How will economists look back on the economy of the 2010s — as the longest economic expansion in US history, or as a period in which the US was stuck below three percent growth for ten years? And looking to the future, how might population growth, trade, Big Tech, and new innovations all affect America’s capacity for economic growth? Peter Klenow and I recently discussed this on last week’s episode of Political Economy.

Peter is the Ralph Landau Professor in Economic Policy at Stanford University, as well as the Gordon Moore Senior Fellow at the Stanford Institute for Economic Policy Research. In the past, Klenow served as a senior economist at the Federal Reserve Bank of Minneapolis. His work focuses on the macroeconomics of growth, productivity, and prices.

What follows is a lightly edited transcript of our conversation. You can download the episode here, and don’t forget to subscribe to my podcast on iTunes or Stitcher. Tell your friends, leave a review.

Pethokoukis: As we record this, we’re entering a new decade. Looking back, how do you think economists will remember the 2010s? As an impressive decade, since we’re in the middle of the longest economic expansion in US history? Or will they look at it as a disappointing decade in which the economy puttered along at about two percent growth, which is slower than the post-WWII average? How will they — and you — look back on the 2010s, economically?

Klenow: Well, cyclically, it looks great, like you said, in terms of the unemployment rate falling more or less continuously and employment rising. But in terms of incomes per worker, it’s been much more disappointing.

Looking back, we had that slow growth period after the oil embargo in the early 1970s — it lasted some twenty years. And then we had this freak period of ten years of rapid growth. And maybe we told ourselves when the economy then grew slowly from like 2005 to 2010 that that was partly the after-effects of the Great Recession and the global financial crisis. So I think we would be disappointed when we look back and say that the growth rate didn’t pick back up again, with productivity growth averaging something like one percent. So in terms of output per worker per hour, it’s been really disappointing I think, from the longer run perspective.

Given demographics, it’s sort of hard to grow fast like we did in the 1960s. People point to that time as an era of very fast growth, but of course we had a lot faster workforce growth and population growth back then than we do now.

And the demographic trends are pretty much baked in the cake, so isn’t the problem really the productivity growth that we haven’t seen pick up after that sort of spurt from the mid-90s to about 2004-2005?

Right. The direct effect of the slowdown in workforce growth — like you’ve said — slows GDP growth. But that doesn’t bother me so much because we have fewer people, so there should be fewer jobs. That’s a natural slowdown, just like Japan grew more slowly at least in part because of, in their case, maybe negative workforce growth. So those effects are relatively benign. But as you’ve said, the weak productivity growth is much more concerning.

Now, it could be that those two are linked. We’ve certainly seen a strong relationship between start-ups of firms and population growth. So just having more people means more people to operate businesses and run them and start them and work for them. So to the extent that innovation is associated with start-ups and businesses, then slower population growth might be slowing the supply of labor to start new enterprises and innovate.

And, of course, it slows the growth rate of the size of the market, which could also affect the profitability of innovation. So there could be some really interesting feedback effects from slow population growth to slow productivity growth that operate through innovation, not just through starting up businesses, but even through incumbent firms. If there’s less population growth, there’s less growth in the number of people to do research, which is why in some of your earlier podcasts and blog posts you talked about things like immigration of high-skilled workers and how that might help the US address its weak productivity growth.

If you read the business pages, it seems that for the past decade there were all of these amazing companies coming out of Silicon Valley that everyone had high hopes for. How many times have we heard the stories about the potential impact of smart phones, everybody has a supercomputer in their pocket? And people thought at some point we would see that stuff somehow result in greater and faster innovation, and that would transfer to productivity growth, which would transfer into faster economic growth and maybe offset those demographic issues.

But, is that not happening, or are we just not able to measure that it’s happening?

I think that it’s more than just the measurement problem, though I do think measurement is part of it. The government is well aware of this. The Bureau of Labor Statistics, for example, is responsible for measuring inflation that they subtract out of nominal GDP growth to arrive at a measurement of real growth that feeds into productivity.

So they’ve known for a long time that it’s really hard to measure the benefits of brand new products, or even improved versions of existing products. Their sense is that they might be missing half a percent per year, one percent per year. But it hasn’t necessarily accelerated, what they’re missing.

It’s sort of always at least for decades, kind of missed it by the same amount. So it’s not getting worse?

Yeah, they don’t see any evidence that it’s getting worse. It may have gotten worse in the IT sector and the heath sector, but gotten better in some other sectors. And it’s hard to estimate something you have a hard time measuring, of course. It’s like measuring the black market. But when measuring the magnitude of the understatement, we don’t find a lot of firm evidence that the problem’s gotten bigger.

So the economic impact of the internet and sort of the information revolution: Was it just a ten-year thing, and that’s it? You know, from the mid-90s to the mid-2000s, we had this higher growth, real things were actually created, and there was real productivity growth. Now, there are a lot of pessimists who say that the Internet was an important technology, but it didn’t match up to some of the great general purpose technologies of the past like electrification or the combustion engine. Does it turn out that the internet just isn’t as important for productivity as those technologies?

That could well be. I’m still hoping in the future, if you broadly include things like artificial intelligence and machine learning, that we’ll maybe get much bigger productivity gains in the future than what we’ve experienced so far.

I’m sympathetic to your statement that maybe we got a ten-year burst of productivity growth from it, but another thing to keep in mind is that this was still pretty good. If productivity grew at two and a half percent or more when it was otherwise on a one percent trend over ten years, that’s accumulating over 15 percent productivity level boosts. Though I grant your point that this would be much smaller than estimates of the productivity benefits of things like electrification.

But I would also say that this kind of assumes we would have hummed along at one percent growth. So maybe we would have fallen below one percent, and computers actually lifted us, not from one to two and a half percent but maybe a half percent to two and a half percent. Because we can’t quite see the counter-factual world of what would have happened if computers hadn’t been improving at the rate they were during the last generation. So it could be that they’re contributing more than fifteen percent cumulative over that time period.

And then there’s still the unmeasured growth. People are saving a lot of time because of a lot of things they can now do with their smartphone. That is worth something even if it’s not showing up in measured output per hour of work.

Do you expect us to get better at measuring some of this output that’s difficult to measure, or is that just always going to be a difficult thing to do?

I am kind of optimistic about that, because I feel like we’re in a golden age for economics. Now maybe people outside economics don’t see it that way.

Well, I don’t know much time you spend on Twitter, but according to Twitter at least, this has been a terrible time for economics — the profession needs to be blown up and rethink everything, it’s failed and now it’s time for a new economics.

Ok, so let me at least say why I think it’s a golden age. I mean this in terms of the amount of data we’re getting. Some of it is administrative data like tax records of companies and individuals, population censuses, and censuses of firms’ activities. So we’re getting much more data than we had before that we can use to understand how people behave and what phenomena are happening. That could be used to measure the benefits of new technology.

I also think it’s a golden age because computing power has enabled us to do modeling better and understand complex aspects of the economy. One of the criticisms of economics that I think is rather out of date is the idea that we aren’t aware that there’s lots of heterogeneity out there in terms of workers, firms, and consumers. For example, simply saying that free trade is good misses all kinds of suffering that happens as a by-product. So because of better data we’re much better at measuring that and much better at coming up with ways to think about how to address it and predict what would happen if we, say, expanded trade adjustments.

And there’s a third strand of this golden age — related to the most recent Nobel prize — which is running experiments. Some of those are natural experiments, and some of them are deliberate policy experiments by researchers or NGOs or some combination. Those are enabling us to get at causal effects of policies much better than we could before. That’s why people have talked about a credibility revolution in economics, where we’ve got more random variation we can use to understand welfare effects of policy changes.

Those are the forces that make me think we’re in the middle of a golden age in terms of economic research and our understanding of economic behavior.

Getting back to the productivity issue, are big technology firms — whether it’s Apple, Amazon, or Google Alphabet — are they helping productivity or are they hurting productivity? People make the case for both.

On one hand, these are certainly very productive firms, and they spend a lot on research. But on the other hand, critics say, “Well, they’re so big that they’re preventing new competitors from starting. Nobody wants to go against them, they buy up these small companies before they can get big, so they’re part of the problem.”

So are Big Tech firms part of the solution or part of the problem, at least if you want faster economic growth?

I think, at least in a direct effect, I’m very confident they’re part of the solution, at least in a sense that the innovations they’ve carried out have benefitted people enormously. Not just their stockholders, but consumers as well.

I think the harder question is what you also touched on, which is whether their ascendancy and their success in terms of their innovation, whether that’s made it more difficult for further innovation to occur, because they have so much market power, or maybe even in terms of Facebook having a network externality, or even Amazon’s marketplace could be viewed as a network that’s hard for some other online retailer to really crack. Or Uber for that matter.

We can talk about network effects or incumbent advantages that could deter future innovation. So I am concerned about that, but that’s much more speculative. So even though I have some research pointing out that is a possibility, and other people do as well, I think that’s more conjecture at this point. We’re more kind of flailing, trying to understand why productivity growth is so weak. One natural candidate is that concentration and market power at the top has increased, and maybe that’s deterring future innovation.

To me there seems to be research on this topic, and there’ll be more research, but activists and policymakers have already seen enough. They’ve already decided that these companies are bad for innovation because they’re suppressing new companies from rising, so people don’t even want to start the next great company if it’s in a sector that those companies are involved in.

So what advice would you give policymakers about this sort of big company, and specifically the Big Tech company issue?

I think that’s a great question. I wish I could tell you exactly what to do, and I guess I don’t know, basically. What I think is hard basically is, I’m tempted to say, “Well, we need to understand it more before we take any action like breaking up one of these big firms.” But that kind of gives a default to no intervention, no regulation. That makes me uncomfortable as well. Why is that the default? We know that the growth rate is down. This is one candidate to explain why, so maybe we should take action.

And I guess one of the biggest difficulties when it comes to something like breaking up a big firm is that I really want to have evidence-based policy, I really want to use systematic evidence to decide what will improve economic outcomes. In these cases, it’s hard to have systematic evidence because you’re talking about a few big players. It’s not like you have a lot of different experiments you’re using to understand.

Here’s an example of something we understand better: the US raised the threshold for anti-trust examination of mergers from $10 million to $50 million and there was this surge in mergers that was between values of $10 and $50 million. And there were horizontal mergers of competitors — they weren’t vertical mergers of buyers and suppliers. So it really looks like anti-competitive behavior.

I would be much more comfortable looking at those cases because there’s so many of them and we can study what did the change in that threshold did to competition in markets that were affected versus those that were unaffected. It’s just much harder to do this systematic analysis when you have a huge player like Amazon or Apple or Google.

If you’re worried about competition in the economy, a lot of talk has been about the very largest tech firms. But if you were a regulator that worked at the Justice Department or the FTC, is that the first area you would look at, or are there other areas that aren’t as sexy because everybody’s involved with Amazon and we love reading stories about Apple? Are there other areas that you would focus on first?

Actually, yes, because we looked at the rise in concentration, and it’s most pronounced in sectors like services. So think of like Southern Health buying up a lot of different local health providers. So you see it in the health care sector, that rise in concentration. You also see it in retail. So services broadly — and retail wholesale — are the sectors where you see the biggest increase in national concentration.

At the same time that I say that though, one of the other things that people are finding is that local concentration isn’t necessarily rising. This might be a kind of corollary to what people found where airlines were deregulated. There was this big shakeout of national airlines, a lot fewer national airlines and national concentration, and airline market share went way up. But they were competing a lot more head to head on individual routes, so you actually had more competition route by route.

Something similar might be going on in these other sectors: national concentration is up, and at the same time local concentration is down. You know, Wal-Mart is facing off some other big retailer more than it did in the past. So it’s in more markets, but it’s also facing off against other non-trivial players. So that’s an example where concentration might be up and is a source of concern for regulators, but they also have to really look at whether mark-ups and market power are up, not just concentration.

We certainly want to have a highly competitive economy where companies rise, but then if there’s a company that does it better they can rise and their founders can get rich — this process of creative destruction.

You’ve written about creative destruction, but also about another kind of innovation, which you call “iterative.” That’s where you’re not creating some brand new product, you’re just making that product better. And they both have a sort of role to play, right?

Definitely. It’s one of these things where, if you just try to look at it through introspection to figure out which is more important, you can’t really come up with an answer. Because it’s easy to think about existing incumbents like coming up with new versions of iPhones or car models every year. And those look like important improvements — they might not be as dramatic as the first iPhone or the first mini-van —

I think of the third camera lens on the iPhone versus the second camera lens.

Right. So it’s easy to look at them individually and they don’t look as transformative as the first smartphone that came in where Apple grabbed a bunch of market share at the expense of Blackberry or Nokia. It looks like those are bigger innovations, but on the other hand, there are tons of these smaller ones happening all the time at all of these firms, and so one can’t use introspection to try to figure out which is more important. It’s too easy to just tell stories like, “Well most R&D and patenting is done by incumbent firms that are pretty large, so maybe they do most of the innovation.”

Now I think that the smaller firms are much less likely to report doing research and development, and a lot of these sectors where this innovation is occurring don’t really report R&D at all. Wal-Mart doesn’t really report doing R&D, for instance. And very few firms patent, so I don’t want to use patenting and R&D as some end-all be-all measure of innovation. But if you did look at that, it certainly looks like big firms and incumbent are very important for innovation.

So ex ante, it’s hard to just decide what we want. Do we want these big, national firms that do a lot of research? Or do we want lots of entry and exit? I think we want both, but I think one of the keys is that —even though by my estimates with some research I’ve done with colleagues, we estimate the majority of growth comes from incumbent research —maybe one of the reasons they keep doing research is that they fear creative destruction from a competitor. If Apple was very comfortable with its market share and felt like it was impenetrable, it might not continue to innovate as much as if it feared losing its market completely.

So creative destruction may play a one-third role directly, but then also play a very important role at spurring incumbents to continue to innovate, continue to make what they do better instead of resting on their current products and services,

And is it different if you’re in an economy, like the US economy, which is supposed to be on the leading edge, the technological frontier? We’re supposed to be coming up with the new inventions and the new way of doing things. Is that role of creative destruction more important for us than some other kind of economy, which is may be a perfectly kind of economy but isn’t the one really generating a lot of new technological progress?

I actually might push back on that and say that countries that are outside the OECD tend to have lower allocative efficiency, meaning they don’t seem to have capital and labor flowing to the highest use as quickly as places like the United States. The United States is kind of like a ruthless, capitalist economy where firms are driven out. Amazon driving out lots of mom and pop retailers or brick and mortar retailers more generally — not just small ones but big ones — the US economy has a lot of that.

Rich countries in general have more of that than some middle-income countries. They tend to, and even some rich countries like Japan, protect a lot of its retail sector and its distribution sector from foreign direct investment. I think Japan’s really shooting themselves in the foot in not allowing that creative destruction that comes from having new foreign retailers come into their market. Mexico opened up to a lot of foreign retailers and there were big gains, on the order of three to seven percent income gains over a decade for the average person from being able to shop at these big box retailers they didn’t have access to before.

So I think creative destruction is in a way being suppressed more in middle income countries and developing countries than it is in rich countries. I’d be more concerned about it there from what I see. You have a lot more monopolies there that should be facing more competition. They should have more entry that they have to deal with.

What would you like policymakers to understand about trade and how it impacts innovation and growth? Obviously, it seemed like we had sort of decided that trade was good, the more free trade the better. Now people don’t seem to be quite as sure anymore. So what would you want policymakers to understand that maybe you think they don’t get about trade?

I would say that trade is a positive-sum game. I don’t think about it like a global Parkers Brothers game, where there’s a winner and loser and a fixed amount of resources. And I think this is classically described as “cloth for wine” but there are dynamic benefits that are even bigger. And by dynamic benefits, I mean, for example, Japan comes up with lane manufacturing and US auto firms imitate it. US firms come up with some leading edge semiconductors and South Korean firms imitate it. So products and ideas are flowing back and forth across countries in a way that is spurred by trade. That’s because when you trade, you can see the people figure out how products are used and then start to figure out how they can develop a better version that serves their purposes even better.

So I think more trade actually spurs more ideas flowing across countries, and that’s a very positive-sum thing. So in the same way I was saying the US could benefit from high-skilled immigration that could do research here, we could benefit from research being done in their countries of origin.

I think the rise of Western Europe, Japan, and Korea has been a boon to the US. There’s a channel through which it’s been a boon, even as there’s been a bunch of people who have suffered. So I think the thing I really want to communicate to policymakers is not to think about this as “This is our national champion versus their national champion, so either we win and they lose or they win and we lose.” We can both succeed at this with an even playing field, good property rights, trying to prevent countries from stealing intellectual property but also subsidizing some firms over others, as well as with policies like what you could implement through things like the World Trade Organization, you could make this a positive-sum game with big benefits.

Do you worry China is going to make the big breakthrough in AI, or the breakthrough in 5G, or the breakthrough in some other technology? Do you worry that they’ll have the technology and we’ll be behind and they’ll be the leading economy and we’ll be the laggard economy? I think that’s a really natural way that a lot of people think about it: Some country comes up with a technology so they’re the winner — and if it’s a big technology then they’re the big winner — and then all the other countries have to buy that technology from the country that originated it, and they’re the followers and no longer the leader.

I think we should worry to the extent that there’s any intellectual property theft. And I think part of what you’re saying is being a leader has some huge benefits associated with it, so of course we want to be the leader. That’s why I’m emphasizing the level-playing field. If they’re the best at coming up with the next generation of cell phone technology, for example, or wireless technology, then at a level playing field and enforced intellectual property rights, that’s going to benefit US consumers more than if the US put barriers on importing or using those technologies from foreign firms.

Now I also have to add another cautionary note there. I’m talking about narrow economic considerations. I’m not considering any broader national security or privacy considerations that obviously need to be factored in, but that I know very little about.

I always think of that 1980s Clint Eastwood movie called Firefox, where the Soviet Union had developed super-advanced airplane technology and therefore the US was at some sort of unbelievable strategic disadvantage so we had to go steal their technology so they didn’t win the Cold War. I think a lot of people think there will be some massive leap forth in technology that will put us at a massive national security disadvantage, though I understand that you’re looking at it more from an economic perspective.

Sort of winding down here, one criticism of economists is that you’re just worried about GDP growth. That the number one goal of economic policy is to generate fast GDP growth. Is that an accurate criticism?

I don’t think so. I think economists are incredibly narrow — so that’s a fair criticism — in the sense that we think about people as maximizing some utility and that we often have a pretty narrow notion of utility that involves utility from consumption, leisure, and maybe over time and maybe for your family. So clearly that misses a lot of things that make life worth living, to borrow one of my favorite scenes from Dead Poets’ Society. I think the line in that movie was something like, “Business and law” — and presumably economics as well — “are really important for sustaining and promoting life, but they’re not necessarily describing why we want to live.” So I think that broad criticism of economics is fair.

But I think the specific criticism that we fetishize GDP growth is more narrow than we are, it’s too specific and too narrow. We have this view that it would increase GDP if you banned retirement, but we think that would lower welfare. Or if you had forced labor of any kind, that would increase GDP but that would be a bad thing. So we’re thinking of welfare in terms of optimal trade off of things like leisure and retirement versus work and consuming now versus consuming later. So we’re not so obsessed with promoting GDP later at the expense of leisure or consumption now, or aspects that people might care more about now. We’re less one-sided in favor of GDP growth than people usually think.

Well, the US economy has been growing at three percent, maybe more, since WWII. More recently, it’s grown more at about two percent, and I think that’s roughly the sort of long term forecast by the Fed or the CBO.

But let’s say they were wrong, and the US economy over the next twenty-five years grows at one percent. What does that look like? The US economy has never had the economy grow that slowly for that long. What does America look like after 25 years of one percent growth versus 25 years of two percent growth, or does it make much of a difference?

I think we experienced for most of the last forty years one percent growth, measured in productivity terms, per worker hour. Maybe you’re saying we’ve sustained faster growth in part because people were working more. So I guess in that sense we have been experiencing it.

I should have emphasized something else when you were talking about narrowly focusing on GDP growth: Another key thing which would become more prominent if the pie is expanding at a slower rate is being concerned about inequality. Inequality is a first-order thing. So when you describe the growth rate going from two to one percent, that’s enormous for progress and human welfare, but another thing we can do to product human progress is to reduce inequality. Now, doing both of those are difficult — both promoting growth and reducing inequality. But economists don’t obsess with GDP growth at the expense of saying, “Well inequality, don’t worry about that.” We’re very concerned about that.

So something like artificial intelligence, which might cause the growth rate to accelerate but at the expense of rising inequality — economists would be very concerned about making sure that the growth that this generated was redistributed, and distributed more equally to prevent inequality from soaring.

So I guess I’m back to your direct question. I do see the pie expanding more slowly, so we’ll naturally be more concerned with inequality. We should be concerned about it anyway — we should be concerned whether growth accelerates or stagnates — but I can see it becoming a more prominent issue if growth is slow.

Is it possible we get an economy that’s this big, given the demographics, to grow at three percent going forward? Whether it comes from AI or robotics increasing productivity, is a three percent economy still possible for the United States?

I think it is, but it’s going to have to come through some sort of technological revolution, like you’re mentioning. Because if I think about the growth rate of resources we can devote to research, even if we allowed twice the H1B visas, that’s not going to double the growth rate of research activity, in a way that could feed the growth rate and push it up to two or three percent in terms of productivity. So where I think the acceleration would have to come form is a rich vein of technological progress in things like AI and machine learning.

You know, it’s somewhat paradoxical, because a lot of people are fearing it, and I may sound like this completely out of touch academic professor when I say, “Hey, this is a great opportunity here! This is like the tractor hitting farming. Yes, it wiped out massive numbers of jobs in farming, but if we can harness it to benefit the population more broadly, this could be a great thing. This could enable us to enjoy more leisure while getting the same goods, reducing pollution, addressing all sorts of resource problems.”

I think people are just focusing on the downside now. I think so much of what we hear about technology, people just think, “That costs jobs. That’s job-replacing, it’s not job-enabling, it’s not job-creating.”

Yes, and I’m a cautious — or conditional — optimist in the sense that, as long as we handle the distributional issues, because there’s going to be lots of pain and disruption associated with this, but as long as we handle that intelligently through things like wage subsidies and rising sectors that make it easier for people to transition, then this is something we should welcome, rather than be fearful of. But I understand that people who have been hammered by the decline of manufacturing might be suspicious of people telling them to be optimistic about this.

Peter, thanks for coming on the podcast.

Thanks for having me! This was really fun.

This article by James Pethokoukis first appeared at the American Enterprise Institute.

Image: Reuters.