How Much Do We Know about the Phenomena That Economics Purports to Explain?

As I've pointed out elsewhere in this blog, Judge Posner argues in A Failure of Capitalism that depressions are poorly understood, and that where hard data is in short supply, preconceptions will fill the void.

That argument came to mind today when I read Judge Posner's latest post on his "Failure of Capitalism" blog. He is describing the causes of possible "aftershocks" of a depression:

The government has created a great deal of money, and borrowed a great deal of money, to finance the bailouts and the stimulus package and increase the amount of money in circulation (to help push down interest rates). If when demand rises the banks lend their $800-plus billion in excess reserves, the ratio of money in circulation to the output of goods and services is likely to rise--and this will mean inflation. The ratio will rise further if the government decides to finance some of the huge additional debt that it is incurring as a result of its anti-depression expenditures by increasing the money supply, that is, by inflation, which is a form of taxation--taxation of cash balances. A low rate of inflation is manageable and does little economic harm, but a high rate is very harmful, and can be broken usually only at the cost of a sharp recession (consequent upon a sharp rise in interest rates in order to reduce the amount of lending and hence the amount of money in circulation). And the recession might (as in 1937) disrupt a recovery from the depression. These costs have to [be] balanced against the benefits of the anti-depression programs; unfortunately only guesses are possible.

The last point -- "only guesses are possible" -- leads one to wonder how we should choose between the "guesses" of Keynesian economists and the "guesses" of "conservative" economists. Both sides make plausible arguments that rely partly on hard research but also partly (largely?) on naked reasoning. What would be worse -- a deflationary spiral (assuming the absence of bailouts and stimulus spending would have resulted in one), or post-depression inflation, or a double-dip depression? Not to mention permanent growth in the scope of government. Whose bogey is more likely to be real? Whose would be worse if real?

UPDATE: After posting this, I noticed that economist Arnold Kling has a post today touching on what appears to be the same uncertainty. He passes along the Three Iron Laws of Social Science: (1) "Sometimes it's this way and sometimes it's that way." (2) "The data are insufficient." And (3) "The methodology is flawed."