The Autocrat's Corruption Dilemma

Abstract

A large body of scholarship shows that autocrats can use corruption strategically to strengthen their political hold, such as by distributing rents to their supporters. However, this scholarship often overlooks how corruption may also politically damage autocrats. I argue that corruption often brings substantial political costs alongside its advantages, resulting in a ‘corruption dilemma’ for autocrats. I show that in recent years, public anger over corruption has led to numerous anti-government protests and has been a major cause of autocrats being ousted from power. How politically costly corruption is depends on factors such as the public's tolerance for corruption, whether the autocrat is accountable to quasi-democratic institutions and whether the autocrat can credibly claim to be fighting corruption. The case of Malaysia illustrates how relying on corrupt practices to stay in power can backfire even in a long-standing authoritarian regime. My analysis advances our understanding of corruption's mixed role in authoritarian durability and authoritarian strategies of rule.

A large body of scholarship shows that autocrats engage in corruption not only to amass personal wealth, but also because they can benefit politically from it. Corruption, commonly defined as the misuse of public office for private gain, allows autocrats to provide their supporters with illicit wealth or opportunities to acquire it in return for political loyalty (Brownlee 2007; Geddes 1999).¹ This is critical, given that an extensive literature argues that authoritarian regimes stay in power by delivering material rewards to non-representative but influential political constituencies (Bueno de Mesquita et al. 2003; Svolik 2012). Whereas political leaders in democracies are accountable – even if imperfectly – to the general public, autocrats can engage in corruption and distribute resources unfairly with relative impunity. For autocrats, as Bruce Bueno de Mesquita and Alastair Smith (2011) argue, ‘bad behavior is almost always good politics’.

The strategic use of corruption to obtain political support is widespread among authoritarian regimes. In Russia, a key strategy President Vladimir Putin has used to consolidate power is making an ‘informal contract’ with wealthy businesspeople and high-level officials who ‘believe that they are entitled to rob the country blind’ in return for supporting him (Krastev and Inozemtsev 2013). Russian clientelism is reminiscent of long-time Filipino president Ferdinand Marcos's strategy of allowing supportive elites to capture lucrative monopolies over national industries and sectors (Wurfel 1988). But corrupt exchanges are not limited to those among elites; autocrats also commonly make strategic use of illicit funding in the electoral arena (Gandhi and Lust-Okar 2009). Besides using electoral fraud, media censorship and repression, autocrats can also manipulate elections by using state resources as campaign funds, bribing legislators or election officials, or directly buying votes. In Egypt under President Hosni Mubarak, the ruling National Democracy Party's electoral candidates were expected to build clientelistic networks to distribute favours and buy votes; election winners could then enjoy parliamentary immunity and make back the money in bribes and other spoils of their office (Blaydes 2011).

However, scholarship arguing that corruption helps autocrats stay in power often overlooks corruption's myriad negative effects on a country's government, economy and society. When officials and bureaucrats engage in bribery and embezzlement, the state apparatus is weakened and government policies become harder to implement or enforce (Heidenheimer and Johnston 2002). Public goods provision suffers (De Vaal and Ebben 2011). Public trust in government institutions declines (Anderson and Tverdova 2003). Corruption can undermine a country's economic performance in numerous ways, such as by raising transaction costs among economic actors, scaring off foreign investment and siphoning funds from critical infrastructure projects (Rose-Ackerman 1999). Corruption can sometimes be a tool to cut through red tape and ‘grease the wheels’ of economic development, but in the long run it is almost always detrimental to good governance and development (Méon and Weill 2010; Mungiu-Pippidi 2015).

Corruption's many negative effects produce citizen anger at powerholders, from which autocrats are not immune. Autocrats cannot be voted out of power through free and fair elections, but popular anger over corruption can erupt into large-scale protests and sometimes results in the ousting of authoritarian leaders. Autocrats can try to censor news about corruption scandals, but the internet and foreign media make this difficult, and corruption's negative effects on government and the economy are often observable. Autocrats can resort to repression to stay in power, but repression is messy and costly and can backfire by triggering more mobilization (Davenport 2007).

I argue that autocrats thus face a ‘corruption dilemma’: while corruption helps autocrats marshal political support, it can also cause public discontent that weakens and sometimes ends their rule. My research shows how, in recent years, public anger over corruption has resulted in significant political costs for autocrats. This threat is often downplayed by scholars perhaps because the political payoffs from bribing officials or vote-buying are immediate and clear, whereas the costs that corruption incurs through contributing to public discontent are indirect and sometimes unpredictable. Will the public protest if ruling party legislators are revealed to have embezzled state funds and bought luxury houses, or will most citizens shrug off the scandal? Will allowing family members and cronies to take control of key industries make an autocrat more secure, or risk a delegitimizing economic crisis? Corruption can have complex effects, leaving autocrats with hard choices.

I suggest that several factors shape how likely autocrats are to pay political costs for corruption, including: (1) the public's level of tolerance for corruption; (2) whether the autocrat is accountable to quasi-democratic institutions (QDIs); and (3) whether the autocrat can credibly present himself or herself as an anti-corruption reformer. QDIs are institutions normally associated with democracy that create limited but meaningful political competitiveness and openness within an authoritarian regime, such as semi-competitive elections or legislatures that allow some opposition parties to compete. Scholars have conceptualized authoritarian regimes with QDIs in different ways, including as ‘hybrid’ regimes (Diamond 2002), ‘competitive authoritarian’ regimes (Levitsky and Way 2002) and ‘semi-authoritarian’ regimes (Ottaway 2003). Authoritarian regimes with QDIs, such as Haiti, Serbia or Malaysia, can be contrasted with ‘fully authoritarian’ regimes, where such institutions are absent or inconsequential, such as China, Eritrea or Oman, though even regimes with QDIs fall short of being democracies. The presence of QDIs makes it more likely that autocrats will pay a cost for engaging in corruption because QDIs create institutional channels – however imperfect – that allow citizens angry over corruption to mount a challenge against the regime.

To support these arguments, this article adopts a two-part empirical strategy. First, I use David Clark and Patrick Regan's (2016) mass mobilization protest data set (V4), supplemented with news reports, to identify corruption-related protests against authoritarian regimes that began between 2000 and 2019. These data help to show the political costs that autocrats face for engaging in corruption. I find that protests have led to an autocrat's ouster from power in eight cases since 2000, and that protests have often been damaging even in cases where the regime survived. And second, I present a case study of Malaysia under the long-ruling United Malays National Organisation (UMNO) to illustrate my argument through process tracing (Mahoney 2010). This case shows that a reliance on corruption to stay in power can be a double-edged sword for autocrats even in regimes thought to be highly durable. Public anger at corruption in the ruling UMNO-led coalition rose over time and eventually exploded following revelations of extreme personal corruption by Prime Minister Najib Razak (2009–18). This public anger was a key reason why opposition forces were able to unite and win a surprise electoral victory in May 2018, ousting Najib and UMNO despite pro-incumbent electoral manipulation and other handicaps.

In this article, a national leader qualifies as an autocrat if he or she is designated as such by Milan Svolik (2012), relying on Andrew Leber et al. (2021) to extend Svolik's (2012) data set temporally. Following other scholars of authoritarianism, Svolik defines an autocrat as the leader of a regime that ‘fails to satisfy at least one of the following two criteria for democracy: (1) free and competitive legislative elections and (2) an executive that is elected either directly in free and competitive presidential elections or indirectly by a legislature in parliamentary systems’ (Svolik 2012: 22–23). As with any definition, this leaves some fuzzy boundary between democracy and autocracy in which there may be contested cases or cases that move between categories, such as Hungary's shift to authoritarianism under Prime Minister Viktor Orban. But regardless of whether these borderline cases are correctly labelled as authoritarian or not, I suggest that any leader using corrupt activities to ensure his or her hold on power will potentially incur political costs for this behaviour, creating a corruption dilemma.

By theorizing the corruption dilemma, this article contributes to the study of authoritarian regime durability and authoritarian strategies of rule. I lay out the political negatives of corruption as a counterweight to the existing literature's focus on how corruption helps autocrats stay in power. Corruption-related protests have become common enough and consequential enough that they should be studied alongside coup attempts, economic crises, factional conflict and other major risk factors in authoritarian regimes. By highlighting the costliness of corruption, this article builds on existing literature about how ‘performance legitimacy’ is a key pillar of support for authoritarian regimes (White 1986; Yang and Zhao 2015). Corruption's corrosive effect on performance legitimacy has been explored extensively in studies on China, but less often comparatively across authoritarian regimes (Zhao 2009). In addition, my finding that corruption is potentially more costly for autocrats who face QDIs cuts against scholarship that argues that semi-competitive elections and legislatures facilitate patronage distribution and otherwise help autocrats to shore up political support (Blaydes 2011; Lust-Okar 2006).

My analysis proceeds in three steps. First, I lay out the evidence that corruption is a risk factor for autocrats. Second, I examine under what conditions the corruption dilemma is more or less severe – in other words, why engaging in corruption is riskier for some autocrats than others. And third, I show how the case of Malaysia under the UMNO illustrates my argument. A brief conclusion summarizes the article's findings.

The political costs of corruption

Engaging in corruption is far from simply beneficial to autocrats, but rather often incurs significant costs. This argument is based primarily on my analysis of 77 corruption-related protests against authoritarian regimes identified between 2000 and 2019, which yields three specific findings. First, public anger over governmental corruption has been a major global factor in forcing autocrats out of power over the last two decades. Second, even when protests over corruption have fallen short of ousting autocrats, they have often threatened the stability of authoritarian regimes. Protests incur costs because repression is costly and so are alternatives to repression, such as co-optation (Gerschewski 2013). Public displays of anger at corruption can also strengthen opposition forces. And third, autocrats have responded to public anger over corruption in ways that show that they recognize how the issue is potentially damaging.

Brought down by corruption

Over the last two decades, the political fallout from corruption has been an important driver – usually in combination with other economic and political concerns – in movements that have brought down eight autocrats. Consider three notable cases. Georgia's Rose Revolution in 2003 was triggered by fraudulent elections, but long-held anger over rampant corruption and economic mismanagement drove citizens to take to the streets and topple the incumbent regime (Kandelaki 2006). Protest leader-turned-president Mikheil Saakashvili used the revolution's momentum to launch a sweeping anti-corruption campaign. In 2011, in the first major rupture of the Arab Spring, President Zine El Abidine Ben Ali fled Tunisia in the face of widespread protests over unemployment, corruption and cronyism, and poor economic conditions (Yerkes 2017). Large-scale protests began in Algeria in 2019 with ‘one of the protestors’ main concerns’ being to catch the ‘“thieves” responsible for people's socioeconomic hardship’ (Volpi 2020: 161). The protests secured the resignation of President Abdelaziz Bouteflika, but not the broader goal of removing the regime as a whole from power.

In addition to the eight cases listed in Table 1, there are others in which public anger over corruption played a less central but still contributing role to an autocrat's ouster from power. Economy-wrecking corruption and mismanagement were key grievances in the Arab Spring protests in Libya and Yemen, which developed into complex conflicts that ultimately toppled both countries’ governments (Levey 2011).

Table 1. Corruption-Driven Ousters of Autocrats, 2000–19

Note:

^a Here I make an exception to Svolik (2012) by listing Armenia as an authoritarian regime. Armenia under the Republican Party was widely seen as a non-democracy. Levitsky and Way (2010) labelled it a competitive authoritarian regime alongside Kyrgyzstan and Malaysia. Freedom House labelled it a ‘semi-consolidated authoritarian regime’.

Some critics might argue that autocrats should not worry too much about the political consequences of corruption because other risks to their rule are more serious. Studies show that the largest threat autocrats face is not from revolution but from other elites, especially coups (Egorov and Sonin 2011; Svolik 2012), but I find that fewer autocrats have fallen to coups than to corruption-related protests between 2000 and 2019. This period saw only seven successful coups against autocrats, in Mauritania (2005), Togo (2005), Fiji (2006), Egypt (2011), Burkina Faso (2014), Zimbabwe (2017) and Sudan (2019).² Moreover, the political consequences of corruption cut across the grassroots/elite divide; anti-corruption protests sometimes pressure other elites to turn against the autocrat, as in the case of Algeria above. Scholars of coups have long understood that public protests can aid or trigger coup attempts (Johnson and Thyne 2018). So without disputing the literature's finding about the severity of elite threats to autocrats’ rule, I suggest that authoritarian leaders who engage in corruption or allow corrupt practices to spread unchecked are taking a serious political risk.

Corruption-related protests pose challenges

Even when autocrats have managed to hold on to power in the face of corruption-related protests, such disruptions have often shaken authoritarian regimes, sometimes impelling them to engage in violent repression. For example, the 1989 Tiananmen Square protests arguably created the most acute political crisis that the Chinese Communist Party (CCP) had ever faced. While most famous for their demands for greater political freedoms and democracy, the Tiananmen Square protests were just as much about public anger over government corruption, especially the privileges officials enjoyed compared with ordinary people (Hsu 2001). Ultimately, the CCP leadership decided to quash the protests brutally, killing hundreds. Most Arab Spring protests did not succeed in unseating autocrats, but still posed serious and in some cases unprecedented challenges to their rule. In surveys conducted across the MENA region after the protests, ‘fighting corruption was mentioned as the most important reason for the Arab Spring by 64.3 percent of respondents’ (Arampatzi et al. 2018: 108).

A related problem for autocrats is that opposition figures can lead corruption-related protests and use them to build support. In recent years, the most prominent opposition figure to Putin's rule in Russia has been the law-maker and anti-corruption activist Alexei Navalny. Navalny's exposés of alleged wrongdoing by those in Putin's inner circle helped to spark an unprecedented wave of protests in 2017, including demonstrations in at least 187 Russian cities. Juan Guaidó, Venezuela's opposition leader and the internationally recognized president, has won public support in part by leading the charge to investigate widespread embezzlement by the Nicolás Maduro government, which has contributed to nationwide food shortages in recent years. Facing mass protests and unrest, Maduro has admitted that corruption is a serious problem in his government and has vowed to catch the ‘thieves’.

Figure 1 shows that corruption-related protests against authoritarian regimes increased in the 2010s compared with the 2000s, though this is true of other forms of protest as well (Carothers and Youngs 2015). These protests have affected regimes in every major region, as shown in Figure 2.

Figure 1. Corruption-Related Protests in Authoritarian Regimes, 2000–19

Source: Clark and Regan's (2016) mass mobilization protest data set (V4).

Note: A protest is ‘corruption-related’ if corruption is mentioned in the data set's short description of the case as a reason for the protest. I do not count multiple corruption-related protests in the same country in the same year.

Figure 2. Corruption-Related Protests in Authoritarian Regimes by Region

Source: Clark and Regan's (2016) mass mobilization protest data set (V4).

Note: A protest is ‘corruption-related’ if corruption is mentioned in the data set's short description of the case as a reason for the protest. I do not count multiple corruption-related protests in the same country in the same year.

Autocrats recognize corruption's risks

Beyond authoritarian leadership ousters and corruption-related protests, evidence that corruption is costly also comes from how autocrats have expended considerable effort to deter public mobilization over the issue. Autocrats’ first line of defence is censoring or spinning unflattering news stories that reveal their corruption. A complete information blockade has become less feasible in the digital age, so some regimes launch sophisticated disinformation campaigns. If the public still protests corruption, autocrats are faced with a choice: repress the protests, which may trigger a backlash, or make concessions, such as promising to enact anti-corruption reforms. There is some evidence that smart autocrats will do both simultaneously, an approach which stops protests while also soothing public discontent (Elfstrom 2019). Certainly, this requires authoritarian regimes to monitor public opinion carefully and to have built up institutions that can deliver targeted coercion and propaganda.

Rather than simply responding to public mobilization over corruption, some autocrats attempt to pre-empt it by launching anti-corruption reforms themselves. Autocrat-led corruption control is often dismissed as an excuse to purge rivals, but recent research shows that substantive reforms are more common than is widely assumed (Carothers 2020a). The most well-known case is probably Singapore under the long rule of Prime Minister Lee Kwan Yew; keeping corruption low helped make Singapore an attractive international business hub and burnished Lee's reputation at home and abroad as a pragmatic and capable leader, despite his rejection of Western-style democracy. But there are also other significant cases. Cuba's revolutionary leader Fidel Castro saw eliminating the corruption of the hated Batista regime as a critical step to legitimizing his revolution and building a new state apparatus (Domínguez 1978: 230; Thomas 1971: 1344). More recently, President Paul Kagame's anti-corruption reforms in Rwanda earned him widespread praise from international aid and development organizations in the 2000s, despite his authoritarian methods. In sum, some autocrats, including globally prominent figures, have rejected the logic that engaging in corruption is uniformly beneficial to their hold on power.

When is corruption costly?

This section of the article explores under what conditions corruption is more or less likely to be politically costly. Those autocrats who have been ousted by corruption-related protests did not necessarily have the most corrupt governments. President Ben Ali, for example, was ousted despite international organizations rating Tunisia as less corrupt than other countries in the region.³ Instead, I argue that we should focus on three factors: the public's tolerance for corruption; whether quasi-democratic institutions allow the public to mount an electoral challenge to the autocrat; and whether the autocrat can credibly present himself or herself as opposed to corruption.

When tolerance is low

How tolerant citizens are of corruption is, commonsensically, a key determinant of how risky it is for leaders to engage in it. People tolerate government corruption for various reasons, such as believing that it does not matter as long as officials do their jobs, that corruption is an inevitable or natural state of affairs, or that corruption is beneficial because it allows people to avoid burdensome government regulations. Though public attitudes towards officials have changed in recent decades, a common saying in Brazil for much of the 20th century was ‘he steals, but he gets things done’ (Senters et al. 2018). In countries where state power is harsh and arbitrary, citizens may routinely rely on paying bribes to protect their livelihoods or even their lives. In North Korea, ordinary citizens engaged in semi-legal or illegal commerce complain about constantly having to bribe government officials, but much prefer this system to the government actually enforcing its nominal bans on capitalist activity (Lankov 2015).

Since the 1990s, several global factors have made publics generally less accepting of corruption (Carothers 2020b). First, the end of the Cold War brought a rapid decline in the importance of ideological battles and the fear of foreign invasion (Cockcroft 2012: 8). Many corrupt regimes that had won support at home and abroad for being anti-communist and being tough on national security came under new pressure by citizens to focus on providing good governance and economic growth (Pharr and Putnam 2000). Second, global levels of wealth and education have risen substantially, though unevenly. Studies show that wealthier and better-educated populations have become less accepting of a wide range of corrupt practices and less susceptible to politicians’ attempts at vote-buying (Carrasco et al. 2020; Jensen and Justesen 2014; Truex 2011). Educated, middle-class youth played a leading role in the Arab Spring protests in several countries. Third, the spread of new technologies, such as the internet and cell phones, has allowed citizens to learn more about government corruption and organize more easily against it. For example, the Russian activist Navalny has used his popular blog to disseminate videos revealing Prime Minister Dmitry Medvedev's immense wealth, including mansions, yachts and a Tuscan vineyard. And fourth, international media and aid and development organizations such as the World Bank began in the 1990s to focus on corruption and highlight its dangers for growth in the developing world. Foreign reporting about corruption can sometimes get around domestic censorship. Autocrats in many countries scrambled to suppress the news that their family members and close associates had been named in the 2016 Panama Papers revelations, for example. All of these global trends have affected authoritarian regimes but are not limited to them; in fact, public anger over corruption has triggered the resignation or ousting of more democratic leaders than authoritarian ones in recent years (Carothers 2020b).

Citizens are the main audience that leaders face, but of course not the only one; in some cases, foreign allies become disillusioned with an autocrat's corruption and withdraw military or economic backing. Corruption may signal that an ally will waste foreign aid or is simply a bad bet who will not be ruling for long. During the Chinese Civil War in the late 1940s, US intelligence assessments suggested that the Nationalist Party's unchecked corruption was losing it the support of the Chinese people and losing it the war to the more disciplined Communist Party (see, for example, Sprouse 1947). These assessments contributed to the US urging the Nationalist Party to replace Chiang Kai-shek as its leader and temporarily suspending military aid to the Nationalist regime.⁴

When there are quasi-democratic institutions

In addition to citizens’ tolerance or intolerance of corruption, a second important factor is whether autocrats face QDIs, such as semi-competitive elections and legislatures that allow opposition parties. QDIs raise the political costs of corruption, making it more of a dilemma, because they strengthen the public's ability to challenge the autocrat over corruption. Where there are semi-competitive elections, public discontent over corruption may contribute to opposition gains in elections, which the autocrat then has to either suppress – which is costly – or accept. In Georgia in 2003 and Kyrgyzstan in 2005, autocrats were forced to cover up embarrassing election results, which in both cases fuelled protests and united opposition parties against them, leading to their ousters (Fairbanks 2004; Kupatadze 2008). In Malaysia in 2018, a newly formed opposition coalition took power through elections, despite the ruling coalition having successfully manipulated elections many times in the past. This argument follows from scholarship that shows that authoritarian elections, even if unfair, can pose serious risks to autocrats (e.g. Treisman 2017). A multiparty legislature can also pose problems, especially if public anger over corruption unifies usually fragmented opposition parties. In Iraq, Prime Minister Adel Abdul Mahdi resigned in late 2019 after failing to address massive anti-corruption protests and related violence that claimed over 400 lives; he was facing a no-confidence vote by a parliament made up of rival political blocs.

A counter-example might be Hungary under Prime Minister Viktor Orban, which has transitioned from democracy to authoritarianism with QDIs. Orban is involved in large-scale corruption – ‘us[ing] European subsidies as a patronage system that enriches his friends and family, protects his political interests and punishes his rivals’ – but does not seem to pay political costs for this behaviour (Gebrekidan et al. 2019). Numerous corruption scandals have hardly dented Orban's high approval ratings. However, some downsides of his corruption are emerging. Orban's misuse of European tax dollars has angered European governments and publics and complicated Hungary's foreign relations. At home, anti-corruption activism, though still limited, is on the rise (Pirro and Della Porta 2020). Moreover, studies show that when exposed to information about political corruption, Hungarian voters are ‘more likely to abstain from voting, less likely to support the incumbent party, and more likely to expect the anti-establishment party to win’ (Snegovaya 2020).

Notably, the countries that international indices rank as among the most corrupt tend to have fully authoritarian regimes, such as Tajikistan or Zimbabwe, not regimes with QDIs. This is not a coincidence; closed and highly repressive regimes make it exceedingly difficult for citizens to hold their leaders accountable for any wrongdoing, including corruption. Anti-corruption protests have sometimes emerged in highly authoritarian contexts, as with the nationwide strike in Guinea in early 2007, but they face harsh repression and a steep road to effect political change.

When autocrats cannot claim to be reformers

Finally, in some cases, autocrats can forestall the criticism that they are ruling over a corrupt regime by launching an anti-corruption campaign and presenting themselves as part of the solution rather than part of the problem. This is difficult to achieve due to widespread scepticism about autocrats’ motives; autocrats often claim that they are dedicated to curbing corruption but then only make superficial reforms and prosecute cases selectively. And if there are already anti-corruption protests in the streets, then an autocrat's promises to clean the house are naturally seen as motivated by self-preservation. In the 1970s, the Shah of Iran, Mohammad Reza Pahlavi, faced rising public anger over his family's extensive corruption, among other issues. The Shah tried to regain the public's trust by launching a ‘panicked’ anti-corruption campaign in 1977, but it was too little and too late (Gillespie and Okruhlik 1991: 89). Similarly, South Vietnam's President Nguyen Van Thieu fired dozens of generals and officials in response to a popular anti-corruption movement headed by Catholic priests in 1974, but succeeded in changing few minds (Goodman 1975). But if an autocrat acts preemptively and demonstrates a sustained commitment to curbing corruption, they may be able to establish a positive reputation, as in the case of Lee Kwan Yew, mentioned earlier. To give another example, Chiang Ching-kuo became premier of Taiwan in 1972 and immediately cracked down on bribe-taking and promoted reforms to curb official privileges. There was soon ‘widespread recognition’ of the positive effect of Chiang's reforms, even among the budding political opposition to the Nationalist Party, known as Tangwai (Newell 1994: 324). Corruption only returned as a major political issue in Taiwan under Chiang's successor in the late 1980s and early 1990s (Fell 2005: 55).

The corruption dilemma in Malaysia

The case of Malaysia under the UMNO illustrates how corruption can be a political tool that helps prop up an authoritarian regime while also provoking public discontent that ultimately undermines it. The long-ruling UMNO-led coalition is widely described as an authoritarian regime with QDIs; it achieved remarkable stability through its dominance in regular but deeply flawed elections (Brownlee 2007; Levitsky and Way 2002; Slater and Smith 2016). A key reason for the regime's stable hold on power was the effective use of patronage distribution, vote-buying and other corrupt practices (Ostwald 2017). But public discontent with corrupt practices rose over time and ultimately was one of the main reasons why the UMNO suffered an unprecedented defeat in the May 2018 elections, despite various unfair electoral advantages. In addition to showing the risks corruption poses, this case also illustrates how QDIs can provide the public with a steep but not impossible path to challenge a corrupt leader and ruling party. This is notable given that some scholarship suggests that multiparty legislatures and other QDIs help autocrats to manage corruption to the regime's benefit, and that scholars have specifically argued that QDIs helped the UMNO maintain its political dominance (e.g. Croissant and Lorenz 2018).

The UMNO and corruption

The UMNO remained in power uninterrupted from Malaysia's independence in 1957 to 2018, regularly holding and winning semi-competitive elections. The ruling Barisan Nasional (BN) coalition, of which the UMNO is the leading party, was born out of a conservative mobilization against the threat of communist revolution in pre-independence Malaysia. This threat brought together a broad alliance that crossed communal divisions among ethnic Malays, ethnic Chinese and other minorities into an ‘exceedingly durable’ political order (Slater and Smith 2016). The coalition won sweeping electoral victories in the early 1950s and through Malaysia's independence in 1957.⁵ A combination of successful economic policies and manipulation of the electoral system secured the UMNO and its allies continued victories. But the BN coalition did not use its dominance to quash all opposition; Malaysia had relative political openness compared with fully authoritarian regimes. While there are other long-lived authoritarian regimes, Steven Levitsky and Lucan Way (2010: 21) identified it as the longest-lived competitive authoritarian regime.

The BN coalition's strategic use of corruption increased from one decade to the next. In the late 1950s, UMNO leaders established ‘covert “special funds” to which firms could make regular contributions’ in part to reduce the party's financial dependence on the Malaysian Chinese Association (Gomez 2012: 1382). Further enhancements to party financing became necessary when the coalition almost lost the 1969 general election. The UMNO increasingly supported itself by engaging in illegal business ventures and through patronage networks among politicians and the wealthy, especially after the launch of the New Economic Policy in 1971 (Gomez and Jomo 1997: 25–27). Prime Minister Mahathir Mohamad allowed corruption to spread exponentially in the 1980s, including the ‘rampant use of cold hard cash’ to buy election results (Wain 2009: 142, 326). While campaigning for the 1982 general election, Mahathir repeatedly promised to combat corruption. He announced that he would lead a movement for ‘clean and efficient government’, and that there would be ‘leadership by example’ to curb the practice of bribe-taking among civil servants (Kapeli and Mohamed 2015: 527; Wain 2009: 142). Instead, the mid-1980s saw a campaign finance arms race, and several high-profile corruption scandals embarrassed UMNO politicians (Reuters 1985).

As in many developing countries, the Malaysian public's tolerance for rampant government corruption declined in the 1990s, especially after the 1997 Asian Financial Crisis. The financial crisis and the government's struggle to respond to it exposed to the public a great deal of economic mismanagement connected to corruption and nepotism (Saw and Kesavapany 2006: 202). By the late 1990s, the BN coalition's image was ‘deeply tainted by its association with corrupt practices’ (Hilley 2001: 97–99). Collusive links between party figures and businesses, privatization that transferred massive wealth to political insiders and now post-crisis bailouts were all targets for criticism, including from within the UMNO. William Case notes that ‘in the 1999 elections, waged in the aftermath of economic crisis, social grievances over the UMNO's corruption helped weaken its electoral appeal’ and fuelled a new rise in support for opposition parties (Case 2004: 92; see also Martinez 2001: 483).

Autocrat-led reform?

Seeing this trend in public opinion, Mahathir and his close ally and successor Abdullah Badawi confronted the autocrat's corruption dilemma. Their response was to present themselves as reformers. Speaking to UMNO colleagues in 1996, Mahathir tearily implored them ‘not to let bribery destroy the Malay race, religion and nation’ (quoted in Wain 2009: 143). He sounded a similar note in in 2001, saying, ‘I have tried asking nicely, begged and even cried … money politics is the worst kind of disease which can cause UMNO to rot from within’ (quoted in Case 2004: 92). The administration enacted an Anti-Corruption Act in 1997, which greatly improved the investigatory powers of the country's Anti-Corruption Agency and allowed for the prosecution of officials even after they left office. Abdullah took over as prime minister in 2003 while promising to make corruption control his top priority (Saw and Kesavapany 2006: 204). The Abdullah administration continued Mahathir's reform push by overseeing the passage of the National Integrity Plan (2004), as well as the establishment of the Integrity Institute of Malaysia and the Malaysian Anti-Corruption Commission.

But these anti-corruption efforts proved largely ineffective – they did not address the core corrupt practices common throughout the regime. After more than a decade in office, Mahathir's credibility as a reformer was questionable. His political motives were nakedly obvious in the administration's prosecution of reform advocate and former deputy prime minister Anwar Ibrahim for alleged corruption and other crimes. Despite making high-minded speeches about the need for clean government, Abdullah increased the use of patronage to win ‘loyalties’ within the UMNO's ranks, including arranging favourable business contracts and direct payments for numerous mid-level politicians (Case 2008: 48). Malaysia specialists assessed that anti-corruption efforts under Mahathir and Abdullah largely failed and noted that existing patterns of corruption continued throughout the 2000s (Case 2017: 639; Kapeli and Mohamed 2015; Siddiquee 2010; Wain 2009: 311). International organizations and foreign media monitoring Malaysian politics came to the same conclusion.⁶

The electoral challenge

Although far from the only factor, corruption played a critical role in the UMNO's eventual ouster from power in 2018. Prime Minister Najib Razak, Abdullah's successor since 2009, had continued his party's traditional patronage practices and ‘the excessive use of state resources for campaigning and vote-buying’ (Welsh 2013: 1). However, Najib's popularity was fatally weakened by revelations of large-scale personal corruption. Najib had allegedly siphoned hundreds of millions of dollars from the state investment fund 1Malaysia Development Berhad (1MDB) into his personal account. Around 7.5 billion dollars went missing from the fund in total. The public was incensed by the lurid details: ‘a 22-carat pink diamond necklace … for his wife’, ‘paintings by Monet, Van Gogh and Warhol’, a ‘megayacht for a family friend’ and so on (Beech et al. 2018). Many citizens saw the 1MDB scandal as representative of the broader corruption and neopatrimonialism that had come to ‘pervade the polity’ (Weiss 2019).

Malaysia's electoral system, though unfairly tilted towards the incumbent coalition through malapportionment, gerrymandering and more direct election manipulation, gave opposition parties the chance they needed to pull off a surprise win. Opposition parties capitalized on the 1MDB scandal. The opposition coalition Pakatan Harapan (PH) helped keep the scandal in the news after it first broke and made corruption control and good governance the core messages of their campaign (and roughly a third of their book-length election manifesto) (Weiss 2019: 43). Citing Najib's corruption, former prime minister Mahathir joined the opposition coalition in 2017, dramatically strengthening PH's chances of forming a non-BN government acceptable to the public. The recently fractured opposition unified around the theme of corruption and the goal of throwing Najib out. The UMNO still went into the 2018 elections with confidence and its time-tested strategy of handing out money (hundreds of millions of dollars) to secure the necessary political support. But this time, the political minuses of corruption outweighed the pluses.

Conclusion

This article has argued that for autocrats, engaging in corruption brings a mix of political benefits and risks, resulting in conflicting pressures to maintain and limit corruption – in other words, a corruption dilemma. This argument is a corrective to the existing literature's focus on how useful corruption is for autocrats, which can lead to the one-sided view that ‘bad behavior is almost always good politics’ (Bueno de Mesquita and Smith 2011). Based on an examination of corruption-related protests and their outcomes between 2000 and 2019, I found that public anger over corruption has been a major cause of public protests and movements that have challenged and in some cases capsized authoritarian regimes. I also found that autocrats expend considerable effort and resources to deter or suppress public mobilization around the issue. In some cases, autocrats have decided that the costs outweigh the benefits and have taken measures to curb corruption. Certainly, many autocrats engage in corruption with seeming impunity, and how much of a risk engaging in corruption poses varies. I identified three important factors that influence how politically costly corruption is likely to be: citizens’ level of tolerance for corruption; whether the autocrat can be held accountable through quasi-democratic institutions; and whether the autocrat is credible as an anti-corruption reformer. One implication of these factors is that the political costs of corruption have probably risen for many authoritarian regimes in recent decades because public tolerance for corruption has decreased globally and the number of authoritarian regimes with QDIs has risen dramatically since the end of the Cold War. Lastly, I illustrated my main argument in this article through a case study of corruption's political effects in Malaysia under the UMNO.

This article's findings suggest several implications for the study of authoritarianism. First, while existing scholarship emphasizes how the strategic distribution of rents to supporters improves authoritarian regime durability, this should be balanced with an analysis of how corruption and its negative effects are major risk factors of authoritarian breakdown. Second, corruption control should be assessed as a key part of performance legitimacy for authoritarian regimes generally, as is already the case in research on China and Singapore. And third, while there is a recent trend in scholarship arguing that QDIs strengthen authoritarian regime durability, in part by facilitating the distribution of patronage, my findings bolster an alternative view that QDIs often increase political risks. Semi-competitive elections in particular provide an institutional channel that publics can and have used successfully to confront corrupt autocrats.