Of Credit Crises, Adverse Selection, and Reverse Auctions

One concept is absolutely critical to understanding the origins of the mortgage crisis, the current seizing up of markets and credit crisis, and to certain problems that the government will have with its plan to buy and resell mortgage-based securities: asymmetric information and adverse selection.

The classic examples are for used-cars and insurance. Suppose I offer a fixed offer price at which I will buy all used cars put on the market. Due to adverse selection, I will only find sellers who think their cars are worth less than my offer -- the "lemons.". If gains from trade are not great enough, then the average value of the cars I buy could be less than my offer. To at least break even, I would have to lower my offer price, but that will mean that more of the better cars drop out of the market, leaving me with even worse of a pool...In the worst cases, markets can cease to exist under these conditions.

The way I put the example should make clear the connection to the original financing of mortgages, the problems in lending to banks who have lots of mortgages on their books, and especially, the problem that the government will have in buying mortgage backed securities.

If the government offers to buy all AAA rated mortgage backed securities, to take an example, who will participate in that market? If the government sets a fixed price, the adverse selection will be severe. If they do a "reverse auction," there will still be a problem: The institutions offering their securities at the lowest price will be the ones who win the auction, and their securities will be the worst of the lemons. The good ones will stay out -- classic lemons market (for those not familiar with the "lemons" terminology, a lemon refers to a low quality item like a bad used car).

Auctions do not necessarily perform well when information is very poor, and that is the case we are in.

All this said, the lesson from adverse selection is that there are tremendous gains to trade that are going unexploited. If some entity (the government, for eg.) can solve the lemons problem, mutually beneficial exchange can explode, and the entity solving the problem can take a spread from each trade and make a mint. This would be the best case scenario for the government's plan -- in which case the cost of the plan will not be $700 billion but will be negative.

The classic way to solve lemons/adverse selection problems is pooling. You cannot let individuals decide whether to enter the market or not. Pooling works great for insurance, as in employer-based pools for health insurance. I suspect it will be critical for the government's plan to have some kind of forced pooling.