The Predictioneer’s Game: Using the Logic of Brazen Self-Interest to See and Shape the Future

by Bruce Bueno De Mesquita

  There is plenty about such a deal that is distasteful. Bankrolling such a horrible human being is no happy task. It would be ever so much more satisfying if we could just persuade him to do the right thing, but then Kim Jong Il wouldn’t be Kim Jong Il if that were feasible. Exactly because he is so horrible, it is important to figure out how to keep him from unleashing a nuclear war in a fit of pique or fear or resignation that he is through anyway and so has nothing to lose. Remember, my task was to find what would work. The desirability of making it work is what we elect leaders to decide.

  I want to emphasize here that I have said “a verifiable means of ensuring that its nuclear weapons program stopped.” I have not suggested that North Korea’s nuclear weapons program and enrichment facilities be eliminated (positions 80 and above on the issue scale in figure 4.1) or dismantled. This is of fundamental strategic importance in making an agreement credible from both sides—so let me explain what I had in mind.

  In the event that the United States or the other participants in negotiations with North Korea insisted on dismantling that country’s nuclear capabilities, I believe agreement would have been impossible or would prove short-lived, leading to inevitable cheating. Dismemberment of North Korea’s nuclear capabilities as an approach overlooks the essence of Kim Jong Il’s interests. It fails to put ourselves, as any good strategic thinker must do, in his shoes, looking at the world from his (perhaps elevated) perspective.

  If he dismantled his nuclear capabilities, he would no longer have a credible threat of restarting his weapons program quickly in the event the international community—especially the United States—reneged on its promises. In such an environment we can be confident that the international community would renege, and, anticipating that, he would never allow his weapons program to be dismantled. After all, once his nuclear threat was completely dismantled, hardly anyone would have any remaining reason to follow through on payments to him, payments that help him keep domestic as well as foreign rivals under control. The international community would have even weaker reasons to leave him in power. They would rather replace him with someone more amenable to their wishes once his threat to use nuclear weapons to defend his regime was no longer meaningful. Remember, game theory takes a dim view of human nature, and that includes our nature as well as his. However high-minded we think we are, we would have scant incentive to continue to pay Kim Jong Il to behave well once his ability to behave really badly was eliminated.

  Thus, as long as his nuclear program is stopped, disabled, placed in mothballs, with inspectors on site at all times, and not dismantled, he has the ability to restart it if the United States or others renege on payments or security guarantees. Conversely, as long as the United States and others do not renege on payments and security guarantees, he has no incentive to throw out the inspectors and restart his nuclear program. It will have achieved its purpose of giving him a life preserver. His likelihood of extracting a higher price in the future by throwing out the inspectors must be balanced against the realization that if he proves untrustworthy there are alternative solutions to the problems he represents—alternatives he wants to avoid. These alternatives, of course, would need to be and could be acted on (and here is where U.S. interests in acting and South Korean interests in avoiding the possible consequences of such action differ enormously) before he recommissioned his nuclear capability.

  So we can see that money and security guarantees for disabling, but not dismantling, Kim Jong Il’s nuclear program create a self-enforcing mechanism. Neither side would have an incentive to renege on its part of the bargain. It is a deal that reinforces each side’s interests. It is the deal that was struck and that, with minor tweaking and continuous jockeying for position, is likely to succeed.

  So that is how simple it is to negotiate a nuclear arms agreement! Just find experts, collect the necessary information, use the computer to discover the impediments to the desired outcome, and work out how to neutralize the impact of those impediments by finding actions that serve the interests of the rival parties. With the help of the computer, we were able to see that neither George Bush nor Kim Jong Il was going to budge on his own. Since they were the two decision makers with the authority to say no to any deal, we had to find a way to overcome their intransigence. To do so, we used the computer model to test the likely effects of alternative degrees of American concessions to evaluate their likely influence on Kim Jong Il’s approach to the negotiations. We found that security guarantees, especially a mix of assurances from the United States (not to attack) and from China (to defend North Korea if necessary), coupled with significant economic assistance (approaching $1 billion or so per year) to North Korea, would induce Kim to mothball his nuclear capability and allow continuous on-site inspections and securing of his nuclear facilities. Thus, we saw that we could move him into the range of 60 to 65 on the issue scale, and that George Bush could accept this compromise as well, and—poof!—using the me-first principles of game theory we examined in Chapters 2 and 3, we saw the path to a settlement that could indeed be both predicted and engineered.

  We’ve seen now that it’s possible to employ this science in nudging North Korea to alternative strategies. But how does this tool really work in the rest of the infinite world of human interaction and conflict? How must we change our base thinking to work out any variety of predictions and, especially, to change the future?

  It is indeed my claim (and part of my livelihood) that outcomes of very big problems with many, many players can be systematically predicted and engineered, but, as you may have noticed, it all starts with asking the right questions. The next chapter will examine what those questions might look like across a wide variety of issues from the worlds of business and foreign affairs.

  5

  NAPKINS FOR PEACE:

  DEFINING THE QUESTION

  EVERY AFTERNOON AT 3:30 the Hoover Institution’s senior fellows get together in the Commons Room for coffee or tea and some of the best cookies on the West Coast. In the summer of 1987, as on many other occasions, I took a break from my research to enjoy a cup of coffee, a large chocolate chip cookie, and a bit of collegial camaraderie. On this particular July day, a distinguished Israeli sociologist, Shmuel Eisenstadt, was visiting. He asked me what I was working on, the standard conversation opener in the Commons Room, and I explained that I was trying to improve my original forecasting model which was then several years old.

  Shmuel asked, “So, you can predict how to make peace in the Middle East?” This being a tall order, I responded cautiously that perhaps I could predict what steps might be taken over the next few years to advance the prospects of peace. I emphasized that this required data, not a crystal ball. He asked what data I needed and then, pen in hand, wrote on my coffee-stained napkin and his napkin as well, listing stakeholders, their potential influence, salience, and positions on a scale measuring possible concessions in the context of a multination peace conference. (The Soviet Union was calling for such a conference in 1987.)

  Eisenstadt’s question to me is pretty close in form to the questions I’m asked to examine—whether by companies pursuing mergers, the Defense Department trying to evaluate a terrorist threat, law firms trying to sort out suits, or the CIA trying to understand Iran’s nuclear ambitions—and almost always, the question asked is not actually the question for which an answer really is wanted or needed.

  WHAT’S THE PROBLEM?

  The big question—how to make peace in the Middle East—is best answered by working out what the many smaller questions are that, taken together, add up to a solution. Framing the problem is usually the hardest part of the prediction and engineering process. I never cease to be surprised that even when billions of dollars or thousands of lives are at stake, decision makers rarely work through what it is they actually need to know. They are usually astonished—fortunately, pleasantly so—to realize that they have not thought systematically enough about their own problem to know what it is they need to know or do
.

  For me, answering a question like “How can we make peace in the Middle East?” involves breaking this question down to specific issues, to specific choices that must be made. Therein lies the key. Questions need to be about actual choices confronting decision makers rather than about abstract ideas like winning or getting ahead. If you want to win, you need to know what it is that signifies you won, you lost, or you did acceptably well, and by how much. Only after we have drawn out exactly what subjects or questions need precise decisions can we start putting the analyses of all of the issues together so that we might successfully identify the underlying stumbling blocks and find ways around them.

  For some problems, framing the problem is relatively easy. Much litigation, for instance, revolves around a settlement price. What a defendant really wants to know is “How much do I have to pay to resolve this lawsuit?” To answer that question, it’s just necessary to know what demands players currently say they support, how salient the price is for each stakeholder, and how much clout each could bring to bear.

  Of course, even lawsuits are sometimes more complicated than just working out a settlement price. Discussing a problem with a client, I might learn that they also need to know whether all charges can be resolved with one global settlement price or whether the price paid for one part of a lawsuit will influence the price that has to be paid for another part. Once in a while, it gets still more complicated. There might be multiple defendants, so it’s important to work out not only how much has to be paid to get the plaintiffs to settle, but how the cost will be distributed across the defendants, their insurers, and other involved parties. This sometimes leads to questions about whether it makes sense to drag in other parties as co-defendants or not. Adding co-defendants involves trade-offs. On the one hand, there will be more potential payers with more defendants. That can be a good thing, of course. On the other hand, more defendants means having to coordinate a strategy with many more businesspeople, their lawyers, and their insurers. That can get confusing and expensive. As you can see, a simple question like “What will it take to settle this lawsuit?” can quickly grow into a complicated, interlaced set of issues in which resolving one question changes how others can be resolved. That means we not only have to figure out what the actual questions are that must be decided, but also the sequence of agreements that leads to the optimal outcome.

  Business mergers are still more complex most of the time. Unlike litigation, mergers rarely hinge just on the price to be paid for putting companies together. Lots of other issues enter into a successful merger. But mergers can also be simpler than litigation in that the range of likely merger issues is usually well defined and tends to be pretty much the same across all mergers and acquisitions. The price to be paid (or to be obtained) is always a question, of course, but merger discussions rarely get anywhere if there is a substantial disagreement about the worth of a property. That means that the price rarely is the reason an attempted merger or acquisition fails. Remember the example of a merger between a French and a German bank in an earlier chapter. That deal hinged on whether the German executives had to move; they were willing to take a lower price for their bank if they could stay put in Heidelberg.

  Several years ago I worked with an international team of analysts on a major effort to create a unified defense industry across Europe, one that could compete with the American defense industry. We looked at virtually every possible combination of defense firms in Britain, France, and Germany, as well as several possible combinations that would have included defense firms in Italy, Spain, or Sweden. The question put to us was “Can we do these mergers?” To answer that question, my team needed to examine approximately seven different issues for each possible combination of firms for a total of more than seventy individual issues, each representing a decision that could make or break a multibillion-dollar deal. Some of the make-or-break questions, the issues on which agreement had to be reached, included (1) the price, (2) allocation of management control between merged units, (3) the scope of businesses to be included or excluded from the merged entity, (4) employment guarantees for workers in various units across national boundaries, (5) government’s role in regulating or sharing in ownership and management of the newly created firm, (6) the timing of the transition to combined working units and teams with shared technology, and (7) where the senior managers would be expected to live.

  Merger efforts are more likely to fail because of the “lesser” issues than because of a price disagreement. Yet few executives seem to recognize this when they initiate the process. As a result, they spend millions of dollars on getting good financial advice to sort out the right price to offer, only to have many prospective deals fall through because too little effort has been invested in working out other issues. Sometimes these “other issues” can seem so absurdly small that they are ignored, only to end up being, in hindsight, the deal breakers.

  A few years ago I worked on a failed merger in the pharmaceutical industry. The prospective deal was expected to produce great efficiency gains that could have dramatically benefited the pharmaceutical market and prescription drug consumers. Almost all of the executives in the two firms involved were most enthusiastic about the opportunity, with “almost all” being the key phrase. The killer issue that did the deal in involved allocating management control between the CEOs in the two firms. That, of course, is no small question. The absurdity came in how the deal was killed.

  The two chief executives hated each other, and, as is the case with many big companies in Europe, there were bitter family issues lurking in the background as well. So deep was the animosity between the two CEOs that everyone agreed a way had to be found for them to work together before going forward. A dinner was arranged for the two of them, and senior aides, who should have smelled a rat, also attended. It took a huge effort just to get this pair to sit at the same table.

  The host CEO finally agreed to have the dinner at his home and then, without my or any of his own aides’ realizing it, he planned a menu of course after course of the other chief executive’s most disliked foods. Amazing as it may seem, this prospective multibillion-dollar deal fell apart over a menu, or, more accurately, over a deeply personal conflict that made all other efforts and points moot.

  Working out what the right issues are takes patience, good listening skills, and the ability to steer a conversation toward what will really drive results rather than the inchoate musings of the decision makers. Fortunately, the dinner menu rarely plays such a prominent part in negotiations. For me, an “issue” is any specific question for which different individuals, organized groups, or informal interested parties have different preferences regarding the outcome, and for which it is true that an overall agreement cannot be reached unless at least a key set of players come to agreement on the question. It helps to know whether there is a status quo, and if there is, to make sure that the issue is not constructed to be biased against the existing situation.

  Recently I taught a seminar in which my students were asked to pick a world crisis that interested them and to model a way to solve the problem. One group of students decided to examine carbon dioxide emissions (a topic I return to in the last chapter). They defined an issue in which one end of the scale reflected current CO2 emissions and the other end the stiffest reduction advocated by any environmental group. Everything in between represented possible agreements among the players. Do you see the problem here? I asked if they thought there were no energy companies or others who felt that CO2 emissions should be less regulated than was the case at the time. Of course there were interested parties who wanted to have more freedom to produce CO2! So the way the students had designed their issue was biased. The only answer that could come out of their exercise was to increase regulatory controls or, with low odds, keep them the same. Their scale allowed no possibility of anyone wanting to produce more carbon dioxide. They had unwittingly created a biased issue, one that made them feel good but would almost certainly lead to a wrong a
nswer. Such fundamental flaws in the design of questions ensure that the answers to them will be useless.

  YOU CAN’T ALWAYS GET WHAT YOU WANT, BUT IF YOU TRY …

  Once an issue is properly framed, we have to think about how to capture the thought process that people go through in working out decisions. Without doing that, without climbing into the heads of your rivals, you’re not likely to get what you want. You’re not even likely to know how to try to get what you need.

  The game structure I use looks at choices as sometimes involving cooperation, sometimes competition, and sometimes coercion. The most complicated part is to try to emulate how people think about changes in their situation as well as what others say and do. Players are always interested in altering the lay of the land in their favor. They want to surround their desired outcome with tall mountains of power that are hard to overcome. They want to tear down mountains of opposition, leveling the power terrain around positions they want to defeat. At every step along the way, everybody has to work out who will help them and who will get in the way. They have to calculate the risks and rewards, costs and benefits of actively trying to change other people’s choices or lying low, trying to stay out of the line of fire. The math can get complicated,1 but let’s look at some examples of the process at work that we should be able to follow pretty easily.

  The table below shows the small data set that resulted from the cookie-and-coffee conversation I had with Shmuel Eisenstadt in 1987, augmented in 1989 by a discussion with Harold Saunders, who by then was the former deputy assistant secretary of state for the Near East and North Africa. The continuum of possible outcomes on the settlement issue ranged from the establishment of a fully independent, secular Palestinian state at one extreme to the annexation by Israel of the West Bank and Gaza at the other extreme. Position 30 on the scale was defined to represent territorial concessions granted by Israel to the Palestinians without establishing an autonomous state but establishing instead a Palestinian political entity federated with Jordan. The 1987 status quo was located at the position equivalent to 85 on this scale (under the “Negotiated Settlement Options” column). There was no semi-autonomous Palestinian territory or government at that time.