Krugman is seeing between the lines. He writes

One thing Mike fails to note is that the recent AEI paper on deficit reduction, which is cited by that JEC study in a way that might make you think that it supports the case for expansionary austerity, actually never provides any evidence to that effect; it focuses only on deficit reduction as an end in itself. In fact, it comes close to conceding defeat on the issue:

Actually the JEC report doesn’t even say that cutting spending is expansionary. It lays out a mechanism through which it might not be contractionary and the layers on a bunch of asterisks. The baseline conclusion of someone reading the report is that all the asterisks are there because they don’t really believe what they are saying is likely. However, someone asked them to put together a report showing that its possible.

I mean anything is possible right. I can read the report again but I don’t even think it describes expansion as likely at any point. Theoretically possible is the most the authors are willing to give.

The authors are the GOP staff, mind you.

My larger point is that economists doing their best to do economics aren’t that far a part. That is not to say there aren’t big policy disputes, but to a large extent these revolve around which elements of the analysis you care about.

The point of the JEC report is that cutting government is great, releases resources for the privates sector and could possibly not damage the overall economy.

In pure economics this doesn’t really contradict Krugman. If your value for government spending is low enough – perhaps even negative for some folks – then this argument makes sense.


Let me go into more detail quickly about the art of the “possible.” From the GOP staff report

Consequently, fiscal consolidation programs that reduce government spending decrease short-term uncertainty about taxes and diminish the specter of large tax increases in the future for both households and businesses. These “non-Keynesian” factors can boost GDP growth in the short term as well as the  long term because:

You might be inclined to read this as if the authors are saying GDP could grow as a result of the fiscal consolidation because of factors they are about to list.

Yet, that’s not what those sentences actually say. They say the non-Keynesian factors can boost GDP.

That is, after the you account for the loss in employment due to government consolidation that you shouldn’t stop there. It is conceivable that GDP will be higher than the depressed level that Keynesian analysis would predict because of non-Keynesian factors.

Its not saying that it would higher than baseline. Just higher than Keynesian analysis would predict and that it is conceivable that this could happen.

Or tellingly they also say

In some cases, these “non-Keynesian” effects may be strong enough to make fiscal consolidation programs expansionary in the short term as well the long term.   A number of developed countries have successfully reduced  government spending,  government budget deficits, and stabilized the level of government debt.  Fiscal consolidation programs in Canada, Sweden, and New Zealand, among others, achieved their goals for government deficit reduction and government debt stabilization and  boosted their real GDP growth rates by reducing government spending

Again you might be tempted to read this as: I am giving you some examples in which austerity caused short term growth. Again, that is not what this says.

Its says three things

  1. There are “some cases” – note the cases need not actually exist in the real world – in which short term non-Keynesian effects are strong enough to overwhelm the Keynesian effects that we readily admit occur.
  2. There are also some countries that have actually reduced government spending as a percentage of GDP. Note, I did not say these countries were the same as the cases I hypothesized above.
  3. In these cases real GDP growth rates were increased by reducing government spending. Note, I did not say when this increase occurred or whether or not it was cyclical or long run. In the cases presented below the evidence is largely than they were long run.