"...the U.S. government employed experts in game theory to analyze strategies of nuclear deterrence. Men with Ph.D.'s in economics, like Daniel Ellsberg, wrote background papers with titles like 'The Theory and Practice of Blackmail.' The intellectual quality of these analyses was impressive, but their main conclusion was simple: Deterrence requires a credible commitment to punish bad behavior and reward good behavior. I know, it sounds obvious. Yet the Bush administration's Korea policy has systematically violated that simple principle."

It disturbs me to hear anyone nostalgize the dangerous era of institutionalized madness when Dr. Strangelove-like academics provided mathematical rationales for the cold war policy of Mutual Assured Destruction. But fear not -- game theory is never again mentioned in Krugman's column. Its role in the introduction is nothing more than to create the impression that any subsequent arguments made by Krugman are to be understood as bearing the intellectual imprimatur of game theory. It's classic Krugman -- the intellectual bully.

And it's also classic Krugman -- the liar.

It's nothing better than a lie to oversimplify game theory as applied to global politics as "...simple: Deterrence requires a credible commitment to punish bad behavior and reward good behavior." Just read the abstract of the 1968 Rand Corporation paper by Daniel Ellsberg that Krugman referred to. The first third would seem to support Krugman's simplification, but read the whole thing.

"Whether it is called blackmail or deterrence, the art of influencing another's choice among alternatives by the use of threats is coercion. To provide a framework for representing and comparing alternatives, a game is developed, employing a payoff matrix, in which the victim has two choices, resist or comply, and the threatener has two, accept or punish. As a rule, a threat has a certain built-in implausibility, that of being costly--or irrational--to carry out. The threatener's problem is to make his threat sufficiently plausible to the victim. He may do so by means of four main techniques: (1) by binding himself irrevocably; (2) by putting up forfeits; (3) by making the victim unsure of what would be rational; and (4) by appearing to be irrational--or, as with Hitler, by [being] irrational. In the last analysis, however, since the estimates of payoffs-- or risks--are subjective variables, the answer to successful blackmail is not within the scope of logic: it is an art."

Nothing -- not one word -- in Krugman's column even begins to hint that Bush is not following Ellsberg's true, complex formulation. The best Krugman can do here is criticize the Bush administration for not using diplomacy -- it "...broke off negotiations as soon as it came into office" -- and then a few sentences later criticize it for not using force: "...even now the Bush administration hasn't done what its predecessor did in 1994: send troops to the region and prepare for a military confrontation."

What's the real problem with Bush's approach, then? Is it that he's being inconsistent? Maybe he's just "appearing to be irrational." Or is the problem Bush's failure to execute -- at the same time -- both of former President Clinton's contradictory approaches to Korea -- "by [being] irrational"?

Posted by Donald L. Luskin at 2:54 AM | link  

Tuesday, December 31, 2002

Posted by Donald L. Luskin at 2:27 AM | link