Scott Sumner has an excellent post on his concerns with IS-LM theory. I endorse most of it, but I want to point out where I disagree. He writes,

Michael Woodford and I agree on one thing; changes in the expected future path of monetary policy are much more important than changes in the current stance of some monetary instrument.

Sumner does a nice job of attacking the notion that the short-term interest rate is a single lever that can determine macroeconomic outcomes. He also attacks the notion that the money supply is a single lever to determine outcomes. But he seems to take the view that the expected path of monetary policy is the magic lever to determine outcomes. I personally do not believe that the typical micro-level decision is based on expectations for macro policy variables.

We should not assume that a policy lever necessarily exists that will fix macroeconomic problems. We cannot undo the housing crash. We can only undo a little of the financial crash, and I am not convinced that what we are attempting there is useful for those of us outside the financial sector. We cannot make labor market adjustments painless and instantaneous.

It would be really nice if we could wave a wand and animal spirits would appear in sectors of the economy where future growth is a realistic prospect. But why government should know better than individual investors where those sectors are eludes me. That makes me skeptical about fiscal policy. Monetary policy presumes a lower level of unrealistic centralized knowledge, but it presumes that we do have a wand that can stir up animal spirits.