Pages

Thursday, January 06, 2022

Applied Game Theory and the Medium Term Reputation Gains from the Recent China-USA Trade War

 The new issue of the Economist includes an excellent Free Exchange column titled Lose-Lose Ordeal with the heading; "New research counts the costs of the Sino-American trade war".  The article does a great job citing the new research by several top economists including my old colleague Pablo Fajgelbaum.  

Empirical research helps to understand the short run dynamics brought about by the Trump tariffs.  But, what do we know about the medium term reputation effects?

Basic game theory teaches us that if you can build up a credible reputation for sometimes not pursuing your short term interest you can improve your bargaining power.  Richard Nixon told his advisers that he wanted to build up a reputation for being "crazy".  Here is a 2017 Slate piece.

During the Presidencies of Bill Clinton and Barack Obama, the U.S was predictable in its interactions with China.   Has President Trump's actions helped future U.S Presidents by inserting some doubt in the minds of China's leaders? Can they confidently predict what strategy the U.S will play in future economic strategic games?

I believe that the answer is no.  So, my question for the game theorists;  when an opponent inserts some uncertainty into her rival's head -- when does this help one party in the game?  The Economist magazine gives the Trump Administration no credit for this medium term "benefit".  Is the Economist magazine, and the empirical trade economists, right to ignore the reputational effects of unilaterally raising tariffs?  

Finally some nuance. I am not saying that the Trump Trade War was "good".  I am saying that it appears to me that The Economist and the academic economists are ignoring an interesting (and hard to quantify) medium term benefit of the Trump Administration's strategic play.    

There is a certain irony that future Democrat Administrations will gain from this policy uncertainty shock.

In our new age of Machine Learning and reduced form empirics and field experiments, what insights does applied game theory offer?  If your rival predicts that you will always comply with "their proposal", what happens next in the game?