Is the U.S. Really Below Potential GDP?

NOTE: The Growth Economics Blog has moved sites. Click here to find this post at the new site.

The CBO just released a new projection of both GDP and the budget out to 2024. In short, the CBO sees the U.S. staying below potential GDP for several years. Menzie Chinn just did a short review of how people use inflation and/or unemployment to try and figure out the difference difference between actual and potential GDP.

From a growth perspective, I wanted to take a look at the projections a little differently. First, I don’t much care about the level of aggregate GDP, I care about the level of GDP per capita. So I took the CBO numbers and combined them with population figures and projections to get actual and projected GDP per capita for the U.S. Note, I’m using the CBO projections for actual GDP, not their potential GDP numbers. I want to look at the expected GDP numbers.

Second, I wanted to consider how this projected GDP per capita compared to long-run trends, rather than using inflation or unemployment to assess whether GDP per capita is “at potential”. I am looking instead whether GDP per capita has deviated from its long-run path. To do this I merged the GDP per capita projections from the CBO with the Maddison dataset on GDP per capita from 1970 to 2008. (The CBO goes back far enough that the two series overlap and I can adjust the actual levels of GDP per capita to match).

I took the trend in GDP per capita from 1990 to 2007, and extrapolated that out from 2008 to 2024. Then I plotted the actual and CBO-projected GDP per capita data against that trend. Here is what you get:
Post 1990 Trend
It’s clear here that in 2007 GDP per capita drops below the 1990-2007 trend line. Moreover, the CBO expects that GDP per capita will stay below that trend line out until 2024. It looks like a distinct “level shift” in the parlance of growth economics. GDP per capita is something like 13% below the 1990-2007 trend.

If you look at the post-war trend in GDP per capita from 1947 to 2007, you get something similar. The gap in 2024, 18% below trend, is actually worse than the gap using the post-1990 era.
Post War Trend

But if you extend your view back even further, and incorporate the whole period of 1870-2007 to form the trend line, things look different. Now, if you plot the projected GDP per capita against the trend, it looks as if the U.S. is spot on.
All Data Trend
GDP per capita is almost exactly where you’d expect it given the historical trend. The CBO expects GDP per capita to be a little low in 2024, about 2% behind the full trend line. Using the 1870-2007 trend, there doesn’t appear to be anything particularly unusual about the projected path of GDP per capita. The U.S. seems to be moving along the same balanced growth path it always has.

What really looks like the anomaly in U.S. data is the extended period from about 1990 to 2010 that we spent above trend. You could think of this as capturing John Fernald’s argument (or see here) that the IT boom of the 1990’s was a one-time level shift up in GDP. We got a big boost from that, but now the economy is settling back to the long-run growth path.

[You should not – NOT – use this as an argument that the financial crash and subsequent recession were necessary, useful, or welfare-improving. It is quite possible for the economy to have managed a graceful slide back to the long-run trend line after 2007 rather than experiencing it all in one dramatic plunge. The long-run trend is like gravity. Yes, it will win in the end, but that does not mean that I have to leap to the ground after cleaning out my gutters. I have a ladder.]

I really thought when I started playing with this data that I’d be writing a post about how the Great Recession had fundamentally shifted GDP per capita below the long-run trend, and that this represented a really fundamental shock given how stable the long-run trend had been until now. But the current path of GDP per capita doesn’t appear to be that surprising in historical perspective.

The big caveat here is that the CBO could be entirely wrong about future GDP per capita growth. If they have been overly optimistic, then we could certainly find ourselves falling below even the very long-run trend. Then again, they could have been pessimistic, and we might find ourselves above trend for all I know. But even with all the uncertainty, the expectation is that the U.S. economy will find itself right where you would have predicted it would be.

9 thoughts on “Is the U.S. Really Below Potential GDP?

  1. I’m not sure I necessarily agree with the conclusion that the economy is on-trend for the 1790 – 20xx period. First, I don’t particularly trust the data before 1947 (and trust it even less before 1929). Second, I see no reason to assume one constant trend for the entire period. (I am, however, open to persuasion.)

    • Data questions, yes, they are subject to errors. Perhaps major ones.

      But the linear trend is hard to reject. That is, if you tried to fit a more flexible model you wouldn’t find any statistical support for different growth rates over time. Could the growth rate be fundamentally lower as of 2008 or 2009? Sure, but we don’t have enough data to detect that yet. I figured this out once, and it’ll be like 2025 before you’ll have enough evidence to statistically be sure that growth rates have changed (even if they honestly did).

  2. The problem here is the attempt to find a single, simplistic explanation (“the end of the 90s IT boom”) for something much broader and complicated (the multi-trillion dollar economic collapse). The impact of IT innovation on national productivity is certainly worthy of study, but Fernald’s papers don’t in any way justify the claim that it singlehandedly drove massive macro shifts. Since Fernald doesn’t recognize any other factors that could have affected productivity data in the 90s (offshoring of manufacturing, the dot-com bubble, financialization, reduced competition and increased rent-seeking in major industries, technical innovation from any non-IT source) there’s no support for the conclusion that the alleged “IT innovation collapse” was way more important than other factors. In fact his data of “IT intensive industries” includes all of finance, so he appears to be arguing that the growth of finance (and similar) industries in the 90s was all due to “IT innovation” and its late 90s collapse was because “IT innovation stopped”, a claim most people would find laughable. Since the conclusion isn’t based on any analysis of how “early 90s IT innovation” actually made society better off, it fails to consider that some of those apparent gains had been ficticious, and others directly led to the destruction of national wealth. The analysis also seems to presume that welfare-enhancing innovation is something that just magically falls from the sky, and can magically stop falling. Fernald makes the unsubstantiated claim that all the “low-hanging IT innovation fruit” magically ran out around 1997. Isn’t it equally plausible that the incentives in “IT intensive industries” like finance and telecoms and military contracting had shifted away from the (difficult, risky) work needed to drive future innovation towards much easier activities offering much higher near-term returns (rent-seeking, creating systemic financial risk), but whose long-term impacts more logically fits the post-2008 economic numbers?
    This argument parallels the many “today’s reduced GDP numbers all explained by demographics” papers. Demographic shifts, like IT technology impacts, are worthy of study, and no attempt to explain the long-term macro shifts should ignore them. But since both sets of papers are presented in ways designed to deny or distract attention from other likely causes (“Stop blaming Wall Street for all those unemployed workers–it was all caused by the magical end of IT innovation! or maybe it was the aging population!”) I think it is fair to questions the motives and the intellectual integrity of the people writing and promulgating them.

    • Fernald isn’t a “mono-causal” guy. His papers are more about how IT pushed us above trend in the 90’s. But he doesn’t use that to explain why we had a big recession. His major point is that we shouldn’t be too surprised if GDP doesn’t return to the 1990’s trend after this recession. You’re right about the demographics stuff – it is certainly *one* factor in keeping GDP from growing more rapidly, but hardly the *only* one.

  3. Pingback: Blogs round-up: 13 February 2015 | FODA RECORD

  4. Pingback: Significant Changes in GDP Growth | The Growth Economics Blog

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s