Source: Getty

Bet on Big-Tent Opposition Electoral Coalitions to Defeat Democratic Backsliding

As illiberal leaders continue to degrade democracy around the world, some pro-democracy activists and candidates are crossing ideological divides to challenge these incumbents.

Published on March 27, 2024

Half of the world’s population will be electing new leaders in 2024, making efforts to safeguard democracy all the more important. As illiberal leaders continue to degrade democracy in countries such as El Salvador, India, and Tunisia, pro-democracy activists and candidates grapple with how best to challenge a democratic backsliding incumbent—and whether to participate in tilted elections in the first place. They have reason for hope, however; some surprising bright spots have emerged, showing that it is possible for political oppositions and civil society to defeat these incumbents.

When a backsliding leader or party creates an unlevel electoral playing field, some analysts believe that the opposition’s best bet in challenging them is to assemble a coalition that encompasses a wide ideological range. And this viewpoint seems to hold water. In analyzing elections in such contexts over the past decade, it is apparent that broad-based, preelection coalitions have had significantly better success rates against backsliding leaders than oppositions who have remained divided.

Notably, in most cases, these broad coalitions formed after a substantial amount of backsliding had already happened, making it more difficult to prevail. Diverse opposition groups may therefore have even greater success if they unite under a big tent as soon as moderate backsliding becomes obvious. The small window before backsliding becomes so severe that the opposition can no longer win often closes quickly.

Understanding and Measuring Backsliding

Democratic backsliding occurs when the quality of democracy gradually erodes under elected governments (including those elected in tainted elections). To measure backsliding, we relied on data from the Varieties of Democracy Institute (V-Dem), an expert survey that analyzes countries’ governance and provides scores on hundreds of indicators of democracy. The key variable of interest to us was the Liberal Democracy Index (LDI), comprised of sixty-five indicators aggregated into a single number to capture the strength of a country’s liberal democracy. V-Dem clarifies that liberal democracy “is achieved by constitutionally protected civil liberties, strong rule of law, an independent judiciary, and effective checks and balances that, together, limit the exercise of executive power.”

We gathered data on every election for a head of state or head of government (president or prime minister) in every country from 2013 through mid-2023—a total of 448 elections. In eighty of these elections, the opposition was running against either a backsliding candidate, a backsliding party, or both. These cases were found across Asia (including Thailand 2023, Fiji 2014, and Armenia 2013); Africa (including Benin 2021 and Zambia 2021); Europe (including Serbia 2022, North Macedonia 2016, and the Czech Republic 2018); and Latin America (including Nicaragua 2021 and Ecuador 2017).

In analyzing the eighty elections, three key takeaways emerged:

  1. Oppositions have better chances of victory in contexts of backsliding when they form a coalition.

We defined the “coalition” strategy as one in which a significant share of the country’s opposition groups successfully join in a preelectoral alliance or coalition for the purpose of ousting the country’s backsliding incumbent leader and/or party. These political unions also had to be unusual for some reason (for example, their constituent parts had ideologically diverse or even hostile, nonoverlapping political priorities/interests, or there were taboos against such alliances in the country’s political system).

We considered a coalition strategy to be “partial” in cases where the opposition took steps to build an unusual electoral coalition for the purpose of ousting the backsliding incumbent, but it only partially succeeded. In these instances, generally the groundwork was laid for a coalition, but it did not attract a significantly broad group of partners, or multiple coalitions were formed and did not consolidate under one big tent.

One example of a successful coalition was found in Brazil. Prior to the country’s October 2022 presidential election, Luiz Inácio Lula da Silva rallied a very large and diverse alliance, including parties from the political far left, center-left, and center-right into his coalition against backsliding incumbent Jair Bolsonaro. Lula chose as his running mate Geraldo Alckmin, a center-right leader whom he had defeated in the 2006 presidential election. Commentators argued that this move drew in many moderates who might not have otherwise supported Lula’s Workers’ Party. In the runoff election, a broad cross-section of society, including academics, former political rivals, economists, and the third and fourth place winners rallied behind Lula’s message of hope and the need to defend democracy to defeat Bolsonaro.

Lula’s opposition coalition was not the only one to win in the last decade. The data, summarized in Figure 1, show that in the vast majority of elections, political oppositions have not formed preelectoral coalitions. Nevertheless, when they did, oppositions had a far higher success rate: when full coalitions were formed, the opposition defeated the backsliding incumbent in 50 percent of the elections; when partial coalitions were formed, the opposition was successful in 13 percent of the elections; and when no coalitions were formed, the opposition was victorious in 23 percent of the elections. Other cases of electoral success by full opposition coalitions were found in Sri Lanka in 2015, the Gambia in 2016, Maldives in 2018, and Honduras in 2021.

However, success depends on the degree of backsliding. Figure 2 shows that all opposition strategies failed in the context of severe backsliding (defined as any total drop of -0.30 or more in a country’s V-Dem liberal democracy score over the tenure of an incumbent candidate, ruling party, or both). Only five countries in our sample fit this category: Belarus, Hungary, Tunisia, Türkiye, and Venezuela. (Poland also passed that threshold in October 2023, although this was after we stopped collecting data and therefore is not part of our dataset.) Oppositions won in 30 percent of mild backsliding cases (defined as any drop of less than -0.15) and 16 percent of moderate backsliding cases (defined as drops between -0.15 and -0.29) (data not shown). Strikingly, Figure 2 shows that coalition building helped in these cases: when oppositions formed a broad-based coalition in elections that took place in contexts of moderate or mild backsliding, they had a success rate about three times higher than when they did not form a coalition at all.

  1. Oppositions tend to form coalitions when media freedom worsens or backsliding becomes severe.

Unlikely political actors tend to join a coalition only when there is sufficient cause for concern and the upcoming election is seen as the last best chance to win. A poor media environment contributes to this perception of threats. When there is serious pressure from the government to rein in opposition newspapers, turn the state-run broadcaster into a propaganda mouthpiece, or ban forms of dissent online, it becomes strikingly apparent that a more uneven electoral playing field could soon move the opposition into unwinnable territory.

We analyzed data from the Reporters Without Borders Press Freedom Index, which uses a scale of 1 to 100 to measure not only journalist safety but also the legal, economic, and political constraints on media autonomy and independence. In the best media environments (scores of 85 to 100), we have no elections in backsliding contexts. Figure 3 shows that in satisfactory media environments, where the Press Freedom scores were between 70 and 85, not a single opposition formed a full coalition in the past ten years. But in problematic media environments, where the scores were between 55 and 70, one in seven oppositions formed a coalition. And in difficult or very serious media environments, where the scores were below 55, one in four oppositions formed a coalition.

For example, in Hungary in 2022, the country had a problematic score of 59.8, and this undoubtedly contributed to the opposition rallying behind a single candidate and new coalition: Péter Márki-Zay and the United for Hungary Alliance, which formed in 2020 specifically to contest the 2022 parliamentary election. Commentators have pointed out that by 2022 in Hungary, “every broadcasting outlet and almost all print media regularly repeated government campaign slogans.” The emergency was clear; it was time to break glass.

Likewise, coalition building is more likely in contexts of more severe backsliding. As a country declines in liberal democracy, it becomes less politically tolerant of opposition groups, which signals that extraordinary strategies are necessary to win. For example, by Türkiye’s 2023 election, the ruling Justice and Development Party had already overseen a period of democratic backsliding in which the country’s V-Dem Liberal Democracy Index score dropped -0.386. That year, two smaller coalitions joined the six-party Nation Alliance to support opposition presidential candidate Kemal Kiliçdaroğlu in a bid to oust the incumbent president, Recep Tayyip Erdoğan. Although the coalition lost, it was abundantly clear in the run-up to the election that extraordinary strategies had become necessary to stand a chance.

Figure 4 shows that in the past ten years, an opposition formed a full coalition in one-third of elections in countries with severe backsliding. In another third, an opposition formed a partial coalition (although as seen in Figure 1, none won). Meanwhile, among elections in countries with mild backsliding, only about one in ten oppositions formed a full coalition.

  1. Oppositions have sometimes won in even very challenging and constrained situations.

Despite several daunting indicators, there are at least a few cases where oppositions defeated backsliding incumbents despite the poor odds, giving hope to oppositions in even highly constrained political contexts around the world. A challenging media landscape is one such indicator. Figure 5 shows that the odds of an opposition victory significantly wane as the media environment becomes more damaged. One exception in our sample occurred, however: Mahinda Rajapaksa took Sri Lanka in a backsliding direction during his presidency from 2005 to 2015 (while a mild amount of backsliding occurred, the V-Dem LDI score started at a low level and got worse under his tenure to reach a low of 0.27). By 2015, the year in which he ran for a third term, the country’s Press Freedom Score stood at just 39.72. But, then, Maithripala Sirisena, the minister of health at the time, defected from the ruling party and won the presidential election by building a broad coalition on an anti-authoritarianism platform. 

Likewise, oppositions can still win even in undemocratic contexts. Figure 6 shows that when the backsliding country was still above a V-Dem LDI score of 0.75, indicating a relatively strong liberal democracy, the odds were unsurprisingly 50–50 for an opposition to defeat the incumbent. Those odds fell marginally in contexts where the country’s LDI score was between 0.74 and 0.50, indicating a democracy with some deficits. Once the country fell below the 0.50 mark, often indicating a slide toward electoral authoritarianism, the odds of an opposition victory fell precipitously by more than half. But even in more clearly autocratic contexts (the bottom quartile of the LDI score), oppositions won 13 percent of the time.

The country with the lowest V-Dem LDI score whose opposition nonetheless ousted the backsliding incumbent was Gambia in 2016. Adama Barrow defeated Yahya Jammeh for the presidency. The V-Dem LDI score for Gambia in 2015 was just 0.10, yet Barrow still defeated Jammeh in the vote. Jammeh seemed poised to annul the election until a force from the Economic Community of West African States credibly threatened a military intervention, forcing him to flee the country. This election was an extraordinary case that demonstrates the role that external actors can play in ensuring that democratic elections, even in autocratic contexts, can lead to meaningful change.

Finally, an even electoral playing field is helpful but not necessary for an opposition to defeat a backsliding incumbent. Figure 7 shows that the integrity of elections is more evenly distributed across the quartiles, in comparison with the LDI which includes broader aspects of rule of law, freedoms, and democratic rights. Some countries, such as Hungary, will have a relatively high clean elections score while scoring in a lower quartile on the LDI. When elections were in the top two quartiles of V-Dem’s Clean Elections Index, the opposition won about one-third of elections. The odds were worse, but not impossible in the lower half of the Clean Elections Index: the opposition still won one in eight times.

One example of a surprising opposition victory in a country with a dirty elections score was found in Honduras in 2021, when Xiomara Castro defeated the National Party’s candidate for president despite systemic unfairness in the electoral system. That year, the country’s elections score stood at just 0.218. Castro built a successful coalition in this election, largely to combat an administration with visible corruption scandals; notably, Salvador Nasralla, a key opposition figure, dropped out of the race and endorsed Castro, who then made him her running mate.


Opposition parties and candidates face a daunting task in challenging an incumbent candidate and/or party that is responsible for democratic backsliding. The backsliding often includes the closing of media and information spaces and tilting of the electoral playing field dramatically in the incumbent’s or party’s favor. However, despite autocratic contexts, poor media environments, and unclean elections, opposition candidates have still sometimes won in the past ten years.

Two opposition victories that occurred after our data collection was completed further illustrate that it is possible to defeat backsliders by joining a coalition before or after the election. First, in Poland in October 2023, opposition parties defeated the governing PiS party that had overseen severe democratic backsliding (the country’s LDI had dropped 0.42 since the party took office in 2015). Strong youth and women’s mobilization in the wake of severe judicial backsliding and a near-total ban on all abortion rights helped propel a record voter turnout overall and among youth and women specifically. A formal preelection pact in the senate election and a postelection coalition in the lower house led to opposition victories in both.

An even more impressive opposition victory occurred in Guatemala’s presidential election in August 2023, when a little-known anticorruption reformer, Bernardo Arévalo of the Seed Movement, defeated the ruling elites. Guatemala’s political system represents a form of elite collusion where political parties with ties to organized crime seek to govern with impunity. In the 2023 elections, the governing authorities had disqualified on dubious pretenses four reformist opposition candidates, including an Indigenous candidate working to build a broad coalition. As a result, Arévalo (initially polling below 5 percent) emerged as the surprise second place winner and forced a runoff. Despite numerous government efforts to disqualify Arévalo and his party both before and after the runoff election, targeted U.S. sanctions and a broad civil society mobilization led by Indigenous groups and including the leading business association eventually resulted in Arévalo’s victory at the polls and inauguration on January 15, 2024.

Defeating backsliding incumbents does not guarantee smooth sailing after the election, as autocratic enclaves in the government structures commonly remain and may impose enormous obstacles to democratic reformers, as in Poland today. Or reformist governments may not last and backsliding incumbents and parties may return, as happened in Sri Lanka.

Nevertheless, the first step to reversing backsliding is an electoral victory by pro-democratic political forces. Our analysis shows that the hard task of building a big-tent, cross-ideological electoral coalition substantially increases the probability of defeating a democratic backslider. But it also shows that oppositions often wait until media freedom and democracy erode significantly before they form coalitions. By that time, the odds are stacked even higher against them.

More research should be done to analyze how to form and maintain highly effective coalitions, as well as what specific electoral messages and strategies produce the best outcomes.

Appendix A: Methods

We used V-Dem’s Liberal Democracy Index, V. 13 (ranges from 0 to 1) to identify cases in which democratic backsliding occurred under the incumbent party or leader running for reelection. Specifically, we identified cases where a country’s LDI score at the time of the national election of interest had fallen below the lower bound of the original confidence interval from the start of the tenure of the ruling party, the incumbent candidate, or both. We excluded elections in countries whose LDI regressed after starting out as closed autocracies because we consider these to be examples of autocratic hardening rather than democratic backsliding.

Figure A1 illustrates our case selection and measurement. In this example, we count the 2020 U.S. elections as occurring during backsliding under incumbent Donald Trump and the ruling Republican Party because in the year before the election (2019), the country’s LDI score (0.72) was below the lower bound of the confidence interval (shown in yellow) for 2015—the year before Trump and the Republican Party were elected (0.82). Thus, democratic backsliding occurred during the incumbent candidate’s and party’s time in office prior to the 2020 election.

The backsliding episodes start in the year before the incumbent candidate’s or party’s mandate began and end in the year before the election in question. This is because elections can usher in new leaders who influence the quality of a country’s democracy, and thus its ranking, within the same year. Slovenia’s Prime Minister Robert Golob, for example, won parliamentary elections in April 2022 against Janez Janša and immediately made reforms that gave his country a large boost in its 2022 LDI score.


The authors would like to thank two Cornell University students, Shaye Butler and Kidest Shemeles, for research assistance in coding the election coalitions.

Carnegie does not take institutional positions on public policy issues; the views represented herein are those of the author(s) and do not necessarily reflect the views of Carnegie, its staff, or its trustees.