2004 Economics Nobel

This year the award was given to Finn Kydland and Edward Prescott for their work on time-inconsistency and real business cycles. Here is Marginal Revolution's explanation of time-inconsistency.

While I'm not familiar with the relevant work, time inconsistency certainly makes sense as a basic result of game theory, which has the notion that a true optimal equilbrium strategy must be subgame-perfect. What that means is that your strategy now must assume that in later situations, you will also make the optimal response. This means that you cannot make threats which you will not carry out.

As an example, consider the single-play "divide-the-dollar" game, where one party offers a split of $100, and the second party can accept it, or reject it, in which case both get nothing. The second person might threaten "If you don't give me $50, I'll reject it!". But this is not a credible threat. Suppose an $80/$20 split is offered. At this point, the attempted threat has failed. The second person might as well just take the $20 (ignoring future reputational issues, which are not part of the game). The threat meant nothing, because the first person knew it would be irrational to carry it out.

So the optimal strategy may well be to say or threaten one thing now, but do something quite different later. This is a time-inconsistent strategy, and it often comes up in policy issues (ie consider monetary policy).

(Aside: this is why it can be a huge advantage to discover methods of committment. If the second person can somehow commit himself to rejecting anything but a 50/50 split, his threat is no longer empty. For example, he could deposit $50 with a trusted third-party, along with a public agreement to forfeit the money if he accepts a less than $50 split. Alternately, consider the winning strategy for the game of chicken, which is to unscrew your steering wheel and throw it out the window in full view of your opponent. Now he must swerve.)

The real business cycle theory demonstrates that without any of the standard causes of cyclical behavior (sticky prices, confusions of nominal vs. real prices, insufficient demand), a recession can be the optimal response of an economy to a natural shock. Both Alex and Tyler over at MR explain in more detail.

Share this

This assumes both that the

This assumes both that the second actor is acting rationally, and considers winning achieving an optimum solution. In actual practice, varying levels of irrationality enter decision-making, which make such simplistic analysis with overtures of complexity less than useful.

(Tit-for-Tat, for example, is so highly simplified as to be useful in understanding strategies in the IPD. This more complicated model requires far more assumption, making game theor much hazier in application.)