Export performance and economic development: an empirical analysis.

The relationship between exports, economic growth and development, and the differences between export promotion and import-substitution industrialization have been a subject of much interest in the development literature. A consensus emerged among neoclassical economists in the 1970s and the theory of export-led growth evolved into a "new conventional wisdom" (Tyler 1981). Export-led growth has also shaped the development policies of the World Bank. "Which trade strategies have enabled countries to attain high growth and to develop their industrial potential?" This was the main theme of World Bank's 1987 World Development Report. The World Development Report concluded that "rapid growth and efficient industrialization [were] usually associated with outward-oriented [export-led growth] policies on trade" (World Bank 1987:92).

The purpose of this paper is to challenge the results of the empirical literature in support of the neoclassical theory of export-led growth and to provide a theoretical and empirical alternative. Contrary to the neoclassical theory, we will argue that both exports and economic growth are preceded by a long and complex process of structural change and economic development. In the following section we will discuss and critique the empirical literature on export-led growth. Section m will present an alternative framework in studying economic development and exports. Section IV will report the results of testing the neoclassical model and our alternative. Section V will contain our concluding remarks.

Exports-Led Growth Thesis: A Critical Assessment

Export-led growth has been celebrated as the rational and efficient alternative to other strategies of development. "Outward orientation" and export-led growth are argued to generate the necessary "flexibility in shifting the economy's resources to take account of the changing pattern of comparative advantage" [World Bank 1987: 81]. This changing pattern has been explained by the replacement of comparative advantage in land/resource intensive commodities to a comparative advantage in unskilled-labor intensive commodities. Developing countries have, therefore, been advised to specialize in the production and export of unskilled-labor intensive products. Exports, it is argued, will lead to faster economic growth by a) increasing the rate of capital formation; b) increasing specialization and expanding the efficiency-raising benefits of comparative advantage; c) offering greater economies of scale; d) affording greater capacity utilization; e) and inducing faster technological change [Ram 1987; Kavoussi 1984; Bhagwati 1978; Krueger 1978].

The neoclassical/World Bank scenario of growth has drawn on a vast body of empirical research in the past two decades [Balassa 1985; Emery 1967; Kavoussi 1984; Michaely 1977; Tyler 1981; Ram 1985, 1987]. Following Tyler (1981), Feder (1982), and Kavoussi (1984), an important strand of this research has consisted of production function-type models in which exports are included as an additional factor of production.(1) The following procedure has been used in most studies to test the thesis about the positive impact of exports performance on economic growth.

Yt = f(Kt, Lt, Xt)

where

Y = GNP 

K = capital stock,

L = labor force.

X = exports.

Reconstructed in growth terms, the following testable growth equation is obtained.

RY = bo + a I/Y + b2 RL + b3 RX

where

RY = growth rate of GNP,

I/Y = investment-income ratio, a proxy for the growth of capital stock,

RL = growth rate of labor force,

RX = growth rate of exports,

(a) is the marginal physical product of capital, and

(b2) and (b3) are output elasticities with respect to labor and exports.

The above growth equation has been tested by using both time-series and cross section data for various sub-groups of developing countries. Although the results of these studies have varied in some respects, their overall conclusion has been in support of the export-led growth thesis. Using various per capita income benchmarks to disaggregate the data, many studies have run separate regressions for low-income and middle-income or semi-industrial developing countries. Rati Ram found a weaker effect of exports on economic growth of the low-income developing countries than the middle-income countries for the period 1960-72. Repeating the same experiment for the period 1970-77, he found no evidence of such difference, concluding that "the regression results for 1970-77 seem to justify abandoning the conventional wisdom that export performance is unimportant for growth in low-income LDCs" [Ram 1985: 41]. Using per capita income as an independent variable in the growth equation (a proxy for economic development), Bela Balassa found a negative sign for the coefficient of output growth and per capita income, leading him to the conclusion that "for a given increment of capital, labor and exports, the rate of economic growth will be higher the lower is the level of development" [Balassa 1985: 26]. All studies have concluded with varying degrees, that exports lead to superior economic performance (higher growth of output) in developing countries.

The neoclassical scenario of export-led growth has been challenged on a number of theoretical and empirical grounds. For the most part, the neoclassical theory attempts to universalize a non-universal process [Schmitz 1984--85; Evans and Alizadeh 1984]. Export-led growth has been promoted as a universal strategy of industrialization and a superior alternative to import-substitution industrialization. But, it has been demonstrated by a number of researches that, in most developing countries, the process of industrialization and capitalist development has been initiated with import-substitution industrialization. In many cases, import substitution has been a prelude to export promotion. It paved the way for the development of a capitalist home market in developing countries, while setting the stage for export-led industrialization and direct production for the world market [Streeten 1986; Bienefeld 1982; Westphall 1978; Syrquim 1989; Yaghmaian 1989].

In its 1987 World Report, the World Bank identified four types of countries with respect to their trade policies. The World Bank categories included strongly outward oriented, moderately outward oriented, moderately inward oriented and strongly inward oriented countries. For the period between 1953 and 1985, Republic of South Korea, Singapore and Hong Kong were the only three countries listed as strongly outward oriented. Brazil, Chile, Israel, Malaysia, Thailand, Turkey and Uruguay were the eight countries listed as moderately outward oriented between 1973 and 1985.

In the case of Hong Kong and Singapore, industrialization began directly through production for exports. It has been argued that small countries like Hong Kong and Singapore had to enter the manufacturing exports market at an earlier stage than resource-rich countries due to their lack of an extensive base of natural resources [Syrquim, p. 233]. But, on the other hand, South Korea, Turkey, Chile, Israel, and almost all countries praised by the World Bank as outward oriented in their trade policy had experienced a long period of import substitution prior to entering the exports market. In fact, South Korea built a strong industrial base and competitive edge in the period preceding the phase of export promotion. Having achieved extensive industrialization and high economic growth, Korea and similar nations entered the exports market in search of demand for their booming industries.

The neoclassical argument about the superiority of export-led industrialization lies on a causal relationship between economic development, export expansion, and the growth of output. It has been argued that the empirical evidence used in the neoclassical literature fails to prove the causality between exports growth and economic performance as postulated by the neoclassical theory [Jung and Marshall 1985, Sheehey 1990; Dodaro 1991, 1993]. Using the causality test developed by Granger [1969], Jung and Marshall investigated the direction of causality between exports growth and the growth of output for 37 countries and found statistical support for the export-led growth thesis in only four cases [Jung and Marshall 1985]. Challenging the validity and the universality of the export-led growth thesis, Edmond Sheehey replaced exports with a number of variables, including private consumption, government consumption, agricultural output, manufacturing output, and other components of national income in the commonly used production function-type models. Sheehey found equally significant statistical support for all these variables as "determinants" of the growth of output. By finding statistical support for the "promotion" of all major components of GDP, Sheehey concluded that the evidence provided in the neoclassical literature had "no bearing at all on the export-promotion/import-substitution controversy" [Sheehey, p. 11].

Output Growth, Development, and Exports: An Alternative Formulation

Development is a dynamic process of interrelated economic, social, cultural and institutional transformations, leading to changes in the composition of production and sectoral distribution of resources [Kuznets 1979; Chenery 1979; Syrquin 1989].(2) Capitalist development and industrialization in developing countries begin with the gradual weakening and erosion of precapitalist economic relations and modes of production. The process of capitalist development results in the separation of subsistence producers from land (artisan tools and other traditional means of production) and their transformation into wage-laborers. Precapitalist (subsistence) production (mainly agricultural) is replaced with capitalist production based on wage-labor. Coupled with nascent industrialization in the urban areas, this leads to the gradual out-migration of now wage-laborers from the traditional sector of the economy (including the agriculture), a fundamental shift in the distribution of population between the rural and urban areas, and a consequent change in the sectoral distribution of employment. Successful industrialization will potentially result in the growth of manufacturing employment at the cost of employment in agriculture and the traditional sector, and subsequent changes in the composition of economic activity. The process results in a shift towards manufacturing on the supply side and related changes in the composition of consumption on the demand side.

Economic development leads to an increase in "total factor productivity" by shifting resources from low-productivity sectors and industries to those with higher productivity [Feder 1985].(3) Though this process occurs unevenly, nevertheless, changes in resource allocation and composition of economic activity lead to a higher overall growth of output. The process of development and the increase in "total factor productivity" will lead to cost advantages in certain productive activities (industries or firms). Cost competitiveness is the result of a complex set of interrelated factors including learned know how, managerial skill, material cost, location, and cost of labor. Such advantages usually come about in the process of development and industrialization.

Having gained the competitive edge and cost advantage in certain products, the more productive firms in developing countries enter the world market in search of demand for their products. The entry in the exports market is the logical consequence of successful industrialization and economic development. It is only at this stage that developing countries can successfully take advantage of the economies of scale and other benefits that are provided by a larger export market. Thus, as a general tendency, development eventually culminates into a geographical change in the composition of demand and the increase in the share of exports in total demand. In this scenario, exports growth is preceded by economic development and structural changes in the economy. We therefore argue that, while exports may lead to higher economic growth, an equally plausible thesis is that both exports and economic growth are caused by the process of development and structural change.

Empirical Models and Findings

We tested the above thesis and the conventional (neoclassical) export-led growth thesis in a cross section analysis of 30 developing nations. The number of countries was determined by data availability. Our sample includes countries in all of the four categories of trade orientation that are used in the 1987 World Report

In addition to theoretical and methodological differences between our approach and the neoclassical approach, our empirical study also departs from the existing literature in the following ways.

All previous studies of the export-led growth thesis have used labor force or population for the labor variable in their growth equations. Given the existence of large structural unemployment in most developing countries, we content that neither of these variables are appropriate proxies for labor (as used in production function-type models). It is not the labor force but only its employed section that contributes to productive activities and growth. In fact, the process of development has historically given rise to a substantial displacement of (subsistence) producers without, at least in its initial stages, being able to productively recruit them in the manufacturing (modern) sector. The rampant growth of the informal sector in most of the Third World (including some very industrialized developing countries like Brazil and Mexico) is witness to this structural dislocation. To remedy this deficiency, we have used employment instead of labor force in this paper. The lack of available data on employment was a major factor in scaling down our sample size to 30 countries.

In addition, most empirical work in support of the export-growth thesis have limited their studies to the relatively high-growth years preceding 1973, while a few have also included the period between 1973 and 1980 [Balassa 1985; Kavoussi 1985]. Even more recent studies in support of the export-led growth have limited their research to the period prior to 1980s [Mbaku 1989]. Periods of slow growth have been excluded from many studies on the ground that the positive impact of exports on economic growth was weaker in these years [see Sheehey 1990].

We argue that, as a general theory and universal strategy of development (as postulated by the World Bank and the neoclassical economists) the export-led growth thesis must hold true even in economically stagnant periods. If export expansion is to be an engine of growth, the causal relation must find statistical support in growth regressions independent of the actual performance of the economy. This is how other causal relations are approached in econometrics research and the export-led growth thesis must be no exception to the rule. To take account of this contention, we have chosen the sluggish years of 1980-1990 as the time period for this study. We ran the following two sets of growth regressions.

I--Alternative growth equations determining the growth of output.

(1) RY = ao + a1 I/Y + a2 RL + [e.sub.1]

(2) RY = ao + a1 I/Y + a2 RL + a3

RX + [e.sub.2]

(3) RY = ao + a1 I/Y + a2 RL + a3

RYm . Ym/Y + [e.sub.3]

(4) RY = ao + a1 I/Y + a2 RYm .

Ym/Y + a3 RLm . Lm/L + [e.sub.4]

II--Alternative growth equations determining the growth of exports.

(5) RX = ao + a1 RYm . Ym/Y + a2

RLm . Lm/L + [e.sub.5]

(6) RX = ao + a1 I/Y + a2 RYm .

Ym/Y + a3 RLm . Lm/L

+ [e.sub.6]

Where:

Y = GNP in constant prices,

RY = average annual rate of growth of GNp,

Ym = output of the manufacturing sector in constant prices,

RYm = RYm = average annual rate of growth of manufacturing output,

I = investment in constant prices,

I/Y = investment-income ratio, a proxy for the growth of capital stock,

Lm = manufacturing employment,

RLm = average annual growth of manufacturing employment,

L = total employment,

RL = average annual rate of growth of total employment,

X = exports in constant prices,

RX = average annual rate of growth of exports,

RYm. Ym/Y = growth of manufacturing output weighted by the share of manufacturing in total output,

RLm . Lm/L = growth of manufacturing employment weighted by the share of manufacturing in total employment,

[e.sub.1], [e.sub.2], [e.sub.3],

[e.sub.4], [e.sub.5], and [e.sub.6], = stochastic error terms.

Equations (1) to (4) are used to test both the neoclassical theory of export-led growth and our thesis that economic growth is determined by structural transformation and economic development. Growth equations (5) and (6) are used to verify the thesis that exports growth is also preceded by prior economic development and structural change. In addition to conventional neoclassical variables, we have also introduced two new variables that capture changes in the composition of output and distribution of resources in the process of development in order to take account of structural transformations. The variables are the growth of manufacturing output weighted by the share of manufacturing in total output, and the growth of manufacturing employment weighted by the share of manufacturing in total employment respectively.

Data and Sources:

The countries included are Brazil, Chile, Costa Rica, Cyprus, El Salvador, Guatemala, Hong Kong, India, Jamaica, Jordan, Kenya, Republic of South Korea, Malawi, Malaysia, Malta, Mauritius, Nicaragua, Niger, Pakistan, Panama, Philippines, Sierra Leone, Singapore, Sri Lanka, Thailand, Togo, Trinidad, Turkey, Uruguay, and Venezuela. The sources of data used in this study are the World Bank's World Tables, 1992, and ILO's Year Book of Labor Statistics, 1992.

The variables from the World Tables are all in constant 1987 prices. All variables are converted to the U.S. dollar. We have used GDP for the Y variable, share of the manufacturing sector in GDP for Ym, fixed domestic investment for I, and the dollar value of exports divided by the exports price index for X respectively. The labor data was extracted from Table 3-C of International Labor Statistics.

RY, RYm, RX, RL, and RLm are average growth rate that are calculated from the time-series data for the period 1981-1990. They are obtained by regressing LnQt = a + bT, where Qt is the variable under consideration at time T. The rate of growth of Q, lets say RQ, is calculated by RQ = ([e.sub.b]-1). Average ratios (I/Y, Lm/L, and Ym/Y) are calculated as simple averages for the period under study.

The growth equations include a sample of thirty observations--one observation (average rate of growth or simple average) for each variable.

Econometric Results:

Regression results are Reported in Table I and Table II.

TABLE I
Alternative Growth Equations Determining the Growth of Output

Equation   C             I/Y        RL          RX

(1)         -0.291        0.826      0.398
           (-3.14)(**)   (2.2)(*)    (.432)
(2)         -0.28         0.727      0.401       0.289
           (-2.0)        (1.67)     (0.43)      (0.46)
(3)         -0.115        0.20       1.45
           (-2.32)(*)    (1.02)     (3.1)(**)
(4)         -0.088        0.22
           (-1.56)       (0.96)

Equation   RYm . Ym/Y   RLm . Lm/L   R2      F-Statistics

(1)                                  0.156    1.72

(2)                                  0.16     1.69

(3)         3.264                    0.802   35(**)
           (9.23(**)
(4)         3.14         0.72        0.73     6.3(**)
           (6.68)(**)   (0.589)

Figures in parentheses are t-statistics.

(*) indicates statistical significance at the 5% level.

(**) indicates statistical significance at the 1% level.

TABLE II Alternative Growth Equations Determining the Growth of Exports: RX as the Dependent Variable N = 30

Equation   C             I/Y    RYm . Ym/Y

(5)          0.043               0.857
            (4.99)(**)          (4.10)(**)
(6)         -0.201       0.23    0.686
           (-0.45)              (2.29)(*)

Equation   RYm . Ym/Y   R2           F-Statistics

(5)         2.044        0.421       9.8(**)
           (3.5)(**)
(6)         1.87         0.518       6.34(**)
           (3.3)(**)    (3.47)(**)

(*) and

(**), the same as Table 1.

Equation (1) is the typical neoclassical growth equation, while equation (2) includes exports as an additional determinant of growth to test the statistical soundness of the export-led growth thesis. Our results indicate that neither versions of the neoclassical model are statistically significant for the period under consideration. The export variable is not only insignificant, but it also fails to improve the model's goodness of fit (R2 improves from 15.6% in equation 1 to 16.3% in equation 2). This casts a serious doubt on the general conclusions that have been drawn on the basis of previous empirical research. Our results question the universality or the export-led growth thesis. We have shown that during the periods of a sluggish growth or downturn the positive impact of exports on growth cannot be statistically supported for the sample used in this research.

The drop in the estimated t-statistics for the variable I/Y after the inclusion of RX in equation (2) points to the possible presence of multicollinearity in this equation. An often used rule of thumb for detecting the existence of multicollinearity is the comparison of the coefficient of correlation between the independent variables and the correlation between each independent variable and the dependent variable. In our case, the correlation between RX and I/Y is 0.49 and is larger than the correlation between RX and RY (.26), and I/Y and RY (.38) respectively. In some cases, this can be an indication of the problem of multicolinearity in the regression. This rule can be quite reasonable if there are only two independent variables in the regression, but can be unreliable in cases of more than two explanatory variables.

A more appropriate way to detect whether or not multicolinearity is in fact a problem is to examine the standard errors of the coefficients. Multicolinearity is likely to be a problem if estimated coefficients have high standard errors and dropping one (or more) variables from the regression lowers the standard error of the remaining coefficients. The following table presents the standard error of coefficients for equations (1) and (2).

Standard Errors of Coefficients
               Constant   I/Y    RL     RX

Equation (1)   0.09       0.38   0.92
Equation (2)   0.09       0.43   0.93   0.62

As can be seen from the table, the introduction of RX in equation (2) marginally increases

the standard error of the estimated coefficient of I/Y, while leaving others basically unaffected. This can be a possible explanation for the drop in the t-statistics of I/Y in equation (2). Although one can detect some degree of multicolinearity in equation (2), this does not alter the lack of statistical significance of the coefficient of RX. To investigate the effect of RX on RY without the problem of multicolinearity caused by the coexistence of I/Y and RX in equation (2) we dropped I/Y from the regression and obtained the following results.

RY = 0.138 + 0.38 RL + 0.80 GX

(0.43) (0.96) (0.56) t = 0.30 .39 1.428

The numbers inside the parenthesis are the standard errors of the estimated coefficients. The exclusion of I/Y from equation (2) causes a drop in the standard error of the coefficient of RX from 0.62 to 0.56. But, even without the problem of multicolinearity, the coefficient of RX is statistically insignificant (t = 1.48).

Can growth be explained by structural changes (during the economic downturn of the 80s)? Based on our results the answer to this question is affirmative. Equations.(3) and (4) are two variants of our development-based growth thesis. Our results support the thesis that structural change (economic development) is a major determinant of growth in developing countries. We found no statistical support for the positive impact of the redistribution of resources on growth. Although the sign of RLm . Lm/L is positive, it is, nevertheless, statistically insignificant. The variable indicating changes in the composition of economic activity (RYm . Ym/Y) is both positive and statistically significant at the 1% level. Including the structural-change variables also significantly improved the goodness of fit of the growth regression by increasing the R2 to 80%. Our best result was obtained by replacing the export variable with the variable showing the shift of economic activity towards manufacturing (equation 3). We conclude that, at least for our sample and the time period under study, it is not exports but the change in economic activities that causes growth. Thus, the appropriate policy (or the superior policy, using the World Bank terminology) for developing countries could be import substitution, export expansion, or any other policy that enhances their structural change and the process of development.(4)

Table II reports the statistical results of testing the thesis that the successful exports performance is preceded by previous economic development and structural change. We have regressed exports growth on the structural change variables, and the ratio of investment to GNP. The best fit is obtained in equation (6) where exports growth is regressed on ratio of investment to GNP, and the variables capturing both the redistribution of resources and the composition of economic activity. All variables are statistically significant at the 1% (for RYm . Ym/Y and RLm . Lm/L) and 5% (for I/Y) levels.(5)

Our results indicate that the more developed the countries are the more successful they are in competing in the exports market and achieving a higher rate of exports growth. The very same countries also enjoy a higher rate of growth of output. We have shown that the growth rate of output and exports are both determined by the degree of economic development already achieved by the developing country. Thus, we conclude that, the positive relationship between exports and economic growth that was reported by earlier studies has been erroneously understood as an evidence of the positive effect of exports on economic growth. In fact, one can argue that, in periods of expansion (the time period used in most of the earlier studies) exports and output had grown independently of one another but due to other factors in the economy.

V-Conclusions

In this paper we presented an empirical challenge to the World Bank/neoclassical export-led growth thesis. We applied a cross-section regression analysis to test the neoclassical thesis that exports lead to superior economic performance (higher growth of output), and our alternative thesis that both exports and economic growth are determined by prior economic development and structural change. The regression results strongly confirmed our thesis while failing to support the neoclassical thesis. We do not argue against the importance of exports in developing nations. But our results suggest that the empirical evidence in support of the export-led growth thesis must be viewed with some degree of caution.

A more comprehensive empirical analysis of the relationship between exports, development, and economic growth requires the development of a behavioral simultaneous equations system. This system of equations can also include appropriate lag structures to test whether or not exports or any of the structural change variables affect the rate of growth of GNP with a lag. In addition, cross-section analysis assumes away important differences between the structures of various developing countries. This deficiency can be remedied with the analysis of exports, development, and growth for individual countries over time. A time-series analysis of this nature will be meaningful if sufficient data were available for a long span of time in order to take account of the process of development and structural change.

Notes

(1.) Bivariate Spearman Rank Correlation was the main method used in the empirical research on the export-led growth thesis prior to the application of production functions. Most researchers used the average annual real grow;h rate of merchandise exports as an index of export expansion, and the average annual growth rate of GNP as a measure of economic performance. A strong positive correlation between the two growth variables was treated as the empirical proof for the export-led growth hypothesis. The bivariate test has been criticized for its omission of other important determinants of GDP growth. Adding exports in the production function and testing the export-led growth thesis by using multiple regressions has been the main remedy presented in the neoclassical literature to this criticism.

(2.) Our goal in this paper is not to construct a theory of development as such. Our focus is on how development effects exports or to what extent exports are independent of economic development and growth. Capitalist development is an uneven process both locally and internationally. A comprehensive study of the development process requires the study of social and economic transformations of developing nations within the context of a changing global economy (Weeks 1985), Without having a theory of global development, any analysis of trade strategies or strategies of development will be incomplete and ultimately misleading (Yaghmaian 1989). In addition, the study of economic development must be centered around the impact of economic and social transformations on human lives. A proper analysis of this question requires a multidisciplinary study of interrelated social, cultural, political and economic changes that emanate from the process of development. This unquestionably important task is beyond the objectives and scope of the present paper.

(3.) Feder, Syrquim. Chenery and others in the tradition of structuralist school have correctly argued that development is a process of disequilibrium adjustments. Development occurs in conditions of disequilibrium marked by continuous inequality between factor returns across sectors.

(4.) It should be noted that structural change and industrialization have profound and contradictory impact on the environment, social traditions, and ways of life in developing countries. It is beyond the scope of this paper to address these questions in any meaningful way. Suffice it to say that the process of development has always been accompanied by a host of positive and negative effects on the society. Development should no longer be blind to the question of the environment and social norms. These considerations must be incorporated into the theory and practice of economic development.

(5.) All growth equations were tested for the possibility of heteroscedasticity error. We conducted the Breusch-Pagan test and did not detect any sign of heteroscedasticity.

Bibliography

Balassa, Bela, 1985. "Export, Policy Choices, and Economic Growth in Developing Countries after the 1973 Oil Shock," Journal of Development Economics, Vol. 18.

Bienefeld, M. A., 1982. "The International Context of National Development Strategies--Constraints and Opportunities in a Changing World." In M.A. Bienefeld and M. Godfree (ed.), The Struggle for Development: National Strategies in an International Context. Chichester. John Wiely.

Bhagwati, J. I. and T. N. Srinivasan, 1978. "Trade Policy and Development." In Dornbusch and J. A. Frenkel (ed.), International Economic Policy: Theory and Evidence. Baltimore: Johns Hopkins University Press.

Chenery, H. B., 1979. Structural Change and Development Policy. New York: Oxford University Press.

Dodaro, Santos, "Exports and Growth: A Reconsideration of Causality," The Journal of Developing Areas. Vol. 27 (1993), pp. 227-244.

Dodaro, Santos, 1991. "Comparative Advantage, Trade and Growth: Export-Led Growth Revisited." World Development, Vol. 19, No. 9.

Emery, R., 1967. "The Relations of Exports and Economic Growth," Kyklos, Vol. 20, No. 4.

Evans, David and Parvin Alizadeh, 1984. "Trade, Internationalization and Visible Hand," The Journal of Development Studies, Vol. 21, No. 1.

Feder, Gershon, 1982. "On Exports and Economic Growth," Journal of Development Economics, Vol. 12, No. 1-2.

--. 1985. "Growth in Semi-Industrial Countries: A Statistical Analysis," in Hollis Chenery, Sherman Robinson, and Moshe Syrquim (ed.), Industrialization and Growth: A Comparative Study, New York Oxford University Press.

Ghorashi, G. R. 1990. "Measuring Terms of Trade of LDC's: A Different Approach," American Economist, Vol. 34, No. 2.

Granger, C., 1969. "Investigating Causal Relations by Econometrics Models and Cross Spectral Me hods," Econometrica, Vol. 37, No. 3.

International Labor Organization, 1992. Year book of Labor Statistics.

Jung, Woo S., and Peyton Marshall, 1985. "Exports, Growth and Causality in Developing Countries," Journal of Development Economics, Vol. 18, No. 1-2.

Kavoussi, Rostam, 1984. "Export Expansion and Economic Growth: Further Empirical Evidence," Journal of Development Economics. Vol. 14, No. 1-2.

--. 1985. "International Trade and Economic Development: The Recent Experience of Developing Countries," Journal of Developing Areas, Vol. 19, No. 3.

Krueger, Anne, 1978. Foreign Trade Regimes and Economic Development: Liberalization Attempts and Consequences, Cambridge: Ballinger.

Kuznets, S., 1979. "Growth and Structural Shifts," in W. Galenson (ed.) Economic Growth and Structural Change in Taiwan: The Postwar Experience in the Republic of China. Ithaca, NY: Cornell University Press.

Mbaku, John, 1989, "Export Growth and Economic Performance in Developing Countries: Further Evidence from Africa," Journal of Economic Development, Vol. 14, No. 2.

Michaely, Michael, 1977. "Exports and Growth: An Empirical Investigation," Journal of Development Economics, Vol. 4, No. 1.

Ram, R., 1985. "Export and Economic Growth: Some Additional Evidence," Economic Development and Cultural Change, Vol. 33, No. 2.

--. 1987. "Exports and Economic Growth in Developing Countries: Evidence from Time Series and Cross Section Data," Economic Development and Cultural Change, Vol. 24, No. 2.

Schmith, Hurbert, 1984. "Industrialization in Less Developed Countries: Some Lessons of Historic Experience," The Journal of Development Studies, Vol. 21, No. 1.

Sheehey, Edmond, 1990. "Exports and Growth: A Flawed Framework, " The Journal of Development Studies", Vol. 27, No. 1.

Streeten, Paul, 1986. "A Cool Look at Outward Looking Strategies for Development." In Robert Baldwin and David Richardson (ed.), International Trade and Finance, Boston: Little, Brown and Company.

Syrquin, M., 1989. "Patterns of Structural Change." In Hollis Chenery and T. N. Srinivsan (ed.), Handbook of Development Economics, North Holland.

Tyler, William, 1981. "Growth and Export Expansion in Developing Countries," Journal of Development Economics..cs, Vol. 9, No. 1.

Weeks, John, 1985. "Epochs of Capitalism and the Progressiveness of Capitals Expansion," Science & Society, Vol. XLIX, No 4

Westphall, Larry, 1978. "The Republic of Korea's Experience with Export-Led Industrial Development," World Development, Vol. 6, No. 3: 347-382

World Bank, 1987. World Development Report 1987, New York Oxford University Press.

--. 1992. World Development Report 1992, New York. Oxford University Press.

--. 1991. World Tables, Baltimore: Johns Hopkins Press.

Yaghmaian, Behzad, 1989. "Development Theories and Development Strategies," Review of Radical Political Economics, Vol. 22, No. 2-3.

Behzad Yaghmaian, Ramapo College of NJ, Mahwah, New Jersey 07430

Reza Ghorashi, Stockton State College, Pomona, New Jersey 08240.

We wish to thank the anonymous referees of The American Economist for their useful comments on an earlier version of this paper. We also like to thank the Ramapo Foundation of the Ramapo College of New Jersey for the partial funding of this project.