Stimulus

With the inauguration just over a week away, the talk of Obama's "stimulus" plan is reaching a fever pitch.  Some of the coverage is funny, some mean, some thoughtful, some just plain stupid (ever heard of the broken window fallacy?).

What ever it is, I can't get enough of the debate.
While some might be blinded by ideology, for those of us who can see clearly, the debate centers around an empirical question regarding the multiplier.  One's view of the size of the multiplier on tax cuts versus the multiplier on government spending drives one's policy choice.

What I find interesting is how little goes into one's evaluation of the credibility of some of the multipliers being reported.  There are two major problems I have in believing that the multiplier on government purchases is higher than the multiplier on tax cuts.  The first is a standard public choice argument- what kind of rent seeking is a $1 trillion dollar spending package going to "stimulate"?  Second, there is a "knowledge problem".

What strikes me about the second argument is how conveniently economists forget it.  Many ignore public choice issues all the time.  But even most of these people know the standard arguments for prices over quantities as exemplified in Weitzman's (1974, REStud) piece.  Indeed, much of the left-leaning academy (who are pushing a spending package), favor a cap and trade system on emissions.  In addition, there is a long academic literature citing the many circumstances in which taxes dominate quotas (check out this great paper by a Texas PhD candidate on the topic, for an example).  Much of the reason prices dominate quantities is because of the "knowledge problem".  Regulators don't know how best to achieve their goals because they are unsure about the costs and benefits individuals and firms face.  Fundamentally, this the reason economists say that "incentives matter" and stress second and third order effects of policy.

I think the problem with the stimulus plan is very similar.  In order for government spending to have a larger multiplier than the multiplier on tax cuts, the government must know how best to spend the money.  For example, if it spends it by employing people to dig holes in the ground, the multiplier will certainly be negative.  In a sense, government spending is like a quota.  The stimulus package would allocate $X and $Y there, based on what government officials think are useful area (still ignoring public choice here...).  A tax cut, on the other hand, would allow individuals and firms to pick exactly what they think money should be spent on.  They are the ones that best know what benefits them (and by extension, the economy).

Worried that individuals and firms might just stuff some the tax cut under their mattress, resulting in a lower multiplier effect?  This might happen.  I'm not arguing that all tax plans have a larger multiplier than all spending plans, just that there exist tax plans with a larger multiplier than any spending plan.  Did Romer and Bernstein find the multiplier to be larger on spending?  Maybe they just need a more creative tax cut.