Efficient Institutions are Context Dependent

Peter T. Leeson

I’m grateful to Daron Acemoglu and James Robinson for their most recent post in our exchange. In it they pose several important questions, two of which in particular go to the heart of our disagreement about the reason for pirate democracy and the efficiency of institutions more generally.

Acemoglu and Robinson ask:

Leeson notes that a democratic organization, with a more equal distribution of political power and better checks on leaders, was efficient for pirates. Probably right. But the same reasoning suggests that it would have also been efficient for many other businesses and social and political organizations. But we see nondemocratic organizations persist in many of those spheres. So where are the efficient institutions?

I’ve tried to address this question in The Invisible Hook by comparing pirate ships’ organization to that of the legitimate ships from which pirates were drawn: 18^th-century merchantmen. As Acemoglu and Robinson suggest, although the former were organized democratically, the latter were not. Indeed, merchantmen were organized quite “autocratically”—i.e., via appointed officers endowed with substantial authority over sailors. If democracy was efficient for pirates, why not for merchantmen too?

The answer, I contend, is that democracy’s cost was far higher for merchantmen than for pirates.

Merchantmen were organized and outfitted by external financiers—wealthy landlubbers who had commercial expertise and capital, but weren’t sailors and thus hired seamen to sail their ships. To make sure crewmembers didn’t shirk, embezzle cargo, or steal the vessels they sailed on in owners’ absence, owners appointed officers to monitor and control them. Allowing crewmembers to democratically elect their officers instead would’ve been extremely costly in this context. Merchant sailors who could choose their officers democratically would have an incentive to elect the opposite kind of officer from what owners wanted—the kind of officer who would let sailors do whatever they pleased, destroying voyages’ profitability.

Pirate ships, in contrast, weren’t organized or financed by external landlubbers. Pirates stole their vessels jointly: they were both the owners and employees of their ships. Because of this, democracy’s major potential cost on merchantmen—the prospect of crewmembers electing lax officers and thus undermining voyages’ profitability—was absent on pirate ships. Pirates who elected lax officers, qua employees, would’ve undermined their own interest, qua owners. Their incentive was therefore to elect the kind of officer that maximized profit. Consequently, for pirates, democracy was cheap.

Efficient institutions are context dependent. The contexts—and thus costs and benefits of alternative institutional forms—that pirates and their legitimate counterparts faced were different. So it’s natural that efficient institutions in these "societies" would differ too.

Exploring differences or similarities in the specific contexts various societies operate in, which in turn generate differences or similarities in the constraints and prices their members face is, in my view, essential to understanding why we observe a variety of institutors in the world, but institutions that are the nonetheless efficient.

Moreover, it’s possible to use this approach to evaluate alternatives to the efficient institutions hypothesis, such as the de facto power hypothesis put forward by Acemoglu and Robinson. Consider, for instance, their suggestion that more equal power drove pirate democracy.

As Acemoglu and Robinson point out, “all pirates had cutlasses.” But pirates weren’t alone in this respect. For example, all sailors on 18^th-century Royal Navy vessels were armed too. Notoriously, however, Royal Navy vessels were organized autocratically, similar to merchant ships, not democratically, as were pirates. Or consider privateers, where, again, everyone "had cutlasses," but governance was considerably more democratic than on naval ships yet considerably more autocratic than on pirate ships. At least on the surface, these observations pose a challenge to the de facto power hypothesis. In contrast, the efficient institutions view can comfortably explain the institutional variation we observe across floating societies—from pirates, to explorers, to privateers.

The other important question Acemoglu and Robinson raise in their piece, which goes to the core of our broader disagreement about institutional efficiency, is the following:

How can we understand, if not as highly inefficient for economic prosperity and for the majority of the population, regimes such as North Korea?

Acemoglu and Robinson’s excellent and thought-provoking book, Why Nations Fail, sheds critical light on the reasons for why some countries languish with very low incomes (while others do not). The point I want to make is simply that incredibly low income need not imply institutional inefficiency.

In every society the “institutional opportunity set”—i.e., range of feasible institutions—is constrained. At least in the short run, a society’s “history” and “cultural” features, for example, limit the institutional choices its people can make. Since history and culture vary significantly across societies, institutional constraints do too. As I’ve argued elsewhere, these constraints suggest that we should think about institutional efficiency in terms of constrained, not unconstrained, optima.

Just as it’s not sensible to reason from the fact that I own a used Subaru instead of say, a clearly “superior” car, such as a Porsche, to the conclusion that my automobile choice is “inefficient,” I don’t believe it’s sensible to reason from the fact that in some societies we observe “inferior” institutions and thus low incomes instead of the much higher incomes we know “superior” institutional arrangements can generate to the conclusion that these societies’ institutional choices are “inefficient.” My budget constraint prevents me from owning a Ferrari: a used Subaru is my constrained automobile optimum. Likewise, institutional constraints can prevent some societies from adopting “superior” institutions that generate higher incomes: their “inferior” institutional regimes can reflect constrained institutional optima.

To say an institution is efficient is not to say its outcomes are ideal, or even that they’re “good.” It’s merely to say they’re the best that is currently possible given the constraints confronted in a particular case. Efficient institutions, in other words, are context dependent.

Comments

[Mark Lutter]

To make the claim of institutional efficiency meaningful, it must be a pattern prediction, not a point prediction. However, given the example of North and South Korea, not to mention China and Taiwan, and Singapore, such a pattern prediction does not seem to hold. While I agree a countries institutional opportunities set is constrained, there seems to be a very real possibility of multiple equilibria.

[Gwern Branwen]

Back in the 1950s, there was an incredibly dirt-poor sub-Saharan-Africa-level country called Korea; resource-poor and ravaged by war, it was noticeably poorer than its other half, which was successfully industrializing (whose propaganda directed at its poorer controlled sibling made no bones of that either), ruled by a tyrant propped up by foreign aid from an interested empire and receiving direct military reinforcements from said empire when necessary combined with official corruption favoring empire-building elements of that society, and attacking any elements of a civil society like churches or newspapers while constantly proclaiming the need to defend against foreign subversion and invasion from the other richer Korea, backing up his propaganda with absurdly harsh laws like prison for anyone caught in possession of propaganda from the other, richer, Korea (and his secret police were notorious for using torture and often killing people), and one of the largest militaries in the world with a coercive draft long after many other countries abandoned it (the Koreans who managed to flee to America would sometimes discover the state considered them draftdodgers and they didn't dare set foot in their homeland).

With such a starting point, who could hold out any hope for this Korea? Whatever tyrannical & extractive institutions were 'efficient' in this scenario must be far from what one could hope for, and if we jumped forward to look at this Korea in 2013, who could hope that things would've improved substantially? After all:

> institutional constraints can prevent some societies from adopting “superior” institutions that generate higher incomes: their “inferior” institutional regimes can reflect constrained institutional optima.

The Korea I have been discussing is, of course, South Korea; the foreign empire America; the rich rival Korea was North Korea before it stagnated; and the dictator is Syngman Rhee.

[Petrik Runst]

Their views are not that different after all! Institutions are context dependent (Leeson) and one such context is the power structure in society (Acemoglu, Robinson). In other words,North Korea does not undergo institutional change because the power structure restricts its choice set.

[Rafe Champion]

The pirate ship sounds like a company after a worker buyout. What is the international experience of firms owned by the workers?

[Roger McKinney]

Remember the Yugo? It was produced in a worker-owned factory. There don't seem to be many examples, though.

[Roger McKinney]

Hernando de Soto pointed out in one of his books that most states in S America copied US institutions, such as the separation of powers and the Constitution when they won freedom from Spain, but the result has been very different. The culture has to embrace the institutions; they can't be tacked on.

Harrison and Huntington argue for the importance of the culture in making institutions work in "Culture Matters."

[Practicingresurrection.wordpress.com]

@Gwern: that is fascinating. I did not know that history.

[Gene Callahan]

Nice analysis of when democracy works!