Dani Rodrik, surrounded by anarcho-sympathizing libertarians at Cato Unbound, thoughtfully discusses the limits of self-enforcing agreements. In response to Peter Leeson’s claim that we don’t see “shriveling international commerce in the absence of supranational commercial law,” Rodrik writes:
It may be objected that the operation of the global economy is proof in itself that a high level of economic activity can be maintained without political institutions… But Leeson is overlooking several things. First, there is in fact a significant global institutional architecture that supports the international economy: globalization would not have reached this far in the absence of the WTO, IMF, World Bank and a host of regional supranational institutions. The global public sector is not non-existent.
Now the World Bank and IMF cannot claim credit for inducing developed countries to liberalize their trade policies, nor do they get much credit for Indian or Chinese policy shifts, so it seems that the bulk of responsibility for the relatively open global trading system would fall upon the WTO. But even those (Subramanian & Wei) claiming that the WTO promotes trade concede that it doesn’t force countries to liberalize, which is why developing countries remain relatively closed and sensitive sectors “did not witness liberalization.” The WTO is not an effective supranational institution if that term means enforcing policies contrary to participants’ wishes.
In fact, a leading theoretical explanation for the WTO is that it is a successful self-enforcing agreement! Bagwell & Staiger:
[T]he optimal unilateral tariff for a national-income maximizing government of a large country is positive. If both governments behave this way and set positive import tariffs, a Prisoners’ Dilemma situation is created. In the Nash equilibrium, tariffs are too high and trade volumes are too low; hence, a trade agreement that facilitates a reciprocal reduction in tariffs could be mutually beneficial…
The terms-of-trade theory is easily generalized to include political considerations, and it may be directly interpreted in the context of the market-access language that trade-policy negotiators use. This theoretical perspective offers a means by which to interpret the rules of GATT/WTO. For instance, it suggests that a government may hesitate to liberalize unilaterally, since it does not want to face the terms-of-trade loss that such behavior would imply…
Likewise, a government would hesitate to liberalize as part of a reciprocal negotiation, if it were concerned that its negotiating partner might later “cheat” and raise its tariff. We argue that the GATT/WTO enforcement provisions can be interpreted in this light…
As there are no GATT/WTO police, agreements between governments achieved through GATT/WTO negotiations must be self-enforcing. Indeed, the rules of GATT/WTO may be interpreted as a codification of supergame strategies.
Rodrik may be right about the limits of self-enforcing agreements in most circumstances, but the liberalization of international trade seems to be a counterexample to his generalization.
[Of course, Rodrik is also writing about the protection of property rights and enforcement of contracts across national borders, which may indeed depend upon states making legal arrangements, but that's not the role of the IMF, WB, nor WTO.]