The Fates of Third-Wave Democracies
This article offers the first comprehensive analysis of the outcomes of third-wave democratic transitions. It charts where democracy broke down, stagnated, advanced, and eroded, and it investigates the initial starting conditions associated with democratic deepening and breakdown. The findings are sobering: Among the 91 new democracies that (by our count) emerged from 1974 to 2012, 34 experienced breakdowns, often in short order. In 28 cases, democracy stagnated after transition, usually at a fairly low level, and in two more it eroded. Democracy advanced relative to the starting point in only 23 cases. Few countries have succeeded in creating robust liberal democracies. Regimes that started off with a higher level of liberal democracy, that were geographically surrounded by democracies, and that experienced better rates of economic growth were less likely to break down. Regimes that started off with a lower per capita GDP and those that experienced lower economic growth, as well as regimes that started off with a higher level of liberal democracy, were less likely to deepen democracy.
Since the "third wave" of democratization began in 1974, analysts have devoted much attention to studying the birth of new democracies. But how have these regimes fared over the longer term? Beyond the up-or-down question of survival versus breakdown, new democracies may encounter a variety of fates. These democracies may deepen, with once-fragile institutions increasingly safeguarding a broad panoply of liberal-democratic rights; they may remain mired in long-term struggles to overcome illiberal practices and institutional dysfunction; and they may slip toward the borderline of competitive authoritarianism or break down completely. Without considering the full range of outcomes, it is impossible to fully grasp the achievements and the disappointments of the third wave.
Despite broad interest in global patterns of democratization, scholars have not yet systematically tracked the fortunes of all third-wave democracies. This article offers the first comprehensive analysis of the outcomes of all democratic transitions from 1974 to 2012.1 While many studies have focused on democratic transitions and breakdowns, our analysis also tracks changing levels of democracy in regimes that have steadily held democratic elections. In other words, we consider not only changes of regime—from democracy back to dictatorship—but also changes within democracies. In comparison to the conceptual work on transitions and breakdowns, frameworks for understanding what leads democracy to deepen, stagnate, or erode are far less developed. At a time of growing alarm over democratic backsliding around the globe, these questions deserve careful consideration. [End Page 99]
We survey posttransition trajectories by grouping the fates of the 91 democratic regimes that (by our count) emerged in 1974–2012 into five mutually exclusive and jointly exhaustive categories: democratic breakdowns, erosions, stagnations, advances, and cases where regimes have remained highly democratic without major advances. The results of the analysis are sobering. Among the 91 new democracies that emerged during the third wave,2 34 experienced breakdowns, often in short order. In 28 cases, democracy stagnated after the transition, usually at a fairly low level, and in two others it eroded. There have been some successes; 23 regimes achieved major democratic advances between their first year of democracy and 2017. Four others attained fairly high levels of democracy at the outset and were still high-level democracies in 2017. Still, unqualified success has been uncommon, and the abject failures have been frequent.
Across third-wave cases, there are huge differences in levels of democracy. Stagnation is found mostly among low-level democracies, although a handful of midlevel democracies have also met with this fate. In the former set of countries, the effective enjoyment of rights by citizens is highly uneven, opposition rights are routinely infringed, and mechanisms of accountability are stunted. In most cases, the electoral playing field is somewhat tilted, although not as much as in competitive authoritarian regimes. Moreover, these regimes are comparatively vulnerable to breakdown. The threat of democratic collapse looms even larger for countries that have undergone erosion. In contrast, high-level democracies are largely immune to breakdown. Citizens' rights are more evenly guaranteed, governments respect opposition rights, and mechanisms of accountability function.
We also investigated the background conditions associated with democratic deepening and democratic breakdown. The countries where democracy survived and advanced differ systematically from those in which democracy broke down or failed to register significant improvements: Democratic survival and deepening are both more likely with a solid economic performance. The odds of democratic advances are also higher in wealthier countries, and the odds of breakdown are lower in a more democratic neighborhood. Countries with a higher initial level of democracy are both more likely to survive and less likely to see improvements in democratic quality.
Breakdowns, Erosions, Stagnations, and Advances
To capture the range of outcomes discussed above, we make use of data generated by the Varieties of Democracy (V-Dem) project, which allows for a more nuanced measure of regime characteristics than other measures of democracy.3 Specifically, we build on the typology developed by Anna Lührmann, Staffan Lindberg, and Marcus Tannenberg [End Page 100] for identifying democratic transitions and breakdowns using V-Dem scores.4 These authors distinguish among four kinds of regimes: closed autocracies, electoral autocracies, electoral democracies, and liberal democracies.
In keeping with this typology, we code as a transition a shift from either kind of autocracy to either kind of democracy. The inverse change, from either kind of democracy to either kind of autocracy, we code as a breakdown. Like Lührmann, Lindberg, and Tannenberg, we drew the line dividing the two kinds of democracies, on the one hand, from the two kinds of autocracies, on the other—thus marking regime transitions—on the basis of scores from V-Dem's Electoral Democracy Index.5 We modified the rules set out by Lührmann, Lindberg, and Tannenberg (see online Appendix 1 at www.journalofdemocracy.org/articles/supplemental-material): We do not count some short-lived increases marginally above or drops marginally below their cutoff point for electoral democracy as a transition followed by a breakdown or vice-versa. Based on our rules, from 1974 to 2012, there were 91 transitions to democracy in 79 independent countries. (Even under our modified rules, there are twelve countries in which democracy arose, broke down, and then emerged once more within this time span.) Of course, as is true with all indicators of democracy, the coding of some cases is open to dispute.
Most major democracy indices, including those produced by V-Dem and Freedom House, recognize thinner and thicker conceptions of democracy—often called electoral democracy and liberal democracy, respectively. We use the thinner concept, the basis for V-Dem's electoral-democracy scores, to identify transitions and breakdowns. Scholars, policy makers, politicians, activists, and citizens typically think of a transition as occurring with the establishment of electoral democracy, and a breakdown as occurring when electoral democracy ceases to exist. To measure the level of democracy in surviving democracies, however, we employ the thicker conception, represented by scores on V-Dem's Liberal Democracy Index. In addition to being based partly on electoral democracy, V-Dem's liberal-democracy scores also reflect the degree to which personal liberties are respected, the rule of law prevails, and the judicial and legislative branches of government constrain the executive. Most discussions about the level or quality of democracy take into account this more comprehensive set of features.6 V-Dem's summary scores for both electoral and liberal democracy range from 0 to 1, with 1 representing the highest possible level of democracy.
Except in the case of breakdowns, we define the outcome of each transition to democracy by comparing the level of liberal democracy at the time of the transition to the level of liberal democracy in 2017. Because we are measuring the regime's trajectory (the change from the starting point to 2017) rather than the 2017 level of democracy, some [End Page 101] regimes that we code as registering a major advance have lower liberal-democracy scores than some we code as cases of stagnation.
Breakdowns: The 34 regimes in which democracy fell followed three different paths to breakdown. In some cases, democratic rule ended due to a military coup (as in Mali in 2012). The Peruvian breakdown of 5 April 1992 took the form of a "self-coup," in which President Alberto Fujimori announced that he was dissolving Congress and the judiciary. The most common route to breakdown among third-wave democracies, however, has been an incremental path without a clear breaking point. The countries that have traveled this path include Russia, Turkey, and Nicaragua. Although analysts widely agree that these countries are now authoritarian regimes, the exact year of breakdown in such cases is disputable. V-Dem dates the three breakdowns as occurring in 2000, 2014, and 2008, respectively.
Erosions: In some cases, levels of liberal democracy eroded substantially while reasonably free and fair competitive multiparty elections continued, so that the regime remained a democracy per our classification. In such cases, the electoral playing field became more uneven; the executive partially undermined mechanisms of accountability; and the government infringed on some opposition rights. We consider democracy to have eroded if there was a major decline in V-Dem liberal-democracy scores from the transition year to 2017 without a reversion to authoritarianism. Under our coding rules (on the coding of outcomes other than breakdowns, see online Appendix 2), the only two cases of this outcome are Ecuador and Poland.
Stagnations: This category encompasses regimes that 1) became democracies after 1974 and remained such as of 2017; 2) without achieving major advances or experiencing major declines in the level of democracy from the transition year to 2017; and 3) without being high-level liberal democracies (meaning a liberal-democracy score of .70 or above) in 2017. Stagnation here means that levels of liberal democracy in 2017 are relatively close to where they were in the year of transition. It does not necessarily mean a flat line between these points.
Some regimes (such as Lebanon) have stagnated at a very low level of liberal democracy; others (such as Greece) at a much higher level. Many found themselves stuck at intermediate levels of democracy for a long time. "Stagnation" might be a welcome scenario in the cases that we call highly democratic without major advances, which were relatively robust democracies at the outset. This is generally not the situation with the countries classified as stagnant, although in poor countries with weak states, preserving a low-quality democracy for an extended time is itself an achievement.
Advances: The opposite of democratic erosion is democratic advancement or deepening. When democracy advances, rights are more respected, checks and balances are strengthened, and elections become [End Page 102] more competitive, freer, and fairer. Empirically, democracy becomes less vulnerable to breakdown. Some cases where advances have occurred are still fairly low-level democracies (such as El Salvador, Guyana, and Romania), whereas others have become robust liberal democracies (such as Chile, Latvia, Spain, and Uruguay).
Highly democratic without major advances: Four democratic regimes registered a high initial level of liberal democracy and also ended with high liberal-democracy scores (at least 0.7), but without much change relative to their starting point. In these countries—the Czech Republic, Estonia, Lithuania, and Slovenia—citizens' rights received relatively even protection and institutions supplied effective checks and balances from early on.
Outcomes in the Third Wave
If measured against the unrealistically lofty expectations that arose shortly after the fall of communism in Europe, the third wave has been a disappointment. Building democracy entails an arduous process of constructing a solid state that can ensure order while also protecting citizens' rights, allow opposition forces to function, and check the executive. The high incidence of breakdowns and stagnations—which together make up 62 of our 91 outcomes—suggests that these challenges proved too difficult for a large majority of countries to surmount.
In 2017, the average liberal-democracy score across the 79 countries included in the analysis was only .48. This is roughly the 2017 score for Paraguay, a country with deep democratic deficits. If we consider only the countries that were still democratic in 2017, the average score is considerably higher (.56). Nonetheless, this brings the average up only to approximately the 2017 level in Brazil, which has been facing major political challenges.
Breakdowns have been the most common outcome. As Table 1 on page 104 shows, most occurred in relatively short order. The mean longevity of the democratic regimes that emerged and then failed between 1974 and 2012 was only 10.6 years, and the median was only 9. Because we did not count some short-lived shifts marginally below Lührmann's cutoff point as breakdowns, our estimate of the share of breakdowns is conservative, and our estimate of the median longevity is generous. Adhering strictly to Lührmann's rules would bring the share of breakdowns up to 58 out of 114 transitions (as opposed to 34 out of 91 under our modified rules).
As the third wave rolled from the southern European countries where it began on to Latin America, most scholars and policy makers were skeptical that the democracies then emerging would prove durable. Past waves of democracy in Latin America had been ephemeral, eventually giving way to periods of renewed authoritarianism. High inequality and [End Page 103] historically weak democratic institutions in most Latin American countries, widespread poverty in many, and the severe economic stresses facing the region in the 1980s and 1990s also augured poorly for stable democracy. Yet in Latin America as well as Southern Europe, the path to democratic survival ultimately proved smoother than in parts of the world later reached by the third wave. Out of the 43 democracies initiated between 1974 and 1992, many located in these two regions, only 14 (32.5 percent) had broken down by 2017. The annual breakdown rate, meaning the total number of breakdowns divided by the total number of country-years coded as electoral democracy between transition and 2017, was almost as low for democracies that emerged during the early third wave as it was for those that came into being between 1849 and 1944 (breakdowns occurred in 1.3 percent of country-years among the former group, compared to 1.1 percent for the latter group).
The early 1990s were a time of widespread democratic optimism, but many of the democracies that emerged from this period onward proved highly fragile and broke down. The annual breakdown rate among the democratic regimes initiated between 1993 and 2012 was 3.3 percent, nearly three times higher than that for the regimes that became democracies between 1974 and 1992. Out of the 48 regimes in the former group, 20 broke down by 2017. This later segment of the third wave included the postcommunist and African cases, in which democracy usually faced less favorable initial conditions. After the end of the Cold War, many African countries bowed to soft Western pressure and liberalized, sometimes leading to the birth of weak democracies. But few of these regimes grew into solid liberal democracies, [End Page 104] and many broke down, often relatively quickly. Among the ten African breakdowns included in Table 1, the average number of years from transition to breakdown was only 11.
Surprisingly, there were only two cases of democratic erosions without breakdown: Ecuador and Poland (see Table 2 above; on the classification of Hungary, see page 109 below). Ecuador's liberal-democracy score remained fairly stable from 1980 (.48), the first year of its democratic regime, until 2007 (.47), but gradual democratic erosion under President Rafael Correa (2007–17) brought this number down to .36 by 2017. In Poland, a country that had been one of the most successful cases of postcommunist democratization, the Law and Justice party has tightened government control over the judiciary, civil society, and the press since taking power in 2015. In that year, Poland scored .78 on the Liberal Democracy Index, a value nearly identical to its transition-year score of 0.77 (in 1991). Over just two years of Law and Justice rule, this score deteriorated sharply to .60. In addition to these cases, a few countries approached democratic erosion by 2017, and four regimes experienced erosion the year immediately before a breakdown (Fiji, Nicaragua, Serbia, and Ukraine).7
These findings suggest that erosion rarely ends in a stable equilibrium in which democracy persists at a reduced level. Many leaders who start down the path of weakening liberal democracy want to further undermine checks and balances. If they are successful, an electoral authoritarian regime emerges. If they are thwarted (as was Colombia's Alvaro Uribe in 2010), democracy remains intact, without experiencing a major erosion. Either way, a large-scale erosion of democracy without a wholesale descent into authoritarianism has been uncommon.
With 28 cases, stagnation is the second most common outcome among third-wave transitions (see Table 3 on page 106). Three notable patterns characterize the regimes in this group. First, they have exhibited little net change from the transition year to 2017 despite having lasted a long time on average—a mean of twenty years. Second, most exhibited glaring democratic deficits in the first year of democracy and still did in 2017. The mean 2017 liberal-democracy score among the 28 cases was .50, similar to the 2017 level of Colombia and of Burkina Faso. These weak democracies gave rise to Guillermo O'Donnell's seminal theorizing about "brown" areas within democratic and semi-democratic regimes—geographic zones in which the protection of democratic rights is truncated, clientelism is pervasive, checks and balances are weak, and the rule of law is precarious.8 Third, if on average these [End Page 105] countries are inching toward becoming more robust democracies, they are doing so at a snail's pace. Between the year of transition and 2017, the average stagnant democracy's liberal-democracy score increased by only 0.03, and the average annual increase per country was a meager .0016. If we calculate based on the mean 2017 score for this group and the mean annual rate of change since the transition year, it would take the average stagnant democracy 125 years to become a high-quality liberal democracy (that is, to reach a score of .70).
There is considerable variation across these 28 cases. As of 2017, Benin, Ghana, Greece, South Africa, and Suriname came fairly close to being coded as having achieved democratic advances. Conversely, democracy has deteriorated in Hungary, Lesotho, and Liberia following their transitions, but not enough for these cases to qualify as instances of erosion.
The fact that many countries did not exhibit much change from the transition year to 2017 indicates that stagnation is often a stable equilibrium. Yet stagnation can also frequently be a way station on the road [End Page 106] to breakdown. Of the 34 regimes that broke down, 28 would have been classified as cases of stagnation if measured the year prior to breakdown (see online Appendix 3 for details). Because so many regimes have stagnated over a protracted period of time, and because these regimes are vulnerable to breakdown, this understudied outcome of transitions deserves more attention than it has received.
Notwithstanding the disappointing overall record of the third wave, it includes 23 cases, spread across six continents, in which major democratic improvements took place after the transition (see Table 4 above).
This list of advances requires some sobering caveats. Most of the 23 regimes that fall into this category had strikingly low initial liberal-democracy scores: The mean initial score was a meager 0.425. Consequently, in most cases, significant score increases did not mean that these countries had become high-level democracies, but simply that they had moved up from a low starting point. As of 2017, many of these regimes remained what Fareed Zakaria called illiberal democracies and Larry Diamond described as electoral democracies.9
Achieving a high level of liberal democracy has been the exception, but it is feasible. The set of 23 advances includes eight countries that received [End Page 107] a liberal-democracy score of .70 or above in 2017—Cape Verde, Chile, Latvia, Portugal, Slovakia, South Korea, Spain, and Uruguay. Portugal, South Korea, and Uruguay were the countries that saw the largest improvements from the transition to 2017.
Highly democratic without major advances
Only the Czech Republic, Estonia, Lithuania, and Slovenia fall into this category (see Table 5 above). These four postcommunist countries share with the cases of stagnation the fact that their V-Dem liberal-democracy scores have not changed greatly since the year of transition. They do not, however, exhibit the same large democratic deficits that we see in the regimes classified as having stagnated. At an intuitive level, it is not helpful to label high-level democracies as stagnant.
All four regimes began with high V-Dem liberal-democracy scores. The average initial score for this group was 0.78 (for comparison, the United Kingdom scored 0.80 in 2017). Three of the four countries—all but Slovenia—had some history of democracy before the Second World War, a legacy that likely gave them an advantage in rebuilding democracy after the fall of communism.
Adding these four cases to the eight cases of democratic advances leading to high liberal-democracy scores yields a total of only twelve robust liberal democracies in place as of 2017 across the 79 countries that underwent third-wave transitions. Pitted against the 34 breakdowns and 28 stagnations, this is a disappointing record. Yet it is worth keeping in mind that many of these breakdowns and stagnations occurred in inhospitable terrain with inauspicious starting conditions.
Among those third-wave democratic regimes that have endured, low-level democracy has been the prevalent outcome. The absence of advances after transition often results from a stalemate between forces that favor deepening democracy and others opposed to this course. One manifestation of this push-and-pull is the existence of subnational hybrid and authoritarian regimes. Under national-level democratic regimes, Argentina, Brazil, Mexico, and Peru have all hosted hybrid regimes on parts of their territory, as did Russia before it reverted to clearly authoritarian rule under President Vladimir Putin. Subnational hybrid regimes have elections that are not fully free and fair even if the vote count is accurate. The playing [End Page 108] field is tilted. The media and the courts are largely under the control of the provincial executives, and subnational authoritarian leaders use public resources to cultivate support and undercut opposition. These pockets of hybrid or authoritarian rule weaken democracy at the national level.10 Large within-country inequalities in the exercise of democratic rights are common in low-quality democracies.
Since the 1990s, another distortion has sprung up in some democratic regimes, mainly in sub-Saharan African and Latin America: Criminal organizations have undercut democracy at the local level, especially in poor urban neighborhoods. Residents of neighborhoods that are controlled by criminal organizations cannot exercise freedom of speech or other basic democratic rights; they cannot vote freely and fairly; and they might not be able to vote at all.11 Where organized criminal groups subvert democracy in this way, even if a country achieves other democratic gains, the overall level of liberal democracy will remain modest.
A third common scenario leading to low-quality democracy occurs when national executives block democratic deepening or work to weaken checks and balances, but fall short of instigating full-blown democratic breakdowns. The resulting regimes allow for elections with an accurate vote count, although usually not a fair playing field. In the 2010s, Hungary and Bolivia have exemplified this path. Both countries made major advances during their first two decades as democracies, but power-hungry executives have since presided over the corrosion of liberal democracy. Hungarian prime minister Viktor Orbán and Bolivian president Evo Morales have chipped away at opposition rights and checks and balances since taking office in 2010 and 2006, respectively. Between transition and 2017, levels of democracy in each country rose and then declined, but they did not fall far enough to qualify as a democratic breakdown or erosion. Instead, they have reverted to roughly what they were in the transition year.
The democratic setbacks that took place under Morales and Orbán are not breakdowns since, as of 2017, Bolivia and Hungary still had electoral-democracy scores above our threshold. They do not constitute erosions because the two countries' 2017 liberal-democracy scores, though lower than before their current leaders came to power, did not drop significantly below the scores assigned at the time of transition. If we coded erosions based on change from a country's highest post-transition liberal-democracy score to 2017, Bolivia and Hungary would qualify. But because we measured change from the transition year to 2017, they are classified as cases of stagnation. (In this regard, Hungary's comparatively low initial liberal-democracy score of 0.55 helps to explain why its classification differs from that of Poland, which received a liberal-democracy score of 0.77 in its year of transition). [End Page 109]
In parts of the world where the United States or the European Union have strong leverage and extensive ties (or "linkage"), international dynamics that raised the cost of openly abandoning democracy have paradoxically contributed to some cases of low-quality democratic survival.12 After the end of the Cold War, coups and overt dictatorships became less common because the United States, the Organization of American States, and the European Union exerted pressure to support democratization and promote human rights. Even before the recent erosion of the prodemocracy foreign-policy consensus in Western Europe and the United States, however, it was usually difficult for Western actors to prod unwilling countries to deepen their shallow democracies. Thus, in many cases where leaders lacked a principled commitment to democratic rule, democracy survived without deepening.
The Role of Initial Background Conditions
To better understand the conditions associated with democratic survival, breakdown, and advance, we performed some simple multivariate-regression analyses. While many posttransition factors influence regime outcomes, here we focused on some of the initial structural and political circumstances that confronted democratic leaders. To this end, we examined primarily indicators measured in the year in which countries transitioned.
The factors that might make breakdowns less likely and major democratic advances easier include a higher level of economic development, greater state capacity, a more democratic neighborhood, and a more democratic past. A higher initial level of democracy should make breakdowns less likely. We also tested whether presidential systems made countries more vulnerable to breakdown in order to see if Juan Linz's classic argument about the "perils of presidentialism" holds up.13 In addition to these measures of starting conditions, we analyzed one contemporaneous factor: average GDP per capita growth from the first year of democracy to the last, which is either 2017 or the final year of democracy before a breakdown. Poor economic performance might make democracies more vulnerable to breakdown and less likely to achieve major advances. To minimize the number of observations that we lose because of missing data, we include state capacity and per capita GDP separately and then put both in the same regression. (For details, see online Appendix 4.)
Our calculations showed that regimes that started off with a higher level of liberal democracy, that were geographically surrounded by democracies, and that experienced better rates of economic growth were less likely to break down (see Table A1 in online Appendix 4). These findings align with expectations. Wealthier countries might also have been less prone to breakdowns, but the numbers are inconclusive [End Page 110] (the p value is less than 0.14). Against expectations, prior democratic history and levels of state capacity show no association with breakdowns.
When we test for associations between the initial conditions and democratic advancement, a slightly different picture emerges. If we consider which of the 57 democracies that survived until 2017 were among the 23 that saw major advances in levels of liberal democracy, we find that regimes with a lower initial per capita GDP, as well as those that experienced lower economic growth, were less likely to deepen democracy. The same is true for regimes that started off with a higher level of democracy—a pattern that presumably reflects a "ceiling effect." Among regimes that survive, those that began with lower scores had further to go before bumping up against this ceiling, and thus found it easier to make big gains.
The fact that countries with a higher per capita GDP are more likely to register major advances is consistent with modernization theory. While many new democracies are born in countries at a low level of development, these regimes are less likely to deepen. Against expectations, greater state capacity did not show a strong association with democratic advances. Also contrary to expectations, countries with a stronger earlier democratic heritage and those in a more democratic neighborhood were not more likely to achieve major democratic gains. Of the other factors we tested, initial levels of inequality showed no clear association with either breakdowns or advances. Likewise against expectations, a transition from a military dictatorship did not result in a higher probability of breakdown than a transition from any other kind of dictatorship.
If one lesson from our findings is that major democratic advances are difficult even under favorable circumstances, another is that such advances were particularly elusive for those democracies encumbered by poor economic conditions. The democracies that had the advantage of a higher per capita GDP in the year of their birth and those that experienced strong growth were more likely to achieve major gains. Regimes that experienced solid economic growth were also more likely to avoid breakdowns.
In this regard, compare the paths taken by the two Peruvian democratic regimes initiated between 1974 and 2012. The first emerged in 1980 (or, by our coding, 1981), but fell to President Fujimori's 1992 "self-coup." After Fujimori fled the country and resigned under popular pressure in 2000, Peru restored democracy in 2001. Since that year, Peru's liberal-democracy scores have improved enough to land this case on our list of democratic advances.
Strong economic growth gave Peru's current democratic regime notably greater odds of survival and deepening than its crisis-ridden precursor. In a country with a growth record as dismal as that of the [End Page 111] Peruvian democracy of 1981–92 (an average of -0.8 percent per capita per year), the odds of breakdown were 90 percent higher compared to a country that has grown as rapidly as has Peru since redemocratization in 2001 (5.7 percent per capita per year on average). Similarly, the odds of advancement are about 3.6 times greater for a country with the growth numbers of Peru's current regime than they were for this regime's ill-fated predecessor.
Although initial background conditions affect the probabilities of different outcomes, they are not destiny, as successful democracies in such challenging settings as India since 1947 and Costa Rica since 1949 demonstrate. Among third-wave cases, Benin (0.61), São Tomé and Príncipe (0.61), Senegal (0.58), Ghana (0.54), and Mongolia (0.53) had by 2017 attained considerably higher levels of liberal democracy than one would have expected based on these countries' very low initial per capita GDPs. Indeed, the mere survival of democracy in these cases is a success story. Likewise, in 2017, Slovenia (0.79) and Benin (0.61) had liberal-democracy scores far higher than one would have expected given their very low state capacity in their first years as democracies. Democracy can flourish despite remarkably adverse initial conditions.
The third wave generated hopes that democracy had triumphed decisively over dictatorship. Global politics was transformed; never before have so many people in the world lived under electoral democracy. Still, the most common outcomes of third-wave transitions have been breakdown and stagnation. The cases of substantial democratic deepening leading to robust liberal democracies are isolated exceptions. Not surprisingly, democratic breakdowns have been considerably more likely in countries with a less democratic starting point, worse economic performance, and a less hospitable neighborhood. Major democratic advances have been more likely in countries that started off with a higher per capita GDP and those that enjoyed stronger economic growth, as well as those with a less democratic beginning point.
We conclude with two optimistic notes at this challenging moment for democracy's champions. First, on average, the countries caught up in the third wave were much more democratic in 2017 than they were before their first post-1974 transitions. The average liberal-democracy score for our 79 countries in the year prior to their first transition was .278, about the score of contemporary Honduras. In 2017, the average was .482. It would be foolhardy to overlook the setbacks and stagnation, but it would be equally misleading to ignore the dramatic democratic advances that come into view when we consider long sweeps of history. Second, although democratic survival and advances are more difficult when regimes face adverse starting conditions, the third wave has shown once again that, even under such conditions, new democracies can thrive. [End Page 112]
Scott Mainwaring is the Jorge Paulo Lemann Professor for Brazil Studies at the Harvard Kennedy School. Most recently, he is editor of Party Systems in Latin America: Institutionalization, Decay, and Collapse (2018).
Fernando Bizzarro is a doctoral candidate in government at Harvard University and a graduate-student associate at the David Rockefeller Center for Latin American Studies.
1. The Portuguese revolution of 1974 is usually considered the beginning of the third wave. The democratic government in Portugal, however, did not take office until 1976, while its counterpart in Greece took office in November 1974—the first transition in our dataset. Because it codes the first calendar year that was mostly democratic, V-Dem dates the Greek transition to 1975.
2. We did not track transitions that occurred after 2012 because we wanted to assess how regimes evolved for at least five years. Due to V-Dem's coding rules, Lührmann et al. often code transitions as occurring a year later than the actual transfer of power.
3. Michael Coppedge et al., "V-Dem Country-Year Dataset v8," Varieties of Democracy Project, 2018; Daniel Pemstein et al., "The V-Dem Measurement Model: Latent Variable Analysis for Cross-National and Cross-Temporal Expert-Coded Data," V-Dem Institute Working Paper 2018:21, 3rd ed., University of Gothenburg, April 2018.
4. Anna Lührmann, Marcus Tannenberg, and Staffan I. Lindberg, "Regimes of the World (RoW): Opening New Avenues for the Comparative Study of Political Regimes," Politics and Governance 6, no. 1 (2018): 60–77.
5. V-Dem's understanding of electoral democracy includes elections with a fair vote count, freedom of association, and freedom of expression.
6. Larry Diamond and Leonardo Morlino, eds., Assessing the Quality of Democracy (Baltimore: Johns Hopkins University Press, 2005), and Daniel H. Levine and José E. Molina, eds., The Quality of Democracy in Latin America (Boulder, Colo.: Lynne Rienner, 2011).
7. There was also one case of a major advance prior to breakdown—Burkina Faso between 1993 and 2014.
8. Guillermo O'Donnell, "On the State, Democratization and Some Conceptual Problems: A Latin American View with Glances at Some Postcommunist Countries," World Development 21 (1993): 1355–69.
9. Fareed Zakaria, "The Rise of Illiberal Democracy," Foreign Affairs 76 (November–December 1997): 22–43; Larry Diamond, Developing Democracy: Toward Consolidation (Baltimore: Johns Hopkins University Press, 1999).
10. Carlos Gervasoni, Hybrid Regimes Within Democracies: Fiscal Federalism and Subnational Rentier States (New York: Cambridge University Press, 2018).
11. Andreas Schedler, "The Criminal Subversion of Mexican Democracy," Journal of Democracy 25 (January 2014): 5–18.
12. See Steven Levitsky and Lucan A. Way, "International Linkage and Democratization," Journal of Democracy 16 (July 2005): 20–34.
13. Juan J. Linz, "The Perils of Presidentialism," Journal of Democracy 1 (Winter 1990): 51–69.