Earlier today, Greg Maniw posted a three-point defense of the Paulson bailout plan from someone to whom he referred only as "a smart friend". I want to address point 2 of Mankiw's friend's argument:
2. "Taxpayers will be better off if Treasury gets warrants."
This is essentially the assertion made in David Leonhart's column in the NY Times on Wednesday. And it again illustrates that we would all be better off if high schools taught the Modigliani-Miller theorem. MM implies that the price of the asset (again,assuming the auction gets it right) will adjust to offset the value of any warrants Treasury receives. In this case of a reverse auction, imagine that the price is set at $10. If Treasury instead demands a warrant for future gains of some sort, then the price will rise in the expected amount of the warrant -- say that's $2. Then the price Treasury pays for the asset will be $12. Some people might prefer to get $12 in cash and give up a warrant worth $2 in expected value. Fine, that's a choice to be made. But the assertion that somehow warrants are needed is simply wrong.
For a smart guy, Mankiw's friend is making a pretty dumb argument. Sure, if everyone has the same information, then an asset with a value of $10 will cost $12 if it's required that a $2 warrant comes with it. But that totally misses the point. Based on what I've read, it appears that if everyone understood what these assets were worth, there wouldn't be any need for a bailout: the bailout is necessary because people with capital are scared witless that anything they buy will just be crud. Folks like Mankiw are constantly reminding us (and they're often right) that there's no reason to think government has better information than private parties. So how will Hank Paulson or his agents know any better than folks risking their own money?
In more concrete terms, the right way to think about the equity issue is as follows. (I'll abstract from debt-deflation spiral issues, which are no doubt important but are beside the point for this discussion.) Suppose a bailee has two assets, each of which has face value of one unit. However, actual value and face value diverge, say because the underlying security may default. The first asset is actually worth $10, and the other is actually worth
$9 $8. The Treasury can't tell the difference, but the bailee can.
Now suppose the Treasury declares that it is willing to pay $10 for one asset with a one-unit face value, with no equity transfer required. The bailee will certainly prefer to sell Treasury the $8 asset, as it will profit by $2. The bailee now has $10 in liquid capital, and Treasury books a real, long-term loss of $2.
Now suppose that Treasury insists on getting warrants whose value is an increasing function of the loss Treasury books when it sells off the bailee's assets. Let's suppose for simplicity that the warrants' value have expected value equal to the difference between the price Treasury pays the bailee and the price for which Treasury sells the asset. Now if the bailee hands Treasury the $8 asset for $10, the bailee will expect to pay $2 later in warrants, so the bailee and Treasury each break even, though the bailee gets liquidity in the short run, which is the point. On the other hand, if the bailee hands over the $10 asset, then there will be no warrants issued later, since Treasury won't lose anything on the deal.
In this example, the use of warrants has (1) gotten liquidity to the bailee, (2) made the bailee indifferent between transferring the two types of assets, and (3) ensured that Treasury doesn't get stuck with an adverse selection-induced loss. The moral of the story is that the use of equity claims makes truth-telling incentive compatible.
This is an example of simple mechanism design, a topic that is well understood by many microeconomists, no doubt including many smart friends of Greg Mankiw's. Obviously the real world is much more complicated than the simple example here (which I emphasize I took from Mankiw's smart friend). But the idea that the MM-type argument Mankiw's smart friend makes is at all relevant to a world full of informational asymmetries strikes me as bankrupt, tough to credit, and more than a little deflating.