiently resilient: democracy was merely a mirage. 23But this assumption appears to be too strong when more consolidated democracies are also under threat.We must contend with the possibility that even strong institutions can be successfully targeted by incumbents, precisely because they have a legitimate and democratic mandate to adapt institutional rules and procedures.Incumbents can leverage their mandate to attack democracyovertly or covertly.In response to incumbent-led democratic recession, we must therefore not only look at institutional resilience, but also consider the actors involved in democratic defence.
In other words, actors and their actions matter for democratic resilience: popular support for democracy correlates with the resilience of that democracy. 24Citizens play an important role in checking and balancing the regime. 25Support for democratic checks and balances, however, is dependent on factors like economic prosperity, partisanship, polarization, and ethnic divisions. 26As such, under some circumstances, preferences other than democracy are more salient, making citizen support for democracy a necessary but likely not sufficient condition for democratic resilience.
Between incumbent, institutions, and citizens, however, there are elite actors who can also prevent or revert autocratization. 27There is only limited research on the opposition strategies against incumbent-led autocratization. 28We aim to expand this literature by asking under which circumstances institutional elites and citizens oppose incumbents and engage in democratic defence: "all activities, be they formal provisions or political strategies, that are explicitly and directly undertaken to protect the democratic system from the threat of its internal opponents". 29When incumbents are the main autocratizing actoras is often the case in the so called "third wave" since the 1990s 30 -, defenders' access to formal provisions to counter anti-democratic actors (such as bans on anti-democratic parties or anti-democratic speech) are limited.Rather, the formal provisions that check and balance incumbents' actions are precisely incumbents' primary targets of autocratization.The defenders' political strategiestheir interaction with autocratizing incumbents and how they respond to autocratic actionbecome all the more important.Therefore, democratic resilience is the outcome of successful democratic defence.
In this article, we present a theoretical framework that models this interaction between autocratizing incumbents and elite and citizen democratic defenders, to explain when democratic defence occurs.

A two-level model of democratic defence
We conceptualize democratic defence on two levels.The first level of democratic defence are elites operating within the democratic institutions.The second level of democratic defence are the citizens.In most cases of incumbent-led democratic recession, both levels will be needed: elites on their own do not have sufficient momentum to persuade a committed autocratizing incumbent to stop; and citizens on their own face an informational disadvantage and classic coordination problems.Elites can help resolve the citizens' informational disadvantage, while citizens rallying in defence of democracy provide elites with more momentum to counteract the autocratizing incumbent.
There are two exceptions to the need for both levels of democratic defence.First, when the autocratic effect of actions is unintended, elite democratic defence on its own can be successful.Democratic defence, in these cases, simply works because the autocratic effect of the proposal was neither intended nor seriously considered up until that point.Indeed, democracies often have checks and balances built in for this very purpose: to critically examine policies and legislation and adjudicate whether they adhere to democratic norms.We disregard such "unintended autocratization" in the remainder of this article, as it is an important part of democratic resilience, but cannot be understood as democratic defence in response to incumbent autocratization attempts.
Second, when the autocratic effect of an action is so blatantly clear to citizens that they face no informational disadvantage, citizen defence on its own can be successful.The 2023 protests in Israel against Netanyahu's judicial overhauls are an example of large popular mobilization without strong central organization. 31Elite opposition actors voiced concerns about the judicial overhauls, but citizens understood the unambiguous autocratic effects of the proposal without further elite cueing. 32As we argue, however, incumbent-led autocratization is often characterized by high levels of ambiguity.This makes it less clear for citizens when an action is, in fact, an autocratic action, and increases the need for elite cueing.
We argue when incumbent-led autocratization is intentional and at least somewhat ambiguous, both levelselites and citizensare necessary for democratic defence.Here, we explore the core considerations for these two types of actors when deciding whether to engage in democratic defence.We focus on four elements: elite self interest in preventing the autocratic action, perceived threat of repression of elite opposition by the autocratizing incumbent, elite credibility as a rallying point for citizens, and ambiguity of the autocratic action.For the first level of democratic defence, we argue that elites weigh their self-interest in preventing the autocratic action from being implemented against the perceived threat of repression by the incumbent.For the second level of democratic defence, we argue that citizens weigh the elite defender's credibility against the autocratic action's ambiguity.
Our starting point is when an incumbent proposes or implements an autocratic action.The model, however, does not aim to explain the incumbent's motives for engaging in autocratic actions.Research on autocratization has been rather lopsided: while the actions of autocratizing incumbents, and their choices to engage in repression versus co-optation during autocratization have been amply studied, 33 research on democratic defenders attempting to counter autocratization is largely absent.We therefore limit our focus on the trade-offs that go into the potential democratic defenders' considerations to engage in democratic defence.

Level 1: elite democratic defence
The first level of democratic defence is elite action within institutions: they stand up against the autocratizing incumbent, refuse to cooperate, speak out against the undemocratic nature of the action, or actively counteract the action.We understand institutional elites to be a broad category, encompassing the actors who work in the public institutions that are essential for the functioning of democracy.This includes, for example, oppositions in parliament, judges within the judiciary, or bureaucrats within ministries; but excludes civil society leaders and private media organizations.Their defining characteristic is that they have the authority, stature, and credibility to claim that a specific action has adverse consequences for democracy.By extension, elites do not have the informational disadvantage citizens have when it comes to recognizing autocratic actions: judges know the law and constitution, bureaucrats know how administrative processes are supposed to run, and legislators know when parliamentary procedures are violated.As such, the primary problem for elites is not assessing how blatantly autocratic an action is.Rather, it is about how much they are willing to risk in an attempt to stop the autocratizing incumbent.
Therefore, the dominant aspects elites weigh when they decide to defend democracy are the threat of repression by the incumbent and the elite's self-interest in maintaining the democratic status quo.While repression by the incumbent can disincentivize elite democratic defence, elite's self-interest can incentivize it.

Incumbent repression
Research shows that incumbents employ repression to intimidate the opposition to further their anti-democratic agenda. 34Repression can come in the form of physical sanctions against oppositions by the incumbent, like harassment or intimidation by instigating unwarranted and undue criminal proceedings. 35Incumbent repression is relevant for understanding democratic defence, as repression can stop the opposition in its tracks.Moreover, it can have the compounding effect that other (potential) opposition members fear a similar fate and start to self-censor. 36Repression can also occur in the form of a credible threat against livelihood, job security, or family-members.Such more opaque repression is especially salient in democracies, where autocratizing incumbents attempt to maintain a veneer of democratic credibility, and where outright, violent repression is more difficult.Examples of more clear-cut repression include the firing of critical journalists from public media in Turkey, 37 illegal surveillance by security services of the media in Serbia, 38 or charging academics and journalists with trumped up charges of sedition in India. 39An example of more opaque repression is the decision by the Fidesz government in Hungary to relax the termination policies for bureaucrats, which then created opportunities to fire state employees when they no longer appeared sufficiently loyal. 40This severely restricts the space for opposition, and effectively represses state employees, even though the new termination policies have not been put into practice yet.
Credible threats can also be issued to elites or organizations that are aligned with potential defenders.Hence, even when a (potential) democratic defender is not directly targeted, they can see which personal costs they might bear if they continue opposing the autocratizing incumbent.If repression is successful, the incumbent does not need to formally exclude voices from the debate as long as they get muffled sufficiently.

Elite self-Interest
Given the high risk of resisting the autocratizing incumbent, elites assess their selfinterest in opposing the incumbent and weigh it against the likelihood and strength of incumbent repression.If the autocratic action of the incumbent threatens to directly affect the ability of (members of) the elite to perform their professional duties or even affect their livelihood, professional and material self-interest can incentivize elites to withstand the incumbent and defend democracy.For instance, when incumbents threaten to fire a large proportion of judges, judicial elites are more incentivized to stand up against this threat. 41hen repression is comparatively low, elites can afford to try to safeguard their selfinterest by defending democracy.When repression is comparatively high, the cost of standing up against the autocratizing incumbent may become too high for elites, making it less likely they will defend democracy.However, when the incentive to defend their self-interest is higher, elites are more likely to bear the personal costs associated with defending democracy.As such, elites adjudicate between the risk of incumbent repression and the extent of their personal and professional self-interest in defending democracy, as they self-select into democratic defence.
Whether it is in the personal interest of elites to defend democracy also depends on the scope of the autocratic measures already in place.When democracy is still fully intact, elites have stronger incentives to resist democratic recession.But when democratic safeguards have been removed, it can be in the self-interest of elites to be coopted by the autocratic incumbent and facilitate their power grab.As institutions are increasingly swaddled by autocratization, elites' self-interests may shift.Rather than defending democracy as a means to defend ones' material and professional interests, elites may be better served by switching sides and supporting the autocratizing incumbentparticularly when the threat of repression is high.When elites are coopted in this way, autocratization has progressed to such an extent that democratic defence is increasingly less likely.This suggests that the progression of autocratization decreases the window of opportunity for democratic defence.For instance, the Polish Constitutional Tribunal was captured by PiS loyalists after 2016, turning them from an activist court into a government enabler. 42n short, depending on the course of autocratization, either democratic defence or co-optation serves the self-interest of elites.In both cases, the threat of repression reduces the likelihood for democratic defence.

Level 2: citizens' democratic defence
The second level of democratic defence concerns citizens mobilizing in defence of democracy.Citizen mobilization only counts as democratic defence when there is an explicit connection to democracy.Citizens must not only stand up to defend a material public good, but also direct their attention to a democratic common good.This does not preclude citizens from defending democracy by primarily protesting against a certain policy issue, such as bills limiting citizenship (e.g. in India, 2019) or rolling back abortion-rights (e.g. in the USA, 2022), but requires that citizens at least understand their protest in a broader, democratic context. 43If not, they cannot be said to defend democracy.While high, this definitional bar is not too high: in India and the USA, protesters and commentators explicitly connect the topics to extant democratic rights and freedoms.Citizens acknowledge both roles: a more direct effect on people's daily lives and a broader effect on democracy, freedom, or the constitution.
When the autocratic nature of the incumbent's action is unambiguous, citizens can take to the streets without relying on elite cues, as the Israeli example shows.When there is no "clear evidence" 44 that an action is anti-democratic, the autocratic nature of certain actions can be ambiguous to citizens: elites become instrumental in mobilizing popular opposition to the autocratizing incumbent.Then the core trade-off for citizens to mobilize is between the autocratic action's ambiguity and the elite defender's credibility.Is the elite defender genuinely concerned about the fate of democracy, or is their supposed democratic defence merely a self-interested, political ploy?Ambiguity of autocratic action Before citizens join an elite defence, they must first determine if the incumbent's action is autocratic in nature.While some autocratic actions have a clear autocratic intention and effect, the intention and effect of other actions can be more ambiguous.This "vexing ambiguity" 45 contains two components: the action's blatancy and its justification.
First, autocratizing actions vary in the degree to which they are clearly and obviously autocratic.Some actions blatantly harm democracy: outright abolishing elections or extending presidential term-limits. 46Yet, other actions are more ambiguous.Whether requiring voter identification is harmful to democracy (because it disenfranchizes some voter groups disproportionally) or beneficial (because it limits some voter fraud), is context-dependent and debated. 47econd, incumbents can also cloak the autocratizing effects of their actions by justifying their actions.Blatant and unjustified autocratization has a negative effect on the incumbent's popularity. 48Therefore autocratizing incumbents justify their autocratic actions by arguing why they are necessary or desirable.In many cases, the incumbent gives a seemingly reasonable justification for their action.In Hungary, autocratic action was justified because it combats corruption. 49In Rwanda, the Republic of Congo, and Colombia the justification was populist in nature: the incumbent called for a referendum to circumvent presidential term limits. 50And in Bulgaria the banning of certain opposition parties was justified with reference to a contested and arguably misinterpreted constitutional phrase. 51In other cases, limitations on democracy might not only seem reasonable or legitimate, but actually be perceived as necessary by citizens to curb a crisis (for example in response to war, natural disasters, economic crisis, or a public health crisis such as COVID19).Crises can therefore be an important window of opportunity for autocratizing incumbents to argue that autocratic actions are necessary and legitimate, increasing the autocratic action's ambiguity. 52ecause autocratic actions are not always blatant and often justified by incumbents, it can be harder for citizens to identify incumbent actions as unambiguously autocratic.Research has shown that partisanship or polarization, policy preferences, and contextual factors cloud citizens' judgments of whether actions are democratic or autocratic. 53Incumbents can exploit the ambiguity of an autocratic action and appeal to fellow partisans to limit opposition against it.If it is sufficiently unclear that a certain action is indeed autocratic, citizens are unlikely to support elite democratic defenders.

Credibility of elite democratic defenders
If ambiguity surrounds the autocratic action, elite cueing is likely necessary for citizens to become democratic defenders.When citizens consider joining elite democratic defenders, they must believe that the elites are credibly committed to democracy.That is: that the defence is not merely a political manoeuvre to gain some (electoral) leverage over the incumbent, but a genuine democratic defence.We argue that a core part of elite credibility is based on the degree to which citizens perceive the institutions themselves to be important for democracy, and therefore likely depends on citizens' specific support for and existing institutional trust in these democratic institutions. 54The higher the trust in and support for the institution, the more likely it is that elite's democratic defence is deemed credible by citizens.
Yet, elite democratic defenders may not be only motivated by democratic concerns or commitment to democracy.As noted, it can be in the professional and personal selfinterest of elites to mount a defence of democracy.From the citizens' perspective, this obscures to what extent the elites are actually defending democracy or acting purely out of self-interest.For instance, a judge who opposes suspending the president of a court (as happened in the Maldives and Hungary) might publicly state their aim is to preserve democracy and the rule of law.But they risk being perceived as just trying to save their own jobs, or as having their own political agenda as "cronies" 55 of the opposition.In more strongly polarized contexts, it likely will be harder for democratic defenders to appear not self-interested, especially to non-partisans or anti-partisans.Hence, perceptions by citizens of elite defenders' self-interest can undermine elites' credibility.
In addition, autocratizing incumbents can actively undermine the credibility of elite defenders by framing democratic defenders as treasonous, fake, criminal, or corrupt, while emphasizing their own democratic mandate.Delegitimation of opponents will likely have an even larger effect in strongly polarized circumstances, especially for the incumbent's supporters.In their quest to mobilize citizens, elite democratic defenders therefore invariably fight an uphill battle.They must assert their own credibility against the autocratizing incumbent's democratic mandate and attempts to delegitimize them.
We argue that citizens' adjudication between the incumbent's action's ambiguity and the elite democratic defender's credibility explains when citizens join the elite democratic defence.When the autocratic nature of the action is comparatively more ambiguous, it is likely that the public does not engage in democratic defence.When the elite democratic defenders are comparatively more credible, however, it is likely that citizens join the elite's democratic defenceraising the cost for the autocratizing incumbent to persist.

A roadmap for democratic defence
Our theoretical framework stipulates that democratic defence can occur on two levels: elite democratic defence and citizen democratic defence.Figure 1 summarizes the theoretical model and visualizes the trade-offs that institutional elites face in Level 1 and citizens face in Level 2. The starting point of our framework is an autocratic action initiated by the incumbent.The solid arrows (pathway 1) show the mostlikely occurrence of democratic defence in response to incumbent-led autocratization in democracies.The dotted arrows (pathway 2) show citizen can defend democracy without elite cues. 56On the first level of democratic defence, we show the trade-off for institutional elites between repression and self-interest.On the second level we show the trade-off for citizens between credibility and ambiguity.In the next section we illustrate the utility of our two-level framework of democratic defence with the cases of Senegal and Poland.

Two illustrations: democratic defence in Senegal and Poland
To demonstrate the utility of our two-level framework of democratic defence, we now turn to the illustrative cases of Poland and Senegal.They provide two diverse cases to illustrate a range of variation on the autocratizing incumbent and their repression, elites' self-interest, and citizens' willingness to mobilize.Both Poland (before 2017) and Senegal have been described as relatively well-functioning electoral democracies in Europe and Africa respectively, prior to (attempted) autocratization.In Poland, DEMOCRATIZATION autocratization started when a new executive came to power, while in Senegal, the sitting incumbent tried to extend his hold on power.In Poland, the autocratic action targeted the judiciary, while in Senegal, it targeted constraints on the executive.In line with Level 1 of our model, the judiciary in Poland spoke out against the incumbent, while in Senegal political oppositions tried to defend democracy.Lastly, in Senegal civil society is "the cornerstone of democratic development and a key indicator of the political climate" 57 (providing a case where citizen democratic defence is comparatively likely), while in Poland civil society has been described as relatively small scale and since 2015 increasingly limited (a case where citizen democratic defence is less likely). 58Combined, these two illustrations are a useful probability probe and illustration of our modelalthough we welcome further empirical testing.
The following illustrative case studies below build on extensive analyses of research on Poland and Senegal, as well as on primary source materials such as newspaper articles, NGO reports, and statements by various (judicial) institutions.While not a thick description of autocratization in either 59 all quotes that we rely on in our analysis are triangulated over different and independent sources.
We divide autocratization in Poland into two empirical units.In the first empirical unit, covering March to July 2017, we start by exploring how low repression and high self-interest incentivize judges to speak out about the undemocratic nature of judicial reforms suggested by the governing Law and Justice party (Prawo i Sprawiedliwość, PiS).Then we show how, despite this high self-interest, citizens perceived the judicial system as credibly committed to democracy.In addition, we show that PiS aimed to increase the ambiguity around the reforms, making an effort to show they were justified and not autocratic.In the second empirical unit, covering August to December 2017, we show that elite level democratic defence did not change.However, PiS shifted gears and not only insisted on the necessity of the reforms, but also tarnished the credibility of the judiciary.In the end, this meant that the trade-off for citizens favoured the autocratic action's ambiguity over the elites' credibility, resulting in a fading of democratic defence on the second level, capture of the Constitutional Court, and subsequent (further) autocratization.
Similarly, we divide autocratization in Senegal into two empirical units.First, covering the beginning of 2011, we show how president Wade proposed to lower the threshold for a first-round win in presidential elections, to ensure his tenure.This action proved sufficiently and blatantly unambiguous autocratic that citizens rallied in defence of democracy, even though the opposition coalition had limited credibility.This induced Wade to back down and retract his proposals.Lastly, covering July 2011 to March 2012, we show Wade introducing again a blatant autocratic action, trying to overstep presidential term limits.This time, however, ambiguity was increased when the Senegalese Constitutional Court greenlighted Wade's bid.This slight increase in ambiguity proved enough to overcome the mostly uncredible opposition, resulting in Wade being able to run for a third term.
"Inglorious constitutional revolutions": judicial repression in Poland (2017-2018)   When high self-interest trumps opaque repression Initially heralded as a great example of successful democratic transition and consolidation, by 2018, Poland was the first country where the European Union used its Article-7 sanctioning mechanism over the degradation of the rule of law.In 2017, right before the Article-7 sanctions were initiated, PiS attempted to capture the entire judicial system with three new bills.These bills would (1) potentially dismiss all judges of the Supreme Court; ( 2) dismiss all members of the National Judicial Council (Krajowa Rada Sądownictwa, KRS), the agency which exists to ensure judicial independence and nominate judges for vacancies; and (3) create a new disciplinary chamber with virtually unchecked mandate.If the bills were signed into law, PiS would be able to appoint new members for the KRS and the new disciplinary chamber, as well as all Supreme Court judges and up to 10% of all lower court judges. 60n the first level of democratic defence, judicial elites spoke out in reaction to the proposals.They did so, first, with formal letters according to the procedures (that is: following the way checks and balances were set up and meant to be used).When that failed, they voiced their opposition publicly.From these statements, it is clear they are not only worried about the harm to democracy, but also about the acute risk they face personally: the bills would "subject judges to politicians", 61 "deprive" 62 the judiciary of powers and responsibilities, and "foresee the end of the term for all judges" 63 currently on the KRS.Repression, on the other hand, is present but more opaque and indirect than their acute personal interest.Amnesty International, for example, describes a case where a critical judge opposed a PiS-loyal judge.The latter was subsequently promoted to president of the regional court, overturning verdicts of the former.In April 2017, the critical judge was indicted (by the PiS-loyal now-president) in a disciplinary proceeding and even in criminal proceedings. 64n addition, the purpose of the three bills obviously increased the potential for future repression.
When high ambiguity is trumped by even higher credibility Despite judicial elites' opposition, PiS remained steadfast in their proposals, and justified them by claiming that they would improve the efficiency of a backlogged judicial system, curb corruption, and end the communist heritage of the Polish judiciary. 65he combined technicality of the bills and extensive justification created large ambiguity for citizens.
Nonetheless, citizens rallied to defend democracy.In the weeks following the bills' submission to the Sejm on July 12th, thousands of people gathered in front of courthouses in more than fifty cities around the country.The protestors chanted "Freedom, Equality, Democracy!" 66 and are quoted saying: "[PiS] is about to finish democracy", 67 and "[this proposal is] against the constitution and democracy". 68Former Polish prime minister Lech Wałęsa also called on the public to "defend democracy". 69The judges' statements that the proposals harm democracy, strengthened by other domestic and international actors, proved credible enough to overcome the ambiguity introduced by PiS.
Initially, the alignment of elite and citizen democratic defence seemed to provide sufficient momentum to make the PiS-autocrats concede.While the Sejm approved the bills, president Duda vetoed two of the bills on July 24th.The third bill, which would remove 10% of all lower court judges, was signed and came into effect at once -"waiting for PiS-friendly KRS members to name their replacements". 70In other words: the combined persuasive force of opposing elites and citizens supporting them proved sufficient to make Duda concede (at least partially).

When ambiguity trumps tarnished credibility
After a "disingenuous measure to outmanoeuvre democratic activists" 71 in September 2017, Duda reintroduced virtually the same bills, with the same justifications and the same technicalities.Again, judicial elites spoke up, through formal reports and public statements. 72But instead of joining their protests, citizen support is "limited in effect and reach" 73 and "not as numerous as in July". 74Right after the original bills were vetoed by Duda, PiS started a billboard smear-campaign against the judiciary in August 2017.The campaigns described the judiciary as a "privileged caste" 75 and showcased examples where judges were caught shoplifting, drunk-driving, or starting bar fights.PiS prime minister Mateusz Morawiecki openly accused the judiciary of taking bribes. 76The judiciary was presented as a corrupt, inefficient body, serving the elites at the expense of ordinary Poles.In particular, the message that "communist judges remain in place" 77 and that "courts need to be cleansed of judges who have lost the public trust" 78 was repeatedly voiced in the media, official statements, and speeches.
In effect, PiS portrayed the elite defenders as so self-interestedcommunist judges trying to prevent removing communist judgesthat they could be in no way credibly committed to democracy.These deliberate attempts to delegitimize the judiciary helped tilt the balance in favour of already high ambiguity.While the protests that remained show that judges tried to and succeeded in solving the informational disadvantage of some citizens, they no longer reached and convinced a group of citizens large enough to provide sufficient momentum to make PiS concede, likely at least partially due to this smear campaign.Without citizen support and despite domestic and international elite condemnation of the revised bills, the Sejm approved the bills and Duda signed them into law in December, resulting in a judiciary that is now under complete and direct control of the parliament and the president. 79on't touch my constitution": unambiguous autocratization in Senegal (2011-2012)   When low credibility is undercut by even lower ambiguity Abdoulaye Wade was elected president of Senegal in 2000, in generally free and fair elections.Since then, Senegal has been classified as one of Africa's most stable electoral democracies.80 From the outset, Wade appeared to be committed to democracy, reforming the presidency by limiting the number of terms to 2 and the length of the term to 5 years in 2001.81 In 2007, his commitment to democracy was reaffirmed, when Wade pledged not to run for a third term, 82 while it was still unclear whether a third term could be constitutional (the debate was whether Wade's first term, under the old constitution, counted towards his total number of terms under the new constitution).In 2008, however, his pro-democracy and pro-turnover stance changed, as the presidential term was again lengthened to 7 years.83 Then, in early 2011, Wade introduced a bill that would lower the required proportion of votes to win in the first round of presidential elections from 50% to 25%.84 The opposition immediately claimed the lower threshold effectively guaranteed his win.85 Indeed, if the proposal would have been implemented, Wade would have won the 2012 elections in the first round without the need for a run-off, with 34.5% of the vote.86 In response to Wade's attempt to lower the electoral threshold, the opposition tried to coordinate and present a single front against Wade.A united opposition arguably signals Wade presented a larger threat to democracy than before, as internal differences could be bridged to oppose a substantial threat to democracy.However, the Senegalese opposition had often "cried wolf" and accused Wade of anti-democratic behaviour, for example, during and after the generally free and fair 2007 elections.87 Initially the opposition's unity should strengthen their credibility as democratic defenders, but the coalition quickly broke down when the election came closer.Tellingly, the opposition leaders refocused their attention and finances on their own campaigns.88 They showed their hand: the aim was not only to protect democracy, but rather to use a democratic defence against the proposed changes as a way to oust Wade. 89As such, their credibility as democratic defenders was rather low.
On the other hand, Wade's electoral manipulation has blatant autocratic effects.It is clear that citizens indeed perceived this action as clearly autocratic: ambiguity was low.Nearing the end of Wade's second term, dissatisfaction with his presidency was already quite high: there were protests against overspending, poor economic performance and (youth) unemployment. 90The protests intensified after Wade's proposals and no longer rallied solely around economic mismanagement, but now also revolved around slogans such as "My voting card, my weapon", 91 "Don't touch my Constitution", 92 and "I am a part of this, so I must act". 93Citizens thus rallied explicitly in defence of democracy in the face of this blatant autocratic action.
Despite the comparatively low credibility of elite democratic defenders, citizens mobilized with explicit reference to democracy.In other words, it was so unambiguously clear for citizens that Wade's proposal threatened democracy, that they mobilized in defence of democracy, even though the opposition was more politically motivated and un-credibly committed to democracy.In June, Wade withdrew the proposal to lower the threshold to win in the first round of presidential elections. 94en low credibility is trumped by ambiguity Right after, in July 2011, Wade reneged on his 2007-pledge to not run for a third term and indicated that he would stand for re-election. 95While the scholarship is generally in agreement that extending presidential term limits is blatantly and unambiguously autocratic, this is not always clear for citizens.Especially in this case, ambiguity was introduced when the Senegalese Constitutional Court greenlighted Wade's candidacy. 96While blatantly autocratic in nature, the third term limit was not perceived as autocratic by a sufficiently large group of citizens.
The opposition denounced the Court as partisan and biased and threatened to boycott the elections. 97No longer united, and focussed on their own presidential campaigns, however, their credibility as democratic defenders remained as low as when they opposed lowering the electoral threshold.Opposition leaders sought to organize protests, but the protests were largely "sporadic". 98Above all, the protests were first and foremost anti-Wade and not as explicitly pro-democratic as the first wave of protests.
Compared to Poland, in Senegal credibility appeared to play a lesser role.For citizens, this means that the slight increase in ambiguity through the Constitutional Court's approval of an otherwise blatantly autocratic action trumped any credible cue of democracy that elites could have given them.In other words: up until June 2011, autocratization was so unambiguously clear that even a largely un-credible defender could provide enough information to citizens for them to mobilize.From July 2011 onwards, a slight increase in ambiguity trumped that low credibility.

Taking stock: democratic defence in Poland and Senegal
With the case studies of Poland and Senegal, we aim to illustrate how the trade-offs in democratic defence can change over time and differ between countries.As such, these brief case studies are not meant to provide conclusive evidence of the empirical validity of our model.Rather, they illustrate how our theoretical model works in practice.
In Poland, the PiS government provided many justifications (the focus on efficiency, anti-corruption, and the role the proposals would play in eliminating the "communist heritage" 99 ) to make the autocratic action as ambiguous as possible.Despite such ambiguity, the legal system and the judicial democratic defenders initially proved credible enough for citizens to rally in defence of democracy.Only when the judges' credibility was tarnished did citizens' democratic defence diminish.In other words, PiS tried to push through the autocratic action by diminishing the credibility of opponents.
In Senegal, Wade more blatantly violated democratic norms, signalling clear, unambiguous autocratization.The opposition was clearly more motivated by politics than by genuine commitment to democracy.Even so, autocratization was so unambiguously clear that even hardly credible political elites proved enough to rally citizens in defence of democracy.A month later, however, Wade managed to introduce more ambiguity when the Constitutional Court greenlighted his overturning of term limits.This increase in ambiguity in a context of consistently low elite credibility resulted in only sporadic mobilization in defence of democracy.In contrast with Poland, Wade tried to push through his autocratic actions by increasing the ambiguity of the autocratic action.
From a comparative perspective therefore, these two cases illustrate the trade-offs citizens and elites face in choosing to become democratic defenders.They suggest that elite defenders' credibility and the autocratic action's ambiguity partly compensate for each other in producing the occurrence of joint elite and citizen democratic defence.

Conclusion: countering autocratization and ways forward
Democracies around the world face the challenges of democratic recession.Increasingly, democratically elected incumbents undermine democracy "from within".Existing research has devoted a great deal of attention to explaining why autocratizing incumbents come to power, and which circumstances facilitate their executive aggrandizement, while less attention has been paid to democratic defence against it.In this article, we present a novel theoretical framework of democratic defence that focuses on the interaction between the incumbent, institutional elites, and citizens.Developing a two-level model of democratic defence, we show how the democratic defender's personal interests, repression by the incumbent, the autocratic action's perceived ambiguity, and the defender's perceived credibility interact to affect the occurrence of democratic defence.
On the first level of democratic defence, elites aim to maintain the democratic status quo.They weigh their personal interest in preventing a specific autocratic action against repression by the autocratizing incumbent.When their selfinterest is comparatively high, they mount a defence by opposing the incumbent.Often, however, this is not enough to make the autocratizing incumbent concede.In these cases, citizens need to mobilize in defence of democracy.
On the second level of democratic defence, citizens decide to mobilize by weighing the credibility of the democratic defender against the ambiguity of the autocratic action.When the democratic defender is considered not sufficiently credible, or when the autocratic action is sufficiently ambiguous, citizens are less likely to engage in democratic defence.
Using the illustrative cases of (attempted) democratic recession in Senegal (2011-2012) and Poland (2017-2018), we illustrate these trade-offs and show that successful democratic defence is most likely when elites and citizens align in support of democracy.The cases also suggest that elite defenders' credibility and autocratic action ambiguity can partly compensate for each other in producing the successful occurrence of joint elite and citizen democratic defence.Joint democratic defence in Poland was mainly driven by credibility and in Senegal by (lack of) ambiguity.
We argue that the two-level framework of democratic defence presented in this article will provide useful for further research on the role of actors in defending democracy in other case studies.Future research should examine how autocratizing incumbents respond to democratic defence.In our roadmap, we focus on the role of democratic defenders, and when they decide to engage in democratic defence.However, defence of democracy is a necessary but not sufficient element to counter autocratization.The response of the autocratizing incumbent is the other half of the strategic interaction between autocratizer and democratic defenders.We leave to future research to investigate the factors that influence an incumbent's considerations on when to concede and when to persist, and especially how this interacts with the defenders' considerations.
Another avenue for future research on democratic defence is to consider the endogeneity of our model over time.For the actors involved in democratic defence, the two levels imply that they must be aware of incremental changes by the autocrat that co-opt or otherwise repress domestic defenders' capability to mount a defence.Democratic defence is increasingly harder every time autocratization is successful.Capture of democratic institutions by the incumbent can limit a defender's ability to mount a defence, to be credible, or to counter an incumbent's delegitimation attempts.Defenders can end up in a vicious circle: the context in which defenders operate (negatively) shapes their capacity to act. 100 In the Polish case study sustained attacks on democracy wore down elites' capability and citizens' willingness to protest: emotions faded away by continuous attacks and constant manipulations or cover-up techniques. 101In the case studies, we abstracted the autocratic actions away from some pre-existing level of repression.However, successful autocratization in both cases would have increased the possibilities for future repression.As such, autocratization is a cumulative, self-reinforcing process.
Conversely, successful democratic defence could lead to a virtuous circle: strengthening the credibility of elite defenders and bolstering the democratic commitment of citizensreducing the likelihood of success of future autocratization attempts by the incumbent.Future empirical research should examine the manifestation of such vicious circles of autocratization and virtuous circles of democratic defence.
Lastly, future research should further unpack the dual role of elite democratic defenders as both institutional elites and democratic defenders.We suggest that the most likely elite actors to defend democracy are self-interested elite actors, when their self-interest coincides with protecting democracy.Yet, for citizens to perceive elites as credible democratic defenders they must be perceived as primarily interested in preserving democracy.This means that elites face a daunting balancing act, providing the autocratizing incumbent with many options to undermine the credibility of democratic defenders by delegitimizing them.Democratic defenders must reiterate again and again that their goal is primarily to preserve democracy, not some ulterior political or self-interested motive.And even thenas we showed for the Polish casetheir credibility might be undermined, and democratic defence might fail.
To conclude, an actor-based approach of democratic defence is crucial to understand what actions domestic and international actors can take to prevent (further) democratic recession.The framework presented in this article provides a first step towards better understanding how and when democratic defence occurs.When autocratization in democracies occurs in covert and ambiguous ways, "ordinary folks" are necessary to join elites and together build momentum to defend democracy.

Figure 1 .
Figure 1.When does democratic defence occur?The two levels of democratic defence.