Latest Idea to Fix the Banking System

Offer to buy $700 billion in preferred stock in a single auction (where banks quote dividend yields to bid on funds, and the bids are accepted starting with the highest until all the funds are used up); put a tax (temporary in principle but with no specific time limit) on bank profits; and regulate banks aggressively to make sure they don't make excessively risky loans.

The tax serves a triple function: first, since banks have an incentive (due to the moral hazard of stockholders with limited liability) to bid too low, the tax recoups the additional risk premium that taxpayers ought to receive (essentially an after-the-fact insurance premium); second, the tax encourages banks to participate in the auction and to bid higher, since they are going to be paying part of the price for the funds whether they get any or not; third, by effectively taking part of the payment as equity, it reduces the incentive for banks to take excessive risks.

The preferred stock will tend to give banks an incentive to take risks, since they need to make a high return on their assets before they get to keep any of the return. It's not clear whether encouraging banks to take risks is a good thing or a bad thing. On the one hand, we want to protect the banks' creditors (mostly the FDIC in the case of commercial banks). On the other hand, we want banks to lend, so as to keep the economy going and enable maximum private investment. To the extent that the FDIC -- effectively the taxpayer -- is the primary creditor, this is essentially a type of fiscal stimulus, although you only have to pay for the stimulus if the banks fail. Considering the likely weakness in the economy over the next few months, it is arguably a good idea to encourage banks to take risks.


UPDATE: More details. Banks should be allowed to make multiple bids, so a bank, for example, might be willing to sell $1 million worth of preferred stock with an 8% dividend and then another $1 million worth at a 7% dividend. Then for any bank that has more than one winning bid, the dividend yield would be the average of the bids, weighted by the size of the bids. The minimum bid would be the 30-year Treasury yield. Every bank should be willing to make some bid, since the minimum bid would be a very good deal for any bank. (Imagine selling a CD that yields the 30-year Treasury rate; you don't ever have to pay back the principal; and if you're a year or two late with the interest payments, it's just fine as long as you've got a good reason. If I were a bank, I sure would be interested in selling that CD.)

Another issue. One (perhaps inevitable) flaw in this plan is that it requires good banks (those which don't need additional capital or liquidity and therefore won't submit winning bids) to pay in part (via the tax) for the sins of their prodigal brethren. Of course, just about any kind of bailout is going to reward the bad banks relative to the good ones, unless (as in my earlier plan) the Treasury can force terms on the banks. To the extent that the good banks are "good" because of good lending and investment practices over the past 5 years, it's unfortunate that they get penalized. On the other hand, to the extent that good banks are "good" because they have been hoarding cash during a crisis (and therefore don't need additional cash), we actually should be penalizing the good banks. During good times, we hope that nobody expects to be bailed out, so they won't take risky actions. During a crisis, we hope that everybody expects to be bailed out, so they will be willing to lend. Given that we are more than a year into the crisis, it's arguable that, overall, penalizing the good banks is not such a bad thing.

Avantages of this plan over those being considered in Congress:

• It's simple -- much simpler than any of the plans currently being considered.
  
  
• It's quick. Since the initial disbursement requires only one simple auction of a well-defined instrument, it will get the cash to the banks much more quickly than a plan that would require auctions on many difficult-to-define asset classes.
  
  
• It's cheap. Taxpayers are likely to make more profit with less risk. Perhaps more important than the interest of the taxpayers is the reputation of the US Treasury and everyone else who borrows in dollars. With so much government borrowing already, every dollar that the Treasury can save is a larger cushion against the point where its (and the dollar's) reputation is in peril and it (along with all other dollar borrowers) has to pay uncomfortably high interest rates.
  
  
• It's effective. It's more effective than plans currently being considered, because it solves both the liquidity problem and the capital problem (and does so without deliberately overpaying for assets).
  
  
• It's Constitutional. It does not give dictatorial powers to a single cabinet officer.
  
  
• It requires less bureaucracy.
  
  
• There is less potential for favoritism. There won't be billions of dollars worth of business that Secretary Paulson can give to his former employer. (There may be hundreds of millions, though. This would be the biggest business deal in history by an order of magnitude -- and by several orders of magnitude if you exclude the AIG deal. It's not something that the Treasury department has the resources to do on its own.)
  

UPDATE2: To mitigate the problem with the tax, there should be a time delay before the tax becomes effective. That way, hopefully, you wait until the bad banks become profitable again, so a greater share of the tax is paid by those who are taking advantage of the benefits.


UPDATE3: The preferred stock should be callable, to facilitate the unwinding of the whole mess once banks become profitable again. Note that banks which have reserved conservatively (as GAAP requires) are more likely than not to make substantial profits as some of their assets turn out to be worth more than book value, and as those assets mature they will be taking in substantial amounts of cash with which to retire the preferred stock.