More Krugman Obfuscation

Steven Horwitz

I haven't gone after Krugman for awhile, so why not today, especially when he's repeating myths about Herbert Hoover?  But that's not the real problem.  The real problem is how he spins the graph below, which shows the rate of change in consumption expenditures and gross investment by various levels of government:

 

Krugman's interpretation is:

The green line shows the rate of growth of federal G, which did shoot up, although it’s starting to fade out. The red line shows state and local G, which moved in the opposite direction. And the blue line in the middle shows the total, which did nothing much.

Now, this omits tax cuts and transfer payments, which presumably did something. But I think it’s fair to say that state and local cuts largely offset federal stimulus.

Now I'm no math whiz, but this seems wrong in four ways (and please correct me if you think I've interpreted this wrong or unfairly):

1. The red line did indeed "move in the opposite" direction of the green line.  But since this is rates of change, moving in the opposite direction does NOT mean that there was an absolute decline in state and local expenditures (with the exception of the brief period they were below the zero line).  The rest of the recession period state and local expenditures were increasing or constant.  "Moving in the opposite direction" when graphing rates of change does not imply the two things "offset".  Only time below zero does, and that is only a portion of the whole recession period here.

2. And this calls into question the conclusion that the total "did nothing much."  To the contrary:  the blue line, though falling, shows a consistently above zero growth rate.  Yes, the second derivative is negative, but the first is positive and that means total government expenditures were clearly growing over this period, rather than doing "nothing much."  There is no technical definition of "nothing much," but one would assume that means pretty close to a zero first derivative.  This one seems to average around 2% over the period.  When two growth rate lines are above zero but for a small portion for one, the sum of them cannot be zero or "nothing much."

3. By doing this in percentage terms, we don't get a sense of absolute magnitude.  For example, an average 2% growth rate of the total amount of federal, state, and local government spending is "real money" as they say. 

4. The growth rates may not look especially large when put up against early last decade, but those expenditures were surely 9/11-related.  Discounting that effect and looking at 2003-forward, the average height of the blue line during the current recession seems distinctly higher than it was in years prior.

No question state and local government expenditures have flat-lined recently, but where Krugman wants to read that as an explanation for why the recession is lingering (i.e., not enough boost in G), isn't it more plausible to argue that it's an effect not a cause?  With 10% unemployment and private investment way down, the tax revenues aren't there and most states can't deficit spend.  State and local G is down because of the recession.

So I'm not exactly sure what is point here is other than his usual call for more government spending, the facts be damned.

Comments

[Pietro M.]

Excellent post.

However, I think it's not enough Orwellian.

I recently discovered that keynesians have redefined fiscal expansion in such a way that they can claim that fiscal policies are relevant even though the recession persists.

The first example I met was Kuttner and POsen's paper on Japan (2001, maybe) on the Brookings papers. It defined fiscal expansion as "actual fiscal deficit MINUS the cyclical increase in fiscal deficit". The result was that fiscal policies in Japan (with some data massaging, like forgetting the FILP, whose weight on total spending I don't know) were CONTRACTIONARY. Despite a 200% debt.

Recently on VoxEU two economists from the central bank of Canada did the same, but without using vector autoregressions: they ran ARMA models and found that the present fiscal policy was contractionary, for similar reasons.

Let's learn doublethink: have a 10% deficit and a contractionary fiscal policy AT THE SAME TIME. That's the way to become great keynesian economists.

I don't get all the microeconomics behind the reasoning, probably they are trying the isolate the unexpected component of fiscal policy in order to avoid lucas critiques. The problem is that the result does not make much sense.

[Frank]

Something else that is obscured by looking at rates of change: I'm pretty sure the federal G is larger than the state and local G. Not only would state and local G need to fall (not just grow more slowly) to offset growth in federal G, state and local would have have a larger (negative) rate of change to offset federal bloat.

[Troy Camplin]

I think your final conclusion is correct. He's not that stupid. And if he's not so stupid as to be able to honestly misinterpret the data, then he's dishonestly misinterpreting the data. It is a rare bird who actually knows what the good is and actively works against it, but Krugman does so at every turn. He KNOWS that his interpretation is wrong. If this is rate of growth, then the state and local governments may not have grown much of late, but they haven't gotten smaller, either. (I would also note that much of the federal spending was supposed to be for state and local governments -- where did that money go?) More, if this is rate of growth, then it easily lends itself to such underhanded misinterpretation because most people look at a line going down on a graph and interpret that as a decrease. There is a huge difference between a decrease and a decrease in growth. What does the graph look like when we turn it into real dollars?

[Max]

IF he didn't understand the concept of rate of change and absolute change, I think he wouldn't be Prof. of anything. However, I also think that you can tend to forget it from time to time, especially the implications.

It is one of the things that always bothers me about politicians and many economic commentators. They always talk about rate of change and how a bump (up or down but still above zero) is a good/bad thing, while this still means that the economy grows/shrinks.
Either they don't understand the concept of rate of change or they don't know how to interpret it.

F.e. this is especially dire in the context of federal deficit, where they actually applaud the constraint on the growth rate, which means it is already a win if the growth is below inflation!

What would they write if a country really succeeded in REDUCING the deficit? I think we have hear a quite disturbing disconnection from reality and meaning...

[Barkley Rosser]

Steve,

I see nothing in Krugman's piece that says he would disagree with your view that the flatlining of state and local spending is a result of the recession.

[topills.com review]

Not only would state and local G need to fall (not just grow more slowly) to offset growth in federal G, state and local would have have a larger (negative) rate of change to offset federal bloat.