Michael Woodford has a piece in today's FT warning against QE3 today in the FT. There are a few problems with it, however.
Woodford's arguments are:
1. QE and QE2 were supposed to work by raising the price level, which is theoretically unlikely. Agreed -- and the Fed specifically said that the LSAP was intended to be different from QE(*) precisely because it was not going to lead to a permanent expansion of the money supply. So an impact on the actual price level was never intended to be the mechanism through which QE helped the economy. Note that Woodford's argument here is exactly at odds with the primary criticism of QE2 that we were hearing a couple of months ago, when the inflationistas were breathlessly proclaiming that the US was entering a period of runaway inflation.
2. QE and QE2 were supposed to work by flattening the yield curve and reducing long-term interest rates... but while QE1 accomplished this, QE2 did not. Woodford does not explain why he thinks the first round of QE worked but the second round didn't. Furthermore, there are a number of empirical studies that provide consistent estimates that QE2 did in fact have a significant impact on long-term interest rates. (See here for a summary of some of them.)
3. QE2's main impact was through altering the market's expectations about future policy... but this is a bad way for the Fed to communicate with the markets. I agree that probably the most powerful single weapon left at the Fed's disposal right now is its ability to change market expectations. But now that the Fed has explicitly described its intended policy with respect to short-term interest rates (i.e. zero for the next two years), why discount the possibility that QE3 could change expectations regarding long-term interest rates?
In fact, QE2 did reduce long-term interest rates, and there's no reason to think that QE3 wouldn't do the same. How much of a boost this would provide to the economy is another matter, and if you want to argue against QE3 because you think that lower long-term interest rates won't do much for the economy, then we can probably have a reasonable discussion. But to say that it won't work because it won't work (which is the best summary I can come up with for Woodford's piece) does not seem a helpful way to advance the debate.
(*)Note that, as I've written before, "quantitative easing" is a bad name for the Fed's program of Large Scale Asset Purchases (LSAP), because the program is not really intended to increase the money supply in the way that, say, Japan's QE program did in the early 2000s. But I'll use the common shorthand for convenience.