The literature on democratization has experienced radical mood swings in recent decades, from extreme optimism in the 1990s to extreme pessimism today. These mood swings have resulted in not only misguided claims about the state of democracy in the world but also a muddied understanding of what drives both democratization and democratic erosion.

The extraordinary success of early Third Wave transitions—almost all of the transitions in Southern Europe in the 1970s, South America in the 1980s, and Central Europe in 1989 led to democracies, most of them stable—engendered a wave of optimism about global democratization. This optimism was reinforced by the collapse of the Soviet Union and the end of the Cold War, which contributed to the breakdown of autocracies across the former Soviet Union, Africa, and parts of Asia. In some cases (e.g., Albania, Benin, El Salvador, Ghana, and Mali), democracy emerged in countries with none of the conditions that scholars considered favorable to democracy. This unprecedented global expansion of democracy, and its reach into new and unexpected settings, led some observers to throw out a half-century of social science research on structural conditions that help to produce and sustain democracy—such as economic development, large working and middle classes, robust civil societies, low levels of social inequality, a dispersion of societal resources, and functioning states—in favor of voluntaristic approaches that highlighted the role of political elites and institutional design (cf. Di Palma Reference Di Palma1990). This voluntarism often was accompanied by a teleological view of history in which countries were largely assumed to be moving toward democracy. So, rather than concluding from the Third Wave that democracy can emerge (almost) anywhere, many observers began to expect that it would emerge everywhere. When confronted with hybrid regimes such as those in Cambodia, Haiti, Kyrgyzstan, Tanzania, and Russia, observers routinely categorized them as incomplete or “transitional” democracies (Collier and Levitsky Reference Collier and Levitsky1997) or regimes undergoing “protracted” democratic transition.Footnote ¹

The early-twenty-first century brought a dramatic mood shift. The global democratic expansion ended. Some countries (e.g., China and much of the Middle East and North Africa region) remained stubbornly authoritarian. Many post–Cold War hybrid regimes that observers described as “in transition” to democracy (e.g., Belarus, Cambodia, Cameroon, and Russia) reverted to full authoritarian rule. A few new democracies in hard places (Masoud and Mainwaring Reference Masoud and Mainwaring2022), such as those in El Salvador, Mali, and Nicaragua, broke down.

These developments should not have surprised us. From almost any social science perspective, global democracy dramatically overperformed in the 1990s and early 2000s (Treisman Reference Treisman2020). Not only did democracy emerge in almost every country with conditions that social science theories suggest are highly favorable to democracy (Singapore was perhaps the leading exception). It also emerged in many countries with highly unfavorable conditions for democracy, including Albania, Benin, El Salvador, Mali, Mongolia, and Nicaragua. By 2010, then, few autocracies remained in countries where we would expect democratization, but now there were many democracies in countries where we would expect regime failure and instability.

However, the stagnation in the level of global democracy was interpreted by many observers as evidence of democratic recession, a reversal of the Third Wave, an “authoritarian resurgence,” and even “the beginning of the end for democracy” (Battison Reference Battison2011; Diamond Reference Diamond2008, Reference Diamond2014; Puddington Reference Puddington2009). These claims were overstated—and based largely on an illusion. The failure of democracy to continue to expand was interpreted by some disappointed observers as evidence of decline.

In the second decade of the 2010s, a few relatively established democracies experienced backsliding (i.e., Hungary, India, and Poland) or collapse (i.e., Thailand and Venezuela). These developments, together with the surprising crisis of democracy in the United States, crystalized a new conventional wisdom: global democracy was in decline and authoritarianism was resurgent. In 2022, Freedom House claimed that the world had suffered 16 consecutive years of democratic decline.Footnote ² Another major index, V-Dem, warned of an “intensifying wave of autocratization.”Footnote ³ The most recent V-Dem report concluded that the “Global level of democracy is back to 1986.”

In this context, Little and Meng make a refreshing contribution to our understanding of the state of democracy in the world. Their primary contribution is methodological. All major indices of democracy rely on largely subjective expert assessments. It is true that V-Dem scholars (in particular) put considerable thought into validating this method. Scores generally reflect the opinion of more than one expert and are mostly accurate assessments of countries’ level of democracy. In addition, V-Dem reports the standard deviation of coding values given by the experts. Nevertheless, expert coding has important downsides. One is that only the expert coders know which specific events and factors motivate their coding decisions. The scores are impossible to replicate or falsify. Some dubious scoring decisions highlight the cost of this lack of transparency. For example, V-Dem coded Ukrainian democracy as declining after the 2014 Euromaidan protests toppled Viktor Yanukovych, an autocratic president who jailed major rivals. The new regime was characterized by relatively free and fair elections, a pluralistic media environment, and no arrests of opposition leaders. So why did V-Dem code Ukrainian democracy as declining? Were V-Dem coders responding to the emergence of local mini-dictatorships (e.g., the Donetsk People’s Republic and the Luhansk People’s Republic) in a few Russian-controlled breakaway regions of Ukraine? Was it the fact that Russian TV was shut down after Russia invaded Crimea? We can only guess.

Second, as Little and Meng argue, subjective measures are vulnerable to biases that may shift over time given the news cycle, the performance of democracy in high-profile cases, and the heightened public discussion of democratic backsliding. Another plausible source of coder bias in addition to those mentioned by Little and Meng is the election of populist- or autocratic-minded politicians including Donald Trump in the United States, Jair Bolsonaro in Brazil, and Giorgia Meloni’s Brothers of Italy. The election of an autocrat is an obvious cause for concern; however, it does not, in itself, constitute backsliding. Due to the strength of institutions or opposition forces or their own political weakness or ineptitude, autocratic politicians sometimes fail to undermine democracy. Bolsonaro in Brazil is a recent example.

Subjective measures also may reflect very different criteria for democracy among country experts who cover different parts of the world. For example, divergent benchmarks for democracy may explain why V-Dem coded Sierra Leone as democratic in 2022 and Malaysia as an electoral autocracy, even though there were more documented abuses in Sierra Leone than in Malaysia. Indeed, for reasons that are difficult to identify, V-Dem places Malaysia—which has experienced three democratic turnovers since 2018—in the same category as Russia, where leading opposition figures are routinely killed, jailed, and barred from running in elections.

Little and Meng address these deficiencies in subjective coding by relying on objective data—in particular, electoral turnover and outside reports on the treatment of journalists. Unlike the expert assessments that serve as the bases of Freedom House and V-Dem indices, data on election results and attacks on journalists can be replicated and falsified. We do not need to guess what was in the mind of a coder. Moreover, as Little and Meng note, their data are less vulnerable to problems of subjective bias.

Drawing on these data, Little and Meng reach a conclusion that diverges from recent analyses by V-Dem, Freedom House, and other observers of global democracy. They focus on the simple but important measure of electoral turnover, which they find has remained relatively unchanged since the late 1990s following a marked increase in turnover in the 1980s and early 1990s. Electoral turnover—or incumbents’ loss of power to the opposition—should not be seen as equivalent to democracy (Levitsky and Way Reference Levitsky and Way2010, 12–13). However, the loss of incumbent power hampers the ability of leaders to build durable patronage networks or establish firm control over institutions such as the judiciary, the electoral authorities, and the armed forces. The regularity of turnover suggests that many regimes are not “autocratizing”—and, in fact, remaining fairly competitive.

Little and Meng’s data on incumbent turnover—suggesting, in effect, a flat line in the twenty-first century—challenge claims of a large-scale democratic reversal or authoritarian resurgence and belie V-Dem’s recent claim that global democracy has returned to its 1986 level (when all of Eastern Europe and almost all of Africa were under single-party or military rule). Although Little and Meng do not provide a full picture of civil-liberties violations over time, the maintenance of meaningful electoral competition suggests that any recent decline has been modest. As Little and Meng note, a substantial decline in civil liberties almost certainly would have affected the level of incumbent turnover.

However, if Little and Meng’s data run starkly counter to V-Dem and Freedom House reports, which claim a steep democratic decline, they actually diverge little from these indices’ underlying data. Looking at their figure 8, it is difficult to see a substantial divergence between Little and Meng’s objective scores and V-Dem’s subjective scores.

This suggests that the problem that Little and Meng highlight lies less with the V-Dem and Freedom House data than with these organizations’ interpretation of the data. Stated bluntly: the dramatic claims by V-Dem and Freedom House of democratic decline are starkly out of line with their own data. For example: Freedom House claimed in 2022 that global freedom had declined for 16 consecutive years—a seemingly devastating figure. Yet, according to its own data, the number of democracies in the world declined by only six—a more modest figure—during that period. Similarly, whereas V-Dem prominently claims that democracy has declined to the level it was before the end of the Cold War in 1986, their own data suggest a more modest decline from a high of 96 electoral and liberal democracies in 2017 to 90 in 2022. By way of contrast, in 1986, according to their data, there were half that number (44).

The stark gap between data and rhetoric in the V-Dem and Freedom House reports suggests that their authors consciously tried to squeeze every bit of pessimism from their data. V-Dem’s most extreme conclusions, for example, are based on nonstandard measures of global democracy that are weighted by population—and thus hinge heavily on scoring of the single case of India. Indeed, V-Dem only began to draw prominent attention to population-weighted measures in its reports in the late 2010s when, following the deterioration of democracy in a few large countries, this indicator made the state of global democracy seem worse.Footnote ⁴ Today, all of V-Dem’s most prominent and widely quoted findings are in population-weighted terms.

Similarly, V-Dem reports use the term “autocratization” to refer to cases such as Brazil and the United States, which—despite mild backsliding—remained democratic according to V-Dem’s own measures. By placing highly competitive regimes such as India, Malaysia, Nigeria, Serbia, and Ukraine in the same “electoral autocracy” category as dictatorships including Cambodia, Egypt, and the Republic of Congo, V-Dem reports—misleadingly—that citizens in Nigeria, Serbia, Ukraine, and Malaysia (!) are “living in autocracy.”Footnote ⁵

We only can speculate about what drives this gap between data and rhetoric. We suspect that claims of large-scale democratic decline draw more attention from journalists and policy makers than reports concluding that the level of democracy has declined slightly during the past decade and remains more or less the same as it was in the early 2000s—which is what V-Dem’s data actually show.

Thus, both subjective and objective indices show that democracy has remained remarkably resilient in recent decades. If we believe V-Dem and Freedom House, democracy has declined slightly during the past decade. If we believe Little and Meng, it has not declined at all. Either way, the level of democracy remains at world-historic highs—and levels of outright autocracy remain at world-historic lows. According to V-Dem’s own data, only 18% of the world’s regimes (33 of a total of 179) were closed in 2022, compared to 54% (86 of 158) in 1973, on the eve of the Third Wave. Likewise, a striking 50% of the world’s regimes were electoral or liberal democracies in 2022, an increase from only 22% in 1973—and slightly less than the peak of 54% in 2017.

This is surprising. Global democracy has confronted serious threats during the past two decades. Most important, the geopolitical balance of power has shifted dramatically. The post–Cold War era of Western liberal hegemony—in which the United States and Western Europe were the world’s undisputed military, economic, and ideological powers—is over. The rise of China and the renewed power and aggressiveness of Russia and other illiberal states (e.g., Iran and Saudi Arabia) have created an international environment that is far less favorable to democracy—and more favorable to autocracy—than that which existed during the heady 1990s. Unlike during the immediate post–Cold War period, democracy is not “the only game in town.” Not only have Western liberal powers weakened; they also have scaled back their efforts to promote democracy (Diamond Reference Diamond2019). Finally, the 2008–2009 financial crisis and the European migration crisis, combined with longer-term problems of growing inequality and social immobility, generated serious problems within Western democracies (Mounk Reference Mounk2018; Norris and Inglehart Reference Norris and Inglehart2019). Indeed, polarization, the rise of right-wing populism, and the crises triggered by the Trump presidency in the United States suggested that some of the world’s oldest democracies were not safe. The rise of Trump challenged democracy both within the United States and abroad. Never in the modern era has global democracy confronted such hostility from the United States.

Given these developments, the fact that a large majority of Third Wave democracies have survived during the past two decades suggests a striking degree of democratic resilience. What explains this resilience? Two broad factors seem to be at work: the strength of societal pro-democratic forces in some cases and the weakness of state authoritarian forces in others. First, as decades of social science research have shown, economic development remains a powerful source of democratic persistence. High levels of capitalist development generate vast independent sources of economic and social power, which disperses resources and shifts the balance of state–society power away from the state, making it almost prohibitively difficult for leaders to monopolize political control. There are almost no dictatorships in the world’s wealthiest societies. If we exclude oil-based rentier states (where vast wealth is produced without the societal pluralism or resource dispersion generated by most forms of capitalist development), then of the 53 countries classified as high income by the World Bank, all but two–Hungary and Singapore—were scored as free by Freedom House in 2022.

Modernization helps to explain the resilience of two sets of Third Wave cases. First, many early Third Wave democratizations in Southern Europe, Central Europe, and parts of South America (i.e., Argentina, Brazil, Chile, and Uruguay) and East Asia (i.e., South Korea and Taiwan) occurred in relatively developed societies. As decades of social science research would predict, democracy has survived in almost all of these cases; the one exception is Hungary.

Second, in other cases, the structural conditions for democracy improved dramatically during the post–Cold War era. Thus, although countries including the Dominican Republic and Panama in the Americas and Bulgaria, Romania, Slovakia, Latvia, and Lithuania in Central Europe were still relatively poor when they initially democratized, they modernized considerably during the late-twentieth and early-twenty-first centuries, which helped to secure their democracies. Indeed, the number of (non-oil–based) high-income countries has more than doubled—from 25 to 53—since 1987. One reason why democracy has persisted in the more than three decades since the end of the Cold War, then, is that socioeconomic conditions improved dramatically during that period.

However, modernization cannot explain the survival of democracy or near-democracy elsewhere in the world. Indeed, about 20 low- and lower-middle-income countries are now democracies. In much of Africa and parts of Asia and Latin America, the persistence of competitive politics is rooted primarily in authoritarian weakness (Way Reference Way2015). For example, autocracies in weak states face the same governance and governability challenges that governments in all weak states face. Weak states suffer from economic volatility, chronic fiscal crisis, corruption, public insecurity, poor public goods provision, and a limited capacity to redistribute wealth. As a result, they frequently confront public discontent, protest, and regime challenges. Thus, in weak states such as Albania, Benin, Ecuador, Honduras, Sri Lanka, and Zambia, embryonic autocracies often have proven as weak, ineffective, and ephemeral as the democracies they replaced.

State and ruling-party weakness also limits the capacity of autocratic governments to thwart opposition challenges (Way Reference Way2015). Weak ruling parties often leave autocrats vulnerable to regime-threatening internal schisms, and weak bureaucratic and coercive agencies often undermine autocrats’ efforts to manipulate elections or crack down on opposition protest. This repeatedly has been the case—for example, in Ukraine. Autocrat Viktor Yanukovych fell from power amid the 2014 Maidan Revolution because the security forces abandoned him and refused to repress. Later that year, Petro Poroshenko won the presidency without an established party. When he tried to declare martial law in 2018 in an apparent effort to delay presidential elections, he confronted intense opposition, including from his own allies. As a result, the elections went forward and he was soundly defeated by Volodymyr Zelensky. Thus, it was authoritarian weakness that made Ukrainian democracy possible.

Little and Meng provide an important reality check to a conventional wisdom gone awry. Describing the state of global democracy during the past two decades as a flat line may gain less public attention than warning of “authoritarian resurgence” or a global wave of “autocratization,” but it is closer to reality. Democracy has proven surprisingly resilient in the twenty-first century, despite an unfavorable international environment. Rather than downplay that resilience, we should seek to better understand it—and, in that way, perhaps help to sustain it.