Except in moments of international conflict or tragedy, such as the still-recent invasion of Ukraine, foreign aid tends to play a supporting role in the national policy discussion, settling down to a quiet hum next to louder policy concerns that have a more tangible impact on daily life for the American public. Some worry that too much is being spent—and can find comfort in knowing that it has hovered around 0.5 to 1% over the past decade—while others seek to spend more, seeing it as an ethical responsibility to support other nations and groups with fewer resources. Infrequently, the discussion will turn to its relevance to the international balance of power; this type of commentary is usually limited to a radical fringe or the ivory towers of academia, perhaps ironically (or perhaps not).

Putting aside inflammatory rhetoric, US foreign aid is critically relevant to the workings of international politics—but not for the reasons most think. This article looks at some of the unintended consequences of altruism in the form of US foreign aid through a game theoretical lens and finds that, despite reaching for the collectively optimal outcome, the games it generates unleash latent political instability in the recipients of this aid, which is then reflected in the timing of populist swings in several G20 countries. The advantage of this game theoretical approach is that it allows us to identify a critical weakness in foreign aid mechanisms: there can be no mutually dependent outcome in unilateral loan agreements. This results in a more nuanced and practicable argument for reform than those who malign the aid outright: improved agreement structures—ones that foster mutual interdependence and are more conscious of the feasibility of their conditions—could mitigate these undesirable outcomes and foster the cooperation for which aid packages were intended. Rethinking these agreements is vital as multinational conflict looms on the international political stage.

Game Theory in International Relations

Game theory has applications in a number of disciplines, including economics, computer science, political science, and international relations. In international relations, one of its most well-known uses is to describe and predict the outcome of cooperation between nations via the prisoner’s dilemma. This game quantifies a situation in which both players are confronted with a problem that tempts them to act selfishly; however, if they cooperate, they sacrifice less than they would lose when acting in their own interests and the collective outcome is relatively better. The ideological relevance of this dilemma is clear in this age of globalized problems, such as environmental issues, transnational security crises, and international production chains. However, this formulation of the problem is missing certain essential features that make its application inaccurate and, as we shall see later, harmful.

Image

The prisoner’s dilemma and the reach for the collective optimum in international cooperation—the maximum total reward, added between all players—became popular through Robert O. Keohane’s text on the possibilities of liberal internationalism, After Hegemony, and has informed cooperative thinking in the late 20th and early 21st centuries. Foregoing the need for a global hegemon, the idea that sustainable agreements could be reached through mutual interests and low-cost communication seemed promising. This was based on the idea that international cooperation follows the structure of a repeated prisoner’s dilemma game and that cooperation can be generated by a tit-for-tat strategy, one where each player mimics the action taken by the other player immediately before his or her turn (Axelrod, 1984).

Image

These ideas were intended to be applied to “advanced market-economy countries…. [which] hold views about the proper operation of their economies that are relatively similar [to the United States’]” (Keohane, 2005, p. 6). Keohane admitted that they may not be fully applicable to North-South or East-West relations, stating that “[c]areful extension of this argument into East-West and North-South relations, including security as well as economic issues, would be most welcome” (Keohane, 2005, p. 7). Unfortunately, the countries to which this analysis could apply represent only a small portion of the world’s nations and available literature applying these concepts to “East-West” or “North-South” relations is scarce. Given that most countries are outside the Western trading bloc and all of these have participated in foreign aid agreements since the end of the Cold War (North Korea received aid until 2009), one can assert that there is little theoretical development surrounding the proper design of cooperation agreements for the majority of the recipients of US foreign aid.

One of the most important differences in these relationships is the dynamics of foreign aid agreements, which can be defined as “money, services, or physical goods that a country sends to another to help it in some way,” making them, essentially, lending agreements. Amounts vary based on need and compliance with loan conditions, which are set in relation to the perceived risk of misuse or default. The US receives no foreign aid from any country; as such, the US is not subject to the same loan conditions as any recipient country. This means that foreign aid agreements are unilateral lending agreements, precluding the mutuality criterion necessary for a prisoner’s dilemma.  

Image

Additionally, for certain countries, this principal-agent relationship may logically be compounded by a difference in economy sizes: the amount may be large enough relative to the cooperating country’s economy that it would rationally decide to cooperate under almost any conditions. As a result, the foreign aid game is really a mixture of two different games: smaller nations play the subjects in an ultimatum game while the US plays the volunteer in its own volunteer’s dilemma.

This by itself is not a surprising finding, especially for realists and economists. However, when following and solving this game over its various stages, this analysis reveals a secondary—but extremely important—effect: confusion and indifference by the receiving country in the wake of the agreement, rather than (or at least prior to) the irrationality or antagonism often implied in whitepapers, discourse, and the academic literature. 

The Effects of the Foreign Aid Game on Cooperation Stability

While, in a sense, this is a repeated game, the payoffs of each player’s strategies differ slightly in each round (each congressional year) as the budget is modified yearly and the realities of policy implementation follow their imperfect trajectories. In a true prisoner’s dilemma, this would be of no consequence; however, given the structure of this game and the substantial economic asymmetries, this means extreme volatility and game dissolution. This point is best illustrated by analyzing the Merida Initiative, a bilateral security agreement between the US and Mexico from 2007 to 2019, where the US played what would have been the collective optimum in a prisoner’s dilemma. Given that the real game was not a prisoner’s dilemma, but a mix between an ultimatum game and a volunteer’s dilemma—hereinafter called “the foreign aid game”—this led to confusion in the receiving country (Mexico), then withdrawal from international cooperation and into populism.

The Merida Initiative was structured so that the US would provide equipment and training to Mexico, conditional upon a full-scale reform of the Mexican justice system from an inquisitorial to an accusatorial system, as well as other stipulations. It also encouraged information sharing and institutional cooperation, but the only resource flow was unilateral, from the US to Mexico, and Mexico was the only actor whose actions were materially conditioned. Thus, there was a dual process at play, where the cooperative objective was established outside of the concrete rules of the game. The material game only supported the strategy that would lead to the collective optimum on Mexico’s side, so for the US, acting rationally on a collective level meant acting irrationally on an individual level (in the economic sense of the term). Looking at a combination of theoretical expectation and empirical evidence, we can identify mechanism contradictions and consequences.

Table 1. Simplified representation of the first stage of the Merida Initiative game using security resource allocation as payoff

Image

Source: Hirsch (2021)

Above is a simplified representation of the agreement as a game in its initial state, constructed based on Mexican and American government materials outlining the terms of the agreement. Given this deal structure, the payoffs to this game are represented by the final resource allocation as a function of the resources transferred and a change in the security production function. Changing the security production function would amplify the effect of resource transfer, which would increase Mexico’s final payoff beyond the simple addition of utilities. In essence, Mexico could only benefit by cooperating and could not benefit by defecting; meanwhile, although the United States would lose a relatively small amount in resources, it would gain this back (and then some) when considering the collective optimum.

Image

Solving the above game for the maximum payoff given the opponent’s strategy, we can see that Mexico would always prefer to cooperate, while the US would never prefer to cooperate. By summing the payoffs in each box, we can also see that the collective optimum is where both Mexico and the US cooperate. Keeping in mind the real historical actions taken by both players, we know that the US continued to provide aid in hopes of achieving the greater collective payoff, making its actions individually irrational in the context of the game, but collectively rational in terms of the theorized collective optimum. At this point, the loss of individual utility for the collective benefit could still be justified as a rational tradeoff. 

Nonetheless, looking at the game from Mexico’s perspective, it is easy to see where the incongruency caused practical problems. First, the fact that the United States acted in an individually irrational manner meant that it acted unpredictably; although its actions were altruistic, threats to defect (stop aid) upon perceived non-cooperation from Mexico were credible because defection was the only individually rational strategy. This would naturally generate doubt in the Mexican government about US commitment. If this were not the case and the US received some tangible benefit from cooperating, this problem could be overcome by Mexico playing a mixed strategy, contingent on the expected payoff for the US. However, as the US’ real actions were at odds with its expected actions Mexico could not predict US behavior as a response to its own. 

This issue by itself would not prevent Mexico from cooperating; however, overhauling a justice system in eight years is quite an undertaking. As long as justice reforms were effective, cooperating would always be preferred to defecting for Mexico. However, the assumption that Mexican justice reform had its intended effects is not trivial: when the agreement was working well and the utility of cooperating was noticeably higher than defecting, this was not a problem, but opposition in the Mexican Congress, overly-ambitious loan conditions, and administrative failures on either side reduced the utility of justice reform and called into question the dominance of cooperation as a strategy for Mexico. With neither internal incentives nor external cues promoting foreign involvement, strategic interaction started to border on symbolic at best.

This is how the game unraveled: with moving strategic guideposts, domestic political costs became a better guide and latent political instability took center stage. Predictably, the agreement fell through on the Mexican side in 2019 (which, notably, was not officially recognized by the US until 2021). The current Mexican government is widely recognized as populist, with anti-neoliberalism and centralizing nationalistic policies being some of its hallmarks. 

Asymmetric Foreign Aid and Populism

If the game structure that led to populism in Mexico is representative of other foreign aid agreements in other countries, we can expect two related patterns regarding the stability of agreements and the appearance of populist leaders. First, the strategic instability mentioned above would lead to greater instability in cooperation agreements, observable by the variation in the proportion of GDP formed by foreign aid in medium-sized economies, as opposed to large or small economies. Countries with large economies—and presumably smaller proportions of their GDPs represented by foreign aid—would be more likely to experience indifference between cooperating and not cooperating and thus would be less dependent on aid, making the choice of the collective optimum less consequential and cooperation more stable. Meanwhile, countries with small economies and a greater proportion of their GDP represented by US foreign aid would be more likely to always cooperate, precisely because of the low probability of indifference between cooperating and defecting: in this case, the benefits of cooperation could be so high as to induce cooperation under practically any conditions, also resulting stable cooperation. Medium-sized economies would be of interest because their incentives to cooperate are more likely to dance around a threshold that sometimes rewards cooperation and sometimes defection, while never quite reaching a level of indifference that would allow them to ignore the material game entirely and focus on the collective optimum.

The second pattern is related to the appearance of populist leaders. As a political ideology or theory of governance, populism can take on many forms and span the spectrum from left to right; however, at its core is the veneration of a group labeled the “common people” over the vilified elites. This focus on the common people tends to manifest through nationalist rhetoric, in contrast with the internationalism required by cooperation. Thus, if the full game is generalizable and countries do, in fact, turn to populism when faced with mixed signals and varying incentives in the international arena, we would expect to see a pattern of a period of a relatively high proportion of GDP formed by foreign aid followed by a swing toward populism.

This is because, as the positive outcomes from change erode, the benefits of reform tend towards zero, disincentivizing cooperation for its own sake. In addition, the US acts altruistically but credibly threatens defection, making it an unpredictable game partner. With these two concurrent phenomena, the idea of a strategic equilibrium that would dictate an outcome becomes meaningless because there is neither an internal pull nor a reliable external cue to interact. On the material game level, it becomes an "unprofitable" venture to form a political strategy around international cooperation; what is more profitable is paying attention to the domestic political base and directing policy and rhetoric to more nationalistic values.

Both of these trends are reflected in the data. Looking at the G20 members, there is a progression of the standard deviation in aid amounts, with middle-low and middle-high income countries at higher standard deviations and high-income countries at lower standard deviations:
 

Figure 1. Standard Deviation in Foreign Aid for G20 Countries as Proportion of GDP, 1990-2021

Image

Source: ForeignAssistance.gov and World Bank Data: GDP (Current US$). Aid amounts averaged across budget requests, appropriations, initial allocations, disbursements, and obligations to capture all facets; GDP provided by calendar year and aid provided by fiscal year, as available. Created by author.

Additionally, the following graph shows the appearance of well-known populist leaders after peaks or spikes in foreign aid in several countries outside of the Western European bloc.

Figure 2. Fluctuation in US Foreign Aid by Year, 1990-2021

Image

Source: See Table 2. (PM) and (President) refers to the start of the respective term for those leaders who have held both positions.

Although this pattern still requires formalization and hypothesis testing, this offers some evidence that suggests that high proportions of US foreign aid relative to GDP may be correlated with instability and populist trends, with faster responses in countries with a parliamentary system.

Conclusion

Now, to confuse this volunteer/ultimatum game with intentional malice from the United States or to consider this article an argument against US foreign aid would be fallacious; neither an overly cynical nor an overly naïve perspective on foreign aid would be helpful or realistic. Foreign aid does have concrete results and benefits through many programs devoted to citizen well-being. Additionally, corruption and defaulting on international debt are often concerns, justifying a certain level of conditionality. However, that does not mean it is not worthwhile to look under the hood and reexamine the mechanisms that drive our international relations and determine where they may be producing undesirable effects.

The benefit of this type of analysis is that, by determining that international cooperation is not a prisoner’s dilemma at present, we can start to consider how to steer it in that direction and thus be able to apply the tools so carefully designed for that situation. For example, borrowing or receiving services from medium-sized economies, even at very small amounts, would create interdependence and make rational play more feasible for the non-US country. Examining the feasibility of policy conditions is another important step that could mitigate—but not solve—the inherent problem. Further specifying the pattern behind this phenomenon will also be an important tool for identifying and mitigating risk cases, as individual contexts will determine the severity of the change, the time-to-populism post-peak, and the direction (left or right) of the swing. While more research is certainly needed, this should be a point for serious consideration given the consequences.

Bibliography

Axelrod, R. M. (1984). The Success of TIT FOR TAT in Computer Tournaments. In The evolution of cooperation (pp. 27–54). Basic Books.

Hirsch, S. (2021) El juego de la cooperación internacional: una propuesta metodológica para observar el efecto de la asimetría en las relaciones internacionales sobre las políticas públicas domésticas a través del control de armas ilegales en México y los Estados Unidos.  Master’s Thesis available at http://repositorio-digital.cide.edu//handle/11651/4485

Keohane, R. O. (2005). After Hegemony: Cooperation and Discord in the World Political Economy. Princeton University Press.

Krueger, J. I. (2019). The Vexing Volunteer’s Dilemma. Current Directions in Psychological Science, 28(1), 53–58. https://doi.org/10.1177/0963721418807709

Mankiw, N. G. (2009). Macroeconomics (7th ed). Worth Publishers.

Ultimatum game https://www.cs.mcgill.ca/~rwest/wikispeedia/wpcd/wp/u/Ultimatum_game.htm
 

The views expressed in this article are those of the author and do not represent those of any previous or current employers, the editorial body of SIPR, the Freeman Spogili Institute, or Stanford University.