More on Romer and Romer 2007
Brad DeLong has a question. He is impressed by Christine Romer’s latest speech but he doesn’t believe the huge estimate (on the order of 3) of the tax cut multiplier in Romer and Romer 2007 (warning pdf).
He suspects that they left something out.
I agree that the evidence for large multipliers (on the order of 3) in Romer and Romer is striking. But I find it disturbing: I don’t see why some of the channels they find are as strong as they are–the tremendously strong investment accelerator, for example. And thus I fear that their estimates too suffer from some opposite omitted variable bias that I am not smart enough to identify.
Regular readers of AngryBear will not be puzzled. Cactus has been there also pgl and this guy (I forgot honest).
After the jump I explain 3 ways in which the 3 might be overestimated.
Before the jump, I’d just like to mention that it was an excellent speech (just click the link at Brad’s post). She politely responded to Mankiw noting (as he has noted himself) that the R&R estimates of the effects of taxes can’t be compared to standard estimates of the effects of spending and that they couldn’t have dealt with spending (spending bills are too long for even the Romers to be willing to read every one and classify all changes in spending — sheesh). She also noted that addition to looking at rationed consumption and planned private investment during wwII (to support the claim that the multiplier is 0.8
OK enough niceness. First, Brad is using his tone for dealing with on leave colleagues whose courses I am teaching so I better make sure they know they are welcome back at Berkeley. So he doesn’t say that the Romers’ data was data (given in Italian and Latin) to the Romers by the Romers reading and deciding. Cactus was pretty ruthless on that point.
However, the discussion also makes it clear that smart people interpreted Romer and Romer 2007 as a vindication of supply side economics. Kevin Drum was clearly also sincere. It sure isn’t, because R&R looked at short run effects and assumed they were Keynesian. However, you don’t have to join the Club For Growth to believe that the estimates are biased up by supply side effects. In particular supply side effects can imply that the tax cut multiplier is greater than the spending multiplier (you have to go pretty much the full Laffer to believe that, but Mankiw was willing to go there (and to the Romers’ wedding where he was the best man)).
I find more convincing the possibility that tax cuts tricked managers into a bit of irrational exuberance, because they really do believe the supply side line deep down (or at least aren’t sure it is wrong).
In particular recall that the Romers found a huge effect of taxes on investment.
This would happen if some businessmen bought into the supply side mumbo jumbo (they don’t have to go the whole Laffer, they just have to believe, for example, that tax cuts will cause a permanent increase in the growth rate).
Now represenatives of business do argue for tax cuts all the time. It is tempting to assume that their calculation doesn’t go beyond their paymasters’ personal bottom line (I mean the people who call the tune not the shareholders). But maybe they aren’t hypocrites and really think that exogenous tax cuts cause an increase in trend growth making it a good time to invest.
So tax cuts have Keynesian effects and cause a bit of irrational exuberance. Seems to fit the facts no ?