Defining Terms and Concepts

Two terms in this article warrant attention in an effort to ensure precision in interpretation. First, I use the phrase colored revolutions to refer to the “Bulldozer Revolution” in Serbia in 2000, the “Rose Revolution” in Georgia in 2003, the “Orange Revolution” in Ukraine in 2004 and the “Tulip Revolution” in Kyrgyzstan in 2005.¹¹

Although only one of these revolutions actually had a color as its symbol, the term colored revolutions has become a popular shorthand for referring to the events in these countries among regional specialists and local politicians. The use of the word “revolution” is not meant to imply any long-term consequences of these events, but rather only to identify that the anti-regime forces were in fact successful in overthrowing the current regime.¹²

Second, the term electoral fraud is intended to refer to situations in which electoral results were knowingly tampered with in an effort to advantage one candidate (or set of candidates) over another. One can conceive of two different types of fraud: minor electoral fraud, which affects elections in which results were tampered with but in which the tampering is perceived to have had little effect on the overall outcome of the election, or major electoral fraud, which affects elections in which electoral fraud is suspected to have influenced the overall outcome. In the context of the colored revolutions, this would imply a belief that either a different candidate would have been elected president or a different party would have controlled the parliament if the fraud had not occurred; and indeed, all four colored revolutions featured instances of what I have labeled major electoral fraud. The ideas contained in the remainder of this article focus on circumstances following this kind of major electoral fraud. Put another way, the arguments that follow assume that if the fraud is corrected, there is a strong chance that a different group of political forces would come to power.

Additionally, it is worth clarifying what exactly I mean by “grievances against the government” or “government abuses” of its population. Here I build on Barry Weingast's concept of citizens who are or are not able to place “appropriate limits on the state.”¹³

One would expect citizens living in such countries to have “grievances against the government” and to suffer from “government abuses.” Examples of such behavior could include state agents engaging in acts of petty corruption, such as demanding bribes to perform routine state functions, or grand corruption, such as outright theft of state resources by upper level members of the government. It also could include a state that does not respect its citizens' political or civil rights. Even beyond examples of corrupt behavior or specific violations of rights, this sense of a non-limited or abusive state may extend to the feeling that citizens are not treated with the respect and dignity they deserve. This sentiment is well captured in a quote from Katya Kishinskaya, a 32-year-old pro-Western marketing executive for Nestlé in Kiev who before the Orange Revolution said of her trips abroad, “every time I return to Ukraine, I feel how different the regime is here. I feel it at the border, at customs. The people who have power don't respect you.”¹⁴

Obviously this is a fluid concept, and one that can be operationalized in many ways. We could think of it dichotomously: countries where Weingast's conditions for restraining the state either are or are not met. We also could think of it as more of a continuous variable, with citizens in countries such as the Khmer Rouge's Cambodia or Idi Amin's Uganda facing a much more abusive state than in Kuchma's Ukraine, which in turn was more abusive of its citizenry than, say, the Swedish government at the time. While there is no perfect measure of government abuse, certainly one potential proxy would be the overall level of corruption in a country; another could be the extent to which basic civil rights are respected. But regardless of how the concept may best be measured for empirical tests in the future, this general description should be sufficient for characterizing the basic argument of this article. For people with serious grievances against their government, it is difficult to act upon these grievances in day-to-day life; electoral fraud and, especially, major electoral fraud can change this situation significantly.

The Colored Revolutions

Before presenting my argument, I want to first provide some background information on the colored revolutions discussed in this article. Each shared the common feature of being centered around a fraudulent election. In all four cases, an election was held and results were widely viewed to have been seriously manipulated by the current regime. As a result, mass protests broke out in all four countries, although the size of these protests varied. After some period of uncertainty, the incumbent president either resigned from office and/or the election results were overturned, resulting in a member of the opposition becoming the new president of the country.¹⁵

Of course, in practice this played out in a different manner in each of the four countries. In Serbia, incumbent president Slobodan Milošević scheduled early presidential elections for September 24, 2000, assuming that he would easily win, even though his term as Yugoslavia's president did not expire until July 2001. His primary opposition was Vojislav Kostunica of the Democratic Party of Serbia, who was the nominee of the Democratic Opposition of Serbia (DOS), a broad alliance of Serbian opposition groups.¹⁶

On September 26, the Yugoslav Election Commission announced that Kostunica had gained the most support with 48.2 percent of the vote and in a second round run-off with would face Milošević, who had come in second place, since neither candidate had received the necessary 50 percent of the vote to be declared the winner.¹⁷ The DOS, on the basis of its 25,000 election monitors, announced that by its count Kostunica had won 55 percent of the vote to Milošević's 35 percent with 97.5 percent of the ballots counted, and Kostunica immediately announced that he would not participate in the run-off because the official results represented “political fraud and obvious stealing of votes.”¹⁸ The following day, opposition leaders called for a “total blockade” of all institutions in Serbia and over 200,000 people took to the streets of Belgrade.¹⁹ On October 5—the day after the Serbian Constitutional Court had annulled the election results, effectively denying Kostunica victory—over half a million people marched on Belgrade and quickly seized control of major government institutions including the parliament and Serbian state run television. Protestors used a front-end loader to break through barricades, thus earning the events the name of the “Bulldozer Revolution.”²⁰ The following day, Milošević resigned in a late night television appearance and the Yugoslav Constitutional Court declared Kostunica the winner of the election.²¹

The Georgian Rose Revolution unfolded in a similar manner almost three years later, although the crisis was sparked by a legislative, as opposed to presidential, election.²²

Following the November 2, 2003, Georgian parliamentary elections, small-scale protests began in Tbilisi as various reports of fraud began to accumulate. However, after the release of a set of particularly suspect results from the region of Ajaria, the size of the crowds began to increase substantially.²³ By the time the Georgian Central Election Commission proclaimed the “official” results, handing victory to the pro-President Eduard Shevardnaze “For a New Georgia” bloc, the crowds had swelled to over 100,000 people.²⁴ On November 22, these protests came to a head as Shevardnaze tried to address the inaugural session of the newly “elected” parliament.²⁵ Led by opposition leader Mikhail Saakashvili—holding a rose to show he was unarmed²⁶—protestors stormed the parliament demanding Shevardnadze's resignation; Shevardnaze was ushered out of the building by his bodyguards without a shot being fired.²⁷ After originally claiming that he would not step down, Shevardnaze resigned the following day.²⁸ Shortly thereafter, Georgia's Supreme Court annulled most of the results from the parliamentary election, and new presidential elections were set for January 4, 2004, which Saakashvili would go on to win with 96.2 percent of the vote in what was essentially an uncontested election.²⁹

The following winter, a presidential election set the stage for Ukraine's Orange Revolution. On October 31, 2004, Viktor Yanukovich, the officially anointed successor to the regime of outgoing president Leonid Kuchma, and opposition leader Viktor Yushchenko received 39.3 percent and 39.9 percent of the vote, respectively, in the first round of the Ukrainian presidential election.³⁰

As neither surpassed the 50 percent threshold necessary for a first-round victory, the two advanced to a November 21 run-off. Following a second round of voting marked by wide-spread instances of voter fraud—including the illegal expulsion of opposition representatives from election commissions, multiple voting by busloads of people, absentee ballot abuse, and an extraordinary high number of mobile ballot box votes³¹ as well as dramatic changes in turnout figures in eastern Ukraine, where Yanukovich's support was stronger³²—the Central Election Commission declared Yanukovich the winner by a 49.5 percent to 46.6 percent margin.³³ Armed with nonpartisan exit polls suggesting that actually Yushchenko had won by a 52 percent–43 percent margin, supporters of the challenger took to the streets of Kiev in protest, now famously adorned in orange clothing representing the color of Yushchenko's “Our Ukraine” parliamentary bloc.³⁴ The extremely well-organized opposition settled in for the long haul, complete with tents for housing, stages for bands and speeches, bans on the use of alcohol and even regularly scheduled garbage removal.³⁵ Over the following weeks, the protests continued in Kiev despite frigid conditions as the crisis was resolved peacefully through the use of existing institutions. Following a parliamentary resolution declaring the results invalid (November 27) and a vote of no confidence in the government (December 1), Ukraine's Supreme Court on December 3 somewhat stunningly—and in marked contrast to the actions of the Serbian Constitutional Court during the Bulldozer Revolution—declared the second round election results to be invalid and ordered that the round be run again on December 26.³⁶ This “second” second round went more smoothly with over 12,000 international observers present, and Yushchenko won by a 52.0 percent to 44.2 percent margin.³⁷

As in Georgia, the Kyrgyz Tulip Revolution took place after allegations of major electoral fraud in a legislative election, although again it culminated with the resignation of a president. The February 27 and March 13, 2005, parliamentary elections followed a change in the composition of the Kyrgyz parliament and the election rules used to select the members for that body; both sets of changes were seen as an effort by current Kyrgyz President Askar Akayev to stack the parliament with his allies, including members of his own family.³⁸

After two rounds of elections marred by accusations of fraud and regime interference—Kilmo Kiljunen, the Finish head of the OSCE observer mission, reported that the election was “undermined by vote buying, deregistration of candidates, interference with media and worryingly low confidence in judicial and electoral institutions on the part of voters and candidates”³⁹—the opposition won a total of only 6 out of 75 seats.⁴⁰ On March 15, protests began in the southern Kyrgyz city of Jalalabad demanding Akaev's ouster, and over the next week would spread to other parts of the country including Talas and Osh.⁴¹ These protests would culminate nearly two weeks later with 30,000 protesters converging in the main square of Bishkek, Kyrgyzstan's capital, and the subsequent flight of Akayev to Russia.⁴² Unlike the Georgian and Ukrainian protests, however, the Tulip Revolution had its violent aspects, with the reported death of six protestors and looting associated with the protests causing an estimated $100 million in damage.⁴³ Although Akayev eventually agreed to resign on April 4, leading to the election of opposition leader Kurmabek Bakiyev as president on July 10 with over 88 percent of the vote, the fraudulently elected parliament was allowed to continue to operate without new elections as part of negotiations among the various opposition figures.⁴⁴

One additional common feature across the four colored revolutions is that they all took place in extremely corrupt societies. At the time of the Bulldozer Revolution, Serbia ranked 89th out of the 90 countries in Transparency International's Corruption Perception Index (CPI), scoring a 1.3 on a 0 (highly corrupt) to 10 (highly clean) score; only Nigeria was judged to be more corrupt.⁴⁵

The year of the Rose Revolution, Georgia ranked 124th out of 133 countries, with a score of 1.8.⁴⁶ In 2004, Ukraine scored 2.2, which was good enough to tie with five other countries for 122nd–128th out of 145 countries.⁴⁷ And in 2005, Kyrgyzstan tied for 130th–136th place out of 159 countries.⁴⁸ So not only were all four of these countries plagued by corruption, they were actually some of the most corrupt countries in the entire world.⁴⁹ Put in the context of the terminology of the previous section, therefore, there would be little reason to expect citizens of these countries not to hold significant grievances against their respective regimes.

Explaining the Colored Revolutions

Broadly speaking, two general sets of explanations for the colored revolutions have been proposed in the academic literature.⁵⁰

The first set of explanations places these events in a larger framework of the post-cold war world of East vs. West in Europe. This argument usually takes one of two related forms. The first approach stresses the lure of the West for members of the opposition fearful that their countries' current leadership could be leading the country down a path of permanent exclusion from “Europe” generally and European institutions such as the EU more specifically.⁵¹ For example, Kuzio notes that in Ukraine, “inspired by the transformation of their country's western neighbors from Soviet satellites to EU members, the young generation supported Yushchenko's vision of a democratic and ‘European’ Ukraine.”⁵² In Georgia, a bodyguard of one of the opposition leaders reflected on the success of the Rose Revolution by noting that “we have done it without bloodshed. This is the first time such a thing has happened in Georgia. It means that finally we deserve our place in Europe.”⁵³ This draw of Europe was not lost on the opposition leaders either. In Georgia, Saakashvili described the Rose Revolution as a “European-type velvet revolution,”⁵⁴ while in Ukraine Yushchenko promised in his inaugural speech to lead the country “into the mainstream of Europe” stating that “we are no longer on the edge of Europe… We are situated in the center of Europe.”⁵⁵

Especially popular in Russia is a second form of this East-West story, which places these revolutions squarely in the context of post-cold war geopolitical struggles for dominance between Russia and the West, particularly in areas that had long been considered in the Soviet sphere of influence. For example, noted Russian political consultant Vyacheslav Nikonov commented before the Ukrainian elections:

The more nefarious view of this struggle attributed the success of the various colored revolutions to CIA plots carried out by American-sponsored NGOs, and in particular George Soros and his Open Society Institute.⁵⁷

As an article in Russia's Nezavisimaya Gazeta noted, “Russian political analysts are fond of referring to divisions of the Soros Foundation and Carnegie Center that ‘force democracy’ all over the world.”⁵⁸ Kuzio also highlights “the fact that [Ukraine's student-led movement] PORA! was modeled on Serbia's OTPOR and Georgia's Kmara gave substance to the allegation that the Orange Revolution was imported from the United States via Serbia and Georgia…. Claims that it was an American plot … were widely accepted in Russia.”⁵⁹ The fear in Moscow was that all of these revolutions were merely a prelude to the big prize—the overthrow of the Russian government. As Andrei Vladimirov put it in Itogi, a Russian political weekly, at the time of the Orange Revolution, “The day before yesterday: Belgrade. Yesterday: Tbilisi. Today: Kiev. Tomorrow: Moscow.”⁶⁰ A slightly less conspiratorial and geo-strategic version of this explanation nonetheless continues to place the focus on the role played by external actors in assisting domestic opposition leaders, especially in terms of election monitoring.⁶¹

The second general approach to explaining the colored revolutions has been to focus on the nature of the opposition movements in each of the countries. In particular, commentators have highlighted how opposition groups learned over time to become more effective, be it through their own experiences or through the transmission of information from external actors (and in particular actors from countries that had already experienced their own colored revolution). For example, in referring to the “Kuchmagate” scandal which began with revelations of corruption and impropriety in the regime of President Leonid Kuchma in November 2000, Kuzio notes that in Ukraine:

Similarly, Bieber focuses on the importance of unified civil society in Serbia, and McFaul notes the importance of a unified opposition more specifically.⁶³

Likewise, Beissinger applies the concept of modular revolution he previously used for understanding nationalist movements in the former Soviet Union to the colored revolutions and highlights the effects of “structural advantages” for revolution (which includes the presence of transnational NGOs) in conjunction with “the power of example,” by which opposition elites learn from prior revolutions.⁶⁴ Evidence of links across opposition elites are prevalent in media descriptions of the revolutions as well. For example, during the anti-Shevardnadze protests in Georgia, supporters of Saakashvili were seen waving a sign with the Serbo-Croat words “Gotov Je” (“He is through”),⁶⁵ and both Ukrainian and Georgian activists had traveled to Serbia for training in the period before the Rose and Orange Revolutions but after the Bulldozer Revolution.⁶⁶

A third important line of explanation has been advanced by Lucan Way. Building on his previous work on the phenomenon of “unconsolidated authoritarian regimes”,⁶⁷

Way has argued that the success of the Orange Revolution can best be understood as a consequence of Kuchma's inability to successfully “consolidate” an authoritarian regime in Ukraine.⁶⁸ Here, the emphasis is on the inability of government elites to prevent a successful uprising, as opposed to either the ability of international forces or opposition elites to deliver one.

My goal in this article is not to deny the importance of any of these explanations—and clearly, all of them do add in important ways to our understanding of the colored revolutions—but it is interesting to note that they are all essentially elite-based analyses. In the following section, I explore what happens when we place the mass publics that participated in the protests leading to these revolutions closer to the center of the analysis.

Collective Action Problems, Major Electoral Fraud, and Grievances against the State

One way to think about the challenges faced by citizens with strong grievances against their governments is the framework of the collective action problem.⁶⁹

The idea of a collective action problem is a very familiar one for social scientists; indeed, typing “collective action” into Google Scholar results in over 100,000 hits. Simply put, collective action problems characterize situations in which a group would benefit from cooperation, but the lack of individual incentives to engage in the actions necessary to achieve this cooperation prevents the goal from being attained. Each individual faces a cost to participating in the action necessary to achieve the group goal, a benefit that they will enjoy from the group goal being attained, and a belief about the likelihood of the group successfully attaining that goal. If individuals believe that their cost of participation outweighs the benefits to be gained from the group goal, then they will choose not to participate. And if enough individuals choose not to participate—the threshold obviously varies widely across different collective action problems—then the goal will not be achieved. Similarly, if individuals value the group goal but believe that the likelihood that it can be achieved is sufficiently low, then again they will choose not to participate and the goal will not be achieved. Solving collective action problems, therefore, depends on the costs of participation, the benefits of the goal being sought, and beliefs about the likelihood that the goal can be achieved.⁷⁰

Abusive or unrestrained states present a classic collective action problem.⁷¹

Most members of society would likely agree that society as a whole would be better off with a less abusive and appropriately restrained state. This is not to deny that there are individual actors in society who clearly benefit from these types of arrangements. Nevertheless, the assumption that most citizens would prefer not to have to pay bribes to policemen, health care workers, and government bureaucrats, and would also prefer not to have government officials stealing public funds and using their public positions for private financial gain, seems to be a reasonable one. Achieving this goal in states where such abusive actions regularly take place, however, requires confronting these abuses and attempting to stop them. This can take a variety of forms, but all share two common features. First, there is a cost to any individual in taking any of these steps, from the relatively minor loss of time to the potentially major loss of livelihood or life. Second, the likelihood of success is always questionable, and this is especially so for individuals facing petty corruption in the course of daily life. While it may be possible to avoid one bribe at one point in time, the chance of this action affecting the overall nature of state behavior is beyond minimal. This combination yields the familiar result predicted by the collective action framework: individuals “shirk” and tolerate whatever actions on the part of the state that have given rise to their grievances, and as a result everyone is worse off from having to continue to live under an abusive regime. Readers should note that in order for the following argument to hold, it is not in any way necessary that all citizens perceive the same level of grievances, or even the same grievances for that matter. The key point is that enough citizens hold grievances against the regime that they could effect change if they managed to work together, but are customarily prevented from doing so by the logic of the collective action framework.

Major electoral fraud, however, can help solve this collective action problem.⁷²

It can do so by both lowering the costs of participating in anti-regime actions or increasing the likelihood of a successful result stemming from those actions. In terms of costs, we need to compare the likelihood of punishment from combating an unrestrained state in standard situations with the likelihood of punishment from attempting to combat such a state in the circumstances following major electoral fraud. If an individual refuses to pay a bribe to a policeman, it is possible that the policeman will just get frustrated and eventually walk away. But it is also possible that the individual will be hauled into the police station, spend time in jail, and end up being forced to pay an even bigger bribe to be released.⁷³ If an entrepreneur refuses to pay a bribe to fire inspectors, then she may find the opening of her new restaurant delayed inevitably while she continues to face all of her overhead expenses. And if an investigative journalist chooses to expose corruption in the press, then it is possible that she will pay for this decision with her life; indeed, in Ukraine, the “Kuchmagate” scandal referred to earlier in this article focused on the decapitation of journalist Georgi Gongadze, “a longtime critic of the [Kuchma] regime and a crusader against corruption.”⁷⁴ While the “costs” associated with combating an unrestrained state differ significantly across these examples, they all share the common feature of almost complete certainty that if there is punishment associated with defying the regime, it will be borne by the person doing the defying. There is usually little room for group solidarity in any of these actions, whether it involves being alone with a policeman in a dark alleyway or having one's name in a newspaper byline.

Major electoral fraud, however, changes this calculation dramatically. For once, the entire country is experiencing the same act of abuse simultaneously; in the language of the collective action literature, major electoral fraud provides an obvious focal point for action.⁷⁵

People no longer have to choose whether to react alone. Especially as crowds grow, individuals know that they will only be one of many, many people protesting, and thus much less likely to be punished individually. (Indeed, observers of the Orange Revolution in Ukraine suggest that during the first days of protest some supporters held back precisely to gauge the likelihood of punishment in the form of an armed response.)⁷⁶ This does not mean that there is no chance of punishment—and it is certainly not meant to deny the bravery of citizens who risked harm to participate in the revolutions described above—but only to note that major electoral fraud presents an opportunity to act on grievances against the current regime without a high degree of certainty that punishment—if it is forthcoming—will be felt by you individually.

Simultaneously, major electoral fraud followed by large scale protests can dramatically increase the likelihood of a successful “result” from one's participation in an anti-regime protest. As Thompson and Kuntz noted in Serbia, “people showed greater readiness to take part in antiregime protests because such activities were now considered much more likely to succeed. The sense of powerlessness that had kept Serbian society paralyzed for so long gave way to a pervasive feeling of political possibility, as Milošević's image of omnipotence was fundamentally shaken.”⁷⁷

This is of course not to say that all large scale protests against abusive regimes are successful. But as my point here is not to explain successful regime change but rather the individual-level motivation that leads people to participate in protests, a belief in at least the possibility of success is important. Again, it is instructive to consider the conditions created by major electoral fraud in comparison to the conditions faced by people in their day-to-day lives. There is almost no chance that the refusal to pay a bribe will lead to any noticeable change in the overall level of abusive behavior by a regime. Perhaps a particularly influential journalist could effect important change—and here the case of Gonganze in Ukraine is at least somewhat illustrative—but this is not an opportunity that most citizens have every day. By contrast, major electoral fraud offers hope for greater success in two important ways. First, if protests follow the fraud, then immediately there is the opportunity to speak out with a much stronger voice than anyone could have alone. But perhaps even more importantly, fighting major electoral fraud holds open the hope of changing who actually wields political power in a country: if you are successful, the bums actually can be thrown out.

Belief about the likelihood of success is also likely to be affected by past experiences. Previous mass protests—such as the Kuchmagate protests in Ukraine or the Serbian protests in the winter of 1996–97—can teach citizens important information about the likelihood of punishment for participating in future protests. Similarly, the experiences of other countries can provide important lessons about the potential likelihood of success. Once the Serbs have successfully thwarted Milošević's attempt to steal a Serbian election, Georgians who have to decide whether or not to protest Shevardnaze's attempt to steal a Georgian election have that much more reason to be optimistic about success in their own country. So the collective action framework allows for learning across revolutions as well, but emphasizes learning that goes on at the mass level as opposed to merely at the elite level.⁷⁸

Moreover, there are additional reasons why elections can be particularly effective focal points for solving collective action problems. First, they come with a limited time frame—the matter generally needs to be resolved before the fraudulent winners take office—which again can encourage citizens to participate by enhancing the belief that other citizens will participate at this point in time as well.⁷⁹

Put another way, this can explain away a part of the free rider problem associated with collective action problems. I may prefer to sit on my couch and let someone else do the protesting, but I know that if enough protesters do not appear in a very short time frame, then the opportunity to confront the regime will be lost. If I care about confronting the regime, then the costs of inaction at this particular moment are significantly higher than they normally are. Second, elections draw international attention. This can take many forms, including international elections observers, foreign media, statements by foreign leaders, and even direct attempts at mediation of post-election crises (as was the case in the Orange Revolution). Regardless of the form it takes, international attention—especially for regimes holding elections in part to seek international legitimacy—should serve to decrease the perceived likelihood of an armed response to protests. To the extent that citizens believe an armed response by the regime to be less likely, this should also help decrease the perceived costs of participating in protests against the regime. This, too, can be thought of in terms of reducing the appeal of free riding; I will likely find the free rider option much more attractive if I have concerns that participation will lead to me being shot. Additionally, elections can help provide an incentive for reducing divisions among opposition leaders. Crucially, this incentive develops in the time leading up to the election, especially if opposition leaders are able to agree on a single opposition candidate. Thus when the electoral fraud does occur, citizens are more likely to get a clear signal from opposition leaders regarding both the logistics and goals of the protests, which in turn should help increase their belief in the likelihood of the protests succeeding.⁸⁰ Finally, elections occur at regular and repeated intervals, which increases the likelihood that both of these events—international attention and opposition coordination—can occur.

It is important to note what major electoral fraud does not do as well, which is to completely solve the free rider problem associated with all collective action problems. There is nothing inherent in the nature of elections or electoral fraud that explains why once there are 100,000 protestors on the street, the 100,001st protestor might not prefer to stay home where it is warm than to join the protests. That being said, thinking of citizens with grievances against abusive regimes in terms of a collective action framework points us towards the importance of individual considerations of the expected costs and benefits of confronting an abusive regime. This in turn leads to all of the preceding observations in this section regarding how major electoral fraud both lowers the costs and increases the expected benefits of anti-regime activity as compared to normal day-to-day life. So, put another way, major electoral fraud may not be able to explain the presence of the 100,001st protestor, but it can help us to understand why all 100,001 people are more likely to be engaging in anti-regime action at that time than any other time. This is of course not to say that changes in individual cost-benefit analyses are the only reason why people take to the streets. There are obviously many reasons why people could have chosen to participate in the protests of the colored revolutions, including feelings of duty, the sense that it was something fun to do, and the desire to be on the winning side of a conflict. Nevertheless, the fact that major electoral fraud can help solve a collective action problem for citizens with grievances against an unrestrained regime deserves scrutiny as an important factor as well. Furthermore, if the mechanism I have described here is important, then it leads to a number of interesting implications for how we ought to interpret these colored revolutions.

Implications of the Collective Action Framework for the Colored Revolutions

In this section, I highlight five particularly important implications of thinking about the colored revolutions through the lens of major electoral fraud as a means of solving collective action problems. First, this framework clearly warns against overstating the likely reservoir of public support for the new regimes that emerged after the colored revolutions. For example, the Orange Revolution in Ukraine should not be interpreted as outrage over the fact that Yushchenko specifically was denied the presidency, but rather over the abuse of power by the Kuchma regime that could allow Yushchenko to be denied the presidency. This subtle distinction is meaningful, as it implies that Yushchenko should have a shorter leash from the Ukrainian public than we might normally expect from a conquering hero swept into office by an adoring public.⁸¹

And in fact events in Ukraine seem to confirm that Yushchenko was indeed on such a short leash. As early as March 2005, Ukrains'ka Pravda had published an article entitled “Did We Win?” noting that “What specifically has changed around here? Only the faces.”⁸² Within half a year of the conclusion of the Orange Revolution, the Kiev Post—an English-language Ukrainian newspaper that strongly supported the Orange Revolution—was running editorials entitled “An Utter Disgrace” and asking “Just how is President Yushchenko's government more honest than former President Kuchma's?”⁸³

Similarly, we should be cautious about assuming that whenever pro-democracy forces have overthrown illiberal regimes in post-communist countries, pro-Western policies will follow. If the revolution is in large part a reaction on the part of the mass public to a domestic problem (grievances against an insufficiently restrained state), then attaching excessive geo-strategic significance to the change in regime may not be warranted. Thus, it is not surprising that in Serbia Kostinica could over throw Milošević, a long-standing foe of the West, and yet continue to pursue a brand of Serbian nationalism that was at times hostile to Western interests. Nor should it be surprising when the new foreign minister of Kyrgyzstan following the Tulip Revolution declares that “Russia is our close ally” and visits Moscow within weeks of coming to power.⁸⁴

Moreover, it suggests that those who want to read a Western victory over Russia into the Orange Revolution ought to pay closer attention to opinion polls taken in Ukraine around the time of the rerun of the second round of the election showing that 83 percent of Ukrainians had a favorable opinion (generally good or very good) of Russia.⁸⁵

The third implication of applying the collective action framework to these revolutions is that it suggests that so-called “delegative democracies” or “competitive authoritarian” states may be somewhat more fragile than originally thought.⁸⁶

The events in these four countries demonstrate that there are inherent risks in trying to falsify election results. As I have noted previously, this article makes no predictions concerning when protests will successfully dislodge regimes from power. That being said, it is clear from the colored revolutions that one consequence of protest following major electoral fraud can be regime change. Thus trying to repeatedly carry out “fake” elections so as to maintain the international appearance of a democratic regime could turn out to be a problematic strategy. There are of course many examples of authoritarian or quasi-authoritarian countries that have carried out “fake” elections without losing power. Nevertheless, the examples of the colored revolutions demonstrate that such outcomes are not necessarily guaranteed a priori. Ironically, this may end up pushing some of these competitive authoritarian regimes back towards the old Soviet-style elections where even the pretense of competition is abandoned as these regimes learn that trying to manage exactly how many votes the opposition can accumulate on election day may prove dangerous. Or, put another way, leaders may learn that the safer way to manipulate election results is in controlling access to the ballot box (and media) as opposed to trying to change results once voting has begun. Just as pro-democracy forces have learned from past experiences, so too can leaders of quasi-authoritarian regimes.⁸⁷

With this in mind, a fourth implication of this article is that it calls attention to the importance of the way in which the existing regime reacts to opposition protests. More specifically, it highlights the fact that in all four of the colored revolutions addressed in this article, the military and security services refrained from carrying out any armed attacks against opposition protesters.⁸⁸

Such an outcome was not foreordained, but it all four cases it kept the cost of participating in protests for individuals lower than it would have been had security forces interfered.⁸⁹ This observation, too, likely has not escaped the notice of leaders of other quasi-authoritarian regimes. Indeed, it seems quite plausible that future potential victims of colored revolutions will have concluded from the cases discussed in this article that a commitment to using force if necessary will be a crucial component of maintaining power. Although it did not follow an election, the May 2005 crackdown by security forces in Andijan, Uzbekistan, where hundreds if not thousands of people were killed, may forebode similar actions in the wake of post-election protests of the future.⁹⁰

Finally, the argument somewhat more optimistically suggests that even people in states with little prior tradition of restraining their government in the manner suggested by Weingast can in fact come together to do so in the right situation. Weingast was primarily interested in the role that constitutions can play in this process, but he did note as well that “a galvanizing event such as a major riot, or a pact can serve to coordinate citizens' reactions so that citizens can police the state.”⁹¹

The experience of the colored revolutions reveals that major electoral fraud can also be such a galvanizing event.

Contributions to the Literature

In addition to offering new insights into the colored revolutions, applying the collective action framework to these events also offers a modest contribution to the existing literature on protest. Simply put, electoral fraud, and especially major electoral fraud, can be a remarkably powerful device for solving the collective action problems normally associated with preventing citizens from taking action against a regime towards which they hold serious grievances and, thus, facilitating protest. The many ways in which electoral fraud can accomplish this have already been detailed earlier, but in this section I highlight a number of important contributions of this insight for the existing literature on protest.

First and foremost, such an observation can help provide an answer—in certain circumstances—to what Sidney Tarrow calls the one of the big three questions of social movements: why do people protest when they do?⁹²

This is always a difficult question for any explanation of protest that revolves around structure: if structure is ripe for protest, why does the protest not take place a month earlier or a month later? In a similar vein, Timur Kuran argues that rational choice theories do a very good job of explaining why protest should not break out, but notes that “from time to time revolution does break out, and this presents a puzzle that the standard theory of rational choice cannot solve.”⁹³ While my contribution is far from a general theory, it does offer one specific answer to this question: publicized major electoral fraud can alter both the perceived costs and benefits of participating in protest in a way that ought to make participation more likely after the fraud has been committed but before the results of the fraudulent election have been implemented. In this way, it does help to point to at least one answer to the question of why people protest when they do.

Kuran rightly notes that perhaps the biggest challenge for rationalist explanations of protest is to explain why the first protestor takes to the streets. I would argue that under the framework put forward here, the first protestor takes to the streets for the same reason that subsequent protestors take to the street: the knowledge of major electoral fraud has altered her beliefs about the likelihood of success and punishment following anti-regime actions. Although these beliefs should change even more as crowds grow, the mere knowledge of the fraud should change the calculation enough for some to allow them to begin the protest, especially if there are prior examples of protest upon which they can draw.⁹⁴

While Kuran has a more complex mechanism for how subsequent protestors choose to join a protest, it is very interesting to note that one of the reasons he puts forward for that first protestor is that he or she “has an unpleasant encounter at some government ministry.”⁹⁵ This is of course similar to what I have labeled grievances against the state. So while in Kuran's story a grievance causes the first protestor to take to the streets, in my account grievances create a reservoir of potential protestors should external circumstances change. In both cases, though, grievances provide the spur to protest. In my account, however, I also highlight a specific event—major electoral fraud—that can change external circumstances.

As I will discuss in the following section, electoral fraud is of course not the only event that can alter external circumstances. However, in focusing on an event related to elections, I augment Debra Javeline's arguments on the relationship between elections and protest.⁹⁶

Javeline's work on the relationship between blame and protest in Russia identifies three ways in which elections can facilitate protest: by providing a “purposively informative political environment,” by providing this information in a finite period of time and by providing people with a limited set of options for expressing blame.⁹⁷ I suggest a fourth way that elections and blame can lead to protest, which is by the election itself—once it has been fraudulently manipulated—becoming a source of blame against the current regime, and therefore functioning as a focal point for resolving collective action problems in the manner described previously.

Additionally, scholars of social movements have long noted that protest tends to occur in waves. Previous attempts have been made to explain the temporal dimensions of protest from a collective action framework. Tarrow writes that “once a cycle [of protest] begins, the costs of collective action are lower for other actors; new movements that arise in such contexts do not need to depend as much on internal resources as on the generalized opportunities of cycles of protest.”⁹⁸

Chong argues that “the availability of successful models of protest raises expectations about the chances of success,” which in turn makes people more likely to participate in future protest.⁹⁹ I add to this line of argument by proposing that protests to overturn electoral fraud can be another of these “successful models of protest” that can spur additional protest in the future. Furthermore, I suggest that it is not merely the likelihood of success from past instances that is important, but rather beliefs about the expected costs of participation that can be learned as well. When Ukrainians see that international attention to Georgian elections helped prevent an armed response to protest by the Georgian state, they can update their beliefs about the likelihood of similar restraint in a similar context in Ukraine. Moreover, unlike Chong's work, which is focused explicitly on experiences in one country (the United States), I have identified a context in which this phenomenon may be occurring across national borders.

Finally, and perhaps a bit more broadly, in an argument I have referred to previously in this article, Weingast has argued that in order to maintain democracy citizens need to be able to restrain the state. This can happen “only if they react in concert to violations of fundamental limits by withdrawing their support from the sovereign or political officials”¹⁰⁰

and “have the ability to act in concert against political leaders who transgress constitutional rules.”¹⁰¹ Clearly there must be variety in the types of transgressions political leaders carry out. Some will facilitate and some will hinder the ability of citizens to react in the way Weingast suggests they must; the occurrence of the colored revolutions suggests that attempting to manipulate election results belongs in the former category.

Directions for Future Research

A final benefit of applying the collective action framework to the relationship between electoral fraud and anti-regime protest is that it yields a number of interesting avenues for future research, both in terms of specific hypotheses and some more general areas for research. At the micro-level, these hypotheses are quite obvious: we ought to expect citizens with higher beliefs about the probability of success and lower expectations of costs of participating to be more likely to participate in protests following electoral fraud. Additionally, we ought to expect—all else being equal—that citizens with more grievances against the regime would be more likely to participate in protests following electoral fraud.

But perhaps more interesting is the set of macro-level hypotheses that can be generated by the emphasis on the cost-benefit analysis inherent in any individual's decision to join a protest following electoral fraud. These hypotheses can be formulated based on beliefs about the set of circumstances that ought to make individuals more likely to participate in protests, and thus ought to make protests more likely to occur. First, we can hypothesize that protest should be more likely following major electoral fraud than other forms of electoral fraud; the logic here is that major electoral fraud holds out the promise of regime change if it is reversed, which would be a more valuable “success” in the eyes of anti-regime protesters. Similarly, we can hypothesize that protest will be more likely as knowledge of major electoral fraud is both more certain and more widespread. Put another way, it is not the fraud itself that should lead to protest, but rather certainty in the belief that the fraud has occurred.

The collective action framework also leads us to hypothesize that prior examples of successful protest following electoral fraud will increase the future likelihood of protest following electoral fraud. Again, the emphasis here is on the individual's calculation of the likelihood of success from protests following the current instance of electoral fraud; anything that increases her belief that success is likely should increase her chances of participating in the protests. Prior instances of protest following electoral fraud that are unsuccessful in effecting change should correspondingly decrease the likelihood of protest following the current instance of electoral fraud. To put this in context, success in Georgia and Ukraine may have encouraged the protestors in Belarus that took to the streets following elections in March 2006, but the failure of the Belarusian protests to overturn these results may make protest in Russia in 2007 less likely. Thus the collective-action approach can explain both why we might observe a sequence of cross-national protests following electoral fraud along the lines of the colored revolutions, but also, importantly, why they might stop as well. This leads directly to a fourth hypothesis, which is that prior examples of citizens bearing costs from participating in protests—and here the most obvious example would be an armed response by the regime—will make protest less likely in the future. Of course, the tricky question for all three of these propositions is what exactly constitutes a prior “example.” Is it only within one's own country? If so, the earlier failed Serbian protests would seem to falsify the hypothesis. Or is it events in other countries within a certain time period? Clearly, many Ukrainians appeared motivated by the experiences of Serbians and Georgians to believe that victory was possible. Perhaps most intriguing is the question of whether there is some sort of regional boundary to examples. Can Egyptians be affected by what happened in Ukraine? Or is this mass-level learning limited only to experiences in countries of a similar type? Was there perhaps some sort of post-communist boundary to this learning effect?

We can also hypothesize that protest following electoral fraud should be more likely to occur in countries where citizens have more longstanding grievances against the state.¹⁰²

Operationalizing such a test would of course be difficult, although I suggested earlier that perhaps levels of corruption could function as a reasonable proxy. Understanding the functional form of this relationship would seem to be particularly important. Is it essentially a tipping model, where electoral fraud leads to protest in certain types of societies, but not in others? Or is it more linear, where incrementally more grievances lead to an incrementally greater likelihood of protest following electoral fraud?

In addition to generating specific hypotheses, applying the collective action framework to the colored revolutions also yields a number of interesting general questions for future research. First, I have argued here that electoral fraud is a particularly useful tool for solving collective action problems because of a number of special characteristics of elections—the fact that elections occur at repeated intervals, encourage opposition cooperation, invite international attention, etc.—but are there other events that can also fill this role? Certainly national holidays would seem to have the advantage of being regularly scheduled, and national disasters or political assassinations tend to invite international attention. How might such events compare to electoral fraud as facilitators of anti-regime protest?

Another interesting question is whether protest spurred by electoral fraud can be a long-term solution to Weingast's need for the citizenry to restrain the state, or whether there is something inherent in these types of protests that limit them simply to one-shot games. In the framework I have laid out here, the perceived chances of success in throwing the bums out motivates some of the individual decisions to act. It is interesting to consider what might happen a second time around. Would citizens emboldened by their previous success be that much more likely to act in the future? Or would the fact that the previous revolution had so quickly given way to yet another government willing to commit major electoral fraud decrease belief in the likelihood of real success, and therefore decrease the likelihood of a new set of protests?

Finally, history clearly offers multiple examples of major electoral frauds that have not led to protests, even in the post-cold war world.¹⁰³

As my argument is not meant to be deterministic (e.g., electoral fraud always leads to protest), this raises the interesting question of whether there are systematic factors that can be identified that increase the likelihood of protests following electoral fraud. Many such potential factors have already been discussed in the article in the context of individual cost-benefit analyses—previous examples of successful protests, belief that the regime will not resort to force to counter protests, a unified opposition—but there may be other interesting factors to explore both within and beyond this framework. More generally, why are some regimes able to pull off fraudulent elections with much less response from the citizenry than others?

Regardless of what the answers to these questions turn out to be, they share the common feature of arising from thinking about the colored revolutions not primarily in terms of the actions of elites, but instead from considering the motivations of the individuals that came out on to the streets to participate in them. Doing so can add both to our understanding of the causes and implications of these revolutions, but also to our understanding of the phenomenon of protest more generally. In this way, the colored revolutions serve as another example of the importance of, to quote Pam Oliver's aptly titled article, “Bringing the Crowds Back In.”¹⁰⁴