Paul Krugman in his recent op ed in the NY Times titled “Stimulus timing”, (@NYTimeskrugman) appears to be criticizing the administration’s stimulus package. Considering Paul has won the Nobel Prize in economics, this should be something to be concerned about, but is it really as bad as Paul suggests?
Although the timing of the stimulus package is his tag, he spends little time discussing the timing of the package relative to the real world and focuses on how the numbers graph out. First he subtracts out the underlying GDP to show the effect of the package. Although this makes good academic talking, it doesn’t take into account any possible change in the underlying real world GDP, it doesn’t even discuss it. That means it also ignores the possibility that he stimulus could, uh, stimulate and it seems to assume the money actually leaves the money supply the day after we get it.
He goes on to show how the stimulus is an inverted “U” (a somewhat lopsided one in this case) with maximum effect up front. He briefly describes the elements of his graph and then displays a nasty looking chart and comments “You can see why I and many others are worried about the second half of next year.” I’ll grant him that we are going to get less and less new benefits from stimulus from here on out but he makes it look like the world is going to end next summer. But what does the graph really show us? It shows us how horrible the day after Christmas really is.
Let’s use an example, a much simpler one but using the same numbers. Over the next 9 days I’ll buy things from you, not because I need them but because business is bad and I have cash to burn, so I just want to help you out, let’s call it a “stimulus”. I’ll spend $770, but I’ve decided you need a real boost right now so I’m not going to give it to you evenly distributed but in a sort of lopsided upside down “U” . Let’s tabulate these as PK does, P1, P2, etc. are the payments:
Rate Change Cumulative
P1 35 +35 35
P2 80 +45 115
P3 110 +30 225
P4 120 +10 345
P5 125 +5 470
P6 120 -5 590
P7 100 -20 690
P8 50 -50 740
P9 30 -20 770
I’ll try to manage posting a graph for this later but, guess what, there’s a perfect one at PK’s article and it happens to be the very graph I want to critique, so now I will borrow very liberally from that article:
“Now think about three questions you might ask. The first is, how much higher is [your income] this [day] than it would be without the stimulus? This should depend on “Rate” — on the quantity of goods and services [I am] buying right now.” In other words, your income is higher today than it would have been without the payment by exactly the amount of the payment! Amazing isn’t it, now if you really want to see something take the month of your birth and multiply by 7 . . .
“The second question is, how much faster is [your income] growth this [day] than it would be without the stimulus? This should depend on “Change” — on the extent to which [I’m] buying more stuff than [I] did [yesterday].” In other words, your income growth will be faster (or slower) than it would be without my purchases, by exactly the amount by which I change my payments from one day to the next. Some days I’m going to give you a lot of money but by showing you the change, you will be so scared you won’t spend it and if I show you this beforehand, you may even ask me not to give you the money.
“Finally, you can ask, how much of the stimulus money has been spent?” At PK’s column a Nobel laureate will explain to you how to figure this out, but here’s a test to see if you can do it on your own: take $10 to the store, pick out a couple candy bars and a bottle of soda but don’t look at the price, then hand the clerk the ten, look at what you have left and tell me how much you spent, for this example it doesn’t even matter if the clerk gave correct change. Come on PK, really?
As we’ve discussed, the “Rate” is an inverted “U”, and that PK says, is exactly his point. “The peak effect on the level of [your income] comes at the top of the curve, but the peak effect on growth comes earlier, before the curve flattens out. In the table above, [my] spending peaks [on the fifth payment], but the peak impact on growth is in the [second payment], [if we look at this at the end of day three], it’s behind us. That’s true even though by the end of [day three] less than a third of the money has been spent.”
“And when the spending begins to tail off, the effect on growth turns negative.” Uh, yep, that’s what an inverted “U” graph does, on the back end it goes down.
Another way to put this is that if I give you a present today, tomorrow you’re going to get 100% less than you would have gotten if I’d never given you anything! That’s even worse than an inverted “U”, it’s an inverted “V”! So, PK, don’t sugarcoat it, let us have it straight and stop charting change and start charting percentage change so we can see just how bad it really is. Cancel Christmas next year, we can’t possibly deal with December 26.