Incentivizing Cooperation in Afghanistan
By: MAJ Greg Merkl, US Army
The United States' prospects for success against the Taliban insurgency in Afghanistan seem inexorably linked to its ability to inspire cooperation among disparate actors in a land of dynamic alliances. To centralize power in a way that will keep the Taliban from establishing sanctuary—much less overt control—in significant regions of the troubled state, the United States and its allies must find ways to foster a capable Kabul government. The last decade of conflict has only increased the complexity of trying rapidly to build a poor, weakly affiliated agrarian society into the likeness of a modern democracy. Moreover, lessons from the Afghanistan war demonstrate how the United States has made a difficult task monumentally more complicated through poor strategic decision-making.
The United States' profligate use of money to influence loyalties in Afghanistan has confused and hindered efforts at establishing legitimate governance, at the levels of both the tribal shura and the National Assembly. Financial aid certainly serves a vital purpose in rebuilding a war-ravaged land. Too little aid can leave a broken state in the violence of desperation. Conversely, too much aid can drown a fledgling legal and political system in corruption. A "sweet spot" exists where the structure of the financial incentives encourages cooperation according to the rule of law. The United States has substantially overshot this sweet spot.
As long as the Department of Defense is going to espouse the concept of "money as a weapons system," it would do well to reestablish the left and right limits of its financial weapons.1 A realist's perspective on the nature of cooperation highlights the fact that the United States has spent hundreds of billions of dollars inspiring counterproductive behavior within Afghan society. The influx of cash into the Afghan system of social and commercial relationships has not only overwhelmed its modes of accountability, but actually encourages unethical and anti-social behavior.
Solving the Prisoner's Dilemma
The paradigm of the "Prisoner's Dilemma" illustrates how natural impulses tend to favor competition over cooperation. In its simplest outline, the dilemma postulates a scenario whereby two individuals can choose to cooperate for the best mutual outcome, or compete and risk the worst mutual outcome. The incentive structure paradoxically leads them to compete, because each believes he can gain a unilateral advantage through competition.2 This scenario plays out frequently in most societies.3
To illustrate this concept, consider the hypothetical scenario of Daoud and Omar, two farmers in modern-day Afghanistan. Each has a large herd of sheep, and they share grazing lands. If they share the land reasonably and responsibly, both can adequately maintain their flocks. Daoud could, conversely, try to force Omar out of the pasture so Daoud's sheep get all the grass. If Omar submits without a fight, Daoud's herd will increase by five at the expense of 10 of Omar's flock, for a net loss to the community of five sheep. The reverse is also true. If, however, they both attempt to seize the entire pasture, they will get into a costly conflict. Some of their sheep will die from starvation, and they will have to sell others to finance their feud. In this eventuality, each will lose five sheep, for a net loss of 10. The matrix in Figure 1 shows the "normal form" of the game: the first number listed in each cell is the change in the number of sheep owned by Daoud and the second number is the change in the number owned by Omar.
One can quickly see that Daoud should always compete. If Omar is foolish enough to cooperate, then Daoud will see a gain of five sheep by competing. If Omar chooses to compete while Daoud remains passive, then Daoud will lose 10 of his flock to starvation in a naively benevolent gesture of cooperation. By competing, he still loses sheep, but he loses five instead of 10. The exact same logic applies to Omar. As a result, the tendency for both will be to compete, despite the fact that this will lead to the worst possible mutual outcome. This incentive structure bodes poorly for the community. The cooperate–cooperate strategy clearly offers the most total utility for the society in which Omar and Daoud live, but the incentive of personal gain will reliably drive a self-interested player to compete. So, how can society balance the incentive scale to ensure that it does not bear the opportunity cost of 10 sheep in this ill-fated conflict?
The Beauty of Punishment
The simplistic model described previously clearly does not tell the entire story. Although history recounts tales of horrific conflict, communities have managed to control violent contests between individuals or groups in order to build a better future for all members. Nevertheless, numerous studies across multiple disciplines have drawn the same conclusion: People are not naturally motivated to cooperate. They cooperate largely because society enforces cooperation.
Heuristically, one may assume that a person cooperates as a spontaneous outpouring from a kind heart. Most of us need only to look within our own family to see that bonds of trust and shared interests can mitigate the competitive urge ascribed to the self-interested player in the Prisoner's Dilemma. Then again, everyone has heard a story or two about the time when Omar broke into Daoud's piggy bank to splurge on whistle pops and comic books at the corner store. It does not take many incidents of cheating within a family, a neighborhood, or a community to erode the bonds of trust. In fact, as the bonds of community expand past the family and the neighborhood, the temptation to cheat increases in a vicious cycle.4 This leads to the "tragedy of the commons," in which self-interested players pursue outcomes that enrich themselves in the short term but hurt the community at large for the long term.5
Why, then, do members of a community generally cooperate? Simply put, people cooperate because social institutions force them to cooperate. Even though Daoud sees an incentive to cheat Omar, his local shura has a clear interest in ensuring that he cooperates. The normal form of the game shows that, taken together, a society maintains its status quo through cooperation. The compete–cooperate and compete–compete strategies promise a loss of five and 10 sheep, respectively, to the community. To forestall this, successful societies force cooperation when it serves the interests of the community, by punishing those who attempt to compete.6
Fortunately, Daoud and Omar have a powerful local shura. If either decides to cheat the other by usurping the grazing fields, the shura will punish the infraction at a cost of 20 sheep. Figure 2 shows the new normal form of the game with a penalty of 20 sheep for counterproductive decisions to compete.7 In this scenario, both Daoud and Omar should clearly cooperate.
By imposing a penalty of 20 sheep for failing to cooperate, the local shura will inspire productive cooperation. But, if eliciting cooperation is so simple, why do real people in relationships like Daoud's and Omar's so often choose to battle?
Defying Society for the Risk of Gain
Daoud might cheat for the same reason a burglar might rob a bank: He thinks that he will get away with the crime. To prevent him from taking this risk, the local shura must maintain a good track record of punishing infractions. Again, the normal form of the game gives insight into why the shura must consistently punish cheaters. If Daoud believes that he can compete without incurring the penalty of 20 sheep with a probability of α, his prospects change entirely. Daoud assumes that Omar will follow the law. Therefore, if Daoud cooperates, he expects no change in sheep. If he cheats, he expects to gain five sheep with a probability of α and to lose 15 sheep with a probability of 1 – α. The utility of cheating is then 5α – 15(1 – α). To incentivize cooperation, Daoud's shura must make certain that he expects a better outcome from cooperation. Mathematically, 0 > 5α – 15(1 – α). Therefore, if Daoud believes that α, the probability that he will get away with cheating, is less than 75%, he will cooperate. Note that a stiffer penalty with the same value of α will further reduce Daoud's incentive to cheat. For this reason, many shuras would take all of Daoud's sheep, his favorite goat, and possibly a hand for stealing from the community!
Societies must keep the value of α (the chances of getting away with the crime) low enough to favor the community. In doing so, they encourage a rational Daoud to follow the rules. But sometimes societies break down. And with them goes the incentive to cooperate.
Captain America: The Breaker and Builder of Societies
To emphasize the relevance of these basic incentive structures to current U.S. policy, consider the United States' mission in Afghanistan. Growing up in rural Afghanistan, neither Daoud nor Omar had much vested interest in religious extremists. Following 9/11, however, their district's acquiescence to the local Taliban militia brought hypothetical Captain America— an extremely motivated U.S. Army company commander—to their village. Captain America is an expert in hybrid conflict, so he seamlessly transitions from breaking to building as the situation requires.
Initially, Captain America focuses on breaking. He controls a vast array of 50-caliber machine guns mounted on armored vehicles, automatic grenade launchers, and laser-guided missiles. Moreover, he has 120 motivated soldiers dedicated to eradicating the Taliban from every corner of the little village. In short order, Captain America has used his lethal weapons systems to eradicate—however temporarily—the Taliban. He has shown admirable discretion with his fire control measures, but the fighting has still taken a toll on the village. Much of the mud-engineered infrastructure has crumbled from the over-pressure of 500-pound bombs, and the members of the village shura have gone into hiding.
Under these conditions, Daoud may now be tempted to cheat the system for two reasons. First, the absence of his shura might lead him to believe that he can get away with seizing the pasture from Omar without punishment. That is, he faces the incentive structure of the Prisoner's Dilemma. Second, he may cheat out of desperation. Assume that the war has truly ravaged the community and that Daoud's family is living in poverty. He and his family will starve if he cannot gain more sheep. A despondent Daoud has no choice but to compete. He knows that he will get no additional sheep from cooperating, so he does not even consider choosing this option; it leads to guaranteed starvation. Instead, he sends a prayer toward heaven and competes. He does not expect a good outcome, but it is truly his only chance at survival.
Fortunately, Captain America has a generous war chest for rebuilding the community, and he easily voids both of these conditions. He quickly reestablishes the shura and reconstructs the demolished government buildings. He helps to finance a police force that can establish the rule of law. He also supports aid projects directly aimed at protecting Daoud and his family from starvation. In no time, Captain America has reestablished the pre-war status quo, and he should stop here.
Eager to exceed the standard and inspired by the concept of money as a weapons system, however, Captain America takes it a step further. He initially broke the system through combat. In doing so, he broke the society's ability to enforce cooperation. In rebuilding, he fulfilled his debt. Now, however, by surpassing the standards of restoration, he will actually break the community through his naïve benevolence.
Breaking with Benevolence
A catastrophic loss clearly undermines the power of a community to enforce cooperation, but a sudden major gain similarly ruins the incentive structure. Afghanistan provides a detailed case study to illustrate the ways in which this occurs. The primary problem with quick riches lies in their affect on the community's ability to enforce punishment in dynamic economic systems.
Captain America meets Daoud in a shura meeting and decides that he would like to purchase his heart and mind. Daoud appears amicable, intelligent, and relatively influential in the community, so it seems perfectly logical for Captain America to expend his CERP (Commander's Emergency Response Program) funds on Daoud's behalf. Soon, an Afghan contractor appears and digs a well close to Daoud's farm at a price of about $1,000, an amount Daoud would never have managed on his own. Daoud is now perceived to have the backing of Captain America—a force far more powerful than his local shura. The next day, Daoud ruminates while minding his flock, and considers forcing Omar to accept less of the grazing grounds. Daoud is a big deal in the village now, as he sees it, and Omar should feel grateful to get any land at all for his sheep to graze. Ultimately, however, his gentler sensibilities prevail, and Daoud settles for the status quo.
Over time, Daoud's relationship with Captain America blossoms. Daoud always speaks in support of Captain America at the shura, and he provides Captain America with frequent—and occasionally accurate—reports on the town's business. Daoud soon takes note of the American captain's affinity for digging wells. Never a scrupulous businessman, Daoud decides to take a chance and calls a cousin in Kandahar who owns a drilling rig. Soon, the two have a booming business digging a well a week for the development-minded commander. Daoud starts to wonder how the community ever procured water prior to the infusion of benevolent aid money. While Captain America has purchased Daoud's heart and mind, Daoud has followed suit, just like countless other Afghan contractors, and purchased a mansion on Kabul's Street Six with his newly generated wealth.8 One day, Daoud decides to force Omar completely out, and he seizes their shared pasture. At this point, neither the shura nor Omar can do much to discourage this behavior.
Two factors influence the ability of a community to use punishment to force cooperation in this environment. First, Captain America's implied support of Daoud reduces the likelihood that the shura will censure Daoud for his misdeeds. Second, when Daoud's income suddenly makes herding sheep a pastime instead of a livelihood, he can purchase the consent of the shura with side payments. In the first case, a direct increase in α, the probability that Daoud will be able to cheat without penalty, results from his relationship with Captain America. In the second case, Daoud can make a side payment to the shura, and they will turn the other way while he steals Omar's land. This hypothetical scenario paints the endemic corruption plaguing the artificially bloated Afghan economy in painfully vivid detail.9, 10
In the normal form of the game, one may argue that Daoud has no incentive to offer a side payment to the shura. After all, if the shura confiscates 20 sheep as a consequence for cheating, Daoud would have to pay at least this sum for a modest gain of five sheep over Omar. Here, one must take careful note of the nature of the side payment. Daoud has gained control over the shura. If Daoud values this power at 15 sheep, then he breaks even. In politics, this type of influence carries a value far greater than 15 sheep and a dozen camels. In fact, at this point in the game, Daoud really does not even care about the sheep. Usurping the power of the shura is the prize that he seeks.
Prior to the change in his status from a common sheep farmer to a powerful businessman, Daoud did not have the wherewithal to buy influence because he needed the sheep that he gained through cooperation to support his family. He simply did not have the capital to invest in the political machine. Moreover, a wise shura takes its bribes under careful consideration. If things should go badly for a conspiring member of the shura, Daoud's deep pockets and cozy relationship with Captain America would help smooth things over. Prior to his accumulation of wealth, Daoud could not have mitigated these risks.
Daoud's display of power incurs some risk as well. He must ensure that Omar has enough sheep to survive so that he does not become desperate, which might lead to a direct confrontation between the two. With Daoud as the power broker, the game structure mirrors that of warlord politics. In it, Daoud represents a strong player and Omar represents a weak player.11 A strong Daoud has a dominant strategy of competing, yielding a submissive Omar. Daoud simply must not allow Omar to sink into the realm of despair, because if Omar does, then Daoud will face a belligerent and desperate opponent. Omar has little chance of defeating Daoud, but if confrontation is his only chance of survival, he will unquestionably take the risk. A dominant strategy of conflict emerges in this circumstance. Fortunately, Daoud can evade this dynamic.
Daoud now controls the resources throughout his village, and if Omar needs some sheep, Daoud can easily afford to act as the benevolent dictator in return for continued submissive support. Should the community meet with utter calamity, so that everyone needs more sheep than Daoud can provide, he has nimbly and wisely hedged his bets with his Kabul residence. In all likelihood, Daoud moved his wife and kids to Kabul long ago. The guarantee of a permanent change of scenery and a comfortable early retirement ensure that Daoud will continue to usurp power within his community without fear. Although Captain America certainly bought the heart and mind of Daoud, and was able to report impressive non-lethal aid numbers to his higher headquarters, he simultaneously had an immeasurably negative impact on the culture of mutually beneficial cooperation within his area of responsibility. Daoud, Omar, and their lineage of ancestors survived without Captain America's wells. Because he failed to understand the true needs of the village or the effects of superfluous cash on an economy concerned with grazing sheep, Captain America nullified the community's ability to use punishment as a means for evading the pitfall of the Prisoner's Dilemma.
Captain America's Lesson
Aid money plays a crucial role in a war zone. Yet, the United States seems to advocate spreading money around as a way for its commanders to buy support. While Captain America purchased the support of Daoud, he destroyed the shura—and the community that depended on it—in the process. Instead of using money as a weapons system to buy friends, Captain America should have used just enough aid to restore power to the community government and to avert catastrophic loss. Investments beyond this sweet spot may curry favor with a minority of the community, but they do not build an enduring, self-reliant society.
Admittedly, some circumstances do call for buying friends. When the United States engages in this type of behavior, it must do so with clear-eyed realism regarding the negative effects that using money as a weapons system has on indigenous communities. Current Army doctrine gives little consideration to this phenomenon, which helps explain the current state of affairs in Afghanistan. We can imagine that every new mansion springing up in Kabul's suburbs memorializes a dying community shura in the heart of the Afghan conflict.
About the Author(s): MAJ Greg Merkl is a Special Forces Officer who currently serves as a mathematics instructor at West Point.
1. Commander's Guide to Money as a Weapon System, Handbook No. 09-27 (Fort Leavenworth: United States Army Center for Army Lessons Learned, 2009).
2. Martin J. Osborne, An Introduction to Game Theory (New York: Oxford University Press, 2004), 14–18.
3. Howard Raiffa, The Art and Science of Negotiation (Cambridge: Harvard University Press, 1982), 346–349. Raiffa describes a similar scenario when discussing Thomas Schelling's "N-Person Prisoner's Dilemma Game."
4. Jack Hirshleifer and Juan Carolos Martinez Coll, "What Strategies Can Support the Evolutionary Emergence of Cooperation?" Journal of Conflict Resolution vol. 32, no. 2 (June 1988): 367–398. Hirshleifer and Coll show that evolutionary behavior does not generally lead to cooperative behavior in the prisoner's dilemma. Instead, a threat of punishment can favor the emergence of dominant strategies.
5. Steven Kuhn, "Prisoner's Dilemma," in The Stanford Encyclopedia of Philosophy, Edward N. Zalta, ed. (Spring 2009): http://plato.stanford.edu/archives/spr2009/entries/prisoner-dilemma/. Kuhn looks at many games that emulate the normal form of the prisoner's dilemma, including the tragedy of the commons.
6. Robert Boyd and Peter J. Richerson, "Punishment Allows the Evolution of Cooperation (or Anything Else) in Sizable Groups," in Ethology and Sociobiology vol. 13 (1992): 171–195.
7. Note that this payoff structure is simply the previous payoff structure with a penalty of 20 assigned to any player who chooses to compete.
8. "Afghan yuppies get rich on ten years of war," Dawn Online, October 3, 2011: http://dawn.com/2011/10/03/ afghan-yuppies-get-rich-on-ten-years-of-war/
9. Karen DeYoun, "U.S. finds Afghan anti-corruption efforts ‘deeply troubling,'" The Washington Post, December 11, 2012: http:// articles.washingtonpost.com/2012-12-11/world/35768044_1_ kabul-bank-afghan-government-cash-counters
10. Kevin T. Carroll, "Afghan Corruption-The Greatest Obstacle to Victory in Operation Enduring Freedom," Georgetown Journal of International Law vol. 43, no. 3 (Spring 2012): 873–898. Carroll's work gives a detailed account of the problems of corruption and how United States' funding has encouraged it.
11. Gordon McCormick and Lindsay Fritz, "The Logic of Warlord Politics," Third World Quarterly vol. 30, no. 1 (2009): 81–112.