Download PDF

On Democratic Backsliding

Democratic backsliding (meaning the state-led debilitation or elimination of the political institutions sustaining an existing democracy) has changed dramatically since the Cold War. Open-ended coups d’état, executive coups, and blatant election-day vote fraud are declining while promissory coups, executive aggrandizement and strategic electoral manipulation and harassment are increasing. Contemporary forms of backsliding are especially vexing because they are legitimated by the very institutions democracy promoters prioritize but, overall, backsliding today reflects democracy’s advance and not its retreat. The current mix of backsliding is more easily reversible than the past mix and successor dictatorships are shorter-lived and less authoritarian.

Scholars have devoted huge amounts of attention to explaining why democracies break down, but systematic and explicitly comparative work on precisely how they break down has been less common. Political scientists have focused more often on economic and institutional correlates than on choices and choosers, even though these may be more amenable to direct influence and rapid intervention.

What kinds of concrete actions transform a regime from one type to another? Which techniques of transformation are most common? Analyzing what has come to be known as democratic backsliding moves us toward answers to these questions, for it forces us to focus on the actual choices that change regimes.

The term democratic backsliding is frequently used but rarely analyzed. This explains why a careful recent survey concluded “we know very little” about it.1 Part of the problem is the term’s extraordinary breadth. At its most basic, it denotes the state-led debilitation or elimination of any of the political institutions that sustain an existing democracy. Since the political institutions that sustain democracy are myriad (including all the institutions that enable people to formulate and signify preferences and then have them weighed by their elected representatives), the term embraces multiple processes. Since the state actors who might initiate backsliding are themselves diverse (ranging from monarchs to presidents to military men), the term embraces multiple agents. In sum, the concept has so many referents that it needs immediate specification to have practical meaning. Like an old steamer trunk, it is opaque and unwieldy but yields much that proves useful when it is unpacked.

This essay unpacks the concept of democratic backsliding by exploring [End Page 5] six of its major varieties. It illustrates that forms have varied in frequency over time; that some of the most blatant forms of backsliding are now less common; and that more vexing forms of backsliding are becoming more common. Ironically, we now face forms of democratic backsliding that are legitimated through the very institutions that democracy promoters have prioritized. Overall, trends in backsliding reflect democracy’s slow progress and not its demise.

A close historical look at the varieties of backsliding reveals that the classic open-ended coups d’état of the Cold War years are now outnumbered by what I call promissory coups; that the dramatic executive coups of the past are being replaced by a process that I call executive aggrandizement; and finally, that the blatant election-day vote fraud that characterized elections in many developing democracies in the past is being replaced by longer-term strategic harassment and manipulation.

According to the Oxford English Dictionary, the earliest use of the English word “backsliding” dates from 1554, when the Scottish Protestant theologian John Knox (1513–72) employed it in a pamphlet entitled A Faythfull Admonition to the Professors of God’s Truth in England. More famously, “backsliding” also appears in the King James Bible (1611), where it translates a prophetic plea (Jeremiah 3:22) for Israel to drop its “faithless” or “wayward” habits in order to resume a relationship of loyalty to God. When linked with the word democratic, the term’s current secular meaning is in keeping with its origins in that it denotes a willful turning away from an ideal. But where does backsliding from democracy lead?

Backsliding can take us to different endpoints at different speeds. Where backsliding involves rapid and radical change across a broad range of institutions, it leads to outright democratic breakdown and to regimes that are unambiguously authoritarian. Where backsliding takes the form of gradual changes across a more circumscribed set of institutions, it is less likely to lead to all-out regime change and more likely to yield political systems that are ambiguously democratic or hybrid. Democratic backsliding can thus constitute democratic breakdown or simply the serious weakening of existing democratic institutions for undefined ends. When backsliding yields situations that are fluid and ill-defined, taking action to defend democracy becomes particularly difficult.

Positive Trends

Democratic backsliding has changed dramatically since the Cold War. Three of the most dramatic and far-reaching varieties of backsliding seem to be waning. Coups d’état, executive coups by elected leaders, and blatant election-day vote fraud all have declined in frequency.

The decline of classic coups d’état

Coups are illegal attempts by military or other state elites to oust a sitting executive. Historical analysis shows [End Page 6] a dramatic decline in all coups and especially the open-ended military coups that gave rise to long-lasting and brutal dictatorships during the Cold War. As Figure 1 shows, the probability that a democracy will be targeted by any sort of coup has dropped dramatically. The probability reached a thirty-year low after 1995, and although it rose slightly as the first decade of the new century ended, it is still significantly less than it was during the 1960s.

Figure 1. C). A detailed explanation of the axes for this figure and how the data sources were used may be found at .
Click for larger view
View full resolution
Figure 1.

Coup Frequency in Democracies, 1950–2014

Source: Regime data are from Polity IV, supplemented with Polity IV data modified by Kristian Skrede Gleditsch (see endnote 6). Coup data are from Jonathan Powell and Clayton Thyne’s “Coups d’état, 1950 to Present” dataset (inline graphic). A detailed explanation of the axes for this figure and how the data sources were used may be found at

The likelihood of a democratic government being the target of a successful coup has also declined markedly, dropping to nearly zero in the early 2000s. Though it has recently risen slightly, the drop in the success rate that began during the Cold War has not been reversed.

The decline of executive coups

Alongside the decline of classic coups d’état there has been a decline in executive coups. These “self-coups” or autogolpes involve a freely elected chief executive suspending the constitution outright in order to amass power in one swift sweep. Executive coups associated with dictatorships (such as that of Ferdinand Marcos [1965–86] in the Philippines) were fairly common during the Cold War and the decade after its end. During the 1990s, there were a full five executive coups—in Peru in 1992 (under Alberto Fujimori), in Armenia in 1995 (under Levon Ter-Petrosian), in Belarus in 1995 (under Alyaksandr Lukashenka), in Zambia in 1996 (under Frederick Chiluba), and in Haiti in 1999 (under René Préval).2 Since then, happily, the number of executive coups in democracies has plummeted: Between 2000 and 2013, Niger was the only democracy in the world to experience an executive coup.

The decline of election-day vote fraud

Alongside the declines in these two types of coup-based backsliding, there has also been a reported drop in blatant election-day vote fraud. Electoral malpractice as [End Page 7] a whole has not diminished, but there is near-consensus that open fraud on election day has decreased. Figure 2 shows the trend. The extent to which this trend is being driven by normative change, the rise of election monitoring, or the deterrent effects of parallel vote tabulation remains under debate, but the decline itself has been widely noted. Seasoned election observers report that cheating has “become more subtle,” that “blatant manipulation on election day seems less and less common,” and that fraud in polling stations has been reduced.3 A recent study of African elections found that count falsification, ballot-stuffing, and ballot-box fraud were relatively rare and that “the vote count process was the most highly regarded dimension in the whole electoral process.”4 In the words of another firsthand observer, “Today, only amateurs steal elections on election-day.”5

Figure 2. V-F A W M P-1975 D, 1991–2012. Additional election data are from Susan Hyde’s NELDA project (). A detailed explanation of the axes for this figure and how the data sources were used may be found at .
Click for larger view
View full resolution
Figure 2.

Vote-Fraud Allegations by Western Monitors in Post-1975 Democracies, 1991–2012

Source: See source note in Figure 1. Additional election data are from Susan Hyde’s NELDA project ( A detailed explanation of the axes for this figure and how the data sources were used may be found at

Continuing Challenges

The decline in the three varieties of backsliding outlined above is certainly gratifying. Unfortunately, other varieties of democratic backsliding either remain unchanged or are on the rise. These have been understudied and merit our immediate attention.

Promissory coups

A first persistent variety of backsliding involves what might best be called promissory coups. Promissory coups frame the ouster of an elected government as a defense of democratic legality and make a public promise to hold elections and restore democracy as soon as possible. Whereas Cold War coupmakers usually cast their seizures of power as open-ended, most coupmakers today emphasize the temporary nature of their intervention and frame it as a necessary step [End Page 8] toward a new and improved democratic order. The share of successful coups that falls into the promissory category has risen significantly, from 35 percent before 1990 to 85 percent afterward.6

Analyzing the aftermath of the twelve successful promissory coups that took place in democracies between 1990 and 2012, we see a dismal picture. Few promissory coups were followed quickly by competitive elections, and fewer still paved the way for improved democracies.

The general who spearheaded the 1991 Haitian coup blandly called it “a correction of the democratic process,”7 but military violence soon showed that the promise of elections was never going to be kept. Haiti’s freely elected president Jean-Bertrand Aristide finally returned in 1994 to complete his term—but only as part of a costly six-month international intervention that reversed the coup.

In each of the other cases, coupmakers did hold elections, but the lapse of time between coups and balloting varied widely. A vote came just five months after the 2009 putsch in Honduras, but took more than six years following the coups in Gambia (1994), Pakistan (1999), and Fiji (2006). Regardless of their timing, the elections that follow promissory coups turn out to be surprisingly favorable to those who backed the coups in the first place. Fully half of all the postcoup elections that Western observers rated acceptable were won either by the actual coup perpetrators or their favored candidates. Elections are not a reliable route to democratic reinstatement.

The electoral victories of coupmakers and their allies were not limited to a particular period or region. The May 2000 coup against the multiracial government of Mahendra Chaudhry in Fiji was followed by September 2001 elections, but the winner was the civilian politician whom the military had handpicked to head the interim government. Gambia’s October 2001 elections yielded a victory for Yahya Jammeh, the military officer who had led the coup against Gambia’s elected government in 1994. Though Jammeh had organized sham elections in 1996 and 1997, he won the 2001 polls with 53 percent of the vote in an election that EU and Commonwealth observers deemed free and fair.

The 2009 elections following the coup that ousted President Manuel Zelaya in Honduras were questionable since the candidate allied with Zelaya withdrew while Zelaya himself organized a boycott, but many still read the poll as a win for the coup coalition. With turnout just 5.4 percentage points below what it had been for Zelaya’s election in 2005, National Party presidential candidate Porfirio Lobo won nearly 57 percent, while his party gained an absolute legislative majority. The National Party had been a key player in the coup coalition.

Madagascar and Mali both voted in 2013, and these ballotings too were coupmakers’ triumphs. Madagascar had taken more than four years to hold elections after its 2009 coup. In the interim, both the coup leader and the elected president whom he unseated had been banned from running. [End Page 9] When surrogates took their respective places, the surrogate for the coup leader won 54 percent in the December 2013 runoff, while his party won a plurality of seats in Parliament. The presidential elections held after military officers toppled Mali’s democracy in March 2012 were won by Ibrahim Boubacar Keïta, a civilian politician whose brief membership in an anticoup coalition did not eclipse his longtime status as a military favorite. In August 2013, he won the presidency in a runoff landside, while his party took an overwhelming legislative majority in November.

In a more recent case, Fiji’s 2014 postcoup parliamentary election was won by Frank Bainimarama, the very officer (and onetime commander of Fiji’s tiny patrol-boat navy) who had headed the 2006 putsch. Though he ruled by decree for over seven years, his Fiji First party won 59 percent of the vote and 64 percent of the seats in Parliament.

Had coupmakers and their allies fulfilled their promises for improved democracy, this subtype of backsliding might be said to have an upside. But an example of democratic deepening after a coupmakers’ victory is yet to be found. It is too early to know if the new government in Fiji will be able to claim such an achievement. Regrettably, none of the other cases mentioned above has even matched (let alone exceeded) the level of freedom they enjoyed before their coups.8

The promised improvement of democracy has remained elusive even when coup opponents have won postcoup elections. Promissory coups in Lesotho (1994), Niger (1996), Pakistan (1999), Thailand (2007), and Guinea-Bissau (2012) were all followed by elections in which coup opponents proved victorious. Yet in only one case—that of Lesotho—has a substantial improvement in political and civil rights been recorded. In 2003, Freedom House’s rating system moved Lesotho from Partly Free to Free, but Guinea-Bissau, Niger, and Pakistan remain Partly Free and even this status remains precarious.

Freedom House deemed Thailand a Free country before its promissory coup in 2006, but it has since returned to dictatorship. The coup coalition that ousted the freely elected government of Thaksin Shinawatra made good on its promise to hold free elections (in December 2007) and even allowed Thaksin allies to regain power through the ballot box. But tolerance was short-lived. The military seized power again in May 2014, and, ominously, made no promise of elections at all. Unlike other forms of backsliding, promissory coups sometimes raise expectations at home and abroad, but these expectations are nearly always dashed.

Executive aggrandizement

Executive aggrandizement contrasts with all forms of coupmaking in that it takes place without executive replacement and at a slower pace. This more common form of backsliding occurs when elected executives weaken checks on executive power one by one, undertaking a series of institutional changes that hamper the power of opposition forces to challenge executive preferences. The [End Page 10] disassembling of institutions that might challenge the executive is done through legal channels, often using newly elected constitutional assemblies or referenda. Existing courts or legislatures may also be used, in cases where supporters of the executive gain majority control of such bodies. Indeed, the defining feature of executive aggrandizement is that institutional change is either put to some sort of vote or legally decreed by a freely elected official—meaning that the change can be framed as having resulted from a democratic mandate.

Executive aggrandizement occurs in a broad range of countries. The career of Turkey’s former premier (now president) Recep Tayyip Erdoğan and his Justice and Development Party (AKP) provides an illustrative example. Erdoğan led his party to a resounding victory in the 2002 national elections and then attracted increasing shares of the vote in both 2007 and 2011. The AKP’s strength in parliament provided the infrastructure for the “quiet revolution” that Erdoğan promised his supporters, enabling the passage of a record number of new laws (including more than five-hundred during his first two years in office).9

Many of these laws undercut institutions of accountability. Media freedoms and judicial autonomy became prime sites for democratic backsliding. In 2004, for example, the government revised the penal code to allow the criminal prosecution of journalists for discussing any subject deemed controversial by state authorities. Later came a series of defamation laws, both civil and criminal, that the state (and Erdoğan himself) began using widely to silence critics. Other laws facilitated the blocking of websites and the identification of Internet users, while still others allowed the Radio and Television Supreme Council to forbid coverage of certain issues altogether. Because media outlets are so often owned by holding companies dependent on government contracts, journalists must choose between free expression and having a job. During the 2013 Gezi Park protests alone, more than eighty journalists were fired.10

Turkey’s judicial system has been a site for executive aggrandizement as well. In 2010, Erdoğan passed two-dozen constitutional changes via national referendum. The president received power to name fourteen of the seventeen Constitutional Court judges,11 while decisions about which parties are legal and allowed to field candidates for office were shifted from the courts to the legislature. In 2014, the government passed legislation giving the justice minister power to directly appoint members to the High Council of Judges and to control the inspection board that disciplines judges. Within six months, more than three-thousand sitting judges had been removed.12 The courts suffered another blow from a law that gave the National Intelligence Organization (headed by a presidential appointee) power to collect “all information, documents or data from any entity in Turkey” without having to seek judicial permission or submit to judicial review.13

All these changes were made by democratically elected officials with a strong popular mandate to rule. Because many of the new measures [End Page 11] challenged military and civilian elites with less than perfect democratic credentials of their own, they cut through the old order with what even critics describe as “a democratizing edge.”14

The same can be said of many of the initiatives taken by President Rafael Correa in Ecuador. Like Erdoğan, Correa has changed the institutions of democracy in basic ways, but always with an electoral mandate that has averaged almost 56 percent across three elections spanning 2006 to 2013.

True to his vow to lead a “Citizens’ Revolution,” Correa has made profound changes in a panoply of democratic institutions. These began with his 2007 initiative to hold elections for a Constitutional Assembly. A stunning 82 percent of the electorate endorsed his proposal in a referendum and a weighty 64 percent of the public endorsed the new constitution in a 2008 plebiscite. Correa cautioned that he and his supporters “had won the elections, but not power,” and he immediately set out to consolidate the latter.15

Correa convinced the newly elected Constitutional Assembly to force the seated Congress into permanent recess and to assume legislative functions itself. Many of the established parties, discredited by corruption and poor performance, never recovered. Correa sealed their fate through recentralizing measures meant to undermine conservative opposition elites in Guayaquil, and through changing the rules for licensing parties, drawing electoral districts, and allocating seats.16

Coupled with the undeniable success of a series of redistributive programs that led to unprecedented drops in poverty and income inequality, these legal initiatives crippled Correa’s legislative opposition. Correa’s party, Alianza País, won a whopping 73 percent of the seats in the 2013 legislative elections while its strongest competitor won less than 9 percent.17 Since the 2008 Constitution allows full amendments with a two-thirds majority, the legislature’s December 2015 vote to eliminate presidential term limits came as no surprise.

Correa has used his strong mandate to make other major changes as well. Banks and bank shareholders may not own media outlets; broadcast frequencies must operate in “the collective interest”; media outlets and journalists are legally liable if the information they disseminate is not deemed “true, verified, opportune [and] contextualized”; and no coverage may be aimed at “destroying the prestige of a natural or juridical person or reducing their public credibility.” Any citizens’ organization can be dissolved if the state deems it divorced from its original purpose, harmful for state security, or disruptive of the “public peace.”18

Though both Erdoğan and Correa are categorized as populists (and their parallels with Hugo Chávez and Viktor Orbán are obvious), leaders of varied ideological hues have engaged in executive aggrandizement. Between 2006 and 2008, President Abdoulaye Wade of Senegal curtailed protest rights, tampered with the electoral calendar, changed legislative rules to hamper a potential rival, and created a new upper house dominated by his own appointees. Before his February 2014 ouster, Ukraine’s President [End Page 12] Viktor Yanukovych meddled with the courts and police; maneuvered for a parliamentary supermajority by inducing floor-crossings while banning multiparty blocs; and engineered a return to the 1996 Constitution so that he could hire and fire governors and cabinet members (including the premier) on his own. We find further examples of executive aggrandizement across a range of countries as diverse as Sri Lanka and Mozambique.

Figure 3. O H L D, 1991–2010. Additional election data are from Susan Hyde’s NELDA project (). A detailed explanation of the axes for this figure and how the data sources were used may be found at .
Click for larger view
View full resolution
Figure 3.

Opposition Harassment and Leader Disqualification, 1991–2010

Source: See source note in Figure 1. Additional election data are from Susan Hyde’s NELDA project ( A detailed explanation of the axes for this figure and how the data sources were used may be found at

Manipulating elections strategically

Strategic election manipulation is a third form of backsliding. It too is on the rise, being often joined with executive aggrandizement. Strategic manipulation denotes a range of actions aimed at tilting the electoral playing field in favor of incumbents. These include hampering media access, using government funds for incumbent campaigns, keeping opposition candidates off the ballot, hampering voter registration, packing electoral commissions, changing electoral rules to favor incumbents, and harassing opponents—but all done in such a way that the elections themselves do not appear fraudulent. Strategic manipulation differs from blatant election-day vote fraud in that it typically occurs long before polling day and rarely involves obvious violations of the law. It is “strategic” in that international (and often domestic) observers are less likely to “catch or criticize” it.19

A number of important studies explain strategic manipulation as an unintended consequence of the rise of international election monitoring. They argue that politicians found new ways to ensure victory once better monitoring made straight-up fraud “more costly.”20 Whatever the explanation, scholars agree that much if not most election-related backsliding now occurs before election day.21 There is also widespread agreement that electoral misconduct “is not declining in the aggregate.”22 Blatant election-day fraud is rarer, but other and subtler forms have filled in. Figure 3 above illustrates [End Page 13] the increase in opposition harassment and the use of legal maneuvers to exclude opposition leaders from presidential races.

Confronting Backsliding

A change in pace

The decline of coups means that de-democratization today tends to be incremental rather than sudden. Dramatic breakdowns will probably still occur, but troubled democracies are now more likely to erode rather than to shatter—to decline piece by piece instead of falling to one blow. Democratic erosion may be better than democratic cataclysm because it is less likely to be violent, but incremental decline still presents us with important challenges.

A first challenge is scholarly. Social science has focused mostly on clear cases of democratic collapse—paying “scant attention” to the “incremental”23 regime changes that color many countries’ histories. Research on “hybrid” regimes has been a step forward, but we need to know more about how the slide backward into hybridity takes place.24 Focusing on democratic erosion will require more scholars to see that democracy is “a collage” of institutions crafted and recrafted by different actors at different times.25 It is put together piece by piece, and can be taken apart the same way. Politicians who engage in executive aggrandizement and strategic electoral manipulation already know this. Political scientists must learn it too, or risk their own slide into irrelevance.

Incremental forms of backsliding create profound political challenges as well. Domestically, alterations in electoral laws, district boundaries, electoral commissions, and voter-registration procedures may seem too arcane to be the stuff of mass mobilization. Court-packing and media restrictions are probably easier to frame as dangers to democracy, but the jurists and journalists who are likely to mobilize in opposition to these maneuvers can easily be counterframed as “special interests” or tools of a discredited old order. Civic organizations representing disadvantaged groups of other sorts can be framed and silenced as tools of foreign forces. The fact that they often are funded from abroad makes this especially likely and effective. Piecemeal erosions of autonomy may thus provoke only fragmented resistance.

At a more general level, slow slides toward 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. Executive aggrandizement takes place precisely where a majority that supports it is already taking root. Strategic electoral manipulation takes place where incumbents already deem themselves capable of either securing or reinforcing majority support. Since both forms of backsliding emerge precisely where oppositions are already weakened by performance failures and internal divisions, mustering the power of numbers to reverse them is especially hard. Even when opposition leaders succeed in mobilizing mass action against a stolen election, their success is often heavily dependent on foreign allies.26 [End Page 14]

Rational responses

Whatever means they muster, opponents of backsliding are more likely to be successful if they recognize that current trends are not random events but rational responses to local and international incentives. The state actors who choose to restructure their institutional environment are often acting defensively to prevent reenactments of past assaults. It is significant, for example, that each of Rafael Correa’s three elected predecessors was driven from office early—undone by fierce media criticism, spectacular civil society protests, and the decisions of legislators and jurists. Correa’s aggrandizing policies—like those of Erdoğan in Turkey—are predictable reactions to recent history and to an institutional landscape that was already deeply troubled.

Today’s trends in backsliding are rational reactions to international incentives as well as domestic history. The emphasis that democracy promoters put on elections, the rule of law, and other institutions deemed essential to democracy are all reflected in the stories told here. Indeed, today’s most popular varieties of backsliding offer ironic proof of democracy promotion’s partial success. There is strong evidence that the aid sanctions that now follow the illegal ouster of democratically elected governments have driven down the number of coups.27 The few successful coups that now take place usually frame themselves as promissory precisely to escape or limit these sanctions.

We have already outlined how the international resources and prestige associated with clean elections have forced electoral malpractice to become more subtle. Here too is testimony to democracy promotion’s impact: If elections were not widely seen as “the only game in town,” politicians would not spend scarce resources trying to manipulate them.

The partial embrace of the liberal-democratic project is also reflected in executive aggrandizement. The push for media that are free, but also private (and commercially competitive), has put elected executives under increased scrutiny. The push to strengthen parties and legislatures has made the creation of a legislative majority a more critical task for chief executives.28 The push to make civil society more active, plus the wide sense of reverence toward the spectacular mass mobilizations that felled dictators in the 1980s and 1990s, has given citizens everywhere lasting models and leaders everywhere lasting worries. Current attempts to control the media, legislative majorities, and associational life are perfectly predictable, especially in highly polarized polities.

Vexing ambiguity

A third quality to reckon with as we weigh contemporary forms of backsliding is their profound ambiguity. We now face forms of democratic backsliding that are legitimated through the very institutions that democracy promoters have prioritized: national elections, voting majorities in legislatures and courts, and the “rule” of the laws that majorities produce.

Trying to deploy what Steven Levitsky and Lucan Way call “leverage” is especially vexing under these conditions. First, challenging laws [End Page 15] crafted by democratically elected executives and legislatures is highly risky: Foreign pressure will inevitably be seen as an affront to sovereignty. Second, proving that a change in institutions has a nefarious purpose is often difficult. Changes in laws governing elections, communications, or even associations occur routinely in established democracies and are not, in themselves, antidemocratic. Third, not all the targets of backsliding’s institutional changes are unambiguously democratic themselves. Defending institutions that shield corrupt politicians, ruthless media moguls, or associations that perpetrate hatred may or may not foster democracy in the long run, but will surely weaken the defenders’ credibility in the short run.

The tactical challenges posed by ambiguity also emanate from promissory coups. Every such coup seeks to legitimate its assault on a flawed democracy by vowing to produce a better democratic regime. Moreover, promissory coups often occur with the backing of jurists, legislators, and mobilized citizens. The pledge of future elections plus the often dubious behavior of the targeted executive makes responding to these coups especially difficult.

Sanctions are often effective when imposed, but these are sometimes lifted or lightened due to security concerns. This is very likely to happen now, as the War on Terror grows in scope, and it may explain why the number of successful coups has recently increased: Though the 2013 military putsch in Egypt took place in a nondemocratic regime and was thus, strictly speaking, not a “promissory coup,” the West’s muted reaction to the ouster of Egypt’s freely elected president highlights a deeply troubling backsliding-security tradeoff and the depth of ambiguity we now confront.

Contemporary forms of democratic backsliding are most ambiguous and most difficult when they marshal broad popular support—and they often do. As recent events in Thailand illustrate, huge numbers of citizens may support an elected official’s unlawful removal. This uncomfortable truth means that those seeking to reverse backsliding must cope not only with the state actors who engineer backsliding but with their mobilized supporters. Silencing or simply ignoring these citizens’ preferences may stoke reactionary fires and undercut the quality of democracy. Yet changing their preferences is devilishly difficult and a long-term project at best.

The policy challenge may be greatest when supporters of backsliding have a credible democratizing agenda of their own. This can occur when the impetus for institutional change comes from marginalized groups that rise up to demand a more inclusive and responsive democratic model, as indigenous groups in the Andean region did during the era of neoliberal reform. Democratic backsliding is the weakening or disassembling of a given set of democratic institutions. Thus it can sometimes occur with the intention of deepening rather than destroying democracy. As we struggle to craft helpful responses to backsliding, we must consider what motivates the citizens involved. Each country’s experience requires an individual, historically informed response. [End Page 16]

Figure 4. T D D A S R, 1950–2009. A detailed explanation of the axes for this figure and how the data sources were used may be found at .
Click for larger view
View full resolution
Figure 4.

The Declining Durability and Authoritarianism of Successor Regimes, 1950–2009

Source: See source note in Figure 1. A detailed explanation of the axes for this figure and how the data sources were used may be found at

A better mix

The challenges summarized above are certainly daunting but not insurmountable. The mix of backsliding we see today is preferable to the mix of the past. Incremental and ambiguous change preserves mixed landscapes wherein one set of institutions or ideas can correct others. As long as some electoral competition takes place, power can be clawed back. When civil society is allowed some space, countermobilization can occur. Because backsliding reflects incentive structures, changed incentives can reverse negative trends. We are seeing all these factors at work at the time of this writing as 2015 draws to a close. In Venezuela, countermobilization, hard campaigning, and economic crisis have enabled opposition forces to defeat the party founded by Hugo Chávez and win a supermajority in the national legislature. In Ecuador, mass mobilizations, local-election defeats, changed economic incentives, and the desire to “not go down in history” as seeking to “perpetuate himself in power” have led Rafael Correa to announce that he will not seek the presidency in 2017.29 The drama in both cases is still unfolding, but the changes are meaningful. The possibilities for reversing backsliding before an unambiguous regime change occurs are real.

But the best news about contemporary forms of backsliding concerns what happens when they do lead to regime change. The varieties of backsliding that lead to breakdown and autocracy today produce outcomes that are less onerous than those produced by backsliding in the past.

As Figure 4 shows, there has been a marked decrease in the length of time that a country with a toppled democracy stays autocratic. Moreover, the dictatorships that follow failed democracies today are, on average, less authoritarian than their predecessors. How these positive trends are related to changes in the varieties of backsliding remains an open question. But the decline in the most egregious forms of backsliding [End Page 17] and the decline in the longevity and brutality of successor regimes suggest that democracy’s prospects are still good. More systematic thinking about how to cope with backsliding will make them even better.

Nancy Bermeo

Nancy Bermeo is the Nuffield Chair of Comparative Politics at Oxford University and PIIRS Senior Scholar at Princeton University. Her most recent book (coedited with Deborah Yashar) is titled Parties, Movements and Democracy in the Developing World (Cambridge University Press, 2016).


The author thanks Alexander Gard-Murray, Mario Rebelo, and Adam Brodie for their research assistance.

1. Ellen Lust and David Waldner, “Unwelcome Change: Understanding, Evaluating, and Extending Theories of Democratic Backsliding,” U.S. Agency for International Development, 11 June 2015, 1.

2. See “Subversion by Ruling Executive (Adverse Regime Change)” in Monty G. Marshall and Donna Ramsey Marshall, “Coup d’Etat Events, 1946–2014: Codebook” (Center for Systemic Peace, 14 January 2015), 15–17,

3. Thomas Carothers, “The Rise of Election Monitoring: The Observers Observed,” Journal of Democracy 8 (July 1997): 22; Eric C. Bjornlund, Beyond Free and Fair: Monitoring Elections and Building Democracy (Baltimore: Johns Hopkins University Press, 2004), 143, 304; Susan D. Hyde, The Pseudo-Democrat’s Dilemma: Why Election Observation Became an International Norm (Ithaca: Cornell University Press, 2011), 42; and Daniela Donno, Defending Democratic Norms: International Actors and the Politics of Electoral Misconduct (New York: Oxford University Press, 2013), 27.

4. Max Grömping and Ferran Martínez i Coma, “Electoral Integrity in Africa,” Electoral Integrity Project, 2015, 22,

5. Diplomat quoted in Brian Klaas, “Bullets over Ballots: How Electoral Exclusion Causes Coups d’Etat and Civil Wars” (PhD diss., Oxford University, 2015), 2.

6. Author’s coding available at Originally from Marshall and Marshall, “Coup d’Etat Events, 1946–2014”; and Kristian Skrede Gleditsch, Modified Polity P4 and P4D Data, Version 4.0, 2013,

7. Tom Harvey, “Junta Chief: Aristide May Return,” Sun-Sentinel (Florida), 6 October 1991.

8. Freedom House, “Individual Country Ratings and Status, FIW 1973–2015,”

9. Sultan Tepe, “Turkey’s AKP: A Model ‘Muslim-Democratic’ Party?” Journal of Democracy 16 (July 2005): 69–82.

10. Testimony by Nate Schenkkan at hearing on “The Future of Turkish Democracy” before the Subcommittee on Europe, Eurasia and Emerging Threats, Foreign Affairs Committee of the U.S. House of Representatives, 15 July 2014, 113th Cong, 2,

11. Berna Turam, “Turkey Under the AKP: Are Rights and Liberties Safe?” Journal of Democracy 23 (January 2012): 109–18.

12. Yesim Arat, “The Decline of Democracy in Turkey: Popular Support, Institutional Decay,” unpubl. ms., Princeton University, March 2015. [End Page 18]

13. Schenkkan Testimony, 3.

14. Turam, “Are Rights and Liberties Safe?” 111.

15. Catherine M. Conaghan, “Ecuador: Correa’s Plebiscitary Presidency,” Journal of Democracy 19 (April 2008): 47.

16. Kent Eaton, “The Centralism of ‘Twenty-First-Century Socialism’: Recentralising Politics in Venezuela, Ecuador and Bolivia,” Journal of Latin American Studies 45 (August 2013): 436–41. See also “Ecuador Politics: Eleven Make the Cut,” Latin American Weekly Report, 18 October 2012, 4–5,

17. Jason Eichorst and John Polga-Hecimovich, “The 2013 Ecuadorian General Elections,” Electoral Studies 34 (June 2014), 364.

18. Catherine M. Conaghan, “Surveil and Sanction: The Return of the State and Societal Regulation in Ecuador,” European Review of Latin American and Caribbean Studies 98 (April 2015): 14–19.

19. Emily Beaulieu and Susan D. Hyde, “In the Shadow of Democracy Promotion: Strategic Manipulation, International Observers, and Election Boycotts,” Comparative Political Studies 42 (March 2009): 393.

20. Susan D. Hyde and Angela O’Mahony, “International Scrutiny and Pre-Electoral Fiscal Manipulation in Developing Countries,” Journal of Politics 72 (July 2010): 690–704.

21. Judith G. Kelley rejects the tradeoff argument in Monitoring Democracy: When International Election Observation Works, and Why It Often Fails (Princeton: Princeton University Press, 2012), but agrees on the latter point along with Sarah Birch, Electoral Malpractice (New York: Oxford University Press, 2011), 163; and Alberto Simpser and Daniela Donno, “Can International Election Monitoring Harm Governance?” Journal of Politics 74 (April 2012): 501–13.

22. Donno, Defending Democratic Norms, 56.

23. Lust and Waldner, “Unwelcome Change,” 4; José Alemán and David Yang, “A Duration Analysis of Democratic Transitions and Authoritarian Backslides,” Comparative Political Studies 44 (September 2011): 1125.

24. For work on democratic erosion, see Guillermo O’Donnell, “Democracy’s Future: Do Economists Know Best?” Journal of Democracy 6 (January 1995): 27; Scott Mainwaring and Aníbal Pérez-Liñán, “Cross-Currents in Latin America,” Journal of Democracy 26 (January 2015): 114–18; Larry Diamond, “Facing Up to the Democratic Recession,” Journal of Democracy 26 (January 2015): 147.

25. Nancy Bermeo, “Interests, Inequality, and Illusion in the Choice for Fair Elections,” Comparative Political Studies 43 (August–September 2010): 1120.

26. Donno, Defending Democratic Norms, 184.

27. Nicolay Marinov and Hein Goemans, “Coups and Democracy,” British Journal of Political Science 44 (October 2014): 799–825.

28. Aníbal Pérez-Liñán, Presidential Impeachment and the New Political Instability in Latin America (New York: Cambridge University Press, 2007).

29. “Se pueden agregar textos a enmiendas, dice Rafael Correa,” El Universo (Guayaquil), 21 November 2015, [End Page 19]