Intelligence Analysis: Behavioral and Social Scientific Foundations
Chapter 3 (pages 57-78) were used for the summary
What is Game Theory?
“Game theory is a body of reasoning, grounded in mathematics but readily understood intuitively as a reflection of how people may behave, particularly in situations that involve high stakes for them. It is part of a family of theories that assume people are rational, meaning that they do what they believe is in their best interest” (58). All games have shared and varied characteristics, involve a minimum of two players, and are solved by looking ahead by anticipating the opponents (rational) response. Chess or checkers is an example of a game theory model that compels the player to think about counterfactual circumstances, and not just about what actually happens. Ignoring counterfactual actions may result in misleading inferences about the process leading to the outcome and the content of the outcome. Utilizing game theory mitigates the risk by ensuring insight into what happened and why the alternative action was not taken. An example of game theory applied to a historical event is President Kennedy’s action to use a naval blockade in response to the Soviet Union’s introduction of long-range ballistic missiles into Cuba. President Kennedy knew the blockade would not remove missiles already stationed in Cuba, and other military actions would have likely resulted in the destruction of the missiles or compelling the Soviet Union to remove the weapons. However, any other actions like a tactical airstrike or invasion were inferior to the expected net gains from the chosen approach. Thus, those actions were removed from the equilibrium path because they would not achieve the expected results.
Categorizing Constraints on Foreign Policy Actions:
This paper uses game theory to approach five constraints related to national security problems: Uncertainty, risks, distribution of costs and benefit, coordination, and patience.
Uncertainty about an event is a concern in almost every situation, as information about rival actions, capabilities, intentions or resolve is often times difficult to find. Uncertainty also allows the opponent to bluff about their capabilities, like in a game of poker. Game theory approaches uncertainty in two ways. The first method is looking at a given situation with a degree of random changes that may affect player expectations. For example, a key political figure may die or a natural disaster may occur which will change the focus of a decision maker. Creating models that allow for these changes may provide a way to think about unanticipated or random events that may change the outcome of the situation. The second method game theory approaches uncertainty is by identifying what critical piece of information is missing. For example, not knowing your opponents preferences, capabilities or expectations. Game theory attaches probabilities to player types having nature in accordance with explicitly assumed probability distributions.
While uncertainty is not knowing whether a government entity will negotiate with terrorist, risk looks at the probability of alternative results. For example, making a bet a 6 sided die will come up with 6 on the first roll involves no uncertainty of the probability. However, the bet is risky as there is 1/6 chance of winning. Everyone responds to risks differently, and estimating a player’s risk aversion or risk acceptance in a foreign policy context is difficult, but crucial to solving the problem. Thus, assessing your opponent risk-proneness draws attention to how risks, weighted by the value or utility attached to alternative outcomes, shape expected pay-offs.
Distribution of Costs and Benefits:
Distributional conflicts arise over the relative costs and benefits associated with different outcomes of a game. For example, wars are fought for territory, wealth or regime change. Each of these factors involves distributional issues between rivals, so they can be assessed in a game-theoretic framework. When combined with distributional issues and uncertainty, situations arise when an enemy may bluff to steer action toward their desired outcome. A country threatening to mobilize its troops to a border with a troubling neighbor is an example of a player attempting to steer a situation in which forces the rival to sacrifice what they want in order to avoid costs. Uncertainty about costs and benefits not only provokes bluffs, but can also provide a means to reduce the odds of being taken in by a bluff. Consider the differences between bluffs that are costly, and bluffs that cost nothing. Talk is cheap unless it is backed by action. Therefore, higher self-imposed costs to a threat like mobilizing troops to a border indicates the action is more serious and not a mere bluff. In other words, when self-inflicted costs are small, the adversary is unlikely to take the rival’s threats seriously. Furthermore, distribution issues often reveal commitment problems. For example, the Taliban promised to not disrupt the 2009 Afghan elections but later reneged on the promise. What commitment shows is that if analysis treats promises as meaningful, even when carrying them out is contrary to their maker’s interests, is bound to lead to overly optimistic conclusions.
Coordination arises when players work together towards a common issue or goal, like Russia stationing troops in Syria. Coordination issues are widespread in international crises, as with the example above, Russia coordinating with Syria complicates US foreign policy and actions in the Middle East, particularly in the context when dealing with ISIS. Thus, outcomes and goals of a game change when coordination occurs between rival players.
Patience calibrates the value a given cost or benefit has tomorrow compared to the same cost or benefit today. The higher the patience a player has, the closer the future value is to the current value and vice versa. Thus, repeated strategic situations are more important than a single incident. It is important to understand how patient or impatient players are and what the sequence of gains and losses looks like, especially when situations are repeated as players will likely behave differently each time. For example, the prisoner’s dilemma will have different outcomes each time the scenario is employed. In the prisoner’s dilemma scenario, repetition can lead to cooperation if the players are patient, and if they value continuous modest benefits more than they value larger immediate gains followed by ongoing greatly reduced benefits. In contrast, an arms race has the opposite effect. Arms races are characterized by absorbing costs not to prevent defeat later. In this context, impatience makes players more reluctant to sacrifice today for tomorrow’s gains. What game theory models of patience suggest is to be careful to not leap to general conclusions from specific insights. Patience may not always lead to cooperation or conflict, but depends entirely on the situation. “An analyst can capitalize on the conditional predictions of model strategic interaction to provide insight into what might look like unique circumstances in any specific case” (70).
Prediction of Future Events:
The problem intelligence analyst face is most know what happened in the past but must figure out which historical patterns are relevant to the current problem, a problem with a resolution that is still unknown. A few quantitative models stand out for their success rate when analyzing national security problems. What applied game theory does is provide a useful alternative to conventional statistical analyses in that applied games have greater case-specific qualities. Game theory is also helpful in looking at ongoing situations and cases involving the prospect of discontinuity, and provides means of modeling how uncertainty alters the strategic interplay among decision makers, and provides a means for taking learning into account. Finally, game theory’s track record on international relations is phenomenal. Stanley Feder, former CIA intelligence analyst, stated one model of game theory was tested more than 1,200 times and produced accurate results 90% of time. Likewise, other literature reviews report the same reliability of game theory in academic papers (Ray and Russett, 1996).
The author notes several limitations of using game theory. Feder (2002) highlighted a cultural divide between humanistic and social science approaches to intelligence analysis which hinders the adoption of statistical or game theoretical methods. Humanistic modes provide insights into intelligence problems, coupled with social science, the two methods provide more insight on a particular situation than when used separately. However, humanistic modes of analysis face limitations of their own as well. While game theory forecasting can be used to evaluate political decisions, and can combine the benefits of detailed case assessment while exploiting the advantages of broad hypothesis testing, the methodology makes strong assumptions about information and people. Games require that some critical knowledge is known to all players, meaning that some information must be known to each player, who must know that each other player knows that same information, and so forth. Standard game models have not overcome the common knowledge conundrum, especially when dealing with assumptions about the probability that players hold a specific belief. Finally, the goals or outcomes in game theory are well defined, but in the real world, situations are a lot more complex, random and involve more steps.
The author does an excellent job of explaining game theory and applying the methodology to the intelligence field. I highly recommend everyone reads the entire chapter, as more in depth information and examples are included which I couldn’t summarize. It seems game theory greatly improves forecasting, but is extremely difficult to utilize. Another significant downside of game theory is the complexity and effort involved in employing the technique correctly, and the potential learning curve involved for the analysts or players.