Registration for the 2012 Allied Social Sciences Meetings has just opened up today. The ASSA meeting is the annual “winter meeting” in which hordes of economists descend on a rotating list of cities to spend a weekend shuffling papers around and stiffing cab drivers.
It was by sheer luck that last night I noticed that today would be the first day to register. And so this morning I was one of the first to login to the ASSA hotel registration system and reserve one of the better suites (we will be in the Fairmont) in one of the more central conference hotels where Northwestern Economics will conduct its job market interviews. (New PhD recruitment is one of the main, perhaps the main, activity at the ASSA meetings.) Had I been just a few hours later we would have been relegated to a remote hotel making it harder for interviewers and interviewees to get to and from the interviews.
It’s funny that a conference run by economists uses a qeueing/rationing system to allocate scarce hotel space. The system doesn’t even allow ex post exchanges between departments which would undo inefficient misallocations. If MIT gets stuck in the Embassy Suites and Podunk U is in the Hyatt Regency, then tough luck MIT (maybe Podunk can build a stronger theory group, ha ha ha.)
The problem is that ASSA negotiates discounted rates for the suites by reserving them in bulk. Obviously that is good for everyone. But the discounted rate is below market clearing and therefore there will be excess demand for the best hotels. It would seem that the resulting inefficiency is the price we have to pay for our monopsony power.
Indeed it would not work to have ASSA negotiate hotel space at discount rates and then turn around and use an efficient auction internally to allocate it. The reason is somewhat subtle. Here’s one way of seeing it. An efficient auction for a single suite is (essentially) a second-price auction. It works efficiently becuase when I know that I will have to pay the second highest bid then I will bid exactly my willingness to pay. Therefore the winner will be the bidder who values the suite the most. However, because ASSA bought the suite at a rate that is below market clearing, the second highest bid for a suite is going to be more than ASSA paid for it.
That means ASSA makes money. Sounds good right? No. In fact it is a problem precisely because we all benefit from ASSA making money (we get lower registration fees, lower journal subscription fees, etc.) You see I internalize the benefits of ASSA’s revenue which essentially means that I get back some of the price I pay when I win the auction. In other words I am not really paying the second highest bid, I am paying something less. And because of that I no longer want to bid my true willingness to pay, and the mechanism breaks down. In the jargon, the efficient auction is no longer incentive compatible.
But there is a solution. The basic mistake ASSA is making is to negotiate discounted suites. Why does it do that? Well, it has monopsony power and it has different hotels in the area compete with one another for the business. Since we are buying hotel space it seems natural to make hotel discounts the currency of that competition. But we saw the problem with that. Instead ASSA should ask for lump sum cash bids from the hotels. The highest bidding hotel gets the right to auction off their suites to ASSA members using an efficient auction. The hotel keeps all revenues from the auction.
That way I don’t internalize any of the revenues from the auction. The mechanism is incentive compatible again and therefore gives an efficient allocation. The hotels make some money. And the amount of money they can expect to earn is exactly how much they are willing to pay to ASSA in advance for that right. So in fact ASSA comes away with their monopsony rents without having to sacrifice efficiency.