The Problem of “Picking Winners” in the Study of International Cooperation

Regarding the first point about the observability and quantifiability of existing international agreements, it behooves us to step away from the headlines and news flashes about imperiled international agreements. Rather than focusing on the latest announcement in the Brexit saga or the next bit of breathless news about yet another international agreement that U.S. president Donald Trump looks set to shred, we need to appraise these developments through a framework that considers the issues of inference and selection.Footnote ²

What social scientists call “selection bias” bedevils the study of international cooperation at every stage, both in the formative years when countries decide on what and with whom to cooperate and throughout an IO's entire life span. Our view of the overall health of international cooperation is biased not only by focusing solely on the “successful” organizations and glossing over the ones that work less well but also by overlooking similar agreements among similar groups of actors that died off or never fully materialized.

In other words, most empirical studies of IOs only focus on those that we observe as being active in the world today. They do not take into account proposals for international cooperation that never got off the ground, or plans that got derailed, or organizations that have been disbanded. There is an understandable set of reasons for this: it is much easier to measure things that occur than it is to measure things that do not, and it is difficult to know how seriously we should take failed proposals for international cooperation.

But if we focus only on organizations that we perceive as “working” without looking closely at similar agreements that might have failed, we are not taking into consideration the whole sample of international agreements. That means that any lessons we attempt to draw from the “good” organizations are incomplete because we are not drawing comparable lessons from agreements with similar designs and membership constellations that did not work as well; we are only considering one side of the coin.

So, for example, rather than wringing our hands over whether NATO would survive the withdrawal of the United States, we should take a broader view that investigates the conditions under which similar endeavors in the Middle East and Asia collapsed.Footnote ³ Additionally, we might examine why it was that NATO found a way to repurpose after the end of the Cold War despite losing its ostensible raison d’être, while the Warsaw Pact did not.

Similarly, rather than worrying about what looks to be a trade war between the United States and China, and agonizing over what this might mean for cooperation through the World Trade Organization, we would do well to recall the conditions that led to the demise of an earlier trade organization—the International Trade Organization—which the United States proposed and designed but then subsequently killed in congress.Footnote ⁴

The lessons from some of these “dead” agreements are that domestic interests and having the agility to reshape an agreement's mission in the face of crisis can determine whether international agreements get salvaged or killed off. But scholarship has tended to sweep these agreements under the rug, so we are left with an incomplete understanding of how they compare to the agreements that managed to survive.

International Cooperation along Four Paths

As noted above, there is another problem with only looking at the organizations that survive and glossing over the ones that do not. There is often an unspoken/unwritten assumption in the literature that IOs more or less operate as planned after they are founded. For this reason, much scholarship focuses solely on the early days of an IO's existence—the enlisting of initial members and the writing of the initial charters, with attention being paid only to formal changes in membership and charters that might occur when the IOs get further down the line. To that end, our understanding of how IOs evolve over time tends to be limited to an organization's expansion to include new members, or the decision to renegotiate and revise the organization's charter.

It is, of course, understandable to focus on those big moments; adding new members and producing new charters are procedures that are easy to observe. But fixating only on these highly visible occurrences in an IO's long life span overlooks the subtler dynamics of how it functions over time.

Most IOs are better analyzed not by using binary metrics (they either exist or they do not; a country is either a member or it is not; the charters have either changed or have not) but instead by using four different categories that provide a more complex picture of their existence. If we survey the actual trajectory of international cooperation, we see that most IOs fall into one of these categories of existence: uninterrupted life, death, inertia, or change.

I have already described the “death” category above: organizations that either never came to be, or that have but have since ceased to exist. But even organizations that are technically alive can still be drifting in a state of “inertia.” For example, if at some point after its founding an organization hits a major road bump in its attempts to orchestrate cooperation among its members, it may nevertheless remain in existence, but over time it may have fewer and fewer meetings and an ever shrinking budget. Once built and staffed, organizations can be surprisingly resistant to disbanding.Footnote ⁵ A large number of organizations and over a third of international economic agreements fall into what I have elsewhere called the “zombie” zone, where they still continue to operate but make few meaningful advances in cooperation.Footnote ⁶

Of course, as is the case with the death of an organization, it is difficult to tell if inert zombie organizations are good or bad for the state of international cooperation more generally. There may be no harm in keeping IOs around in the event that they might be useful someday; conversely, inert organizations might be a drain on countries’ resources, and they might obstruct more useful attempts at cooperation.

Similar problems plague the remaining two pathways: “change” and “life.” Scholars tend to assume that when organizations change, they do so for the better and in pursuit of a more efficient outcome. This can sometimes be the case: If cooperation proves difficult in one area, an organization can either formally or informally shift its mandate to a more productive mission. But it does not take much for that change to drift toward pathological outcomesFootnote ⁷—a mission far different from what its founders intended, or a structure that becomes a hive for bureaucrats who find ways to prolong their jobs regardless of whether the organization itself is fulfilling its mission. Thus, change in an IO can either be a positive adaptation toward better outcomes or result in “mission creep” for the organization.

This normative ambiguity holds true for organizations that maintain uninterrupted life as well. Life without adaptation is a luxury that most IOs cannot afford. Eking out a long, uninterrupted existence is a far easier task for agreements with circumscribed mandates than for organizations that try for more ambitious modes of cooperation. Those organizations that do end up surviving without any change to their mission usually cover relatively narrow technical issues, such as currency unions or navigational agreements.

The stumbling often seen in more ambitious and more complicated organizations should therefore not be surprising. Life without change may be too high a bar to which to hold those organizations.

Life, Death, Inertia, and Change in International Trade

The myriad organizations that have attempted to govern international trade provide some good illustrations of the four categories of IO existence detailed above. As a starting point, it is worth recalling that the classical rationale for liberalizing international trade hinges on the logic of universal participation. The principle of comparative advantage tells us that the more countries that agree to liberalize trade, the greater the consumer surpluses, as more and more individuals gain access to a wider diversity of goods at lower prices as a result. In other words, trade is meant to work best when every country is on board. Furthermore, the remit of governments in the classical logic of trade cooperation is meant to be fairly restricted: they simply agree to inaction. That is, they agree not to protect domestic goods or put up barriers to foreign products, and instead let firms and markets do the work. In this naïve view, reciprocal trade cooperation should be fairly straightforward and trading institutions should be relatively simple and inclusive.

In reality, things quickly get complicated: domestic producers are not keen to have their goods compete with foreign ones, and they can lobby and vote for protection even in the face of freer trade. And thus the history of trade cooperation is anything but simple. After the death of the International Trade Organization (ITO)—an agreement designed in the international spirit of Bretton Woods but killed in the U.S. Congress—a far more modest organization emerged in the form of the General Agreement on Tariffs and Trade (GATT). The GATT had been negotiated in parallel with the ITO, but included far fewer countries and a more limited set of commitments. While fifty-four countries had signed on to the ITO, the initial GATT group contained only twenty-three members, and they were far more similar to one another in terms of level of development than had been the proposed ITO members. It also, crucially, left agriculture—a contentious issue area in trade, with entrenched interests opposed to its liberalization—off the table, focusing instead on opening markets in manufactured goods.

What we might think of as the “uninterrupted life” of IOs according to the typology above—the uninterrupted fulfillment of their initial mandate—was easy enough for the GATT in its early days. At the time of the agreement's formation, the average level of tariffs among members was around 20 percent, so they had nowhere to go but down. And the countries that signed on to the GATT had already expressed a willingness to drop tariffs among those trusted partners, while a broader and deeper set of commitments had been killed off with the ITO. This is another good illustration of the “selection” problem described earlier, where countries sign agreements that require commitments to things that they wanted to do in the first place. The GATT is, perhaps unsurprisingly, largely regarded as a victory in trade cooperation during the time period when it stuck to its original mandate, up until its ambitious sixth session, the Kennedy Round, a set of trade negotiations that kicked off in 1964.

That success meant that change was coming for the GATT. Excluded members from the developing world, some of which had been part of the original ITO, wanted to reap the gains from trade, but they wanted to participate on different terms than those defined by the initial grouping of rich member states. These new potential members wanted a different set of institutional rules including, for example, a formal dispute-settlement mechanism rather than the series of “gentlemen's agreements” that were the standard for resolving conflicts in the GATT, as well as different schedules for the liberalization of sensitive products. In addition to pressure from less developed countries, once the GATT members had successfully reduced tariffs in the relatively politically uncontroversial area of manufactured goods, this left the more complicated and sensitive areas of services, intellectual property, standards harmonization, and, of course, agriculture—the third rail of trade liberalization that had doomed the ITO in the first place.

Although the GATT accommodated some of these issues through eight rounds of internal negotiations over the course of forty years, the organization could only withstand so much change without transforming into a new organization altogether. The final agreement in 1994 made way for the World Trade Organization (WTO), which involved a more institutionalized version of trade that accommodated nearly three times the countries of the initial GATT agreements, with an attendant level of complexity that was meant to match the needs of those new members.

Change was not easy, and the newly formed WTO was quick to stumble just a few short years after its founding. The 1999 talks in Seattle were meant to usher in a “development round” of trade deals that focused particularly on asymmetries between poorer countries and the initial GATT members. But protests derailed those discussions, and the subsequent ministerial meeting in Doha ended in gridlock. When new, economically powerful members such as Brazil and China threw their weight behind the developing countries that were arguing for the United States and the EU to scrap their substantial farm subsidies, the negotiations collapsed.

In this respect, the changes that the WTO sought to undertake to accommodate its new members proved to be too much to bear, arguably leaving that organization today in a state of inertia. In the twenty-five years since the WTO was founded, little progress has been made on the organization's main remit of trade liberalization. Although its dispute-settlement mechanism—a judicial body initially conceived of as being a way to level the playing field between powerful rich countries and weaker, poorer ones—remains active, the majority of its cases are between those rich countries that can afford the legal costs of keeping one another locked in court. And even when developing countries win suits against rich ones, the WTO has no means of ensuring that those powerful countries comply with rulings and, perhaps unsurprisingly, the record of compliance is uneven at best.

The problematic record of the ITO, the GATT, and the WTO illustrate the four avenues of IO life cycles: death for the ITO, life giving way to change in the GATT, and inertia in the WTO. But this picture, too, is incomplete, as it does not include the many other institutions governing trade, starting with the web of preferential trade agreements (PTAs) around the world. Meant to be merely a complement to the WTO, allowable only when such agreements would not detract from the broader goal of multilateral liberalization, PTAs are now the rule rather than the exception, with every country in the world having at least one and often scores of such agreements. Negotiations for trade liberalization have now pivoted away from the WTO itself and back to these smaller, more limited agreements, with countries choosing the issue areas and partners that reflect the types of deals that best reflect their interests.

But even among these PTAs, we can observe life, death, inertia, and change. Witness the recent near dissolution of the Trans-Pacific Partnership; the renegotiation of NAFTA into the United States–Mexico–Canada Agreement (USMCA); and the inertia among many of the PTAs between countries that had little to trade in the first place (such as the trade agreement between, say, Singapore and Peru). And even those more limited bilateral trade commitments may be set aside when new leaders with different priorities enter office.Footnote ⁸ Thus, throughout the trade regime, we see cooperation in many different permutations, some more successful than others.

Conclusion: International Cooperation as a Continuum

Perspectives on the robustness of the international liberal order necessarily carry an assumption about how well that liberal order worked in the first place,Footnote ⁹ as well as which benchmarks for success or failure should be used. If one perceives IOs to historically have been in good health, this era of renegotiations and exit will appear to portend doom for cooperation between nations. If, however, we acknowledge that death, inertia, and change are the rule in international cooperation rather than the exception, and we understand that the vast majority of IOs stumble in some form or another in the face of crisis, the current period looks less exceptional. This is not to say that countries should not strive to find ways to cooperate. But we should be aware that the history of international collaboration is far more nuanced than what is often assumed.

The U.S. withdrawal from the Paris Climate Agreement, its renegotiation of NAFTA, and Britain's attempts to extricate itself from the European Union seem like notable events because they look to be mortal blows to the associated agreements, dealt by key parties to them. Furthermore, these are precisely the types of agreements that are included in conventional mental maps of international cooperation: they are the types of agreements that get counted as successes.

However, even these seeming paragons of stable and successful international cooperation have gone through various stages of IO existence. The EU was in the throes of crisis back in the early 1990s, with not only the U.K. but also many Southern European countries such as Italy, Spain, and Portugal being forced to abandon the exchange rate mechanism; the U.K. never returned, and the other countries faced a second reckoning during the 2008 European sovereign debt crisis. The U.S.-led renegotiation of NAFTA, much like the U.S.-proposed ITO, tended to center on provisions that the United States itself had insisted on including in the original NAFTA agreement. And the Paris agreement is only the latest in a round of climate accords that seem to reaffirm what countries had already established as national goals, rather than truly charting new areas for cooperation—and similar agreements, such as the Kyoto Protocol, have also witnessed withdrawals and defections over the years from major parties such as Canada.

This further illustrates how doing the more complicated work of measuring cooperation has always been more important than simply acknowledging that an IO exists; cooperation can take many different avenues and organizations very rarely have an unswerving life span. It also illustrates how little we know about these life spans. We still lack a real understanding of when organizations can rebound from crises and how cooperation can become more durable. This necessitates further scholarly attention and research.

We also lack a sense for when the death of an organization is a good or a bad thing for overall levels of international cooperation. If an organization does its job so well that it writes itself out of existence, this may be a welcome development; if it really has ensured cooperation in the issue area that it was set up to service, there might not be any further need for that organization to continue operations. Or, if an organization is simply failing altogether, shutting it down may be better than draining resources to try to keep it afloat. These examples should give us pause when looking at the state of international cooperation today; sometimes, an honorable death may have its advantages.

To that end, some of our current crises may only be pointing out the longstanding fault lines in existing organizations. If organizations are heading toward inertia, then change or dissolution might not be a negative outcome. Much depends on whether by following such change or dissolution, countries can find new grounds for cooperation. They may or may not do so by setting up or preserving an existing IO—but history shows us that simply having an IO in place is no guarantee for harmony.