• Print •  • E-mail Alerts • Search Articles

September 2002

Interview with Robert Solow

As a young professor in the 1950s, Robert Solow rejected prevailing theories about wobbly “knife-edge” economic growth paths governed rigidly by capital and labor. If they were accurate, he later wrote, all one could expect to find on earth was “the wreckage of a capitalism that had shaken itself to pieces long ago.” Instead, Solow created a robust model of “steady-state” growth in which economies progress on essentially stable paths. He then demonstrated that technological change—ignored by mainstream theory—accounts for a huge proportion of that growth.
	Solow's model fundamentally changed economic analysis, providing “a framework within which modern macroeconomic theory can be structured,” according to the Royal Swedish Academy of Sciences, which selected Solow as the 1987 recipient of the Nobel prize in economics. His theory—born as Sputnik was launched—also altered history as national governments came to realize they must harness technology as an engine of growth. Solow's revolutionary model became the new orthodoxy.
The Keynesian interventionist policies that Solow advocated as senior economist to the Council of Economic Advisers during the Kennedy administration also became conventional wisdom for a time, but in recent decades Keynesians have been under stiff fire. Solow delights in the controversy, skillfully defending his position and disarming his opponents with humor. “The best thing you can say about Reaganomics is that it happened in a fit of inattention,” he once quipped. Dismissing a rival's theory, he said, “Intellectually soft as a grape.” With The Region, Robert Solow discussed growth in growth theory, his tenure on the Boston Fed's board and the lessons economists can learn from sailors, among other things. Our conversation reflects the wit, intellect and commitment to human welfare of an economist who, often as not, makes progress by sailing against the prevailing winds.

Current growth theory

REGION: The field of growth theory blossomed after your contributions in the 1950s, but you observed that by 1970 it had reached a point of “intellectual diminishing returns.” Over the last 10 or 15 years, Paul Romer, Robert Lucas and others have brought about something of a renaissance. How far have we gotten with the “new growth theory”?

SOLOW: I'm probably in a minority here, but my perspective is that many more questions have been asked than answered. People are now willing to take a shot at issues that 40 or 50 years ago would have been thought to be a little too hard. But I don't have the impression that we have useful answers to the important issues that were pursued.

The prize example in my mind is not so much the Lucas case as the Romer case of endogenous technological progress. If you think about it, nobody in the world ever doubted that there was a large endogenous element in technological progress. All those people in Eastman Kodak or General Motors laboratories, drawing down salaries and doing work, were clearly not believed to be wasting their time.

On the other hand, I thought from the very beginning and still think now that there is an element of sheer chance in technological innovation that we've come nowhere near catching.

What little experience I have in observing industrial research says that R & D groups work on problems because those problems are interesting to their employers. But they may actually come up with something different from what was intended. They may fail where they thought they were going to succeed, and vice versa. Nor is it a pure probability thing. Every science has its own internal logic. There are questions you can answer and questions you can't, and that logic may have very little to do with the economic motive for trying to answer the questions. So I think there is a lot more of the exogenous in what actually comes out of serious technological research than the models allow. The models tend to be fairly simplistic. They have to be simplistic, of course: They're models. But not every simple model is a good one.

I thought from the very beginning that the so-called AK models [which assume that output is the product of capital, K, and a constant positive level of technology, A] were not only foolish but dangerous, in the sense that they would suggest to economists that they understood something that they didn't. I also thought they were intellectually threadbare. They came so close to assuming what they wanted to prove: that fiddling some easy-to-fiddle level could control a rate of growth. And in fact it has worked out that way. The AK models are gone. But what's left is almost always something very much like the AK assumptions, only more elaborate.

REGION: You've called them “arbitrary linearity assumptions.”

SOLOW: Exactly, and the reason is that if you want to talk about “the rate of growth” as an independent intellectual object, you need a linearity assumption; but talking about a rate of growth as an object may itself be an interference with intellectual progress.

If I show you a chart of a local red-tailed hawk gaining altitude through time as it flies, you'll see a graph with some wiggles but generally going up. If you ask, “What's the hawk's rate of altitude growth?” the right answer is that a hawk doesn't have a rate of growth. When the hawk has gone as high as it's going to go, you could say that on average it gained altitude at 20 feet per second, but you're telling me nothing about the hawk's path. In fact you're falsifying the hawk's path.

I think that economists—and I was partially responsible for this—have fallen into the trap of thinking that in order to talk about growth at all, you need a model in which the rate of growth is an independently defined object. Doing that may have been costly. It's why we have all these AK-type and other linearity assumptions, because growth theorists want to be able to say they have “determined” a
steady-state rate of growth. But nothing significant is lost if there is no steady-state rate of growth, if growth paths are not piecewise exponential.

I think it was really worthwhile opening up the question and saying that it can't be right to treat the level of technology—or the level of human capital, in Bob Lucas' version of this—as exogenous. We all know that to some extent ordinary economic incentives and motives play a role. So, we should study it as a piece of economics. But I think that the part of the profession that was interested in these subjects fell a little too easily into simple-minded but convenient theories. Maybe it was a necessary stage. But I don't have the impression that the next steps have yet been taken.

My own pet suggestion is that it would be a good idea for economists who are interested in endogenous technology and growth theory to do a little observational work on industrial research laboratories (or at least read what others have done). Part of the difficulty is that a lot of the improvement in technology has nothing to do with R & D, but emerges in a different way from the factory floor.

Even within R & D, it would be a good idea to pay closer attention to how, say, IBM allocates its research resources. How do they measure their output? If you were actually trying to describe the process of endogenous generation of new computer technology at IBM or Compaq or somewhere like that, how would you do it?

It might even be interesting to observe the difference between promises and results in industrial research. The reason not much of this has happened is that the kind of person who's good at observing things like this is not necessarily the kind of person who's good at making models.

The Solow paradox

Trade and growth

Still a Keynesian?

REGION: You feel, I believe, still comfortable with the label “Keynesian,” but as you know, some economists consider Keynesianism irrelevant, outmoded. What parts of Keynesian economics do remain relevant? Does fiscal intervention still work? Does the Phillips curve theory that you and Paul Samuelson brought to American shores in 1960 still live as an exploitable trade-off? Or do you have something broader in mind about what it means to be a Keynesian?

SOLOW: The broad way in which I would associate myself with Keynes is this. When you think about a modern capitalist economy, there are two polar perspectives you might take. It's possible to focus on the beautiful mechanics of the market and the way in which it decentralizes decision-making and therefore decentralizes information and yet manages to make all of those things cohere through this beautiful instrument of the price system.

Or you can focus on the deficiencies, on the way in which the beautiful mechanism of the market doesn't work. There are asymmetries of information. (By the way, nobody in the world today, after Enron and [Arthur] Andersen and all that, is ever going to be able to argue that information gets around. It doesn't.) There are substantial elements of monopoly in the modern world. Even international trade, though it certainly helps, doesn't necessarily clear that up. And there are inflexibilities, rigidities in the price and wage mechanism.

I think it's foolish to think that these are imperfections you could hope in time to get rid of. I think those rigidities in wages and prices correspond to needs that social institutions are trying to meet, and those social institutions don't function very well if every wage is open to nontrivial movement at every instant of time.

So there are these two perspectives. You could say the system works so well: Why doesn't it make sense just to suppose that the real world is like this beautiful textbook world. Maybe you can convince yourself that the things that follow in the textbook follow logically in the real world too. Markets clear all the time. Imperfections work themselves out. Competition eventually triumphs, especially if the state keeps its paws off of things.

The alternative view is to say the whole thing is hopelessly flawed. Capitalism cannot be made to work well, and therefore we might as well have the state or some other agency run everything. That's the other extreme.

I, of course, stand squarely in the common-sensical center. [laughs] My view is that we know no better way of running an economy than market capitalism. Nothing else seems to do. But there are many ways in which a market capitalist economy malfunctions. And one of the ways it malfunctions—and now I'm getting very close to Keynes—is it can have excess supply of goods in general and not be able to fix that by itself. Or at least not in a short enough interval of time so that it makes sense to say, let's just keep hands-off and let the system solve these problems the way it's supposed to.

I think there are periods of time when we have persistent excess supply or excess demand, and it is a function of government to repair the malfunction. There are times when there will be inflation and governments ought to try to control it. There are times when there will be prolonged unemployment and governments ought to try to fix that. I also think that there's every reason to believe that, in order to fix problems like that, fiscal policy, as well as monetary policy, is both necessary and effective.

Now it might sound as if I think that governments work just beautifully. Of course I don't believe that. For really short-run macroeconomic problems, monetary policy is the tool of choice precisely because we have learned to operate it in a way that doesn't get caught in political hang-ups.

By contrast, when you try to do economic stabilization through fiscal policy, you always (probably necessarily) end up with conflicting interests. You can't just say we need higher taxes or lower taxes. You have to have higher taxes in some particular way or lower taxes in some particular way and the people in society and their elected representatives are going to care quite a lot about who is favored by a tax reduction, who is damaged by a tax increase. The same is true on the expenditure side.

Because of this inevitable political process, it's hard to get fiscal policy done promptly, in a timely fashion. That's why I think that for short-run processes we do have to rely on monetary policy.

Now, getting back to the big-time macro theory dispute, we know that if a country has more than one economic objective, more than one goal it's trying to pursue, it has to have more than one instrument. Monetary policy is one instrument. You can measure it by what happens to the short-term interest rate or by some characteristic of the money supply or whatever, but it controls one thing and it can only have one objective at any moment of time.

If you listen to the enthusiasts (I'm thinking mainly of the old Bundesbank—this is more European than American generally, although in some ways it was true of American monetarism as well), you would think there was only one policy problem, namely, the control of inflation. And therefore you only needed one tool and monetary policy was the logical one. To say that there's only one policy problem—to control inflation—is to say that if only you control inflation, the rest of the system would work itself out very well.

The Bundesbank people used to say this almost with dead seriousness, but it strikes me as the most utter foolishness. The history of our economy since the Second World War, without going back to the Great Depression, tells us that even when inflation is under control, there are often other macroeconomic problems. If politics would allow fiscal policy to operate, then we could make good use of fiscal policy. Whether the United States can find a way to make timely fiscal policy, to master this fact that there are always conflicting interests and political bargaining, is uncertain. Maybe we can, though I am not optimistic.

My own particular view is that we have done something foolish regarding fiscal policy. We used to depend a lot—more than people other than economists knew—on what were called automatic stabilizers. Just as an example, if the economy began to boom, the first part of income to gain would be profits. Profits were very heavily taxed. So when the economy boomed, the federal budget moved in the direction of surplus.

On the other hand, if the economy began to turn down, transfer expenditures would rise rapidly, profits would turn into losses, and corporate tax revenues would fall. So you'd get the kind of shift in federal budgets that any good Keynesian type would have wanted, but you got it without legislation.

Over the years we have weakened those stabilizers. We've weakened them by, for good or evil, diminishing welfare and other transfer expenditures. It might have been right or wrong to do that, but it was not done in order to weaken automatic stabilization. That was a side effect. We also tax corporate profits relatively much less. We depend much more on personal income taxes now than we used to.

The beauty of these automatic stabilizers was that they didn't depend on the stupid Congress getting its act together. It just happened. It wasn't exactly a policy rule but it operated almost as if it were. And so one of the things I think we could do right now is to go back and try to find ways of strengthening those automatic stabilizers again.

Inflation-targeting

The Phillips curve

Real business cycles

No one has done it. It's not easy to do, of course. But without that, I don't think there is any good reason to think that real business cycle theory is getting anywhere. In other words, the low-hurdle tests have very low power against plausible alternatives.

All is not lost, however. It does seem to me that some real business cycle modelers have gone at it in a different way. They've said, well, since the world does seem to have imperfections, whether market imperfections or information imperfections that are not in the simple model, why don't we try to keep as much as we can of the real business cycle theory but allow for the imperfections? Some people have done that, Marty Eichenbaum at Northwestern and [Susanto] Basu and [John] Fernald, [Craig] Burnside and [Sergio] Rebelo, and others. I think that they're going in the right direction.

As I say, I'm not a doctrinaire Keynesian. I'm not trying to defend J.M. Keynes, whom I undoubtedly would have disliked intensely if I'd ever met him. And I'm not against dynamic stochastic general equilibrium models. But what I want them to do is to explain not only the good things, but also the important pathologies of market economies. So to the extent that you introduce imperfections and you explain pathologies, that's great, that's the way it ought to go.

It seems to me that the original model was OK for a start, but there was a tendency to claim too much. I have an intellectual aversion to models that start with a representative consumer optimizing over infinite time. It just seems to me to be asking too much. Real business cycle theorists start there and then say let's whittle away elsewhere in this model to try to make it look more like the world. It's one way to go, but I don't think it's necessarily the best way.

The Boston Fed

REGION: The Boston Fed was fortunate enough to have you on its board during the mid to late '70s.

SOLOW: I was on the board for, it was supposed to be six years, but I was actually on it for eight. I guess they couldn't find Class C directors. Who would want to be a Class C director? And I was the chairman for three years or so, and vice chairman before that.

REGION: What insights could you share about how district bank boards work, and also about the relationship between district bank boards and the Board of Governors?

SOLOW: I found that what made the Boston bank work was the then-president, Frank Morris, who was himself (a) a very good economist and (b) a very undoctrinaire economist. He had both university experience and he had worked in the financial services industry. The useful role that someone like me could play was doing two things. First of all, just talking with Frank. He liked to talk and collect ideas. And secondly, to act as an intermediary between Frank Morris, the president, and the rest of the board that didn't always have much of a grasp of the logic of monetary policy or the nature of the current macro problem. Having someone on the board who could translate, so to speak, those things from the staff—and Frank was very good about using his staff—to the board, that was a really useful function to perform.

That it was part of my education is unimportant, but I learned a lot about the banking system and a lot about the way the Federal Reserve System worked. And I loved going down to Washington and extracting out of the Board of Governors a really high salary for Frank Morris. I thought he deserved every bit of it and more.

It was then, I suppose, that I got a serious feeling for the way in which part of the problem of doing monetary policy is the necessity of juggling multiple objectives with a single tool. When do you focus on this, when do you focus on that? I don't know whether that has changed internally, but as I've been saying right along here, I think that is still a besetting problem for makers of monetary policy.

It is very important to remember that during that time Paul Volcker was the chairman. And he, like Alan Greenspan, was someone you could talk to, a person of real economic sophistication. It was a lot of fun.

I don't know that the experience I had way back then has very much to say to current boards. But I do think that the things that I learned about running a Federal Reserve bank, I learned from Frank Morris. And they included such things as getting and keeping the staff involved. And using the staff to brief the board on current events, current policy issues. Not so much on analytical issues. There I think that what every Federal Reserve bank board needs is one or two academic economists who can intermediate between the staff, including the president, and the rest of the board.

But keeping the staff involved and exposing the board to the staff so that they realize that there's a lot of hard work going on, to brief the president for his functions on the Open Market Committee, that's really very important.

Economic literacy

REGION: You once wrote a provocative essay with a wonderful title, “Why Economic Ideas Turn to Mush,” explaining how difficult it is to convey economic ideas clearly. It sounds as if you played that role between the staff and board of the Boston Fed, and you demonstrate it in the pieces you write, for example, for the New York Review of Books. What have you learned about how to maintain the integrity of an economic idea, yet still convey it clearly?

SOLOW: It's the hardest thing in the world. It really is very, very difficult. What have I learned apart from the fact that it's really difficult to do? Well, first of all, focus. Try to formulate an economic problem in a very clear, focused way. Try to answer one question at a time, and insist on that. And above all—this is really what's difficult—at least I know that I tend to forget it: Don't omit qualifications. Never claim more than you actually believe or can justify.

What makes that hard is that what people want—especially if they're being fed it in sound bites on a television program or in a two-sentence quotation in The Wall Street Journal—what they want is something very definite. They don't ever want those qualifications. And you must never let them off that hook. The interesting thing is that I think it's useful. An economist trying to talk to the general public gains respect by insisting on the qualifications, by not appearing as a pundit, as someone who knows all the answers.

All my academic life I have tried to be a good teacher. I've spent a lot of time at it. And communicating economic ideas to the general public bears a lot of relation to communicating economic ideas to freshmen or sophomores. If you're worth your pay you ask yourself before every class, what is the best way to get this stuff across to these kids? You have to ask yourself, what is the best way to get this across to the reader of The Region, or the Minneapolis StarTribune? Think of it as teaching. At least, think of it as teaching if you're a decent teacher. [Laughs]

Welfare reform

Global warming

Lessons from the sea

REGION: The economic world seems infused with aquatic metaphors, from liquidity trap to cash flow to exchange rate float. Even “leaning against the wind”—a favorite Fed slogan—sounds vaguely nautical. You're an avid sailor and have been for years. I believe you even used some of your Nobel prize money to buy a jib for your boat here on Martha's Vineyard.

SOLOW: That's right, I did.

REGION: Do you ever relate these two worlds? Is there anything you've learned from sailing the Atlantic that has taught you about economics, or vice versa?

SOLOW: Let me remind you, I'm an inshore sailor. I never even talk about sailing a transatlantic sail. If it ever comes up in the family, the next sentence is, “Well, you can do that with your next wife.” But I have no wish to do that. I'm an inshore sailor. I just like sailing a boat. Apart from the activity itself, the main thing I like about sailing is that it teaches you that the water and the wind out there don't give a damn about you. They're doing whatever the laws of physics tell them to do and your problem is to adjust as best you can.

And learning to adjust, to adapt, is not a bad thing for economists to learn either: Adapt to changes in the world. If things that used to be in inelastic demand suddenly acquire a lot of substitutes and demand for them becomes elastic, there's no point in wishing it would go back to the old situation which was so simple. You've got to fit your model to the world, not the world to your model. (We're back to real business cycle theory again.)

I don't build a lot of my life around sailing a 28-foot boat; it's just one thing I take some pleasure in. But I do think as a metaphor for other things sailing teaches you that there's the water and there's the wind and if you'd like to get from here to there, well, figure it out, but don't expect any help. And don't try to impose your will on the world because it's going to go the other way.

REGION: Thank you, Mr. Solow.

—Douglas Clement