Over-diagnosing democratic decline

After a massive expansion in the wake of the Cold War, the past decade seems to have been tougher for democracy. Optimism about democratization in China has been all but completely extinguished. Countries that once seemed on the brink of democratic future, like Russia or Turkey, have devolved into authoritarianism. Populism has taken hold across the West, and the number of populist leaders in power soared to an all-time high. Optimism about global democratization in the wake of the Cold War has given way to talk of democratic recession.

Still, there are reasons for pause about such alarmism. While the number of populist leaders was up to an all-time high in the mid-2010s, it has decreased to a 20-year low in 2022 and 2023. Most of these countries had been noted as examples of democratic backsliding, but voters became tired of the populists’ heated rhetoric and replaced them with more moderate leaders. And contrary to the United States’ experience with Donald Trump in the wake of the 2020 presidential election, most of these populist leaders, like Boyko Borisov of Bulgaria and Andrej Babis of the Czech Republic, left office without much fanfare.

How much democratic backsliding could there have really been if voters replaced the accused leaders through an electoral process that, in most cases, functioned smoothly? Recent commentary and the major global indicators of democracy on which they have based their judgments have been arguably over-diagnosing democratic decline. The conceptions of democracy from which these various sources work are often very broad, including not just whether electoral institutions function or are likely to function smoothly, but whether the country protects a variety of liberal rights and has healthy sociopolitical norms. While these are important aspects of good governance, and are likely to affect the health of democracy in the long run, diagnoses of the current state of democracy should focus on the likelihood that free and fair elections can occur in the short run.

Voters became tired of the populists’ heated rhetoric and replaced them with more moderate leaders.

When diagnosing the current state of democracy, we should work from a minimalist conception of democracy that focuses on the likely performance of electoral institutions. A country experiences democratic backsliding only if there’s a significant decline in open competition for office or in the translation of vote results into office holding. To be sure, this requires some rights and institutional safeguards. There must be minimal restrictions on who can run for office; freedom of speech, press, and assembly must be protected so that dissenters can air their views; and the institutional machinery for running and verifying elections must work in a relatively even-handed way. But while protection of further liberal rights and institutional checks and balances to constrain executive power may be desirable in themselves and make for a more stable political regime, they are not necessary for a country to be considered fully democratic in this view.

We must scrutinize both whether recent elections and upcoming elections are likely to be relatively free and fair before we accuse a country of being a democratic backslider. Backsliding is a powerful accusation, and if not well-founded, it may threaten to undermine faith in the experts who made such judgments and increase political polarization. Although heated rhetoric and political polarization may eventually cause politicians to try to degrade electoral institutions, in the short run, elections can remain open and competitive despite them. And it is possible that the electorate will tire of heated rhetoric and polarization and elect politicians who promise to dial them down. Many will be dissatisfied with this process, but we should not lose sight of the fundamental virtue of democracy, that it allows the people to be self-correcting in how they’re governed.

Defining “democracy” and “democratic backsliding”

How should we define democracy? This has been the subject of at least a few centuries of lively debate, and despite the likely millions of words written on it, debate continues to be lively. A central debate is between minimalist and maximalist conceptions of democracy. Minimalist conceptions of democracy, associated with Joseph Schumpeter and Adam Przeworski, focus on the electoral process. To be considered a democracy, there must be open contestation for high office. Incumbents must occasionally lose elections, and they must leave office when they do. This contrasts with a more maximalist conception of democracy, which is concerned with democracy’s ability to guarantee individual freedom and includes a variety of liberal rights and institutional constraints on power to help ensure this. 

When focusing on democratic backsliding, we should work from the minimalist conception’s focus on functional electoral institutions. By functional electoral institutions, I mean electoral institutions in which elections for key leadership offices are based on open competition, there are minimal significant restrictions on voting, vote tallies reflect and are roughly proportional to voters’ intended votes, and candidates with the highest vote tally according to the electoral procedure assume office. While the translation of votes into representation need not be strictly proportional, changes which significantly enhance the voting power of certain groups relative to others, such as gerrymandering, would constitute democratic backsliding.

To be sure, some liberal rights and institutions are necessary for the realization of open competition. Different groups must be able to make their political views and criticisms of the government public, meaning that a free press is a necessary precondition. Therefore, increasing media censorship, restricting groups’ access to media, or consolidating control of the media under political allies could count as democratic backsliding. But a case where a leader takes a hostile attitude toward the media without successfully censoring it would not.

Major global democracy datasets, like Polity, Freedom House, and the Varieties of Democracy project (V-Dem), include the elements of this minimal conception of democracy as part of their definitions, but also include several additional components, including liberal rights, constraints on the executive, and social norms. This broader conception forms the basis of the headline democracy indicators on which they have based recent diagnoses of democratic backsliding. V-Dem, for example, bases its yearly reports on a combination of its Electoral and Liberal Democracy Indices, which contain critical elements of the minimalist conception like government censorship of the media and party bans, but also include broader “good governance” elements such as the presence of media bias, freedom of foreign movement, and executive responsiveness to legislative questioning.

Including a wide range of substantive rights and institutional checks on power in a conceptualization of democracy risks over-diagnosing democratic decline because it removes focus from the performance of electoral institutions. Views about which rights people should have vary around the world, and a conception that includes many liberal rights risks classifying a country that is conservative but has robust political competition as non-democratic. Respect for pluralism is an essential element of democracy, and incorporating it into democratic practice means allowing each group to take control of political office and shift policy in their preferred direction if they can convince enough people to support them to win an election.

In a liberal country, there’s also a strong chance that a politician who tries to take away valued rights or remove institutional checks on their power will inspire voter backlash. If electoral institutions are functional, voters will be able to remove these leaders from office. It’s crucial that a conception of democratic backsliding can distinguish cases in which leaders have insulated themselves from removal from ones in which they’ve taken controversial actions but have not significantly affected voters’ ability to remove them. 

Distinguishing degrees of democratic decline

Democratic breakdown and democratic backsliding

Working from this minimalist conception of democracy, we can separate recent diagnoses of democratic backsliding into three categories: democratic breakdown, significant democratic backsliding, and over-diagnosed democratic decline. 

Democratic breakdown has become less common in the past few decades because one of the major sources, military coups, has declined. But there have been military coups that toppled nascent democracies recently in Thailand and Myanmar. And several countries that had been competitive or were becoming competitive democracies have effectively extinguished electoral competition—Presidents Daniel Ortega of Nicaragua and Nicolas Maduro of Venezuela, for example, banned their primary opposition from running in the most recent presidential elections, a common tactic in authoritarian countries like Belarus and Russia. Moreover, prominent opposition leaders in both countries boycotted the elections, providing ex ante evidence that they would not be competitive. Opposition boycotts are a strong signal that the prerequisites for open competition are not in place and that there has been almost complete democratic breakdown.

Two examples of significant democratic backsliding that have featured in many recent discussions are Turkey and Hungary. And for good reasons; each has limited open competition by placing significant restrictions on the media, with Turkey becoming one of the world’s leading jailers of journalists and Hungary’s leading media coming almost entirely under the control of Viktor Orban and his political allies. Furthermore, Turkish President Erdogan has jailed political opponents and Orban implemented electoral reforms that favored his Fidesz party in the translation of votes into representation.

But there’s still reason to believe that electoral institutions have not completely broken down in Turkey and Hungary and that Erdogan and Orban could lose significant national offices through elections. Each has lost significant local elections, with Erdogan’s AKP party losing 2019 mayoral elections in Ankara and Istanbul, the latter a historic gateway to national power, and Orban’s Fidesz losing the 2019 mayoral election in Budapest. While it appears that the primary opposition parties in Turkey will be able to freely contest the 2023 parliamentary election, Erdogan’s recent jailing of Ekrem Imamoglu, the mayor of Istanbul, who was seen as Erdogan’s leading opponent in the upcoming presidential election, suggests that Turkey might be close to full democratic breakdown in its presidential election. Orban, on the other hand, has not jailed his political opponents.

Over-diagnosing democratic decline

While most commentators agree that the above cases constitute democratic backsliding, other examples of backsliding according to some of the leading democracy indicators, including Brazil, the Czech Republic, El Salvador, India, Poland, the Philippines, and Slovenia are more questionable. In these cases, accused leaders have either lost elections and subsequently left office or face highly competitive upcoming elections that show few signs of being corrupted. 

Much of the discussion of backsliding in Brazil and Slovenia has been based on their former populist leaders’ (President Bolsonaro and Prime Minister Jansa) hostility toward the media and use of divisive rhetoric, which increased political polarization. But despite their hostile rhetoric and the worsening political climate, both had limited effects on media freedom and political competition. If anything, they increased voter engagement. While voter turnout in the Brazilian presidential election was similar to previous elections, hostility toward Jansa caused a massive increase in parliamentary election turnout—up from 52 percent in 2018 to 70 percent in 2022. Both leaders lost and neither seriously contested the results when they did. While there was political violence in the run-up to the Brazilian election and many of Bolsonaro’s supporters did not accept his loss, the election was generally considered to have been free and fair and there was no serious threat that a transition in power would not occur.

A more controversial case is Poland, which had the largest degree of backsliding in the V-Dem dataset from 2011 to 2021—from being in the top 10 percent of countries on the Liberal Democracy Index to only in the top 40–50 percent. Since taking power in 2015, the right-wing populist Law and Justice party (PiS) has increased control over the public media and implemented reforms which increased political control over the appointment of judges. But despite increasing bias in the main public broadcaster, it has had less influence over media ownership than in Hungary, and the media landscape remains competitive. While the judicial reforms raise concerns about judicial independence, the government has backtracked on some of them in response to European Union criticisms, and there appears to be little evidence that these have affected electoral competition. PiS has proven vulnerable in elections, losing the 2019 senate race, which has limited its ability to pass legislation, and narrowly winning the 2020 presidential race. Continued conflict with the EU has harmed the government’s popularity, and there’s a substantial chance that PiS will lose the upcoming parliamentary election.

How about the US? Since Donald Trump took office, the US has been one of the largest backsliders in the V-Dem data and even fell below the threshold to be considered a democracy, according to Polity. But the US is a mixed case. Some of the recent developments are examples of democratic backsliding while other accusations are overwrought.

An example of the latter is Donald Trump’s heated rhetoric and attacks on the media, which together are a large part of why he lost the 2020 election. But reforms in several states are more concerning. The most concerning is gerrymandering, in which partisan state legislatures draw electoral maps that maximize the translation of their votes into seats. While several states have given legislative district map drawing to independent commissions, gerrymandering has been legitimized by the Supreme Court and has created several highly disproportionate state legislative and congressional maps.

Other concerns about restrictions on voting are less significant. Restrictions on early voting are not common, with multiple days of early voting available in all but four states. Likewise, voter ID requirements, while providing an impediment to voting for some, are common around the world and popular among voters. While commentators have noted that threats to democracy have become more subtle and that the real danger is in the cumulative effect of small changes, we still must assess whether each of these changes constitutes a significant restriction on voting in its own right and ensure that we are not overplaying their importance.

Why have we over-diagnosed democratic decline?

One commonality among countries accused of democratic backsliding is that they had a populist leader during the period in question. There is concern that populists, who claim to represent the will of the “true people” and delegitimize their opponents, are proto-authoritarians, having the will if not always the opportunity to dismantle democracy. But while it’s important to try to assess populists’ intentions, it’s ultimately more important to assess their actions and the ability of the country’s institutions to constrain them. And most recent populist leaders have not significantly affected electoral competition, evidenced by their loss of subsequent elections. In the above cases and others like Bulgaria and the Czech Republic, the populist leader’s divisiveness generated significant backlash, which pushed the opposition parties to forge themselves into a unified front that defeated the populist. 

Another recent argument for why there has been an over-diagnosis of democratic backsliding is methodological. Measures of democracy have multiple components (freedom of expression, strength of the rule of law, and so on) that are based on subjective expert judgments rather than objective indicators. Political scientists Andrew Little and Anne Meng argue that if there’s a general perception that backsliding is occurring, reporting on issues related to these components will increase and the expert assessors may take this increased reporting as evidence of backsliding. When Little and Meng focus on objective indicators like the rate at which incumbents lose elections or the number of journalists arrested, they find little recent evidence of backsliding. Still, even if we had objective indicators for many of the democracy measure components, there would still be reason to question whether a decline would constitute democratic backsliding because many of these components go beyond what is necessary for ensuring open electoral competition.

Why should we care about over-diagnosing democratic decline?

We should be concerned about over-diagnosing democratic backsliding for a few reasons. First, it can confuse us about how upcoming elections are likely to play out. If the electoral process remains functional, voters can remove politicians who have attacked liberal rights or helped degrade sociopolitical norms, as recent election results in several countries have shown. Perhaps even more reassuringly, US voters in the 2022 midterms showed that they can distinguish politicians who threaten the sanctity of electoral institutions from those with conservative preferences, with voters in swing states rejecting almost every Trumpist who was running for a state-level office involved in election administration but electing several mainstream Republicans. Democracy analysts must be similarly judicious in distinguishing policies that degrade electoral competition from those that are undesirable, but don’t in themselves affect electoral competition.

US voters in the 2022 midterms showed that they can distinguish politicians who threaten the sanctity of electoral institutions from those with conservative preferences.

A second reason is that accusing a politician of being anti-democratic or a country of undergoing democratic backsliding is a strong charge, and if such proclamations fail to pan out repeatedly—the “anti-democratic” politician loses the next election and the “backsliding” country continues to replace politicians through them—the public may begin to lose trust in the judgments of the experts, activists, and politicians who made such charges. This lack of trust may be a serious problem if a leader comes along who undermines media freedom or electoral institutions but the public has become desensitized to such accusations. For politicians and activists, attacking opponents as anti-democratic can also appear to voters as an attempt to avoid substantive issues, particularly if leveled against right-wing populists, who argue that mainstream parties ignore their issues. This is arguably what happened in the 2022 parliamentary elections in Italy and Sweden, where charges against right-wing populist parties for being anti-democratic backfired and the center-left lost.

Worse, if politicians or activists repeatedly declare that democracy has declined and the system is rigged against them, their supporters may believe them and decide that political activism and voting aren’t worth it. Consider the Senate races in Georgia in 2020. After Donald Trump lost Georgia by just over 10,000 votes in the 2020 presidential election, he disparaged the state’s electoral process, giving his supporters the impression that the voting was rigged and their votes weren’t being counted. This likely reduced their turnout in Georgia’s runoff election for its two Senate seats six weeks later, costing the Republican candidates both elections and giving the Democrats control of the Senate.

Looking forward

Concern about the over-diagnosis of democratic backsliding should not be taken to disparage concerns about illiberal tendencies among political leaders or the strength of sociopolitical norms around the world. Many populist leaders elected in open competition have subsequently become autocrats, including those in Belarus, Nicaragua, and Venezuela. And the loss of liberal rights and the decline of sociopolitical norms may be just as harmful to individuals and damaging to society as a loss of electoral competition. Of course, there’s also likely to be a causal relationship between sociopolitical norms and open electoral competition. If sociopolitical norms remain degraded, this will embolden politicians to take steps to undermine electoral institutions. But until politicians have actually taken these steps, we should allow for the possibility that voters can correct some of these pathologies through the electoral process.

It also shouldn’t be taken to disparage the global democracy datasets, which have been invaluable to researchers. These concerns wouldn’t require them to change anything about their data collection efforts—each democracy indicator contains necessary components to measure open electoral competition and there’s nothing wrong with publishing additional measures relevant for good governance. But when these organizations announce their headline findings for their data updates, they should limit their assessment of democratic backsliding to the subset of their indicators most relevant for electoral competition. While it is also important to call out countries for a decline in rights or sociopolitical norms, they should distinguish between these and democracy as the state of electoral institutions.