Nestled in a capacious house on tree lined boulevard in Bonn, is the Hausdorff Research Institute of Mathematics (HIM). Yes, he of the distance, dimension and measure. Less well known, is that Felix was also a modestly successful playwright and critic. But, that is another tale.
Bonn, itself, was a capital of intrigue, making a star turn in a Smiley novel. But, the diplomats, parliamentarians have decamped to Berlin leaving behind a modest city on the banks of the Rhine dotted with cavernous buildings and, of course, the Hausdorff Institute.
What brings me to HIM is the special trimester program in Mechanism Design organized by Benny Moldovanu. In spite of what one may have read on other blogs, Benny is NOT mad. Excitable, yes, but not mad. So, Benny, if you read this, I hope it makes you feel better. Benny the gracious host, has brought us together for the summer to talk about mechanism design. To keep us on task he has organized a couple of workshops, the first of these I attended was on Dynamic Mechanism Design. Three intense days of stimulating talks.
Alessandro P. described his joint work with Toikka and Segal. Essentially, they work out the conditions under which the techniques for static mechanism design with one dimensional types can be pushed through to the dynamic case. In some ways it reminds one of the anecdote of the mathematician asked to boil water (Felix is given an empty pan and asked to boil water. Felix fills pan, places on stove, lights it and voila boiled water. Next Felix is given a pan filled with water and asked to boil it. Felix pours the water out and now notes that he has reduced the problem to a prior problem for which the solution is known). In particular, the paper raises the question of whether one can avoid having to rely on agents posting a bond. This makes life easy because it allows one to enforce the IR constraint in period 1 only. Drop it, and one must have IR period by period.
Johannes H., gave an amusing presentation of his joint work with Larry S. in which anything one thinks is obviously true is either wrong or does not admit an obvious proof. The set up is ridiculously simple. A monopoly seller with one unit of supply, many periods, who must sell the unit to agents whose valuations are private information. The catch is that the seller cannot commit to a mechanism. Surely, the best the seller can do is a Dutch descending is the immediate reaction. Surely, but…………….
Peter C. was there as well. Apparently, the clock auction can now cure cancer. I know, wicked of me. Peter’s fault. He brings out the green monster in me. Peter has everything I don’t:a mind like Aristotle and a body of mortal sin (you should have seen his pecs under a sweat drenched t-shirt at the top of the Drachenfels). Peter was describing his auction for liquidity needs, which provoked a deal of discussion.
We were also entertained to an abstruse discussion of genericity in the context of Cremer, Mclean and Neeman inspired by a paper of a Alia Giytaluna and Martin Hellwig. Stphen M. the discussant, made a valiant attempt to prevent some of us from reaching for our lugers at the mention of the word genericity. I was not convinced, so this matter goes into the folder along with the cosubstantiation debate (is the holy trinity 3 in 1 or 3 of 1?).
David R. offered a charming interpretation of the negative cycle-cyclic monotonicity condition in terms of undetectable deviations. His advisor, Ostroy, would be proud, since the job of an economist (for those of you who did not get the memo) is: 1) Write down a primal. 2) Take it’s dual. 3) Interpret Jusso V. says he asked Bengt H. if this was the case. Bengt H. with a disdainful glance asked what else he thought one was supposed to do. Long live Farkas!
Philippe J. ended the conference with an uncharacteristic, for him, talk on matching, enlivened by many denigrating asides about Toulouse.
My list of speakers is not exhaustive, but I certainly am.