Foreign Policy magazine recently published the 2012 edition of the Fund for Peace‘s Failed States Index (FSI), and the response in the corner of the international-studies blogosphere I inhabit has been harsh. Scholars have been grumbling about the Failed States Index for years, but the chorus of academic and advocacy voices attacking it seems to have grown unusually large and loud this year. In an admirable gesture of of fair play, Foreign Policy ran one of the toughest critiques of the FSI on its own web site, where Elliot Ross of the blog Africa is a Country wrote,
We at Africa is a Country think Foreign Policy and the Fund for Peace should either radically rethink the Failed States Index, which they publish in collaboration each year, or abandon it altogether. We just can’t take it seriously: It’s a failed index.
As Ross and many others argue, the core problem with the FSI is that it defines state failure very broadly, and in a way that seems to privilege certain forms of political stability over other aspects of governance and quality of life that the citizens in those states may prize more highly. In a 2008 critique of the “state failure” concept [PDF] that nicely anticipated all of the recent sturm und drang around the FSI, Chuck Call wrote that
The ‘failed states’ concept—and related terms like ‘failing’, ‘fragile’, ‘stressed’ and ‘troubled’ states—has become more of a liability than an asset. Foundations and think tanks have rushed to fund work on ‘failing’ states, resulting in a proliferation of multiple, divergent and poorly defined uses of the term. Not only does the term ‘failing state’ reflect the schoolmarm’s scorecard according to linear index defined by a univocal Weberian endstate, but it has also grown to encompass states as diverse as Colombia, East Timor, Indonesia, North Korea, Cote d’Ivoire, Haiti, Iraq, and the Sudan.
In that essay, Call advocates abandoning the now-hopelessly-freighted concept of “state failure” in favor of a narrower focus on “state collapse”—that is, situations “where no authority is recognisable either internally to a country’s inhabitants or externally to the international community.” I agree.
In fact, in 2010, while still working as research director for the U.S. Government–funded Political Instability Task Force, I led a small research project that aimed to develop a workable definition of state collapse and coding guidelines that would allow researchers to know it when they see it. The project stopped short of producing a global, historical data set, but the coding guidelines were road-tested and refined, and I think the end results have some value. In light of the FSI brouhaha, I’ve posted the results of that project on the Social Science Research Network (SSRN) in hopes that they might be useful to a broader audience.
In those materials—a concept paper and a set of coding guidelines—I argue that we can get to a more workable concept by moving away from Max Weber’s aspirational vision of modern states as legitimate and orderly bureaucracies. Instead, I think we get further when we recognize that real-world states are a specific kind of political organization associated with a particular realization of global politics. That realization—the “Westphalian order,” or just “the international system”—constitutes states and delegates certain forms of political authority to them, but national governments in the real world vary widely in their ability to exercise that authority. When internationally recognized governments cease to exist, or their actual authority is badly circumscribed, we can say that the state has collapsed. That kind of collapse can happen in two different ways: fragmentation and disintegration.
When the failure to rule involves the national government’s territorial reach, we might call it collapse by fragmentation. The ideal of domestic sovereignty presumes final authority within a specific territory and international recognition of that authority, so situations in which large swaths of a state’s territory are effectively governed by organized political challengers whose authority is not internationally recognized represent a form of collapse. In practical terms, these situations usually arise in one of two ways: either 1) a rebel group violently pushes state agents out of a particular area, or 2) a regional government unilaterally proclaims its autonomy or independence and becomes the de facto sovereign authority in that region. In either situation, the rival group directly and publicly challenges the national government’s claim to sovereignty and effectively becomes the supreme political authority in that space. State military forces may still operate in these areas, but they do so in an attempt to reassert control that has already been lost, as indicated by the primacy of the rival organization in day-to-day governance…
State collapse also occurs when the national government fails to enforce its authority in the absence of a rival claimant to sovereignty. This type of failure might be called state collapse by disintegration. The ideal of domestic sovereignty presumes that a central government is capable not just of making rules but also of enforcing them. Dramatic failures of a state’s enforcement capabilities are indicated by widespread lawlessness and disorder, such as rioting, looting, civil violence, and vigilantism. In the extreme, central governments will sometimes disappear completely, but this rarely occurs. More often, a national government will continue to operate, but its rules will be ignored in some portions of its putative territory.
To distinguish state collapse from other forms of political instability and disorder, we have to establish some arbitrary thresholds beyond which the failure is considered catastrophic. Saying focused on the core dimensions of domestic sovereignty—territory and order—I do this as follows:
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.
If you’re interested, you can find more specific language on how to assess challenger control and lawlessness in the coding guidelines.
Applying this definition to the world today, I see only a handful of states that are clearly collapsed and just a few more that might be. In the “clearly collapsed” category, I would put Libya, Mali, Somalia, and Yemen. In the “probably collapsed” category, I would put Afghanistan and Democratic Republic of Congo. Those judgments are based on cursory knowledge of those cases, however, and I would be interested to hear what others think about where this label does (Chad? Haiti? Ivory Coast? Sudan? South Sudan?) or does not (Afghanistan? Mali?) fit. Either way, the list is much shorter and, I believe, more coherent than the 20-country sets the Failed States Index identifies as “critical” and “in danger.”
More important, this is a topic that still greatly interests me, so I would love to have this conceptual work critiqued, put to use, or both. Fire away!
Dart-Throwing Chimp 
alex@aipr
/ July 5, 2012Good one, Jay. I posted it on our Facebook page (https://www.facebook.com/genprev). Thought you might like to know. Alex Z.
dartthrowingchimp
/ July 5, 2012Thanks, Alex. I appreciate it. If we had better data, it would be really interesting to look at the association between degrees of state collapse and the occurrence of genocide and mass atrocities.
Amy Dowler
/ July 6, 2012I wonder whether you could include Pakistan or Myanmar, both in the fragmented sense? Although I understand there’s been considerable progress in Myanmar on peace negotiations between the centre and ethnicity-based revel groups over the last 12-18 months.
Pakistan also has non-state groups controlling parts of its territory (no idea if anywhere near 50% though), but it also raises the question, what is ‘the state’ in the definitions above? It seems like its sovereignty is fragmented not territorially, but institutionally, between the political, military/intelligence and judicial classes.
dartthrowingchimp
/ July 6, 2012I think the territorial scope of the fragmentation in Pakistan and Myanmar/Burma is too narrow to hit the 50% threshold, and in neither case is the capital city under rebel control, either. There’re lots of states with pockets of space that are under rival control or simply lawless, but not on this scale. Ideally, we would be able to measure those features continuously and not have to make a binary yes/no call, but in my experience, the information required to do that is very hard if not impossible to come by.
Your point about institutional fragmentation is an interesting one. I guess I see that kind of rivalry as a normal feature of politics in most situations, but it does seem like something that, when amplified, could lead to collapse by disintegration. I suspect, though, that it more often leads to plain old coups. The more dangerous situation may be when you get fragmentation within security forces, especially along regional or communal lines.
Grant
/ July 8, 2012Rebel forces may not hold the national capital, but that rarely occurs in separatist wars*. In both Pakistan and Burma there are strong separatist forces that control a good deal of territory and I’m not entirely sure how well either nation (especially Pakistan) will survive the coming decades. As for coups, the older ones led by generals seem to occur either in politically disunited states where the military has also become disunited or in nations where the civilians don’t seem to be doing enough about a separatist threat.
*In fact, Slater argues that it was separatists (Karen I believe) threatening Rangoon that actually led to a brief period of political unity among Burmese elites.
David Mandel
/ July 6, 2012Jay, two years ago, a colleague of mine and I factor analyzed 20 “governance indices”, including the FSIs (i.e., FP’s failed and CIFP’s fragile). The analysis provided a three factor solution that accounted for 90% of the variability in the measures. Most indices had a clearly dominant factor loading. The FSIs, e.g., both loaded on Factor 1, which we called Stability. The second factor, which we called Freedom, had, e.g., high loadings from the Freedom House indices. Factor 3, which we called Humanity, had only two indices that loaded dominantly on it, namely, the UNDP’s gender and human development indices. We didn’t provide factor scores for the countries that entered into the analysis (65 countries, as I recall), but the idea we had in mind is that the factor scores could be viewed as integration of the 20 separate indices that might have greater indicative value than the separate indices for some warning or forecasting situations. If we knew which factors made better predictors of key outcome variables (such as genocide as you noted in an earlier reply), we could zoom in on the relevant factor scores. Whether these would be more informative than the constituent indices is something we haven’t looked at. If anyone is interested, the tech report can be downloaded here:
Click to access tr2011022.pdf
David
Grant
/ July 8, 2012I think the ultimate problem is that state failure seems inclined to making simple diagnosis’ of complex state/local population interaction, an area of political science that probably is at least as complex as two or three others put together. For example where should we put North Korea? There clearly is a state and it clearly has a monopoly on violence. However at the same time the state is not providing many vital services that the population needs and probably only barely represents the population.
I prefer the idea of a loose spectrum of state power or tiers of state collapse*. That nicely avoids trying to say ‘nation X is clearly the number one failed state while nation Z is a bit less failed and is number 2’. Instead nations X and Z would both be at the far end of the spectrum towards state collapse or in the lowest tier of state power without needing a (probably) arbitrary decision of which one is worse than the other.
*That also works better than, say, trying to argue that the Soviet Union was the number one farthest left leaning state in the world while Cuba was number two or something of the sort. We should leave precise numerical rankings to the hard sciences.