Just about all economists agree that a sudden dramatic increase in oil prices is likely to be bearish for the US economy. Of course that’s not always the case (for “never reason from a price change” reasons), but it’s usually true that sudden oil price spikes have supply-side causes.

So is it symmetrical? Does a sudden plunge in oil prices boost the economy? I see two views:

The Keynesian view sees plunging oil prices as a sort of “tax cut”, which puts purchasing power in the pockets of consumers. Because the US is a net importer of oil, this should boost the economy.

I’ll call the alternative view the Sumner/Kling hypothesis.

This alternative view holds that aggregate demand is not likely to be affected by lower oil prices, for standard monetary offset reasons. (That’s the Sumner part.) In addition, the impact of oil prices shocks is not symmetrical. Plunging oil prices cause resources to be re-allocated out of energy and into other sectors. During this reallocation, output growth may actually be slightly depressed, depending on the size of the industry being impacted. (That’s the Arnold Kling part of the theory.)

James Hamilton is perhaps the leading energy economist in America, so let’s see how he summarizes the impact of the oil price plunge:

Christiane Baumeister and Lutz Kilian presented an interesting paper at the Brookings Institution last week that takes a detailed look at the effects on the U.S. economy of the dramatic oil price decline of 2014-2015.

An oil price decrease loosens consumers’ budget constraints, and historically this tends to show up as higher consumption spending. I had done some earlier analysis suggesting that might not be happening this time. But Baumeister and Kilian . . . conclude that . . . real consumption spending grew on average by 3.1% over the two years since oil prices began falling in 2014:Q3 compared with only 2.0% during the preceding two years. . . .

But gains to consumer spending were mostly offset by a drop in oil-related investment spending. Nonresidential fixed investment had been growing at a 4.3% rate prior to the oil-price drop but has only increased at a 0.8% annual rate since, due to a 50% drop in investment spending in the oil sector.

. . . In idealized economic models, the resources that had been producing oil should now shift to producing other goods, and with the new terms of trade we should as a result be even more productive than before.

But in the real world shifting resources is easier said than done. We now have a large stock of capital that was being used to develop U.S. shale oil, and contrary to the predictions of simple economic models, there is not some other more productive place to use that equipment. . . . Somewhat surprisingly, there wasn’t even any significant gain in the transportation sector, the one area you might have expected to enjoy the biggest windfall from lower oil prices. The weak performance in transportation comes from a big drop in rail shipments, dominated by lower rail transport of petroleum. Idealized economic models treat capital as if it’s a big fungible lump that you can use wherever it’s most productive. But it turns out to be not that easy to try to carry milk in an old oil tanker rail car.

Those last two sentences could have been penned by Arnold.

. . . The bottom line is that there seemed to be little net stimulus to the U.S. economy from the collapse in oil prices.

Here’s my takeaway. You don’t want to treat RGDP as an undifferentiated mass of “output”. Inputs cannot be costlessly and quickly moved from one sector to another. Hence a sudden oil price shock might actually depress RGDP growth, regardless of whether it’s an increase or a decrease.

On the other hand, I think you do want to treat NGDP (or better yet nominal labor compensation) as a “fungible lump”. That variable tells you pretty much all you need to know about how AD is impacting the labor market.

To summarize, in my view NGDP is a homogenous mass of nominal spending/income, while RGDP is a heterogeneous collection of very different industries.