The Coup Trap, Mali Edition

Mali had another coup d’etat yesterday, just 10 months after one that brought down the country’s elected civilian government and stunned a lot of observers in the process.

I’m not going to try to analyze the specifics of Mali’s latest coup or its repercussions, a task best left to area experts who actually know about those things, like Gregory Mann. Instead, I want to talk in broadly comparative terms about how and why this second coup is less surprising than the first.

The fact is, coups are often recursive. In a classic of the genre called “Poverty, the Coup Trap, and the Seizure of Executive Power” (alas, behind the dreaded JSTOR paywall), political scientists John Londregan and Keith Poole note that “the aftereffects of a coup include a heritage of political instability in the form of an increased likelihood of further coups.” This is the coup trap their title references; once you’ve had the first event, the risk of the next (and then the next, and then the next…) goes up.

This recursive pattern shows up loud and clear in statistical models I’ve used with some success to assess coup risk in countries worldwide. As noted in a previous post, my assessments are the average of estimated probabilities from two models, one of successful coups and the other of any coup attempts. In the former, the occurrence of a coup attempt at any time in the past five years roughly doubles the risk of a successful coup over the next year. In the latter, I use the natural log of the count of coup attempts in the previous five years minus one, but the strength of the association—and thus the general pattern—is essentially the same.

What’s more, it’s often not just the occurrence of the coup itself that affects the models’ estimates of the risk of a recurrence, but the ripple effects of that event on other risk factors. One of the models I use includes a nonlinear form of a scalar measure of democracy, the Polity scale. According to this model, countries in the mid-range between stark dictatorship and full democracy are at highest risk, and that’s often where countries wind up immediately after a successful coup. The other model uses Polity’s measure of the durability of a country’s political institutions. In this version, it’s the countries with recent institutional ruptures that are at higher risk, but again, the basic effect is the same.

We can see how this plays out in a case like Mali. If we adjust the inputs to our forecasting algorithm to catch up with the March coup and its ripples, Mali’s risk score jumps from about 4 percent, which was already high enough to land it in the Top 20 for 2012, all the way to 10 percent. Ten percent probably sounds small if you’re used to consuming probabilistic forecasts about routine things like the weather, but for rare events like coups, that’s a huge jump. Instead of being 11th on the global list, Mali would be 3rd, behind only Guinea-Bissau and Niger.

These models weren’t designed to test specific hypotheses about why coups recur, so I won’t make any bold assertions on the causal front. For what it’s worth, though, I will say that the patterns highlighted by these and many similar models strengthen my own belief that politics are, in no small part, a matter of confidence.

Whether they succeed or fail, coup attempts often disrupt established relationships among political elites. These disruptions increase elites’ uncertainty about the intentions of their potential rivals, and the proximity of the last attempt may lead them to overestimate the likelihood of the next one. In a kind of self-fulfilling prophecy, this intensification of uncertainty strengthens incentives to try to seize power. In game theoretic terms, the occurrence of a coup attempt pushes political elites out of a world resembling Stag Hunt, where coordination of action is their chief concern, to one more like Prisoner’s Dilemma, where uncertainty about the the other player’s intentions overwhelms incentives to cooperate. Once the original trust network has fallen apart, no one wants to be the sucker who keeps cooperating while the other guys are all planning to fink.

A Mexican Standoff in Georgia

Bidzina Ivanishvili’s Georgian Dream coalition has only held power for a few weeks since its surprise win in last month’s parliamentary elections, but some of its first steps already have me worried about the risk of a reversion to authoritarian rule there. A week ago, Georgian authorities brought criminal charges against Bacho Akhalaia, a former defense minister and close ally of President Mikheil Saakashvili’s, along with two sitting officials in the Defense Ministry. This week, we hear that authorities have arrested five senior officials in the Interior Ministry on charges linked to the October election.

I’m worried about these arrests because I’m watching them through the lens of a theory that sees strategic uncertainty as one the leading killers of new democracies. In my mental model, democracies can revert to authoritarian rule three ways: 1) an executive coup, whereby the ruling party quashes its rivals or otherwise rigs the political system in its own favor; 2) a military coup, whereby state security forces install themselves in government; or 3) a rebellion, whereby one or more opposition parties successfully seizes power by means other than a fair election. Rebellions occur rarely and almost never succeed, but executive and military coups are historically common, and most attempts at democracy worldwide have failed within a decade or two of their start by one or the other of these means (see here, here, and here for some previous posts on these broader points).

The spoils of state power often play a strong role in enticing incumbent officials and military officers to attempt coups, but they aren’t the only force at work. Political factions may also be lured into undemocratic behavior by uncertainty about their rivals’ intentions and fear of the steep costs of guessing wrong.

A game-theoretic model demonstrates this point in a formal way, but you can get the same idea by thinking about a Mexican standoff (and if you don’t know what that is, watch the embedded clip below from The Good, the Bad, and the Ugly for the classic cinematic example). In democracies, the three gunfighters are the ruling party, the opposition, and the military. In some cases, each of these factions may be itching to knock off its rivals as a way to win sole control of some treasure at hand. In other cases, though, some or all of the dueling pistoleros might genuinely prefer to cooperate with the others. Maybe there’s an even bigger treasure up the road that they can only capture if they work together, or maybe they’re just tired of shooting. Whatever the reason, the problem is that this desire for cooperation can’t always overcome the fundamental problem of mutual distrust. Because the stakes are so high, every little turn of the eyes or twitch of the finger is liable to get misinterpreted as a sign of bad intentions, and no one wants to be the sucker who waits a little too long to shoot in hopes that things will work out okay on their own.

Turning back to Tbilisi, the early arrests of Saakashvili loyalists in the state security apparatus are raising concerns that Georgian Dream means to punish its former overlords in ways that push the boundaries of democratic practice. An executive coup is the typical trajectory for new democracies in the post-Cold War period, especially in countries, like Georgia, where politics is sharply polarized. Even if they aren’t aware of those general facts, many observers seem quite sensitive to this risk. “The gloves are off in Tbilisi as the new ruling power takes aim at President Mikheil Saakashvili’s allies,” Molly Corso wrote of the arrests in Business News Europe. According to the New York Times, “some lawmakers feared [the arrests] presaged a wave of reprisals against members of President Mikheil Saakashvili’s defeated government.” And on today’s TBLPOD podcast from Tbilisi, Camrin Christensen noted that, “People here are now probably thinking, ‘Oh, no, am I also on this list?’ and thinking about taking family vacations.”

It’s not the handful of arrests themselves that are so worrisome, of course. It’s what they imply about Georgian Dream’s underlying intentions. In the Mexican standoff metaphor, these arrests are a menacing turn of the eyes and hips in the direction of Saakashvili’s United National Movement (UNM) and its sympathizers in the police and defense establishment.

Maybe those fears are overblown. Maybe Georgian Dream is being straight with us when it says it’s just pursuing legitimate investigations into abuses of power during President Saakashvili’s tenure. Even if that’s the case, though, the resulting uncertainty about its true intentions and growing fear of a self-coup will increase the risk of a military coup or a rebellion by the UNM as these factions grow more concerned about their fading prospects under Georgian Dream. The stronger their belief that Ivanishvili has it in for them, the stronger their incentive to respond fast, before the bullets arrive and score some serious damage.

My judgment might be clouded my affection for the place—I was a Soviet area-studies major as an undergrad; one of my oldest and closest friends is an American expat now living in Tbilisi, and I loved what I saw when I traveled there for his wedding a few years ago—but I’m optimistic that this budding standoff will wind down without any grave injuries or fatal mistakes. Georgia’s rival factions might not like each other, but they hate and fear Russia even more. (The cartoon to the right sums up many Georgians’ views of their 2008 war pretty nicely.) Because of the omnipresent threat from its neighbor to the north, Georgia badly wants into Europe, and entry into NATO is seen as the first door through which it must pass. NATO has sound geostrategic reasons not to admit Georgia while the threat of renewed war with Russia lingers, but Saakashvili’s authoritarian tendencies didn’t help its case, either. The risk of further alienating Europe with a blatant demolition of democracy will probably be powerful deterrent to would-be rebels or coup plotters.

European officials are keenly aware of this desire and already making good use of their leverage. A few days after the former defense minister’s arrest, RFERL reported that NATO Secretary-General Anders Fogh Rasmussen said he was “extremely concerned” about Georgia’s post-election politics. “It’s for the legal system, the judicial system in Georgia, to sort out these cases,” he told a meeting of NATO’s Parliamentary Assembly in Prague, “but of course it’s important that such trials are not undermined by political interference and we will of course follow that development very, very closely.” As will we.

Why Is Libya’s Transitional Council Thumbing Its Nose at the ICC?

What better way to welcome a new blogger to the blogosphere than to rebut his inaugural argument?

In the first substantive post on his promising new Causal Loop blog, Georgetown University political-science student Anton Strezhnev applies game theory to present-day Libya to try to explain why that country’s National Transitional Council (NTC) is not cooperating fully with the International Criminal Court (ICC). Strezhnev focuses on the NTC’s welcoming of Sudanese president Omar al-Bashir, for whom the ICC has issued an arrest warrant, but the Council is arguably failing to cooperate with the ICC by dragging its feet on a trial for Saif Ghaddafi as well. Strezhnev starts with a nice recapitulation of one body of relevant theory:

Political science theory suggests that a state with a recent history of civil war that has a strong commitment to rule-of-law will be less likely to bind itself to the ICC. The explanation lies in the intersection between the role of [international organizations] and Libyan domestic political imperatives. Leaders are very careful about surrendering state sovereignty to international organizations and only do so when there is a clear political benefit. Credible commitment or “hand-tying” theories of international institutions emphasize the advantages of being restrained by an external actor. Governments may want to convince domestic audiences that they will refrain from a particular behavior but lack the ability to make that commitment believable. International agreements provide a means of signalling credibility since enforcement is no longer in the hands of the (untrustworthy) government.

He then applies this theory to the Libyan case as follows:

If the commitment explanation for state behavior is accurate, then the NTC’s tenuous relationship with the International Criminal Court may suggest a belief by Libyan transitional leaders that their domestic reforms are a sufficient signal that they will not return to Gaddafi-style repression. Given the NTC’s professed goal of establishing democratic and accountable institutions, one would expect Libya to be less likely to turn to the ICC as a post-civil war commitment mechanism, given that the sovereignty costs are still high, but the signalling benefits are not uniquely advantageous. However, the task of disarming militias and integrating fighters remains daunting and if not done properly, could increase the risk of renewed violence.  Indeed, if the NTC begins to lack credibility in the eyes of some factions, then it may start looking outward to international organizations as a means of reassurance.

Strezhnev concisely summarizes an important theory, but I think the Libyan case actually illustrate the limits of this theory rather than its validity. From the stories I’ve read, I get the sense that the NTC’s political commitments are widely regarded as dubious at best. Libya’s transitional government was born with a credibility problem, and that problem hasn’t gotten much better since Ghaddafi’s forces were finally defeated. If anything, the NTC’s credibility seems to have eroded in recent weeks, because the council is now expected to actually govern, and so far, it has proved unable to do so. As the International Crisis Group summarized in a December 2011 report on the Libya transition,

The NTC’s desire to bring the militias under central control is wholly understandable; to build a stable Libya, it also is necessary. But obstacles are great. By now, [the militias] have developed vested interests they will be loath to relinquish. They also have become increasingly entrenched…They also have advantages that the NTC and the National Army lack, notably superior local knowledge and connections, relatively strong leaderships and revolutionary legitimacy. In contrast, the NTC has had to struggle with internal divisions, a credibility deficit and questions surrounding its effectiveness. But the heart of the matter is political. The security landscape’s fragmentation – and militias’ unwillingness to give up arms – reflects distrust and uncertainty regarding who has the legitimacy to lead during the transition.

This does not sound to me like a body exuding the kind of credibility Strezhnev inferred from its decision to host Bashir.

To understand why the transitional council is still keeping the ICC at arm’s length, I think we need to think about the multiple games it’s trying to play at one time. The locus classicus on the subject is George Tsebelis’ Nested Games, in which the “games” in question are the kinds of strategic interactions Strezhnev describes. As Tsebelis puts it (p. 7),

If, with adequate information, an actor’s choices appear to be suboptimal, it is because the observer’s perspective is incomplete. The observer focuses attention on only one game, but the actor is involved in a whole network of games–what I call nested games. What appears suboptimal from the perspective of only one game is in fact optimal when the whole network of games is considered.

It would be a stretch to describe anything the NTC has done so far as “optimal,” but I think the metaphor of nested games is very useful here. The NTC wants and needs some things from the states that have endorsed the ICC, but it also wants and needs things from the militias that emerged during the civil war, and from neighbors like Sudan. Moreover, the members of the NTC are themselves presumably engaged in lots of internal haggling. In other words, the transitional government is simultaneously engaged in bargaining at four levels–internal, domestic, regional, and global–and actions that look like the prudent play on one of those levels will often look wrong-headed on others.

The decision to welcome Bashir is a great case in point. To backers of the ICC, Bashir’s visit seems like a thumb in their eye, but as Multilateralist blogger David Bosco points out in a recent post, the NTC’s decision to welcome the Sudanese president “is utterly unsurprising” in regional context.

As a weak player in a rough neighborhood, Libya’s new authorities need the support of powerful states as the move forward. They don’t need the ICC anymore (if they ever did); the court is now a nettlesome complication. So Libya will happily endure blistering press releases from the human rights community in order to cement relations with a rich and powerful neighbor. The only thing that would change that equation is the insistence of other powerful states that there would be serious consequences for welcoming the Sudanese president. Bashir’s arrival suggests that message either was not sent–or was not received.

In the case of Saif Ghaddafi, the domestic game seems to be the crucial one. When rebels from Zintan captured the son of the deposed dictator was captured in November, they initially refused to hand him over to the NTC, and reports (like this one) suggested that the high-profile prisoner was being used “as a bargaining chip in the contest between rival groups for power in the new Libya.” Seen from this perspective, the NTC’s recalcitrance looks less like a marker of the council’s domestic credibility than a function of its inherent weakness.

Whatever the exact sources of the NTC’s decisions may be, it’s clear that the ICC’s arrest warrant for President Bashir and its demand to host the trial of Saif Ghaddafi have become points of contention in all of the political games the NTC is attempting to play, and incentives at the various levels seems to be pulling the council in different directions. Even if it could make up its own mind, the NTC doesn’t yet have the power to pick a winner, and arresting Bashir or trying to force a handover of Saif before it’s capable of doing so could shred its already-gossamer authority.

Democratic Consolidation, Meet Prospect Theory

Psychologist Daniel Kahneman was awarded the Nobel Prize in Economic Sciences in 2002 for some of the work he did with Amos Tversky over many years on the study of human decision-making. Dubbed prospect theory, the part of Kahneman and Tversky’s research that led to a Nobel Prize identifies certain cognitive biases that cause actual humans to deviate in their decision-making from the mathematical rationality of expected-utility theory. In his fantastic new book, Kahneman (p. 282) discusses the three “cognitive features” underpinning prospect theory.

  • We evaluate potential gains and losses in relation to a neutral reference point. Under expected-utility theory, $10 is always $10, and $20 is always twice as much as $10. It turns out that’s not really how we think about gains and losses. Prospect theory shows us that, to a real-live human decision-maker, the value of $10 and the magnitude of the difference between $10 and $20 depend on their relation to some prior reference point that shapes the decision-maker’s expectations. As Kahneman writes, “For financial outcomes, the usual reference point is the status quo, but it can also be the outcome you expect, or perhaps the outcome to which you feel entitled, for example, the raise or bonus your colleagues receive.”
  • Our sensitivity to changes in wealth diminishes as values increase. “Turning on a weak light has a large effect in a dark room. The same increment of light may be undetectable in a brightly illuminated room. Similarly, the subjective difference between $900 and $1,000 is much smaller than the difference between $100 and $200.”
  • We dislike losing more than we like winning. Psychologists call this principle loss aversion. As Kahneman reports, “When directly compared or weighted against each other, losses loom larger than gains. This asymmetry between the power of positive and negative expectations or experiences has an evolutionary history. Organisms that treat threats as more urgent than opportunities have a better chance to survive and reproduce.”

I first encountered prospect theory in graduate school, when Kahneman and Tversky’s groundbreaking article in Econometrica (pdf) was one of many required readings in an excellent course on organization theory taught by Jonathan Bendor. Most of my work since then has involved statistical analysis of cross-national data, so I haven’t had much cause to think about how these cognitive biases shape political behavior.

A couple of year ago, though, I wrote a book on democratic consolidation that uses game theory to help understand when and why attempts at democratic government usually fail. The formal model at the heart of that book construes those failures as the result of choices made by political organizations–rival parties and the military–in the face of uncertainty about future gains and losses and the likely actions of their competitors. Re-reading the outlines of prospect theory now, I can see a couple of ways that the cognitive biases it identifies might help explain some of the empirical patterns I observed in failures of democracy.

First and most interesting to me, I can see how loss aversion would strengthen the temptation for incumbent office-holders to diminish civil liberties and manipulate elections in order to preserve their power. Prospect theory teaches us that fears of large losses drive people to do all kinds of things to protect themselves against those losses, even when the costs of that protection seem to outweigh the expected gains. One of the two ways democracies usually fail is by executive coup, where the party that won the last election fixes the game to ensure it doesn’t lose the next one. (The other, of course, is military coup.)

It’s not hard to imagine situations where the spoils of power would tempt incumbent officials to try to lock themselves into office, but it’s also not hard to think of significant risks and costs associated with those efforts, especially if they fail: never-ending payoffs to cronies, loss of property, prison time, and so on. Prospect theory helps me see how the experience of holding power might reset actors’ expectations and then lead them to weigh the expected costs of losing power more heavily than the actual costs of trying to sustain it. In the world described by prospect theory, the spoils of power wouldn’t have to be all that grand, and the threat of losing the next election or being ousted in a military coup or rebellion all that large, to tempt incumbents to try to lock their status down. As Kahneman writes (p. 316),

When you pay attention to a threat, you worry–and the decision weights reflect how much you worry…The worry is not proportional to the probability of the threat. Reducing or mitigating the risk is not adequate; to eliminate the worry the probability must be brought down to zero.

In the context of democratic government, bringing the probability of a loss of power down to zero means breaking the democracy.

Second, I can also see how what Kahneman and Tversky call the possibility effect might increase the temptation to attempt a military coup. Military coups fail much more often than they succeed, so the puzzle here is why officers keep trying in spite of the long odds. Apparently, loss aversion doesn’t mean that people never chase long shots. Instead, psychologists have found that people tend to prefer big payoffs with long odds over smaller payoffs with much better odds in paired choices when the outcomes of the safe play is framed as a loss.

This cognitive quirk leads us to make decisions that rational-agent theory says we shouldn’t make, like chasing unlikely profits from business ventures into which we’ve already sunk a lot of money. Military leaders facing slashed budgets or loss of status and prerogatives confront a similar choice: stomach the losses associated with continuation of the status quo, or gamble on a coup bid that probably won’t succeed but might stem those losses if it did. Through the lens of prospect theory, we can see more clearly why they might pick Door Number 2, even if the expected payoff from that pathway is worse.

The possibility effect is connected to a pervasive human bias toward optimism. As Kahneman puts it (p. 255), “Most of us view the world as more benign than it really is, our own attributes as more favorable than they truly are, and the goals we adopt as more achievable than they are likely to be.” In addition to being overly optimistic about our own skills and prospects, we tend to reward other people who exude self-confidence, even when that confidence is unwarranted by past performance. As a result, “Optimistic individuals play a disproportionate role in shaping our lives. Their decisions make a difference; they are the inventors, the entrepreneurs, the political and military leaders–not average people.”

The reference to military leaders caught my eye when thinking about the risk of coups. Military leaders are not a representative slice of the general population. As Kahneman notes, they are more likely to be risk-seekers who overestimate their own skill and downplay the chances of failure. These traits probably serve them well in combat and make them appealing leaders, but they could also lower the cognitive barriers to making a grab for power in shaky democracies.

Those are two ways that prospect theory might deepen our understanding of the forces behind the two most common forms of democratic breakdown. Structural conditions might determine the size of the gains and losses political decision-makers can expect to realize under different future scenarios, but cognitive biases will affect how they weigh those gains and losses, and how they evaluate their chances of insuring against or capturing them. Prospect theory doesn’t “prove” anything about the causes of democratic breakdown any more than my own little model does, but it’s comforting to see that an informal application of the former to the latter seems to amplify rather than undercut the lessons we might draw about why democracy fails so often.

“Magic Democracy Words” Don’t Tie Their Speaker’s Hands

In an August 30 piece for BBC News, Shashank Joshi, a graduate student at Harvard University and associate fellow at a major U.K. think tank, argued that strong statements from American officials about Syrian president Assad’s loss of legitimacy would help advance the Syrian revolution by committing the U.S. to stronger courses of action. Joshi writes (emphasis added):

The Syrian revolution of 2011 could also have been one more of those many abortive uprisings whose blood flecks the history of the modern Middle East, yet could not change its course. Things are no longer so clear. The outside world is slowly getting its act together. The US finally issued its “magic democracy words” (a term coined by US Middle East scholar Marc Lynch) and called for President Assad to go. No-one expects that the words will wound themselves, but they tie American hands and thereby force the machinery of US foreign policy to churn out fresh ways of hounding Damascus.

This isn’t the only place I’ve seen it said that sharp pronouncements from American officials about a foreign leader’s right to rule or the need for regime change “tie American hands.” This might sound nit-picky, but that phrasing’s not quite right, and it makes a difference for how effective we might expect those “magic words” to be.

The language about hand-tying comes from game theory. In multiplayer games, each player’s course of action often depends, in part, on its expectations about what other players will choose to do. This interactive aspect of the game means that one player can influence the others’ choices by committing him or herself to following or eschewing a specific course of action. For that commitment to be credible, it has to be visible (or audible) to the other players. More important here, it also has to be something its maker can’t undo, or, if he or she can undo it, something that would obviously be costlier to undo than to follow.

A classic example of hand-tying comes from the game of chicken. Imagine a contest with two cars hurtling toward each other. If the cars smash into each other, both drivers lose badly. If both cars swerve, neither driver wins, and they both look a little cowardly. The only way to win the game is to hold the course longer than the other guy. To scare your rival into swerving first, you might commit yourself to holding course by, say, visibly locking the steering wheel into a fixed position. (To see this idea in action, watch Kevin Bacon on a tractor. Technically, that’s foot-tying, but you get the point.)

Credible commitments differ from weaker forms of signaling. Signals don’t foreclose any courses of action; instead, they affect other players’ beliefs about what course of action the signal’s issuer will choose. Game theory tells us that signals should have a weaker effect on other players’ actions than credible commitments do. They don’t lop any branches off the game tree; they just modify receivers’ beliefs about which branch of the tree they are probably heading down.

“Magic democracy words” are not credible commitments; they are signals. They are audible, but they neither lock in nor foreclose any specific policy options. After saying that a ruler like Assad must go, the U.S. government might do more to make that happen, but it can also do nothing, and it can even work to support that ruler’s continuation in office. Whichever path it chooses, it can also change course at any time. Doing so might somehow diminish America’s reputation, but the costs of a diminished reputation must be balanced against all kinds of other interests, many of which will probably weigh more heavily than ephemeral concerns about consistency and likeability. International relations is replete with flip-flops, hypocrisy, and duplicity, so it’s hard to imagine many situations in which reputational concerns would compel a government to pursue a course of action that was otherwise judged to be counter to its national interest.

To my mind, magic democracy words are more like trash-talking than hand-tying. They might get players and fans a little hot under the collar, but they don’t really tell us much about the action to come. Smart players and coaches will ignore the jawboning and will look for their signals in the play that follows instead.

Follow

Get every new post delivered to your Inbox.

Join 5,814 other followers

%d bloggers like this: