Matt Yglesias has a post discussing the new cold war with China. In the end, he’s skeptical of calls for more military spending to counter China, but along the way I think he concedes too much to the other side:

But when defense spending went down in Obama’s second term, that was part of an across-the-board austerity agenda that was economically harmful. Under Trump, Congress cut deals to lift the sequester, and both military and non-military spending went up. That, it seems to me, was good.

We should not let concern about China deprive us of funds for useful domestic spending. But whether or military spending actually crowds out or crowds in domestic spending depends on the situation.

I strongly disagree with the first paragraph, and the second is also somewhat misleading.

As I’ve pointed out in numerous previous posts, the 2013 austerity program represented a spectacular failure of Keynesian economics.  Keynesians predicted that the fiscal austerity would slow growth, perhaps triggering a recession, and instead both real and nominal GDP growth sped up after January 1, 2013.  That’s because the impact of fiscal shocks is normally offset by monetary policy, as was the case in 2013.

I don’t accept Keynesian models of fiscal policy, but even if we accept the standard textbook Keynesian model, a persistent increase in military spending, say from 3% to 4% of GDP, completely crowds out domestic spending.  That’s because in the standard Keynesian model, military spending can only avoid being zero sum by reducing the volatility of the business cycle.  In other words, it would need to be “countercyclical.”  That’s presumably what Matt means by “depends on the situation”.  But that phrase doesn’t mean what many people think it means.

If you permanently raise military spending from 3% of GDP to 4% of GDP, it doesn’t even have the effect of boosting GDP during a recession period.  The economy adjusts to the new trend rate of military spending, wages and prices reflect those levels, and shocks are just as likely to produce big recessions at 3% of GDP military spending as at 4% of GDP military spending.  In the more sophisticated Keynesian models, it’s not government spending that matters, it’s changes in government spending.  Again, to boost growth the military budget would have to be countercyclical.

And here’s the basic problem, it’s very unlikely that military spending would be meaningfully countercyclical.  In any case, that would not be an argument for a bigger military, it would be an argument for a more countercyclical military.  If we are talking about long run changes in the military as a share of GDP, we need to assume 100% crowding out, if not more.

In practice, the crowding out is more than 100%.  That’s because higher military spending requires higher taxes.  As taxes as a share of GDP rise, other things equal, the level of GDP per capita declines.  While rich countries often have fairly high taxes (i.e. Sweden), between two otherwise similar rich countries the one with lower taxes usually has higher per capita GDP.  (And even Sweden’s economy improved after it cut back on its extremely high tax rates during the 1990s.)  That’s one reason why America’s per capita GDP is higher than Europe’s—we have lower tax rates. In general, rich countries with lower taxes (the US, Canada, Australia, Switzerland, Ireland, Singapore, etc.) tend to be richer than those with higher taxes.

PS.  A tiny part of military spending may involve growth generating R&D.  But that’s an argument for the government funding some R&D, not an argument for massive military spending.