The past couple of years have produced an unusually large number of collapsed states around the world, and I think it’s worth pondering why.
As noted in a previous post, when I say “state collapse,” I mean this:
A state collapse occurs when a sovereign state fails to provide public order in at least one-half of its territory or in its capital city for at least 30 consecutive days. A sovereign state is regarded as failing to provide public order in a particular area when a) an organized challenger, usually a rebel group or regional government, effectively controls that area; b) lawlessness pervades in that area; or c) both. A state is considered sovereign when it is granted membership in the U.N. General Assembly.
The concepts used in this definition are very hard to observe, so I prefer to make probabilistic instead of categorical judgments about which states have crossed this imaginary threshold. In other words, I think state collapse is more usefully treated as a fuzzy set instead of a crisp one, so that’s what I’ll do here.
At the start of 2011, there was only state I would have confidently identified as collapsed: Somalia. Several more were plausibly collapsed or close to it—Afghanistan, Central African Republic (CAR), and Democratic Republic of Congo (DRC) come to mind—but only Somalia was plainly over the line.
By my reckoning, four states almost certainly collapsed in 2011-2012—Libya, Mali, Syria, and Yemen—and Central African Republic probably did. That’s a four- or five-fold increase in the prevalence of state collapse in just two years. In all five cases, collapse was precipitated by the territorial gains of armed challengers. So far, only three of the five states’ governments have fallen, but Assad and Bozize have both seen the reach of their authority greatly circumscribed, and my guess is that neither will survive politically through the end of 2013.
I don’t have historical data to which I can directly compare these observations, but Polity’s “interregnum” (-77) indicator offers a useful (if imperfect) proxy. The column chart below plots annual counts of Polity interregnums (interregna? interregni? what language is this, anyway?) since 1945. A quick glance at the chart indicates that both the incidence and prevalence of state collapse seen in the past two years—which aren’t shown in the plot because Polity hasn’t yet been updated to the present—are historically rare. The only comparable period in the past half-century came in the early 1990s, on the heels of the USSR’s disintegration. (For those of you wondering, the uptick in 2010 comes from Haiti and Ivory Coast. I hadn’t thought of those as collapsed states, and their addition to the tally would only make the past few years look that much more exceptional.)
Annual Counts of Polity Interregnums, 1946-2010
I still don’t understand this phenomenon well enough to say anything with assurance about why this “rumble” of state collapses is occurring right now, but I have some hunches. At the systemic level, I suspect that shifts in the relative power of big states are partly responsible for this pattern. Political authority is, in many ways, a confidence game, and growing uncertainty about major powers’ will and ability to support the status quo may be increasing the risk of state collapse in countries and regions where that support has been especially instrumental.
Second and related is the problem of contagion. The set of collapses that have occurred in the past two years are clearly interconnected. Successful revolutions in Tunisia and Egypt spurred popular uprisings in many Arab countries, including Libya, Syria, and Yemen . Libya’s disintegration fanned the rebellion that precipitated a coup and then collapse in Mali. Only CAR seems disconnected from the Arab Spring, and I wonder if the rebels there didn’t time their offensive, in part, to take advantage of the region’s current distraction with its regional neighbor to the northwest.
Surely there are many other forces at work, too, most of them local and none of them deterministic. Still, I think these two make a pretty good starting point, and they suggest that the current rumble probably isn’t over yet.
Dart-Throwing Chimp 
Peter Dörrie
/ January 4, 2013Thanks for this post!
Regarding the question if the rebellion in the CAF is related to the situation in Mali/Libya: I actually doubt it. The biggest regional player in the CAF is Chad and that country is not deeply involved in Mali. The other big player for the rebels to consider is France but the French unwillingness to intervene on Bozizé’s behalf is not so much rooted in their obsession with Mali but rather with the new political stance the Hollande administration has taken towards their involvement in african affairs.
The latter shift could be one aspect of the rebel’s calculation. As for the immediate timing, I would agree with the expert interviewt by UN Dispatch that it was likely determined by the end of the rainy season and the state of the road network in the north of the country.
dartthrowingchimp
/ January 4, 2013Thanks, Peter. Very helpful details.
Re France, the broader shift you describe is part of what I had in mind when talking about changes in the policies of major powers. Instead of intervening everywhere, these powers are now picking their battles much more carefully.
And re timing, I meant why 2012 instead of, say, 2011 or 2010, not why December. It sounds like the strategy is mostly parochial, but I still wonder if the regional tumult isn’t figuring in the calculations at all.
Felix
/ January 5, 2013I do not really get your “fuzzy” take on the concept. I think being “crisp” is one of the main advantages of the concept state collapse compared to similar concepts such as state failure or fragility, which indicate a process or a continuum of “stateness”. With the latter concepts you end up pretty opaque, like the many composite indicators of state failure (i.e. Failed State Index). Those measurement tools only capture the fragility of states in relation to other states and thus are not able to identify certain states as collapsed.
I think using a crisp approach for the concept is better when you want to do historical comparison or identify causal effects. Anyway, the interregnum variable is indeed a crisp (binary) measurement of the concept.
Best
Felix
dartthrowingchimp
/ January 5, 2013It’s not the concept that’s meant to be fuzzy here; it’s the statement about membership in the set of collapsed states. Put another way, I am looking to declare an either/or status for the sound reasons you note, but I can’t be 100% sure of that status in some cases, so I express my judgment probabilistically instead of definitively. If I could clearly observe all of the elements of the definition, I would make a crisp call—but then I (probably) wouldn’t be talking about a complex social phenomenon in hard-to-reach places. If I could, I would measure democracy similarly. With both concepts, the result would probably be a trimodal distribution. Most cases would take values very close to 1 or 0, but a small set of especially hard-to-observe or hard-to-judge cases (e.g., Georgia before and after its recent parliamentary elections) would cluster around 0.5.
I don’t use the interregnum indicator if I can help it because I don’t think it’s reliable. Some cases get the -77 tag for reasons I can’t understand (e.g., Comoros 1975, Czechoslovakia 1968, Dominican Republic 1963); some seem like they merit the tag but the timing is strange (e.g., Angola 1992, Cambodia 1975); and some cases don’t get the tag at all that seem like they should (e.g., Albania in the early 1990s or in 1997). In other words, I think Polity’s interregnum offers a poor measure of a valid concept, so I try not to lean on it.
Grant
/ January 5, 2013What about sectarian issues? In Libya, Syria, Yemen and Mali a good deal of the violence came from sectarian divides as opposed to the broader cross-group revolutions we saw in other parts of North Africa and the Middle East. Maybe sectarianism makes a nation far more likely to face civil war in times that are more revolutionary across the region*.
As for the Central African Republic, we’d need to know more about the makeup of the rebels.
Alternatively maybe the collapse was simply from weak governments being overwhelmed by revolution (note that Bahrain and Saudi Arabia were able to crush Shia protests) and the lack of strong state or civil institutions meant that it turned to civil war rather than revolution. In that case Syria would just be an odd coincidence where a strong state for once was unable to crush the opposition and the opposition was unable to push it from power. Personally sectarianism seems more likely to me but the other’s a possibility.
*Though the presence of revolutionaries in all four nations clearly disgusted with their leaders makes it a confused matter.
dartthrowingchimp
/ January 5, 2013Sectarian divides do seem to be playing an important part in the organization of the conflicts in many or all of those cases, but they don’t really help explain why collapses happened in these cases instead of other ones with similar sectarian structures, or why they happened in 2011-2012 instead of some other time. I was trying to focus on factors that provide cross-sectional and cross-temporal contrast to get at why this rumble is happening here and now.
Fran
/ January 6, 2013I’m more interested in how some countries or governments managed to rule over the whole population. I mean, take Mali for instance. How did Bamako rule successfully over the northern territories? (Although I don’t know if it really did, I know next to nothing about Malian history) And I think the same can applies to several African countries, in the sense of externally imposed borders, which usually included confronting ethnic groups within.
dartthrowingchimp
/ January 7, 2013Yes! How governments expand and deepen authority throughout their territory is the flip side of state collapse, and one of the most important things to understand about state collapse is that it often happens when central authorities attempt to make big gains on those fronts. And then there are some cases, like CAR, where this has really never happened, but this fiction mostly goes unnoticed by those of us far away because everyone’s just getting on with their lives in spite of it.
Grant
/ January 9, 2013In other words intrusion by a central government to increase power that isn’t seen as ‘legitimate’ by the locals increases the odds of rebellion and state collapse?
dartthrowingchimp
/ January 9, 2013I wouldn’t bring “legitimacy” into it, but yes, I am arguing that in many states, attempts to expand the writ of the central government in ways consistent with the norm of domestic sovereignty can be a catalyst for state collapse. Afghanistan in the late 1970s is a good example of this process. Ironic, isn’t it?
Amy Dowler
/ January 7, 2013Could it not just be random? See this article I happened to read right before yours: http://www.empiricalzeal.com/2012/12/21/what-does-randomness-look-like/
dartthrowingchimp
/ January 7, 2013It could be, yes, but the low base rate of these events–on average, state collapses only happen once every several years–suggest that the probability of a cluster this large occurring at random is very small. There’s a statistical technique called Bayesian change point analysis that would help us estimate that probability, but I haven’t applied it here because I don’t feel like I have a reliable time series of observations on state collapse.
Grant
/ January 7, 2013The timing, geography and similar political issues of the nations that did see a collapse of state authority in 2011-2012 would argue against it really being random.