South Africa's manufactured international trade in the post-sanctions epoch: patterns and potentials.

Weiner, Ross D. ; Roxo, Trevor ; Kellman, Mitchell 等

I. INTRODUCTION

The decade of the 1990s marks a key epoch in South Africa's international trade. The first Democratic elections in 1994 marks a convenient date from which point the country became able to freely trade with the rest of the world. This opportunity was eagerly embraced. The newly democratic elected government of South Africa committed itself to an outward oriented policy stance (Edwards and Golub 2004), and in turn was accepted as a member of the WTO.

South Africa whose world trade had long been stymied by trade restrictions and sanctions, was expected to grow rapidly. In order to properly benefit from the new potential gains opening up, it became essential for the government of South Africa to be able to identify the likely sectoral and industrial patterns of the newly growing foreign sector. The issue is a classic one. Wherein lies South Africa's likely comparative advantage? A straightforward if simplistic working hypothesis would be that since the main bottleneck standing in the way of South Africa's exports (i.e. the sanctions) had now been removed, the country would be expected to follow the classic pattern demonstrated by many countries in South East Asia. This would involve an initial focus on the traditional gateway to industrialization--the labor-intensive product groups, textile-clothing and miscellaneous manufacturing. In turn, the miscellaneous manufacturing would likely begin with simple nondurable consumer goods. Later, a second phase would be a shift to consumer durables and intermediate and capital goods. The later phase, now observed in countries such as Taiwan, would be a shift to sophisticated high-wage, high tech manufactures.

The Nordas study calculated, for the year 1990, an international competitiveness index, in which it compared the South African relative ratio of value added per wage to that of the U.S., taken as a proxy for the technological envelope. (1) Note that the indices calculated by Nordas for detailed product groups are a form of the Ricardian chain of comparative advantage. As such, the actual competitiveness is not as interesting as the ranking of the indices, since the ranking is invariant with respect to the exchange rate.

The inferences to be drawn from them may be considered of special interest in shedding light on the potentials and prospects of South Africa's international competitiveness patterns as it newly emerged into the new WTO world.

Nordas' conclusions are summarized in the last column of her Table 5 (our Table 1 below):

A higher value for this index corresponds to a higher degree of international competitiveness. Note that for the Wage dimension, the highest level of international competitiveness was found for the "Medium-wage" product group. For the Technology dimension, the Low-technology product group was found to be the most competitive. Finally, utilizing the Orientation dimension, the product group most internationally competitive was the Resource-intensive product group.

If we assume that by 1990 the new government-to-be had already formed its general trade regime strategy, then it would have surmised, given the results of Nordas' study (as summarized in Table 1), that those product groups with the highest innate international competitiveness indices were likely to be the ones most likely to succeed as growth engines within the export sector. We are able to empirically test the extent to which Nordas' results (and methodology) provided a sound basis for forecasting the actual trade patterns during the decade of the 1990s.

Using her results as our benchmark, we examine the actual unfolding of South Africa's exports in the decade following Nordas' study in order to evaluate her results as indicators of the country's trade potential. Since we rely on a different empirical methodology, and use a different data set, we first "calibrate" our study to note the degree of concordance with her 1990 results. In short we will calculate actual measures of South Africa's international competitiveness (in manufactures) utilizing the same three sectoral criteria as did Nordas.

We must first introduce the methodology we use to determine actual levels of international competitiveness, the relative revealed comparative advantage (RRCA). This will be done in Section II below:

II. THE REGIONAL REVEALED COMPARATIVE ADVANTAGE (RRCA) MODEL

1. The Theoretical Model

Many alternative indices and measures may be employed to examine the level and changes in a country's international pattern of international competitiveness and comparative advantage, such as Yeats (1992). In this paper we focus on a newly developed variant of the RCA index. The use of the RCA index has had a long and persistent history since its development by Balassa (1965). Its theoretical underpinnings have been questioned despite efforts by Balassa and others (e.g., Hillman 1980). Yet in practice it is one of the most widely used measures of comparative advantage (e.g., Leamer 1997, Wolf 1999, and Richardson and Zhang 2001 ).

2. The Empirical Framework

Only recently has it become evident that slight variations in Balassa's RCA index enable this model to identify sectoral as well as geographical niche comparative advantage of a country. We here follow Richardson and Zhang (2001) to illustrate this recent theoretical development.

The RCA model may be summarized by an index of a ratio of relative trade shares; that is a ratio of ratios. Specifically in its generic (Balassa's 1965 article) form it calculates the relative competitiveness of a country's industry to that of its other industries, relative to global norms. The reference group here is the World. An application of such a generic RCA index, using the USA as a focus is:

RCA = [usxp.sub.i/usxp]/[wxp.sub.i/wxp] (1)

Where

[usxp.sub.i] : U.S. exports of [product.sub.i]

usxp = U.S. exports of all products

[wxp.sub.i] : World exports of [product.sub.i]

wxp = World exports of all products

This is calculated for each sector (or product). If the value of the index exceeds unity (or 100%) this "reveals" a sector (or product) in which the U.S. enjoys a comparative advantage relative to the rest of the world.

The higher the value of the RCA calculated from equation (1), the greater is the revealed competitiveness, though technically, comparative advantage is indicated if and only if the value of the RCA index is above unity. However, the changes in the values of RCA over time also impart valuable information. Such changes in RCA values may be interpreted as an indicator of the dynamic changes in evolving patterns of comparative advantage over time (Valentine and Krasnik, 2000).

3. The Model Estimated in This Study--Using the OECD as the Reference Group

The index, as presented by Balassa may be generalized. Each italicized term in the ratio above may be replaced by some other choice. Instead of all sectors, the researcher may choose a subset, such as all exports of goods and services. Or perhaps only manufactured goods. Or perhaps all semi-manufactured goods. Or perhaps all manufactured goods excluding non-ferrous metals. Or perhaps all merchandise exports. Or, perhaps only Chemicals. Clearly each such choice narrows the interpretation of the resultant index in terms of identifying a country's products representing its comparative advantage. Instead of World, the researcher may wish to narrow the definition to peer exporters. Perhaps all exporters everywhere in the world. Or perhaps only exporters outside of a regional integrated market. Or perhaps only a group of close rivals against which one wishes to compare the exports of the focus country. To date empirical applications of the RRCA model are relatively new and rare in the literature. Early examples of similar applications are in Kreinin and Plummer (1994), and Richardson and Zhang (2001). The latter of these, noting the pioneering aspect of the use of this model for inter-temporal, inter regional comparisons coined the term Regional Revealed Comparative Advantage (RRCA) Index.

In this paper we calculate South Africa's RRCAs, comparing its export levels (and changes over time) with those of the Industrialized OECD countries of the European Union, the United States and Japan, serving as proxies for the variables [wxp.sub.i] and wxp in equation (1).

RRCA = [X.sub.j/X]/[W.sub.j/W] (2)

Where [X.sub.j] = exports of South Africa of product j

X = total South African manufactured exports

[W.sub.j] = exports of OECD of product j

W = total manufactured exports of OECD.

We chose the OECD (as opposed to perhaps the intuitively appealing choice of other African exporters) as the reference group for the following reason. In order to prosper, if not survive, in today's increasingly competitive international environment a country cannot rely primarily on intra-regional trade. Such trade tends to be "protected" from forces of international change and competition, and hence tends to diverge over time from optimal technological norms. Hence, it tends to fail to reflect a country's true comparative advantage (e.g. Yeats, 1998). Given the speed at which globalization is moving, it may be argued that as a matter of long term viability, South Africa needs to focus on the major markets lying outside of the continent in competition with major world manufactured exporters.

Additional considerations bolster the appropriateness of this choice for reference group. It has been pointed out that South Africa's industrial development pattern is similar to that of the OECD. The following from Nordas (1996) is a good example: R&D expenditure was 1.04% of GDP in 1991, which is relatively high even compared to some OECD countries ... a remarkably sophisticated technological capacity has been built up in production of weapons, nuclear energy, computers, electronics and radiation-therapy to mention a few. (Nordas 1996, p4)

Hence, it is of interest to compare its manufactured-trade patterns with those of that reference group, especially in light of excellent recent work in the literature using the Developing Countries as a reference (e.g. Alves and Kaplan, 2004).

4. The Data

An extensive set of trade data was used for this study. The data obtained as raw files from the U.N. were subsets of the Comtrade data set, compiled and maintained by the United Nations Statistical Office in New York. We applied SAS (Statistical Analysis System) directly to these data files obtained from the U.N.

The data set includes all annual manufactured exports from South Africa and from eight major OECD countries (the original E.E.C. including UK) plus the U.S. and Japan; to "the World" for the years 1992 through 1999.

Each year's data include, in thousands of $U.S. values for 101 manufactured traded commodities, ranging from Standard International Trade Classification (SITC, Rev 1) categories 5 through 8. These were aggregated at the three digit level, which closely corresponds to the widely used four digit ISIC industrial classification. The advantage of using Rev. 1 is that it renders this dataset consistent with historical data, thus allowing for long term analyses.

Though the product groupings are based on South African data sources, the selected product characteristics used in this study are based on U.S. trade data. The scale measure $2 is from David and Kellman (1997), and the others on Hufbauer (1970). Unfortunately corresponding South African measures based on South African trade data are not available at the detailed level of disaggregation used here. As pointed out by a referee, this requires that we assume a high correlation between US and South African values. Such a correspondence would of course be in agreement with the Heckscher-Ohlin model and has a "venerable" history in the empirical trade literature, starting with Leontieff's application of U.S. input- characteristics to U.S. imports in 1950. We believe that the correspondence required here is in fact not far from the truth. However, to strengthen the likelihood of this assumption, we did use the Spearman rank correlation coefficients in Table 4. Finally, note that the main part of the paper in which RRCAs are calculated and compared over time does not require the use of the product characteristics.

The actual products used are all of the available data for manufactured trade (exports) at the 3-digit level of disaggregation for SITC Rev. 1. These are all the products between 512 and 899. There are a maximum total of 102 products in this category. However, not all appeared in either OECD, or in South Africa's exports. The numbers ranged from 100 (for OECD) to 91 (for South Africa).

5. The Empirical Results

The following table shows the estimated RRCA values for each of the product groupings for the year 1992, the closest to the year (1990) used for the RULC figures in Table 1 above.

In a strictly Ricardian world, one need only rank an array of products by descending relative unit labor costs in a chain of comparative advantage to determine ranges of comparative advantage. In a model in which the number of products exceeds 2, as long as each of the products in the chain are traded, then (regardless of the exchange rate) the country must unequivocally enjoy revealed comparative advantage in the first product (or product category) on the left, with the highest RULC (termed the "Competitiveness index" by Nordas) (2). If labor unit costs dominated total average costs, the country could sell these products cheaper than could international competitors. Following this logic 'Nordas' calculations (for 1990) would lead to the expectation that a sustained shift away from anti-protectionist bias, that characterized the decade of the 1990s should have resulted in relatively high RRCA values for products in the Resource Intensive category since this category had the highest relative Unit Labor Cost (RULC)--or "Competitiveness Index" of 0.68, as seen in Table 1 above. Similarly when products are organized by Wage Costs, the product grouping with unequivocally highest Competitiveness index is the Medium Wage grouping, whose RULC equals 0.70--the highest in its category. Finally, in the Technology category, the grouping found in Table 1 to have the highest RULC in 1990 was the Low Technology grouping, with a Competitiveness index of 0.65.

If we now compare these findings with the "actual" RRCA results for 1992, as summarized in the last column in Table 2 above, we find that the product group with the highest RRCA for those products classified by "Orientation" is Resource Intensive (RRCA = 5.28). When the products are classified by Wage Costs, the product grouping with the highest revealed comparative advantage is Medium Wage (RRCA = 1.11, the only one with a value higher than unity). When classified by Technology, the best international competitiveness is found for the Low Technology group (RRCA = 2.66, again the only one greater than unity). We must make clear that these (RRCA) measures did not make any use of wage costs or of measures of value added. They utilized solely and only trade values as their basic inputs. Despite the totally unrelated methodologies, and total different data sets and data requirements, precisely the same product groupings are identified as the most internationally competitive for South Africa as were found using the RULC approach. We find this result very interesting (if not amazing). At a minimum these findings strongly support Yates' (1992) argument that there is a need for further study of the linkages between these two (seemingly unrelated) approaches to measuring competitiveness in international trade.

In addition, the findings here are interesting from several perspectives. First, they indicate that Ricardian labor-cost-differential advantages do indeed translate into actual world competitive stances. (3) One reason this is notable is that it is commonly believed in the Trade Theory literature that international competitiveness in the modern world is the result of complex interactions of scale, product differentiation and research and development, and that the "new" theories of comparative advantage have displaced the older Classical and Neoclassical models as reasonable explainers of the sources of international competitiveness and comparative advantage, assigning them as it were to the "dustbin of history". (4) Nevertheless, we must not forget the statistically significant results consistent with the "simple" Ricardian model obtained by McDougal, Stem and Balassa. In a sense we replicate these "remarkable" findings (5) by demonstrating that South Africa's export pattern of specialization is consistent with relatively lower labor costs.

And this last point is another reason why these results are of special interest. Most work in this area assumes that in South Africa the critical factors which determine its present, as well as future pattern of comparative advantage are associated with scale, or R&D, or resources; as opposed to the case of South East Asia which built its initial export successes on the basis of its relatively inexpensive (skilled) labor. The findings we make here suggest that this latter avenue is not a priori shut in the face of South Africa's future export successes as it continues to find its proper role in the global trading system.

III, SOURCES OF SOUTH AFRICA'S REVEALED COMPARATIVE ADVANTAGE

The discussion above has noted that South Africa is increasing the diversification of its successful manufactured exports. The product group with a clear revealed comparative advantage, the Resource Based products, has shown a clearly decreasing level of RCA indices over the 1990s, while several other product groups, such as Science Based, in which South Africa does not enjoy an international comparative advantage, have gained RCA scores over the decade.

A clear inference one may draw from this is that the sources of South Africa's international comparative advantage are relatively complex. As such, rather than seek an explanation in one simple factor such as cheap labor, one is likely to find the underlying explanation of its international pattern of competitiveness in a combination of factors. In fact, one may posit that several factors may either reinforce each other, or counter each other in determining the optimal mix of manufactured exports in which South Africa will tend to find itself especially competitive in international markets.

An initial approach one may take to identify the likely product characteristics underlying South Africa's comparative advantage is to apply a correlation analysis relating individual product revealed comparative advantage indices with various product characteristics. The product characteristics we will examine are gathered from Hufbauer (1971) and David and Kellman (1997):

1. Average per capita wages. This may be seen as a proxy for capital intensity.

2. Human capital variable. This skill variable is the percentage of the industry's labor force accounted for by professional, technical and scientific personnel.

3. Scale variable. This is the exponent in the regression equation V = [KN.sup.a] where V is the ratio between value added in plants employing N persons and the average value added for the industry. This is obtained from David and Kellman (1997).

4. Consumer good ratio--the percentage of industry output directly purchased by final consumers.

5. First trade data, a measure of sophistication. This is the first date (year) in which the product was traded internationally.

6. Product differentiation variable. This is measured by the coefficient of variation in unit values of industry goods destined for different countries. Differentiated goods have higher coefficients of variations.

The Spearman rank correlations, summarized in Table 3 below were calculated for cross sections of 91 products respectively for each year from 1992 to 1999. The results for the first and last years are in the table. The figures in parentheses are the probability-values, or "p-values". These indicate the probability of being wrong if one rejects the null hypothesis. The correlation is statistically significant if its respective p-value is less than 0.05.

The results above show practically no difference between the association of various embodied characteristics and revealed comparative advantage in the beginning as compared to the end of the 1990s. This would be consistent with the general observation that country comparative advantage tends to change slowly over time; hence also that time series of RRCA indices would be expected to be characterized by high degrees of autocorrelation.

South African comparative advantage is positively associated with capital intensity, and inversely with scale economies, consumer good ratios, and product differentiation indices.

The negative correlation with the consumer good ratio is interesting. It suggests that South Africa has succeeded in excelling with high RCA values in products that are primarily servicing producers, rather than directly for consumers. This is exactly the opposite of the experience for all developing countries, South Asian exporters, or for South East Asian (NIC) exporters of manufactures (Yeats 1992, Table 3). This suggests that the development path being followed by South Africa is not similar to that which propelled the South East Asian exporters from LDC to Small Tiger Miracle status.

Hence, a study of the underlying explainers, or characteristics of South Africa's international competitiveness, is not consistent with typical patterns in the developing world. It is unlike the typical developing country with respect to capital intensity and the consumer good ratio.

IV. CONCLUSION

We identify those sectors of the economy in which South Africa has a comparative advantage We accomplish this through the application of a recently developed variant of Balassa's RCA index. Our findings are consistent with previous studies of the South African economy. An interesting aspect of this paper is the high degree of compatibility of the findings utilizing a variant of Balassa's Revealed Comparative Advantage (RCA) approach, with an alternative approach utilizing the Ricardian chain concept of the Relative unit labor cost. These two approaches have rarely been compared and jointly applied in the literature. The source of South Africa's revealed comparative advantage stems from relatively high capital/labor product embodiments, and a concentration in products servicing producers. This places South Africa in an interesting "middle-ground" between typical developing nations and industrialized countries. In particular, it suggests that the typical "flying geese" pattern observed in East Asian shifting patterns of comparative advantage do not represent a useful tableu for understanding South Africa's pattern of evolving comparative advantage. In East Asia, initial international competitiveness was initially based on relatively low capital/labor embodiments, reflecting relatively inexpensive labor; as well as an initial replacement of Japan in the production of relatively simple manufactures of consumer goods. This is in clear contrast to the evolving South African comparative advantage. APPENDIX A: OECD CLASSIFICATION OF MANUFACTURING INDUSTRIES--PRODUCT COMPOSITION OF SECTORS, BY 3-DIGIT SITC CODE (REV 1) SITC Code(s) Low Technology 521, 611, 612, 631, 632, 641, 642, 651-657, 671-679, 681-689, 691-698, 735, 821, 841, 842, 851, 892 Medium Technology 512-515, 521, 531-533, 551-554, 561, 571, 581, 599, 621, 629, 681-689, 711, 712, 715-719, 731, 732-733, 893, 894, 899 High Technology 541, 714, 721-726, 729, 734, 861 Low Wage 611, 612, 631, 632, 651-657, 721-723, 725, 726, 729, 731, 733, 821, 841, 842, 851, 899 Medium Wage 581, 621, 629, 641, 642, 661-667, 671-679, 681-689, 691-698, 711, 712, 715-719, 724, 726, 735, 891-894 High Wage 512-515, 521, 531-533, 541, 551-554, 561, 571, 599, 714, 732, 734 Resource Intensive 521, 631, 632, 661-667, 821 Scale Intensive 512-515, 521, 531-533, 551-554, 561, 571, 581, 599, 621, 629, 641, 642, 671-675, 731-733, 735, 892-894 Labor Intensive 611, 612, 651-657, 676-679, 691-698, 841, 842, 851, 899 Science Based 541, 714, 726, 734, 861 Specialized Supplier 711, 712, 715-719, 721-726, 729, 891 Source: Nordas (1996), 731.

REFERENCES

Allenye, T. and Subramananian, A. 2001. What does South Africa's Pattern of Trade say about its Labor Markets? International Monetary Fund Working Paper, no. WP/01/48.

Alves, P. and Kaplan, D. 2004. South Africa's Declining Export Shares: The Developing Country Challenge, Trade and Industry Monitor, June, 2-5.

Ariovich, G. 1979. The Comparative Advantage of South Africa as Revealed by Export Shares, South African Journal of Economics, vol. 47 (2), 188-197.

Balassa, B. 1965. Trade Liberalization and 'Revealed' Comparative Advantage, The Manchester School of Economic and Social Studies, vol. 33, 99-123.

Barnes, J. and Kaplinsky, R. 2000. Globalization and the Death of the Local Firm? The Automobile Components Sector in South Africa, Regional Studies, Vol. 34: 9, 797-812.

Bell, T. 1997 Trade Policy, in Michie, J. and V. Padayachee, V. (eds), The Political Economy of South Africa's Transition, Dryden Press, London.

Centre for Investment and Marketing in the Eastern Cape (CIMEC), Annual Report 2000, Eastern Cape: South Africa.

David, B. and Kellman, M. 1997. Scale Economies and International Trade in a Rapidly Growing Region, Journal of Economic Integration, March, 26-46.

Deardorff, A. 1984. Testing Trade Theories and Predicting Trade Flows, in R. Jones and P. Kenen, Handbook of International Economics, Vol. 1, 467-517, New York: Elsevier.

Edwards, L. and Golub, S. 2004. South Africa's International Cost Competitiveness and Exports in Manufacturing, World Development, August, 1323-1339.

Edwards, L. and Schoer, V. 2002. Measures of Competitiveness: A Dynamic Approach to South Africa's Trade Performance in the 1990s, The South African Journal of Economics, September, 1008-1045.

Jenkins, C., Bleaney, M., Holden, M. and Siwisa, N. 1997. A review of South Africa's trade policy, paper presented at the Trade and Industrial Policy Annual Forum, Muldersdrift, September.

Hillman, A. 1980. Observations on the Relation between Revealed Comparative Advantage and Comparative Advantage as Indicated by Pretrade Relative Prices, Weltwirtschaftliches Archiv, vol. 116 (2), 315-21.

Holden, M. 1993. Lessons for South Africa from the New Growth and Trade Theories. South African Journal of Economics, vol. 61: 4, 215-228.

Hufbauer, G. C. 1970. The Impact of National Characteristics and Technology on the Commodity Composition of Trade, in Vernon, R. (Ed), The Technology Factor in International Trade, New York: Columbia University Press, 145-231.

Kreinin, M. and Plummer, M. 1994. Natural economic Blocks: An Alternative Formulation," International Trade Journal, vol. 8 (2), 193-205.

Leamer, E. 1997. Evidence of Ricardian and Heckscher-Ohlin Effects in OECD Specialization Patterns, in Maskus et al eds., Quiet Pioneering: Robert M. Stern and his International Economic Legacy, University of Michigan Press.

Levy, P. I. 1999. Sanctions on South Africa: What Did They Do?, American Economics Review, vol. 89: 2, 415-420.

Loots, E. 1998. Job Creation and Economic Growth, South African Journal of Economics, Vol. 66: 3, 319-336.

Naude, W. 2000. The Determinants of South African Exports: An Econometric Analysis. South African Journal of Economics, Vol. 68: 2, 246-265.

Nordas, H. K. 1996. South African Manufacturing Industries-Catching Up or Falling Behind?, The Journal of Development Studies, vol. 32: 5, 715-733.

Richardson, J. and Zhang, C. 2001. Revealing Comparative Advantage, in Blomstrom and Goldberg, eds. Topics in Empirical Inter-national Economics, University of Chicago Press, 195-227.

Strydom, P.D.E 1995. International Trade and Economic Growth. The South African Journal of Economics, September.

Tsikata, Y. 1999. Liberalization and Trade Performance in South Africa," World Bank informal discussion papers on the South African economy, 13, The Southern African Department of the World Bank.

Vollrath, T. 1991. A Theoretical Evaluation of Alternative Trade Intensity Measures of Revealed Comparative Advantage," Weltwirtshaftliches Archiv,.

Valentine N. And Krasnik, G. 2000. SADC Trade with the Rest of the World: Winning Export Sectors and Revealed Comparative Advantage Ratios, The South African Journal of Economics, vol. 68: 2, 266-285.

Van Zyl, G. and Kotze, E C. 1994. An Analysis of the Tariff Structure of the Motor Vehicle and Related Industries in South Africa. Journal for Studies in Economics and Econometrics, vol. 18: 2, 27-39.

Viner, J. 1955. Studies in the Theory of International Trade, London: George Allen and Unwin.

Wolf, E. 1999. Specialization and Productivity Performance in Low, Medium and high-tech Manufacturing Industries, in Heston A. and Lipsey, R., International and Inter-area Comparisons of Income, Output and Prices, University of Chicago Press.

Yeats, A. 1992. What do Alternative Measures of Comparative Advantage Reveal about the Composition of Developing Countries' Exports? Indian Economic Review, vol. XXVII, no. 2, 139-154.

--1998. Does Mercosur's Trade Performance Raise Concerns about the Effects of Regional Trade Arrangements? World Bank Economic Review, Vol. 12 (1), 1-28.

NOTES

(1.) These competitiveness indices are similar to those pioneered by Lary (1968) who used the relative value added per worker as a proxy for capital intensity at the product level.

(2.) Since Nordas placed the SA Unit Labor Cost in the denominator (relative to that of the U.S.), the higher the calculated RULC, the higher the degree of SA competitiveness.

(3.) The only systematic study of the relationship between RCA, and relative labor intensity indices in the literature is Yeats (1992). That study focused on Asian exporters. This remains a relatively little explored aspect of empirical work in comparative international competitiveness. A recent study of South African trade patterns based on the Ricardian framework is Edwards and Golub (2004). This excellent study, however, does not utilize the RCA framework.

(4.) For examples pertaining to South Africa, see Naude (2000) and Holden (1993).

(5.) The fact that the results obtained from the simple Ricardian model were verified empirically (unlike the preponderant results obtained from more recent and elegant models) was termed by Bhagwati "remarkably successful" (Deardorff 1984).

Ross D. Weiner, * Trevor Roxo, ** and Mitchell Kellman ***

* Department of Economics The City College of the City University of New York rweiner@ccny.cuny.edu 212-650-6213

** Faculty of Economic Sciences University of Transkei Eastern Cape, South Africa

*** Department of Economics The City College of the City University of New York and The City University of New York Graduate Center tiger998@hotmail.com 212-650-6203

**** We wish to acknowledge our thanks to the participants of the Conference on South African Trade and Development held in August 2001 at Unitra, sponsored by the Trade Projection Project of the University of Transkei; and for financial assistance from the USAID/UNCF Terciary Education Linkage Project (TELP). TABLE 1. Relative Unit Labor Costs Calculated by Nordas, 1990 Average VA Average wage per worker per worker RULC Index USA RSA USA RSA (3/4)/(1/2) Sector (1) (2) (3) (4) (5) Low-wage 61.92 9.73 20.58 6.13 0.53 Medium-wage 70.46 17.22 28.28 9.90 0.70 High-wage 122.97 28.62 36.40 13.00 0.65 Low-technology 67.32 14.58 23.51 7.82 0.65 Medium-technology 82.87 17.60 28.92 10.56 0.58 High-technology 86.36 17.37 32.71 10.89 0.60 Resource-intensive 82.01 18.03 23.47 7.56 0.68 Labor-intensive 46.03 8.54 20.51 6.36 0.60 Specialized supplier 67.35 16.22 28.58 10.65 0.65 Scale-intensive 87.20 19.39 28.86 11.46 0.56 Science-based 96.36 16.63 35.81 11.45 0.54 Source: Nordas (1996), Table 5 TABLE 2. Regional Revealed Comparative Advantage of South-Africa By Nordas' Functional Product-Groups, 1992 * Product Group Nordhas Competitiveness RRCA Wage Costs High Wage 0.53 0.53 Medium Wage 0.70 1.11 Low Wage 0.65 0.53 Technology High Technology 0.65 0.13 Medium Technology 0.55 0.73 Low Technology 0.60 2.66 Orientation Resource Intensive 0.68 5.28 Specialized supplier 0.60 0.22 Labor intensive 0.65 0.91 Scale intensive 0.56 1.40 Science based 0.54 0.12 TABLE 3. Spearman Rank Correlation between RRCAs and Selected Explanatory Variables: All Manufactured Exports 1992 1999 Capital per worker 0.41 0.33 (0.001) (0.001) Skill variable -0.15 -0.8 (0.15) (0.44) Average per capita wage 0.18 0.17 (0.07) (0.11) Scale variable -0.28 -0.25 (0.007) (0.002) Consumer goods ratio -0.32 -0.29 (0.001) (0.01) Product differentiation -0.26 -0.26 (0.008) (0.01)