Arguments Against Infrastructure Stimulus – Economics
In my last post I touched on the political arguments against a big infrastructure stimulus, which boil down to illogical arguments about what we can and can’t afford. In this one I point out the problem with the most prevalent argument against such stimulus I’ve seen put forward by economists.
The reflexive neo-classical Austrian argument seems to me to come down to Ricardian equivalence, which I’ve seen bandied about quite a lot lately. David Ricardo noted back in 1820 that taxation and government borrowing are functionally the same, as the government would have to collect taxes in the future to repay the debt, the present value of which would be the same as taxes collected today.
Modern advocates have built on this thought (most notably in Robert Barro’s 1974 work) to argue that increased deficit spending can achieve no stimulative effect in the economy, as individuals will modify their behavior immediately, planning for the increased future taxation they know must occur to repay the new debt. Saving goes up and consumption falls, so private spending is effectively crowded out by the deficit spending. The argument is that this effect is complete, that deficit spending is completely offset in the economy as capital is taken away from the private economy and as private spending falls in anticipation of higher future taxes.
Ricardo’s central argument that debt financing would crowd out private investment clearly doesn’t fit today; with real interest rates on government bonds either near-zero or negative, there simply aren’t other investments that could be crowded out. The world’s capital prefers the safety of government debt to the assumption of any risk. So the notion that government borrowing would reduce private investment today is clearly wrong; a conclusion that should be enough to lead people to start thinking outside the box for solutions.
Interestingly, Ricardo didn’t believe in the crowding out of spending. The work used to establish the equivalence was on a choice between taxation and debt to finance war spending, and Ricardo noted that even though there was no financial difference between funding by immediate taxation or by increasing debt, the behavior of people under the two scenarios would not be the same. If people expected future taxes to increase to pay for the incurred debt, rationally speaking they should increase their savings by exactly as much as the deferred tax would have been. Ricardo didn’t believe that this happened in reality.
In any case, the spending at issue is fundamentally different from Ricardo’s war spending for two reasons. First, infrastructure spending may spur significantly higher economic benefits. Second, infrastructure spending is not unexpected today.
On the first point, war spending can have a productive component; production capacity may be expanded and useful skills may be developed – this is also true of infrastructure spending. But the work product of a war economy is used to destroy the enemy rather than to enrich the citizenry, and lives lost are a dead-weight loss (hardy har har) to the economy. Compared to infrastructure spending, the benefit to society of war spending is likely much smaller. So while Ricardo was considering spending that results in little or no increase in the general welfare, the deficit spending at issue today will generate a positive economic return inasmuch as it expands beneficial infrastructure (one could argue that defensive war spending maintains the status quo, serving much the same purpose as maintaining old infrastructure).
Here’s the big problem with the Ricardian equivalence argument as put forth against massive infrastructure spending: when the logic is extended just a little further than the anti-stimulus folks take it, the argument falls apart.
The Ricardian equivalence advocates say that infrastructure stimulus won’t work because the private economy would react to higher expected future taxes by cutting back on current spending. But it’s not the act of borrowing that causes the change in private sector behavior – it’s the change in expectations. The Ricardian equivalence can only be true if the additional government borrowing was previously unexpected. If the deficit spending is already expected, the theory says that the private sector’s current behavior already reflects this expectation.
Today, we’re talking about infrastructure spending that’s widely viewed as necessary and inevitable. The private sector therefore must expect that taxes in the future will rise to pay for these costs, and therefore current behavior must have incorporated this expectation. What I and the Paul Krugman axis are advocating is moving infrastructure spending forward in time to take advantage of absurdly low borrowing costs and abundant available labor. It’s not new spending; it’s future spending accelerated.
In the Ricardian equivalence framework, this acceleration of infrastructure spending would not change the size of expected future taxes; it would change the timing of future taxes that are expected already. So the crowding out of spending based on the Ricardian equivalence theory would be much smaller than opponents claim, because the private sector already expects taxes to rise in the future to pay for infrastructure.
Perhaps, if the Ricardian equivalence holds in reality (there are lots of reasons why it doesn’t, which Elmendorf and Mankiw outlined), a massive infrastructure stimulus would be somewhat offset by increased saving in the private sector. However, this effect would necessarily be much smaller than is generally argued, because advocates assign rational expectations only as far as fits their purposes.
The bottom-line facts remain: we have huge infrastructure needs that must be met within a relatively small number of years; we have large amounts of underutilized resources today; and we have borrowing costs at bizarre, market-failure levels.