Evan Soltas
Apr 2, 2012

The Unnecessary Recession

Investment, and not consumption, is the culprit -- and we can fix it.

One of the central principles of Keynesian economics that gets widely taught is that underconsumption causes recessions -- and that a sudden coordinated increase in autonomous saving explains aggregate demand volatility. This is something that, as recently as a few days ago, I would have thought, would have taken as a base assumption. Now I don't. The Keynesian story is not implausible, or impossible; it's just that the only recession in the last three decades which really matches this description in the early 90s recession. Instead, the most significant cause of cyclical volatility has been investment.Perhaps this surprises nobody. John Hicks once said that "[i]nvestment is a flighty bird which needs to be controlled," after all. Even though private investment is, in the United States, usually between a third and a fourth of private consumption, the volatility of the former is often more than three to four times that of the latter, meaning that the change in constant dollars of investment is often greater than that in consumption.

This is increasingly true -- variance in real GDP less investment over the past decade now is a fifth of real GDP variance over the same period, down from half in the 1950s. That spike you see in the early 90s in the recession I already mentioned, which my understanding has been for a while was caused by a sudden drop in consumer confidence during the Gulf War.

Consumption, in fact, has been remarkably constant in much of the post-Volcker era. Indeed, if investment had grown exactly as consumption did during this period, the United States would have not seen a recession since 1981. Hopefully, you're beginning to see where I'm going with this: investment stability = far fewer recessions.

Now, I'm not trying to imply that the fiscal-monetary policy mix should seek to stabilize real GDP, or real investment -- there are obviously costs to inflation. (This would have not been an optimal policy during the 70s real shocks to supply.) But I think the variance in investment should suggest to us that we should look to primarily monetary, and not fiscal, policy for a cure -- presuming that demand is not perfectly substitutable between categories in the long term, which is an important reason to remember that aggregates never capture the whole story.

We can say, however, that a more aggressive monetary policy response to recessions since 1981 (which have all been driven by nominal shocks) would have resulted in substantial gains in real output stability. Given stable long-run inflation expectations, I think the case that there would have been a net benefit from such a policy -- i.e. that short-run inflation costs would have been low -- is pretty evident. In particular, I think a good base assumption for how an NGDP target would work in response to a nominal shock is that real investment would be somewhere between stable, with no effect on consumption, and real investment growth of 2 percent, with a 1.5 Keynesian multiplier on consumption.

Now let's apply this to the 2008 recession, which, as we can see in the two graphs above, was the most investment-driven recession in the postwar era.

Looking below now, if we assumed that, starting in January 2007, monetary policy held real investment constant, and that there was no multiplier effect on any other component of output (this is to get a low-side estimate of the effect of such a policy), there is basically no recession in 2008. Maybe a soft landing in 2009. This is why I called the recession "unnecessary" in the title -- it's that nominal stability of aggregate demand would have plausibly stabilized real investment, meaning that there would have been no recession in either real or nominal terms. It also implies that the amount of inflation needed for this outcome would have been minimal -- all that was needed was the pre-existing and credible commitment on the part of the central bank to nominal stability!

If we make a higher-end estimate of the effect of the policy, and say that investment would have grown at 2 percent annually (this is below trend for the time series, but I think plausible), and that there would have been a 1.5 multiplier on the rest of output, then there really was no need for the 2008 recession. Notably, we would have also "hit bottom," if we can even call it that, in January 2009 -- the period of falling real output would have been almost a year shorter.I want to look a little more closely, even, at this recession. We can do this by imagining what would have happened had all components of GDP but a given variable been held constant, without considering multipliers. (That's what the graph below is.) A whopping 7 percent of the drop in RGDP, as we can see, came directly from investment, as compared to the 2 percent which came from consumption. Importantly, government only contributed 1 percent in real terms to cushioning the recovery in terms of the direct impact of their expenditure (this is without considering the multiplier, but a rough estimate of all effects is 1.3 percent, given a 20 percent share of government in real output and the 1.5 multiplier I think is standard.) The second benefit of monetary easing would have been the increase in net exports which came from the depreciation of the dollar, so if we think about (investment + net exports) stability, the burden of monetary policy is smaller, and the idea that an NGDP target would have stabilized this becomes even more obvious.
One more point about consumption during this recession is that the drop can be entirely explained by the multiplier effect from the drop in investment, and not the other way. ("T" refers to my model of consumption with a 1.5 multiplier from how investment changed, given their relative fractions of real output; "A" is the actual consumption.) This relationship is robust from 2008 through late 2010, when I think there would have been distinct shocks to consumption and investment (increasing consumer confidence, decreasing investor confidence).A few conclusions we can walk away with -- recessions are primarily investment phenomena, and this is increasingly true in the US, and this recession was particularly unnecessary, as the drop in investment would have not happened, or affected consumption, with better monetary policy.