Bryan Caplan  

Two Skeptical Questions About State Capacity

PRINT
You Should Compromise With Me... Non-voting is a right too...
Many economists I know, especially the fans of Douglas North, appeal to the concept of "state capacity" to explain economic development, human rights, health, and other social outcomes.  In their stories, with few exceptions, high state capacity leads to good things, low state capacity leads to bad things.

Lurking in the background is what I call the the naive Statist Theory of Everything.  Why does society X have good thing Y?  Because society X's awesome government provided Y.  Why doesn't society Z have good thing Y?  Because society Z's crummy government failed to provide Y. 

As far as I can tell, the state capacity story is supposed to differ from the Statist Theory of Everything.  But only obvious difference is packaging: advocates of the state capacity theory tend to decorate their story with mild libertarian sympathies.  More substantial differences are hard to detect.  Yes, you could say, "I'm not saying that society X has good thing Y because its awesome government provided Y.  I'm saying that society X has good thing Y because its government has the capacity to provide good things of all sorts."  But this answer seems at once vague and evasive.

Of course, the state capacity story's tension with the libertarian worldview isn't much of an argument against it.  Yet faced with the state capacity story, you would at least expect social scientists - regardless of their political leanings - to raise their two standard skeptical challenges to every explanatory concept.  Namely:

1. Appeals to "state capacity" sound tautologous: "Why did state X provide good thing Y?  Because it has the capacity to provide good things like Y."  So if you're going to use state capacity to explain the world, you first need to convince me it's more than mere post hoc tautology.

2. Assuming state capacity isn't mere tautology, it's presumably going to correlate strongly with a bunch of other good things - current per-capita GDP, a long history of per-capita GDP, social trust, public opinion, cultural closeness to other successful societies, etc.  So if you're going to use state capacity to explain the world, you also need to convince me that state capacity, rather than any of its blatant correlates, has a big independent effect.

I admit I haven't read widely in the state capacity literature.  Most of what I know comes from listening to lunch conversations from aficionados of the approach.  Perhaps my elementary challenges have long since been acknowledged and addressed.  If so, please share in the comments.  I'm all ears - but please tell me you've got something better than the settler mortality stuff.


Comments and Sharing






COMMENTS (11 to date)
David Condon writes:

Was there an original paper explaining the theory somewhere? What's the formal definition of state capacity? What percentage of researchers working on the issue could state this definition by heart?

Enial Cattesi writes:

This idea of state capacity is on every page of "Why Nations Fail", by Acemoglu and a guy from Harvard. For some weird reason, it was considered an excellent book by some of the contributors here, although it has a ton of problems, lies, omissions, stupid correlations etc.

Any way, it is a very good start on the concept of "state capacity."

Daniel Kuehn writes:

I'm not deeply read in this stuff, but in the little I am familiar with I thought stability, rule of law, property rights, and especially security in certain areas where the important things that states provided. In the context of those conditions, non-state actors of all sorts can flourish and innovate without state provision per se. That seems quite different from just a change in packaging. It suggests that the state has a limited but important role.

Daniel Kuehn writes:

Anarchists can still quibble about whether the above is necessary, of course. But that's different from saying it's a repackaged theory of everything.

Nathan W writes:

I'm not convinced that the relationship is well explained. A snapshot (the present situation of state capacity and other indicators) is never good at telling a story (how they all got to be that way).

But since the case is so easily made in areas like health, education and transport infrastructure, it would be surprising to find that there were not other areas where there is social and/or fiscal and economic benefit linked to the state's role in areas of activity where the quantity or quality of services provided by the market is "too low".

Chicken and egg questions abound, however. For example, has it an initial growing of the pie which enabled the state to develop more capacity or the some state managing to capture enough that it was then able to develop/implement areas of activity which increased the net social benefit?

I think there are specific areas of activity which may be questioned, but I'm fairly convinced that some pretty huge sectors of the economy are best managed by the state (health, education, infrastructure, justice/security), and that presupposes a capacity to tax which should easily run to a quarter to a third of GDP.

The USA and Sweden didn't become rich because of their environmental agencies, but previous interplays between economic development and state capacity made it likely that in both case, such an agency could fill in gaps in terms of information and liability, for example. The point there is that the difference between 35% and 35.1% of GDP by the public sector probably isn't a big deal because you might be working in areas of activity where the average gains of the activity are economically marginal assuming some reasonable distribution of payouts across various projects. Compare this to the situation when comparing 10% of GDP to 10.1% of GDP. The additional 0.1% of GDP in this situation could find any number of activities which would increase total social benefit with this 0.1% of GDP, whether invested in health, education, infrastructure, or informational, bureaucratic and decision making capacity.

Consider it this way: when 10,000 hungry children are taken off the streets and put in school, almost certainly there is a total social benefit in this spending, but when 10,000 more people get a PhD, beyond a certain point the economy may have done better if they had instead developed their skills while working for five years in the private sector instead of being funded to research. The parameters can be juggled to create almost any combination of those situations and will differ by actual context.

In other countries, however, 10,000 PhDs could turn around the course of history. Consider if only half of the Rwandans who left in the 1990s and who now have above-average income in their new countries were to dedicate an average of one-quarter of savings to fund investment opportunities in their home country.

Alex writes:


I'm read a little on this sometimes 'state capacity' is called 'state infrastructural power' (if you're doing a google scholar search), which is the ability of governments to enforce laws.

One theory of why strong states emerge in some countries but not others has to do with the role of war and geopolitical competition. The theory goes as follows: when countries are threatened by war, their leaders are pushed find ways to raise money and pay for armies and national security infrastructure. The theory says that the reason why “strong” states sprung up in Europe, is that there was so much war in European history – and so the only political units that survived were the ones where the leaders found ways to collect lots of taxes and build armies.

Applied to the post-colonial setting – the theory says that there are more weak states in Africa, Latin America, and some parts of Asia because these parts of the world did not experience the same levels of geopolitical conflict, and so not only were the pressures that created states in Europe absent, but also weak states were not weeded out by conquest. Of course, this doesn’t explain why some post-colonial countries (north America, Australia, Chile etc..) produced stronger states. A political scientist named Cameron THies has a bunch of studies that show countries with strong states faced more geopolitical competition even in the modern times and so these mechanisms, he says, may account for it.

Weak states and bad development outcomes: I think the mechanisms are things like, the increasing likelihood of conflict in weak states (which have more trouble repressing/deterring rebels, terrorists etc...); lack of property rights enforcement etc... In a way, it's right out of Hobbes, where he says that in the state of nature people are too busy protecting themselves to get anything done. Measures of state capacity that i've seen include: taxation as a percent of GDP; police per capita; roads per km (to measure state control of territory).

Anyway, just food for thought; I’ve never found this completely convincing for the reasons Bryan and other commentators mention.

pseudoerasmus writes:

But even without the Albouy criticism, I still don't see Acemogluism as particularly compelling. Basically proto-colonial settler mortality determines whether the institutions set up are inclusive or extractive, and thereafter path dependence and multiple equilibria keep the countries with extractive institutions continuously saddled with extractive institutions. The few escapes from the extractive institutions traps are explained ad hoc.

Nathan Smith writes:

Francis Fukuyama's *Political Order and Political Decay* is very much in this vein. It's not a bad book, but as a defense of the statist theory of comparative development, I'm surprised at how unpersuasive it is. For example:

He devotes a lot of attention to the failure of Italy to develop the kind of effective state institutions and Weberian bureaucracy that Germany has. Fine, but why is he such a fan of Prussian/German bureaucracy? It made Germany much more effective at destructive militarism than Italy, but contemporary German GDP is only slightly ahead of Italy's.

He devotes a lot of fawning attention to civil service reform in the late 19th century USA, to the rise of the ICC to regulate railroads, and of the forestry service and the national parks. But it's far from clear that the late 19th-century USA suffered much from the lack of a Weberian civil service such as Western European states had at about the same time. On the contrary, it was probably the most flourishing and technologically dynamic society in the world at that time, and what's more, we're still living on its technological legacy.

He devotes some attention to the determinants of state effectiveness among developing countries, e.g., he explains how Nyerere's *ujamaa* (African socialist) ideology has given Tanzania more of a national identity than neighboring Kenya. But, so what? Tanzania and Kenya are about equally poor.

Again and again, I find myself bewildered by why Fukuyama is so fascinated by, and such a fan of, bureaucracy. It's pretty obvious that the story of comparative development is elsewhere.

One thing I will say for Fukuyama: he's much better than Acemoglu. His book is much less tendentious, but the result is simply that a critical reader has an easier time weighing his thesis and finding it wanting. Acemoglu, you want to interrupt and contradict. Fukuyama, you want to interrupt and ask him, "Why do you care about this? Stop taking for granted that bureaucracy is a good thing for its own sake. Let's focus on figuring out why some countries are rich and others are so poor. Can't you see that all this stuff you're writing about doesn't help much?

Ben Kennedy writes:

It seems pretty obvious that causation runs the other way. A functional state that is at least somewhat responsive to citizens can't use it's power to do stuff until the underlying economic development exists to afford whatever the state is providing. Banning child labor doesn't work well in Bangladesh because families need the income, it works quite well in the US because adults are so much more massively productive that the incremental gain from child labor is small so that banning it is of no real consequence. A central planner or a benevolent dictator is not some kind of free lunch, and the actions of states are to some degree constrained by what people are actually willing to put up with

ScottA writes:

Agree with many of the comments about the weird obsession with state capacity for its own sake. Think there is a stronger form of the argument though; it would be much more like the Weingast book on violence (the violence trap, I believe it's called).

The point of state capacity in this view is to manage and contain violent forces in society that steal/destroy/disrupt attempts to invest capital and develop longer-term economic thinking. A low capacity state can't do these things. Potential negative: a level of state capacity that can contain violence is also sufficient to prevent development, if leaders are so inclined. I've never understood why people don't stress this more. Capacity isn't everything. China had massive potential for development far before the West, but was run by a state that more or less decided that development and innovation weren't in their interest. No external threats were strong enough to force them to change their minds. Strong capacity actively prevented development.

The counter-point to the China case might be that we need two things for 'good' state capacity: enough capacity to contain violence and an open-access order (in the DA term). The former doesn't guarantee that latter; the former can actually prevent the latter in a lot of scenarios, so probably tough to get both.

Think the problem with Fukuyama and those like him is that they assume the open-access part will happen automatically. Or, less charitably, they assume that bureaucrats will always inherently have "the public interest" in mind. A term which means nothing but which many people like Fukuyama seem to like to use a lot.

The state is like opium; it causes dormition because it has a dormitive virtue.

Comments for this entry have been closed
Return to top