[Abstract] [1] The dream of all politicians is to remain forever in office and to use it to do whatever they want. Most governments attempt to advance these goals by building popular support within the established institutional framework. Some, however, seek to protect their tenure in office and to remove obstacles to their discretion in choosing policies, by undermining institutions and disabling all opposition. The striking lesson from successful cases of backsliding is that governments need not take unconstitutional or undemocratic steps to secure complete domination. The effect of what we call ‘stealth’ is that if the opposition fails to stop the government from taking a particular series of legal steps, it will eventually be unable to prevent it from taking illegal ones. We investigate whether a government will take anti-democratic steps, whether it can be stopped short of realizing its complete domination and whether it is likely to be removed at any stage of the process.

1. Introduction

The dream of all politicians is to remain forever in office and to use their tenure to do whatever they want. Most governments attempt to advance these goals by building popular support within the established institutional framework. Some, however, seek to protect their tenure in office and to remove obstacles to their discretion in choosing policies, by undermining institutions and disabling all opposition. Prominent recent examples are Turkey under the government of the Justice and Development Party (AKP), Venezuela under Hugo Chávez and Nicolás Maduro, Hungary under the second government of Fidesz, and Poland under the second government of the Law and Justice Party (PiS).

Democratic ‘deconsolidation’ or ‘backsliding’ is a process of gradual erosion of democratic institutions and norms. Huq and Ginsburg (2018a: 17) use the term ‘authoritarian retrogression’, distinguished from outright ‘reversion’, and define it as ‘a process of incremental (but ultimately still substantial) decay in the three basic predicates of democracy-competitive elections, liberal rights to speech and association, and the rule of law.’ As ‘backsliding’ or ‘de-consolidation’ or ‘retrogression’ (whatever one wants to call it) advances, the opposition becomes unable to win elections or assume office if it does win, established institutions lose the capacity to control the executive, while manifestations of popular protest are repressed by force. This process is propelled by the desire of the government to monopolize power and to remove obstacles to realizing its ideal policies. Yet it is a process of interaction between the government and the various actors that seek to block it. Hence, the strategy of governments that enter on this path concentrates on disabling potential blockers, who differ from case to case but typically include the opposition parties, the judicial system, the media, as well as street protests.

The thinking is as follows. A government deeply committed to some ideological goals, such as Islamization in Turkey, ‘Bolivarianism’ in Venezuela, ‘preserving the purity of the nation’ in Hungary, or ‘defending Christianity’ in Poland, wins an election [2]. This government decides whether to take steps to increase its incumbent advantage, or steps that would increase its discretion in policy making. As Lust and Waldner (2015: 7) put it, ‘Backsliding occurs through a series of discrete changes in the rules and informal procedures that shape elections, rights and accountability. These take place over time, separated by months or even years.’ Examples of such steps include changing electoral formulae, re-districting, changing voting qualifications (age, eligibility of citizens residing abroad), harassing the partisan opposition, imposing restrictions on NGOs, shifting power from the legislature to the executive, reducing independence of the judicial system, or using referendums to overcome constitutional barriers. Having observed such steps, citizens who value democracy may turn against the government, even if they support its policies or enjoy the outcomes that they attribute to its policies. If the opposition rises, the government may be removed from office, or it may decide to stop taking further steps in anticipation of its rise.

In principle, the opposition could prevent the government from taking the next step, either by stopping it through legal measures, such as defeating a bill in parliament, obtaining a presidential veto, or a favourable court ruling, or by removing it from office, by means ranging from elections, impeachment, or a vote of no confidence, to mass protests and coups (which failed in Venezuela and Turkey). The history of the four cases we have referred to above shows, however, that governments regularly overcome legal obstacles. In Turkey, when in 2007 the President vetoed a constitutional amendment passed by parliament for direct election of the president, the government organized a referendum and won. In Venezuela, when the opposition won a legislative election in December 2015, Maduro replaced the Congress by a newly elected Constitutional Assembly. In Hungary, when the Constitutional Tribunal invalidated an electoral reform in 2013, the government passed a constitutional amendment curbing the power of the tribunal. This is not to say that governments always prevail: in Poland, for example, the government withdrew from fining an opposition TV station that is owned by US interests. Nevertheless, it seems that legal counter-moves by the opposition at most slow down the process, but are ineffective in stopping it. This is why we think of the opposition in terms of an effective threat to remove the government.

The first obvious question is why some governments decide to enter on this path while most refrain from it. The second is whether once a government takes such steps it can ever be stopped short of realizing complete domination. The third is whether a potential opposition would be able to remove the government and reverse this process.

It is worth emphasizing that we are trying to understand the ‘how’ rather than the ‘when’: how deconsolidation evolves when it does, rather than the conditions under which it is likely to occur. We learn from the experiences of the four countries mentioned above because we want to understand how democracies can be destroyed through small steps; not why backsliding occurs in some countries and not others – for which see Maeda (2010), Svolik (2015), and Graham, Miller and Strøm (2017). Our question is not whether it will happen here, but if it can happen anywhere?

1. Stealth

The puzzle associated with destroying democracy by ‘backsliding’ is how a catastrophic state of the world can be gradually brought about by small steps, against which people who would be adversely affected by it fail to react in time. As Ginsburg and Huq (2018b: 91) put it, ‘The key to understanding democratic erosion is to see how discrete measures, which either in isolation or in the abstract might be justified as consistent with democratic norms, can nevertheless be deployed as mechanisms to unravel liberal constitutional democracy.’ In the story of ‘the frog in the pot,’ if a frog is put suddenly into hot water, it will jump out, but if it is placed in cold water which is then heated slowly, it will not perceive the danger and will be cooked to death. Yet the story is not true: recent experiments show that the frog will get uncomfortable when the water is heated and will try to jump out (see ‘Boiling frog’ in Wikipedia) [3]. How, then, can gradual backsliding succeed in destroying democracy?

The first lesson that we are learning from recent experiences is that democratic institutions do not provide safeguards against subversion by duly elected governments observing constitutional norms. When Hitler came to power, through an ‘authoritarian gap in the Weimar Constitution’ (Bracher 1966: 119), the possibility of a legal path to dictatorship was seen as a flaw of this particular constitution. Yet such gaps may be generic. The father of constitutionalism, Montesquieu (1995: 326), insisted that ‘For the abuse of power to be impossible, it is necessary that by the disposition of things, the power stop the power.’ But, pace Madison (Federalist #51), checks and balances do not operate effectively when different powers of government are controlled by the same party: as Madison himself was almost immediately to discover (Dunn 2004: 47–61), a constitutional separation of powers is vulnerable to partisan interests. Courts, constitutional as well as ordinary, can be packed, intimidated, or circumvented. Wholesale changes of constitutions, amendments, or referendums can constitutionally overcome extant constitutional obstacles. Public bureaucracies, including security agencies, can be instrumentalized for partisan purposes. Public media can be controlled by partisan regulatory bodies, while private media can be legally intimidated or destroyed financially. All such measures can be taken legally. As Landau (2013: 192–3) observes, ‘The set of formal rules found in constitutions is proving to be a mere ‘parchment barrier’ against authoritarian and quasi-authoritarian regimes. There is even worse news: existing democracy-protecting mechanisms in international and comparative constitutional law have proven ineffective against this new threat.’

Democratic deconsolidation need not entail violations of constitutionality. Thinking about the United States, a constitutional lawyer writes, ‘If it happens here, it won’t happen all at once…. Each step might be objectionable but not, by itself, alarming …there will have been no single, cataclysmic point at which democratic institutions were demolished… the steps toward authoritarianism will not always, or even usually, be obviously illegal…. In fact, each step might conform to the letter of the law. But each step, legal in itself, might undermine liberal democracy a little bit more.’ (Strauss 2018: 365–6). In a broader context, another constitutional lawyer concludes, it is difficult to identify a tipping point during the events: no single new law, decision or transformation seems sufficient to cry wolf; only ex-post do we realize that the line dividing liberal democracy from a fake one has been crossed: threshold moments are not seen as such when we live in them’ (Sadurski 2017: 5). This is what we mean by ‘stealth’: ‘the use of legal mechanisms that exist in regimes with favourable democratic credentials for anti-democratic ends’ (Varol 2015).

Some measures that backsliding governments adopt do not even require legal acts, just a change of practice. For example, the Polish ruling party, PiS, gradually altered the parliamentary procedure for introducing new bills: parliamentary rules say that bills proposed by the government must be subject to public hearings, while private members’ bills do not, and the government switched to presenting its proposals as private bills of its deputies (Sadurski 2017: 6). Moreover, nothing is constitutionally wrong with legal measures such as a parliamentary act easing restrictions on the teaching of the Koran (June 2005 in Turkey), or anti-terror laws (June 2006 in Turkey, May 2016 in Poland), or a statute that requires non-governmental organizations to register as foreign organizations if they receive funding from abroad (June 2017 in Hungary). These are ordinary laws, passed in accordance with constitutional provisions, by legally competent bodies – a prerogative of any democratic government. Even changes to constitutions are valid, as long as they observe constitutional provisions, as they did in Hungary in April 2011, and in Turkey by the referendums of October 2007 (direct election of the president, after the incumbent president vetoed a parliamentary act), September 2010 (increased civilian control over the military and the courts) and April 2011 (introducing a presidential system). As a result, several observers agree that ‘there is usually no single event or governmental conduct which may mobilize the resistance by sending a clear signal that democratic norms are imperilled’ (Huq and Ginsburg 2018b), and that ‘slow slides towards authoritarianism often lack both the bright spark that ignites an effective call to action and the opposition and movement leaders who can voice that clarion call’ (Bermeo 2016: 14).

Protests against legal measures taken by a government that has just won an election only show that the opposition is a sore loser and that it does not respect democratic norms. Even more perverse are the situations in which governments succeed to pack or control the constitutional tribunals and then use judicial review to legitimize their actions, as they did in Venezuela, Turkey and Hungary. They are perverse because they allow governments to portray actions of the opposition as anti-constitutional: in President Trump’s tweet the appointment of special counsel Robert Mueller was ‘totally UNCONSTITUTIONAL’ (@realDonaldTrump 6 April 2018, 4.01 pm). Some of the rhetorical games employed in backsliding involve competing for the mantle of ‘democracy’ and ‘constitutionalism’, in which the opposition does not always prevail. Thus a pro-Putin Russian journalist, Mikhail Leontiev, disingenuously observed: ‘I do not understand what is undemocratic in that some force enjoying overwhelming social support wins elections’ (an interview of 19 January 2008 for a Polish daily, Dziennik).

Note that claims of ‘unconstitutional’ require, in principle, a more rigorous test than those of ‘undemocratic.’ Declarations that a particular statute or an action of the government violates the constitution are issued by specialized bodies designated by the constitution and are couched as interpretations of its text. But courts are populated bodies (Ferejohn and Pasquino 2003): their decisions are made by particular people appointed by politicians. Hence, if a government succeeds in stuffing these bodies with its partisan supporters, they will issue decisions favourable to the government. Venezuela is a flagrant example. Moreover, constitutions can be amended or completely replaced, just by constitutional provisions.

Constitutionality can be questioned even when a court has been constitutionally formed and when it has ruled that the actions of the government are constitutional. One issue is whether an act adopted following constitutional provisions that abrogate constitutionality altogether can ever be considered constitutional. Hence, a literal interpretation of ‘constitutionality’ does not provide a decisive criterion. Some broader conception [4] must be invoked to deem actions of a government to be violating constitutionality when the pertinent bodies declare them to be constitutional and, given that such conceptions are inevitably vague, partisan-based disagreements are unavoidable.

The notion of ‘undemocratic’ is even more permissive. There is nothing ‘undemocratic’ about the election of Donald Trump: in the words of his advertisement, ‘Our movement is about replacing a failed and corrupt political establishment with a new government controlled by you, the American people.’ (https://www.youtube.com/watch?v=vST61W4bGm8). It is even more paradoxical to claim the same about referendums, in which Marine Le Pen, advocating a vote on ‘Frexit,’ promised ‘You, the people, will decide.’ Not holding scheduled elections or committing blatant fraud are almost universally seen as violations of democratic norms, but short of this instinctive marker, the norms of what constitutes an undemocratic practice are less crystallized and diverge even more across partisan lines. For example, in the United States, almost everyone believes that fraud-free elections with equal voting rights are important for democracy, but fewer people think it important that districts should not be biased, or that the government should not interfere with the press, with divergence occurring between supporters and opponents of President Trump (Bright Line Watch Wave Survey 5, April 2018).

Moreover, it is often unclear whether certain steps taken by governments are anti-democratic. Imagine that a government extends voting rights to citizens residing abroad (as both Recep Tayyip Erdoğan and Silvio Berlusconi did), or adopts legislation to require additional documentation at the polling place, or re-maps electoral districts. Are these steps anti-democratic? The government says: we want to extend rights to all citizens, we want to prevent fraud, we want every vote to have equal weight. The opposition says: the government does not care about rights and is extending the vote to Turks in Berlin only because they will vote for it, the government does not care about fraud and only wants to prevent poor people who do not have the necessary documents from voting, the government is redistricting in its favour. All these steps are adopted observing constitutional provisions, so they are not ‘undemocratic’ in the sense of violating procedural norms. The disagreements are not about facts but about intentions, and intentions are not directly observable.

Now, governments may take actions that are flagrantly unconstitutional and in some cases undemocratic. Refusing to comply with court rulings is clearly unconstitutional, and, in the United States, subjects the perpetrator to contempt-of-court charges. Banning an opposition newspaper flagrantly violates democratic norms. Moreover, some are offensive whether or not they constitute either: the shooting of peaceful demonstrators, as at Kent State University on 4 May 1970, may provoke widespread outrage, whatever the legal or normative niceties. Stealth is a process by which a government takes steps, none of which are flagrantly unconstitutional or undemocratic, and which cumulate in enlarging its discretion in making policies or undermining the capacity of the opposition to remove them.

1. Dynamics of subversion from above

Parchment barriers are not sufficient to prevent erosion of democracy by governments that proceed by stealth. The question, then, is whether a government intent on backsliding can be dissuaded from pursuing it, or be removed because of the rise of popular opposition.

Suppose that a government decides whether to take a step that increases its incumbent advantage, or that brings a policy closer to its ideals, or both, anticipating the size of the opposition. In their turn, individual citizens decide whether to switch against the government, or to support it. The probability that the government will remain in office depends on the size of the opposition. The government starts on the backsliding path if it expects to be better off in terms of staying in office and implementing its preferred policies than under the institutional status quo.

Individuals attach different weights to the extent in which they care about democracy and about certain policies, such as anti-immigration measures, or outcomes, such as income growth (Svolik 2017, Graham and Svolik 2018). People who do not like the outcomes the government generates, say environmentalists under Trump’s government, oppose it regardless of the value they place on democracy. In turn, people who support government policies trade off at different rates their benefits from policies against the damage to democracy, where ‘democracy’ means that the government can be removed by elections (or some constitutional provisions, such as impeachment, or a vote of no confidence), when a sufficient majority opposes it. Obviously, there may be people who do not care about democracy at all.

While this much is general, there are many possibilities. The opposition may not rise at all; it may remain dormant at first and only then suddenly increase; it may remain at some constant level or just rise sporadically in reaction to particular measures taken by the government. By making appropriate assumptions, one can obtain any result one wants – generating conclusions by assuming them. Therefore, our strategy is to keep these possibilities as open as possible by exploring the generic conditions under which a government enters on a path of deconsolidation, as well as the conditions that cause it to desist under the threat of opposition, or to be removed from office while deconsolidating.

Now, a government comes into office by winning an election only if some qualified majority believes it would benefit from its policies. In turn, a government is certain to backslide successfully if a majority benefits from its policies and does not care at all about democracy. Stealth matters when a majority benefits from government policies, but would oppose the government if it was certain that the government was undermining democracy. Under such conditions, if the government cannot sufficiently hide the damage that it inflicts on democracy, then it does not backslide. If it can hide it to some intermediate extent, it does backslide and enjoys a high level of support – there is then a high probability it will backslide successfully. Yet, if it is obvious that the government can hide its actions, people are more suspicious and the probability of successful backsliding declines. Hence, backsliding by stealth is successful when the people who benefit from government policies are not too distrustful of the government and the government can hide the effects that its actions have on democracy.

Defending democracy imposes thus a difficult challenge on individual citizens. To act against the government that may be destroying democracy, individuals who enjoy its policies or certain outcomes that they attribute to those policies [5] must consider the effect of the government’s actions on democracy. Even if individuals have consistent time preferences (Ackerloff 1991), they must be able to calculate the effects of seemingly democratic steps – they must be able to see through the stealth. This is a formidable task and, even if the incapacity to anticipate the future violates the assumption of full rationality, it should not be surprising that people are unable to perform it. Consider a sequence of events in which a government first adopts legislation requiring additional documentation at polling stations, then has its cronies buy a major opposition newspaper, then re-maps electoral districts, and then stuffs the bodies that administer and supervise elections. What people need to see is that, although each of these measures is perfectly legal, their cumulative effect is to protect the incumbent from being defeated by even a largely majoritarian opposition. Moreover, policies have interactive effects that are even more difficult to calculate. Scheppele’s (2013) example is the interaction of Article 48 of the Weimar Constitution that permitted the president to declare a state of emergency, subject to the check that parliament could reject such a declaration, while Article 25 permitted the president to dissolve parliament once for any reason; the effect of which was that once parliament was dissolved, the president could declare an emergency at will. Even the most competent constitutionalists who authored this constitution failed to see the potential effects of this combination, and it proved fatal.

One might think that people would be informed by the political leaders of the opposition. But their potential role is limited. Exhortations by partisan oppositionists are not likely to be effective in influencing beliefs. People know that the opposition leaders’ goal is to replace the incumbents, whether they are acting for good or bad reasons. If opposition leaders criticize every action of the government, people tend to dismiss their messages: as Austen-Smith (1992) put it, ‘speech predictable by interests is not credible.’

The fact is that backsliding governments have enjoyed continued popular support. To our best knowledge, the only case in which a backsliding government has lost an election and left office is that of Sri Lanka in 2015, and this outcome was the result of massive defections from the ruling coalition, with the winner a previous minister in the outgoing government. Other backsliding governments have suffered temporary reversals, but were able to recover and continue: with 40.9 per cent of the vote, AKP failed to win a majority of seats in the election of 7 June 2015, but it called for a new election and won 49.5 per cent of the vote five months later. Three years later, in June 2018, Erdogan won the presidential election with 52.6 per cent. In Poland, a majority of survey respondents thought that the government was ‘performing badly’ when it began to tinker with the Constitutional Court in 2015, but two years later, in October 2017, it received a positive evaluation from the majority (Kantar Public). In Hungary, Fidesz and its allies won a re-election in April 2018 with 44.9 per cent of the vote. In Venezuela, Chavez won a re-election in 2006 with 62.8 per cent of the vote and again in 2102 with 55.1 per cent. He enjoyed majority support in the polls and the opposition became majoritarian only after his death (Venezuelabarometro). Also in the United States, President Trump’s popularity hovers narrowly around 40 per cent, regardless of everything. The implication must be that either many people do not care at all about democracy, or they fail to see the consequences for democracy when they vote or when they answer survey questions.

1. Could It Happen Here?

All conclusions must be speculative. But the optimism that citizens might effectively threaten governments that commit transgressions against democracy and thus prevent them from taking this path (Montesquieu 1995: Book XIX, Chapter XIX [6]; Weingast 1997, 2015; Fearon 2011) is sadly unfounded. These views are based on an assumption that the government commits some acts that flagrantly threaten liberty, violate constitutional norms, or undermine democracy. But when a government proceeds by stealth, citizens turn against it, only if they see what its actions are leading to in the long run. Hence, resistance against a backsliding government is a difficult challenge for individual citizens.

The effect of stealth is to obscure the danger. And if the opposition does not stop the government from taking a particular series of legal steps, it will be too late to prevent it from taking illegal ones.

Adam Przeworski,
Department of Politics, New York University
18 March 2019

[1] Parts of this paper are based on joint work with Zhaotian Luo. For comments and suggestions we thank John Ferejohn, Joanne Fox-Przeworski, Diego Gambetta, Roberto Gargarella, Steven Holmes, Beatriz Magaloni, Bernard Manin, Pasquale Pasquino, Milan Svolik and Andrea Vindigni.

Footnotes:
[2] The definition of ‘victory’ is not as simple as it may appear. Electoral laws play an important role: in Turkey, the AKP won 34.3 per cent of votes to obtain 66 per cent of seats when it first assumed power in 2002; in Hungary, Fidesz won 53 per cent of votes and 68 per cent of seats in 2010; in Poland, PiS got 37.5 per cent of votes and 51 per cent of seats. Chavez’s ascension to office in Venezuela was convoluted: the traditional parties actually won the legislative election of 1998, while Chavez won the presidential election with 56.4 per cent. The referendum for a new constitution was passed by 71.8 per cent. Then Chavez won a new election with 59.8 per cent, while his party gained 44.4 per cent of votes and 55.7 per cent of seats in the legislative election of 2000.
[3] We are most grateful to Milan Svolik for forcing us to refocus the original version of this paper by referring to this story.
[4] Thus Landau (2013: 195) considers ‘the use of mechanisms of constitutional change in order to make a state significantly less democratic than it was before’ as ‘abusive constitutionalism,’ while the Colombian Constitutional Court ruled that even duly adopted constitutional amendments can be unconstitutional (Ginsburg and Huq 2018b: 188). Strauss (2018) argues that courts should be looking for consequences down the line when adjudicating particular cases.
[5] Note that in Turkey per capita incomes grew at the annual rate of 4.4 per cent under the AKP government, Venezuela enjoyed spectacular growth between 2004 and 2011 (except for 2009) due to oil prices, Hungary grew at the rate of 3.5 per cent under Fidesz, and Polish incomes continue to grow under PiS. (Data from Penn World Table 9.0, end in 2014).
[6] If power were to be abused, Montesquieu hoped, ‘everything would be united against it’, there would be a revolution, ‘which would not change the form of government or its constitution: for revolutions shaped by liberty are but a confirmation of liberty’.

Bibliography:
D. Austen-Smith, ‘Strategic Models of Talk in Political Decision Making’, International Political Science Review 13, 1992, 45–58.
N. Bermeo, ‘On Democratic Backsliding’, Journal of Democracy, 27/1, 2016.
Fearon, James. 2011. ‘Self-enforcing democracy’, Quarterly Journal of Economics 126: 1661–708.
Ferejohn, John and Pasquale Pasquino. 2003. ‘Rule of Democracy and the Rule of Law’, in Jose Marfa Maravall and Adam Przeworski (eds.), Democracy and the Rule of Law. New York: Cambridge University Press. Pages 242–60.
Ginsburg, Tom and Aziz Z. Huq. 2018a. ‘How to Lose a Constitutional Democracy’ UCLA Law Review 65(1): 78–169.
Ginsburg, Tom and Aziz Z. Huq. 2018b. How to Save a Constitutional Democracy. Chicago: University of Chicago Press.
Goldsmith, Jack. 2018. ‘Paradoxes of the Deep State’, in Cass R. Sunstein (ed.), Can It Happen Here? New York: HarperCollins. Pages 105–34.
Graham, Benjamin A.T., Michael K. Miller, and Kaare Strøm. 2017. ‘Safe-guarding Democracy: Powersharing and Democratic Survival’, American Political Science Review 111(4): 686–704.
Graham, Matthew, and Milan W. Svolik. 2018. ‘Democracy in America? Partisanship, Polarization, and the Robustness of Support for Democracy in the United States’, Unpublished Manuscript, Yale University.
Kantar Public. 2018. ‘Ocena działalności rządu, premiera i prezydenta.’ [Assessment of the work of the government, prime minister and president], online K002/18, January 2018.
Landau, David. 2013. ‘Abusive Constitutionalism’, University of California Davis Law Review 47: 189–260.
Lust, Ellen and David Waldner. 2015. ‘Unwelcome Change: Understanding, Evaluating and Extending Theories of Democratic Backsliding’, pdf.usaid.gov/pdf_docs/PBAAD635.pdf.
Maeda, Ko. 2010. ‘Two Modes of Democratic Breakdown: A Competing Risks Analysis of Democratic Durability’, Journal of Politics 72(4): 1129–43.
Magaloni, Beatriz. 2017. ‘Thought Piece on How Democracies Fail’, Paper presented at the conference How do Democracies Fall Apart (And Could it Happen Here?), Yale University, October 6.
Minow, Martha. 2018. ‘Could Mass Detentions Without Process Happen Here?’, in Cass R. Sunstein (ed.), Can It Happen Here? New York: HarperCollins. Pages 313–28.
Sadurski, Wojciech. 2018. ‘How Democracy Dies (in Poland): A Case Study of Anti-Constitutional Populist Backsliding’, Legal Studies Research Paper No. 18/01. Sydney Law School.
Scheppele, Kim Lane. 2013. ‘The Rule of Law and the Frankenstate: Why Governance Checklists Do Not Work’, Governance 26: 559–62.
Strauss, David A. 2018. ‘Law and the Slow-Motion Emergency’, in Cass R. Sunstein (ed.), Can It Happen Here? New York: HarperCollins. Pages 365–86.
Svolik, Milan. 2015. ‘Which Democracies Will Last? Coups, Incumbent Takeovers, and the Dynamic of Democratic Consolidation’, British Journal of Political Science 45(4): 715–38.
Svolik, Milan W. 2017. ‘When Polarization Trumps Civic Virtue: Partisan Conflict and the Subversion of Democracy by Incumbents’, Working Paper, Department of Politics, Yale University.
Varol, Ozan O. 2015. ‘Stealth Authoritarianism’, Iowa Law Review 100 : 1673–1742.
Weingast, Barry R. 1997. ‘Political Foundations of Democracy and the Rule of Law’, American Political Science Review 91: 245–63.
Weingast, Barry R. 2015. ‘Capitalism, Democracy, and Countermajoritarian Institutions’, Supreme Court Economic Review 23: 255–77.

This work is licensed under a Creative Commons Attribution (BY) 3.0 Unported License (users are free to share and adapt the material providing attribution is made to the author and source)