The Crisis of Democracy as De-democratization


Thank you, your comment will be visible after it has been approved.


April 29, 2022

The Crisis of Democracy as De-democratization

Looking back at transitions to democracy and the quality of democracy thirty years later

  • Bolsonaro
  • Brazil
  • crisis of democracy
  • Democracy
Image from Shutterstock

There is a global crisis of democracy in play. It raises the possibility of a potential reversal of the third wave of democratization and puts at risk even consolidated regimes (Mounk, 2019; Levitsky and Ziblatt 2018; Przeworski, 2019; Rosanvallon, 2020). Steven Levitsky, Daniel Ziblatt and Adam Przeworski, among many other authors, set up the main elements of a theory on the crisis of democracy as situations in which status quo institutions are in some kind of disaster: “…no change occurred, but it may…the current situation is in some ways threatening…” (Przeworski, 2019:10). For Levitsky and Ziblatt two elements are at stake, institutions and legality in the internal deterioration of democracy (Levitsky and Ziblatt, 2018:5).

 Yet, two major caveats remain to be discussed in the way this debate has been unfolding during the last few years:

The first one is a differentiation of cases among well established and new democracies, which requires separating democracies according to the degree of institutionalization of conflict and balance of power, in addition to polarization (Avritzer and Renno, 2021). The crisis of democracy can be and should be separated from the analysis of populism or polarization (Rosanvallon, 2020), turning instead to institutional features and political values which are ingrained and long-lasting, placing countries at greater or lesser odds of authoritarian reversal.

The second issue is related to the comparison, or lack thereof, between cases of crisis of democracy. So far, literature has concentrated on the US case, with incidental incursions in Chile, Peru, Venezuela, and Germany (Levitsky and Ziblatt, 2018; Przeworski, 2019). The aim of this paper is to develop a concept of democratic erosion that could move the current debate from the opposition between democracy and autocracy (Mounk and Foa, 2021) to more nuanced scenarios, having as its starting point shortcomings in the formulation of transitions to democracy (O´Donnell and Schmitter, 1986; O´Donnell, 1991; Linz and Stepan, 1996). The concept is based on two elements: the decrease in trust and satisfaction with political institutions and democracy. The second aim of the paper is to operationalize the concept of democratic erosion and apply it to the crisis of democracy in Brazil and Argentina. 

Transition to democracy, consolidation and the quality of democracy

Theories of transition to democracy emerged in the Latin American context of the 70´s. Departing from a purely analytical perspective, transition theory acknowledged the responsibility of non-democratic elites for the existence of authoritarian regimes in Latin America. It began from the crystal-clear assumption that Latin American elite anti-democratic or semi-democratic convictions combined with economic conflicts to undermine the region’s democratic systems. Thus, from a theoretical point of view, transition theory also acknowledged that democratization does not consist in a set of variables to be empirically identified, but rather in a long process in the course of which there is the possibility that circumstances will force, persuade, or even fool non-democrats into assuming a democratic attitude in relation to democratic political institutions (Rustow, 1970:344-345). Thus, Rustow proposes neither accepting nor ignoring elites’ anti-democratic convictions, but rather thinking about a process through which it will be possible to transform non-democratic into democratic elites. Such a process encompasses an institutional dimension and it is this process, in a long-term perspective, that we will evaluate throughout the book. 

Building on the assumption of a relationship between the role of democratic and non-democratic actors in the decision-making process, transition theory understood the breakdown of democracy throughout the “second reverse wave of democratization” as an inability to build satisfactory institutions (Linz,1973). In this sense, the institutional dimension constitutes an inherent concern of transition theory from the very start. Transition theorists believe that “institutions have an impact upon outcomes of conflicts” and that “institutions that would provide the relevant political forces with reasonable security can be found under some circumstances” (Przeworski, 1988:66). This discussion allowed a definition of transitions: they represent a period of time in which an institutional arrangement aiming at the re-establishment of political competitiveness between democratic and authoritarian political actors is reached. In the course of this process, institutional arrangements as well as mutual trust are built in such a way that at the end of a period of transition, authoritarian political actors withdraw from the political scene (O’Donnell and Schmitter, 1986:IV). This definition constitutes the core of transition theory’s conception of democratization. Yet, in order to adapt this feature of democracy to late developing contexts, transition theory introduced a few changes to this conception of democracy.

First, it renewed democratic theory by problematizing the convictions of political elites. For transition theory, elites do not a priori possess democratic values. Transition theory assumes the possibility of different types of actors within the political system. There are those which it calls hard-liners: those who “believe that the perpetuation of authoritarian rule is possible and desirable, if not by rejecting outright all democratic forms, then by enacting some facade behind which they can maintain inviolate the hierarchical and authoritarian nature of their power” (O’Donnell and Schmitter, 1986:16). Next, there are the blandos: those who believe that authoritarian regimes should make use of some form of electoral legitimation in the near future (O’Donnell and Schmitter,1986:16). And, in the third place, the democratic opposition: those who advocate a full return to electoral competition but accept the need to negotiate the rules of competition with authoritarian powerholders. Thus, from the point of view of political elites, the process of transition is both a restricted form of electoral competition between the authoritarian powerholders and the electoral opposition and, at the same time, a negotiation of the rules under which this competition will take place. The issue that remained unsolved was how to deal with non-democratic elites after democratization. We will return to this issue below. 

The second is proceduralism linked to the transitions pact. According to O´Donnell and Schmitter “…a pact can be defined as an explicit but not always publicly explicated or justified agreement among a select set of actors which seek to define rules governing the exercise of power on the basis of mutual guarantees for vital interests…” (O´Donnell and Schmitter, 1986:IV:37). 

We see two problems with the theory as we analyze the current crisis of democracy. The first problem is that proceduralism and the capacity to craft new electoral institutions do not have the capacity to bind social actors and public opinion in their support for democracy. Thus, the second problem is that flawed proceduralism leads to low quality democracy. Eventually, as imperfect or low quality democracies based on these pacts get stuck, the whole democratic project is put at stake. 

The “quality of democracy” theory

The theory of the quality of democracy build directly on the theory of transition both from a historical and analytical perspective: from a historical perspective the issue is once you have a large spread of democracy in the world the conceptual issue should be “evaluating and explaining the character of democratic regimes “ (Diamond and Morlino, 2005). From an analytical perspective the problem became more complicated because it involved several dimensions of democracy such as rule, responsiveness and procedures. (Diamond and Morlino, 2005). Thus, the definition of what is a quality democracy operates on the assumption of a teleological improvement of democracies capable of providing its citizens with “ a high degree of freedom, political equality and control over public policies…”. Though the theory help to advance democratization debates it fitted initially into an overly optimistic framework of improvement and deepening of democracy worldwide. 

Morlino advanced the model by trying to put together two dimensions that are viewed in terms of alternatives rather than teleological evolution. They are: “democratic consolidation and crisis and democratic qualities deepening or worsening. “ (Morlino, 2012:19). Thus, Morlino points out from the very beginning that a theory on the quality of democracy does not mean a general and teleological theory that takes for granted that democracy will always improve its quality. In an additional remark Morlino compares democracy to an industrial product and disaggregates two of its quality dimensions. For him ”… quality is the result of an exact, controlled process carried out according to precise, recurring method and timing: the emphasis is on procedure… Finally…the quality of the service is indirectly derived from the satisfaction expressed by the customer …” (Morlino, 2004:6).

I would like to highlight a few considerations on procedure and quality in order to understand the limits of transition theory in the explanation of the current crisis of democracy. The theory of transition to democracy places its focus on procedures that can receive different definitions, the most important among them being Robert Dahl’s rules for the existence of polyarchy. This definition, which is entirely procedural, includse rules for free, fair, competitive and recurrent elections and a series of other procedures for electoral competition and the exercise of political rights (Dahl, 1972). The theory of transition adopted a mixed approach based on negotiations for the withdrawal of authoritarian powerholders. Its central idea is a pact as defined above. (O´Donnell and Schmitter, 1986:IV:37).

When we look into this formulation from the perspective of 30 years onwards we see two limitations in relation to the procedural formulation: the first one is that despite the attempt to propose a procedural conception of pact in the negotiations for the transition from authoritarianism to democracy, transitologists acknowledged that procedures would have to abide to “vital interests” of the parties and the non-authoritarian power holders have had vital interests that would eventually collide with democracy. In Brazil these interests were related to an amnesty that precluded transitional justice. A similar element was proposed by the Argentinian military but was later struck down by president Alfonsin (Avritzer, 2002). In Chile, a broader continuity with the authoritarian political order was inserted into the Constitution and only now is being revised (Cuoso, 2011; Heiss, 2019). Thus, pacted proceduralism implied in the heavy toll of allowing remnants of the previous authoritarian order to remain in place.

The second issue is that transition to democracy assumed that once authoritarian powerholders left power there was no way back, because authoritarianism was ingrained into the political system but not in society. The breakdown of democratic regimes was a consequence of an institutional deficit (Linz, 1973; Linz and Stepan, 1996). Though I would not deny the institutional element in the breakdown of democratic regimes, I would point out that it is far from encompassing all the dimensions of authoritarianism that are reconstructed through several different dimensions, mainly the anti-political one (Urbinati, 2019). 

The theory of the quality of democracy solves some of these issues by associating proceduralism and satisfaction. Proceduralism for the quality of democracy perspective is linked to the enforcement of political freedom and some dimension of accountability. Yet, the quality of democracy theory points out a dimension of satisfaction that cannot be downplayed, especially if we understand the possibilities of democracy in terms of pairs with opposite directions as table one below shows.

Road to democracyProcedural elementsQuality elementsResult 
First stage transition to democracyFocus on procedural pact or transitional justiceQuality elements are partially compromisedLow quality with or without justice
Second stage: democratic installationFocus on procedures for political equality, elections, rightsQuality element may expand in cases of transitional justice, accountability and participation Partial improvement in quality and satisfaction with democracy
Third stage: democratic deepening with or without quality commitment Elections intertwine with accountability and participationQuality may improve if the two dimensions mutually reinforce the otherMinimal level of democratic quality
Fourth state: crisis if accountability and participation do not overlapElections are challenged and accountability collide with sovereigntyQuality receded because founding issues were not dealtChallenge to democracy’s procedural elements

Hence, the problem with the relation between transition and quality of democracy or proceduralism and satisfaction is that eventually, satisfaction is so low that it compromises minimal procedures for fair competition. I will give an example of that relation based on the Brazilian case. 

Brazilian democracy: from low satisfaction to crisis

Brazil is a case of long-term low satisfaction with democracy even by South American standards, since the beginning of the first decade of the century. It is interesting to point out that the Brazilian levels of satisfaction with democracy are lower than the Argentineans level though Argentina suffered a major political crisis in 2001-2002, with the blockage of savings and the resignation of elected and appointed presidents (Peruzzotti, 2017). 

Table 2: satisfaction with democracy in LA

Satisfaction with Democracy 2006
Very satisfied5,10%3,70%14,40%4,10%
Very Dissastified9,70%13,50%5,40%11,30%
No opinion0,40%0,70%0,40%1,80%
Don’t know0,60%4,70%3,30%2,20%

Thus, Brazil has had a lower level of satisfaction than its neighbors, some of them, like Urugua , who are deeply satisfied with their democracy. My point is that already at that moment in which there was a virtuous cycle in operation in Brazil with three successful democratic administrations, the remnants of authoritarianism were already charging their costs in Brazil; inequality and deep violations of human rights, the two most important legacies of authoritarianism, were influencing the new democracy. Levels of inequality decreased at the end of that decade. Yet, a dysfunctional democracy was already in operation and this will determine the fate of Brazilian democracy in the 2011-2020 decade.

June 2013 was a watershed in the construction of democracy in Brazil. At that point, Brazil had been democratic for 28 years, the left had governed the country for 10 years, poverty had been decreasing for more than a decade and inequality started to decrease at the end of the century’s first decade. June 2013 started unequivocally on the left side of the political spectrum (Goncalvez, 2019) with the Free Pass Movement launching its most ambitious move, calling for demonstrations against the recent elected mayor of São Paulo, Fernando Haddad, who failed to convince president Dilma to change her fuel policy and had to authorize against his best interest a hike in São Paulo´s bus fare (Haddad, 2017). 

The São Paulo`s free bus fare demonstration of June 7th, 2013, would become famous for several reasons: it was a left-wing movement that unleashed the public reorganization of Brazilian right-wing sectors that were s already under way in small groups and think tanks (Solano et al, 2021); it consolidated a negative vision of both the P.T. and the political system. (Amaral,2016 ). Last but not least, a process of judicialization already under way, with a strong criminal focus that would be strengthened after June 2013 with the approval a few months later of 12.850/2013, on criminal organization that would focus on the P.T. and the political system and further erode political stability. Thus, June 2013 was triggered by the low quality of the Brazilian political system, which implemented electoral and accountability institutions but could not generate an equilibrium among them.

June 2013 altered the correlation of forces, but it could not avoid the reelection of Dilma Rousseff in 2014 (Avritzer, 2016). The most significant post- election fact was the non-acceptance of the result first by Aécio Neves and later by a large and diverse portion of the Brazilian Congress. Aécio Neves did not accept his electoral defeat and gave an interview with the headline “I was defeated by a criminal organization,” a successful attempt to link Dilma Rousseff`s electoral victory with corruption investigations in Curitiba (O Globo 29 de outubro de 2014). In addition to that, he immediately initialized a plan to put an end to Dilma Rousseff´s mandate in association with his longtime partner Eduardo Cunha,¹ who was elected Speaker of the Lower House.

Thus, we have a second step that is interesting to analyze. As I have showed above, both the theory of transition and the theory on the quality of democracy place proceduralism ahead of quality of democracy and they are both right to do so. However, there is an additional point that I would like to make and that is missed by both theories: that once quality declines , as it has been the case in Brazil, proceduralism will also be challenged. 

Demonstrations against Dilma Rousseff eroded her capacity to govern and to re-organize her basis of support in Congress. Eduardo Cunha the House Speaker accepted the charge for impeachment of Dilma Rousseff on December 4th, 2016. The impeachment was filed by three lawyers linked or close to PSDB in São Paulo, which again shows that proceduralism and acceptance of electoral results were a past practice. The charges were legally questionable because they were administrative. Dilma Rousseff was impeached by the House on April 17, 2016, a vote confirmed by the Senate in late August. After Dilma Rousseff’s impeachment the door was open for the return of non-democratic practices that would take place in 2018.

The year of 2018 started with a very low level of trust in political institutions that were the consequence of the impeachment, the corruption scandals investigated by the Lava Jato operation and the disastrous performance of the post-impeachment administration of Michel Temer whose approval rate sunk even below Dilma Rousseff´s . In the first poll carried out by “Instituto da Democracia”² in March 2018³ it was possible to find two interrelated movements at the level of public opinion.

The first one was a further relativization of the support for democracy. Brazil has had low levels of support for democracy in comparison to both consolidated democracies and the Southern Cone democracies. Only after the broadening of access to social services and protections and the huge decrease in poverty levels did it start to climb in Brazil by the end of the first decade of the century. However, all these advancements were rapidly reversed by the political crisis described above . In March 2018, for the first time in a few decades, the percentage of Brazilians who thought that certain circumstances would justify a break with democracy was slightly above 50%, s as Graph 1 below shows.

Graph 1: Which circumstances justify breaking with democracy?

This was the initial point for two additional important breaks with the democratic tradition built after 1985. On April 4th, 2018 on the eve of an important Supreme Court decision on the so-called “presunção de inocência” , a technical term that is part of A rticle 5 of the 1988 Constitution that limits arrest to those caught in the act of committing a crime. The idea corroborated by Article 5 is that judicial appeals guarantee answering felonies and misdemeanors without being arrested. This was the case of ex-president Luis Inácio Lula who was being targeted by the Lava Jato operation at that point. On the eve of the Supreme Court decision a writ by the commander of the army, General Villas Boas, pointed out that the army was closing watching the Supreme Court vote and would not renounce its Constitutional mission . Though the language of the writ was general and abstract it was immediately read as a message to the Supreme Court, given the content of A rticle 142 of the Constitution discussed above.

The election of Jair Bolsonaro as president of Brazil became possible due to the break with the three dimensions singled out above: first of all, the break with informal and formal ideas about the legitimacy of election results. Even though the Electoral Court dismissed Aécio Neves’ challenge to the elections of 2014, the challenge disassembled one of the major accomplishments of the period known as the “ New Republic”. Though Brazilian democracy could have survived this issue, it came associated with two other issues that created space for open challenges to democracy: an impeachment with questionable legal grounds that put in place the defeated economic program of the opposition, and last but not least the return of the political influence of the military. Together, these three democratic drawbacks made the Bolsonaro candidacy viable.

As a conclusion, we can analyze the Brazilian case as one of the partial collapse of democracy despite a successful transitional pact. The transition pact was successful insofar as it guaranteed, through mutual agreement between the authoritarian regime and the opposition, the rules for the return of competitive elections. . The problem with these kinds of pacts is that the new order is tainted by the old, and the procedures for free elections most of the time intertwine with the remnants of the old order. This issue has been dealt in the literature in terms of a dispute between consolidation and non-consolidation. I tried in this paper to expand the issue using the elements of the theory on the quality of democracy. This theory tries to move the debate forward by adding a second dimension to proceduralism, namely the quality of democracy. Indeed, this is a new and important dimension. However, the issue I tried to approach in the paper is whether there is a necessary stage or if low quality could commit the whole procedural project. I tried to show that this is the Brazilian case. Because the pact of the transition was successful, there was a lot of leeway for authoritarian actors to ignore the new democratic rules and keep pockets of authoritarianism untouched. The result was that the quality of democracy and the support for democracy were low in the long run. This would eventually reverse procedural gains made during democratization. The question that remains to be answered is: is Brazil a unique case or should the theory be revised?

Leonardo Avitzer, is professor at the Federal University of Minas Gerais. He was president of the Brazilian Political Science Association (ABCP) from 2012-2016. He published Democracy and the Public Space in Latin America, Princeton University Press and O pendulo da Democracia no Brazil that just won the award of best book of the year in policy. He coordinates the Instituto da Democracy an academic think tank on Democracy in Latin America.


¹ It is very important to understand the type of MP that Eduardo Cunha has been. Cunha entered politics with Collor de Mello and was president of the state-run telephonic company during Collor´s administration. At that point he belonged to Collor´s newly formed party, PRN. After the impeachment, Cunha belonged to PPB and after 2006 to PMDB. Cunha was the leader of centrão and has limited loyalties vis-a-vis PMDB. His disputed with Dilma began because both him and Temer has interest contradicted by the ports reform. Cunha coordinated Aécio Neves presidential campaign in Rio de Janeiro despite PMDB support for Dilma Rousseff.

² Instituto da Democracia is a joint Project by UFMG, Cesop, Unicamp.

³ The Instituto da Democracia’s survey was applied in a sample of 2500 interviewees. The sample was representative of whole countries population and xxx.


Avritzer, 2002. Democracy and the public space in Latin America. Priceton university press

Avritzer, 2016. The impeachment of Dilma Rousseff. Critical policy studies

Avritzer, Leonardo and Lucio Renno. 2021. The pandemics and the crisis of democracy. Journal of Politics in LA. 

Diamond and Morlino. 2005. Assessing the quality of democracy. Johns Hopkins university press

Goncalves, Izabella. 2019. Brazil em movimento. Pd.D dissertativo UFMG

Levitsky and Ziblatt, 2018. How democracy dies. Crown publishing group

Morlino 2004. Good and bad democracies: how to conduct research on the quality of democracy. Journal of communist studies and transition politics. Vol 20.

Morlino 2011. Changes for democracy. Oxford. Oxford university Press

O´Donnel and Schmitter, 1986. Transitions from authoritarian rule. Johns Hopkins University Press. 

Mounk and Foa. 2019. Sign of de-consolidation. Journal of democracy. 

Peruzzotti. 2017. Populism as democratization menesis: the politics of regime hybridization. Chinese journal of political science

Przeworski, Adam. 2019. Crises of democracy. Cambridge. 

Rosanvallon, 2020. The populist century. Cambridge. Polity press.

Urbinati, Nadia. 2016. Me, the people. 


Leave a reaction with this article