One criticism, made Paul Krugman and many others, is that it simply wasn't big enough - in retrospect, it looks like a field goal when we really needed a touchdown.
A natural villain is Larry Summers, in this case because he prevented Christina Romer's case for a bigger program from getting to the president (this doesn't look so funny now). New York Magazine's Krugman profile has Summers' response:
"[T]here is some element of [Krugman] that is like the guy in the bleachers who always demands the fake kick, the triple-reverse, the long bomb, or the big trade."
Summers concedes that a bigger stimulus would have been the optimal policy in 2009. “The Obama administration asked for less than all that it recognized pure macroeconomic analysis would have called for, and it only got 75 cents on the dollar. But political constraints and practical problems with moving spending quickly constrained us. The president’s political advisers felt, and history bears them out on this since the bill only passed by a whisker, that asking for even more would have put rapid passage at risk.”Ezra Klein suggests we should be asking a different question:
[T]he interesting counterfactual is not “what would have happened if the stimulus had been a bit bigger” but “what would have happened if Barack Obama had been inaugurated a couple of months later?” By June, unemployment was over 9 percent, and the full scope of the emergency was a lot clearer. If that had been the context behind the initial stimulus, I think it’s plausible to think it could’ve turned out very differently.His post illustrates one of the problems with discretionary economic policy - the "recognition lag" - that the state of the economy only becomes clear in retrospect, after the data comes in. This is particularly difficult because, by the time sufficient information to identify trends and turning points is available, the economy may have changed directions again. In most cases, that's a good argument against "fine tuning" macro policies. But I think it was plenty clear in February 2009 that there was serious trouble - payroll employment had declined by over 400,000 in each month from September 2008 through January 2009, and the shock of the fall 2008 financial panic was still fresh in the collective consciousness.
What was unclear was the shape the eventual recovery would take. Some of us hoped that the economy would bounce back quickly, as it had from previous severe recessions (e.g., in 1982, unemployment peaked at 10.8%, but the recovery was brisk; real GDP grew 4.5% in 1983 and 7.2% in 1984). However, the 1990-91 and 2001 recessions, which were much milder, had been followed by sluggish "jobless recoveries." Moreover, as Reinhardt and Rogoff showed, the typical historical pattern is that recoveries in the wake of financial crises are very slow.
Therefore, looking back, in my capacity as another "guy in the bleachers," the main flaw I see is that the stimulus should have been state contingent. That is, the aid to states and many other spending provisions, as well as tax cuts, could have been designed to stay in place as long as they were needed. The act could have contained a trigger to phase out after some recovery benchmark had been achieved, e.g., after the unemployment rate has been below 7% for six months, the stimulus steps down by 50%, with the rest coming off after 6.5% or less unemployment is maintained for a period. Some parts of it could have been tied to state, rather than national, conditions. That might have spared us the ugly spectacle of severe cuts in state services, even as unemployment remains at an appallingly high level. Moreover, knowing that the fiscal support would be in place as long as needed might have served to create more confidence in the recovery, leading to a stronger improvement in private activity. (The natural counter-argument is that such an open-ended fiscal commitment would undermine confidence in the government's ability to handle its debt burden, but I don't think it would have been a problem).
I don't recall the idea of a state-contingent stimulus being raised anywhere at the time, and I have no idea whether it would have been politically feasible or not. Arguably, its just an amplification of the "automatic stabilizers" already built in to the system. Hopefully, there is no next time, but when it comes, perhaps we should give this aspect of the design of policy some further thought.
The Economist's Ryan Avent has another question:
[W]hat if Congress had failed to pass a stimulus at all? Would the Fed have acted sooner or more aggressively or both, and how might recovery have gone differently?Which reminds me of the other thing the administration should have done differently (and this one is harder to explain) - they have allowed several seats on the Federal Reserve Board to remain unfilled for long periods. I agree with Avent that "QE2" appears to have worked. A different Board might have done more of it, sooner, and for longer, and that would have been better. On this point, Brad DeLong is yelling from the bleachers.