American growth could be higher than the pessimists think!

Robert J. Gordon has a new NBER working paper essay that really should cite Tyler Cowen, since it's all about a Great Stagnation in U.S. growth. Unlike Cowen, however, Gordon believes that the Stagnation will persist into the indefinite future. The paper is really two essays in one - the first part speculates that most of the big discoveries and inventions have already been made, and the second part identifies social and institutional constraints on U.S. growth. (Here are thoughts on the paper by FT Alphaville and Paul Krugman). Although there are some parts of the paper with which I agree, overall I think Gordon overstates the case for pessimism.

First, the technology part. Gordon claims that there have been three big "industrial revolutions" that have driven growth over the past two centuries, but that these bursts of innovation have diminishing returns:

A useful organizing principle to understand the pace of growth since 1750 is the sequence of three industrial revolutions. The first (IR #1) with its main inventions between 1750 and 1830 created steam engines, cotton spinning, and railroads. The second (IR #2) was the most important, with its three central inventions of electricity, the internal combustion engine, and running water with indoor plumbing, in the relatively short interval of 1870 to 1900. Both the first two revolutions required about 100 years for their full effects to percolate through the economy. During the two decades 1950-70 the benefits of the IR #2 were still transforming the economy, including air conditioning, home appliances, and the interstate highway system. After 1970 productivity growth slowed markedly, most plausibly because the main ideas of IR #2 had by and large been implemented by then...

The computer and Internet revolution (IR #3) began around 1960 and reached its climax in the dot.com era of the late 1990s, but its main impact on productivity has withered away in the past eight years. Many of the inventions that replaced tedious and repetitive clerical labor by computers happened a long time ago, in the 1970s and 1980s. Invention since 2000 has centered on entertainment and communication devices that are smaller, smarter, and more capable, but do not fundamentally change labor productivity or the standard of living in the way that electric light, motor cars, or indoor plumbing changed it...

The article suggests that it is useful to think of the innovative process as a series of discrete inventions followed by incremental improvements which ultimately tap the full potential of the initial invention. For the first two industrial revolutions, the incremental follow-up process lasted at least 100 years. For the more recent IR #3, the follow-up process was much faster. Taking the inventions and their follow-up improvements together, many of these processes could happen only once. Notable examples are speed of travel, temperature of interior space, and urbanization itself.

As I always say, we economists don't really know where technology is going. No one really does, and we probably know less than most. I agree with most of Gordon's claims about the three industrial revolutions. But what does that tell us about future revolutions?

There are many hypothetical technologies that have the power to transform the human experience even more profoundly than any that have gone before. Fusion power and self-replicating networked nanotech are two popular ones, but there are others that have the potential to be even more transformational. For example, personality upload (also called "brain emulation"). Why spend all your wealth rearranging the world we live in, when you can just create your own? An even bigger one is desire modification. Why spend all your wealth creating a better world, when you can make yourself like the one you're in? In fact, inventions like this are so transformational that they would call into question the very definition of economic growth, economic value, GDP, etc. The "industrial revolutions" from personality upload and/or desire modification would dwarf anything that has gone before.

But even on a more modest level, there are technological improvements already in the works that will help sustain growth. The fall in solar costs is one of these. Improvements in robotics/automation/A.I. are a second. Genetic technologies are a third, and biomechanical engineering (artificial eyes and the like) is a fourth. These things are real.

After delivering a long discourse on technology, Gordon pivots on a dime and lists six non-technological "headwinds" that will supposedly drag down U.S. growth in the near future. I'll go through these one by one...

(1) The “demographic dividend” is now in reverse motion. The original dividend was another one-time-only event, the movement of females into the labor force between 1965 and 1990, which raised hours per capita and allowed real per-capita real GDP to grow faster than output per hour. 

This is correct. Next:

(2) The second headwind already taken into account in the 2007-27 forecast is the plateau in educational attainment in the U.S. reached more than 20 years ago, as highlighted in the path-breaking work of Claudia Golden and Lawrence Katz (2008). 

This is also correct, BUT, then Gordon turns around and says this:

At the secondary level the OECD PISA test results for 37 nations had the U.S. recently ranked as ranked as 21st in reading, 31st in math, and 34th in science. There is an ongoing achievement gap between whites and Asians on the one hand and Hispanics and Blacks on the other, while the Hispanic percentage of our nation’s schoolchildren keeps increasing, dragging down the national average. Making matters worse is a new and growing gap between the educational preparation and achievement of American girls and boys; the female share of college graduates is now up to 58 percent.

First of all, this paragraph contains the disturbing assumption that Hispanic educational underachievement is something permanent (racial/genetic?). Fortunately the evidence seems to say otherwise.

On a more basic level, this paragraph doesn't support Gordon's contention. Gordon is arguing that growth from educational attainment is finished. As evidence he then points to low educational attainment. But low levels mean more room for growth. Doesn't everyone know this? You wouldn't point to Vietnam and say "Look how little capital this country has; slow capital accumulation will be a drag on Vietnamese growth"...right?

In other words, U.S. educational attainment has little room to grow in terms of quantity (years of study), but has absolutely huge room to grow in terms of quality. That is a potential source of accelerated growth.

Gordon goes on...

(3) The most important quantitatively in holding down the growth of our future income is rising inequality. The growth in median real income has been substantially slower than all of these growth rates of average per-capita income discussed thus far...If what we care about when we talk about “consumer well being” is the bottom 99 percent, then we must deduct 0.55 percent from the average growth rates of real GDP per capita presented here and elsewhere.

This is moving the goalpoasts, to be sure, though I agree with the sentiment. But notice that again, Gordon is claiming that levels are correlated with growth rates, and that's just not right. For example, suppose overall U.S. GDP growth slows to zero, but inequality decreases. By Gordon's own preferred metric, that would speed up growth for the 99%. Presto.

So once again, Gordon implictly assumes that recent trends will continue; if the trends reverse, the "headwind" becomes a tailwind.

And here is yet a third example:

(4) The interaction between globalization and ICT is a daunting headwind. Its effects include outsourcing of all types, from call centers to radiologist jobs. Foreign inexpensive labor competes with American labor not just through outsourcing, but also through imports. And these imports combine lower wages in emerging nations with growing technological capabilities there. This is nothing more than the Hecksher-Ohlin-Samuelson factor-price equalization theorem at work, and it inevitably has a damaging effect on the nations with the highest wage level, i.e., the United States.

I strongly agree with Gordon that factor-price equalization has held down American middle-class income, and contributed heavily to the inequality he discusses in his Headwind #3. But again, there is no certainty that this trend will continue. Chinese wages are rising rapidly, and unless India fixes its political problems, there will be no equally big source of cheap labor to replace it. If factor-price equalization goes into reverse, it will be a big tailwind for America's 99%.

Next we have:

(5) Energy and the environment represent the fifth headwind...India and China are both growing more rapidly than the U.S. and taken together those two nations are responsible for double the carbon emissions of the U.S.

Yes, global warming will probably restrain global growth.

Finally, we have:

(6) The twin household and government deficits represent the final headwind. Already in 2007 U.S. households suffered from an unprecedented overhang of debt equal to 133 percent of disposable income. The government debt was then manageable but has since begun to explode. Consumers have gradually been paying off debt, and this is one reason why the economic recovery has been so tepid. As a matter of arithmetic the ratio of government debt to GDP can be reduced by a mix of higher taxes, lower expenditures, and lower entitlement benefits (including higher retirement ages). But the same arithmetic implies that higher taxes and/or lower transfers reduces the growth rate of real household disposable income relative to that of real GDP.

There are two issues here. One is the "balance sheet recession"/liquidity trap. But that is fixable by policy. Next is the distortionary effect of taxation needed to make the government deficit sustainable. But I don't believe that this is nearly as large as conservatives claim, if the taxes are some combination of income, sales, and property taxes.

So in conclusion, I think that Gordon overstates the pessimistic case. Yes, there are headwinds for American growth, but there also potential tailwinds. These include:

• Improved educational quality
• Slowing factor price equalization
• New technologies like cheap solar, robotics, and biotech
• Increased high-skilled immigration

And farther in the future, there is the potential for further industrial revolutions that will dwarf the ones we've seen in the past.

Addendum: I seem to be the only person talking about Desire Modification as a transformational technology. Greg Egan and Vernor Vinge have written books in which this technology plays a central role. In my "spare time" I'm writing a couple of sci-fi short stories based on the idea. It's a really big deal, and I'll write a post about it soon. (Update: That post is now up.)

Update: John Cochrane makes a similar critique of Gordon's conflation of level effects with growth rate effects.