Don’t Bother Predicting North Korea’s Demise

Imagine you’re out to dinner one night with a group of friends and you’re trying to decide whether to eat Indian food or Thai food. You prefer Indian food to Thai food, but everyone else in the group says they want Thai food. What do you do?

If you chose to go along with the group despite your preference, you have engaged in “preference falsification:” the act of concealing one’s private preference (i.e. the choice one would make if given the opportunity to vote by purely secret ballot) due to perceived social pressure. The operative word here is perceived; one need not be bullied or threatened directly to engage in preference falsification, but one need only hold the belief that the consequences of believing what you believe to be an unpopular preference will have negative social ramifications. In the example above, objecting to eating Thai food when everyone else is unanimous in their opinion might cause others to think you are difficult and argumentative, and if you start earning that kind of reputation, people might stop wanting to hang out with you; no one wants to be “That Guy” (or Girl).

Besides a night out on the town, we encounter preference falsification frequently in our lives. The norms of polite society dictate so. Every male instinctively knows what to say when his wife or girlfriend asks, “Do these jeans make my butt look fat?”—lest he venture to spend a lonely night on the living-room sofa. And when you’ve been invited over to your boss’s house for the evening, it’s probably better to compliment his wife’s interior design aesthetics, regardless of how tasteless the decor. However, preference falsification is not limited to the social realm; it also plays an important role in politics and economics, and it plays an especially important role in the theory of revolutions.

Just months before the 1989 Revolutions, in a startling moment of prescience, Timor Kuran, a Turkish economist at Duke, wondered why revolutions tend to take everyone by surprise, and in doing so he developed a brilliant mathematical model based on his idea of preference falsification. Mathematically-inclined readers can read the original paper here, but for everyone else, I’ll try my best to summarize the results intuitively. (See here as well.)

First, we assume that an individual is faced with two choices: support the government or support the opposition, and that he has a private preference such that his preference for the government or the opposition lies somewhere between 0 and 1 where values closer to 0 indicate greater preference for the government and values closet to 1 indicate stronger preference for the opposition. An individual has to make a preference declaration. Whether or not he reveals his private preference publicly depends on the sum of two “utilities.” The first is the utility gained from having a reputation for believing a certain preference. The second is the utility gained from integrity. An individual’s reputational utility is a function of the amount of popular support each position has: the more popular the position, the greater his utility. Returning to our Indian/Thai food example, since the rest of the group was unanimous in their support for Thai food, you would have a higher reputational utility by choosing Thai food, and lower reputational utility if you had chosen Indian food. The second component, integrity utility, is the difference between your private belief and your public belief. Think of this as how bad you feel for betraying your values. If your preference for Thai food was only slightly greater than your preference for Indian food, then the “cost” of going along with the group wouldn’t be as great if you were—say—a hardcore vegetarian.

Putting these two components together we learn two things: first, as expectations about the share of collective support increases from the government to the opposition, individuals are more likely to publicly side with the opposition (and vice versa), while at the same time, as an individual’s private preference increases from the government to the opposition, they are more likely to side with the opposition. The boundary at which all supporters of the government switch to the opposition is called the threshold function. The threshold function tells us the levels of expected support for which it is it is optimal for individuals of a certain private preference to support either the government or opposition. Note that we’re talking about expected shares of support: how much support individuals believes either position has, not how much support it actually has. The latter is denoted by the weighted average of private preferences, weighted by the relative importance of each individual. The graph below depicts this situation. The top axis shows the actual share of the opposition, the bottom axis is the expected share, and the side axis are private preferences. In this particular situation the expected share in favor of the opposition in 70%, meaning it’s optimal for individuals whose private preferences are greater than 0.5 to support the opposition, while the actual weighted share of of individuals whose private preference is greater than 0.5 is 20%. This system is in equilibrium.

Revolution, defined by Kuran, is “a sudden and massive shift in collective sentiment which induces a fundamental transformation of the social order occurs when one or both curves shifts.” The following graph shows just a such a change. Notice that the weighted average of private preferences curve has shifted. Perhaps economic conditions in the country have deteriorated drastically. There are three equilibrium: one at 0, one at 0.5, and one at 0.8. As long as the opposition’s expected share stays below 0.5, the equilibrium will move back to 0, and if it moves above 0.5, it will revise itself upwards to 0.8 (0.5 is an unstable equilibrium). In the latter case, revolution will ensue.

This is just one possible type of change. Kuran summarizes the possible types of change as follows:

First, a rise in the expected collective sentiment can precipitate a revolution if the stage for a revolution has already been set. Second, a leftward shift of the threshold function can set the stage for a revolution, and it can also precipitate one. And finally, an upward shift of the density of private preferences can help set the stage for a revolution, but it cannot precipitate one by itself.

Let’s put this back into the context of the Indian/Thai food analogy. First lets assume there is a sudden jump in expected share of support. As you’re walking to the restaurant, you overhear one group member talking on the phone about his love of Indian food. Would this lead to a “revolution”? Not necessarily. It depends on the structure of the reputational incentives given by the threshold function. If the Thai food restaurant were owned by the uncle of one of the group members and choosing not to go would hurt his feelings, then the group might still choose to eat Thai food. Next, suppose that the threshold function moves. On the way to the Thai food restaurant, you bump into the Indian chef who promises he can get you a good deal if you bring your party over. Suddenly, it becomes optimal for people with lower private preferences to support eating Indian food, thus precipitating the “revolution.” The group eats Indian food. Finally, consider a shift in the weighted average of private preferences curve. Perhaps one of the group members finds an article online touting the health benefits of eating Indian food. Like a shift in expectations, a “revolution” may be precipitated, but it might not occur, depending on reputational utility.

There are some important caveats to this argument, however. Kuran does not explain how people overcome the collective action problem associated with revolutions. From an economic point of view, revolutions are “public goods.” They are non-rival in the sense that one person’s “consumption” of the benefits of a revolution do not take away from another person’s “consumption” of those benefits, and revolutions are non-excludable because you cannot prevent me from enjoying the benefits even if I did not “pay” for them (in the form of protesting or opposing the regime). There is also another problem: the costs of personally participating in a revolution are quite high: death, injury, arrest, financial ruin, blacklisting, and so on, while the benefits of participating are extremely low: one extra person in the main square doesn’t make much difference overall. So the question is: even if I strongly support the opposition, what incentive do I have to go out on the streets? This seems especially true if I believe that opposition support is quite high and that the revolution is likely to succeed. I might as well stay home, take it easy, and let other risk their lives.

Now, let’s bring this back to North Korea. what can this theoretical framework tell us about the prospects for revolution? First, let’s revisit what B.R. Myers wrote in his New York Times Op-Ed:

Membership in the great family is also thought to provide greater access to the elders’ wisdom. This makes the time Kim Jong-un spent away in a Swiss school especially problematic, but the propaganda apparatus may be planning to ignore that part of his life altogether. (The latest reports suggest that he is now being credited with having written, at the age of 16, a treatise on his grandfather’s thought, presumably while in Pyongyang, the capital.) In any case, the notion that army generals or any other important faction would object to Kim Jong-un’s takeover was an improbable one to begin with; no North Korean could oppose the hereditary succession without being opposed to the state itself. Such an attitude is unlikely to be held by anyone in the ruling elite.

and also:

By Communist standards, the North Korean masses would have to judge both the government’s economic performance and the succession in the harshest possible terms. It is because they judge them by very different standards that Kim Jong-un was able to take over so effortlessly while promising to budge “not an inch” from his father’s line. We should therefore not make too much of the fraudulence of all that on-screen wailing. Just because North Korean TV never films anything before rehearsing all spontaneity out of it does not mean the average citizen was unmoved. By ultra-nationalist, militarist criteria, which have more to do with North Koreans’ perception of where the country stands in the world than material living conditions, the Dear Leader did a very good job indeed: the Korean Central News Agency may well be correct in saying he made the country virtually impregnable.

And finally here was Joshua Stanton’s critique of this point:

Brian Myers thinks that the succession of Kim Jong Eun does not presage instability in North Korea, because the people with the power and the guns are invested in his survival. As much as I respect Myers’s understanding of North Korea’s official pathology, I don’t see how he could possibly have enough information to know this. The premise is probably true, but it was just as true that a year ago, men who now make up the Free Syrian Army were invested in the survival of Bashar Assad. When the people rise against a system like this (or at least somewhat like this), as history suggests they usually do eventually, it’s always in defiance of most expert predictions. Repressive regimes are very good at concealing nascent discontent from foreign observers, and foreign observers who get access to repressive countries tend to be selected for how easily they can be fooled.

Stanton’s critique seems right on the money in a number of respects: he notes that repressive governments are “very good at concealing nascent discontent,” and as Kuran demonstrates, the people themselves are very good at concealing discontent as well. He’s correct to point to the example of Syria (and Libya, Egypt, etc.) as a state with seemingly rock-solid regime suddenly suddenly began to dissolve in the face of widespread discontent. All of this is consistent with Kuran’s model. But what about the other factors? What empirical evidence is there regarding private preferences? In a recent Korea Times piece, Andrei Lankov, basing his finding on interviews with defectors, concludes:

[H]aving interviewed well over a hundred North Korean refugees over the last two years, the present author would probably dare to try to answer the question.

People who succeed under the new, de facto market economy, tend to prefer the current system. They remember the regimented life under Kim Il-sung as a nightmare ― especially if they came from underprivileged groups, whose members had no chance of social advancement under the old system. Youngsters are positive about the changes, too. These people take the current situation for granted and don’t feel much in the way of nostalgia for the recent past.

However, the older generation largely hold a different view. They frequently say that they would prefer to live in the times of Kim Il-sung in the 1960s and ‘70s. They are willing to admit that those times were not without serious shortcomings. They are not fond of the political indoctrination sessions and mutual criticism meetings which in those times took a couple of hours on an average working day (separate from the normal working day, meaning a longer day at work).

Other research conducted on North Korean refugees has found a similar patterns. Though noting that attitudes toward the regime are becoming more negative, Haggard and Nolan found that 58% of North Korean in China left the country for economic reasons, while only 27% left for political reasons and 8% out of “fear.” That is, they weren’t leaving because they hated North Korean society, but because they lacked opportunity. All of this is consistent with assertions made by Myers, Cho Myeong-chul, and others.

One must also imagine that the reputational costs of dissent are quite high in North Korea as well. Anyone’s who has spent time in South Korea knows that Korean culture is a high preference-falsification culture. Much import is placed on creating a good group “feeling” (cheong/청/靑) and good “atmosphere” (bunwigi/분위기). The pressure to conform and follow the dictates of seniors is quite strong (though this hasn’t stopped South Koreans from engaging in multiple attempts at revolution and uprising, nor has it stopped South Koreans from developing a vibrant political protest culture). Similar social pressures exist in the North. There’s also the very real and very deadly costs of taking a stand against the regime that can’t be dismissed. In South Korea, rebelling against your boss might be socially harmful, but in the North it could land you and your family in jail.

There’s also the issue of the expected size of the opposition. How do North Koreans form judgments about this? In his 1989 article, Kuran notes the importance of revolutionary leaders in this regard. This was, of course, before the era of the Internet, social media, and Al-Jazeera. The recent Arab uprisings have demonstrated that revolutions no longer require Khomeinis, Vaclav Havels or Lech Walesas. Civil society and telecommunications can help fill this void. But in North Korea, where civil society is non-existent, informants and spies are lurking everywhere, and travel and communications are highly restricted, how would North Koreans know what events were transpiring on the other side of the country? To borrow the phrasing of that oft-quoted thought experiment: if a protest erupts in Chongjin and no one from Pyongyang is around, does it make a difference?

So on the face of it, it seems highly unlikely that North Koreans will be taking up arms against their government any time soon. But to conclude so would be to miss the entire point of Kuran’s body of work on revolutions. His goal is not to predict when revolutions will occur, but to demonstrate precisely why they are impossible to predict. Revolutions, as Kuran shows, are nonlinear systems. A tiny change in one variable can have a profound change on the overall system, while a major change can have limited effect. Due to preference falsification, societies like North Korea’s are imperfectly observable. This is all the more true given outsider restrictions on access to the country. So the best advice I have: don’t bother predicting North Korea’s demise. Oh, and enjoy the Thai food.