Robert Solow on Growth and the State of Economics
Oct 27 2014

Robert Solow, Professor Emeritus at Massachusetts Institute of Technology and Nobel Laureate, talks with EconTalk host Russ Roberts about his hugely influential theory of growth and inspiration to create a model that better reflected the stable long-term growth of an economy. Solow contends that capital accumulation cannot explain a significant portion of the economic growth we see. He makes a critical distinction between innovation and technology, and then discusses his views on Milton Friedman and John M. Keynes.

RELATED EPISODE
Robert Barro on Growth
Russ Roberts interviews Robert Barro, Harvard University Professor and Hoover Institution Senior Fellow, on the economics of growth, what the developed world can do to help poor people around the world, and the role of US assets and the dollar...
EXPLORE MORE
Related EPISODE
Joel Mokyr on Growth, Innovation, and Stagnation
Joel Mokyr of Northwestern University talks with EconTalk host Russ Roberts about the future of the American economy. Mokyr rejects the claims that the we are entering an area of stagnation or permanently lower economic growth. He argues that measured...
EXPLORE MORE
Explore audio transcript, further reading that will help you delve deeper into this week’s episode, and vigorous conversations in the form of our comments section below.

READER COMMENTS

Nyle Kardatzke
Oct 27 2014 at 12:18pm

I haven’t listened to the podcast yet, but I did Google Professor Solow because I have heard his name and read some of his papers for over forty years. He is now 90, and is still active.

When I first studied economics in the mid-1960’s at Ball State University in Muncie, Indiana, one of my professors pointed out the longevity of many economists. I have verified that in several Google searches and now feel entitled to a lifespan of at least 90 years. That’s not the main reason I stayed in economics, but it’s a nice fringe benefit.

Greg G
Oct 27 2014 at 2:11pm

There might be something to what Nyle says. This reminded me of the interviews with Friedman and Coase. Each was astonishingly sharp for his age – or for any age. Many people are struggling to remember the names of family and friends at their age.

I think of myself as fairly old and well past my prime and Solow was a professor at MIT before I was born. This gives me hope for additional mileage.

Today’s podcast was a great example of Econtalk at its very best. I listened to it twice already. As usual Russ did a great job of exploring ideological differences in a way that was lucid, polite, very entertaining and entirely accessible to the layman without being dumbed down at all. Bravo.

pcle
Oct 27 2014 at 6:04pm

Russ good job. I only wish you had asked Solow about Weitzman’s “hybridizing growth theory” which actually looks far more interesting than Romer’s work.

ECharles
Oct 28 2014 at 8:32pm

Solow sounds amazing for 90 years old.

SaveyourSelf
Oct 29 2014 at 4:00am

Robert Solow is formidable. I am still rather awed by the man. His comments on the limits of knowledge in an impossibly complex world could not have found a more favorable audience than Econtalk. His four reasons why macro economics is now and will always be contentious were well thought out and expressed. His views on Milton Friedman, though, were startling when compared to the rest of the interview.

I read through the “Why is there No Milton Friedman Today?” commentary. Like that portion of the interview with Russ, it lacks ANY specifics. It has the flavor of Fox News—artful slander without any substance. That article and his conversation with Russ both give the impression that Milton frustrated him in debates. It is telling, though, that he takes time in the article to cite John Maynard Keynes as Milton’s ideal foil while off handedly hinting that Milton is probably closer to Marx that Keynes.

I suspect that Solow, at 90, is still frustrated at Friedman for successfully opposing the ideas of Keynes, his idol. I further suspect that, when he says Milton was different than other economists because of, “the fact that, in my mind, he was not given to doubt.” (38:00) he is really voicing his frustration that Milton Friedman did not seem to acknowledge any of Keynes Genius. I think it was Milton’s opposition to Keynes, not opposition to Solow’s own ideas, that frustrate Robert Solow to this day because he seemed so genuinely humble throughout the rest of the interview. Additionally, I suspect that, were he not a Nobel Prize winner, Solow’s opinion article would have been unceremoniously censored by the Econtalk logic fallacy and slander filters for multiple violations of both 😉

@Greg G

Greg G wrote, “Today’s podcast was a great example of Econtalk at its very best. I listened to it twice already. As usual Russ did a great job of exploring ideological differences in a way that was lucid, polite, very entertaining and entirely accessible to the layman without being dumbed down at all. Bravo.”

Well said, Greg. I second that. Though I still don’t know what growth theory is about. Perhaps they could have dumbed it down just a little for me ;p

Greg G
Oct 29 2014 at 8:26am

SaveyourSelf,

I don’t see Keynes as being as incompatible with Friedman as you and Solow (and most other people) do. Two of Keynes’ big ideas were that macro analysis is crucial to good public policy and that government should manage aggregate demand. Friedman played an crucial role in popularizing both those ideas. Friedman wanted AD managed with monetary policy, not fiscal policy.

To be sure, they disagreed about fiscal policy. My own view (which certainly cannot be proven) is that Keynes would have been very influenced by Friedman’s work if he had lived long enough to see it. Keynes was famously willing to change his mind. Hayek was constantly complaining about that.

I thought it was Friedman’s unwillingness to change his mind that Solow saw as his worst quality. It is certainly true, as you say, that he might have been frustrated by Friedman in their public debates. No one was a more formidable debater than Milton.

Danny Nhepera
Oct 29 2014 at 12:57pm

I’m only a 23 year old masters student in development econ and names like Solow are just mythical to me and most people my age. we just read about them but much like the renaissance artists their work is all we can ever hope to know about them. thanks for bringing this name to life, never would have imagined i would hear a man whose given me headaches in exam halls actually speak. cheers for that.

D. F. Linton
Oct 29 2014 at 1:18pm

I don’t think the idea that Friedman was unwilling to change his mind has any real basis. One example is his position on the optimal monetary rule. On that issue he changed his mind a number of times.

Perhaps, it nothing but a variant on “I’m principled; you are stubborn; he is pig-headed.”

I found Solow’s critique of Friedman as overly confident and too interested in “marketing” his policy views, too precious for words coming from an admirer of John M. Keynes.

Overall a great podcast, but I’m still not clear whether Prof. Solow believes growth does or does not require savings. Early on he said the savings rate determines whether the growth path was high or low, but later seemed to discount the need for savings. Then during his discussion of the age distribution of capital goods, I kept waiting for even a mention of the need to replace dying capital from savings, but never did. Was his growth theory just lipstick on the “Paradox of Thrift” pig?

Kyle D
Oct 29 2014 at 6:21pm

I thought the moment where Russ asked about how technological innovation was achieved was going to be a golden moment yielding the answer to all of our economic problems (although answers in economics typically amount to trade-offs). I was disappointed when Dr. Solow did not give any complete answers, nor did he directly refute the idea of deferred consumption and investment in future production as a means of achieving innovation. I’m a rank amateur at this econ stuff but can someone explain how deferring consumption of present resources in order to allocate and invest capital in production to yield greater future resources (ie greater output per unit input) is not, at least, one method of achieving innovation?

pe
Oct 30 2014 at 9:44pm

……but can someone explain how deferring consumption of present resources in order to allocate and invest capital in production to yield greater future resources (ie greater output per unit input) is not, at least, one method of achieving innovation?

Russ can correct me if Im wrong but I believe in the growth theory of Lucas-Uzawa this is the case.

SaveyourSelf
Oct 31 2014 at 7:47am

SaveyourSelf wrote, “ I still don’t know what growth theory is about.”

Danny Nhepera wrote, “…names like Solow are just mythical to me and most people my age. we just read about them.”

  • Danny, after your comment I went looking for Growth theory in my Micro textbooks and couldn’t find it. Out of desperation I dusted off my old Macro textbook and discovered a whole chapter on it. (Mankiw, N. Gregory. Macroeconomics. 1997: 80-113) Thanks!

@Kyle D “…can someone explain how deferring consumption of present resources in order to allocate and invest capital in production to yield greater future resources (ie greater output per unit input) is not, at least, one method of achieving innovation?”

  • So, according to the Macro textbook cited above, “deferring consumption of present resources” is called “saving” and all savings are used for investing. Investing, from what I can tell, is productive output used to buy new capital or maintain old capital. Anything not saved/invested is consumed.
  • Once you select a rate of saving, that determines the level of capital available for productive activities in the future. Because capital depreciates, at some point a given level of savings will eventually lead to a steady state amount of capital. The steady state amount of capital is that point where ALL saving goes to maintaining a given amount of capital and no savings are left over to purchase new capital.
  • Now, Solow’s model “…assumes that the production function has constant returns to scale.” The textbook explains that it is convenient to think about “Production functions with constant returns to scale” at the individual level rather than the aggregate level because at the individual level “output per worker depends only on the amount of capital per worker” (82).
  • So, because a given rate of saving/investing will eventually lead to a steady state level of capital and individual-productivity is entirely explained by the amount of capital available to each individual, then an individual’s productive output will also eventually reach a steady state. Since a steady state is not a growing state, the level of saving/investing cannot explain growth in productive output over time.
David Zetland
Oct 31 2014 at 9:39am

Wow. I’d LOVE to be so cogent at 90 years old!

More interesting, I always thought (and wiki seems to support: https://en.wikipedia.org/wiki/Solow%E2%80%93Swan_model) that Solow had “missed” technological change. That seems to depend on your interpretation of “capital,” which he explained as “capital in use” — i.e., a definition that can include improved techniques for using fixed capital. Perhaps he’s revising with hindsight (to Romer), but knowing that Schumpeter was an adviser makes me think that he had, indeed, allowed for innovative combinations of capital and labor. If so, I wonder if (1) that was clear or implied in his papers and (2) why endogenous growth theory (same idea) was even an “event.”

Or perhaps I missed it.

Kyle D
Oct 31 2014 at 2:21pm

@SaveyourSelf,

Thanks so much for the response. Very helpful. To follow up: Wouldn’t any investment beyond steady state yield a potential source for innovation? That is to say, after expending resources to maintain capital stock, any foregone consumption that is directed towards investment of new and improved capital stock will have some probability of achieving innovation.

SaveyourSelf
Nov 1 2014 at 9:05am

@Kyle D “Wouldn’t any investment beyond steady state yield a potential source for innovation?”

  • So you are proposing a situation exactly similar to increasing the saving rate after a steady state has been achieved. In that case, the additional savings would buy more capital which would increase the costs of maintenance on capital repeated over and over until a new steady state is reached. No innovation triggered.
  • I suspect the reason you think buying new capital might lead to new innovations is because, in our world, new capital is so often better than old capital. My first Blackberry was simpler, slower, and had a shorter battery life than my present Galaxy-Note. Lots of improvements have occurred over time. I can do a little more on my Galaxy than I could on my Blackberry. I am a little more productive thanks to those improvements. But think about the question from my individual standpoint: does my personal savings rate have anything to do with those changes to my phone? If I doubled my rate of savings, will that make it any more likely that my next phone will be better than my current one? Where is the connection? Dr. Solow’s growth model suggests there isn’t one.
SaveyourSelf
Nov 1 2014 at 10:33pm

@Russ Roberts. Around 11:07 you said, “…where does technological progress come from?” “If you have no investment, if you have no savings, if all you do is consume your current product, it seems difficult to shift the production possibility frontier or to change the production function…”

Dr. Roberts, it would give me immeasurable pleasure if I could provide you a satisfactory answer to this question.

One of my criticisms with the Solow growth model is that it treats all Capital the same. Even though cash, stocks, a wrench, a car, an oven, a robot, and probably my college degree are all “Capital”, dealing with such wildly different tools as if they are all the same seems…imprecise.

Another criticism of the growth model is how it conceptualizes changes in technology. Dr. Solow in the podcast talks about the 3 major variables in the growth model: Labor, Capital, and Technological-change. But he might as well have just said Technological-change was the only variable in the growth model because, “when I [Dr. Solow] analyzed those time series, those data, it turned out that what the history of that period seemed to be saying was that in fact, almost all–more than 80%–of the long-term average growth of output per person, or income per person, could be traced to this broad sense of technology progress.” (06:22)

Over 80%! Just throw out 2 of the 3 variables and you would have a nearly identical outcome! Perhaps that is why the variable for Capital can be so wildly imprecise. It does not matter one lick! Neither does labor, for the most part. What the model is really saying, therefore, is that “technological change in its broadest sense” = “growth”. Why not just say “change in technology” = “change in productivity”? Perhaps because changes in technology do not always improve productivity. So instead let us say that “change in technology” is somehow related to “change in productivity”–stating a correlation but making no claim regarding causality or the direction of causality. But that still does not answer where the changes that improve productivity come from, or does it?

Change is everywhere and always. It is ubiquitous. Even objects at rest are constantly changing relative to almost everything else in the universe. The real question, therefore, is not where the changes that improve productivity come from. The real question is what mechanisms exist that select for the changes we desire? The answer, of course, is our desire.

In an ideal-market system, desire is often expressed through currency. Income, therefore, is a signal to producers that people desire recent changes to continue. When income is greater than costs, a profit is produced, which “signals” that those recent changes are more desirable than other alternatives. The profitable changes survive whereas the unprofitable alternatives do not. “Growth” is not a mysterious, insightful force pushing good ideas forward from the present to the future [investing]. It is a selection-bias PULLING good ideas along from the past to the present [where the “bias” in selection-bias is human desire.]

SaveyourSelf
Nov 3 2014 at 2:58am

There are many practical consequences of a “selection-bias” model of growth.

1. “Change” is random and diffuse and nearly always decreases productivity. Rarely, though, a change occurs which increases productivity. These rare but desirable changes may occur anywhere, anytime, and to anyone. The greater the number of people who are actively alert for those positive changes and able to respond to them, the more likely they will be discovered and developed and expanded for the benefit of the rest of society. Thus the base level of education in a society may influence growth rate.
2. The ever-present desire of human beings to improve their standard-of-living is the energy that drives the search for improved ways to live and work. The freer people are to act on that desire, the higher the expected growth rate.
3. Societies with strong religious convictions, customs, laws, legislation, or norms opposing actions and thoughts aimed at improving standard-of-living will have lower growth rate when compared with economies that accept or encourage thoughts and activities aimed at improving standard-of-living.
4. Societies with many small businesses spread out geographically will have higher growth rates because the skill set associated with developing and marketing new ideas will be more diffuse, increasing the likelihood that the person who discovers a beneficial change will have access to people and tools necessary for diffusing that information to the rest of society.
5. “Adoption lag” is the time it takes for technological improvements to be integrated in the structure of society. Younger generations have far less difficulty adopting new technology than older generations. Thus, rapid population turnover and high consumption among young people may lead to higher growth rates.
6. After a technological improvement is discovered, it must be nurtured and shared or else risk being forgotten and overlooked. Energy, capital, and time is required to nurture those changes and expand them throughout the society. Thus societies with easy access to capital will have higher growth rates.
7. Prices carry the desires of consumers to producers. Noise in price signals can lead producers to misunderstand consumer desires and thus behave unproductively. So erratic changes in the money supply and price controls both decrease growth rate.
8. Most changes, positive or negative, will not occur in a time and place or to a person capable of recognizing and developing them. If a society has a long memory, however, that memory can serve as a treasure trove of productive insights. Thus the length of a society’s memory, and more importantly the tools it has for searching and sorting through old data, will influence growth rate positively.
9. An absence of security or justice in a society will reduce growth. Some negative outcomes are the result of indirect adaptations to a violent environment like the inability to specialize outside of violent skillsets. The lack of specialization will reduce the ability of the society’s members to recognize or respond to positive opportunities for change. In addition, violence has direct, immediate negative effects on growth and productivity. Dead and injured people are less productive than healthy, living people. Violence also undermines the reliability of price signals, rendering them less effective for communicating desires between strangers and therefore less likely to help producers match their activities to the desires of the rest of society efficiently.
10. Competition in markets carries with it a reward system for rapid recognition and adoption of new ideas. Competition also punishes individuals who are unwilling to adopt improved production techniques. Competition in a society is predictive of higher growth rates.
11. Monopoly carries with it incentives to actively stifle innovation and even punish innovators. It is new ideas, after all, that undermine the position of the monopoly. Monopoly, in all its many forms, is predictive of low and even negative growth rates.
12. “Catch up growth” occurs in underdeveloped countries because the trial and error associated with identifying, testing, nurturing, and integrating technological change has already been done by someone else. All that is left to do is adopt those changes. Societies can display catch up growth without necessarily having any of the features that correspond to a high growth rate. Once catch up is achieved, though, those societies bereft of freedom, competition, security, justice, and reliable monetary prices will demonstrate an alarming freeze in growth.

SaveyourSelf
Nov 4 2014 at 4:26am

It appears I was mistaken.

In my previous post, point 6 above, I stated that “Societies with easy access to capital will have higher growth rates” thinking that technological improvement would need to be shared for a society to benefit enough for the innovation to show up in the growth rate. I later remembered that Dr. Solow said savings/investment did not affect growth rate when he looked at real life date. So I went looking for other real life studies. I found this one titled “When Does Domestic Saving Matter for Economic Growth?”– http://www.nber.org/papers/w12275.pdf–which also found that savings did not affect growth rate in wealthy countries. Interestingly, they found savings did positively correlate with growth rate in poor countries.

So I am accepting that savings/investment do not affect growth in the USA. But that means there must be a low cost way that wealthy countries are disseminating improved technology techniques that poor countries lack. After thinking on it a while, it occurred to me that a healthy price signal could serve this function.

Reasoning: If one producer discovers a new technique that reduces costs, he is able to lower his price below his competitors. Other producers will take notice of their competition’s advantaged price position and spend some time and money investigating. Once they discover the new technique, they will adopt it or else go out of business. Thus, with zero cost to the original producer or to the greater society, the new technology is disseminated. Although it is true that the inventor’s competitors will have to spend time and resources to uncover the new technology and then implement it, doing so is in their best interest because these short term expenses are much lower than the long term alternative—bankruptcy.

The economists in the article I cited above thought the poor countries might show a relationship between growth and savings because the increased savings attracts more foreign investment. I find their idea plausible, but building on my own guess about prices automatically and inexpensively disseminating breakthrough technology in wealthy countries, I postulate the poor countries lack healthy price signals in their markets–which probably would explain why they are poor in the first place. Saving up money to buy technology from wealthy foreigners must be a less efficient work-around to reliable price signals.

Comments are closed.


DELVE DEEPER

EconTalk Extra, conversation starters for this podcast episode:

About this week's guest:

About ideas and people mentioned in this podcast episode:Articles:

Web Pages and Resources:

Podcast Episodes, Videos, and Blog Entries:


AUDIO TRANSCRIPT

 

44:51

Time
Podcast Episode Highlights
0:33Intro. [Recording date: September 29, 2014.] Russ: You're often credited with launching growth theory, a claim you modestly dismiss. But, talk about your papers a long time ago in the 1950s, that at least launched, if not growth theory, your contributions to that theory. What were you trying to achieve, and what can we learn from that work? Guest: I think the thing to remember, if anyone can remember that far back, is that in the aftermath of the Second World War, questions about growth, about growth in various parts of the world, were in everybody's mind. Now, some of that became talk about economic development, and I, as a young economist, had no real understanding of what went on in underdeveloped countries. But I was interested in what governs--like and I was interested like everybody else--in what governs the long-term growth of an economy and why some industrial economies grow faster than others. And what there was on offer in economics was the work of Roy Harrod and Evsey Domar; and the way they had developed an understanding of economic growth had a very peculiar aspect to it, which was that the long-term growth path of an industrial economy tended to be unstable. Intrinsically unstable. And I'm not speaking now just of the fact of business cycles that everyone accepts, but the notion that the reasonably smooth growth was not an attractor for the economy. It didn't tend to be a normal state of affairs, but something that if you got anywhere near it, most of the pressure was to move away. And that seemed to me not to characterize the long-term history of modern industrial capitalism. Even the Depression of the 1930s was a depression. And it ended in the War for perfectly understandable reasons. So, I started to think about that. And there are a lot of ways one could have gone at that question, but I went at it by trying to think about the varying capital intensity of growth, the fact that if an economy that saves a lot can afford to have a capital-intense, highly capitalized kind of production, an economy that doesn't save very much can still grow, but would grow in a more labor-intensive way. So I tried to work that out using fairly standard economic ideas. And it worked out. It worked out in the sense that there was a neat way of representing that and thinking about it. And not only that, but it produced a startling result. The theory seemed to lead in a direction that I certainly hadn't anticipated, which was--and this was rather different from Harrod and Domar indicate in their way--what turned out was that the volume of saving, and therefore capital investment, was not a determinant of the long-term rate of growth. It was a determinant of whether that growth path was high or low. A country that successfully saved an invested a lot would be richer than a country that only managed to save and invest a small amount. But the growth rates would depend entirely on the demography, the rate at which the population and the labor force were growing; and on, in the broadest sense, technological progress. And what I mean by 'broadest sense' was I wasn't thinking primarily--well the truth is, I was thinking algebraically, not in terms of images--but not necessarily mechanical invention, but organizational inventions and things like that. But in any case, it turned out that the long-term rate of growth was a matter of demography and technological progress.
6:20Russ: And this work launched a large literature. Guest: Oh, yeah. It did catch on. I think it caught on because it was neat and simple and it seemed to lead somewhere. It led to things that you hadn't thought about casually. And then, immediately after doing that, I started looking at data for the United States. And it's amazing to think about now, but what I was able to find, for the numbers that I needed to put flesh on that theory, covered the period in the United States from 1909 to 1949. A mere 40 years. And not a very long stretch. But when I analyzed those time series, those data, it turned out that what the history of that period seemed to be saying was that in fact, almost all--more than 80%--of the long-term average growth of output per person, or income per person, could be traced to this broad sense of technology progress. And that even between 1909 and 1949, very little of the fairly enormous improvement in income per head and the standard of living in the United States, even over that period, only a relatively small amount came from the nitty gritty of saving and investment. And more than [?] that you could track, at least the way I devised to track it, you could track to what an economist calls[?] shifts in the production function. Which translates in the capacity to get more output per unit of input--technology progress in the broadest sense. So, that's what I did then. And it did catch on and people worked at it, both of these aspects: the model building and empirical things. For instance, just to give one example, a good friend of mine, Padma Desai, tried to use the empirical apparatus. She was a student of the Soviet Union, what was then the Soviet Union. And what she found was that, unlike the United States--and by the way, you might remember that in the Post-War years, the fact that the Soviet economy was eventually going to collapse was not so clear, and Khrushchev's remark about how we'll bury you, and he meant economically--but Padma Desai found that the Soviet Union was just the reverse of the United States, and that what increases in income per person had been achieved there, had been achieved by grinding down consumption and investing in heavy industry, especially an enormous amount of the national income. Which suggested right away that over the long haul, Khrushchev was not going to bury us, but the other way around. Which turned out to be the case. So, this got to be a popular thing. And of course other people worked at it, and either improved it or at least elaborated it. And that goes on today. You pick up a textbook in macroeconomics or in growth theory and you find, starting with the model the way I [?] started doing it, an enormous amount of elaboration that goes on with that.
11:07Russ: So, I have the great disadvantage--and advantage--of being first of all many years removed from studying your model in graduate school--which of course I did, in the late 1970s. But I haven't thought much about growth theory. I've thought a lot about growth and I've interviewed a lot of people about growth on this program. And the first thought that comes to me as a naive economist is one that I know you're aware of, which is: Well, then the question is, where does technological progress come from? Where does this ability to get more from less come from? And one of the thoughts that I think inevitably crosses one's mind is it comes from that investment. So, explain how that--in other words, if you have no investment, if you have no savings, if all you do is consume your current product, it seems difficult to shift the production possibility frontier or to change the production function, or to speak in laymen's terms, to get technological change. So, in your mind, both then and now, where does technological change, where does productivity come from? Guest: Well, that's not an easy thing to pin down. Technological change itself, I think, comes from, at the beginning, Research and Development activity. In a capitalist economy, our economy, business firms are always looking for an advantage, or looking for a way to increase their profits to get ahead of the competition, if there is competition. And one of the ways they try to do that is to improve, is to reduce costs--put it that way. To find cheaper ways of producing what they now produce. And another path toward improving profits and gaining competitive advantage is by inventing, designing newer and better products. Now, here I want to digress for a minute. It occurred to me very early on, and I wrote about this, that new technology often needs ordinary plain vanilla capital investment in order to become real. If you design a new way to produce whatever you are producing, it may be the case that in order to actually make good, perform, on that new method, you need different machinery, different factory layout. You need ordinary capital investment. And I tried, and succeeded in a way, very early on, in working out a way of incorporating this fact in the theory. The interesting thing is that it still turned out, when you looked at the very long run, that it was still the rate of technological improvement that governed the long term, really long term, rate of growth. And the explanation for this turned out to be that what you can--if you are trying to conceptualize this, if you are trying to think about it--the fact that you need ordinary capital investment to embody, to really have--make good on--make a new technology effective, what you are really doing is making the age distribution of the stock of capital goods, you are shifting it toward the younger end. You can think of the population of machinery and capital goods like a human population, as having an age distribution. Young ones, then middle age ones, and eventually old enough to either be obsolete or physically useless. And what this connection between new technology and plain vanilla capital investment says that in order to move toward a newer technology you have to make the age distribution of the stock of capital goods shift in the younger direction. That can't go on forever. There are limits to the capacity of any economy to product young capital goods, just as there are limits to a human population's capacity to produce young kids. And in the long enough run, what this--when you incorporate this in the story, you don't change the story about the long-term, the permanent rate of growth. But you do change the mechanics of it. So, in the end, I think what you have to say, or as it seems to me, is that the source of technological change is research and development; and then a little beyond that, the act of--I was a student of Joseph Schumpeter's, among others. And he has this half-mystical, half-real thing about innovation and entrepreneurship, as distinct from mere technological maneuvering and an entrepreneur is somewhat different from an engineer or different from an inventor, actually. And I think all that does play a role. We understand very little about innovation as distinct from the development of science and the development even of technology. But that's where the source of it all is--it's creativity and all that. And creativity is motivated, you know, in part by just sheer playfulness, the way people are, but also in very large part by the search for profit and for competitive advantage. Russ: And as you point out-- Guest: So, if you want to encourage long-term growth, somehow you've got to encourage both technological innovation and innovation. Russ: When I used to teach microeconomics, I would give a poor person's version of growth in trying to capture how our standard of living is improved over time by the fact that cost curves shift down. Firms find new ways to make stuff cheaper, the same quality at a lower cost. And through competition, to the extent there is competition, they are forced to pass those savings on to customers. That model, which I think a lot of us teach some version of in micro, doesn't capture the more revolutionary idea of finding a new way to serve that desire that is totally different. It's cheaper in some dimension, but calling it cheaper doesn't really do justice to it. Guest: Yeah. That's right. In later work, I'm doing all this as a theorist, as a modeler. I try to work out a story in which there are big-time major improvements, and then once one of those takes place, it creates an opportunity for a lot of small further improvements. And I think the story of technological progress is a lot like that. One of the things it means, of course, is that it's very rare for even apart from business cycles, it's rare for growth to be smoothly exponential, x% per year forever. There are going to be changes in the rate of growth as these major inventions occur and as experience with the major inventions allows improvement. I spent some time, by the way, a number of years on a science advisory committee to a big automobile company. And one of the pieces of enlightenment that came to me then was to realize how much day-to-day technological progress doesn't involve Research and Development people. It occurs on the shop floor. Russ: Tinkering. Guest: Yes. Somebody realizes, You know, we could attach the bumper to the chassis just as securely using many fewer fasteners than we are using now if we did it this way rather than that way. And if it works out, that lowers costs. That's an improvement in productivity and the research laboratory never thought of anything to do with it. It's done by some foreman, and by trial and error.
21:26Russ: Talk about the work of Paul Romer in endogenous growth theory. Guest: Yeah. That's really interesting. It's interesting both in itself and as part of the history of economics. Paul Romer, a very bright guy, had the notion to try to make a fairly precise model, that is a step-by-step theory of profit motivated innovation. And this got to be popular under the name, under the label of endogenous growth theory. Endogenous meaning that you don't depend on some, as I did, on some poorly understood process of changing, improving technology, improving product, and improving productivity; but you treat creating higher productivity as itself a business, with the costs and payoffs. And you try to incorporate that in the whole story of economic growth. And it was interesting--for me it was sort of amusing in a way because people would say to me: How could you have been so stupid as not to realize that making technological improvements is itself an economic process? To which my answer was: Of course I understood--I knew that; I just had no idea how to think about it. So naturally I didn't think about it. But Romer and a number of other economists after him--and actually, a one-time colleague of mine, Karl Shell, had actually written some papers doing the same thing a decade or two before. This was an idea that was a natural for economics. And so Paul Romer and his [?]school that sprung up doing that--there was a whole proliferation of models of this kind. The interest in that seems to have fallen away. And Paul Romer himself has gone on to other things. And in a way, I almost feel as if the story of endogenous growth theory, of treating innovation, technological innovation as a business process itself, is hard. Russ: Yeah. Guest: It's just not something that [?] to theorize about it; to make a theory of it requires you have to just choose one way of looking at it, and in fact it must happen in a million different ways. And so I don't--it seems to me that endogenous growth theory didn't fulfill the promise that we all expected it would have. And there's not as much interest in it now as there was before because of the realization that it didn't unlock anything interesting that was really terribly useful. But, you know, some smart person may yet find a good way of analyzing that process. We all know that a lot of the innovation occurs as a business process. I keep telling myself we also all know that a lot of innovation comes as a matter of dumb luck. Russ: Yep. Trial and error. Guest: You set out to solve problem A, and you fail totally to solve problem A, but you solve problem B that wasn't in your head at all. And so, I don't know that we're going to get very far down that path[?]. Russ: Yeah. It seems to me it was a nice descriptive theory, but it has not, it seems to me--maybe I'm wrong--led to a much deeper understanding than Schumpeter, which you alluded to earlier.
26:25Russ: Now, in 1987 you famously said, You can see the computer age everywhere but in the productivity statistics. Has the investment in computing made a big difference? A small difference? And have we been able to measure it? Have things changed? Guest: Yeah, that line of mine, which was in a book review in a newspaper, gets talked about more than anything else. Russ: Yeah. Sorry. Guest: Well, that's all right. It's interesting: it was true when I said it. There was talk, and the book I was reviewing at the time was talking about the vast computer revolution and how it changed our lives; and, as I said, we were all very conscious that there were computers and things were going differently. But there was no measured increase in productivity. Well, over the years, there came a measured increase in productivity. And when that happened, it happened in an interesting way. It turned out when there were first clear indications, maybe 8 or 10 years later, of improvements in productivity on a national scale that could be traced to computers statistically, it turned out a large part of those gains came not in the use of the computer, but in the production of computers. Because the cost of an item of computing machinery was falling like a stone, and the quality was at the same time, the capacity at the same time was improving. And people were buying a lot of computers, so this was not a trivial industry, although not an abnormal [?] industry at the time. And you got big productivity gains in the production of computers and [?] and whatnot. But you could also begin to see productivity improvements on a national scale that traced to the use of computers. Interestingly, I got involved in some work, quite a lot of work, actually, with the McKinsey Global Institute, and we discovered that on the national scale, an awful lot of the improvements or the acceleration of the productivity trends that came from computers, came, surprisingly, in wholesale and retail industry. It was not the most sophisticated [?] use of the computer, but it gave you small improvements in output per person, and there are an awful lot of persons working in the retail level sale trade. Russ: The control of inventory costs just as an example was incredible. Guest: Yeah. Inventory costs; at checkout counters. All kinds of gains. And so, eventually you got to see the computer also in the productivity statistics. And you still do. Now the question is, today, this doesn't involve me any more, particularly, but as an observer: today there are beginning to be questions as to whether the main productivity gains from information technology may not be behind us. May not have passed. And the low-hanging fruit so to speak have already been plucked. And what we're doing now is refinement. And so, maybe the decades of substantial gains in productivity as a consequence of information technology and computational capacity have gone by. But I don't know about that. Russ: What do you think about this issue--we've talked about this on EconTalk before--that many of the benefits of the computer revolution and the Internet are not particularly monetary? So, for example, Wikipedia, Facebook, Twitter, Google, just to pick four off the top of my head. I don't think I pay anything for any of those directly: obviously I can pay indirectly in the form of higher costs due to advertising and other things. But the change in the quality of my life from those things is very hard to measure; but it's very large. Guest: Yes. All we can do is measure GDP (Gross Domestic Product). On the other hand, keep in mind that the fact that a large part of the benefits, of the human benefits, from standard-of-living type benefits, from information technology are free and are not measured in GDP--that kind of thing is not new with information technology. My friend Bob Gordon will remind you that an awful lot of the benefits from the invention of the flush toilet don't get measured in GDP either. Russ: That's correct. Guest: And most inventions, I suppose, create what economists call 'consumer surplus.' That is, benefits beyond what you actually have to pay for. And we always seem to be impressed with what is happening to us now; but there was a time when what was happening to us now was that sewage didn't run in the streets, and they didn't stink, and you didn't get sick. And those things also weren't priced in GDP. I think that's the moot[?] point, [?] no doubt prove what is significant.
33:00Russ: Let me ask you about your friend Bob Gordon, and Tyler Cowen and others who have--and recently Larry Summers--we're reading that we're in some period of possibly extended stagnation. Tyler's a little more optimistic, in the longer run. But some have suggested we've picked the low-hanging fruit. Others say that's absurd: there are enormous unimagined gains coming from innovations we just can't foresee. What are your thoughts on that? Guest: Well, I don't know. What am I to say, or what is anybody to say, when I'm told that there are gains coming that are unimagined--and unimaginable? If they're unimaginable, I can't imagine them. It might happen. Russ: You bet. Guest: The fact that I can't say what they might be, that comes from the fact that they're unimaginable. But some you can see--or you can see the potential of. But then you have to worry about their significance. For instance, you are now speaking to a 90-year old retired economist. The extension of the lifetime, of the longevity of people, that could continue still further. I think biologists and medical people have some notion that there is a maximum length to human life somewhere around 130; but I'm not even sure about that. It could be like taxi cabs--every part will be replaced but at least the name remains the same. But you could imagine that there will be, over the next decade, when my children and grandchildren are still around, there will be gains in longevity beyond the very substantial gains that we have. I never expected to live to be 90 and be talking on the phone to you, for that matter. So, they could live longer. You want, hope that they'll live a lot longer productive lives, as well, or at least enjoy their lives as well. And that's hard to foresee. So I don't really know how to answer the people who say, You shouldn't worry about the possibility of a much lower growth of capacity, of potential output in the future because there will be these marvelous technological changes that nobody can think of. If nobody can think of them, then you can't think of them. But that doesn't mean they won't happen. They might. But I don't know what--you're just testing whether you are an optimist or a pessimist. Russ: Yeah. I think there's a lot of truth to that. But as you point out, to combine your earlier remark to what you just said, we do see credible increases and effective changes in Research and Development in the biotech world that do promise to create some serious changes in our life in a good way. So, I am an optimist, I guess. That's probably why I'm not so worried.
36:35Russ: Now, let's shift gears a little bit. We'll get a little more contentious here. You've written recently that Milton Friedman was "bad for economics and bad for society." I want to leave out the society part for the moment; I want to focus on the economics part. Why do you think he was bad for economics? His contributions to the role of money in the economy, his skepticism about the Phillips Curve and the ability to fine-tune the economy--those seem to be kind of good for economics. Do you disagree? Guest: No. Well, the question was a little different. I have no interest at all in denying that Milton Friedman who was a friend made contributions to economics. What I was being asked about was the--the question was asked in this form: 'Why is there no Milton Friedman today?' Well, that can't mean--the question couldn't have meant 'Why is there nobody contributing to our analysis of consumer expenditure or contributing to our analysis of monetary policy,' and so on, because the woods are full of people who are doing that. What the question meant, and the question that I was answering, in my mind, was the things that set Milton as different from other economists; and in particular the fact that, in my mind, he was not given to doubt. And he was--the problem was that I thought an awful lot of other economists' time got spent in either real or side-table debates with Milton over broad issues of economic policy. Very broad issues of economic policy. Russ: Philosophy. Guest: Yeah, philosophy. That seemed to be a diversion, getting nowhere. And so, that's the sort of thing I had in mind. I think that--I'm perfectly happy to have arguments about monetary policy: is a constant rate of growth of the money supply a good target to give the Federal Reserve, or should you do something else? Does the globalization of finance affect the way you would instruct your Central Bank to behave? Should you license taxi cabs in New York City? All of those are questions that I'm perfectly happy to have people discuss and was perfectly happy to have Milton discuss. What I was against was the question, 'Why is there no Milton Friedman today?' is a little bit like the question, 'Why is there no Joan of Arc today?' Or whatever. And that side of Milton, as the ideologue, I thought was a waste for economics and diverted people from doing their daily jobs. Russ: Yeah, that question was raised by the Econ Journal Watch. We'll put a link up to that issue and to your contribution to it. I guess I'd have to say that I think there is a Milton Friedman alive today, of the kind you are talking about: his name is Paul Krugman. And he is also a man who is fairly free of doubt. When I say 'another Milton Friedman,' I'm talking about someone who speaks to the masses about public policy generally as well as some arcane economics and makes it accessible. Milton had many distinctive aspects, but that I think was what was particularly unusual about him. As you said, there are many great economic theorists, empirical economists; but the people who can combine their economic understanding and then speak to a general audience is only a handful of people who have been able to do that. Paul Samuelson was one. Paul Krugman is doing it today--but he isn't doing it from a free market perspective. And so in that sense he's not another Milton Friedman, I guess. Guest: By the way, Paul Samuelson, who I knew infinitely well, who was my closest friend, was utterly different from Milton in the respect that Paul Samuelson always had doubts about the truth of what he was saying. Paul was never a sort of person--in fact he changed his mind quite frequently. Something that I think Milton did not do. So, the art of speaking to a broad public with a foundation in solid economics is a very difficult art, and not many people can do it at all, and even fewer can do it well, I guess. But that's a little different from the lack of any skepticism, the ideological drive. I think there's a difference there. Russ: Well, let me defend Milton for a second; and then we'll move on. Unless you want to hit back. I think Paul Krugman, and Paul Samuelson, yourself, me--just about every economist--has an ideology. I do agree with you that Milton was very confident in his principles, and often confident in the application of those principles. So I take that as a legitimate criticism. But I think you have to consider the fact of where he was coming from. When you are a voice in the wilderness--and he was an extraordinarily lonely voice for a very long time, both philosophically and in terms of economic theory: for example in thinking about the importance of money--you do have to have perhaps a different public persona in how you present your ideas if you want to dent the consciousness of both the profession and the public. And he did that successfully. I think that was a good thing. But I can understand you might disagree. But I suspect his jovial confidence--which definitely was part of his demeanor--was a marketing point for him, as much as anything else. It may have been a personality trait as well, to deal with it. Guest: Well, I accept that it's a marketing point. I think that's right. And marketing is not something that's--the marketing of ideas is an activity that's very iffy. We all know that people who market snake oil, for instance, or market goods, are not above fooling the consumer, the customer. And we like to think that people who market ideas don't do that. And so I don't have much faith in the marketing faith[?]. I don't have much attraction to the marketing analogy at all. But I really don't want to argue about--Milton and St. Peter handle this between them as partners[?]. Russ: Uh, yeah. I'm not sure he's going to be seeing St. Peter, but not because he's heading south.
44:51Russ: Anyway, let's talk about Keynes. You've written that Keynes's General Theory is one of the great contributions of the last 100 years. It certainly is one of the most influential; I would say it's the most influential economics work of the last 100 years. For those of us who are skeptics--of the value, rather than the impact--of Keynes's theories, what evidence might one provide to convince us? I'm really asking a broader question, which is: Why is there so much disagreement in macroeconomics over the most fundamental things? Guest: Well--the question about Keynes and the question about why there's disagreement are separate questions. Let me--I'm going to say one minute on Keynes; and then a few more minutes on the other question. What I think that in the broadest sense, the truest significance of Keynes can be stated fairly simply: For most of the history of economics there was no careful distinction drawn between the capacity of the economy to produce goods and services, and the willingness, or the ability, of the economy to sell those goods and services--to get them bought and consumed. And what, in 1936, in the middle of the Great Depression, what Keynes's fundamental achievement as far as I'm concerned--and it changed economics forever, in my mind--was to say, you really have to distinguish between these things. There is this great economic machine that can produce goods and services which are valuable and which people want. And there is also the question as to whether it will produce them. And it will produce them only if the businesses in the economy can find willing and able buyers for them. And that fact creates a possibility of glitches. You can analyze the glitches till the end of the world, and they won't be the same in one 20-year period from another 20-year period. But the fact that you start--that I start--and I hope many economists today start with that distinction in mind. But in our lingo, Aggregate Supply and Aggregate Demand, we owe that to Keynes. And I think that that's important. And it has nothing to do--it's wrong to say it has nothing to do. That's not a matter of your beliefs about how monetary policy works or how fiscal policy works. Russ: Totally agree. Can I just challenge that for one sec? Guest: Sure. Russ: I would argue that there's a better claim to Robert Solow inventing growth theory--much better claim--than claiming that Keynes invented business cycle theory. Because--I've heard this claim before. It's strange. Economists were obsessed with the business cycle: the fact that--they didn't call it 'aggregate demand didn't equal aggregate supply.' They were very aware of the fact, before Keynes, that there were times when the economy did not fulfill its potential. Going back, I'd say, at least for 50 years before Keynes. It seems to me the distinctive contribution of Keynes is you can get out of that problem by borrowing money and spending it through the government. And that to me is the central question, is whether that is true. Do you disagree? Guest: Well, I disagree with that very much. Remember, you are talking to a 90-year old guy. I was brought up reading all those business cycle theorists. I was brought up on Albert Aftalion and Arthur Spiethoff and all that sort of stuff. And I was brought up on Gottfried Haberler's Prosperity and Depression. And I will tell you that for a person of reasonable, normal intelligence, studying economics, all those business cycle theorists, Hawtrey or Pigou for example, another example; Dennis Robertson. Russ: Mises. Guest: Writing about industrial fluctuations, and optimism and pessimism and whatnot--they were not ever clarifying the fact that even apart from oscillations, from intrinsically determining ups and downs, there is this conceptual difference between [?]. You're not going to deny that most economists as of 1936 believed in Say's Law, believed that except for casual interruptions, you didn't have to distinguish between aggregate supply and aggregate demand. Russ: But I don't think they thought--but they didn't--they believed in Say's Law, but they didn't think it worked instantly. And they thought there was, particularly for labor-- Guest: Well, but they thought that it was the underlying theme that governs it. And reading or learning or in fact--you know, Gottfried Haberler was [?] Keynesian, but he added to Prosperity and Depression a last chapter, after the original edition of the book; and it's clear that without being a Keynesian or without accepting Keynes's notions of how to deal in policy terms with the Depression, it was clear in Haberler's mind The General Theory had opened up this issue in a way that he, who had written the standard, authorized textbook on business cycle, or survey rather than textbook, on business cycle theory, hadn't done before. I think there is a real achievement there. Russ: Okay, fair enough. Let's talk about the--I don't mean to deny that he changed the way people think about it. Guest: Well, that's simple.[?] Russ: I just think--I find it strange--you mentioned a number of names. You could add Mises, you could add Irving Fisher. A lot of people worried about the fact that the economy didn't work so smoothly all the time, that Say's Law didn't work instantly. Guest: Yeah. But the difference between not working in theory [?] or are you seeing clearly that you could draw a curve of potential output and then you could try to analyze demand for that output. Russ: Okay, so-- Guest: You don't like that way of doing it, then you don't like way of doing it. But I do.
52:23Russ: Okay, so the question is, for those of us who don't like that way of doing it--which, the fact that I'm not in that group is irrelevant but a lot of illustrious people are not in that group: Why don't they accept? The real question I'm asking is, coming back to-- Guest: Well, [?] that. The question was: Why is there so much controversy in macroeconomics. And there I think you don't need to talk about specific doctrines in macroeconomics. There's controversy in macroeconomics because a). The problems are very complicated; b). There is a lot at stake. c). There are vested interests on all sides of every issue. And [?] it overflows into everyday politics. And I would have liked it a lot better to work in a branch of economics where those [?] aren't so; and you can sit quietly and do your work and go to sleep. But as a child in the Depression years, I got hooked on those macroeconomic questions. And the controversies arise for the reasons I said. You are talking about a complicated issue; you are talking about an issue in which a lot is at stake. The nature and use of the tax system, the nature and use of public budgets. You are talking about a problem in which there are important vested interests on all sides of the issue who are trying to, would like their interests exemplified in somebody's doctrine. And you are talking about issues which are part of day-to-day politics, which of course means that most of the time all anybody--one of the sayings for which I am famous is that "The length of the shortest true statement about economics is longer than the attention span of most people." And so, it's a natural for controversy. Russ: But let me ask it a different way. That's very well said. It's beautiful. Let me ask it a different way or get you to comment on the empirical side. So, I went off to graduate school in 1976 to Chicago, and my roommate in college went off to MIT (Massachusetts Institution of Technology). And we had of course a very different experience. I learned about how powerful monetary theory was, and he learned how powerful fiscal theory was. And he learned about how the government can fine-tune the economy and I learned that the government can't. And strangely enough, both of us came out of graduate school, I suspect, with a great deal of certainty about the wisdom of our views. I've become--whether this is, I don't know how exemplary this is--much more skeptical about the empirical evidence that supports my world view. And I think everybody on all sides should be skeptical about that. So my question is: The fact that empirical evidence is not decisive in settling these disputes, do you see that as just a question of time, as we get better; or do you see it as an inherent part of point a) that you mentioned: it's just a hard problem? Guest: I think it is in fact a hard problem. And it's not clear to me that the accumulation of empirical evidence will make the problem easy. Because, unlike, say, the velocity of light, the right answer changes over time. Society and the economy changes; the institutions change. The behavior, the way people behave, the way the participants, the important participants in the economy understand what they are doing may change. And so I don't think that--some fairly narrow issues are going to get cleared up with the accumulation of data, no doubt. And some have got cleared up. But I don't think that macro is going to become less controversial in any important way. For the reasons that I ranted about just a minute ago. I don't--you know, I had the experience that I had and I read the books that I read; and I can't say I was taught the macroeconomic side I accept now, because I wasn't, actually. I picked that up in the street, so to speak-- Russ: Sure-- Guest: where I pick up most of what I know-- Russ: We all do. Guest: And time--you know, I could contest your description of the difference between a Chicago education and an MIT education. How could you for instance think that your alternative MIT self would have believed that money was not an important thing to study when you were taking those courses with Franco Modigliani and Stan Fischer? Or if you'd gone to Yale, Jim Tobin? You wouldn't--those extreme versions that, you know, one side is for fiscal policy, the other side is for monetary policy--that's part of the everyday distorted discussion. That's not what goes on in the academic departments [?]. Russ: It isn't. That's a great point. It isn't what goes on. But it is what is yielded from--unfortunately, I think yielded from those departments when it gets into the public. Guest: That's because there are vested interests in politics. Russ: Yep, that's true. There's a lot more doubt in the classroom than there is in the pundit-sphere, or whatever you want to call it. That's for sure.
59:17Russ: We're almost out of time. I'd like you to close with a little thought experiment. If Robert Solow were 24 or so years old and coming out of graduate school today--and let's say, not too full of hubris and a little bit humble, which is hard at 24; I think as we get older we get a little more humble. But if you were launching your career now, what would you be working on? What do you think are the key--not the most effective things: I'm not asking for career advice for young economists. I'm asking, if you want to change the world, what would you work on now? Guest: If I wanted to change the world--oh, boy! That's a question I haven't thought about at all. I don't--you know, I don't think that there is some special thing. By the way, one of the things I might be interested in is the theory of economic growth; we were talking just a minute ago-- Russ: Oh: The more things change, the more they stay the same. I don't think that--I think it's the most important thing. Guest: [?] about the difference between whether we have run out of improvements in the standard of living or whether stagnation either on the supply side or the demand side or both is in the future or not. I think that's an issue that--I don't know whether you could change the world, but you'd sure get to understand something about the world, if you do that. I'd still think of growth theory and what goes into growth theory, which is primarily the role of capital, labor, and knowledge in the economy, is still the important issue. For better or worse.