Why did Pirates Choose Democracy?, and Some Thoughts on the “Efficient Institutions View”

Peter T. Leeson

On WhyNationsFail.com, Daron Acemoglu and James Robinson provide a fascinating discussion of pirate democracy in which they were kind enough to discuss my work on the subject. As they point out, I characterize pirate democracy as an efficient institutional response to a principal-agent problem. Pirates were worried that if they delegated power to their captains, their captains would abuse that power, as their captains had done on the merchant ships pirates sailed on before turning to piracy. Constitutional democracy permitted pirates to address this problem.

Acemoglu and Robinson disagree with my explanation for pirate institutions, however, which they describe as a manifestation of the “efficient institutions view.” Instead, they argue that, like other democracies, pirate democracy is better explained in terms of power. As they put it, “Democracy arises when nondemocratic elites are forced to cede power to the previously disenfranchised.” In pirates’ particular context, Acemoglu and Robinson suggest, democracy likely “arose because the distribution of de facto power was fairly equal” among crewmembers.

So whose theory is right? My theory, based on a principal-agent problem, or theirs, based on power?

Answering this question, or at least shedding significant light on it, may be easier than Acemoglu and Robinson suggest. They state that “Leeson doesn’t present any evidence on why pirates chose democracy and the separation of powers,” and that “they don’t have any evidence either on why pirates chose the institutions they had.” But there is evidence on why pirates chose the institutions they had. And I present it in my 2007 JPE paper on the topic, and in my book, The Invisible Hook.

Before turning to that evidence, though, I want to briefly consider the broader issue of the “efficient institutions view,” which Acemoglu and Robinson critique by way of the former’s important work on the political Coase theorem (or rather its absence). According to that critique, Coasean bargains with institutional rulers can’t be externally enforced since the task of enforcement itself typically falls to rulers. Any such bargains must therefore be self-enforcing and, in general, self-enforcement isn’t up to the task. Coasean bargains thus precluded, inefficient institutions abound.

Acemoglu and Robinson are right to characterize my view of pirate institutions as efficient. Indeed, I follow George Stigler in regarding all long-lasting institutions as efficient. I’ve tried to defend this perspective in my work, which shows that a wide variety of institutions that seem obviously inefficient—and, indeed, sometimes downright absurd—are in fact, on closer inspection, efficient and not so absurd after all. (See, for example, my research on ordeals, trial by battle, human sacrifice, and vermin trials).

Am I then suggesting that the variety of institutions we observe reflect Coasean bargains after all? No. I’m suggesting that the absence of Coasean bargains doesn’t imply institutional inefficiency and often implies just the opposite.

My reason for thinking that long-lasting institutions are efficient isn’t rooted in the Coase theorem, which I agree is often (though, not always) inapplicable in institutional contexts. Rather, as it was for Stigler, my reason is rooted in the simple fact that inefficient institutions can’t survive long in a world of rational people. In this “efficient institutions view,” Coasean bargains aren’t required to get to efficient institutions. In fact, efficient institutions are often efficient precisely because they don’t reflect such bargains.

Coasean bargains aren’t free. Indeed, sometimes they’re very expense. Their expense stems from the cost of arranging and enforcing them, which in some cases isn’t possible at all, where they’re infinitely costly.

Efficient institutions maximize social wealth net of such costs. Where these costs prohibit Coasean bargains, as they often do, efficient institutions won’t involve Coasean bargains. If Coasean bargains were free, many institutions would surely look different than they look now, just as if platinum were free, railroad tracks would look different than they look now. But prohibitive transaction costs no more imply inefficiency in existing institutions than prohibitive platinum costs imply inefficiency in existing railroad tracks.

Returning to the reason for pirate institutions: Did pirates choose a system of democracy and separated powers to solve a principal-agent problem, as I argue? Or did they choose it because of equal power, as Acemoglu and Robinson suggest?

The evidence we have that bears on this question, and I adduce in my previous work, is surprisingly direct. Pirates actually tell us—or rather told their contemporaries—why they adopted their institutions. And their reasoning is preserved in the historical record.

As pirate Walter Kennedy testified at his trial:

“Most of them having suffered formerly from the ill-treatment of Officers, provided thus carefully against any such Evil now they had the choice in themselves . . . for the due Execution thereof they constituted other Officers besides the Captain; so very industrious were they to avoid putting too much Power into the hands of one Man” (Hayward [1735] 1874, 1: 42).

This pirate, at least, explicitly linked the desire to avoid captain abuse—the “ill-treatment of Officers”—which pirates suffered as merchant seamen in their pre-piratical employment, to pirates’ reason for adopting separated powers. This is about the most direct evidence for the principal-agent theory we can hope to get from the 18^th-century historical record.

This isn’t to say that small asymmetries in pirate power didn’t have important influences on pirate organization. For example, roughly equal pirate strength surely helped to make captains’ agreements with their crewmembers self-enforcing, enabling a piratical Coase theorem to operate. But if we take pirates’ own explanation for their institutions at face value, which I think we should, the raison d’etre of pirate institutions was a principal-agent problem.

Comments

[Richard Schulman]

The Leeson explanation seems plausible and well documented for some 18th century piracy, when crew abuse was widespread and the pirates were European in origin, but I would question whether his explanation can be extended to contemporary piracy, which occurs in a maritime context in which crew abuse is relatively infrequent and most of the pirates come from strongly hierarchical societies.

Piracy can be state supported -- and indeed, it was in the 16th through 18th century via letters of marque and privateering. It can also be conducted by hierarchical criminal gangs. The latter form of organization would be my starting hypothesis for much of today's piracy.

[Gu Si Fang]

It seems to me that both theories do not contradict each other and do not apply to the same situations. Pete Leeson talks about voluntary cooperation, whereas Daron Acemoglu talks about coercive extraction. Political institutions in the Acemoglu are by no means efficient, but social institutions in the Leeson sense tend to be.

[Troy Camplin]

One might be able to reconcile the two by recognizing the fact that pirates are going to be of a particular psychology more often than not.

In Clare Graves' emergentist psychology, we evolve from tribal to heroic (think ancient Athens) to authoritative (think Medieval Europe) to classical liberal to egalitarian (think of the line from Rousseau to postmodernism) to integrationist (Hayek) to holistic (Max Borders, me, probably the Bleeding Heart Libertarian people). Pirates seem to be at the second level, which is typified by being interested primarily in power and in rejecting authority. Which sounds like a pirate -- or a gang member. If there are a few around, they will make empires which they will rule, but if society is dominated by them, we will get an Athenian-style democracy.

The institution of democracy is going to emerge precisely because people at this psychological level are not going to put up with having someone over them. Either they are going to be in charge, or power will be shared equally. We see this too in gangs, with rituals emerging to help mediate the levels of hierarchy that do emerge.

It seems, then, that understanding the level of psychosocial of pirates would help one reconcile these two views.

[Jonathan Conning]

Nice discussion but couldn't the same quote be used to support Acemoglu-Robinson's hypothesis that pirates didn't like the abuse of the hierarchical merchant ships, hierarchy that mirrored (and got help at enforcement from) the non-democratic societies under whose flag they flew. Operating as they often did, without the backing of such state institutions, captains would have been forced to cede power to "the previously disenfranchised".
A simple test of this might be to ask whether privateers (who did have that state backing) were less likely to run their ships democratically compared to other types of pirates.