PrintPrint CiteCite
Style: MLAAPAChicago Close


C. Peter McColough Series on International Economics: Productivity Growth and Monetary Policy

Speaker: Ben Bernanke, member, Federal Reserve Board of Governors
Presider: John H. Makin, principal, Caxton Associates
January 19, 2005
Council on Foreign Relations


Council on Foreign Relations
New York, N.Y.

JOHN MAKIN: Good morning. I think we should get under way because we only have an hour. My name is John Makin. I'll be presiding today. And I have the pleasure of introducing Fed Governor Ben Bernanke. Ben and I share something in common. We both have moved on from academic careers to other careers, Ben as a policy-maker and I as Wall Streeter-hedge fund strategist. I think when you move from academia to the Fed, you have to be very careful about what you say. And when you move from academics to Wall Street, you can say anything you want. But today Ben is going to talk about productivity. This is an on-the-record session. And he will probably talk for about 25 minutes, and then we will have a Q&A session. And we will conclude promptly at 9:00. So, without further ado, Governor Ben Bernanke.

BEN BERNANKE: Thanks, John. You don't—you didn't mention that you have Fed association, too, right. So, I mean, you're not embarrassed about that, I trust. [Laughter] Well, thank you for coming this early morning, this cold morning. I'm going to talk today about productivity, which is probably the most, you know, it's not the variable you see every day in The Wall Street Journal, but it's probably the most important, or one of the most important variables of the macro economy.

We've had quite substantial developments in productivity in recent years. Between the '70s and 1995, labor productivity growth—and when I talk about productivity I'll be referring to output per hour—grew between one [percent] and 1.5 percent a year, a relatively disappointing rate of growth, and less than we saw in other major industrial countries. Between 1995 and 2001, however, productivity growth rose to about 2.5 percent a year—a pretty significant change in the context of these things, and a contributor to the view at the time that we're, perhaps, entering a new economy. Of course, the stock market was part of that story as well. Talk of the new economy faded with the stock market and the high-tech evaluations in particular around 2000. But interestingly enough, and perhaps not evident to everyone—certainly to most people here, but not to the general public—productivity growth actually accelerated after 2001. Over the last four years, the pace of productivity gains have been better than 4 percent per year, despite all the adverse developments—9/11, the recession, and all the other factors.

Now, why is productivity so important? You know, I've written a couple of economics textbooks, and we all know that in the long run, productivity growth is essentially the only determinant of living standards. In the shorter run, the link between productivity and living standards is a bit more loose, as we've seen in the last few years. We've had high productivity, but we've also had things like changes in labor participation, we've had changes in the split between capital labor and income, so the link is a little bit looser in the short run. Nevertheless, productivity growth is still very important because it has very significant macroeconomic impact, even in the short run, including effects on growth, inflation, and employment.

What I'd like to do today in the time I have, is first talk a bit about why productivity seems to have changed into a new dynamic, has entered a new era in some sense in the last decade. It's important to try to understand that because we are also interested in knowing where productivity is going to go in the next few years, and I'll talk a bit about that. And then finally at the end, I'll come back and I'll say something about how our productivity projections and expectations influence the economic outlook and monetary policy.

So, let me start and talk first a bit about the last 10 years and why productivity growth seems to have picked up. The economic consensus on this issue has evolved somewhat over time. By around 2000, there was a pretty widespread view among economists that the main driver of the pickup in the productivity growth was advances in information and communication technologies, ICT, during the 1990s. The general analysis showed that ICT improvements had expanded the productivity of the U.S. economy in at least two important and distinct ways. First, technological advances allowed the ICT-producing industries—the industries that produced the computers, the communications equipment, and so on—to expand their rate of productivity growth. For example, there were advances in chip manufacturing processes, and increases in the speed of the product cycle, of Intel, for example, that literally increased the quantity and value of chip production in the U.S., and thereby raised labor productivity in that sector. But, of course, just increasing the productivity of the ICT-producing sector is not really what this game is all about. What's interesting is the extent to which improvements in ICT led to productivity increases in the ICT-using sectors, that is, most of the economy. And here, the evidence suggests that in the United States that a wide range of industries, including service, durable goods, and other industries, were able to use these new technologies to reduce their costs and increase their quality. The McKinsey study in 2001 was particularly informative about some of the details.

I think many people here are at least broadly familiar with some of the sea changes in industry that have taken place in the last 10 years in a wide variety of areas. Take retailing, for example—not what we used to think of as a high-tech industry, but Wal-Mart has used IT tools to improve the management of their supply chains, increased their ability to respond to changes in consumer demand. Wal-Mart typically knows, I think, on Monday afternoon exactly what the composition of purchases was on Monday morning all across the country, and they can respond with lightning speed to changes in consumer demand. In other sectors, here in New York, many of you work for brokers and traders and dealers, and you know something about how automation of trading processes and back-office operations has increased productivity in the financial services industry. In durable goods, General Motors and other automobile companies have, for example, developed programmable tooling systems to increase the flexibility of their production processes. So, for example, now a GM plant can produce vehicles from different platforms all in the same line because of the ability to shift from one setup to another.

More broadly, there's a nice study by Kevin Stiroh, who's at the New York Fed, who's been one of the leaders in recent productivity research, who found that nearly two-thirds of the U.S. industry, comprising about 70 percent of U.S. employment, experienced an acceleration of productivity in the later 1990s. And he found that the growth of productivity across industry was positively correlated with the intensity of ICT use in that industry, given prima facie evidence in favor of the ICT causality. So, there's no doubt, I think, that at this point that the ICT revolution and the productivity revolution [in] the U.S. are connected. But there are some puzzles that have arisen that have made economists in the last four or five years think harder, I think, about exactly how those two phenomena are related.

First, it's interesting to note that the United States, of course, was not the only country to see a rapid expansion in ICT investment. Similar patterns are seen in a number of other industrial countries. And yet with a few exceptions, and there have been some exceptions, productivity growth in other advanced industrial countries has not increased to the same degree that it's happened in the United States. In comparison with members of the European Union, it's particularly interesting in this regard. Throughout most of the post-World War II period, labor productivity growth in Europe grew more quickly, exceeded the rate of growth in the United States. First, there was the post-war reconstruction as Europe rebuilt and, of course, returned to normalcy. And then over time, there was a continued convergence as European business practices, technology, and the like, approached the American standard. Indeed, [Northwestern University's] Bob Gordon, another leader in productivity research, has estimated that European productivity in 1950 was only 44 percent that of the U.S. level, whereas in 1995 it had risen to 94 percent. That is, a convergence was approaching. However, since 1995, because productivity has picked up more in the U.S. than in Europe, there's actually been some divergence. That is, the difference between the U.S. and European productivity levels has begun to expand.

Now, what's the reason for that? Interestingly, there doesn't seem to be much difference in productivity gains in the ICT-producing sectors between the United States and Europe. The United States generally has the lead in information technology, but in Europe, Nokia has done well on the communications side. What is different is that Europe has clearly not been as successful as the U.S. in applying ICT technologies to non-ICT industries. Why would that be? Well, one view that has received some attention, in part because of the proposers, who included [Federal Reserve Board Chairman] Alan Greenspan in a speech in 2000, and also [National Bureau of Economic Research President] Martin Feldstein, suggests that the relatively heavy regulatory burden in Europe may have inhibited flexibility to the extent that adoption of new technologies has not been as successful. For example, taking full advantage of new ICT technologies may involve extensive reorganization of work practices and reallocation of workers among firms and industries. So, to the extent that you have regulations that inhibit hiring and firing workers and reducing employees' ability to change work assignments, as exists in a number of European countries, some of these changes may be difficult to achieve.

Then we also in Europe have a certain number of government-owned, semi-monopoly firms, and restrictions on the entry of new first, which reduces the competitive pressure in industry, which also may reduce innovation and adaptation of new technologies. There's some Fed research that shows that across countries, that if you look at indexes of regulatory burden and labor market regulation, that the rate of adaptation or adoption of ICT technologies is inversely correlated. The more regulated you are, the harder it is to adopt these technologies. And there are a lot of specific examples. For example, Bob Gordon, in his talk about the land use restrictions and the restrictions on shopping hours that has impeded the development of Wal-Mart-type, big box retail outlets in Europe—whatever anyone might thing about the aesthetic Wal-Mart, they have been a tremendous economic boom—boon—and a boom—and they have contributed very substantially to the overall productivity increase in the U.S.

The regulatory burden is probably not the entire explanation of the U.S.-European differential. And one striking case here is the U.K., which has also not had U.S.-style productivity gains, despite the fact that its regulatory regime is closer [to] that of the United States. One possible explanation there may be a shortage of workers with appropriate skills. The U.K. has a relatively low education attainment compared to other industrial countries, and that may be part of the problem. Skilled shortages might be relevant also in some other contexts. For example, Europe has had extensive youth unemployment now for several decades, which probably reduces the opportunity of young people to get on-the-job training and experience that allows them, as everyone knows—young people are the people who know how to run these machines, and if they're not working, that's going to be a problem. So clearly there's a lot of interesting work to be done in understanding the differences between European and U.S. improvements in productivity.

A second puzzle that is worth talking about briefly in terms of understanding the link between ICT and productivity gains is the fact that ICT investment has not, in general, been followed on a short and reliable time period by improvements in productivity. It's a very long and very "lagged," to use Milton Friedman's past terminology from another context. And so, some critics have argued that, you know, that perhaps the connection is not as tight as we think. For example, ICT investments began in a fairly serious way in the United States in the 1980s. In 1987, Bob Solow, one of my thesis advisers at MIT, a well-known economist, made the famous quip that "Computers are everywhere, except in the productivity statistics." And indeed, there seemed to be just a long period where everyone was learning how to use Windows, yet nothing was happening to measure productivity. At the same time, by analogy, as I mentioned before, after the tech meltdown in around 2000, there was a very substantial decline in ITC investment—ICT investments, and yet for the last four years, as I've mentioned, we've seen even a step-up in the rate of productivity gain. So, there is some puzzle here about the temporal link between ICT investment and productivity growth.

One solution is that people who have studied this in great detail find that it's not just enough to buy the computers and bring the box in and hope that suddenly productivity is going to rise. I mean, there's a lot more to it than that. Clearly, you have to have a thought-out plan, and with some expectation of how exactly you're going to use the new technology to make the company more efficient. Indeed, there have been a number of examples of industries, perhaps hotels might be one example, where there's been a substantial investment in technology without too much measurable benefit in productivity. One sometimes hears the view that the reason for the current productivity is that somewhere around 1999, firms bought a huge amount of ICT equipment, brought it into the house, and since then have been trying to figure out how to use it. I think that can't be quite literally right. There's got to be some sense, some planning, some understanding in advance of how the new equipment is going to be used.

A more subtle issue relating the connection between ICT investments and productivity gains flows from the fact that a number of observers have characterized these new technologies as so-called "general purpose technologies," or GPTs. A general purpose technology is a technology that has the potential to essentially revolutionize production and consumption in a broad range of activities. And GPTs in the past have included such major developments as electrification and the internal combustion engine. And if one believes that the computing and communications advances are a GPT, the potential is tremendous, but because they are general purpose, they need to be customized and applied to a specific use. And indeed, managers must supplement their purchases of new equipment with investments in research and development, worker training, and organizational redesign—all of which economists tend to call "organizational capital" or "intangible capital."

For example, Wal-Mart just didn't buy the computers, they had to reorganize their work assignments, they had to retrain their workers, they had to develop new relationships with their suppliers, and they had to modify their management systems. There's a lot of intangible investment that has to go into the process to complement the physical investment in new equipment before the ICT equipment shows a return. This intangible investment is probably very important. There's another Fed study recently that tried to measure intangible investment. You don't have income statistics on this, but you can try to pick out how much R&D [research and development] there is, how much is invested in training, things of that sort. And the—this study suggested that intangible investment in the last decade is on the same order of magnitude as the physical investment in new equipment and software. So that's a very important finding for a number of reasons. One is that, of course, in the national income accounts, we expense training and organizational redesign and research and development, but what this viewpoint suggests [is] that these are really investments, not expense items, and in that respect, investment and saving in the U.S. is a good bit higher in reality than the national income statistics would suggest.

Second, there's been this interesting phenomena I've long noted that when new machines are introduced, the computer is upgraded, and so on, there's actually a decline in productivity for a while, maybe because the kinds of games the employees can play become more sophisticated. But clearly, if you believe that these machines need additional investment in terms of training and reorganization to become productive, then there would be a period when output and productivity actually decline following these kinds of investments. But finally, the importance of investing in intangible capital is, I think, a good explanation, at least, for why there's not this immediate relationship between buying a computer and having higher productivity. It's a drawn-out process. It is a process that has lagged, and that is certainly to be expected given the importance of this intangible investment.

All right. So, those are some of the ideas that have been laid out about the sources of productivity gains and some of the modifications that economists have made in thinking about the last decade. Let me say a couple of words about looking forward. Where do we think productivity growth is going to go over the next two years, the next five years, the next 10 years? It's a hard thing to predict, obviously. And, you know, intrinsically what you're tying to do here is predict innovation, which almost by definition is impossible. And nevertheless, economists are never thwarted from doing impossible tasks, and a number of people have tried to do that.

Now, one complication, which I won't get into in much detail, is the fact that productivity growth has a very substantial cyclical component, which has to do among other things with the fact that firms are reluctant to change their workforces rapidly, so that in periods of low output, typically measured productivity is low because you have less work being spread over more people, whereas in periods of rapid employment growth, we tend to see productivity slow because you're training and hiring new people who are not yet up to speed. So, analysis of productivity does have to look at the cyclical component. And I would just say that as of right now, we are now in a maturing expansion. It's been now more than three years since the official end of the recession. Incidentally, one of my previous assignments before coming on the Fed, I was on the committee that decides these things, and let me tell you, it ain't science, but that's another question. [Laughter]

Be that as it may, we have now about, of course, as I said, more than three years from the trough of the recession. We are beginning to see improvements in the labor markets. And so productivity, predictably, is declining somewhat these last few quarters, below what might be sustainable in the long run. So, we have to take that into account. Looking ahead, though, a couple of quarters, maybe to the end of this year, we should expect productivity to return back to a more normal, what we would call the secular underlying long-term trends, and what is that number? Well, there's an interesting consensus developing on that number, which by itself would tend to make you worry. But, some of the leading economists in this area—Martin Baily, who is the head of the CEA [Council of Economic Advisers] under [President] Clinton, Bob Gordon, who I've mentioned, who has worked on productivity issues for many years, Dale Jorgenson at Harvard, Stiroh—all these people have been trying to estimate what they think the underlying productivity trend is. And they have, to a large extent, converged to a number of about 2.5 percent. Their view is that productivity over the next three to five to 10 years ought to be similar to what it's been in the late '90s, which, if true, is actually quite encouraging.

Now, as they say, past returns are not a guarantee of future results, and so the evidence for this, of course, you know, is of mixed quality. There are some indicators that productivity may slow, and I want to just point out that that is a serious possibility. Certainly ICT investments have slowed in the last few years relative to the late '90s, which is probably in some sense a good thing. But, more generally, there seem to be fewer major developments, "killer apps" [software applications] as they're called. And the improvements in computing power have, while still substantial, have seemed to have slowed somewhat recently. So, there are some indications, if you want to look for them, that technological progress has slowed moderately. That notwithstanding, I think—I personally think that the 2.5 percent range, two to 2.5, something like that is probably a reasonable baseline assumption for forecasting and for policy-making. The—again, computation in the communication technologies are general purpose technologies. We have probably not yet begun to see some of the applications that they will have. There's cross-fertilization between these applications, and other new technologies like biotechnology, so all that's very promising.

The other aspect which suggests that technology will continue to promote productivity gains is the issue of diffusion. To the extent that we can measure it, the extent to which the existing technological improvements and organizational plans have been adopted broadly, is still an incomplete process. So, for example, surveys of managers often—they often apply, but they are still in the early stage of exploiting what they see to be the benefits of these technologies, and more formal econometric studies also find a so-called technological gap—that is that the average performance at firms remains below the state-of-the-art, and so they have a catch-up potential that should continue for a period of time. So, on the medium-term prospects, I would draw, I think, two conclusions. One is that, again, that I think there's some reason for optimism, that level productivity growth in the 2.5 percent range, which, incidentally, would apply potentially to GDP [gross domestic product] growth in the sort of 3.25, 3.3, something along that range, remains a reasonable scenario. But at the same time, of course, one also has to admit that there's a lot of uncertainty around those numbers. And just to cite one example, Jorgenson, Ho, and Stiroh, a paper that got a lot of attention recently because it just—it made the projection of continuing growth at 2.6 percent was their number, so the headline was 2.6 percent, but if you read the paper, you'll see that their range of possibility goes from 1.4 to 3.2, which is broad enough that most likely they'll be right. [Laughter] So, probably good numbers, a lot of uncertainty, that's a necessity.

So, what does that mean for my colleagues and me who are trying to look at the medium term and trying to plan—make the forecasts and plan policy? In particular, if we watch the productivity numbers, what does that suggest we should do? Let me say—talk a bit first about how in the medium term productivity numbers affect the economic forecast. The key here—this is the key point, is that the effect of productivity changes on the economy depend very importantly on how aggregate spending, that is an assumption in investment, respond to the perceived change in productivity. That is, you could have a very strong response, or you could have a modest response, and the difference is going to be crucial.

Now, we know, for example [inaudible] for this country and [inaudible] a slowdown in productivity, which, if the data come in slower than expected, there's some concern that growth is going to be slow going forward, what effects will that have? Well, in the first instance, we can pretty safely say that slower productivity growth is going to reduce spending, in particular consumption spending and capital investment. Consumption will decline because lower productivity growth will reduce stock values, which will affect household wealth. It should also, for those households who are sufficiently forward-looking, they may also recognize that slower productivity growth will reduce their future wages as well. So, generally speaking, the economy will be less rich in the future and people should save more and be a little more conservative in their spending. And likewise, lower returns for capital would suggest weaker investment spending. So, clearly a slowdown in productivity will be—will cause a slowdown in economic growth.

So now—now it gets a little more complicated, because let's talk about inflation and employment. First let's talk about inflation. A lot of the discussion in the analyses one reads says, well, if productivity slows down, inflation's got to go up because slower productivity is going to raise unit labor costs. If wages are pretty sticky, productivity slows down, that means the wage cost of each unit of output is going to go up. That's going to raise costs. That's going to raise inflation. And, that's certainly possible. But it's important to understand that there's a countervailing effect, which is that slower productivity growth also reduces spending and demand, and if that effect is strong enough, you could actually get slower productivity being—reducing inflation. And particularly, what would happen would be, it would be bad for firms because what would happen is the wages and unit labor costs would go up, but because of the weakness of final demand, mark-ups would come down by even more, and so the net inflationary effect would actually be negative.

Likewise, the effects of unemployment are potentially ambiguous. If you had a slowdown in productivity growth, you know, a number of us have blamed productivity growth as being one of the contributing factors to the so-called jobless recovery of the last few years. So one might believe from that, that if productivity growth slows down, that's going to be good for employment because you need more workers. But, of course, again it depends on what happens to final output. If final output were to decline a lot, then the impact on employment would be probably adverse. You can see the differences in these scenarios if you look at the '90s and the early 2000s. In the '90s, there was a productivity gain, about a percentage point, as I mentioned earlier, and there was a very strong response in both consumption and investment, largely because of the stock market and its response. Because that response was so strong, in particular, you got employment gains despite the productivity increases because final demand overwhelmed those productivity increases. And inflation was stable to slightly higher because even though the productivity gains lowered unit labor costs, the demand factors at the other side increased inflation.

In the more recent years, we had quite a different scenario. Productivity rose again, but this time for a variety of reasons, output and demand didn't grow nearly so quickly, particularly investment spending didn't respond nearly so strongly this time, maybe because we were just a little bit chastened from the previous experience. So, with the increases in productivity, but not much increase in demand, you've got two adverse results. First, slow increase in output, but increase in productivity meant that there was no need for firms to hire workers, you get the jobless recovery. That was that important implication. And secondly, with weak increases in demand, but falling unit labor cost, you also got this disinflation phenomenon we saw in 2003, which actually had the Fed pretty concerned for a while, as you know. So, really the key issue then is how strong the final demand response is going to be.

So, anyway, so what does this imply for monetary policy? Certainly, we are going to be watching productivity very closely because it is a variable whose influences permeate the economy. But, as we see changes either up or down, in productivity, our response has to depend in part on how we see the private sector responding to that change. My guess would be, and this is only a guess, is that responses to future changes in productivity, at least in the next few years, will probably be somewhat more dampened than the late '90s, in which case the more likely scenario would be that, I'll say, a decline in productivity would probably raise inflation, would probably be actually slightly beneficial for employment because firms need more workers to meet their output. And so, the implication being that a decline in productivity growth would probably require actually some tightening or quicker removal of accommodation of monetary policy.

Likewise, if we are pleasantly surprised, as we have been fairly consistently for the last 10 years—how many times do you have to be surprised before you need to change your model?—if we are pleasantly surprised, and productivity continues to be strong, that will probably have benign effects on inflation going forward. It would probably slow a bit the interruption of unemployed labor and, therefore, the net effects in that case would be either an easing of policy, by which I probably mean more realistically a slower removal of accommodation.

So, responding to productivity changes is going to be one of the factors that helps determine how the Fed operates over the next couple of years. I should just close by saying that, of course, the world is always more complicated than theory suggests. In particular, the data are very imperfect. It's hard to tell if a productivity change is a permanent one or a temporary one. And the private sector has the same problems that we have at the Federal Reserve at making those determinations. So, it's not going to be a mechanical relationship. We shouldn't overreact to a one-quarter change in productivity, certainly. But if we do detect what appears to be changes in the trend going forward in productivity, then that will have implications for the Fed's mission, and of course, we'll have to respond to that. OK. So, I will stop there, and, John, are you going to be the—

MAKIN: I guess I'll be the—crack the whip here.

BERNANKE: The mediator.

MAKIN: Or the mediator. And thank you, Ben. That was a very thorough cover of the productivity issue. In fact, every question I was thinking of asking you, you've covered, except one. And perhaps I'll ask quickly the first question, and then we'll invite questions from the audience. I'll try to set an example by actually asking the question, and asking only one question, because probably a lot of people want to ask them. The question I have is, if the expected real return on 10-year Treasury notes is quite low, below 2 percent, and during the run-up to the tech boom in the late '90s we actually saw that expected real return over 4 percent, so I usually think of that return as a proxy for the expected real return on capital, because people can arbitrage between those investments. Question: Does the low expected real returns on risk-less assets suggest that investment may be low because expected real returns are lower, and does that tell us anything about where productivity may be going?

BERNANKE: Well, there are a lot of factors that are involved there. I don't know the answer entirely, but let me make a couple of—

MAKIN: Neither do I. [Laughter]

BERNANKE: Let me make a couple of comments. The first is the question of what the nature of the productivity gains is. According to many stories, during the last few years a lot of the productivity gains were what might be called capital saving or labor saving. That is, they involved applying existing technologies, but making a lot investments in intangible capital, that is making organizational changes, retraining, and so on. And, so the nature of those productivity gains was a type that didn't necessarily involve large amounts of new capital investment. We have not seen in this recovery exceptionally strong capital investment. So, the link between real returns and productivity gains is not perfect, and it particularly depends very much on whether or not the gains are capital intensive, capital augmenting, or whether they are capital labor savings, as we've seen more recently. So, I would just say the link there is fairly loose. I would expect eventually that as firms exhaust their potential for reorganization and so on, they'll begin to want to invest more in physical capital; I would imagine their rate of return will probably rise somewhat.

The other factor just to keep in mind, and this will probably relate to the international question, is that there's an awful lot of savings in the world. The United States is the only country in some sense that is accepting international savings at this point. And not only are there other substantial current account surpluses in Asia and a more modest one in Europe, but even countries like Brazil, Argentina, and South Korea—all emerging market economies that in some sense should be accepting savings from the rich countries, have current account surpluses. So, there's a lot of savings in the world. Countries can't seem to make good use of that domestically. And so a lot of that savings is being made available for U.S. investments. So, when you think of supply and demand for savings, since there's a lot of savings available, then the real return is going to tend to be low. So, the bottom line is, I don't fully understand what's determining rate of return right now, but there are some special circumstances, including the nature of the technical change, and the availability of savings in the global economy.

MAKIN: Thank you very much. Let us, we'll have to have people who ask questions identify themselves. Are we going to have a microphone provided? Please, when you get the microphone, identify yourself, and go ahead with your question. This gentleman right here.

QUESTIONER: John Baker. Could you comment on just the open economy we have and the impact of either outsourcing or the movement of lower wage jobs elsewhere, and how that impacts your view of productivity enhancement?

BERNANKE: Well, in this case, two parts. Let me focus on the productivity part. The effects of outsourcing on productivity are actually somewhat ambiguous. The main—it depends on some subtle issues in national income accounting, which I don't necessarily want to get into, but to a first order, a lot depends on the type of job that's being outsourced. If you're outsourcing low productivity jobs so that the average worker in the U.S. is a more skilled worker, then that actually tends to raise measured productivity in the U.S.; whereas, if you are outsourcing more high tech jobs, that tends to lower average productivity just arithmetically because of the mix of workers and the jobs that you have. So, that's sort of the theoretical answer.

I would say that I don't think that—I would say "offshoring" is probably the better word than outsourcing, because outsourcing involves even within the country. I don't think that offshoring is a major factor affecting U.S. productivity. So, this is, in some sense, a theoretical discussion.

MAKIN: Right here.

QUESTIONER: Thank you. I'm Steve Robert of Robert Capital Management. If I could go back in time a little bit, maybe to the early '90s, I remember a number of meetings with well-known economists who were trying to figure out why our productivity over the previous couple of decades had lagged behind what it was before then. And I think there was a fairly good majority of opinions, substantial majority of opinion, that pinned it on the fact that our national savings rate had gone down. And, in fact, if you don't have savings, you don't have investment, putting aside for a minute the fact that you can get from foreign investment, although a lot of that is not in fixed and long-term investment. And, there were a number of components of this decline in savings, including the consumer, including corporations, but by far the big culprit was the government, whose deficit had gone from a very small percentage of GDP to maybe 5 percent or more of GDP. And that was just drawing savings out of the system, and drawing investment out of the system.

So, there was, as I say, a pretty substantial body of opinion at that time who said, "Oh, yes, it's very obvious, it's the decline in savings, particularly because of the federal government in the United States that causes the decline in productivity." Sure enough, as the '90s moved on, the government moved from substantial deficit even into surplus, and there was a big, very big, increase in investment. It did go into ICT, but we assume in a free market it just went into the area that promised the best rewards. With the government deficit now approaching 5 [percent], 6 percent again, why wouldn't we worry that we're going to go back into that lag of investment because of the decline in savings and, therefore, have much lower productivity as we go forward?

BERNANKE: Can I answer that true or false? [Laughter] There's a number of interesting points that you raised. One is that—well, let me just try to respond to a couple of the points. One is that labor productivity is really, in some sense, the sum of two components. One is the ratio of capital to labor, capital intensity of production, which is what essentially you're talking about, but the other is what's called multi-factor productivity, which is in some sense the pure technological change that's occurring which allows a given combination of capital and labor to produce more. Now, I believe, in the last 10 years, both multi-factor productivity gains and capital deepening, as it's called, an increase in the intensity of capital-labor ratio have contributed to productivity. I think, as I recall, of course I was in grade school at the time, as I recall, the slowdown in the '70s and the '80s was more concentrated on the multi-factor productivity side than on the capital-labor ratio side, and there was a lot of discussion about such things as an exhaustion of new opportunities after the spurt, the post-World War II spurt, that used up a lot of the opportunities that existed at that time.

So, I'm not sure that the saving rate issue was ever the mainstream view of the productivity slowdown. I think people talked about energy, they talked about the exhaustion of existing opportunities, they talked about—at that time there were issues of social change and regulation, other things that were talked about. Going forward, it's important to have sufficient investment, and it's certainly true that national savings in the U.S. is relatively low, which is, of course, one of the reasons why we have the current account deficit.

Let me just make a few comments—one is that the savings in the savings, if you look at savings from a wealth perspective, then it looks like a very different picture because of stock prices, and housing prices, and so on. National wealth has increased actually quite substantially in the last few years, and that's part of the reason why people don't see much need to save out of their current income. The other comment to make is that, of course, that in the globalized world, savings from around the world can mobilize to make investments domestically, and that's been happening, we've had a lot of inflows of foreign capital, foreign savings. So, the low savings rate raises a lot of issues. But, I don't think it's going to inhibit productivity growth per se, because we are able to tap global savings in order to make the investments we need to make.


QUESTIONER: Thanks. Bob Hormats, Goldman Sachs. You mentioned a very—

BERNANKE: Can I ask you a question? [Laughter]

QUESTIONER: Go ahead. First of all, thank you for an extremely interesting presentation. You made a very interesting point at the outset, where you talked about how well we responded after 9/11. Let me ask you to expand on a couple of points. One, after 9/11 there was a considerable amount of argument that, well, we would have to spend more money on security-related matters, which would take money out of the productive part of the economy and make it less productive. And we've actually seen that to a degree, and I'd be interested in your reaction as to how that has affected productivity.

But the broader issue now seems to be that the culture openness with respect to foreigners has given way to a culture of suspicion. It's harder to get [inaudible] visas, as you know. It's harder for foreign students to come in, and being a former academician, you know how important they are to research and to the whole university system here. Looking at things over the medium term, how is that going to affect productivity, and should we be concerned, and should [we] be thinking about ways of, within the realm of the security requirements of the country, being a little more open, A, because of the social benefits, and B, because these kinds of restrictions, particularly on the inflow of people, do impact productivity in the medium term. I suspect you would agree with that.

BERNANKE: You'd be right. First of all, on the security cost, there was a lot of discussion around 9/11, as I recall, as to how firms and government were going to have to invest a lot. Part of the Federal Reserve's budget increases in the last—as you can imagine over the last four years, have been going from a sort of night watchman type of security detail to a full police force. If you've ever tried to visit the Federal Reserve, you've seen the changes. And now we're doing—all the banks have done major infrastructure things. So there is a lot of money being put into that, both by private firms—of course, you know the Federal Reserve has asked financial institutions to work on their business continuity, having backup sites. There's a lot of money involved in that. So that is a drag, but for what it's worth, it doesn't seem, obviously, to have made that much difference. It may be partly because of bad accounting reasons—some of this stuff gets counted as output, which is, of course, a mistake.

The second part of your question is much more serious. I'm still on the faculty at Princeton University. We have a top class economics graduate program, we have of course, many good graduate programs, computer science, and other fields. To an increasing degree, Americans are not even the plurality. I mean, that is, there will be other national groups that have more people in the graduate program than Americans. Not just that Americans are the minority, they're not even a plurality. It is an important issue in economics. The Federal Reserve has great difficulty hiring people, hiring Americans to meet the security concerns and so on. So I think a very important part of the productivity gains in the last decade was associated with our open immigration policy. It was said that Silicon Valley was the place where Indian engineers use Japanese money to produce for the European market. [Laughter] So if we don't allow—if we don't make provisions for bright people, whether they be graduate students, undergraduates, or professionals to come, that's a big loss. That's a loss to our society, and it's a loss to our potential productivity. I certainly recognize, now that I'm a government official, I certainly recognize the importance of security issues. They're not to be taken lightly, but to the extent that we can find ways to do the necessary due diligence and bring good people in, I think we're missing a very important opportunity.

QUESTIONER: David Bard, McKinsey and Company. You spoke in some detail about productivity levels and changes in the U.S. and Europe. I was hoping you'd speak to the same situation in East Asia, specifically China, Japan, and South Korea, if possible—any implications that might have for our economy.

BERNANKE: Well, it differs a lot by groups. As you know, East Asia, the Tigers, had a remarkable burst of productivity some decades ago, and they continue to be very productive economies. They are moving towards essentially membership in the world group of industrial nations. As I mentioned, the odd feature is that countries like Korea and Thailand and so on are currently exporting capital, rather than importing capital. I think that is going to be something of an issue for them, going forward. They also are somewhat less diversified than countries like the United States. I mean, for example, East Asia rises and falls to a considerable extent on the high tech demand for—because they are high tech producers. But I think, broadly, the East Asian countries—I 'm including now Japan and China—are real success stories in terms of their increases in income and productivity over the recent decades. Japan, as you know, has been more slow. It's a mature economy. It's had macroeconomic problems, including banking issues and so on. We're hopeful that recovery will be sustained in Japan. But, right now there are still questions about how strong and sustained that will be. China—of course, productivity is growing rapidly in China. This is less, I think, an issue of technological change than it is of a change in the composition of what they're doing. I mean, they're becoming industrialized. This is something the U.S. went through when we became, from agricultural to industrial to finally post-industrial. And China is doing that same path. Their productivity is growing rapidly. They're becoming much richer. And of course, this is the Council on Foreign Relations, you know. We can discuss Asian policy for a long time, major implications, political and diplomatic, as well as economic. But, that clearly is right now where the action is in East Asia and China, where the society is transforming itself, and accordingly becoming much richer. So there's a diverse set of stories there. I don't think any of them are quite parallel to what the U.S. has been experiencing, because they're different sources of productivity growth in different areas.

MAKIN: This gentleman right here.

QUESTIONER: Jeff Shafer, CitiGroup. Let me shift to the short run. I'll try to ask a question that you can answer. If you look at the immediate past, specifically the fourth quarter, things look somewhat slower in the U.S., in Japan, in Europe, maybe even the rest of Asia. And I'm wondering if you would reflect on the dynamics of that, and what you see unfolding going ahead globally?

BERNANKE: Well, as an arithmetic matter, the fourth quarter now appears to have been somewhat less—statement of fact, the fourth quarter now seems to have grown—output in the fourth quarter seems to have grown a bit less quickly than we were expecting. Estimates for the fourth quarter range around 3 1/4 percent real GDP growth on an annual basis, something of that sort. But, arithmetically speaking, a good bit of that arises from the surprisingly large current account or trade deficit. That is, average demand, spending in the fourth quarter of the United States remained quite healthy. It was, for example, a remarkably strong quarter for automobile purchases, to take an example. Housing remained strong. The Christmas season in the end turned out to be a healthy one. So I think the problem is not so much aggregate demand in the U.S., but rather the draining of aggregate demand abroad. And that, of course, raises questions about next year—this year as well.

So I think the question then is—there's two reasons why current account can worsen—one is that foreign growth is slower than we expect, and the other is some kind of trend against U.S. products, whatever the compositional effects might be. So what this all boils down to, is I think that for growth next year to be what we hope it would be, that is above trends, maybe 3 1/2 percent or something better, we need for the trade balance to be no worse of a drag, let's say, than it's been recently. And that, in turn, requires that there not be a significant slowdown in the international growth picture. So the international growth picture is the major risk factor for the U.S. going forward. I think the prognosis differs by region, but I think there is some reasonable expectation of a decent performance in Europe, and some improvement in Japan, and continued strength in China. So our hope is that the trade balance will not worsen and not be more of a drag in '05 than it was in '04. If that's the case, I think the prospects for reasonable growth are pretty good.

The other similar factor is oil prices, which rose substantially during '04, and were a fairly significant drag on real GDP growth. Even if oil prices remain high, that is if they don't fall, they remain high, as long as they don't continue to rise at something like the pace of '04, then that, too, ought to be a net plus to aggregate demand and growth in the U.S. So those are the two major risk factors I see. One is the oil prices, and the second is foreign demand for U.S. goods. If those two things come in—if they don't worsen, then the growth next year ought to be reasonably good.

MAKIN: Thanks very much. We have time for one more question. But, before I take that question, I just want to thank Governor Bernanke for coming and talking to us here, and thoroughly answering all of our questions. So I'll go to this side, this gentleman back over here. Then after this question we will be finished.

QUESTIONER: John Hartzell, KWR International. Governor Bernanke, do you see any issues in the taxation area, either tax rates or tax structure, that ought to be addressed if we want to do things to sustain strong growth and productivity over the next several years?

BERNANKE: Well, a number of the questions, earlier questions, turned on the issue of national savings, and the importance of savings to promote capital formation in the U.S. That is certainly one major concern the fiscal policy ought to address, and that would involve, I think, both trying to reduce the current budget deficit, and also trying to make progress on some of the long-term fiscal obligations, like Social Security and Medicare, for example, by trying to further pre-fund Social Security, so as not to reduce the imbalance in that program. So the deficit, both short term, and long term, is obviously an important issue for generating national savings.

I know, of course we all know, that the president has set up a commission of distinguished individuals from all different walks of life to look at the tax code. My hope is that what comes out of that will be—and out of the legislative process, will be a tax code that is essentially pro-growth, not over complicated, doesn't overtax productive activities, that it broadens the base, and that it promotes economic efficiency. Now there are a number of different models for doing that, and I'm probably not really prepared to choose one or to give you any details. But, I hope that when that program is analyzed, that the growth implications are very carefully discussed.

MAKIN: Governor Bernanke, thank you very much.

BERNANKE: All right. Thank you. [Applause]






More on This Topic


Stimulus in a Volatile Financial World (Audio)

Speaker: Justin Yifu Lin
Presider: Harry G. Broadman

Listen to Justin Yifu Lin, chief economist and senior vice president at the World Bank, discuss how stimulus measures may be applied to boost...


Playing Monopoly with the Devil

Author: Manuel Hinds

Playing Monopoly with the Devil offers sound, practical advice for policymakers on how to deal with the currency problems that developing...