There's been a lot of speculation over reports that the administration held the stress test results while the banks argued for more lenient treatment. But the Washington Post is the first paper I've seen with specifics on those conversations. Citibank, for instance, "successfully pushed to lower the amount of common equity it needs to raise to $5.5 billion by applying $52.5 billion from capital it has not yet reworked. It also was able to get a credit for the sale of a unit that has not been completed."

Those aren't particularly crazy arguments. But they're unbalanced arguments. There are basically three poles in this debate: Independent analysts, who tend to think the situation is darker than the government does, but who arguably have some tendency to err on the side of gloom because they don't bear the downside risk of their solutions. The banks, who argue that the situation is much better than the government projects, but who have a strong tendency to err on the side of optimism because they want to free themselves from government regulation and oversight and present themselves as healthy to investors. And the government, who have some incentives to err on the side of caution (they don't want to be discredited or see their policies fail) and some incentives to err on the side of optimism (they want to avoid nationalization, they want to convince investors that the ground is firm).

But the stress tests are the product of the government negotiating with the banks. There was no hold-up while Dean Baker and Paul Krugman prosecuted their case. Where the government's incentives would naturally put it in a good position to adjudicate between the banks and the outside analysts, they're instead acting as the counterweight to the banks. But their incentive structure isn't well-suited to that role.