Printer Friendly

The catch-up factor in postwar economic growth; Presidential address to the Western Economic Association, June 21, 1989.




JUNE 21, 1989


This talk is about two of the central concerns of our time. One is the pronounced slowdown of productivity growth that has persisted in the whole of the industrialized world since the early Seventies or late Sixties--about two decades. The other is the much slower rate or productivity advance in the United States than in other industrialized countries than has been apparent during almost the whole postwar period--about four decades. [1]

I shall argue that these two developments are connected and that both can be better understood if we think of them as implications of a process of international productivity catch-up and convergence that is, in certain conditions, natural and foreseeable and, in a long-term sense, desirable. I mean desirable not only for the countries that are catching up. That is obvious. I mean desirable also for the old leader, that is, for the United States as well. Needless to say, I do not contend that a tendency to convergence is the whole story behind the two developments and the anxieties they have caused, but I believe it is a large part of it. [1]


Let me begin by asking why and what conditions a tendency to convergence is, as I say, in the nature of things. Consider two countries otherwise similar, but one is a productivity leader, the other a laggard. There are at least five reasons why the laggard will have a stronger potential for productivity growth than the leader.

First, as the leader's capital stock turns over, its embodied technological progress is limited by the advance of knowledge over the lifetime of its capital instruments. But a laggard can make a larger leap by substituting modern, state-of-the-art equipment for instruments that were obsolete even when originally installed. The same applies to modernization in disembodied technology, that is, when a laggard adopts a leader's advanced practices in the organization of production and management.

Secondly, laggards typically suffer from relatively low levels of capital intensity. On this account, as well as because new capital means a large technological leap, the inducement to invest tends to be high. So there is a potential for rapid progress by capital accumulation (Edward F. Denison [1967, Table 21.28]; Robert E. Lipsey and Irving B. Kravis [1987]; Edward N. Wolff $(1987, Table 31$)).

Thirdly, the level of schooling in laggard countries is typically relatively low and, in some degree, obsolete in content. So there is an opportunity for relatively rapid advance by the extension and modernization of education (Maddison $(1982, Table 5.7$); Denison $(1967, Table 21.28$)).

Fourthly, in laggards one usually finds a relatively large fraction of partially or wholly redundant workers attached to farming and petty trade. This is an opportunity for relatively large gains from the better allocation of workers (Denison $(1967, Table 21.28$)).

Fifthly, the chance for rapid productivity advance along all these channels means a chance for rapid growth in aggregate output and size of markets, which brings a productivity bonus from the economies of scale.

This is the simple core of the tendency to convergence. And if that were all there was to it, we should expect to see laggards "always and everywhere" advance faster than a leader. Rates of productivity growth across countries in any period would be an inverse function of their initial levels of productivity, and national levels of productivity would converge towards the leader's. By the same token, as the process of convergence went on, the gaps separating laggards from leaders would become smaller, and rates of growth would decline. The catch-up effect supporting growth would be self-limiting, weakening steadily as catch-up proceeded.

But things are not so simple; there are limitations. When I spoke of two countries otherwise similar, I had in mind two conditions. The first is a somewhat vague, hard-to-pin-down set of characteristics that others and I sometimes call "social capability." It refers, first, to a country's political institutions, its political integration and the effective consensus in favor of development. These matters bear on the ranking of economic activity and of material welfare in the scale of social values, the social sanctions that protect earnings, property and honest trade, and the willingness and capacity of governments to create the physical infrastructure for private activity. Next, there is a country's technical competence for which, at least among Western countries, years of schooling may be a good proxy. Finally, there is the state of its private economic institutions, which means experience with the organization and management of large-scale enterprise and with financial institutions and markets capable of mobilizing capital for individual firms on a large scale. Social capability is what separates less developed countries from advanced countries today and which, in the past, separated the late-comers among the countries that are now industrialized from the early entrants into what Kuznets called "modern economic growth." The upshot is that a country's potential for growth is strong not when it is backward in all respects but rather when it is technologically backward but socially advanced.

My name for the second condition is "technological congruence." The path of technological progress by borrowing from leaders is not equally open to countries with any mix of raw materials, labor skills, capital intensity and scale of operations and markets. It is rather biased in its characteristics and adaptable to different conditions of resource supply and market scale only at a cost. So a country's technological congruence limits its ability to exploit the full potential for rapid growth envisaged by the simple theory of convergence.

There is something still more. The simple theory speaks of potential, but that means potential in a long run. How rapidly the potential will be realized in any limited period is a matter of economic response to the opportunities that technological backwardness holds out. That depends partly on the conditions that govern the diffusion and acquisition of knowledge, professional publications and contracts, travel, trade, multi-national corporate operations, the laws and regulations that govern technological transfer and the protection of knowledge, and the effort expended on research and development itself. The rate at which countries can borrow and adapt the products and methods of others, no less than the advance of practical knowledge itself, depends on R and D.

The realization of potential also depends on the mobility of people and on other labor market conditions that control people's response to the structural changes in occupations, industries and location that growth entails. And, perhaps most important, it depends on the macroeconomic conditions that govern investment and capital accumulation. So the size of productivity gaps defines a growth potential, but the pace of realization of potential is subject to fluctuations from period and perhaps also to secular drift according to a set of other conditions that control the rate of realization and that are at least partly independent of the catch-up potentials themselves. (See Endnote I.)


With these ideas in mind, I turn to a review and discussion of some pertinent historical experience. I confine myself to a group of sixteen presently advanced countries (3) whose productivity levels, that is, GDP per hour, and whose productivity growth rates can be traced back as far as 1870. The fifteen countries (other than the U.S.) are, indeed, those whose relative productivity advances provide the standard against which U.S. postwar growth is usually measured and pronounced unsatisfactory or worrisome. Together with the United States, they are also the countries whose more recent productivity slowdown is a cause of anxiety throughout the industrialized world.

We owe the data to Angus Maddison [1982; 1989] on whose work so much of present-day growth studies have come to depend. So far as concerns his group of countries, the long-term message of his data is clear. Within the Maddison sample, the cross-country variance of the productivity levels, which was over 50 percent of the mean level in 1870, fell to 15 percent in 1986. Moreover, the decline was steady in direction from decade to decade, with only a single reversal across World War II. The (inverse) rank order correlation between the initial productivity levels of countries and their subsequent growth rates also rises steadily as the period beginning in 1870 is extended until, in a measure across 109 years between 1870 and 1979, the association becomes well-nigh perfect. The coefficient is -0.97. Or--using Maddison's revised data--it is -0.96 from 1900 to 1986. In short, the less productive a country in 1870 (or 1900), the faster was its subsequent growth rate. In a regression of productivity growth rates, 1870-1979, on a single variable, the countries' relative productivity levels in 1870, William Baumol and his collaborators [1989] obtained an [R.sup.2] of -0.88. I conclude that, for the countries in the Maddison sample, there has, indeed, been the convergence that the catch-up hypothesis leads one to expect--given adequate social capability and technological congruence. Perhaps another way expressing it is to say that convergence has been a long-term tendency among countries that now make up the industrialized economy of the "West." (See Endnote 2.)

There is more to say, however, if we look at experience over shorter periods. The main thing to observe is that the speed of productivity convergence was much faster--twice as fast--and the strength of the (inverse) association between levels and growth rates of productivity was much stronger in the quarter century after the War, that is, during the great growth boom from about 1950 to 1973, than during the seven decades before the War. I propose four reasons for the difference.

First, in the late nineteenth and early twentieth centuries, some countries--Japan, Italy, Finland, Norway--were still newcomers to modern industralization. They lacked social capability in several dimensions.

Secondly, the economic characteristics of a larger number of the laggards were incongruent with the pat of technological advance. At the frontier, typified by the United States, advance was heavily capital intensive and scale-dependent. Our path of technological progress ran through the extension of centralized power, assembly-line methods, long production runs of standardized products and large markets for durable consumer goods and business equipment. Among the laggards, however, their underdeveloped financial markets and their still low levels of income restricted capital accumulation, while the combination of low incomes and small populations limited their domestic markets, and, therefore, also the base on which large-scale competitive exports could be built. Before World War I, therefore, the catch-up potential of the laggard countries was limited in some countries by deficient social capability and in still more by scarcity of capital and restricted scale.

Thirdly, however, matters were improving in these respects before 1913, and they improved still more between the wars. One might, therefore, have expected a speed-up in convergence after World War I. But this was frustrated in the interwar years. Territorial re-adjustment, political instability, financial and monetary disturbances, protectionism, and the Great Depression did not provide a happy background for technological diffusion, investment and the foundation of new lines of business. Conditions supporting the rate of realization of potential were unfavorable between the wars for many countries that were in the course of catching up.

Finally, there were the two wars themselves. They were economically stimulating in the United States but destructive elsewhere. The result was to widen the productivity gaps between this country and many of the laggards. The average gap was 48 percent of the U.S. level in 1913, but 57 percent as late as 1950, after recovery from World War II had been largely completed.

In all these respects, the prewar limitations on the strength and speed of the convergence process were reduced or removed after World War II. The initial productivity gaps were larger than in 1913 or in the '20s. Social capabilities were stronger. Education levels in Europe and Japan were higher and continued to grow. There was political consensus everywhere supporting growth. When incomes in Europe and Japan began to rise above prewar levels, patterns of consumption coverged towards those in the United States and so did production structures. The economies of laggard countries became technologically more congruent with the techniques and industrial organization of this country. All this strengthened the potential for catch-up. At the same time, the conditions for rapid realization of potential became more favorable. Markets were enlarged by liberalized trade. On that account and because of multi-national corporate operations, technological diffusion accelerated. Government policies were directed to the support of investment, and the combination of the Bretton Woods dollar-exchange standard and U.S. monetary policy sustained international monetary and price stability for many years. The postware international trade and monetary regimes, which, to a large extent, were reflections of U.S. policy, remind us how important U.S. leadership was in the background of the postwar growth boom. Our ability to lead was, of course, based on our postwar economic and political strength, and it is true that our policies were colored by geopolitical calculation. But such calculation was tempered by a generous and progressive impulse and by enlightened analysis that does credit to American economics and diplomacy and to the broader political currents of that heady time. It would be heartening if we could rise to those levels again.

With all this in mind, I would make two assertions. First, the logic of the matter, the broad pattern of cross-country growth rates and a series of growth-accounting and regression studies assure us that the superior growth rates of other industrialized countries after the War can be attributed in good part fo the release of an underlying catch-up process. (See Endnote 3.) And since the U.S. rate of productivity advance itself was at least as fast in the '50s and '60s as in prewar decades, the whole industrialized world enjoyed an extraordinary growth boom for a quarter-century.


Now we must face up to the consequences of that boom and the catch-up and convergence of productivity levels, on which it was based. I have time--barely--to talek about two sides of the subject.

The first is that the growth boom, in several of its most important aspects, was inherently transitory, dependent on a catch-up process that was in its nature self-limiting. Between the years of the growth boom, that is from 1950 to 1973, and the years since 1973, the average productivity growth of our group of sixteen countries fell by more than half. The decline, of 2.3 percentage points a year, falls into two almost exactly equal parts. Half is a reduction in the amount by which European and Japanese growth exceeded that in the U.S. The other half is measured by the slowdown that occurred in U.S. growth, which, of course, is shared by the other countries as well (see Footnote 1, above).

It is the first half, the decline in the superior rate of advance of the laggards, that is the part clearly and directly traceable to the catch-up that took place during the boom and later. In the course of the boom, the average productivity gap between the United States and the other countries was reduced by 49 percent according to Maddison's latest estimates. The gaps have continued to decline since 1973, although at a slower pace. In consonance with the reduction in the productivity gaps, the rate of convergence of labor productivity levels has slowed drastically, the inverse cross-country correlation between levels of productivity and rates of growth has become weaker, and, as said, the amount by which the growth rates of the laggards exceeds that of the United States has become much smaller (as shown in Tables I and II, above). I interpret these observations, together with the showing of growth-accounting work and a number of regression studies to mean that much of the reduction in the superior productivity advance of the laggards can be attributed to the effect of the reduced productivity gaps on their total factor productivity and rates of capital accumulation. (See Endnote 3.)

This, of course, is not the whole story. There is still the other half, the part ofthe slowdown shared by the United States and the rest of the industrialized "West." Here we are up against the limits of present knowledge. If one asks what caused this part of the slowdown, the short answer is that nobody really knows. As in a good detective story, there are plenty of clues and plenty of suspects but no case that can yet get a conviction. Some point to the oil-price shocks. Others, especially in this country, think that management failures are responsible. Still others were persuaded that old lines of technological progress have been exhausted while we are only slowly learning how to apply and exploit the new potentialities of computers and telecommunications. Perhaps the suspects are jointly culpable. I intent to leave all that aside and direct your attention to still another suspicious character because I believe he--or it--was an important accomplice, although not the uniquely guilty party, and because it has a connection with the postwar catch-up process, which is my central theme.

Recall the distinction I drew earlier between the potential for technological advance and productivity growth and the conditions that govern the realization of potential. The suspect to which I point is a set of disturbances and developments that have weakened the supports for realization and, therefore, slowed the rate at which countries can exploit the potential for advance. The weakening of the supports for realization were in many ways the result of exogenous developments. The most prominent were the OPEC oil shocks, the rise of energy prices and all the macroeconomic troubles to which they led. But the disturbances in capital markets, labor markets, product markets and international trade, which have surely hampered investment and innovation during the last two decades, also find their origins, at least in part, in the preceding growth boom and in the catch-up process on which the boom was largely based. It is this additional self-limiting aspect of the catch-up process that I now want to stress. It implies that the growth boom was inherently transitory for reasons that go beyond the fact that catching up reduces technological gaps and, therefore, the potential for further advance. It also undercuts the conditions that support the realization of potential--or, at least, that was the case in the postwar years.

It would take too long to develop this subject now. The best I can do is to mention four ways in which I believe the catch-up process caused the conditions supporting the realization of potential to deteriorate.

First, the rapid rate of capital accumulation during the growth boom, together with the gradual decline in technological advances embodied in new capital, reduced prospective rates of return. The resultant decrease of investment rates helped to produce a poorer macroeconomic climate for investment and innovation. (Shultze [1987, 508 et seq.)

Secondly, the superior productivity gains in the countries catching-up during the Bretton Woods regime tended to undermine the international reserve position of the United States. The United States suffered chronic international balance of payments deficits, losses of reserve assets and accumulation of liquid liabilities to foreigners. Faith in the dollar-exchange rate and convertability of the dollar declined, and this was one of the bases for the eventual overthrow of the dollar-exchange standard (Abramovitz; [1979, 28-30]). The new regie of floating rates may or may not be the best that can be devised to meet the new circumstances, but it has permitted far wider fluctuations in rates of price inflation, in interest rates and in exchange rates, with deleterious consequences for investment and innovation.

Thirdly, the convergent productivity levels produced by the catch-up process also meant radical changes in the competitive positions and markets of large industries in many countries. Where the impact has been negative, there has been a natural resistance to the competition of foreign goods. The postwar trend towards trade liberalization has stalled and has been replaced by a resurgence of protectionism.

Fourthly, labor markets became tighter during the boom as labor reserves in agriculture and petty trade were absorbed in higher paid employment. This had implications for the recovery of labor union strength and earnings aspirations which created inflationary pressures from the cost side and, when inflation was restrained, produced "wage gaps" and unemployment [Schultze, 1987, 230-95]. It is also supported enlarged provision for job security and unemployment compensation and the elaboration of other elements of the Welfare State, all of which are developments that may be desirable in themselves but that act to restrict labor mobility and to make production costlier and investment more risky (Assar Lindbeck [1981; 1982; 1983] also William Fellner [1979]).


The moral of my story, therefore, is that the catch-up process by its own workings undermined the bases of the growth boom to which it gave rise. It weakened the potential for growth among the laggards by reducing the gaps separating them from the leaders--gaps in technology, capital stock, efficiently allocated resources, and market scale. And by its effects on the macroeconomic environment, on the labor market, and on the international monetary and trade regimes, it has hindered, and is hindering, growth in leaders and laggards alike.

As one looks to the future, several things come to mind. Looking to the long run, they are good things. The potential for growth by catch-up is now, indeed, weaker, but only among countries already rich and industrially advanced. That potential now passes to the renew industrializing countries of Southern Europe, Southeast Asia, and Latin America.

Moreover, the loss of catch-up potential among the most highly developed countries carries an offset. The resources in those countries that until recently were devoted mainly to borrowing and adapting existing advanced technology are now being increasingly devoted to pushing out the technological frontier itself. For the moment, of course, that makes us in the United States worry about our competitive edge and about the injuries that our own industries can suffer from rivals who are becoming stronger and may, in some industries, surpass us. In the long run, however, advances in knowledge, from whatever source, benefit all countries. We stand to gain from the technological progress of Europe and Japan and from trade with more productive partners, just as they have gained from us-- provided, of course, that we keep open the channels for trade and for the diffusion of knowledge. And that is why I began by saying that the process of catch-up and convergence is, in a large sense, desirable not only for the countries catching up, but also for an old leader--that is, for us.

If it is true--and it is--that the beneficial effects of convergence and catch-up are not yet apparent, one of the reasons is that we are still living in the trough behind the great postwar catch-up wave. We still have to absorb the backwash of imbalances in trade and finance that the wave itself has left behind. The corrections of those imbalances, if we can make them, will not only create more favorable conditions for realizing our own growth potential and that of the other rich countries, it will give the poorer countries of the South and East a renewed chance to develop.


That is the more or less serious side of my message. I have pointed to the important benefits and to the positive externalities of the catch-up process. We in the United States can trade with more productive and efficient partners, and we can now learn more from them s become more productive ourselves. Now let me close on what, I suppose, is a somewhat lighter note. There are some disadvantages for us when other countries become more efficient. I am thinking now of long-term disadvantages, looking beyond the fact that comparative costs necessarily change, that we are now over-matched in some directions and that our industrial structures of output and employment must adapt to a new distribution of international comparative advantage. I am thinking rather of the simple fact that when foreigners catch up in productivity, they also catch up in relative wages. So we have a repetition on the international scene of a process with which we are all-too-familiar on the domestic scene. When average wages go up with average productivity, the relative price of services, where productivity lags, rises steeply. Translated to the international arena, that means that we can no longer travel abroad as cheaply as we used to do; in all probability, we shall never again be able to do so. No longer can we enjoy French cooking and great clarets and burgundies for small sums. Moreover, those nouveaux riches foreigners can now also afford to travel; so now we have to make hotel reservations months in advance. When we visit Florence or Venice in the summer, we have to buck those crowds of Germans and Swedes in the streets and museums. The queue to enter the Louvre is now two hours long, as I know from a recent visit. And who is in the line? Frenchmen. When we walk in the Alps, even the Swiss are there to clutter up the slopes and to crowd the hutten. How easy and pleasant it was to travel in Europe when European wages were low, when Europeans were poor, and it was an American's birthright to move about like a Victorian English milord. All now gone. It's scandalous! Some of you perhaps are not old enough to share these nostalgic memories. To you I would say: "It's great to be young, but it carries a heavy cost. You were born too late."

So now we have a pretty paradox--which is my closing thought. The rise in foreigners' efficiency gives us the benefits of cheap goods produced in those sectors where their productivity has advanced rapidly. The accompanying rise in wages, however, snatches from us those wonderful cheap services produced in the sectors where European and Japanese productivity is advancing slowly or not at all. But do we--I mean our politicians and the public press--do we bemoan the fact that foreign wages have risen? No! Just the opposite. We complain that foreign wages are too low, that they have not risen enough, that we have to compete against those unfair low-wage foreigners who insist on selling us their goods--their embodied labor--too cheaply.

O, my fellow economist, all of us descendants of Adam Smith, is this what we have to show for two centuries of public education in economics?


1. Potential and Realization Further Considered

The determinants of potential and realization are also interdependent and may, in some conditions and in some stages of the catch-up process, operate to reinforce one another.

As catch-up proceeds, businessmen become more familiar with the organizational forms and practices of large-scale manufacturing or distribution. In response to the opportunities and requirements of large-scale business, facilities for financial intermediation are created, and individuals become accustomed to holding and trading in the securities through which the finance for large undertakings is mobilized. As levels of income converge, so do the patterns of consumption. The technologies developed by richer countries become more congruent with the needs of countries in course of catching up. In particular, capital gods become cheaper and markets for goods subject to economies of scale expand, so it becomes easier to adopt the capita-intensive and scale-dependent methods of the more advanced countries (Denison [1967, 239-45]. As incomes rise, educational levels are lifted. And as industrialization proceeds and firms become larger, corporate and university facilities for conducting research are established.

These feedbacks from catching up work to strengthen the potential for catch-up and to speed the pace of realizing potential. There are others, however, that may work against potential and its realization. The political consensus, which at low levels of income is concentrated on growth may turn in larger measure towards other goals, towards job security, income equality, environmental protection, safety in work and consumption. Public policy may then work in ways that impede productivity advance or, at least, the advance that our measures record.

Our knowledge of the net balance of these concomitants of the growth process is incomplete, to say the least, and our knowledge of how they work at various stages of catch-up and growth and in different social conditions is almost non-existent. It is clear enough, however, that they have a bearing on one element of the simple theory of convergence. They serve to qualify the inference previously drawn that the relative pace of advance of countries catching up will necessarily fall back steadily as productivity levels converge. That inference is drawn from the putative potential afforded by technological obsolescence, low capital intensity and overallocation of resources to agriculture and petty trade. It may be overborne, however, at some times and for particular countries, by feedbacks from higher levels of income and industrialization that serve to strengthen the elements of social capability or the supports for the realization of potential. As we shall see, however, the various feedbacks did not prevent a systematic general reduction in rates of productivity advance relative to that in the United States following the postwar convergence of productivity levels. But the feedbacks may well have been involved in the persistent strength of the catch-up process during a full quarter-century following World War II. And they may well have made a powerful contribution to the exceptional performance of Japan for an even longer period.

2. Limited Significance of the Maddison Sample

Notice that my statement in the text is carefully qualified. It does not assert that the evidence I cite is support for the hypothesis that there is a general tendency, applicable to all countries, for national productivity levels to converge. The countries in Maddison's group are not only few in number, they are specifically the advanced industrialized members of OECD. They are, therefore, by selection, a group of countries that have been relatively successful in adopting modern technologies. Presumably they had, or acquired, the characteristics needed for such success, that is, the ill-defined elements of social capability and the technological congruence that I view as required. In a sample chosen on the basis of some ex ante criteria, some members might have failed to progress, and such a sample would presumably display a lesser degree of convergence than does a group of ex post successes. I recognized the limitations of evidence based on Maddison's sixteen countries in my own first use of these data [1986, 394]. I wrote:

Mine is a biased sample in that its members consist of countries all of whom have successfully entered into the process of modern economic growth. This implies that they have acquired the educational and institutional characteristics needed to make use of modern technologies to some advanced degree. It is by no means assured--indeed, it is unlikely--that a more comprehensive sample of countries would show the same tendency for levels of productivity to even out over the same period of time.

Besides sample selection bias, I also drew attention to the possibilities of error bias because the initial-year levels of productivity are obtained, in effect, by backward extrapolation of late-year levels by national productivity growth rates. Errors in the estimated growth rates will, therefore, tend to produce an inverse correlation between initial levels and subsequent growth. The effect of such bias is presumably relatively serious in very long-term calculations that carry data back for a century. It would be less serious for thepostwar period itself for which estimates of productivity growth are more reliable.

Since my 1986 paper, and since the publication of a parallel paper by William Baumol [1986], J. Bradford De Long [1988], in a comment on Baumol, has shown now important these biases may be. De Long inspects the experience of a somewhat altered and enlarged sample of countries that he believes satisfy ex ante criteria of selection. His goal was to include all nations that "had high potential for economic growth as of 1870, in which modern economic growth had begun to take hold by the middle of the nineteenth century" [1988, 1140]. Construction of the sample admittedly required judgment. In the event, De Long dropped Japan, by far the poorest of the Maddison sample, and added six countries whose 1870 incomes per capita were as high as that of Finland (the poorest of the remaining original fifteen) and which in other wasy displayed promise of development. The countries added were Spain, Portugal, Argentina, Chile, Ireland and New Zealand. De Long then shows that the expanded sample of twenty-two, after allowance for the effect of measurement bias, no longer displays a significant tendency for productivity levels to converge.

Baumol and Edward Wolff [1988] and Baumol, Sue Blackman and Wolff [1989] have now carried the matter further. Selecting their samples ex ante (by initial-year per capita income level), they show that one can find a small number of countries (eight, possibly ten) that do exhibit steady convergence since the last quarter of the nineteenth century. More important, they show that in the period 1950-1980, a variety of samples related in different ways to incomes per capita at the initial date, 1950, do display convergent tendencies. The strongest evidence of convergence, betokened by a decline in the coefficient of variation, emerges for a sample confined to the sixteen countries (out of seveny-two) with the highest per capita incomes. But as the number of countries is extended beyond the richest sixteen, the tendency grows weaker. In samples confined to the poorest countries, there is even evidence of divergence.

As might be suspected, Baumol's sample of the highest income countries in 1950 overlaps strongly with the Maddison sample, Japan and Austria, which are members of the Maddison 16, do not make the Baumol 16. They are replaced by Luxembourg and Iceland.

With all this in mind, I think we may be safe in summarizing as follows. A strong catch-up tendency has characterized the productivity growth of the rich, industrialized "West" in the postwar period. This tendency has a longer history. The same group of countries was tending to converge in a process that can be traced back as far as 1870.

3. Evidence of the Catch-Up Process from Growth Accounts and

Cross-Country Regressions

This note refers to the bearing of growth accounts and regression studies on the assertions that the superior growth rates in Europe and Japan in the earlier postwar decades and the subsequent slowdown can be attributed in good part to an underlying catch-up process. It supplements the discussion of section III of the text and also that of section IV. Both sources of evidence are complicated by the fact that backwardness is a potentially for rapid technological progress, but it also works through several other channels, the inducement to invest, the improved allocation of resources, and the economies of scale opened up by advance along all these paths. The interaction of these several processes is a further complication.

In an earlier paper [1979], I made use of Denison's accounts for France, Germany, Italy and Japan, drawn from his publications of 1967 and 1976, the latter with William K. Chung. I treated three of Denison's growth sources as avenues of catch-up. These were his "changes in the lag in the application of knowledge" (which was his final residual for countries other than the United States), his "improved allocation" and "economies of scale associated with income elasticities." These accounted for 83 percent of the superior productivity performance in three European countries compared with the United States from 1950 to 1962. The same figure for Japan, 1953-1971, was 56 percent. If one adds one-half the difference between the contributions of capital accumulation in these countries and that in the United States, the percentages rise to 95 and 71.

These high figures presumably reflect to some degree the exceptional catch-up connected with recovery from wartime destruction and disruption. Figures adapted from John Kendrick's growth accounts [1981, Table 7] are presumably free of this influence. Kendrick's estimates refer to 1960-73, which covers the culmination of the postwar growth boom, and 1973-79, the initial stage of the subsequent retardation. In Table III, I use Kendrick's estimates to derive estimates of the contribution of various growth sources to the excess of Japanese and Western European productivity growth over that in the United States Again, I count four items as sources of the superior performance in Japan and Western Europe: (1) advances in knowledge (which is Kendrick's final residual); (2) reallocation of labor; (3) capital-labor substitution; and (4) economics of scale. As sources of catch-up, I weight the latter two items each one-half on the ground that non-catch-up factors may have contributed to the differences between the United States and the other countries. The effect of the catch-up sources appears in lines 8-11 of the table (in percentage points) and in lines 12-15 (as shares of the aggregate productivity growth-rate differences).

During the boom period, the account suggest that in Japan, the catch-up factor accounted for about 40 percent of the differences between Japanese and U.S. productivity growth if we associate the catch-up process only with the advance in knowledge incorporated into production. But if we add the other channels of catch-up, the share rises. It becomes about 70 percent when all four channels are recognized. In Western Europe, the comparable numbers are similar; they run from just under 50 percent to about 70 percent. In percentage points, the superiority of the Japanese rate of advance exceeded the European, and so did its advantage in each of the channels of catch-up--all of which is consonant with Japan's greater degree of initial backwardness.

Next, since the pace of Japanese growth during the boom was faster than the European, the gap between the Japanese and U.S. productivity levels was reduced more than the European. It is consistent with this fact that the Japanese productivity slowdown between 1960-73 and 1973-79 exceeded the European. In percentage points also, the Japanese slowdown in each of the variant catch-up measures exceeded the European. But that was not consistently true when the catch-up effects are measured as shares of the aggregate slowdown. In any event, in this growth-accounting approach, substantial portions of the reductions in the superior performances of both Europe and Japan are associated with the measures that we take to be connected with the preceding catch-up. In Europe, these mesures run from 66 to over 100 percent; in Japan from about 25 to 60 percent.

Numerous attempts have alos been made to estimate the effects of the catch-up process among the industrialized countries by cross-country regression. In these experiments, rates of growth of productivity (or per capita output) are regressed on relative levels of the same variables or on the relative gaps between U.S. and national levels. Rates of capital accumulation often enter as a second independent variable. In some experiments, other factors are introduced. The effects of changes over time in relative productivity gaps on the pace of growth is then estimated on the assumption that the influence of cross-country level differences on growth can be used to represent the influence of change in the relative levels of individual countries over time.

All the studies of which I am aware agree that growth rates significantly associated--inversely--with relative productivity (or per capita output) gaps. Among the industrialized countries, therefore, there is consistent evidence of a catch-up effect. The studies differ, however, in the importance they assign to this effect on growth rates within periods and on the slowdown of the '70s and '80s compared with the '50s and '60s. I refer here to three studies to suggest the range of results regarding the slowdown.

Robin Marris [1982] finds that, in a group of nineteen industrialized countries, the increase in per capita output levels from 1965 to 1973 would have caused an average slowdown of per capita growth of 1.6 percentage points between 1965-73 and 1973-79. This was 70 percent of the average actual slowdown, and it exceeds the reduction in the amount by which growth in the other countries surpassed that in the United States. Marris's estimates take account not only of a direct relation between per capita output levels and subsequent growth but also of an indirect effect that operates through the relation between output levels and capital accumulation.

Assar Lindbeck [1983] worked with labor productivity data. He estimates only the direct relations between productivity levels and growth rates in an equation in which investment enters only as a second independent variable. He attributes perhaps one-half a percentage point of the slowdown in Western Europe between 1960-73 and 1973-79 to technological catch-up during the earlier period. This is about 60 percent of the reduction between the two periods in the excess of the European over the U.S. growth rate. For Japan, his figure is about one percentage point which is about 25 percent of the reduction in Japan's superior performance.

A still later study by S. Dowrick and D-T Nyugen [forthcoming] is based on per capita output data for all twenty-four OECD countries including the less-industrialized countries of that organization. It compares 1950-73 with 1973-85. Growth rates are regressed against per capita output levels, employment-deepening (i.e., change in employment-population ratios) and capital-deepening (proxied by ratios of investment to output). Their results attribute about 45 percent of the reduction in the superior performance of Japan over the United States between the two periods to the catch-up factor, that is, the reduction of the per capita output gap. The same figures for the other OECD countries vary widely. An average for Germany, France and Italy is about 40 percent.

(1) The following estimates are frequently cited measures of the high rates of advance in the first quarter-century after World War II, the subsequent slowdown and the relatively slow pace of U.S. growth in both periods.

(2) Notions of catch-up and convergence as an element in the growth process are, of course, not new. Veblen [1915] is often cited as an early exponent. Jan Fagerberg [1987] provides a good brief set of references. I took up the theme in three earlier papers [1979; 1986; and 1989, Chapter 1).

(3) The countries are Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Italy, Japan, Netherlands, Norway, Sweden, Switzerland, United Kingdom and United States of America.


Abramovitz, M. "Rapid Growth Potential and Its Realization: The Experience of Capitalist Economies in the Postwar Period," in Economic Growth and Resources, Vol. I, edited by Edmond Malinvand. Proceedings of the Fifth World Congress of the International Economic Association held in Tokyo, Japan, 1977. London and Basingstoke: Macmillan Press, 1979, 1-30.

Abramovitz, M. "Catching up, Forging Ahead and Falling Behind." Journal of Economic History, June 1986, 385-406.

Abramovitz, M. Thinking About Growth. Cambridge and New York: Cambridge University Press, 1989.

Baumol, W. "Productivity Growth, Convergence, and Welfare: What the Long-Term Data Show." American Economic Review, December 1986, 1072-85.

Baumol, W., Sue Anne Batey Blackman and Edward N. Wolff. Productivity and American Leadership, The Long View. Cambridge, Mass.: M.I.T. Press, 1989.

Baumol, W. and Edward N. Wolff. "Productivity Growth, Convergence and Welfare: Reply." American Economic Review, September 1988, 1155 et seq.

De Long, J. B. "Productivity Growth and Welfare: Comment." American Economic Review, September 1988, 1138-54.

Denison, E. Why Growth Rates Differ, Washington, D.C.: The Brookings Institute, 1967.

Dowrick, S. and D-T Nyugen. "OECD Comparative Economic Growth, 1950-85: Catch-Up and Convergence." American Economic Review, forthcoming.

Fagerberg, J. "A Technology Gap Approach to Why Growth Rates Differ." Research Policy, 16, 1987, 87-99.

Fellner, W. "The Declining Growth of American Productivity," in Contemporary Economic Problems, edited by William Fellner, Washington, D.C.: American Enterprise Institute, 1979, 3-15.

Kendrick, J. W. "International Comparisons of Recent Productivity Trends" in Essays in Contemporary Economic Problems: Demand, Productivity, and Population, edited by W. Fellner. Washington: American Enterprise Institute, 1981, 125-70.

Lindbeck, A. "Disincentive Problems in Developed Countries." Institute for International Economic Studies Reprint No. 171, University of Stockholm 1981.

Lindbeck, A. "Emerging Arteriosclerosis of the Western Economies." Institute for International Economic Studies Reprint No. 191, University of Stockholm 1982.

Lindbeck, A. "The Recent Slowdown of Productivity Growth." Economic Journal, March 1983, 13-34.

Lipsey, R. and I. Kravis, Saving and Economic Growth: Is the United States Really Falling Behind. New York: The Conference Board Report No. 901, 1987.

Maddison, A. Phases of Capitalist Development. Oxford: Oxford University Press, 1982.

Maddison, A. The World Economy in the 20th Century. Paris: Organization for Economic Cooperation and Development, 1989.

Marris, R. "How Much of the Slowdown Was Catch-up?" in Slower Growth in the Western World, edited by R. C. O. Mathews. London: Heineman, 1982, 128-44.

Schultze, C. L., "REal Wages, Real Wage Aspirations, and Unemployment in Europe," pp. 230-91, "Saving, Investment and Profitability in Europe" pp. 508-39, in Barriers to European Growth: A Transatlantic View, edited by Robert Z. Lawrence and Charles L. Schultze. Washington D.C.: The Brookings Institution, Washington, D.C., 1987.

Veblen, T. Imperial Germany and the Industrial Revolution. New York: Kelley, 1915.

Wolff, E. "Capital Formation and Long-term Productivity Growth: A Comparison of Seven Countries." C. V. Starr Working Paper No. 87-37, September 1987.

Presidential address for the Western Economic Association, Lake Tahoe, California, June 21, 1989. The body of the paper is published here substantially as read. I have inserted some references to relevant evidence and literature. I have also added footnotes and a few longer Endnotes to supplement the argument in the text.
COPYRIGHT 1990 Western Economic Association International
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1990 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Abramovitz, Moses
Publication:Economic Inquiry
Date:Jan 1, 1990
Previous Article:Karl Brunner, 1916-1989.
Next Article:Does exchange rate appreciation 'deindustrialize' the open economy? A critique of U.S. evidence.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters