The Wall Street Journal has a useful piece up today about the historical analogy that looms large in the minds of administration economic officials: 1937, the year the Fed, FDR, and Congress prematurely tightened economic policy and sent the economy back into a deep recession after several years of recovery.
What I always find remarkable about these discussions is that proponents of early tightening essentially (sometimes explicitly) argue that the expected loss from too much inflation (that is, the loss weighted by the probability of it happening) is greater than the expected loss from a double-dip recession. So you get lines like this from Allan Meltzer, among the more intellectually serious of the inflation hawks:
[White House economist Christina] Romer is "sending the absolutely wrong message -- that we can't do anything to worry about inflation until the recovery is locked in because of concern for unemployment," says Allan H. Meltzer, a political economist at Carnegie Mellon University. "The reason economists and central bankers have two eyes is so they can do two things at once." ...
Mr. Meltzer says the risk lies not in pulling back too soon but dithering too long.
Let's posit, not unreasonably for a policymaker, that what we're trying to avoid in this case is worst case scenarios. Given that the 1970s and '80s (the worst inflation outbreak in recent memory, along with the recession that tamed it) was far less traumatic than the second half of the Great Depression (the worst example of a double-dip owing to premature tightening), you have to place a much bigger probability on inflation getting out of hand than on a double-dip recession to get those expected losses to favor early tightening. And given how low inflation is at the moment, and how weak the labor market is--to say nothing of how much more vigilant policymakers and the policymaking apparatus are about inflation these days--that seems like a pretty heroic assumption to me.