Nouriel Roubini and David Backus

Lectures in Macroeconomics

Chapter 4. Productivity and Growth

Productivity and Growth

The differences in per capita GDP across countries are enormous, and the differences in growth rates suggest that rankings of levels can change dramatically. Note, for example, that Korea and India had comparable levels of income in 1960, while in 1985 Korea's per capita GDP exceeded India's by a factor of 4. At this rate of growth, per capita GDP doubles every 12 years. Clearly there is something going on in Korea that India hasn't figured out yet. Or consider Italy: In 1870 per capita income was 60 percent less in Italy than in Great Britain, but after more than a century of growth one-half of one percent larger, Italy has now caught up. This is the hoary old "power of compound interest" cliché, which shows that even clichés can contain basic truths.

Our objective today is to try to put this into some perspective (which is asking a lot for one lecture). There are no simple answers, but for a subject this important I think it's useful just to throw out some ideas. One of the things I find interesting about this subject is that many of the issues that arise when you think about countries also pertain to the performance of firms.

Inputs and Outputs: The Production Function

Y = A F(K,N),

where Y is output (real GNP), K is the stock of physical capital (plant and equipment), and N is labor (the number and hours of people working). The letter A measures what we will call productivity. A higher value of A means that the same inputs lead to more output, as clear a definition of productivity as I can think of. [Despite this, the word productivity is used in many different ways. When you run across it in other contexts, your first order of business is to find out exactly what it means.] For future reference, we'll refer to A sometimes as total factor productivity, to distinguish it from, say, average labor productivity, Y/N.

As we'll see shortly, we owe a lot of the growth in the US economy (and other economies, too) to increases in A. For now, let's think of things that might affect A. (i) Technological progress can be thought of as increases in A: invention of the diesel engine, the transistor, the microchip, penicillin, and so on. (ii) The skill level of the labor force is another thing that might be incorporated in A. One of the big differences between rich and poor countries is that the former have better educated and more highly skill workers. For this reason, we would not expect NAFTA (the North American Free Trade Agreement with Mexico and Canada) to result in American wages falling to the level of Mexican wages. (I'm afraid Ross Perot was out to lunch on this one.) (iii) Oil prices. We've seen that an increase in the price of imported oil may leave us with lower GNP, other things equal, since a greater fraction of gross output goes to oil, less to capital and labor (hence less value-added). We can think of this as a downward movement in A (and, in fact, that's what we see in the data). (iv) Weather. A drought or extreme cold snap might lead to lower output for given inputs. Droughts aren't a big deal in the US economy, since agriculture is a small part of the economy, but it gives you the idea that lots of things might affect A. (v) The economic and legal environment might also play a role in aggregate productivity. Most economists think that competitive markets play an important role in allocating resources in an efficient manner, and this kind of thinking is behind many of the changes in the former Soviet Union and Central Europe. Conversely, corruption and red tape are often given much of the credit for India's lethargic performance.

Sources of Growth

This section is dedicated to defining and measuring productivity, and using a theoretical framework to determine the sources of economic growth. The fundamental relation in productivity measurement is, again, the production function:

Y = A F(K,N).

This function says, as we've seen, that we get higher output for three reasons: because more people are working (higher N), because they have more equipment to work with (higher K), or because capital and labor are used more productively (higher A, a catchall category). What I'd like to do is decompose growth in real output Y into components due to each of these three elements -- an exercise that has come to be called growth accounting. If the production function has the form

Y = A K^1/3N^2/3,

this decomposition has a particularly simple expression. [Don't ask! The exponents mean that one third of output is paid to capital in profits, depreciation, etc., and two-thirds to labor, which is about what we see in the US and many other countries.] Then growth in aggregate output Y is

dY/Y = dA/A + 0.33 dK/K + 0.67 dN/N, (*)

where dX/X represent the percentage rate of change of variable X over the period considered (for example one year):

dX/X = (X_t - X_t-1)/X_t-1

In practice we know all the terms but A, which we compute as a residual. We generally do not measure A directly, which adds to its enigmatic character.

We can, with little change, use this to account for growth in output per worker, Y/N, which is more directly tied to living standards than output. Given the structure of the production function, this can be written

Y/N = A (K/N)^1/3 ,

and growth decomposes into

d(Y/N)/(Y/N)= dA/A + 0.33 d(K/N)/(K/N).

What this says is that output per worker can rise for two reasons: because total factor productivity A is increasing and because the amount of capital per worker, K/N, is increasing.

We can get some idea of why the US economy grew over the last thirty years by examining the components of (*). Growth of real output over the 1960 to 1985 period was 3.1 percent per year. [This is higher than the number above, because we are talking about total output here, not output per capita. There are also some minor differences in how the numbers were constructed.] How much of this growth was attributable to increases in inputs? From the data we find that the growth rates of labor and capital over the same period were 1.9 and 3.2 percent, resp. Applying the coefficients as indicated in (*), we find that this gives us about 1.1 (= 3.2 x .33) percent growth per year due to increases in the stock of capital and 1.3 (= 1.9 x .67) percent due to increases in the number of people working. What's left over, so that the two sides of (*) balance, is about 0.7 percent growth in productivity A. To be honest, it's not really this easy to separate the three components: many advances in technology show up in either capital (new and more productive machines) or labor (better trained workers). But this gives you some idea of how important productivity is in aggregate growth. Figure 1 and figure 2 present the same data in a different way, as the annual values of (the logarithm of) A and its rate of change. Note for future reference the spikes: there are large short-term movements in A associated with business cycles.

This emphasis on productivity is well placed, I think, but there are cases in which other factors are more important. Following World War II both Germany and Japan suffered massive losses to their capital stocks (think of Kuwait or Iraq following the Gulf War). In the process of catching up with the rest of the world, they had very high rates of capital growth that led to high rates of output growth. In the case of Japan, this process continued into the 1970s. To see this, let's do the numbers. From a different dataset (constructed by the OECD) we have
 

Digression on growth rates. I've used two properties of logarithms to get these numbers. The first is that, for any two numbers x and y

log(x^y) = y log(x)

here log stands for the natural logarithm (sometimes denoted ln), which is what we'll always use in this class (no unnatural logs allowed!). In words, the equation says that logarithms change powers into multiplication. The second property is that the logarithm of 1+x, where x is small, is approximately x:

log (1+x) = x.

You can verify this on your calculator for x = .03, say.

Here's where the growth rates come from. The annual growth rate g for US output is defined by

3103 = (1+g)¹⁵ 2083.

Using property one we get

15 log (1+g) = log (3103/2083)

or

log (1+g) = [log (3103/2083)]/15 = .0266,

so g = 2.66 percent per year. You might verify some of the other growth rates to see that you've got it. Both of these properties of logarithms are useful when we're dealing with compounded growth rates or compound interest.

Back to our problem. In levels (as opposed to growth rates) we see that the US was much richer than Japan in 1970, in the sense that it had much greater output per worker: 26.5 (thousand 1980 dollars per worker) vs 17.5. Where did this differential come from? One difference is that American workers in 1970 had three times more capital to work with: the ratio of K to N was 108.6 in the US, 36.4 in Japan. If we use our production function, we find that productivity A was also slightly higher in the US in 1970: 5.64 vs 5.35. Thus, the major difference between the countries in 1970 appears to be in the amount of capital: American workers had more capital and therefore produced more output, on average. Of course, you lose a lot of information in such aggregate comparisons (comparisons by industry show Japan more productive in some, the US in others), but it gives you some idea what's been going on.

By 1985, much of the difference had disappeared. It's obvious from the numbers that the biggest difference between Japan and the US over the 1970-85 period is in the rate of growth of the capital stock. From the basic growth accounting equation, labeled (*), we find that for the US the output growth rate of 2.66 percent per year can be divided into 0.93 percent due to capital and 1.26 percent due to employment growth. That leaves 0.47 percent for productivity growth. For Japan the numbers are 2.48 for capital, 1.08 for labor, and 1.13 for productivity. The largest difference between the two countries is in the rate of capital formation: Japan's capital stock has grown much faster than the US's, raising its capital-labor ratio K/N from 36.4 in 1970 to 88.0 in 1985. These numbers are summarized in the following table:
 

Nevertheless, some of the difference in growth rates is associated with pure productivity. By our measure, Japan enjoyed an advantage of 0.66 percent per year in productivity growth over the US in the 1970-85 period, and by 1985 enjoyed a slight advantage. This result is to some extent due to the data set I've used. As always in economics, it's best not to make too much of small differences in fuzzy data. Most studies now suggest that the major industrialized countries (the US, Japan, Germany, and so on) have roughly comparable productivities when measured by the best available methods, with the US in the lead. This is a major change from the 1950s and 1960s, when there were still large productivity differences between these countries.
 

TFP Growth in Asia
The issue of  how much output growth in particular country is due to total factor productivity growth versus growth in inputs is particularly important to understand the Asian Miracle and the recent economic crisis in Asia. In 1995,  Krugman advanced the controversial view that the Asian economic "miracle" was not due to total factor productivity (TFP) growth but rather to intensive use of inputs, i.e. a high growth rate of capital due to the high rates of investment in Asia and a high rate of growth of labor inputs given the increased labor participation rates in the region. This view was very controversial since it implied that very little TFP growth had occurred in Asia; if true, it also suggested that the very high rates of Asian growth were not sustainanle in the long run given the expected fall in the rate of growth of employment and the expected reduction of investment rates. Krugman's views were highly debated and criticized; in this regard,  read the articles in The Economist "The miracle of the sausage makers" and "The Asian Miracle: is it over?"  (both are available in the Reading Package). The economic crisis in Asia in 1997, even if originally triggered by large currency depreciations, appeared to indirectly confirm Krugman's views on the weakness of the Asian economic model and and fragility of the Asian Miracle.
 

The Productivity Slowdown Puzzle

The debate on the causes of this productivity slowdown has turned into a puzzle as the causes of the worlwide slowdown have not been clearly identified. Several explanations of the slowdown have been suggested but none has been found to be fully satisfactory (see Krugman "The Age of Diminished Expectations" Chapter 1 for a detailed discussion):

1. The energy crisis in the 1970s (1973 and 1979 oil shocks).

2. Exhaustion of the post-W.W.II technological boom.

3. Low investment and savings rate.

4. High taxation of savings.

5. Excessive government regulations.

6. Low rate of public investment in infrastructures.

7. Decline of R&D investment.

8. Sociological explanations.

9. Decline in quality of education.

One of the most likely explanations is the oil price shocks we saw in the 1970s, especially 1974 and 1979. We've seen that increases in the price of imported raw materials lead to lower value-added and GNP for any given quantity of capital and labor, so it's not surprising that sharp increases in oil prices were associated with productivity declines. See, for example, the downward spike in productivity growth in 1974 in Figure 2. Most experts agree that oil prices were a large part of the story for the 1970s. You'll note that in the 1980s, when oil prices were stable or even declining, productivity growth picked up. Another explanation for the productivity slowdown is that our measures of productivity are not that good. At some level, I'd have to agree: there's nothing fancy about what we did above. The question is whether better techniques change the picture much. For the most part they don't, but there is some question (as we noted in our discussion of national income accounting) whether quality change is adequately treated in our measures of output. A third explanation is that there is some underlying malaise in productivity. Maybe technological advances come in spurts, and we don't happen to be in one right now. Or maybe there has been a decline in the quality of education, the amount invested in research and development, or the development of infrastructure. Since productivity is so central to our economic well-being, all of these ideas deserve to be taken seriously. We'll return to some of them when we discuss policy options for increasing productivity growth.

The Mystery of the Vanishing Productivity Growth in the 1990s and Its Post-1995 Resurgence

However, the switch in 1995 to the chain-weight method of measuring productivity changed drastically the picture: the new chain-weight data showed that in the 1990s total factor productivity grew at a dismal 0.9% per year rate while labor productivity grew at a 1.4% yearly rate, not much above the 1970s and 1980s rates. So the great resurgence of American productivity in the 1990s suddenly disappeared overnight by a statistical wand.

These numbers looked dismal because many economists believed that the process of corporate restructuring, reengineering, downsizing of the last decade, together with the development and adoption of computers and information technologies in the corporate world, had led to a major resurgence of productivity. The new chain-wighted numbers seem to imply that such productivity resurgence never occurred.

In the debate that ensued in 1996, there were essentially two views. On one side there were those, like Paul Krugman, who argued that the new measures of output and productivity were substantially correct and that the productivity benefits of the Information Revolution had been overstated. On the other side, those arguing that the new chain-weight method underestimated output and productivity because, among other reasons, of mismeasurement of the growth in productivity in the service sector.

In 1997, the debate on productivity growth took a new twist as data for 1996 and 1997 appeared to show a significant increase in the rate of productivity growth. For example the latest data for the third quarter of 1997 showed that productivity was growing at a annualized rate of 4.0% in the business sector and a whopping 9.3% in the manufacturing sector (see the BLS Web Site for the latest productivity report). While the annualized growth rate data might be distorted by a particular good quarter, the actual quarter-on-quarter annual rates of growth showed similar large increases of productivity growth: the actual productivity growth between the third quarter of 1996 and the third quarter of 1997 was 2.4% for the business sector and 4.6% for the manufacturing sector, well above the dismal rates of 1.0% observed in the early 1990s and close to the high rate of the 1960s. These new data led a number of authors to argue that we had entered in a new era of sustained productivity growth; one heard a lot of talk about a  "New Economy" where a "New Paradigm" of high growth and low inflation holds. The homepages on the New Economy on Productivity Growth in the 1990s present an introduction to this recent debate. For a critique of the New Economy hypothesis, see two recent articles by Paul Krugman, one on Slate, and the other on Fortune magazine.

Returns to Education

Another factor is education, which you can think of as investment in people, or what economists call "human capital." There is lots of evidence, at the levels of both countries and individuals, that education is associated with productivity. As a rule, countries that invest the most in education also tend to be the richest and have the highest rates of growth of per capita output. Note the growth rate effect: not only are countries with more education richer, they also seem to grow faster.

Education has clear benefits to individuals, too, as your presence at Stern probably indicates. This includes formal schooling, job training, and work experience. A huge number of studies has established that each year of school tends to raise one's wage between 5 and 7 percent, on average. The numbers vary depending on the quality of school, the type of education, and so on, but there's little doubt that more highly educated workers are better paid and, unless firms are throwing their money away, more productive.

There is a clear connection between education and our catch-all productivity measure. Let us say, to be specific, that educated workers are essentially like extra quantities of uneducated workers. For example, suppose a "standard" worker with a high school diploma has a productivity of one (this is just a benchmark). Then a worker with one year of college is worth, from the studies cited above, about 1.06 standard workers. An increase in the average education level of the workforce by one year then leads to an increase in the effective labor force, N, of 6 percent. From our production function,

Y = A K^1/3 N^2/3 ,

we see that this leads to about a 4 percent increase in output and measured productivity [1.04 = 1.06^2/3]. In short, education shows up directly in aggregate productivity, and its effects are large.

One current fear in the US is that the quality of education has deteriorated in the last twenty years. This shows up in lower test scores and in frequent complaints by college teachers that their students are not as well prepared for college courses as they used to be. This is not just a problem of social policy. Companies in the US and elsewhere spend an enormous amount of money on worker training. Viewed another way, the increase in private sector training suggests that learning is no longer confined to schools or to the young. It's a continuing process.

Technology

One of the troubling trends in the US in recent years has been a decline among US firms in R&D expenditures, patent applications, and other technology indicators. If this trend continues, some fear adverse long term effects on US productivity growth.

Some Puzzling Examples

China. In many aspects of technology China was much more advanced prior to 1400 than Europe. Mokyr mentions: (i) Agriculture: rice cultivation, the iron plow, and so on made Chinese agriculture much more productive than Europe's. (ii) Iron: the Chinese had blast furnaces for casting iron in 200 BC, Europe not until late 1300s. (iii) Ship-building and navigation: the great voyages of the European explorers were predated by the Chinese who, inexplicably, prohibited foreign exploration just as the Europeans got moving. There are many other examples (paper, porcelain, the cross-bow, gunpowder), too. On the whole, they suggest a culture with advanced technology in both pure and applied areas.

In short, China was technologically advanced in 1400, but this didn't lead to an industrial revolution. It's not clear why, but a leading hypothesis is that the centralized system of government was prone to malfunction with inadequate leaders. An ambitious leader could encourage innovation and development, or throttle it completely. Regardless of the explanation, it's clear that it takes more than technology to produce sustained growth in output and productivity. Science and technology alone won't do it.

The Arab World. Maybe some other time. The general idea is that they had better understanding of math and some aspects of technology than the Europeans in the middle ages and before, yet didn't develop, as Europe did, into a successful industrial society.

I could name some other examples. Argentina was as rich as almost any country in the world in 1890, but is far from it now. Japan, on the other hand, has grown dramatically from shortly after the Meiji restoration (1868) to the present, truly a remarkable and, I think, unprecedented achievement. In 1868 they were internationally isolated but still a highly educated society, and thus in a good position to capitalize on Western technology once they were exposed to it. Britain, on the other hand, developed more quickly than France in the 18th century, despite a substantial disadvantage in pure science. They were experts, however, in practical engineering (eg, the steam engine).

This lesson applies to firms, too: advanced technology often fails, for a variety of reasons, to translate into business success. A striking case is IBM, who for years had the best pure and applied research. The corporate culture, however, did not exploit these advantages as well as they might have. One example is RISC based computing, the basis for the exploding work station market. IBM invented this technology in the mid-1970s, but lagged far behind Sun and other smaller companies in bringing it to market. Evidently technology alone isn't enough. Another example is GM. Although technology is arguably more important in manufacturing than in services (and in modern economies services are increasingly important), the advances GM has made are the result more of management methods than technology. GM's experiment in robotics is termed by Maryann Keller (Rude Awakening, HarperCollins 1989, ch 10) as "an experiment that failed," and most observers agree. A similar point is made by Womack, Jones, and Roos in The Machine that Changed the World (HarperCollins, 1991). Their figure 4.11 (p 97) compares Ford's Atlanta assembly plant with GM's Fairfax plant. They found that the GM plant, despite a higher degree of automation, had substantially lower productivity than the Ford plant. They attributed the difference to design (the GM car had many more parts) and plant organization.

In short, good economic performance, for both countries and firms, appears to involve features of the economic and social structure that are difficult to define, let alone measure. We shouldn't be surprised, then, that strategies for raising productivity and growth cover almost every aspect of economic and corporate organization.

Management Strategies: Cooperation and Competition

One of the interesting trends in management philosophy has been toward a greater emphasis on cooperation. The increased weight placed on group work at Stern is an example you may have noticed. At some level the benefits of cooperation are obvious: you should have the quarterback and the wide receiver running the same play. In modern management the suggestion is that there must be active cooperation among the entire production team, from assembly line workers on up to the CEO. Most of these methods require active participation by the people on the line to work, since they are the closest to the process and thus know the most it (hard as that is for senior management to believe). (The center, for example, might know more about what the opposing lineman are doing than the quarterback.)

As an economist I find this emphasis on cooperation fascinating, because we tend to focus on competition---in some ways, the antithesis of cooperation. Deming, for example, argues that firms should have a small number of suppliers, because only then can they enforce quality standards. The competitive approach, followed by US automobile companies for years, is to use many suppliers, so that you can use competition among them to keep the price low. So who's right? I think most people would agree that competition can be an effective tool. The former Soviet Union, for example, would probably have been more productive if incentives had led to greater competition among individuals and firms to supply goods and services. And at IBM, competition between divisions, rather than enforced cooperation with the mainframe division, might have spared them their poor performance in non-mainframe businesses. But some management studies suggest that competition between workers in similar jobs can actually lower productivity. Or that competition among students in a course can reduce the value of their educational experience. The question is where you draw the line. Should IBM be one firm or many? Should Citibank centralize its loan operations, or have several competing divisions? Should computer manufacturers join "cooperative" research programs or go it alone?

The question of where to draw the line between cooperation and competition is not one with a simple answer, but it's one of the basic issues in business and economic policy. For management it involves strategic decisions (enter a new business alone or with a joint venture?), performance appraisal (should individual members of a team be rated and rewarded differently, which might foster competition but discourage cooperation?), etc. For policy it involves questions of anti-trust (treatment of joint ventures to allow more cooperation), tariffs (should we protect ourselves from foreign competition?), and foreign investment (should we restrict foreign purchases of hi-tech firms, a restraint of international competition?).

Economists, by and large, find that competition among firms has been useful. But maybe there should be room for cooperation, too. The tension between these two forces is a continuing theme in management, and I think you'll find that it reappears in your management courses in other forms.

Policy Options

Saving and investment. The first item on most lists is that we could do more to encourage investment, either directly through tax incentives (reduced disencentives?) or indirectly through incentives to increase saving. Recall that saving and investment are connected through the identity: S = I + CA. Generally CA is small relative to I and S, as we have seen, so most investment is financed through domestic saving. We might guess, therefore, that policies to raise saving will also raise investment. The question is how to do this. The 1989 Economic Report of the President suggested a cut in the capital gains tax, lower corporate tax rates, higher limits on IRA contributions, and (this may be the big one) a smaller federal government deficit. Missing from their proposal is something many economists think is the largest distortion in the US tax code: the favorable tax treatment of housing. What seems to happen in the US is that we invest a disproportionate amount in housing, rather than plant and equipment. Since it's the latter that makes us more productive, we may have a lower standard of living (but nice houses!) as a result. It's unlikely that preferential treatment of housing will disappear in the US, but expanded tax-sheltered saving plans are a possibility. We will discuss more these issues in Chapter 5.

Education. Another method of boosting productivity is to invest in people, which has payoffs to both the individuals and the economy as a whole. The evidence we've seen suggests that this can have large returns. The question is how to go about it. I see this as primarily a management issue: how to deliver high quality education on a large scale.

Infrastructure. One of the most important things governments can do, it seems, is make sure the economy has a good infrastructure, esp transportation and communication. Some of this is done by the private sector, but in most countries at least part is the responsibility of the government. Good roads, for example, make it possible for firms to centralize production and exploit economies of large scale production. Communication equipment is vital in many fields (think of finance: where would you be without up-to-the minute information). Historically these have been US strengths, the question is whether we have allowed our roads/airports to deteriorate in an effort to save money short term, or fallen behind in the adoption of new communication technology (a fiber optic network). We may find that investment in these things has large payoffs. The question is which ones.

Scientific research. The US has, over the last fifty years or more, been the world leader in pure science (often, though, with imported talent). If you look at expenditures on what is called research and development narrowly defined (the people in the white coats again), the US spends as much as anyone (measured as a fraction of GDP). But unlike (say) Japan or Germany, part of this money is spent on military applications, and only has civilian value by accident. When we subtract the military, we spend somewhat less than these two countries. We just got done arguing that basic science isn't enough, but that's not to say that it's not useful. So one of the things we might ask ourselves is whether we should be doing more to encourage research, both publicly and privately funded.

Taxes. One of the things you often hear in the US is that high tax rates discourage good economic behavior (saving, for example). That may be true, but the US is, on the whole, a low-tax country. The success of Germany and Japan, where taxes are higher, on average, suggest that taxes aren't our major problem. In fact, we might be better off with somewhat higher taxes and better services (choose from the list above). What might be true, though, is that the structure of the US tax system is inefficient, that we are discouraged from saving and encouraged to put too much of our wealth in real estate. We might come back to this later in Chapter 5 (but let me warn you now that there is no consensus on this).

That's the A-list. A few other things have gotten attention in the US recently, and maybe they're important. Health care, for example, has gotten to be such an important factor in hiring by firms and job decisions by individuals that we have to do something about it (but what?). It seems to me that almost anything would be an improvement. A second issue is the torts system, which may discourage innovation by saddling firms with large and highly uncertain liabilities in new product development. The question is how to balance the incentives for innovation with the incentives to provide safe products. Third, there is some concern that US anti-trust law, particularly the treble damages aspect, may discourage firms from adopting joint projects. The evidence here, though, is that large projects are not generally more effective than small ones, so this may be a red herring (despite the recent wave of cooperation in computing, like that between Apple and IBM). Moreover, large Japanese firms, when you look at them closely, show no particular desire to cooperate, so that's not their secret.

What these policies tell you is that everything is a productivity issue. You don't have to get them all right, but you must get enough of them right to continue to enjoy growth in aggregate productivity and wages. Any country that doesn't faces the possibility of becoming the next Argentina or, as a less extreme example, Great Britain. There are no simple answers, but the stakes are big enough to make the question worth thinking about.

Summary

1. Growth in output comes from increases in factor inputs and growth in productivity (the letter A in our production function). A major reason why Japan has grown more rapidly than the US over the last thirty years has been their larger rate of investment in physical capital (I/Y).
2. Productivity is related to investment in education, research and development, management techniques, and simple experience...
3. ...but there are a number of examples of societies that were technologically but not economically advanced. Evidently technology is not enough, you also need a social/political/cultural/legal environment that fosters its application.
4. Analogously, many productivity increases at the firm level reflect management and organization of people more than advances in pure science.
5. A variety of national policies are likely to lead to higher productivity: education, transportation and communication infrastructure, basic and applied research, and clearly defined property rights.

Further Reading

There's been a lot of work on productivity at the level of both firms and countries. With regard to the former, half the books in the business section of most bookstores include the word quality. Andrea Gabor's The Man Who Discovered Quality is an interesting and highly readable review of Deming's work. With regard to national productivity, Williamson's article ("Productivity and American leadership," Journal of Economic Literature, March 1991) is a little difficult, but will give you a good idea of the range of economists' opinions on sources of growth both historically and around the world.

Further WEB Links and Readings

Read the controversies on Productivity Growth in the 1990s, the New Economy, Inflation and Output Mismeasurement and the NAIRU. See also the additional WEB readings on Productivity and Growth in the home page on Macro Articles and Analysis.

Copyright: Nouriel Roubini and David Backus, Stern School of Business, New York University, 1998.