The Greek Tragedy is a “thing.” And lately it has reemerged. The question at the heart of this post is how one should bargain when between a rock and hard place. This point was raised and discussed very well by Henry Farrell in this piece, which was responding to this op-ed in the NY Times by Yanis Varoufakis, the finance minister of Greece and, in earlier times at least, a game theorist. Varoufakis claims in his op-ed to essentially disown game theory in pursuit of bigger, and of course more noble, goals.
I am not actually interested in what Varoufakis’s true goals are here. Instead, I am going to attack the face validity of the claim that he is not “busily devising bluffs, stratagems and outside options” — because I am going to argue that he was (at least arguably) strategically using that very op-ed as a strategem to make it seem less likely that he is bluffing because the op-ed alters his outside options.
Varoufakis claims in his op-ed that “[t]he trouble with game theory, as I used to tell my students, is that it takes for granted the players’ motives.”
wait…give me a second…
…oh my goodness…
…I just vomited a little in my mouth.
Look, sure… the first 6 weeks of introductory game theory does this, just like physics first starts out neglecting air resistance. But, really, does game theory “take for granted” the players’ motives?
Answer: OH HELL NO, GAME THEORY DOES NOT TAKE PLAYERS’ MOTIVES FOR GRANTED.
The Ironic Emergence of Concerns About Reputation. The classic case of game theory not taking for granted players’ motives is something known as “the chain store paradox,” which poses the question of when and whether someone would be willing to incur losses to themselves so as to establish a reputation for toughness. In the interest of being succinct, that is exactly what Varoufakis is (putatively) attempting to establish in the op-ed. The fact that game theory is entirely and completely consonant with such behavior was established no later than 1982, when both David Kreps & Robert Wilson and Paul Milgrom and John Roberts independently established that game theory can and does predict that individuals will have an incentive to fake having “tough” or “principled” beliefs so that otherwise “irrational decisions” make sense to or be believed by an opponent. As the articles by Kreps, Milgrom, Roberts, & Wilson show, it is often the case that a “pure bottom line” player will have an incentive — in repeated negotiations/interactions — to act as if the player has a purpose other than “the bottom line”, regardless of whether this other thing (in Varoufakis’s case, it’s something called “doing the right thing”) is something that is deemed “irrational.” The reason for this, in intuitive terms, is reputation. In a sense, Kreps, et al. saved game theory 33 years before Varoufakis attempted to throw it under the bus by showing that, against some naive expectations, it is consistent with common sense. The Bully on the playground need not actually like hitting people, he might just be someone who really likes not being hit and accordingly “pays it forward” by beating a few people up so as to make others think that he or she likes fighting, thereby making others in the future less likely to challenge him or her.
Thomas Schelling is a genius, and properly credited by Farrell for offering erudite understanding of the dynamic that Farrell discusses. However, Farrell focuses exclusively on “appearing irrational” in his discussion. Moving beyond simple “Varoufakis versus the EU” narratives, Schelling, and others (including Bob Putnam and Andy Kydd), have commented on the importance of hiring mediators that are themselves biased/irrational. The basic idea is the same as that behind why you hire a hit man that you don’t know and can’t be recalled—hiring a crazy “agent” is a commitment device that makes your negotiating partner change his or her valuation of holding out against your demands.
Following this logic, presumably Varoufakis was installed as finance minister precisely because he is a very good game theorist. And, to boot, he was installed by a government that itself is worried not only about interactions with the EU, but also with the citizens of Greece in upcoming elections. The question, then, is whom do you hire to go bargain your way out of the absolute poop-storm of debt and austerity that surrounds you?
On one hand, you could install a technocrat that wants to make markets easy and handle things in a mechanistic and (economically/technically) efficient way. But, to be short about it, economic/technical efficiency is irrelevant to most voters. Such a technocrat would have a hard time sealing the technocratic deal struck with lenders by selling it to the voters in the form of the ruling coalition. Accordingly, such a technocrat would have little leverage at the bargaining table with the lenders in the first place.
On the other hand, you could hire a true believing, firebrand populist who will quickly and unabashedly pursue a “forgive, haircut, or default” strategy with the EU. That person would cause other problems: short term populist gains, but long term fiscal problems that would probably undermine the ruling coalition. And (unless that person is strategic, see below) the firebrand will also exert little bargaining power because his goals are too extreme and he or she would prefer to walk away.
Finally, you could install someone who is widely believed to be an excellent bargainer. You know, like an internationally recognized game theorist. Then suppose that this individual announces that he or she does not believe in being strategic, that he or she is just committed to getting the “right” outcome for the country. (From the op-ed: Varoufakis promises to “reveal the red lines beyond which logic and duty prevent us from going,” alleges that the “circumstances” dictate that he “must do what is right not as a strategy but simply because it is … right,” and even invokes Immanuel Kant!) Finally, suppose that the voters to some degree “believe” the game theorist insofar as they become more willing to support a somewhat technocratic deal, falling somewhere short of absolute forgiveness.
The arguments of Kreps, et al. imply that a smart game theorist should say the things that Varoufakis said in his op-ed. If voters respond as supposed above (i.e., believe the statements even a little bit), this increases the credibility with which he can negotiate with the lenders. Note that the voters’ beliefs that the game theorist actually has stopped believing in being strategic should be stronger if the game theorist takes a very public stand (say, you know, in an op-ed in a globally read newspaper) to that effect, and especially if, as I have pointed out, he aims his arrow at what at least once was his bread and butter.
Conclusion: Varoufakis Doth Protest Too Much. I actually applaud Varoufakis for the strategy I see him playing (not that he should care, of course). Nonetheless, I think that he went farther than he needed to go by parroting a frequently tossed-about and grossly inaccurate criticism of game theory. Of course this is ironic. I can only hope that at some point in the future, Varoufakis might fess up to the gambit. Regardless of whether he does or doesn’t “believe in” game theory, I am under no impression that he does not believe in being strategic. Especially not after reading his op-ed.
 This post is about game theory, and good game theorists would advise one to think about how not to wind up being between a rock and a hard place in the first, ahem, place.
 I am getting tired of the academic tradition of admitting that perhaps I might not be right. Of course, I might not be right. But, that said, this is one of those “every 18 months or so” arguments where I can say “well, if I’m not right, then I am right, because that’s the crazy kind of bull-hooey that emerges in strategic situations.” And, yes, “bull-hooey” is jarring technical jargon, which is why I put it in a footnote.
 These four giants of game theory are, because of their multiple contributions to this incredibly seminal 1982 issue of the Journal of Economic Theory, sometimes referred to as the Gang of Four, a reference that will hopefully still please at least a few people in the set of “game theory and awesome rock fans.” But seriously, each of these four have contributed huge ideas separately and in combination to game theory for 3+ decades, and for Varoufakis to pretend otherwise is absolutely offensive. I only say that because he has coauthored a textbook on game theory. He should know better (for example, see Secction 3.3.4 of the linked textbook). I also say this because I have the privilege of writing a blog that is at best occasionally clicked on by (some of) my family members. But, again, I LEARNED THIS STUFF AS AN UNDERGRADUATE.
 A great piece about how this works between chambers of Congress was written by Sean Gailmard and Tom Hammond.
 This is arguably an example of what some social scientists call “audience costs.”
 This is akin to the notion of “burning one’s boats,” in which one eliminates or reduces the attractiveness of backing down at some future point so as to make one’s demands more credible in the present.