The Asian and Chinese economic growth models - implications of modern findings on economic growth

The issue of whether China's economic stimulus package, and the 'Asian growth model' in general, is correct, and therefore its success will continue, or whether it will fail, is evidently a question of great importance from the point of view of world economic development and the world economic situation.

China's economic stimulus package, in particular, has led to a major international debate. Taking non-Chinese writers, those holding that China's package is successful, naturally with differences on 'why' and on scale, include the author of this blog, Jim O'Neill, chief economist of Goldman Sachs, Professor Danny Quah of the London School of Economics, Mark Weisbrot and others,

On the other side are Martin Wolf of the Financial Times, Morgan Stanley's Stephen Roach, Michael Pettis of Peking University, and others who consider, again with significant differences on why and scale, that China's economic strategy is wrong, its stimulus package is misconceived, or both, and therefore it will end badly.

Because of the importance of the issue this discussion has involved not only immediate economic assessments but fundamental economic questions with general applicability outside Asia and China.

One of the most important of these is the fact that the rapidly growing Asian economies in general, and China in particular, base fast economic growth on very high levels of investment and an orientation to high levels of exports. Critics of China's economic stimulus package, and the Asian growth model in general, call for such policies to be abandoned, for investment to be cut back in favour of consumption, and for 'export led growth' to be abandoned.

This article therefore sets out why the Asian growth model is correct, consequently some of the key reasons why China's economic stimulus programme is successful, and why the proposals of those calling for investment to be scaled back in favour of consumption, and for a export led growth to be abandoned, are erroneous and would be damaging in slowing down Asian, and therefore world, economic growth.

The article below does this by noting that modern econometric research shows that the high investment/high export policies of the rapidly growing Asian economies are justified not only by evidently successful practical results but by economic theory. Given this combination the success of a number of Asian economies and China, now joined by India which has adopted elements of a similar orientation, will continue. Predictions of crisis in China, or of the ‘Asian model ‘ in general, are erroneous both practically and from the point of view of economic theory.

This article is more technical than those normally appearing on this blog. The reason for including it is because of the importance of the issues. However given this more technical character readers may prefer a summary of the conclusions and for proof they are referred to the article itself. The key points are:

1.Modern econometric research shows that, provided an overall framework of a high level of participation in the division of labour is maintained, which in a modern economy requires a high and expanding level of foreign trade, capital investment is the decisive factor in economic growth. The very high levels of investment in the rapidly growing Asian economies, and China in particular, are therefore correct and maintaining this high level is the key to continued rapid growth. The stress laid by Indian Prime Minister Manmohan Singh on the need for India to achieve very high levels of savings and investment is, for example, fully justified. Other Indian experts stressing the importance of high savings and high investment levels for the country's economy include, for example, Amir Ullah Khan. Proposals to lower the investment levels of China, India and the other Asian economies would, if implemented, have serious negative consequences in cutting their growth rates.

2. A high level of participation in division of labour in a globalised economy, which is a precondition for rapid growth, requires a high level of exports and imports and therefore a situation whereby strategically trade grows more rapidly than the domestic economy. Criticism of the Asian and Chinese economies for ‘export led growth’ is invalid as it confuses two different issues. The first is a high level of trade in GDP, i.e. both exports and imports, which is necessary and desirable, and the second is a large trade surplus – which is unnecessary and, in the case of China, has existed only for a short period during its economic reforms.[1]

The method adopted in this article is a review of the findings of modern economic research on the sources of economic growth and productivity.

*  *  *

Consideration of the findings of modern econometrics on economic growth, and its implications for China and Asia, must necessarily analyse the work of Dale Jorgenson, Professor at Harvard University, and former President of both the US Econometric Society and the American Economic Associations, as the findings of Jorgenson and his co-authors have now been officially incorporated in international standards for national account statistics and for growth accounting set by the US statistical authorities, by the OECD and by the National Accounts System of the United Nations.

At an earlier stage analyses by Angus Maddison played a crucial role in studies of sources of post-World War II economic growth, although in the more recent period Maddison has shifted his focus to studies of very long term international growth.

However, while Maddison's work has been absorbed by almost all writers on economic growth, and forms the standard point of reference in such studies, the work inaugurated by Jorgenson and his co-authors, while enjoying an elevated reputation in analyses of productivity has, for various reasons, not received equivalent adequate international attention in discussion of economic growth.[2] 

The totality of such research has however left no ambiguity as to the main sources of economic growth since World War II. It is that, within the framework of the expansion of the international division of labour, the accumulation of capital and labour, above all fixed investment, is the decisive source of economic growth.

As this finding is evidently of direct relevance for analysis of China and the 'Asian growth model' it is therefore somewhat surprising to find that a considerable amount of discussion of these issues proceeds without reference to modern econometric findings. A widely cited book such as Martin Wolf's Fixing Global Finance, which deals extensively with Asia and China, for example, fails to deal not only with such changes in international statistical methods but with interrelated studies of the causes of economic growth. A widely read and serious blog on China such as that of Michael Pettis similarly fails to adequately address such findings. An exception is the work of Danny Quah, which as already mentioned previously, deals extensively with this literature in regard to Asia - and updates and reviews some of the statistical work referred to below. Personal experience of research and university courses in China shows that insufficient attention is paid to such work there although, as will be seen, it directly pertains to China's economic strategy.

It is therefore worth briefly outlining here the relevant findings of this huge corpus of research. It should be made clear that while the focus is necessarily on Maddison, and then Jorgenson and his co-authors, as the leading figures in modern precise statistical research on post-World War II economic growth, the methodologies and approaches Jorgenson et al have outlined have been examined and received official endorsement from the OECD, and the National Accounts System of the United Nations. While there are, of course, different precise estimates of factors in growth, some of which are dealt with below, the statistical methods referred to have become the official international standard - what is involved is not the views of single individuals.

Particular analysis will be made of the way in which such work aids integration of analysis of the determinants of long term economic growth with practical government and company policy. While some questions involved may appear statistical and theoretical, in reality, as will be shown, they have decisive implications for practical economic policy and economic strategy. Naturally only a brief summary of the issues can be given, and the present article deals only with the overall framework of such work with particular emphasis on the implications for Asia, China and the US. For a more comprehensive account readers are referred to the work of the authors cited.

It should also be emphasised that this brief review is not aimed at those studying productivity or econometrics, who will already be familiar with the work, but is simply to highlight its importance for wider contemporary economic discussion regarding Asia and China. Given the remarks made above about insufficient international attention paid to such work, in particular in a number of countries for which it is most important, rather more than usual factual material is given regarding statistical conclusions. This has the effect of making the text rather dense but it is hoped thereby to convey the full importance of the findings and stimulate desire for more direct study.

The normal caveats in particular apply. The selection of topics for treatment below corresponds to specific issues affecting Asian and Chinese economic growth and far from reflects the full range of issues dealt with by the authors cited - to whom readers should turn for a comprehensive treatment.

As it may aid in understanding points outlined on this blog, the comparison of economic theory to certain key qualitative facts of economic development which it must explain are set out before dealing dealing with much more fine grained statistical research. The method adopted therefore is to proceed from the most general long term qualitative considerations to progressively shorter time frames. Readers who wish to proceed directly to the most recent detailed statistical research may skip over the earlier sections and move to the section on 'Input growth in advanced and developing economies' below

Determinants of long term economic growth versus the theories of Solow and Kuznets

At the beginning of the 1970s, a period when the current author entered economics, discussion of economic growth was dominated by the theory that expansion of inputs of capital and labour (factor accumulation) played a relatively minor role in economic growth and that the decisive determinant of the latter was (total factor) productivity – frequently asserted to be due to advances in technology. The key names associated with such arguments were Solow and Kuznets. These arguments were presented in an updated form, as noted previously, by Paul Krugman in his well known 1995 paper 'The Myth of Asia's Miracle' - which is still widely cited in discussion on China and Asia's growth despite the fact that, as will be seen, its conclusions have been vitiated by modern econometric research.

Solow's 1957 article 'Technical Change and the Aggregate Production Function' was crucial in setting the framework that the decisive factor in economic growth was not investment, or capital and labour inputs, but increases in total factor productivity. Solow stated regarding the US economy: 'over the forty year period (1909-49) output per man hour approximately doubled… about one-eighth of the total increase is traceable to increased capital per man hour, and the remaining seven-eighths to technical change.'

Kuznets similarly argued in his 1971 Economic Growth of Nations that: 'The high rate of growth in product per capita associated with modern economic growth can be credited to a large degree to growth of productivity, that is, of output per unit of input… the rise in productivity amounts to at least eight-tenths of the rise in per capita product in several countries.'[3]

While Solow/Kuznets introduced, and retains, a key role in the establishment of the framework of growth accounting it should be noted that the role of 'technology' was defined from the outset in a statistically highly unsatisfactory way. It was treated as a 'residual' - that is all growth that could not be definitively allocated  to another factor was assigned to the 'Solow residual' and treated as technology.  This  necessarily inflated the role assigned to 'technology/total factor productivity' and indeed had the perverse effect that the less accurate were the statistics, in the sense of the less their grip on the data, the higher became the role of 'technology'. This unsatisfactory state of affairs was famously characterised by Moses Abramovitz as being that what was actually being measured in the 'Solow residual' was 'ignorance'. It was entirely possible in principle that the 'high' role played by technology, as compared to capital and labour inputs, was due to the inadequate state of statistics in early periods of research of economic growth rather than any actual role of technology. To jump ahead, this precisely turned out to be the case.

The dominant theories prevailing at that time were thus accurately described by Dale Jorgenson: 'The early 1970s marked the emergence of a rare professional consensus on economic growth articulated in two… books. Kuznets… Economic Growth of Nations… [and] Solow's… Economic Growth … The resulting professional consensus, now obsolete, remained the guiding star for subsequent conceptual development and empirical observation for decades. .. Kuznets… [argued]… "... with one or two exceptions, the contribution of the factor inputs per capita was a minor fraction of the growth rate of per capita product." For the United States during the period 1929 to 1957, the growth rate of productivity or output per unit of input exceeded the growth rate of output per capita. According to Kuznets' estimates, the contribution of increases in capital input per capita over this extensive period was negative!...

'Kuznets' assessment of the significance of his empirical conclusions was unequivocal: "Given the assumptions of the accepted national economic accounting framework, and the basic demographic and institutional processes that control labour supply, capital accumulation, and initial capital-output ratios, this major conclusion - that the distinctive feature of modern economic growth, the high rate of growth of per capita product is for the most part attributable to a high rate of growth in productivity – is inevitable."'

I rejected such analysis at the beginning of the 1970s for a clear reason - it was not in line with the facts. Or as Jorgenson put it more elegantly: 'the consensus on economic growth reached during the 1970s has collapsed under the weight of a massive accumulation of new empirical evidence.'[4]

This conclusion, in my case, flowed from the study of very long term economic growth, the analysis of which had been greatly facilitated by a number of important statistical analyses that were produced commencing in the 1960s.[5] At that time Maddison had commenced the long series of studies which were to culminate four decades later in The World Economy and Contours of the World Economy 1-2030AD, while authors such as Cole, Deane, Feinstein, Matthews, Mitchell Odling-Smee and others were providing statistical data at a level that had not previously been available.

The advantage of studying such long term economic growth is the same as its disadvantage - the details cannot be seen and only the main trends stand out, thereby making it easier to assess these. It was evident from analysis of such long term economic statistics that the decisive relations were those regarding the division of labour, and that between investment and growth, not the factors identified by Kuznets and Solow.

Given science requires that where there is a difference between facts and theory it is the facts which prevail, therefore, despite the fact that Solow and Kuznets were the 'conventional wisdom' in academic economics, their conclusions were to be rejected as being inconsistent with the principal known economic data. Before proceeding to outline modern statistical conclusions the chief facts flowing from studies of long term economic growth, and some of the practical conclusions which follow from these will therefore be outlined. Placed in that context the significance of modern statistical work in integrating long term economic developments with practical economic strategy will become apparent.

The division of labour

The first crucial issue invalidating the Solow/Kuznets approach might initially appear pedantic but it has, as will be seen, decisive economic consequences - and, to jump ahead, was later vindicated by subsequent statistical work. This issue was that both factual evidence and economic theory are in accord that the economic division of labour is the most powerful instrument in raising output. To instead introduce 'technology' as the determining factor in growth, as Kuznets/Solow did, violated a fundamental principle of economic analysis since its classical foundation.

To take first theory, putting it in a polemical but hopefully clarificatory way, it may be recalled that the first sentence of the first chapter of the founding classic of modern economics, Adam Smith's The Wealth of Nation is: 'The greatest improvement in the productive powers of labour, and the greater part of the skill, dexterity and judgement with which it is anywhere direct, or applied, seem to have been the effects of the division of labour.'[6] According to the Kuznets/Solow hypothesis, however, Adam Smith was in error – he should instead have cited 'technology' as being responsible for the 'greatest improvement in the productive powers of labour'! Kuznets/Solow replaced the central socio-economic driving force identified by Smith, consolidated in subsequent economics, and today exemplified in the international division of labour by the process of globalisation, with the quite different driving force of 'technological change'.

While it may in principle have been possible to integrate 'technological change' within the division of labour, the Kuznets/Solow priority to technology as the decisive element turned reality on its head – it reversed the relation of what should have been determining element (division of labour) with what should have been subordinate (technology). Some of the economic implications of this are considered below.

Trade and division of labour

Factually this issue can be illustrated most clearly over the longest period by considering international trade – which is of course why Smith initially most clearly outlined the fundamental economic mechanisms in this area.[7]

At the beginning of the 1970s Maddison's magisterial quantitative analysis of 2000 years of world economic history, with calculations of GDP per capita for different areas of the international economy, was not available. Nevertheless the qualitative elements of world economic history, and therefore the key facts that had to be explained by any theory of growth and productivity, were entirely clear.

Preceding the rise of capitalism in Europe the most economically advanced, in terms of per capita GDP,part of the world economy was its pre-eminent trading part, i.e. the Arab/Iranian core of the Muslim world - not Europe, China or India. The economic success accompanying this great classical period of flowering of Islamic civilisation was evident. Later, within Europe, once the development of capitalist economy commenced, the succession of the economically most productive powers was clear – from initial leadership by Venice, then to that of the Netherlands, and finally to Britain. Each of these consolidated a trading empire larger than the one that preceded it.[8]

The development of the subsequent, and currently most highly productive economy, the US, does not violate this principle but illustrates it – clarifying that what is essential is the scope of the division of labour and not the specifically international character of trade in the sense that what is crucial is the crossing of national boundaries.

The US created the world's first integrated continental scale economy – China and India are becoming the second and third. The proportion of foreign trade in the US economy was therefore lower than in preceding dominant economies. But because the US economy was far larger than the previously leading economies of Britain, the Netherlands (or Venice) it could develop far greater division of labour, even on the basis of its internal market, than Britain could on the basis of international trade. By the mid-20^th century, however, even the scale of the US domestic market, and its internal division of labour, was inadequate and the US was itself forced down the road of globalisation in order to raise further its productivity.

This correlation of the main trends of world productivity with trade, with trade itself only constituting division of labour carried onto an international scale, precisely as formulated by Smith, is evident. The principles of Smith's analysis, which is naturally one of the many reasons it continues to be one of the greatest classics of economic literature, therefore provided a clear explanation of the main trends of world economic history, and why the highest productivities of labour were achieved at particular times in particular places. Kuznets/Solow framework, to fit the main facts of economic history, required an arbitrary, and essentially unexplained, jump in technology leadership from the Islamic world, to Venice, then the Netherlands, then Britain, and then the US. These conclusions drawn from the key qualitative facts of economic history were, as will be seen, confirmed when periods were studied in which quantitative as well as qualitative trends could be analysed.

Adam Smith, not Kuznets/Solow, in short laid the foundations for the correct analysis of the fundamental determinants of economic growth.

Inward and outward facing economic orientations

This apparently abstract economic issue had, and has, decisive contemporary consequences for economic strategy – particularly the choice between 'outwardly oriented' or 'import substitution' economic strategies. If 'technology', or to take the next issue considered below, the growth of fixed investment, is the most powerful determining element in economic growth, then an inwardly facing, nationally autonomous, policy seeking to maximise fixed investment or promote technology, may be a valid growth strategy. If, on the contrary, division of labour is the most powerful force for raising productivity and growth then 'inward facing' economic strategies cannot be successful – precisely because they cut the economy off from the international division of labour.

This was an extremely practical economic choice, not merely a theoretical one, which was put to the test, as subsequently extensively documented, from the 1970s onwards.[9] All economies with inwardly facing import substitution policies, whether using market economic mechanisms (Argentina), non-market mechanisms (the USSR), or apparently attempting to operate an eclectic combination of the two (India), suffered deep crisis. Similarly the 'opening' policy of China after 1978 was of the greatest economic interest because, in contrast to the inward facing import strategies, it should, if economic theory were correct, bring great economic success – as it did.

Adam Smith therefore achieved not merely theoretical but practical victory over not only 'import substitution' strategies but also over the focus on technology flowing from the analysis of Kuznets/Solow. The correctness of the high level of trade, that is export oriented, character of the Asian economic model is a direct consequences of this fundamental fact.  

Growth of factor inputs

If the first reason for rejecting Kuznets/Solow even in the 1970s might, in principle, have been dealt with by reformulating their theory, and placing technology in a wider framework, the second objection could not be solved by any reformulation because it involved a direct contradiction with facts.

It was entirely possible to calculate, particularly using the long term statistical data that had became far more readily available from the 1960s, that the proportion of the economy devoted to fixed investment had strongly risen historically and that there was a clear relation of this to more rapid economic growth. Studying such long term trends left no doubt that, after the division of labour, the decisive relation was between investment and growth.

The main historical features of these trends have been outlined several times in this blog and do not need to be repeated again here. The key graph summarising the historical increase in the rate of investment is reproduced as Figure 1 and for details of its periodised relation to increasing rates of growth readers are referred to other articles.

Figure 1

Such a historically strongly rising trend of the proportion of fixed investment in the economy meant that investment was historically growing more rapidly than GDP and this was correlated with more rapidly rising rates of economic growth.[10] Such a finding, which was clear analysing long time periods, was evidently in contradiction with the thesis of Kuznets/Solow that capital accumulation played little role in economic growth.

Indeed Kuznets had claimed: 'special factors limit the level and upward trend in the savings and capital formation proportions in total product as the latter grows over time.'[11] This was clearly factually false – on the contrary one of the most striking historical trends was precisely the upward trend in savings and capital formation as a proportion of total product.

It is also evident from this data that there is nothing inexplicable in the very high level of investment in China. It is merely the latest stage of the historical tendency for the level of fixed investment to rise and for this to be associated with faster and faster rates of economic growth.

As Kuznets/Solow was evidently in contradiction with the facts evident from study of long periods of economic development their theory was to be rejected and the decisive relation between investment and growth was evident.

Practical versus theoretical concerns in economics

While the analysis and conclusions outlined above involved considerable scrutiny of statistical material my reasons for undertaking it were, however, practical and not academic – advising companies and attempting to influence government, or potential government, policy. Such advice was necessarily based on facts regarding economic development and not on academic orthodoxy and therefore also based on the perspective that the decisive role in economic development was played by increasing division of labour and inputs of capital and labour - above all by investment.

These differences with Solow/Kuznets necessarily had crucial practical conclusions. To take a major international issue of economic strategy, for example, it informed the assessment that the economic reform policies being pursued in China after 1978 would be highly successful whereas the 'shock therapy' urged by many academic economists, and media commentators, on Russia after 1991 would be an economic disaster (See for example the 1992 article 'Why the Economic Reform Succeeded in China and Will Fail in Russia and Eastern Europe'). Underlying more technical discussion about the structure of competitive and non-competitive markets in Russia and China was an imperative that in the 'reform' period Russia must keep up inputs of capital and labour via methods that had been highly successful in China. The alternative approach, based on academic economics prevailing at that time in the US and Europe, and the writings of Kornai in Eastern Europe, stressed the decisive aim in Russia should be not be to maintain factor inputs but to increase factor productivity – a perspective evidently in line with Kuznets/Solow.

The factual record is that the policies pursued in China led to the most rapid prolonged economic growth in human history whereas 'shock therapy' policies pursued in Russia led both to the largest declines in output in any country in peacetime in history and to an actual fall in productivity. In short the theoretical issues had deeply practical implications.

Asian growth

Such differences also led to directly divergent judgements regarding the so called 'Asian growth model' – i.e the paradigm dominated by high levels of investment, and high levels of trade, pursued by South Korea, Singapore and the other South East Asian Tiger economies, and today most comprehensively followed by China, Such a model is based on massive mobilisation of inputs of capital and labour. If mobilisation of capital and labour inputs was the chief factor in economic growth - provided that an external facing economic orientation underpinning at least average increases in total factor productivity was maintained - then the Asian growth model was correct. The key practical policies which flowed from such a policy were therefore those which allowed such mobilisation of factor inputs – creating high levels of savings to finance investment, raising the rate of participation in the workforce etc.

If, however, increases in total factor productivity were the key to economic growth, as Solow/Kuznets argued, then the Asian model of development was wrong – as Krugman and other authors claimed. In that latter perspective measures to mobilise factor inputs were not crucial, and the key policies are those aimed at increasing total factor productivity – for example incremental improvements in markets to allocate capital and labour most efficiently.

Different positions of matters of economy theory, while apparently dealing with abstract analytical issues, therefore again had decisive practical implications for economic policy. The study of long term economic growth, and the decisive role played by division of labour and investment, led clearly to the conclusion that the 'Asian' model or Chinese model was correct and would be successful – as indeed it has been. And that critics of the Asian and Chinese models, predicting decline and inability to maintain economic growth well above US and European rates, would be invalidated by events – as indeed they have been, and as can be verified to the present period by reading the article by Professor Quah already cited.

While such work carried out in the fields of economic policy and company strategy involved intensive and persistent study of statistical data, regrettably time constraints prevented following academic discussion as closely as would have been desirable. As it was quite clear from earlier study that Solow's and Kuznets argument were not in accord with facts regarding long term economic growth I did not pursue the academic discussion regarding their work at that time.

Progress in the study of economic growth

As Dale Jorgenson has emphasised, the work that transformed debate on research into sources of international economic growth was the publication, in the early 1980s, of Maddison's Phases of Capitalist Development followed by his Dynamic Forces of Capitalist Development.

These works confronted the fundamental statistical problem that in the Kuznets/Solow approach 'technology' was defined as a residual. Maddison synthesised the work of himself and others on economic growth, using the the much more advanced econometric tools that were then available, and showed that most of the residual had indeed been 'a measure of our ignorance' and not technology. Maddison's key findings for the post-World War II period are set out in Table 1 and 2.

As may be seen Maddison showed that the largest role in economic growth was played by increase in factor inputs - i.e. increases of labour supply and capital investment. In addition once other the impact of other identifiable factors were measured - economies of scale, foreign trade, structural changes such as the decline of agricultural employment, and the impact of energy and natural resources endowment changes - were taken into account then the great majority of growth could be assigned to explainable sources other than technology even if the extremely extremely biased assumption was made that all unexplained growth was assigned to technology. In the case of the most advanced economy, the US, 78-96% of growth was due to such quantifiable factors.

In short the 'Solow residual' had indeed been measuring 'ignorance' rather than technology.

Table 1

Table 2

Maddison himself centred his subsequent research on other issues, in particular very long term economic growth, but Jorgenson, who had been developing increasingly sophisticated econometric tools since the 1960s, was able to apply these to even shorter periods of time than those analysed by Maddison. Jorgenson had earlier produced even more detailed results, centred particularly on the US economy, that produced the same conclusions regarding the sources of economic growth in the post-war period as Maddison's studies.

Integration of studies of very long term economic growth and shorter term

A consequence of the work of Maddison, Jorgenson and others is that it is now possible to achieve an integration of contemporary studies of growth with long term historical data - overcoming the split between fact and theory which had led to the reasons for rejecting Kuznets/Solow in the 1970s.

In the 1970s a radical disjunction had existed between on the one hand the large body of statistical research being accumulated, which confirmed the dominance of factor inputs in economic growth, and academic theory as it was being taught in economics departments – which asserted the Kuznets/Solow thesis that such growth of inputs was a minor factor in economic growth. Maddison, Jorgenson and others work overcame this disjunction between fact and theory through demonstrating the dominance of division of labour and factor accumulation in economic growth.

Increasingly sophisticated econometric techniques were deployed to deal with much shorter statistical time periods than those which had originally led to rejection of Solow and Kuznets conclusions – 'short term' in this context, of course, being a relative term as Maddison's studies in the 1980s, and Jorgenson's work, primarily considered the post-war history of the post-war economy whereas the statistical material produced by Feinstein, Cole, Deane, Mitchell, Maddison's own earlier work, and others had dealt with very much longer time frames.

Trends which could be seen immediately by considering very long time periods required increasingly sophisticated econometric methods to reveal over shorter time frames. Econometric microscopes revealed in detail what was clear to the naked eye when very long time frames were considered. Or, to put it another way, in considering the very long term it was easy to see the signal amid the noise, whereas in analysing shorter time frames advances in econometric techniques were required to remove the noise so the signal could be seen clearly.

This closing of the gap between historical studies and economic theory is not only of decisive theoretical and practical significance but renders superfluous books on economic growth which fail to start from the key facts of economic development. 'Pre-Copernican/Ptolemaic' economic analysis, consisting of building mathematical models which bear no relations to the real facts of economic development, is not merely of no use from a practical point of view but is invalid from the point of view of economic theory. Analysis of actual facts of economic development confirms the correctness of the export oriented and factor accumulation, above all investment oriented, economic model of Asia and China.

Conclusion of Jorgenson's studies

Turning now to Jorgenson and his co-authors, and bringing this work up to date to 2009, the fundamental factual conclusion of the work Jorgenson initiated more than four decades ago is clear, unequivocal, decisive and parallels the conclusions arrived at by Maddison. It is growth in division of labour and inputs of capital and labour, above all investment, and not total factor productivity which is determinant for economic growth. Using much shorter time frames, what is clear immediately from much longer term studies is confirmed.

Establishing continuity in the study of economic growth

Jorgenson himself paid generous tribute to Jan Tinbergen as being the first to establish the factual situation regarding the decisive factors in specifically US economic growth: 'The starting point for our discussion… is a notable but neglected article by the great Dutch economist Jan Tinbergen (1942), published in German during World War II.Tinbergen analyzed the sources of U.S. economic growth over the period 1870-1914. He found that efficiency [Tinbergen's term for productivity] accounted for only a little more than a quarter of growth in output, while growth in capital and labour inputs accounted for the remainder.'

Jorgenson, indeed, ironically notes that the result of his more than four decades of economic research has been to come up with essentially the same answer Tinbergen had found more than sixty years previously! In such a perspective, of course, the theories of Kuznets/Solow were a short term interlude in the main course of economic research.

As Jorgenson stated: 'Among the many remarkable features of Tinbergen's study was an international comparison of growth of output, primary factor input, and total factor productivity for France, Germany, the United Kingdom, and the United States for the period 1870-1914. For the United States, Tinbergen found that the growth of output averaged 4.1 percent per year, the growth of input was 3.0 percent, while productivity growth averaged 1.1 percent. Productivity accounted for only 27 percent of US economic growth during the period 1870-1914. The findings here allocate more than three-fourths of US economic growth during the period 1948-1979 to growth of capital and labour inputs and less than one-fourth to productivity growth.' [13]

The factual foundation laid by Jorgenson and his co-authors was, of course, much more detailed and firmer than that available to Tinbergen's inspired initial analysis. In his 2005 paper 'Accounting for Growth in the Information Age' Jorgenson was able to conclude: 'Input growth is the source of nearly 80.6% of US growth over the past half century, while productivity has accounted for only 19.4%.'

Methodology

The specific econometric methods utilised by Jorgenson have been thoroughly vindicated by subsequent statistical examination by a wide range of international bodies. It would take too much space, and it is unnecessary, to recap here the increasingly sophisticated econometric methods Jorgenson, with co-workers, used to dissect economic growth and productivity. That is in any case best followed by reading the various authors themselves.

It is sufficient to note here the overwhelming degree to which this statistical methodology has been vindicated by subsequent research and international statistical methodology brought in line with its conclusions. Few statistical method have been examined in such detail, rightly in the light of their major implications, and eventually received such sanction.

As Jorgenson notes regarding the final outcome: 'The traditional approach of Kuznets (1971) and Solow (1970)… has been replaced by the new framework presented in the OECD (2001) manual, Measuring Productivity… The OECD productivity manual has established international standards followed by Jorgenson, Ho and Stiroh… and the EU (European Union) KLEMS (capital, labor, energy, materials and services) study.'

A comprehensive account of the various statistical developments that led to this detailed overturning of Kuznets and Solow's conclusions may be found in Jorgenson's 2009 introduction to The Economics of Productivity - to which readers are referred to for a full analysis. The fundamental conclusions may, however, be briefly noted: 'the BLS [US Bureau of Labor Statistics] Office of Productivity and Technology undertook the construction of a production account for the US economy with measures of capital and labour inputs and total factor productivity, renamed multifactor productivity…. The official BLS (1994) estimates of multifactor productivity have overturned the findings of Abramovitz (1956) and Kendrick (1956), as well as those of Kuznets (1971) and Solow (1970). The official statistics have corroborated the findings summarized in my 1990 survey paper, 'Productivity and Economic Growth'.… The approach to growth accounting in my 1987 book with Gollop and Fraumeni and the official statistics on multifactor productivity published by the BLS in 1994 has now been recognized as the international standard. The new framework for productivity measurement is outlined in Measuring Productivity, a manual published by the Organisation for Economic Co-Operation and Development (OECD) and written by Paul Schreyer….

'The transition to the new framework for productivity measurement, represented by Jorgenson, Ho and Stiroh (2005)… has precipitated the sudden obsolescence of earlier productivity research employing the conventions of Kuznets and Solow... 

'Jorgenson and Steven Landefeld have developed a new architecture for the US national accounts that includes prices and quantities of capital services for all productive assets in the US economy. The incorporation of the price and quantity of capital services into the revision of the 1993 System of National Accounts (SNA) was approved by the United Nations Statistical Commission at its February–March 2007 meeting. A draft of Chapter 20 of the revised SNA, "Capital Services and the National Accounts", is undergoing final revisions and will be published in 2009. Schreyer, now head of national accounts at the OECD, has prepared an OECD manual, Measuring Capital, published in 2009. This provides detailed recommendations on methods for the construction of prices and quantities of capital services.'

In short the statistical methodology employed has been vindicated in the most thorough fashion. Those who wish to attempt to maintain the approach of Kuznets and Solow, and their conclusions regarding growth, have therefore to overturn what is now an enormous corpus of international statistical work.

Having outlined the most central conclusions, and methodological outcomes, of this work some of its other results and implications will be briefly considered, particularly as they affect Asia and China.

Input growth in advanced and developing economies

Considering these fundamental statistical findings in more detail, and by period, in Productivity and US Economic Growth, Jorgenson, Gollop and Fraumeni noted: 'To analyse the sources of US economic growth for the period 1948-79, we… considered the contributions of capital and labour inputs, and the rate of growth as sources of growth in value added. For the period as a whole the contribution of capital input averaged 1.56 percent per year, the contribution of labour input averaged 1.05 percent per year, and the rate of productivity growth averaged 0.81 percent per year… capital input is the most important source of growth in value added, labour input is the next most important, and productivity growth is the least important.' [12]

Such findings meant that increase of factor inputs accounted for 76.3% of US economic expansion in the period considered (capital 45.6% and labour 30.7% labour), and productivity for only 23.7% of US economic growth in this period.

It should be noted, for discussion of Asian and Chinese economic growth, that this dominance of inputs of capital and labour over factor productivity in economic growth applied not only to developing but to developed economies.

As Jorgenson noted regarding the growth of the world's most advanced economies, i.e the G7: "investment in tangible assets is the most important source of economic growth in the G7 nations. The contribution of capital inputs exceeds that of total factor productivity for all countries for all periods."

More precisely, Jorgenson and Vu found, analysing 'the contribution of capital input to economic growth for the G7 economies,' that: 'Capital input was the most important source of growth [for the G7] before and after 1995. The contribution of capital input before 1995 was 1.28 or almost three-fifths of the G7 growth rate of 2.18 percent, while the contribution of 1.43 percent after 1995 was 55 percent of the higher growth rate of 2.56 percent. Labour input growth contributed 0.49 percent before 1995 and 0.46 percent afterwards, about 22 percent and 18 percent of growth, respectively. Productivity accounted for 0.42 percent before 1995 and 0.67 percent after 1995 or less than a fifth and slightly more than a quarter of G7 growth, respectively.'

Regarding the US economy, Jorgenson summarised the situation regarding the sources of GDP growth for the entire post-World War II period 1948-2002, that is extending the analysis of the immediate post-war period noted above, as follows: 'Output grew 3.46 percent per year, capital services contributed 1.75 percentage points, labour services 1.05 percentage points, and total factor productivity growth only 0.67 percentage points. Input growth is the source of nearly 80.6 percent of U.S. growth over the past half century, while productivity has accounted for 19.4 percent.''

Dominance of capital inputs during the technology boom

This dominance of capital inputs over technology/total factor productivity in growth was, furthermore, not negated during the US 'technology boom' at the end of the 1990s or in the most recent economic period.

In a 2007 analysis Jorgenson, Ho and Stiroh, analysing the peak of the US IT boom, found: 'The growth rate [of the US economy] during the 1995-2000 boom was a remarkable 1.85 percentage points higher than during 1990-95. Capital input contributed 1.02 percentage points of this 1.85 [58.4%]. Labour input contributed 0.44 percentage points [23.8%]… Faster growth in total factor productivity contributes the remaining 0.40 percentage points [21.6%]. In other words, capital continued to be the most important source of US economic growth, as in the earlier decades.'

Jorgenson noted that what applied at the level of the overall US economy also applies at the level of individual industries: 'The perspective on US economic growth suggested by the results… emphasises the contribution of mobilisation of resources within individual industries rather than productivity growth. The explanatory power of this perspective is overwhelming at the sectoral level. For 46 of the 51 industrial sectors… the contribution of intermediate, capital and labour inputs is the predominant source of output growth.'[14]

As Jorgenson concluded in his comprehensive 2009 survey of The Economics of Productivity: 'Turning to the sources of the US growth acceleration after 1995, Jorgenson, Ho, Samuels and Stiroh… find that the contribution of capital input was by far the most important.' Considering post-World Ear II economic development as a whole Jorgenson similarly concluded in his 2009 survey: 'Although the role of innovation is often described as the predominant source of economic growth, the growth of productivity was far less important than the contributions of capital and labour inputs to US economic growth.'

Intermediate inputs and the division of labour

One of the advances in modern econometric techniques that has been introduced is the ability to attribute US economic growth to individual industries. Such analysis necessarily involves analysing intermediate inputs into individual sectors as well as the overall contribution of capital and labour inputs to economic growth.

In this regard one of the most important of findings, from the point of view of considerations analysed above, was that regarding intermediate inputs – i.e. inputs into one industry from another.

As Jorgenson noted: 'For each sector, intermediate input is represented as a function of deliveries from all other sectors.' The conclusion of such analysis is that: 'The sum of contributions of intermediate, capital, and labour inputs is the predominant source of growth of output… Comparing the contribution of intermediate input with other sources of output growth demonstrates that this input is by far the most significant source of growth. The contribution of intermediate input exceeds productivity growth and the contributions of capital and labour inputs. If we focus attention on the contributions of capital and labour inputs alone, excluding intermediate input from consideration, these two inputs are a more important source of growth than changes in productivity.' [emphasis added] Analysing 51 US industrial sectors Jorgenson, working with Gollop and Fraumeni, found that intermediate inputs were the largest source of growth in 37 of these.[15]

The economic significance of this finding is evident and relates directly to issues discussed earlier.The most important single source of economic growth is the increase in deliveries from other sectors – an economic consequence of growth of division of labour. That is, Jorgenson quantitatively confirmed from the flows within the US economy itself that which is also evident, with lower degrees of statistical exactitude, in studies at the international level of process of globalisation – i.e. a process of increase in division of labour through international deliveries.

Not merely is this result a confirmation of the fundamental framework Adam Smith (which is scarcely a breakthrough in itself as Smith was confirmed rather a long time ago!) but most importantly it gives precise quantitative numbers to this process. It confirms that division of labour remains the single most important factor in economic growth - even above increased inputs of capital and labour, let alone technology.

This work also illustrates graphically, in regard to present concerns, that increased division of labour and globalisation should not be conceived of as separate processes.

In this regard there is nothing specific about national boundaries – both the rapid increase of intermediate inputs within the national economy and of the process of globalisation itself are part of a single process of the increased division of labour which continues to remain the most powerful lever of economic growth. Increases in inputs of capital and labour, therefore, will only find their desired result as the second most powerful levers of economic growth provided they are placed in an economic framework developing a position within the (now international) division of labour – precisely the process of 'opening' that, for example, characterises China's economic model and is the strategic core of the rapid export growth of the Asian economies.

Internationalisation of the conclusions of Jorgenson's studies on the US economy

Maddison had from the beginning concentrated on international economic development. It has however already been noted that the most early extensive focus of Jorgenson's work was the US economy. Indeed, curiously, Jorgenson appears to have been somewhat surprised (unless he was writing ironically) when he found the same result he had noted for the US applied internationally as well as regarding the US.

In Information Technology and the World Economy, written with Khuong Vu, he notes: 'We allocate the growth of world output between input growth and productivity and find, surprisingly, that input growth greatly predominates'. More precisely: 'we allocate the growth of world output between input growth and productivity. Our most astonishing finding is that input growth greatly predominated! Productivity growth accounted for only one-fifth of the total during 1989-1995, while input growth accounted for almost four-fifths. Similarly, input growth contributed more than 70 percent of growth after 1995, while productivity accounted for less than 30 percent.'

Considering their period of analysis of the international factors in growth as a whole, that is over almost a quarter of a century of economic development, Jorgenson and Vu concluded: 'we allocate the sources of world economic growth during the period 1989-2003 between the contributions of capital and labour inputs and the growth of productivity. We find that productivity… accounted for only 20-30 percent of world growth. Nearly half of this growth can be attributed to the accumulation and deployment of capital and another quarter to a third to the more effective use of labour…'

'The contribution of capital input to world economic growth before 1995 was 1.18 percent, slightly more than 47 percent of the growth rate of 2.50 percent. Labour input contributed 0.79 percent or slightly less than 32 percent, while productivity growth contributed 0.53 percent or just over 21 percent. After 1995 the contribution of capital input climbed to 1.56 percent, around 45 percent of output growth, while the contribution of labour input rose to 0.89 percent, around 26 percent. Productivity increased to 0.99 percent or nearly 29 percent of growth. We arrive at the... conclusion that the contributions of capital and labour inputs greatly predominated over productivity as sources of world economic growth before and after 1995'

Factor inputs and Asia

The fact that division of labour, and growth of factor inputs, above all capital investment, is the dominant element in international economic growth, of course has direct implications for 'Asian' economic growth. 

Its inevitable conclusion is that the claim by Paul Krugman, repeated by others today for China, that Asian economic growth strategies were not viable because they rested on massive mobilisation of capital and labour, rather than asserting a framework of productivity growth, makes no sense when it is found that economic growth in all major economies, both developed and developing, is based primarily on accumulation of factor inputs.

Provided that at least average/reasonable rates of total factor productivity are maintained by the Asian economies, which major studies show is the case, and which is reinforced by the rejection of import-substitution regimes in favour of outwardly facing ones, then factor accumulation, particularly of capital, of the Asian economies, including China, is not irrational but, on the contrary, an example of their superior growth potential compared to the US and Europe.

The rational strategy for Asian, and indeed all economies, is therefore that outlined by Jorgenson regarding US post-World War II growth: 'The overall conclusion from this evidence is that the driving force behind the expansion of the US economy... has been the growth in capital and labour inputs. Growth in capital input is the most important source of growth in output, growth in labour input is the next most important source, and productivity growth is the least important. Clearly, this perspective focuses attention on the mobilisation of capital and labour resources rather than advances in productivity.'[16] That is precisely the approach taken by the successful Asian economies including China.

The implications of this for discussion of the Asian growth model is therefore quite clear. Such econometric findings in fact justify the Asian growth model. If the decisive quantitative element in economic growth is factor inputs, and not factor productivity, then the Asian countries were right to concentrate on factor inputs – above all on investment. An alternative strategy based on raising total factor productivity could not have worked given the basic quantitative constraints on the sources of economic growth – indeed such a process has not operated in the most advanced countries either.

Growth in the G7 economies

If the above findings are considered in more detail it is possible to quantify the implications of these conclusions rather easily. They may be illustrated by taking the growth pattern of the most advanced economies, the G7, in the most recent period. This is outlined in Table 3, which shows the growth of output, growth of inputs and growth in total factor productivity for the G7 economies, considered as a whole, over the period 1989-2006 broken down into three sub-periods.

As may be seen the growth in total factor productivity in the G7 economies was slow - the annual rate of TFP growth in 1989-95 was 0.42%, in 1996-2000 it was 0.60%, and in 2000-2006 it was 0.59% - an average of 0.54%. This means that, in the absence of growth in capital and labour inputs, the annual rate of growth of the G7 economies would also have averaged 0.54% over this period.

In reality the G7 economies grew considerably more rapid. The actual annual average rates of G7 growth were 2.14%, 3.11%, and 2.06% in the respective periods – an average 2.44%. The reason for this, of course, is that the G7 grew primarily not through increases in productivity but through increases in inputs of capital and labour. The contribution of increase of G7 factor inputs to growth never fell below 71.9% and for most of the period it was above 80.0%. The average contribution of factor inputs to G7 growth in this period was 77.8%. The average contribution of productivity growth was 22.3%.

Table 3

Turning to the more detailed breakdown of growth this is shown in Table 4. As may be seen the increase in capital inputs accounted for the absolute majority of economic growth in the G7 in each period – in other words, it was investment which was the most dominant factor in growth.

Table 4

Factor inputs also determine short term trends in economic growth

It may also be noted from Table 4 that growth in the factor inputs of capital and labour dominated not only strategic growth but also short term shifts.

Three periods of G7 growth may clearly be distinguished – slower growth in 1985-1995 and 2000-2006 with a period of more rapid growth in 1996-2000. The acceleration of growth in the 1996-2000 period was a 1.0% a year increase from 2.1% to 3.1% - a significant acceleration. However the increase in total factor productivity was a small 0.2% – the acceleration in 1996-2000, compared to the previous period,  being only from 0.4% to 0.6%. However factor inputs grew by 0.8%. That is, the contribution of the increase in factor inputs to growth acceleration was four times that of the productivity increase.

Therefore growth of factor inputs in G7 which dominated both strategic growth and acceleration and slowdown. Fixed investment was the single biggest source of growth throughout the period.

The G7 economies, in short, performed just like … underpowered versions of the Asian economies.

Critics, in demanding that the Asian economies should not base their rapid growth on factor inputs, in particular investment, are therefore demanding that they achieve something which the G7 economies themselves are not able to achieve! In reality, given such fundamental economic arithmetic, the Asian economies, including China, are entirely rational to base their economic growth on factor accumulation in general and fixed investment in particular.If not they would be confined to the very slow rate of growth of total factor productivity.

There is, in fact, evidence that total factor productivity growth in the Asian economies and China was faster than in the G7, which would, of course, multiply the effect of more rapid factor accumulation, but even without this a strategy based on high levels of trade in GDP, and high levels of investment, was entirely rational for the Asian economies in general and China in particular.

Rather than arrogantly lecturing the Asian countries for reliance on mobilising factor inputs the US and Europe should be copying them. Nor is there anything peculiarly 'Asiatic' or 'Confucian' about the Asian, or Chinese, growth model based on high investment and high levels of trade. It is a perfectly rational utilisation of universal laws of economics. It is 'Asian' only in the sense of where it is geographically occurring, not in the sense of its fundamental economic principles - which are universal in character.

Conclusions

We may therefore now summarise the conclusions of modern econometrics for study of the Asian and Chinese growth models.

1. In all economies the growth of inputs of capital and labour, in particular fixed investment, is decisive in economic growth. The difference between the Asian and advanced G7 economies in this regard is simply that the rate of growth of investment, and other factor inputs, in the Asian and Chinese economies is much more rapid than that in the G7 economies – itself sufficient to account for the much more rapid economic growth rate of Asia and China.

2. Growth due to technological change, or total factor productivity, in the advanced G7 economies is slow, centring on 0.5-0.6% a year and, therefore, if economic growth were dependent on total factor productivity it would be equivalently slow. There is evidence that the growth of total factor productivity is more rapid in the Asian economies and China than in the G7. Nevertheless the rate of growth of total factor productivity in all economies is far slower than the 7-10% a year growth rates achieved by the rapidly growing Asian and Chinese economies due to their high level of factor inputs. Therefore, a strategy based on very high rates of factor accumulation, and in particular very high levels of investment, is entirely rational, and indeed the only possible, route to rapid economic growth for any economy including those of China and Asia. A lowering of the rate of factor inputs, in particular a lowering of the rate of investment, would necessarily lead to a rapid slowing of economic growth.

3. Increasing participation in (a necessarily international) division of labour remains fundamental to economic growth, and maintaining factor productivity - as division of labour is the most fundamental force in the development of economic growth. The Asian economies, including China, are therefore entirely correct to orient to high levels of trade in their economies. Criticism of ‘export led growth’ is misplaced because the term systematically confuses a high and increasing level of trade, i.e exports and imports, in the economy, which is desirable, with a large trade surplus – which is not necessary and, in the case of China, has only existed for a relatively short period during its reform period.

4. The ‘Asian model’ of high levels of investment and export led growth is therefore not only practically successful but is in accord with economic theory and the findings of modern econometric research.

It is evident, therefore, why critics of China’s economic policy, and of the Asian growth model, do not refer to the findings of modern statistical research on economic growth.  Because it would disprove their arguments given such work demonstrates that in all economies the division of labour and the accumulation of factor inputs, in particular investment, is the decisive factor in growth. The export led growth and high investment levels of the Asian economies is an entirely correct economic strategy.

Notes

1. As a number of criticisms are made of Martin Wolf's writings on China in what follows it should be pointed out in fairness that Martin Wolf, unlike some other writers on the issue, notes that the appearance of a large current account surplus by China is a relatively recent development. He notes in Fixing Global Finance: 'Until 2004 it [China] ran a modest current account surplus.' (p84)

2. To take recent examples, the survey of Modern Economic Growth by Daron Acemoglu contains only two references to Jorgenson's work in nearly 1,000 pages, while a standard textbook on Economic Growth, such as that by Robert J. Barro and Xavier Sala-i-Martin, also fails to accord Jorgenson's work the central significance it deserves.

3. Simon Kuznets, Economic Growth of Nations, Oxford University Press, London 1972 p306

4. Dale W. Jorgenson, 'Investment and Growth' in Econometrics Vol. 3, The MIT Press, Cambridge Massachusets 2002 p259.

5. Taking merely some of the selective highlights of this material, in chronological order, key works were Deane and Cole's British Economic Growth 1688-1959 (1962), Mitchell and Deane's Abstract of British Historical Statistics (1962), Maddison's Economic Growth in the West (1964), Feinstein's Statistical Tables of National Income, Expenditure and Output of the UK 1855-1965 (1972), Mitchell's European Historical Statistics 1750-1975 (first published 1975), Matthews, Feinstein and Odling-Smee's British Economic Growth 1856-1973 (1982), Maddison's Phases of Capitalist Development (1982), The Economist's World Business Cycles (1982), The Economist's One Hundred Years of Economic Statistics (1989), Maddison's Dynamic Forces in Capitalist Development: A Long-Run Comparative View (1991).

6. Adam Smith, The Wealth of Nations Books I-III, Penguin London 1999 p109.

7. Adam Smith's classic work is after all entitled The Wealth of Nations and not The Wealth of Regions, despite the fact that the fundamental principles he outlined apply equally at a regional as at international level!

8. Maddison, whose contribution in the field of very long term economic statistics is the one which matches that of Jorgenson's shorter term analyses, later gave quantitative dimensions to these developments. His calculations indicated that around the year 1000 the GDP per capita of Iraq and Iran was about fifty percent higher than Europe, China or India. Measured in million 1990 International Geary-Khamis dollars Maddison calculated the GDP per capita of Iraq and Iran to be around $650 per capita in the year 1000 compared to $450 for Italy and $425 for Netherlands, $466 for China and $450 for India. By 1500 the GDP per capita of Italy, with Venice as its economically leading region, was approaching double that of Iran and Iraq, and was more than a quarter ahead of Belgium and the Netherlands, the next most advanced parts of Europe. By 1700 the GDP per capita of the Netherlands was more than double that of Italy and seventy percent ahead of Britain. Although Britain's economy was larger than that of the Netherlands, in all periods, due to its much larger population, UK GDP per capita did not overtake the Netherlands until the mid-19^th century - before Britain was itself overtaken in GDP per capita by the US in the first decade of the 20^th century. It should be stressed that such calculations are of GDP per capita, and not productivity in the scientific economic sense, but given the orders of magnitude of the differences involved wholly unreasonable statistical assumptions would have to be made to avoid the conclusion that such differences in GDP per capita reflected the relative development of productivity.

9. Although his work was published later Lardy lays these issues out particularly clearly: 'A wide range of empirical studies supports the view that the more outwardly-oriented economies in the 1960s, 1970s and 1980s achieved significantly higher rates of real growth of gross domestic product. These studies showed that this was because more open economies achieved both higher rates of saving and investment as well as more efficient use of investment resources. These efficiencies arise from greater utilisation of existing plants, economies of scale that are sometimes achieved when production is not for the domestic market alone and from the stimulus that competitive pressure from abroad provides for technological change and management efficiencies. In addition, the export sector confers positive effects on productivity in the non-export sector through externalities that include the development of more efficient management, improved production techniques, training of higher quality labour, and an improved supply of imported inputs and so forth…

'Typically, as import substitution policies persisted and domestic production replaced an ever broader range of increasingly capital-intensive imported goods, incremental capital-output ratios for the economy as a whole rose more rapidly than would have been expected… Efficiency was further reduced because the distortions of the inwardly-oriented regime, such as an overvalued exchange rate, discouraged domestic producers from exporting. But without the export market, the scale of production was sometimes too small to reap advantages of scale economies, resulting in inefficient, high cost production.

'These sources of inefficiency reduced the real output of the economy and thus usually reduced savings and investment as well.' Nicholas R. Lardy, Foreign Trade and Economic Reform in China 1978-1990, Cambridge University Press, Cambridge 1993 p8.

10. Unless completely unreasonable assumptions are made about the rate of change of relative prices of investment goods.

11. Simon Kuznets, Economic Growth of Nations, Oxford University Press, London 1972 p306

12. Dale W. Jorgenson, 'Productivity and Postwar US Economic Growth' in Productivity Vol.1 Postwar US Economic Growth, The MIT Press, Cambridge Massachusetts 1995 p1.

13. Dale W. Jorgenson, Frank M. Gollop, Marbara M. Fraumeni Productivity and US Economic Growth, toExcel New York 1999 p316.

14. Dale W. Jorgenson, 'Productivity and Postwar US Economic Growth' in Productivity Vol.1 Postwar US Economic Growth, The MIT Press, Cambridge Massachusetts 1995 p5.

15. Dale W. Jorgenson, 'Productivity and Postwar US Economic Growth' in Productivity Vol.1 Postwar US Economic Growth, The MIT Press, Cambridge Massachusetts 1995 p17.

16. Dale W. Jorgenson, 'Productivity and Postwar US Economic Growth' in Productivity Vol.1 Postwar US Economic Growth, The MIT Press, Cambridge Massachusetts 1995 p4.