This produced a predictable result: It spooked holders of dollars around the globe, and by Thursday credit spreads had widened to record levels. Over the weekend pressure on Congress was increased to hand over the $700 billion Mr. Paulson said he needs to execute his plan.

The main problem with buying distressed and hard-to-value assets from a bank is that, if the bank is to attract new capital, it is necessary for the government to overpay. And while the value of those assets may eventually move higher, the taxpayer is exposed to great risk, risk that really belongs to the bank and its shareholders. Such a huge federal expenditure also raises the risk of inflation.

A further problem is that the Treasury is itself engaging in hedge-fund speculation. This expands the role of the state in the economy at a time when downsizing that role is more important than ever.

One alternative to the Paulson plan would be to provide secured loans to troubled institutions as a way to allow them to recapitalize. The collateral against the loans would be bank assets (presumably impaired assets) but the transaction would be similar to a "repurchase agreement." In this transaction, otherwise known as a "repo," the borrower is required to repurchase the securities, with interest, in the future in order to retire the loan.

Chile used such an instrument to recover from its 1983 banking crisis. It is true that the government intervened directly in two banks, wiping out shareholders, removing management and nationalizing the firms. Those banks were later re-privatized in a sale that gave tax incentives to encourage Chileans to participate in the offering.

The many other banks that were in trouble were handled differently. For those, the government provided loans that were secured by bank assets, with an agreement that the banks would later repurchase those assets.

These "repos" had conditions attached, including a provision that the shareholders could not take profits out of the company until the loan was repaid. This meant that shareholders were asked to give something in return for getting rescued by taxpayers; and it gave the bank a strong incentive to get back on its feet and return the money.

Another advantage of this model over the Paulson plan is that although the Chilean government took the bank assets as collateral against the loan, it did not adopt responsibility for managing the assets. That role stayed with the bank.

It was important that the government was a subordinated creditor; otherwise it would have been difficult to bring new capital into the bank. But if the U.S. were to use secured loans, it might not be necessary to make the loans at subsidized rates, as Chile did. Chile was in a depression. In the U.S. case, a penalty rate might be preferred in order to ensure that the banks will use the facility only as a last resort.