World Bank Reprint Series: Number 288 REP288 1983 Hollis B. Chenery Interaction between Theory and Observation in Development FILE Co0P Reprinted with permission from World Development, vol. 11, no. 10 (1983), pp. 853-61. World Bank Reprints No. 253. Larry E. Westphal, "Fostering Technological Mastery by Means of Selective Infant-Industry Protection," Trade, Stability, Technology, and Equity in Latin America No. 254. Gershon Feder, "On Exports and Economic Growth," Journal of Development Economics No. 255. Mohan Munasinghe, "Third World Energy Policies; Demand Manage- ment and Conservation," Energy Policy No. 256. Keith Marsden and Alan Roe, "The Political Economy of Foreign Aid: A World Bank Perspective," Labour and Society No. 257. James A. Hanson, "Contractionary Devaluation, Substitution in Production and Consumption, and the Role of the Labor Market," Journal of International Economics No. 258. Christiaan Grootaert, "The Conceptual Basis of Measures of House- hold Welfare and Their Implied Survey Data Requirements," The Review of Income and Wealth No. 259. Guy Pfeffermann and Richard Webb, "Poverty and Income Distribu- tion in Brazil," The Review of Income and Wealth No. 260. Pradeep K. Mitra, "A Theory of Interlinked Rural Transactions," Journzal of Public Eco)1o0711cs No. 261. David L. Lindauer and Richard H. Sabot, "The Public/Private Wage Differential in a Poor Urban Economy," Journal of Development Economics No. 262. J. B. Knight and R. H. Sabot, "Labor Market Discrimination in a Poor Urban Economy," Journal of Development Studies No. 263. Carl Dahlman and Larry Westphal, "Technical Effort in Industrial Development: An Interpretative Survey of Recent Research," The Economics of New Technology in Developing Countries No. 264. Michael Bamberger, "The Role of Self-Help Housing in Low-Cost Shelter Programs for the Third World," Built Environment No. 265. Bela Balassa, "The Adjustment Experience of Developing Economies after 1973," IMF Conditionality No. 266. Bela Balassa, "Outward Orientation and Exchange Rate Policy in Developing Countries: The Turkish Experience," The Middle East journal No. 267. Dipak Mazumdar, "Segmented Labor Markets in LDCs," American Economic Review No. 268. Stephen P. Heyneman and William A. Loxley, "The Effect of Primary- School Quality on Academic Achievement across Twenty-nine High- and Low-Income Countries," The American Journal of Sociology No. 269. James R. Follain, Jr., Gill-Chin Lim, and Bertrand Renaud, "Housing Crowding in Developing Countries and Willingness to Pay for Additional Space: The Case of Korea," Journal of Development Economics No. 270. Bela Balassa, "Policy Responses to External Shocks in Sub-Saharan African Countries," Journal of Policy Modeling No. 271. Jaime de Melo and Sherman Robinson, "Trade Adjustment Policies and Income Distribution in Three Archetype Developing Economies," Journal of Development Economics World Development, Vol. 11, No. 10, pp. 853-861, 1983. 0305-750X/83 $3.00 + 0.00 Printed in Great Britain. X 1983 Pergamon Press Ltd. Interaction between Theory and Observation in Development HOLLIS B. CHENERY Harvard Institute for International Development, Cambridge, Mass. Summary. - The author reviews empirical research carried out over the past 30 years with respect to the process of development. He first examines the extent to which economic theories and generalizations derived from the experience of the developed countries are relevant to the developing countries. He next explores the relationship between the 'structural transformation' that occurs in the process of development and economic growth, summarizing his own current research on several issues. Finally he compares applications of the neoclassical growth model to both the more advanced countries, whose economic transformation is near completion, and the developing countries still in the process of structural transformation. The author argues that further empirical work can reduce the controversy between neoclassical and structural approaches to development by making possible the formulation of computable models in whilch the distinc- tion between developing and developed countries can be largely reduced to observable differences in certain statistical parameters: initial conditions, price and income elasticities, and adjustment lags. 1. INTRODUCTION porated into a more comprehensive form of analysis or will it maintain its focus on the The study of less developed economies has special problems of less developed countries? emerged as a separate field of economics in the The potential benefits of the first alternative past 30 years.1 Like other applied branches, are illustrated in the second section of the development economics has been motivated as paper, which examines the apparent differences much by the needs of policy makers as by in the sources of growth of developed and opportunities for theoretical analysis. Despite developing countries. this practical orientation, the theoretical under- pinnings of the field have evolved considerably from the early application of existing neoclassical and Keynesian concepts in the 1950s to the more recent formulation of distinctive models 2. STYLIZED FACTS AND EMPIRICAL designed to explain the observed phenomena of TESTS development. What has been the feedback from this evolu- The comparative studies initiated by Colin tion to the general body of economics? Most of Clark and Simon Kuznets have identified similar it has come from the interaction between theory patterns of structural change in the historical and observation in which the samples studied experience of the advanced countries and include both developed and less developed compared them to recent trends in developing economies. Under this heading one can distin- countries. This approach has provided a basis guish two types of contribution: (i) generaliza- for testing a variety of hypotheses concerning tions of existing theories from testing them the universality of economic behaviour. In against a larger and more diverse body of some fields the result has been to strengthen experience; (ii) the evolution of new concepts existing formulations and extend their range of to explain empirical phenomena. Both types application. In others the existing theoretical are illustrated in the first section of this paper. base has proved to be too narrow to accommo- Although this interaction has been quite date the range of observed behaviour, and the fruitful, its implications for the field of economic empirical tests have led to alternative hypo- development are not yet clear. Will it be incor- theses. I will consider examples of both types. 853 854 WORLD DEVELOPMENT (a) LExtending existing theory other aspects of the economy beyond income levels and selected prices, similar approaches to Fields in which attempts to apply existing the estimation of production functions have theory to developing countries have been been less successful. Although the a priori case generally rewarding include consumer behaviour, for assuming technological regularities across international trade, fiscal policy, and - with countries would appear to be at least as strong some qualifications - the theory of production. as that for uniform consumer preferences, Starting with Houthakker's (1957) confirma- nothing comparable to Engels' Law has yet tion of the validity of Engels' Law in virtually emerged from similar efforts at cross-country all cases, the theory of consumer behaviour has and time series estimation. been tested in cross-section and time-series In the intercountry study of production studies of a number of developing and developed functions, attention has been focused on two countries. The synthesis provided by Lluch, sets of parameters: the elasticity of substitution Powell and Williams (1977) takes advantage of among inputs and the rate of productivity observations from countries having a wide range growth. The estimates of elasticities of substitu- of income levels to verify the existence of quite tion seem to be very sensitive to the choice of uniform income and price effects for all coun- samples and the definition of sectors, so that no tries as well as some systematic shifts associated stable ranking of sectors by degree of substituta- with the level of income. bility has yet emerged.S Productivity growth Economic theory has also benefited con- seems to depend more on the economic environ- siderably from attempts to analyse trade among ment of different countries than on the differ- countries at different levels of development. ences among sectors. Measurement of the factor endowments of The last observation suggests that the omis- different countries and factor proportions of sion of variables describing other aspects of the different commodities has made it possible to economy may be the source of the difficulties apply the Heckscher-Ohlin analysis of com- encountered. Price distortions, surplus labour parative advantage to a wide range of countries and other country characteristics affect most and commodities.2 The identification of a sectors of the economy and need to be allowed category of semi-industrial countries inter- for in some way. This problem is taken up in mediate between advanced and less developed, the next section. and the emergence of the oil exporters as a significant trading group, have opened up a range of questions concerning the evolution of (b) The concept of development the trade patterns of each group. The greater specialization of less developed countries also The most notable effect of the interaction makes them more susceptible to the effects of between theory and observation is the evolution changes in their terms of trade on growth, of the concept of economic development which has been widely studied in recent years.3 itself. The term came into general use in the In sum, the concentration of a number of postwar period to indicate that something more economists on the relations between trade and than increased output was required to raise the development has enriched the theory of inter- welfare of poor countries. Over time the national trade as well as that of development. term has acquired a number of connotations Most of these advances have been encompassed and has no agreed definition for all purposes. within the existing framework of trade theory. As Little (1982, p. 5) points out, much of the On the development side, the greater role of difficulty in defining development stems from trade in explaining growth is one of the features differences in value judgments concerning social that distinguishes developing from more mature objectives. economies, as will be shown below. The empirical studies of the past 30 years The theory of production has been tested in have made it much clearer how countries have comparative studies of developing and developed developed even though they have not resolved countries at both aggregate and sectoral levels. some basic questions as to how they should. As in the case of consumer demand, these The central concept of development that has studies have led to some extensions of the emerged from comparative studies is the 'struc- underlying theory,4 but at the same time they tural transformation' required to sustain have identified systematic differences among increased output or consumption.6 Kuznets' countries that are not explained within the work demonstrates the usefulness of analysing neoclassical paradigm. While demand functions related shifts in resource allocation by measures have been effectively studied without specifying such as the composition of demand, production, INTERACTION BETWEEN THEORY AND OBSERVATION 855 trade and employment. The extent to which the neoclassical framework to both developing the structural transformation has been similar and developed countries and to test the signifi- among countries in the postwar period is cance of the additional factors suggested by explored in Chenery and Syrquin (1975). disequilibrium hypotheses related to the struc- One of the earliest and most influential tural transformation. Since there is no sharp theories incorporating the main features of the distinction between countries at different structural transformation is Lewis' (1954) stages of the transformation, the study of theory of a dual economy. His basic assump- growth in advanced countries may also be tion is that there is a substantial category of enriched by the adoption of a broader analy- developing countries in which the marginal tical framework. productivity of labour is relatively low in traditional sectors such as agriculture and hence the supply of unskilled labour is quite elastic. 3. GROWTH AND TRANSFORMATION9 The reallocation of such 'surplus' labour to more productive uses and the accompanying The concept of the transformation sketched growth of savings form part of a distinctive above leads to quite a different view of the growth mechanism that can persist so long sources of growth in countries to which it as there are potential gains from reallocating applies than the concept of neoclassical or resources. equilibrium growth. Growth may be retarded A common feature of dual economy theories by the requirements of rapid structural change is that they provide an explanation of some of because of lags or imperfections in the mechan- the central features of the transformation, such isms by which resources are allocated, but it as the reallocation of labour among sectors and can also be accelerated when these imperfections the potential acceleration of growth. Further- can also beaceleratedwenc ts imprfctios are overcome. If such a sequence IS characteristic more, the initial structure of demand and pro- of a significant group of countries, it can duction and their change over time constitute provide a partial explanation of the acceleration an essential aspect of the theory, in contrast to of growth that has been characteristic of a general neoclassical theory in which the sectoral number of middle-income countries in the past composition of growth is essentially irrelevant. 20 years. In this context, the neoclassical growth As will be discussed below, this is the main model can be seen as a limiting case in which point of difference between structuralist and neoclassical analyses of development. 8 there are no departures from competitive The principal factor linking growth to equilibrium and growth converges to a steady sThctral change acsthe time r ireforwthe t rate determined by population growth and structural change is the time required for the technical progress. economy to adjust to changes in demand and On a priori grounds, lags in reallocating supply, which leads to persistent differences in resources and differences in factor returns can factor returnls in different uses. The longest be expected to be more important in countries lags are in complex demographic phenomena, starting from an initial disequilibrium or under- such as the demographic transition from high going rapid structural change. Conversely, the to low birth rates and the migration from rural neoclassical assumptions should fit better in to urban areas, both of which take several the more developed countries where the econo- generations. Lags of ten years or more are mic transformation is closer to being completed common in adapting to changes in comparative Since the neoclassical model has been widely advantage, particularly when these require a applied it provides a useful point of departure. shift from primary exports to light manufactures involving a very different set of infrastructure and labour skills. While a well-functioning price (a) Equilibrium growth mechanism can reduce the time required to bring about shifts in resource allocation, in The neoclassical growth equation as devel- practice lagged responses are a nearly universal oped by Solow (1957) has provided the basis characteristic of the structural transformation for empirical work in at least 40 countries. This and the neoclassical equality of returns to equation is derived from an aggregate produc- factors in all uses is the extreme case. tion function that explains total output as a The remainder of this paper will be devoted function of capital and labour inputs and time. to an examination of the effects of introducing In measuring aggregate capital and labour, each elements of structural analysis into the theory component is weighted by its marginal produc- and measurement of sources of growth. Here it tivity; the same procedure could, in principle, is possible to compare the results of applying be applied in the disequilibrium case in which 856 WORLD DEVELOPMENT factors may receive more or less than their countries for several postwar periods.12 These marginal products. provide a basis for the comparison of growth It is the assumption of competitive equi- processes in these two groups as well as a librium in all factor and product markets that starting point for testing the importance of leads to the well-known equation for equilibrium other factors. growth that can be estimated relatively easily The main question that has been studied in this framework is the relative importance of the G V= KGK + LGL +GA (I) three factors in equation (1) in different countries.13 The answers to this question from where GV, GK, GL and GA are the growth separate country studies provide a starting rates of total output (value added), capital, point for examining the variation in these labour and total factor productivity, and ,BK proportions with the rate of growth and the and ,L are the shares of capital and labour in possibility of differences among groups of the total product.10 countries. Because of the arbitrary nature of In subsequent work the conceptual basis some of the estimates of ,B in equation (I), I for this equation has been extended to include will concentrate on the division of sources of changes in the quality of inputs as measured by growth between total factor inputs (f3KGK + the returns to different types of labour and OLGL) and total factor productivity, GA. capital." From the work of Christensen, Figure I summarizes the results of studies of Cummings and Jorgensen (1980), Elias (1978) 12 developed countries and 18 developing and others, estimates of the three sources of countries that cover periods of a decade or growth - capital, labour, and factor produc- more centred on the 1960s.14 For the world tivity - are now available for a substantial economy this period falls between the postwar number of developing as well as developed recovery of the 1950s and the oil crisis of 1973 7 8 I/°X °/ %_ 8; - 2 6 V5 2- * C~~~K A~~~~~J 0 * 4 5 TotoL factor Input growth Figure 1. Relationship between total factor productivity growth and total factor input growth (1960-1973). Developed countries: BE, Belgium; CA, Canada; DE, Denmark; FR, France;lIT,lItaly; JA, Japan; NE, Netherlands; NOR, Norway; SWE, Sweden; UK, United Kingdom; US, United States; GE, West Germany. Developing countries: AR, Argentina; BR, Brazil; CH, Chile; CHA, China/Taiwan; CO, Colombia; EC, Ecuador; GR, Greece; HO, Honduras; HK, Hong Kong; IRE, Ireland; IS, Israel; KO, Korea; ME, Mexico; PE, Peru; PH, Philippines; SP, Spain; TU, Turkey; VE, Venezuela. INTERACTION BETWEEN THEORY AND OBSERVATION 857 and was characterized by generally high and postwar period it was quite significant in Japan, stable growth. The chart plots the two elements Italy, Germany and France. Other authors in equation (1) so as to indicate the trade-off estimated the sources of growth from cross- between them along lines of constant growth country regressions, which make it possible to and also shows its relation to the rate of growth. test the effects of other variables and avoid the The bulk of the countries observed had growth necessity of assuming that returns to capital and rates of between 4 and 7% during this period. labour are equal to their marginal products.'6 The nine developed European countries form a Since it has only been possible to estimate an distinct group (A) characterized by low labour aggregate production function directly without growth (1%), which is offset by high TFP resorting to the general equilibrium assump- growth. (The United States and Canada have tions in a few cases (e.g. Bruno, 1968), cross- somewhat higher growth of labour and less TFP country studies provide some of the best growth.) evidence on the importance of disequilibrium The typical middle-income developing coun- factors. tries in Group B, by contrast, have high growth Cross-country studies can shed light on the of labour inputs (3%) and hence somewhat following questions: higher growth of total factors. Under equilibrium (i) How important are resource realloca- assumptions, they achieve only a quarter of tion and other disequilibrium factors to the their total growth from productivity increases. explanation of growth? At the extreme of this pattern is Brazil, which (ii) Are there significant differences achieved a growth rate of 7% with TFP growth between developed and developing countries of only 1.5%.15 in this respect? The six outliers in group C have growth rates (iii) What is the effect of dropping the averaging over 10%, achieved in roughly equal assumption of competitive equilibrium on proportions by higher factor inputs (mainly estimates of the Solow growth equation? capital) and higher TFP growth as compared The earlier studies focused on the first of to group B. Not surprisingly, the group consists these questions.'7 For this purpose it was not of Japan plus five semi-industrial countries - necessary to formulate an explicit model. The Israel, Spain, Hong Kong, China/Taiwan, and results show significant associations of growth Korea - that tend to be outliers in most com- with resource reallocation, export expansion, parative studies. In the next section it will be capital inflows, and several other factors shown that their growth performance can be related to the structural transformation. explained much better when the effects of Feder (1982) has carried this methodology resource reallocation are added to the elements a step further by modelling the effects of contained in the equilibrium model. resource reallocation and export expansion These results have been tested in more formal explicitly and estimating regression equations regression analyses based on pooled observations that reflect them. Since his results cover both for earlier periods as well as 1960-1973. developed and semi-industrial countries for Although there were a few significant changes the period 1964-1973, they can be compared in growth rates between the two periods, the to the equilibrium estimates reported above. significance of the differences between devel- The general form of the regression equation oped and developing countries, and between used in all the studies cited is an extension of groups B and C of developing countries, is equation (I) ,strongly confirmed. Adoption of the Denison methodology instead of the Jorgensen-Griliches G V = ao +a () + a2 GL + a3 XA + a4Xf + aiX (1967) approach would accentuate the differ- ence between groups A and B, since the esti- (2) mated quality changes are smaller in group B. where XA is a measure of the shift of resources out of agriculture, XE is a measure of the growth of exports, and Xi are additional measures of (b) Disequilibrium growth structural change. Using only the first two explanatory variables gives a cross-country Attempts to allow for the effects of disequi- equivalent of the Solow equation in which the librium in estimating sources of growth date contributions of capital and labour growth from the 1960s. Denison (1967) allows for the are measured by the regression coefficients.18 continuing shift of labour out of agriculture in Table I illustrates the results of Feder's developed countries, showing that in the early estimates by comparing the sources of growth 858 WORLD DEVELOPMENT Table 1. Alternative specifications of sources of aggregate growth Sample: semi-industrial countries, period: 1964-1973 Specification Source (a) (b) (c) (d) (1) Investment 4.97 2.80 2.14 2.20 (78%) (44%) (33%) (34%) (2) Labour 1.62 0.89 1.39 1.72 (25%) (14%) (22%) (27%) (3) Resource allocation 2.00 0.50 (31%) (8%) (4) Export growth 1.85 1.96 (29%) (31%) (5) Residual -0.18 0.72 1.04 0.01 (11%) (16%) Total growth 6.41% 6.41% 6.43% 6.39% Regression equations Specification ao a, a2 a3 a4 N 2 a 0 0.25 0.78 30 0.46 (4.06) (3.45) b 0.01 0.11 0.38 0.87 30 0.71 (2.12) (1.92) (4.1 c 0.01 0.10 0.59 0.145 34 0.79 (2.67) (3.28) (4.73) d 0 0.11 0.74 0.23 0.90 32 0.81 (3.0 ) (4.15) (3.23) (5.07) Source: Feder (1982). in semi-industrial countries as calculated from countries, in contrast to the 20 to 30% that four different specifications: (a) neoclassical is typical of developing countries. only (terms I and 2); (b) reallocation added to (iii) The main effect of dropping the a; (c) export growth added to a; (d) all four assumption that factors are paid their factors. Together with the results of the earlier marginal products is to cast doubt on the studies, they lead to the following responses to equilibrium estimates of labour's contribu- the three questions raised above. tion to growth in developing countries. Not (i) In all of the studies of developing only is the share attributed to labour lower countries, the addition of structural variables in most of the structural specifications, but improves the explanation of the sources of in a number of them the regression coeffi- growth quite substantially. In general, con- cient is not statistically significant. These siderably less growth is attributed to capital results tend to support Lewis' surplus labour accumulation and more to the reallocation hypothesis. For the developed countries on of capital and labour. (In Feder's estimates the other hand, the cross-country results are the reduction in the contribution of capital not substantially different from the average between specifications a and b is from 78% of the equilibrium estimates. to 44% and in Robinson's it is from 60% to Although the importance of capital forma- 3 1%.) tion is reduced in the structuralist specifications (ii) The differences between the sources of Table 1, it remains the most important of growth estimated for developing and explanatory factor, accounting for 30% or more developed countries were quite significant of the aggregate growth in virtually all specifi- where they were both studied.19 In general cations, time periods, and country samples. The structural variables contributed little to the cross-country findings thus coincide with similar explanation of growth in the advanced conclusions of Christensen, Cummings and INTERACTION BETWEEN THEORY AND OBSERVATION 859 Jorgensen (1980) for the advanced countries. The distinctive features of development Finally, the structuralist specifications of economics have been somewhat exaggerated by Table I explain the performance of the rapidly the rather crude techniques and limited data growing countries of group C in Figure I quite available for the analysis of the structural well. None of them appears as an outlier from transformation. Until recently the only multi- these equations. sectoral framework that could be utilized empirically has been some form of input- output analysis, which exaggerates the inflexi- 4. CONCLUDING SPECULATIONS bility of economic systems and only incorporates price effects in the exogenous elements of the The results of empirical studies of growth model. This model is now being replaced in over the past 25 years illustrate the advantages studies of resource allocation by computable of analysing developed and developing countries general eqililibrium models, which allow for the in a common framework while allowing for effects of substitution in final demand and differences between them. The neoclassical trade as well as factor inputs. This more flexible model has proven to be a useful starting point approach can accommodate neoclassical formu- even though it seems to require more extensive lations for advanced economies as well as more adaptation to fit the developing countries. structural assumptions for the study of transi- Further explanation of the factors underlying tional countries.21 In this framework the the differences between the two groups will distinction between developed and developing clearly require a less aggregated analysis that countries can be largely reduced to observable can incorporate some of the constraints on differences in initial conditions, price and resource allocation that characterize the struc- income elasticities, and lags in adjustment. tural transformation. Once hypothetical assumptions about these The desirability of pursuing the feedback elements are replaced by empirical estimates, from development studies to mainstream the controversy between neoclassical and economics is evident in the several fields cited structural approaches to development should in this paper. Instead of regarding the newly be reduced. industrializing countries in group C of Figure I Finally, one can hope that the rapid accumu- as outliers to be dealt with on an ad hoc basis, lation of both data and computing algorithms they may turn out to be illustrative of policies will reduce the dependence of the economist by which developing countries can complete on the properties of a particular analytical the transformation more rapidly and catch up technique. What is now lacking, in my view, to the presently advanced group.20 Extending' is the ability - or willingness - to generalize the boundaries of conventional economics about 'second best' situations and thus to would seem to be more rewarding than limiting broaden the range of underlying assumptions to it to the special case of the already developed fit a wider range of economic behaviour. countries. NOTES 1. The origins of the field of Development Economics being acceptable to most hberal economists: 'Economic are traced in H. Arndt (1981) and 1. Livingstone development (or economic progress or real economic (1981). growth) occurs if there is a rise in the present value of average (weighted) consumption per head.' This 2. Sec for example Flirsch (1974) and Balassa measure is consistent with my suggestion (1977, p. 6) (1979). that 'the transition from a traditional to a developed economy can be defined in general terms as the set of 3. The World Bank, World Development Report, changes required to sustain a continued increase in 1981, gives a summary of empirical findings. income and social welfare'. Both proposals leave open the properties of the social welfare function to be used 4. For example the CES production function was in weighting the growth of consumption. derived by Arrow, Chenery, Minhas and Solow (1961) from observations on a number of developed and 7. In a later assessment, Lewis (1972) points out developing countries. that there are three separate versions of this theory, treating dualism in closed and open economies. These 5. See Nerlove (1967). were also elaborated by Fet and Ranis (1964). 6. Little (1982, p. 6) suggests the following test as 8. Comparisons between neoclassical and struc- 860 WORLD DEVELOPMENT turalist analyses of development are given by Chenery 16. Cross-country estimation has its own problems, (1975) and Little (1982). including the need to assume the same production relations and other effects in all countries. 9. This section is based on Chapter 2 of Chenery, Robinson, and Syrquin (1983). 17. Hagen and Hawrylyshyn (1969), Chenery, Elkington and Sims (1970), Robinson (1971). 10. Assumptions underlying this equation and alter- native formulations are given by Nadiri (1970) and 18. The coefficient a, for the investment term can Branson (1979). only be identified with the marginal productivity of capital to the extent that capital-output ratios are 11. See Denison (1967), Griliches and Jorgensen the same in all countries, since I/V is used as a proxy (1966), Nadiri (1970). forl/K. In Feder'sformulationXA andXEarereplaced by more complex terms. 12. The data are given in Chenery. Robinson and Syrquin (1983), Ch. 2. 19. These differences are explored more fully by Chenery, Elkington and Sims (1970). 13. Nadiri (1972) gives a comparison based on earlier data. 20. A global view of this phenomenon is given in Chenery (1977). 14. The advanced countries are all taken from Christensen, Cummings and Jorgensen (1980), who 21. Empirical multisectoral general equilibrium use a methodology that raises the weighted growth models are descended from the pioneering Norwegian of capital and hence reduces the estimate of total study of Johansen (1960). The Hudson-Jorgensen factor productivity. Kendrick's (1982) estimates (1974) model of the US is of this general type. Appli- for the same period, using the Denison methodology, cations to developing countries are discussed by show TFP accounting for about two-thirds of aggre- Dervis, de Melo and Robinson (1982). gate growth in 9 developed countries compared to the average of 50% shown here. 22. This proposition is argued in Chenery and Raduchel (1971). Little (1982) holds that structuralist 15. 1 have omitted the seven centrally planned views are maintained by economists who mistrust thc economics studied by Balassa and Bertrand (1970) price system because of a preference for government because the data cover only manufacturing, but their controls and that there is 'no such thing as a struc- results fit with the Brazil-Turkey pattern of high turalist theory of growth'. capital growth and relatively low productivity. REFERENCES Arndt, H. W., 'Economic development: a semantic industrialisation', in B. Ohlin, P. 0. Hesselborn, history', Economic Development and Cultural and P. M. Wijkmian (eds.), The International Change, Vol. 29 (April 1981), pp. 457-466. Allocation of Economic Activity (London: Arrow, K. J., H. B. Chenery, B. Minhas and R. Solow, Macmillan, 1977). 'Capital-labour substitution and economic effi- Chenery, H. B., H. Elkington and C. Sims, 'A Uniform ciency', Review of Economics and Statistics, Analysis of Development Patterns', Hlarvard Uni- Vol. 43 (August 1961), pp. 225-250. versity Center for International Affairs, Economic Balassa, B., 'A "stages approach" to comparative Development Report No. 148, Harvard University advantage', in I. Adelman (ed.), Economic Growth (July 1970). and Resources, Vol. 4 (1979). Cliencry, H. B., and W. J. Raduchel, 'Substitution in Balassa. B. and T. J. Bertrand. 'Growth performance planning models', in H. B. Chenery (ed.), Studies in of Eastern European economies and comparable Development Planning (Harvard University Press, Western European economies', American Economic 1971). Review Proceedings, Vol. 60 (May 1970), pp. Chenery, H. B., S. Robinson and M. Syrquin, Industriali- 314-320. zation and Growth (1983, in preparation). Branson, W. El., Macroeconomic Theory and Policy, Chenery. H. B. and M. Syrquin, Patterns of Develop- 2nd ed. (New York 1979). mient, 1950-1970 (Oxford University Press, Bruno, M., 'Estimation of factor contributions to 1975). growth under structural disequilibrium', Inter- Christensen, L. R., D. Cummings and D. W. Jorgensen, national Economic Review, Vol. 9 (February 'Economic growth, 1947-73; an international 1968), pp. 49-6 2. comparison', in J. W. Kendrick and B. N. Vaccara Chenery, H. B., 'The structuralist approach to devel- (eds.), New Developments in Productivity Measure- opment policy', Papers and Proceedings of the ment and Analysis (University of Chicago Press, American Economic Association, Vol. 65 (May 1980), pp. 595-698. 1975), pp. 310-316. Denison, E.. Why Growth Rates Differ (Washington, Chenery, H. B., 'Transitional growth and world 1967). INTERACTION BETWEEN THEORY AND OBSERVATION 861 Dervis, K., J. de Melo and S. Robinson, General Equi- Lewis, W. A., 'Economic development with unlimited libriurn Models for Development Policy (Cambridge supplies of labour', Manchester School of Economic University Press, 1982). and Social Studies, Vol. 22 (May 1954), pp. 139- Elias, V., 'Sources of economic growth in Latin 191. American countries', Review of Economics and Lewis, W. A., 'Reflections on unlimited labor', in L. E Statistics, Vol. 50 (August 1978), pp. 362-370 DiMarco (ed.), International Economics and Feder, G., 'On Exports and Economic Growth' (World Economnic Development (New Y'ork Academic Bank, 1982). Press, 1972). Fei, J. C. H. and G. Ranis, Development of the Labor Little, 1. M. D., Economic Development (New York Surplus Economy (Homewood, Illinois, 1964). 1982). Griliches, Z. and D. W. Jorgensen, 'Sources of measured Lluch, C., A. A. Powell and R. A. Williams. Patterns in productivity change: capital input', American Household Demand and Saving (New York. Economic Review, Vol. 56, No. 2 (May 1966), Oxford University Press, 1977) pp. 50-61. Livingstone. I . 'The developmcnt of development Hagen, 1E., and 0. Ilawrylyshyn, 'Analysis of world economics', ODI Revievw (London: 1981), pp. income and growth, 1955-1965', Economic 1-20. Development and Cultural Change, Vol. 18 Nadiri, M. J., 'Some approaches to the theory and (October 1969) (Part 11), pp. 1-96. measurement of total factor productivity: a surve'. Ilirsch, S., 'Hypothieses regarding trade between Journal of Economic Literature (1970), pp. 1137- developing and developed countries', in 1-1. Gicrsch 1177. (ed.), The International Division of Labor (Tdbin- Na(liri. NI. J.. 'International studies of factor Inpujts gen, 1974). and total factor productivity a brief survey', liouthakker, H. S., 'An international comparison of Reviewe of Inconme and Wealth (1972), pp 129- household expenditure patterns: commcmorating 154. the centenary of Engels' Law', Econormetrica, Nerlove, MN.. 'Recent empirical studies of the CES and Vol. 25 (October 1957), pp. 532-551. related production functions', in M Braun (ed ), Hudson, E. A. and D. W. Jorgensen, 'US energy policy TIhe Theory' and Empirical Analysis of Production and economic groxwth, 1975-2000', Bell Journal (New York: NBER, 1967), pp. 55-122 of Economics, No. 5 (Autumn 1974). pp. 461- Robinson, S., 'Sources of growth in less developed 514. countries: a cross-section study', QuarterlY Journal Johlansen, L , A Multisectoral Studil of Economic of Economics, Vol. 85 (August 1971), pp. 391- Gro;th (Amsterdam: Nortli-liolland, 1960). 408. Jorgensen, D. W. and Z. Griliches, 'The explanation Solow. R., 'Technical change and the aggregate pro- of productivity change', Reviews of Economic duction function'. Reviesv of Economics and Studies, Vol. 34, No. 99 (1967), pp. 249-283. Statistics, Vol. 39, No. 3 (August 1957), pp. Kendrick, J. W., 'International comparisons of recent 312-320. productivity trends', in W. Fellner (ed.), Essays in World Bank, World Development Report, 1981. Contemporary Economic Problemis (American Enterprise Institute. 1982). At the time of writing, the author was Senior Advisor, Economics and Research Staff of the World Bank. This article is based largely on comparative studies prepared under the Bank's research program. Opinions are those of the author and not necessarily of the World Bank. H. Chenery No. 272. J. B. Knight and R. H. Sabot, "The Role of the Firm in Wage Determination: An African Case Study," Oxford Economic Papers No. 273. William G. Tyler, "The Anti-Export Bias in Commercial Policies and Export Performance: Some Evidence from Recent Brazilian Experi- ence," Weltwirtschaftliches A rclhiv No. 274. Ron Duncan and Ernst Lutz, "Penetration of Industrial Country Markets by Agricultural Products from Developing Countries," World Development No. 275. Malcolm D. Bale, "Food Prospects in the Developing Countries: A Qualified Optimistic View," The American Economic Reviezw (with Ronald C. Duncan) and "World Agricultural Trade and Food Security: Emerging Patterns and Policy Directions," Wisconsini Interniational Law Journial (with V. Roy Southworth) No. 276. Sweder van Wijnbergen, "Interest Rate Management in LDCs," Journal of Monetary Econtomics No. 277. Oli Havrylyshyn and Iradj Alikhani, "Is There Cause for Export Optimism? An Inquiry into the Existence of a Second Generation of Successful Exporters," Weltwirtschaftliches Archiv No. 278. Oli Havrylyshyn and Martin Wolf, "Recent Trends in Trade among Developing Countries," European Economic Review No. 279. Nancy Birdsall, "Fertility and Economic Change in Eighteenth and Nineteenth Century Europe: A Comment," Populationi and Development Reviezv No. 280. Walter Schaeler-Kehnert and John D. Von Pischke, "Agricultural Credit Policy in Developing Countries," translated from Handbuch der Landwirtschaft und Ernahrung in den Entwicklungslandern (includes original German text) No. 281. Bela Balassa, "Trade Policy in Mexico," World Development No. 281a. Bela Balassa, "La politica de comercio exterior de Mexico," Comnercio Exterior No. 282. Clive Bell and Shantayanan Devarajan, "Shadow Prices for Project Evaluation under Alternative Macroeconomic Specifications," The Quarterly Journal of Economics No. 283. Anne 0. Krueger, "Trade Policies in Developing Countries," Hand- book of International Economics No. 284. Anne 0. Krueger and Baran Tuncer, "An Empirical Test of the Infant Industry Argument," American Econiomtiic Review No. 285. Bela Balassa, "Economic Policies in Portugal," Economnia No. 286. F. Bourguignon, G. Michel, and D. Miqueu, "Short-run Rigidities and Long-run Adjustments in a Computable General Equilibrium Model of Income Distribution and Development," Journal of Development Economics No. 287. Michael A. Cohen, "The Challenge of Replicability: Toward a New Paradigm for Urban Shelter in Developing Countries," Regional Development Dialogue Issues of the World Bank Reprint Series are available free of charge from the address on the bottom of the back cover. THE WORLD BANK Headqu(arters: U 1818 H Street, N.W., Washington, D.C. 20433, U.S.A. Telephone: (202) 477-1234 Telex: WUI 64145 WORLDBANK RCA 248423 WORLDBK Cable address: INTBAFRAD WASHINGTONDC European Office: 66, avenue d'1ena 75116 Paris, France Telephone: (1) 723-54.21 Telex: 842-620628 Tokyo Office: Kokusai Building 1-1, Marunouchi 3-chome Chiyoda-ku, Tokyo 100, Japan Telephone: (03) 214-5001 Telex: 781-26838 The full range of World Bank publications, both free and for sale, is described in the Catalog of Publicationis; the continuing research program is outlined in Abstracts of Current Studies. The most recent edition of each is available without charge from: PUBLICATIONS SALES UNIT THE WORLD BANK 1818 H STREET, N.W. WASHINGTON, D C. 20433 U.S.A ISSN 0253-2131