“[Faith] in the spontaneous forces of adjustment [is what] makes the liberal accept changes without apprehension, even though he does not know how the necessary adaptations will be brought about.” – Friedrich Hayek

Here’s what the uber-optimist, techno-solutionist, Up Wing part of me would like to believe: Amid all the death and suffering of 2020, the faintest of silver linings appeared. National economic statistics seemed to provide intriguing evidence that the Great Stagnation — the long-term downshift in productivity growth and innovation, whether dated from 2005 or 1973 — was over. Which would be awesome, of course. The scientific and technological advances reflected in an economy’s long-term productive capacity are the engines of economic growth and rising living standards. When you think about prosperity, think about productivity.

Which is why these numbers are a major bummer: In the ten years from 2010 to 2019, business productivity was only 1.1 percent a year — the slowest productivity growth of any decade in U.S. history. More context: That performance is far below the postwar average of 2.5 percent from 1950 to 2009, even further below the Golden Age rate of 3.2 percent from 1950 to 1972, and the 3.3 percent rate during the IT revolution of 1996 through 2004.

But, but, but … productivity accelerated from 2017 through 2019, to an average annual rate of 1.5 percent vs. 0.9 percent from 2010 through 2016. Then came 2020. Over the four quarters of that year, productivity growth surged to 4.1 percent. “This apparent productivity growth revival has been interpreted as caused by automation, artificial intelligence, and a massive investment by households in the equipment and software needed to conduct work from home,” observe economists Robert J. Gordon (Northwestern University) and Hassan Sayed (Princeton University) in their new NBER working paper, “A New Interpretation of Productivity Growth Dynamics in the Pre-Pandemic and Pandemic Era U.S. Economy, 1950-2022.” (All of the above numbers come for their paper.)

Of course, the co-authorship of Gordon is sort of a spoiler here. He’s a well-known skeptic — a highly respected one, to be sure — of the notion that advances in AI and robots are about to drive a big growth and productivity revival. (For a deep dive into Gordon’s scholarship on the issue, definitely check out his brilliant 2016 book, The Rise and Fall of American Growth: The U.S. Standard of Living since the Civil War.) So it’s hardly a surprise that Gordon and Sayed conclude the US economy is not experiencing a revival of such creative innovation. It’s a view supported by what happened after the 2020 productivity boom. Productivity growth fell to only 0.6 percent in the five quarters of 2021–22.

So what’s happening here, at least according to Gordon and Sayed? Something similar to what happened after the Global Financial Crisis/Great Recession. (Keep in mind that by “productivity” what we are basically talking about is gross domestic product per hour worked — labor productivity. That said, Gordon and Sayed are using a slightly modified version where output is the geometric average of business sector GDP and gross domestic income, or GDI. Here’s a good explainer of GDP vs. GDI.) During that downturn, it legitimately looked like another Great Depression was possible. So businesses freaked out and slashed payrolls much like Hopper slashed the demogorgon with his Conan the Barbarian sword at the Soviet Gulag prison in the just-concluded season of Stranger Things. (Yeah, there’s a lot to unpack in that sentence.)

Or as Gordon and Sayed put it, “In their panicked reaction to the collapse of output in the 2008–09 recession, business firms overreacted with ‘excess layoffs,’ adjusting hours to the output decline with a far higher elasticity than normal.” This overshooting, which led to a productivity surge of 6.4 percent in 2009, was gradually reversed by a slow-but-steady pace of rehiring over the next decade, which finally wound down in the 2017–2019 period. Gordon and Sayed: “We estimate that without that post-recession rehiring spread out over the seven years 2010–16, the annual rate of productivity growth on average would have been almost twice as high — 1.7 percent per year rather than 0.9 percent per year.”

Interregnum: Why might productivity go up during a downturn? That is, why might the percent decline in aggregate hours worked be larger than the decline in output? One possible reason is a changing mix of workers with higher productivity workers keeping their jobs. Another: Retained workers up their game and work harder.

Likewise, “The rapid growth of productivity in 2020 is explained by the same pattern of excess layoffs as in 2008–09, but with a faster adjustment due to the shorter duration of the 2020 recession.” (This also helps explain why hiring has so far stayed strong despite weakening economic growth.) With the rehiring comes a drop in productivity since businesses have actually not become wildly more efficient or innovative — except perhaps for the work-from-home phenomenon. From the paper: “Positive pandemic-era productivity growth can be entirely explained by a surge in the performance of work-from-home service industries, while goods industries soared and then slumped, while contact services recorded strongly negative productivity growth throughout 2020-22.”

So where does this leave their assessment of long-run productivity growth? I can’t say I like the Down Wing conclusion:

The paper has suggested that productivity growth was distorted by excess layoffs in 2008-09, shifting some productivity growth into that two-year period and moving it away from 2010-16 when rehiring cancelled the effects of the excess layoffs. The business-sector data displayed in Table 1 and used in our regressions and simulations records growth rates of 1.3 percent in 2005-07, 3.2 percent in 2008-09, and 0.9 percent in 2010-16. The underlying trend of productivity growth is better represented by averaging across these three intervals, resulting in a 2005-16 growth rate of 1.4 percent. This is remarkably close to the 1.5 percent in the apparent revival interval of 2017-19 and also close to the 1.5 percent achieved in the long “slowdown” period of 1973-95. A cautionary note is that a suggestion of 1.5 percent for the long-term growth of productivity in the business sector translates into only 1.1 percent for the total economy, since productivity in the total economy grew an average of 0.4 percent slower than in the business sector during 1950-2019. This 1.1 percent suggestion is relevant for predictions of future potential output growth and is slower than the current CBO ten-year forecast of 1.4 percent total-economy productivity growth. Our conclusion is thus consistent with the long-term forecast of 1.2 percent for the total economy included in our previous long-term evaluations. … [This leaves] the dot.com achievement of 3.3 percent as a historic outlier as it recedes further into the past.

This is the second gloomy report about the future of US economic growth that I’ve featured recently in this newsletter. (Back in April, I wrote about “Additive Growth” by Thomas Philippon of New York University’s Stern School of Business. And then in June I did a 5QQ chat with him.) And that’s OK. I think we should assume the worst. We should assume disappointing past results suggest disappointing future performance. The history of the Great Stagnation is one of policymakers underestimating the challenge confronting them. So they need to aggressively take advantage of every tool they have at their disposal: more federal R&D, pro-build regulation, pro-investment tax reform, increased economic openness. They should create as fertile an ecology as possible for increased scientific discovery, technological invention, and commercial innovation.

Benjamin F. Jones is an economist and professor at Northwestern University’s Kellogg School of Management. Last summer, he authored a fascinating paper, “Science and Innovation: The Under-Fueled Engine of Prosperity,” one that I’ve referred to several times in this newsletter, including the previous issue of Faster, Please. As such, I thought I would highlight a podcast chat I had with him last December.

James Pethokoukis: The US invests about 3 percent of GDP in R&D. Is that a lot, in the post-war era, a little — where does that rank?

Benjamin Jones: It’s really about the same. For the most part, the US is quite consistent, like other countries, in how much they invest — what share of our total resources we invest in a given year in research and development.

The composition of it has shifted in an important way, which is to say that the private sector makes up an increasingly large portion of that 3 percent. And public investment, which really supports science in particular, makes up a declining share. And, in fact, the public support portion is now at its lowest level in about the last 70 years.

Does that composition, how it’s divvied up between public and private, matter?

It matters in the sense that what a lot of the public money is doing is science, and what a lot of the private money is doing is application and the creation of specific goods and services. And those are, of course, complements. You can think of the science as opening up new doorways, and then the private sector is walking through those doorways and making applications from the new knowledge that’s been generated.

The fact that we’re doing less science as a share of our resources is, I think, concerning because it’s something that’s opening up many fewer doorways and, therefore, not creating as many opportunities.

Why doesn’t government invest in R&D the way it used to?

I think the answer is probably salience to the public. The public, and I think I would include policymakers, don’t fully understand the value of these investments to our future potential and progress. I think that our public interest waxes and wanes. I think it probably becomes more salient in moments of fear and with a sense of competition internationally. So the Sputnik moment was a moment that sparked the Apollo program, a moment where the US suddenly feels behind.

Take a modern context: Today, where there’s COVID, we have a very clear challenge. You get Operation Warp Speed. I think the sense of China rising today is also something that is pressing on Congress, effectively in many ways, to increase our investments in this space. I think a sense of threat can make it more salient to the public.

Walk me through the process of how we figure out how much we should spend. I would guess that might begin by just looking at the returns to public spending on R&D.

Let’s just start with a high-level total number — that 3 percent. Evidence suggests that for every dollar we put into the R&D machine, across the waterfront of R&D policies and investments, on average that’s returning something like $5 or more back in social value. In other words, put $1 in, you get $5 back. That’s an amazing return. It’s an incredible return.

You could imagine doubling it, 3 percent to 6 percent, and that could really elevate the growth rate of the economy. In some ways, the harder question actually is, what exactly would you invest in there? When you study these individual R&D policies, you tend to find over and over again that they have very high returns themselves. A simple policy would be to just expand our investment across the range of inputs into the R&D process, which are very many.

Rather than focusing on whether we should spend it on this area or that area, one of the debates is how much should be very basic research, raw science, versus more applied. Does that matter in these calculations of return?

I think what we know is that the returns look high in both cases. We don’t know which return is higher. Nor are they necessarily separable, because applied research builds on the basic research. But also, this actually may be more surprising to people, basic research often builds on applied research. A lot of the really interesting understandings we’ve developed of nature actually come from people solving very particular applied work.

To try to separate basic research from applied research — in some ways, it’s a natural question. It’s a hard question to answer because they interplay so clearly. What we’d probably do best with is spreading our bets and doing more of both.

So one criticism of basic research I’ve heard is that the United States will do the initial research, and then other countries take that new knowledge and create their own industries. Is that something we should worry about?

There are somewhat contending imperatives in R&D. The question at some level is, do we just want to get as good as we can ourselves, or is it really about how we rate versus somebody else? In a health context, if we can live a longer life and we can solve Alzheimer’s, we’re pretty happy. And we’re probably pretty happy if people in Switzerland and South Africa also live longer lives and solve Alzheimer’s. Whereas in a national security context, it might seem like you need to be ahead. I think that in particular is where that concern might raise itself more acutely.

I think, in a sense, the way you hold onto the advantage of basic research is largely because you’re investing in the people in your country, who are in fact doing it, and then have this leg-up advantage. One way to see that, of course, is: Why do we see clustering of innovative activity on the map? It’s because that’s where the people are, and they cluster with all these different specialties. That becomes a very sticky and self-fulfilling investment.

I think, yes, it will be the case that basic research will spill over, to some extent, to other countries. In many ways, that’s a good thing because other people can benefit. But in terms of keeping ahead, that human capital piece is a key part of why you do keep ahead through those investments.

We can create talented people in this country or we can bring those people in from other countries. How are we doing in both those areas?

It’s a mixed scorecard. The US does, obviously, very well at higher-level education in research institutions. People from all over the world want to come and study in those institutions. If you look at our domestic pipeline, in terms of our K–12 education, we seem to do very poorly compared to many other countries.

The other side, of course, is importing talent through immigration policy. This is, of course, an area where the US, for a long time, gained enormously. If you look at who does a lot of the great research in the United States, who does a lot of the invention and the patenting, who starts a lot of the great companies, the answer is surprisingly often people born outside the US who move to the US. But of course, for a whole bunch of reasons, we have a lot of conflict over immigration policy in the United States and we’ve been stymied for any kind of major immigration reform for some time.

We’ve had Nick Bloom on here, talking about his work on how game-changing ideas are getting harder to find. And to find those, we need to put resources, including more researchers, into discovery. Is that a natural limitation on doing science?

Things like AI will help; it really depends on people. I don’t think we are actually limited in some fundamental sense. Innovation may be getting harder and there’s lots of evidence for that, but again, the US is putting less and less public dollars into R&D.

If we have a productivity growth slowdown, which we seem to have this century, and then we look and we say, “Oh, the government is not investing as much of the share of GDP as it used to. It’s half of what we did in 1980.” Well, maybe we should do a lot more there. It’s scaling funding into R&D.

I think we have lots of talented people who are not going into this space, and there are certainly lots of people abroad who, traditionally, are very eager to come to the United States and participate but cannot do so because of visa category limitations.

I wish science policy was just about coming up with a number. It would be a lot easier to analyze. But it seems like it gets very messy trying to figure out if we need to be doing something and making those changes, as far as the actual practice of science.

Yeah, the practice is quite complex, obviously, and there are effectively many government institutions that are involved. And they have very different kinds of grant systems — ways of reviewing, length of grants, amounts of grants. There are also a lot of philanthropists who are increasingly in the game of funding science in their own, rather idiosyncratic ways.

But, if you were to boil it down for me, I would say that I think diversity of approaches is really key. And the reason you want diversity is because, first of all, nobody has a crystal ball. If we knew what was going to happen when you started an R&D project, we wouldn’t have to do it. The whole point is that we’re going into the unknown. And so, in that sense, spread your bet. You’d be very surprised where some of the big insights come from and the spillovers that come and just go in unexpected directions.

When entrepreneurs try something and fail, they’re still celebrated for their risk-taking. But with government, we don’t seem to have nearly as much tolerance for failure. If we want to expand government research funding, will Americans have to become more tolerant of the failures that naturally come with bleeding-edge research?

Solyndra, of course, is the common example people like to throw around of a government investment misfire. But this is why I go back to salience: People often don’t see the benefits directly. These are very technical things that spillovers happen in unexpected and slow ways. So the public does not really appreciate that.

First of all, venture capitalists waste money all the time, because they know they don’t have a crystal ball. They spread their bets and they’re looking for the big thing, and they miss most of the time. And that’s okay, because that is the nature of R&D.

If you take every patent issued in the United States (this is a study that I did with a co-author), and you look to see what kind of science they build on — do they reference specific science or not? — you actually find that the vast majority of all scientific articles will flow through eventually into some patent. Maybe not directly, maybe not cited directly by a patent, but a science article is built on by another science article, and then that one is cited by a patent. There’s a process where basic flows towards applied and flows into the actual marketplace inventions that become goods and services in front of people.

We need the public to come to an orientation where they don’t expect to understand every detail — this is science — and don’t demand that you’re going to succeed when it’s impossible because there’s going to be failure and you need to fail. And the public needs to recognize that the stakes here are so high for our standard of living, our workforce, our health, our national security, that we can just go and make these investments that drive those things.

Some people think we’re progressing too quickly and we can’t control these technologies that we’re creating. How do you respond to those fears?

I think we do produce problematic technologies. There’s a double-edged sword to a lot of these things. But I think when you look and you step back, and you say, “Well, let’s look at human history and the human experience,” you realize just that on net, the positives that have come from this, historically at least, over a long history now, have really been enormously positive.

When I look at a technology that comes and that looks potentially problematic — which I might do with things like echo chambers in social media (it seems like a problematic aspect of the internet), and there are problematic applications of certain things in the context of weapons of mass destruction, there’s AI, and they have complex and problematic properties in some forms, if not in others — I think the answer to these things: If technology creates a new problem, the answer to that is going to be actually a new technology, a new insight that’s going to solve that problem.

So a different way to think about what science and technology do, in strictly applied applications, we’re trying to solve problems. If technology does good things, like the internet does a lot of great things, maybe it does some problematic things. Okay, so what are those problems? Figure out what those problems are, and let’s try to solve them. I think if you looked at the scope of human history, I’d want to bet on more technology, not on less technology.

Over at my blog and newsletter, I’ve written about the long bet between Robert Gordon and Erik Brynjolfsson about whether we’ll see faster productivity growth over the rest of this decade. Where do you come down on that?

I’m a tech optimist in the sense that I think there are all sorts of problems we know now that we don’t know how to solve. I think just in health alone, there are so many problems we haven’t solved. And we have many, many uncertainties and doors to go through, and biology is producing radically surprising insights into new tools all the time that are incredible. Space travel, artificial intelligence, any number of deep understandings of nature and reality that we still haven’t figured out. Physics is puzzling through very deep questions.

Now, I also think it is getting harder. I think that we’ve plucked the low-hanging fruit first. One thing I emphasize in my work is that there’s just so much we’ve already figured out, that to be an expert now, you have to be very narrow at the frontier. That may be why we have a productivity growth slow down. It is getting harder. But I’m an optimist because I think there’s so much more we can do and I think we will solve it. We’re not trying anywhere near as hard as we could. We’re leaving a lot of talent out of the game; we’re not investing anywhere near what we could.

To finish up, there’s a lot of interest in Washington about doing more on R&D and there’s an ever-involving R&D plan moving its way through Congress. What would be your policy advice on science and innovation investment in R&D?

I think that the more we can do through the NSF, the Department of Energy, or other ways, it will be to our collective benefit. So I think it’s a great investment. When I look at the numbers they’re talking about, I think it’s still very small compared to what could be done.

▶ How Americans Feel About Space – Kaleigh Rogers and Zoha Qamar, FiveThirtyEight | “In a July 2021 poll from YouGov/The Economist, a majority of Americans said the U.S. should send astronauts to the moon and Mars. This was true across political parties, with slim majorities for Democrats, Republicans and independents. Most Americans are on the same page about funding space exploration, too. About a third of Democrats, Republicans and independents said government funding of space exploration should be kept the same, and about 40 percent of each group said funding should be increased. Find me another issue where roughly the same share of Republicans and Democrats agree that the government is not spending enough money. Grandpa Buck would be happy to know that about three-quarters of Americans across the political spectrum — 77 percent of Democrats, 76 percent of Republicans and 75 percent of independents — also agreed that the effort made to land the first astronauts on the moon was definitely or probably worth it.”

▶ The Metaverse Will Reshape Our Lives. Let’s Make Sure It’s for the Better – Matthew Ball, Time | “The idea of the metaverse means an ever-growing share of our lives, labor, leisure, time, wealth, happiness, and relationships will be spent inside virtual worlds, rather than just aided through digital devices. It will be a parallel plane of existence that sits atop our digital and physical economies, and unites both. As a result, the companies that control these virtual worlds and their virtual atoms will be more dominant than those who lead in today’s digital economy. The metaverse will thus render more acute many of the hard problems of digital existence today, such as data rights, data security, misinformation and radicalization, platform power, and user happiness.”

▶ The wreck of Bidenomics – Noah Smith, Noahpinion | “America isn’t dealing with a crisis of jobs right now; we’re dealing with a crisis of costs, both in the macroeconomic inflation sense, and in the long term everything-costs-too-damn-much sense. Making life less expensive for regular Americans is a mission that fits the times we’re in. Government investment — and encouragement of increased private investment — has got to be a big part of this. High-quality infrastructure and public goods like R&D funding are important for boosting growth in real incomes. And a big push for abundant green energy — whose costs are already lower than fossil fuels and falling by the day — is key to making all kinds of things cheaper.

▶ The Blockchain Revolution: Decoding Digital Currencies – David Andolfatto and Fernando M. Martin, St. Louis Fed | “Cryptocurrencies and decentralized finance have grown considerably since the publication of the white paper on bitcoin in 2009. This article presents an overview of cryptocurrencies, blockchain technology, and their applications, explaining the spirit of the enterprise and how it compares with traditional operations. We discuss money, digital money, and payments; cryptocurrencies, blockchain, and the double-spending problem of digital money; decentralized finance; and central bank digital currency.”

▶ Is Inflation Still Low In The Digital Economy? – Marshall Reinsdorf, Progressive Policy Institute | “But one sector that has moderated rising prices is the digital economy. Prices for goods and services that are either entirely digital, like wireless phone service, internet services, and electronic information providers, or related hardware, like computers and smartphones, have either fallen or risen much less than prices in the economy overall. … [The] rapid development of new goods and services in the digital economy is hard for traditional measures of inflation to keep up with. New product models are not added to price indexes quickly enough to fully account for improvements to quality or reductions in prices. … These errors in measurement of the digital economy tend to overstate the overall inflation numbers. Correcting this mismeasurement of prices in the digital economy will require more frequent addition of new digital goods and services into price indexes.”

▶ Is the fear of an addictive metaverse overdone? – Elaine Moore, Financial Times | “In the four years I have been testing out virtual and augmented headsets, I have yet to try one that feels comfortable. ‘Like strapping a brick to your forehead,’ as one friend put it. It is possible to buy upgraded head straps that attempt to redistribute the weight, but even so the sets remain heavy. This would be easier to accept if there was more to do in the metaverse, yet after 30 minutes untethered from reality, I’m more than happy to log off.”

▶ On another planet: designing commercial space travel – Fred Scharmen, The Architectural Review | “In the immediate term, customers will continue to pay well for the privilege of taking the trip, and they will expect to be comfortable and pampered for the price. Broken toilets and cramped quarters will not be tolerated for long. Form matters here, almost as much as function. The styling of Tesla’s vehicles flatters the company’s customer base, reassuring the car’s driver and everyone around them that this is a high value artefact from the future, and that the object’s owner has both the wisdom to purchase for electric efficiency and the taste to make that decision look cool. The clean lines of SpaceX’s Dragon and Starship vehicles do the same.”

Learn more: The US Should Dramatically Ramp-Up R&D. And It Might Not. | “My God, It’s Full of Stars!” | What Happens When You Warn of Civilizational Collapse, but Civilization Doesn’t Collapse?

Weekly analysis from AEI’s Economic Policy Studies scholars

Thank you for your submission.