Assessing
optimality of export portfolios for five Brazilian economic regions
_________________________________________________________________________________________________
Raul Gouvea, Jana Hranaiova and Sul Kassicieh*
Summary: This paper uses a portfolio approach to examine efficiency of export portfolios in five Brazilian regions. The Markowitz efficient frontier model is used to assess the mean-variance properties of regional export portfolios and the efficiency improvements attainable by rebalancing the proportions of exports in primary, semi-manufactured and manufactured products. The Single Index Model is then used to evaluate the instability of regional exports as a result of global fluctuations. Results indicate that regions with a higher proportion of export earnings coming from manufactured and knowledge-intensive products are less sensitive to global disturbances but offer worse risk-return tradeoffs than the regions relying mainly on primary export products. The current export portfolios are in general efficient, that is they are positioned close to the regional efficient sets.
1.
Introduction
Economic development of individual regions in Brazil has long been a
focus of policy makers in the country. The economic literature argues
that diversifying export portfolios and shifting towards exporting more
manufactured and knowledge-based products will improve the economic
well being of the regions. This paper presents a portfolio approach
that evaluates the success of Brazilian export development policies.
Risk-return tradeoffs available to policy makers are evaluated for five
economic regions and the optimality of existing portfolio structures
is assessed. Sensitivity of regional export earnings to global fluctuations
is then examined.
The paper is organized as follows: Section 2 discusses implications of export promotion strategies for economic development. Sections 3 and 4 talk about export efforts in Brazil and the evolution and structure of Brazilian regional export portfolios. Section 5 presents the methodology used in evaluating performance and instability of regional export portfolios. Section 6 presents the results and section 7 concludes.
2.
Export performance
Export promotion and diversification have been a constant concern of
emerging economies’ policy-makers for the past four decades. Traditional
economic wisdom accepted the approach that concentration of a country’s
exports on a few primary commodities contributes largely to the instability
of that country’s export earnings (Brainard and Cooper, 1968;
Hirschman, 1945; MacBean and Nguyen, 1987; Love, 1989; Montenegro and
Soto, 1996). This traditional view states that the more highly concentrated
a country’s export structure, the lower the probability that movements
in earnings from some of its exports will be offset by countermovements
or stability in others (Labys and Lord, 1990; Love, 1983).
Export earnings instability has traditionally had a negative impact on region’s growth and development through introducing uncertainty about economy’s import capacity, income, employment opportunities, and inflation rate (Athukorala and Huynh, 1987; Dutt and Ghosh, 1996). Acknowledgment that these negative economic impacts have severe short-term and long-term growth and development implications has provided the rationale for implementation of export diversification and promotion strategies. These strategies often translate into an increased share of manufactured and knowledge-intensive products in a country’s export structure. This is based on the expectation that export earnings from manufactured and knowledge-intensive products are more stable than export earnings from primary commodities. This is so because, as the argument goes, the correlation between earnings from manufactured and knowledge-intensive products and primary commodities is likely to be less positive (or negative) than the correlation between earnings of primary commodities (Gersovitz and Paxson, 1990; Alwang and Siegel, 1994). However, the benefits in terms of lowering the risk for a given expected return occur only as long as the move towards manufactured products provides diversification of the export portfolio and the correlation with the incumbent primary products is less than one (Gouvea, and Vasconcellos, 1991; 1993; Gutierrez and Ferrantino, 1997; Wood and Berge, 1997).
The export product categories of a given region are treated as a portfolio of earning assets. An increase in export earnings is desirable, while increasing the variance of those earnings is undesirable. The corollary is that as new products are added to an export portfolio, each product’s contribution to portfolio risk should be measured by the covariance of its earnings pattern with the existing portfolio, rather than the earnings’ returns covariance with individual commodities. In principle, regions should be able to choose how to position their export portfolio on the efficient frontier. Given this, regions familiar with their resources, endowments, markets, and ability to attract foreign direct investment can choose a different composition of their export portfolios.
3.
The Brazilian experience
In 2001, Brazil is still struggling to increase its exports. Brazilian
exports currently account for less than 1% of global exports. During
the 1990s, the country suffered from a lack of attention to strategic
policies towards export promotion (Kleber, 2000). The result has been
a slow growth for the exporting sector throughout the 1990s and early
2000s. In 2000, smaller Latin American economies, such as Mexico, showed
a larger share of global exports, around 2.4%, than Brazil (Secex/Decex,
2001).
This lack of attention to exports, however, is a recent trend. Between 1960s and 1980s, export promotion was one of Brazil’s main policy-makers’ goals. The main objective behind this export promotion drive was to increase trade surpluses and reduce regional imbalances (Baer, 2001; Ipea, 1997; Galvao 1984). This objective was to be obtained through an increase in exports of non-traditional primary, semi-manufactured, and manufactured/knowledge-intensive products. The export promotion/diversification drive also had in mind narrowing the income gaps between these three economic regions. Brazil, like the majority of emerging economies, suffers from acute regional income disparities. The roots of these were established in the long process of development and growth of the Brazilian economy (Guimaraes Neto, 1997; Souza, 1997). Especially the Northeast region was lagging behind the national income level during the period under analysis.
Regional economic development through regional export promotions became a major agenda during the same period, 1960s-1980s. Explicitly and implicitly Brazilian policy-makers played an important role in deciding the spatial allocation of economic activity (Redwood III, 1978, 1979). The idea was to promote the industrialization of regions that were lagging behind the national levels, and promote the exportation of these regional manufactured and knowledge-intensive products.
Contrary to regional economic objectives designed in the 1960s, the Brazilian government's industrial development polices have over the years strongly encouraged the concentration of secondary sector activities in the Southeast and South regions, further reinforcing those regions’ comparative advantage (Redwood III and Jatoba, 1984). By the late 1990s, the most competitive regions in Brazil are still the Southeast and South regions. The Mid-West region is also rapidly becoming Brazil’s third most competitive region in the country.
Export promotion programs have also had a significant impact on the regional allocation of economic activity (Gouvea, 1992). As mentioned by Clements (1988) commenting on the Brazilian experience, an export promotion strategy that emphasizes products that are produced in the South and Southeast region is highly regressive and further enhances the regions’ relative share of national regional exports. Moreover, the battery of export incentives guided towards exports of manufactured and knowledge-intensive products set up to boost regional exports also promoted additional pressure on regional disparities. For instance, Braga (1981), Tyler (1984), and Clemente (1988) pointed out that the allocation of fiscal export incentives added another dimension to the issue of regional disparities in exports further contributing to the lack of regional convergence of the Brazilian export promotion policy.
Table 1 illustrates the growth of Brazilian and regional exports over the period 1980-2000. The Northeast region for instance showed the lowest growth rates, an increase of 60% in export earnings over the 1980-2000 period, increasing from US$ 2.31 billion to US$ 3.72 billion. The South observed a 152% increase over the same period, jumping from US$ 4.94 billion in 1980 to US$ 12.46 in 2000. The Southeast region observed a 146% increase, with exports increasing from US$ 12.16 billion in 1980 to US$ 30.00 billion in 2000. In the same period, the North and Midwest regions exhibited the highest growth rates, 335% and 1,858% respectively. These rates mirror the recent insertion of these two regions in the Brazilian economy, and the increasing dynamism of these regional emerging business frontiers.
Table
1: Evolution of Regional Export Earnings 1980-2000, by product category
(US$ 1,000 FOB)
| 1980 | 2000 | ||||||
| Regions | Primary | Semi-manuf. | Manuf. | Primary | Semi-manuf. | Manuf. | |
| North | 268 | 121 | 202 | 1,115 | 1,082 | 1,114 | |
| Northeast | 1,256 | 495 | 505 | 737 | 1,461 | 1,753 | |
| Southeast | 4,012 | 1,172 | 6,804 | 5,027 | 4,464 | 20,949 | |
| South | 2,921 | 538 | 1,420 | 3,901 | 1,256 | 7,583 | |
| Midwest | 26 | 19 | 16 | 1,479 | 227 | 123 | |
| Total | 8,487 | 2,348 | 9,027 | 12,561 | 8,499 | 32,950 | |
Figure 1 shows relative shares of the five economic regions in Brazil’s total exports for the period 1980-2000. The Southeast and the South regions were Brazil’s leading exporters, accounting for close to two thirds of the country’s total exports for the entire period. Figure 2 shows the breakdown of regional exports by main category of products: primary, semi-manufactured and manufactured products. The South and Southeast regions, accounted for the largest share of manufactured products in their exports structures. The North and Midwest regions exhibited the highest dependency on exports of primary products. As a general trend, all Brazilian regions saw an increase in the share of manufactured and semi-manufactured products in their export structures, and a reduction in the share of primary products, with the exception of Midwest.
Figure 1:
In sum, the Southeast and South regions were export leaders during the period under analysis, accounting for approximately 79.7% for the year 2000. The concentration of most of the Brazilian manufacturing industry and non-traditional agricultural products in these two economic regions largely explains the dominant role of these two economic regions.
Figure 2:
a)
b)
c)
d)
e)
4.
Export portfolios
Export portfolios for five Brazilian economic regions are analyzed:
North, Northeast, Southeast, South, and Midwest. Figure 2 illustrates
the contribution of different regions to total Brazilian exports during
the period of 1980-2000. The proportions remained relatively stable,
with the Southeast region contributing the highest percentage of total
exports, 60%, and Midwest region the lowest percentage, below 5%. At
the same time, exports from the Southeast region were diversified into
manufactured products, while those from Midwest were concentrated mostly
in primary products (Figures 2a-2e). The importance of manufactured
products in the exports from the Northern region declined from 30% to
10% in the late 80’s only to come back in the late 90’s.
Currently, the North region export portfolio is comprised of approximately
equal proportions of all three product categories.
At the same time, the regions with higher proportions of manufactured products are expected to be less susceptible to fluctuations due to global export earnings variability. Thus, we would expect the Southeast region portfolio to be the most efficient in the mean-variance sense and with the lowest beta. The Midwestern export portfolio should then lie on the opposite side of the spectrum.
Three product categories are considered in the export portfolio of each of the five economic regions of Brazil: primary, semi-manufactured and manufactured. This approach avoids the complication associated with considering individual export products. Namely, some products were discontinued as export products during the last twenty years, while others were newly added, resulting in incomplete time-series for these products for the period under analysis. Considering only three general product categories enables us to encompass 100 percent of the regional export portfolios for all twenty consecutive years.
Table 2: Summary Statistics for Regional Export Portfolios
| Primary Products | Semi-manufactured Products | Manufactured Products | Total | |
| North | ||||
| Average Return | 8.28% | 16.58% | 11.32% | 10.23% |
| Std dev | 15.03% | 41.78% | 23.20% | 17.57% |
| CV | 1.82 | 2.52 | 2.05 | 1.72 |
| Northeast | ||||
| Average Return | -0.979% | 7.383% | 8.382% | 3.824% |
| Std dev | 18.05% | 20.30% | 24.35% | 14.68% |
| CV | -18.43 | 2.75 | 2.91 | 3.84 |
| Southeast | ||||
| Average Return | 2.68% | 9.46% | 6.71% | 5.43% |
| Std dev | 17.62% | 24.72% | 14.76% | 11.92% |
| CV | 6.57 | 2.61 | 2.20 | 2.20 |
| South | ||||
| Average Return | 2.81% | 11.84% | 9.88% | 5.92% |
| Std dev | 17.11% | 45.37% | 16.32% | 14.71% |
| CV | 6.08 | 3.83 | 1.65 | 2.48 |
| Midwest | ||||
| Average Return | 32.20% | 19.08% | 14.17% | 24.62% |
| Std dev | 57.88% | 41.99% | 27.99% | 44.47% |
| CV | 1.80 | 2.20 | 1.98 | 1.81 |
Table 2 summarizes the average risk-return tradeoffs for export earnings’ returns for the five regions during years 1980-2000. Coefficients of variation indicate that on average the North region performed the best, followed by Midwest. The export portfolio of the Northeast exhibited the largest average risk per unit of return. In terms of individual product categories, manufactured products represented relatively stable risk-return tradeoffs across the regions, with the coefficient of variation ranging from 1.65 for Midwest to 2.91 for Northeast. Primary product category performed well for the Northern and Midwest regions but poorly for the other three regions. The Northeastern region averaged a dismal – 1% return while maintaining a volatility of 18%.
5.
Methodology
This paper uses the Markowitz model to examine the best portfolios of
export product categories and the Single Index Model (SIM) to estimate
the sensitivity of these export categories to world markets. The Markowitz
mean-standard deviation model (1952, 1959) enables a region to select
portfolios consisting of risky assets, in our case export product categories.
This model can be utilized to determine the percentages of a portfolio
that should be allocated to each product category. The SIM allows us
to estimate how sensitive export portfolios are to global trade oscillations.
Basic microeconomic knowledge tells us that demand and supply forces will combine to cause changes in prices and/or quantities. Murray (1978) states that variations in both price and volume constitute sources of export earnings instability. However, factors affecting the supply of and demand for any particular product are outside the scope of this study. This paper considers export earnings, not prices or volumes.
Mathematical expectations, variances, and covariances of export commodity returns are computed for each region’s export portfolio. They are used as initial parameters for simulations that generate country’s efficient frontier in the Markowitz mean-variance space. Two efficient portfolios are then selected to represent two extreme risk-return choices for export portfolio diversification. Attainable improvement in terms of lower risk is evaluated for the actual export portfolio in year 2000. That is, the standard deviation of an efficient portfolio with the same return as the actual portfolio is compared with the actual standard deviation. Return is measured as a rate of change in the annual dollar earnings.
The major drawback of the Markowitz portfolio variance measure is that the number of covariances to estimate becomes extremely large with an increasing number of portfolio items. This problem is not of major concern to us, since we consider only three product categories. Thus, only a 3x3 variance-covariance matrix has to be estimated.
The efficient frontier is a benchmark measure of success for an export diversification program. Portfolios located on the frontier dominate the portfolios located below them in a mean-variance sense. The efficient frontier can be divided into two equal parts, the top and the bottom, where the dividing point is the Minimum Variance Portfolio (MVP). The general MVP presents the lowest feasible level of standard deviation. The selected optimal portfolios are all on the upper part of the frontier, the efficient set. Thus, they represent choices that for a given level of risk, as measured by standard deviation of export earnings returns, observe the highest available rates of return.
The Single Index Model (SIM) is used to estimate the sensitivity of export portfolio earnings to global export fluctuations. The following ordinary least square regression is estimated for export portfolio returns of each economic region:
,
where
rj,t is the rate of return in year t for export portfolio
of country j, and rm represents the market return, here measured as
a rate of return of the world export earnings. Beta value shows the
degree to which an export portfolio is affected by the oscillations
of the world trade markets.
SIM makes an important assumption about the process of generating returns
on exporting goods. It assumes that returns on exporting products are
correlated with one another because they reflect the global pull of
the global exporting market. The model makes strong assumptions about
the covariance between the residuals. Thus, the SIM’s estimated
variance is just a proxy for the true variance.
6.
Results
The Markowitz Model was used to map different points of the efficient
set; i.e., proportions of different export product categories were changed
to obtain portfolios that minimize risk for a given average return.
Efficient sets with and without short constraints were considered. Efficient
Set I allows for negative weights on product categories and Efficient
Set II does not. That is, Efficient Set I allows for changing an export
category into an import category, if that is the mean-variance optimal
decision. Only the results for the Efficient Set II are reported here,
as the results were similar for both efficient sets. Also, we believe
that turning an export category into an import one is a relatively unrealistic
assumption from a policy-making point of view.
Insofar as there is an infinite number of points on the efficient set, each one of them being an optimal export portfolio allocation, we discuss only a limited number of combinations. We report two extreme portfolios for each region, the minimum variance portfolio (MVP) and the portfolio with the highest attainable return (HAR). These two portfolios represent two extreme policy choices with respect to the risk-return tradeoff of the export portfolio. Policy makers can select a portfolio with more or less risk depending on their policy choices and obtain the highest possible returns for a given level of risk. In addition, the optimal portfolio with the same return as the actual export portfolio in year 2000 is presented. This enables us to evaluate how inefficient the actual portfolio is and what improvement in terms of lower risk can be achieved at the current export earnings’ returns.
Table 3 summarizes the average returns, standard deviations and coefficients of variation for the three optimal portfolios for each region.
Table 3: Results For Optimal Portfolios 1980-2000
North |
Northeast |
||||||
| Portfolios | Return | Std dev | CV | Return | Std dev | CV | |
| MVP | 9.19% | 12.48% | 1.36 | 3.42% | 14.97% | 4.37 | |
| Actual optimized | 12.02% | 18.21% | 1.51 | 6.23% | 16.27% | 2.61 | |
| HAR | 16.58% | 41.78% | 2.52 | 8.38% | 24.35% | 2.91 | |
Southeast |
South |
||||||
| Portfolios | Return | Std dev | CV | Return | Std dev | CV | |
| MVP | 5.39% | 10.75% | 1.99 | 6.68% | 14.45% | 2.16 | |
| Actual optimized | 6.44% | 11.77% | 1.83 | 7.86% | 14.71% | 1.87 | |
| HAR | 9.46% | 24.72% | 2.61 | 11.84% | 45.37% | 3.83 | |
Midwest |
Total Brazil |
||||||
| Portfolios | Return | Std dev | CV | Return | Std dev | CV | |
| MVP | 15.74% | 23.62% | 1.50 | 2.82% | 9.71% | 3.44 | |
| Actual optimized | 29.36% | 49.49% | 1.69 | 6.60% | 12.94% | 1.96 | |
| HAR | 32.20% | 57.88% | 1.80 | 8.85% | 23.07% | 2.61 | |
Coefficients of variation, which measure risk per unit of return, indicate that the two regions with the lowest proportion of manufactured products provide the most favorable risk-return tradeoffs. The minimum variance portfolios (MVP) for the North and Midwest regions yield the lowest risk per unit of return, with coefficients of variation at 1.36 and 1.51, respectively. CV’s increase with increasing risk and variance but still retain a better risk-return tradeoff relative to the other regions. Thus, the regions with a high proportion of manufactured and knowledge intensive products lag in performance of their efficient portfolio choices, with the Northeast region performing the worst. The MVP portfolio in the Northeast region yields a coefficient of variation of 4.37, while the portfolios further along the efficient frontier yield the standard deviation of export earning returns of more than double the average return.
These results support the findings of Alwang and Siegel (1994) that go against the claim of improved export performance when the proportion of manufactured and knowledge-intensive products is increased. Switching to manufactured exports and knowledge-intensive products is not a panacea that guarantees the success of an export diversification program. Export earnings from manufactured products are likely to have a low correlation with a portfolio based on a few primary commodities in the initial stages only. As more and more manufactured products are added to an existing portfolio, the covariance of any new product with the existing portfolio is likely to increase.
Comparison of the efficient portfolios with the same return as the actual portfolio in year 2000 and the actual portfolios shows that the North and South regions had the most inefficient portfolios in year 2000 (Tables 2 and 3). They could achieve the largest improvement in efficiency by rebalancing their current export portfolios and thus placing them on the efficient set. The coefficients of variations can be decreased by 8.4% and 6.5% for North and South, respectively. The Midwest export portfolio appears to be the most efficient, with the coefficient of variation for the actual portfolio only 0.5% from that of the optimal portfolio with the same return.
Table 4 reports the SIM Betas for individual product categories and the shares of exports allocated to the product categories for Brazilian MVP and HAR. Semi-manufactured category contributed the most to the instability of Brazilian export earnings, while primary products proved the most resilient to global export earnings’ fluctuations. Thus, if the goal of Brazilian policy makers is to minimize risk, most exports should actually be allocated to primary products, then manufactured products and the least to semi-manufactured products, 72%, 26% and 2%, respectively. To increase return, at the cost of higher risk, majority of exports should be allocated to manufactured products, 61%. Exports of semi-manufactured products should also increase and those of primary products should decrease.
Table 4: Estimated Model Parameters: Brazilian Total Exports
| Product Categories | Beta | MVP | HAR |
| Return | 3.39 | 7.14 | |
| SIM Std Dev | 8.62 | 12.00 | |
| True Std Dev | 10.02 | 13.94 | |
| Shares | |||
| Primary | 0.50 | 71.33 | 14.71 |
| Semi -Manufactured | 1.74 | 2.21 | 23.84 |
| Manufactured | 0.62 | 26.46 | 61.45 |
Table 5 lists the SIM Betas for individual regions, as well as the shares of exports for these regions for the efficient MVP and HAR. The results indicate that regions with higher shares of manufactured and knowledge-intensive products in their export portfolios are less sensitive to global fluctuations than those more dependent on natural resource-based goods (NRBGs). The South and Southeast regions show the lowest Beta levels, 0.64 and 0.65, respectively. On the other hand, the high Betas of the regions more heavily dependent on NRBGs such as North and Midwest indicate that NRBG-dependent regions are more affected by the global market. With Beta of 1.07, the systematic risk of the Northeast region export portfolio is very close to that of the global export portfolio.
Table 5: Estimated Model Parameters: Total Exports by Region
Region
|
Beta
|
MVP
|
HAR |
| Return | 5.75 | 11.52 | |
| SIM Std Dev | 8.90 | 14.07 | |
| True Std Dev | 11.75 | 16.30 | |
| Shares | |||
| Midwest | 1.45 | 0.88 | 21.22 |
| North | 1.47 | 8.00 | 40.80 |
| Northeast | 1.07 | 21.52 | 0.00 |
| Southeast | 0.65 | 42.97 | 19.92 |
| South | 0.64 | 26.56 | 18.06 |
Results for optimal shares of exports from individual regions indicate that if the policy objective for Brazil is to minimize risk, majority of Brazilian exports should come from the Southeast, South, and Northeast regions. The country can improve its export earnings’ returns, at the cost of higher volatility, by boosting exports from North and Midwest, mainly at the expense of Northeast.
These results suggest that primary products were more conducive to export earnings instability. However, examination of the Betas for individual export product categories (Table 6) reveals that semi-manufactured products contributed the most to the regional export earnings instability. Primary products in the Northeast and North region, in fact provide the lowest Betas overall, at 0.36 and 0.25, respectively. But since they represented a low percentage of export earnings in the two regions, their stabilizing effect was relatively weak. The decline in the importance of primary products as export items in the Northeast region and the simultaneous increase in the semi-manufactured product category served to mimic the variability of the global portfolio, resulting in the overall Beta close to one.
Table 6: SIM Beta by Product Category and by Region
| Primary Products | Semi Manufactured | Manufactured | |
| South | 0.78 | 1.94 | 0.45 |
| Southeast | 0.74 | 1.37 | 0.42 |
| Northeast | 0.36 | 1.21 | 0.60 |
| North | 0.25 | 1.23 | 0.77 |
| Midwest | 1.45 | 1.43 | 0.73 |
| Brazil total | 0.50 | 1.74 | 0.62 |
7. Conclusions
This paper uses a portfolio approach to evaluate export portfolios of
five economic regions in Brazil. The Markowitz model is used to assess
the risk-return choices available to policy makers in the individual
choices and to evaluate efficiencies of existing export portfolios.
Single Index Model is used to evaluate the sensitivity of regional exports
to global fluctuations.
Contrary to the conventional wisdom, regions with the highest proportion of manufactured and knowledge-intensive products perform the worst in the mean-variance sense. The North and the Midwest regions in general offer the best risk-return tradeoffs, allowing policy makers to reposition their export portfolios along the efficient set without compromising return. The regions with the highest proportion of export earnings coming from manufactured products, like South and Northeast perform the worst. However, the main source of instability in these regions was the semi-manufactured product category.
Also, it may be optimal to increase the share of manufactured products in the Brazilian export portfolio, depending on the policy objectives. If risk minimization is the main goal of policy makers, the highest share of Brazilian export portfolios should be allocated to primary products. If Brazil is willing to take more risk, and thus achieve higher return, the share of manufactured products should be increased.
These results suggest possible guidelines for policy makers. It is possible to improve mean-variance performance of export portfolios by rebalancing the shares of export product categories in the Brazilian export portfolio as well as changing the proportions of exports coming from different economic regions. The manufactured products should be emphasized only if Brazil is willing to take more risk to achieve higher returns. If the goal is to minimize stand-alone risk of export portfolios, primary products should be prevalent in the regional export.
If, on the other hand, the goal is to minimize the sensitivity of regional exports to global fluctuations, the proportion of the primary and/or manufactured products should be increased at the expense of the semi-manufactured products.
References
ALWANG, J.; SIEGEL, P. Portfolio Models and Planning
for Export Diversification: Malawi, Tanzania, and Zimbawe. The Jounal
of Development Studies, Vol. 30, No. 2, p.405-422, 1994.
ATHUKORALA, P.; HUYNH, H. Export Instability and Growth: Problems and
Prospects for the Developing Economies. London: Croom Helm, 1987.
BAER, W. The Brazilian Economy. New York: Praeger, 2001.
BRAINARD, W.; COOPER, R. Uncertainty and Diversification in International
Trade. Food Research Institute Studies in Agricultural Economics, Trade
and Development, Vol. 8, No. 3, p.256-285, 1968.
CLEMENTE, B. Foreign Trade Strategies, Employment and Income Distribution
in Brazil. New York: Praeger, 1988.
DUTT, S.; GHOSH, D. The Export Growth-Economic Growth Nexus: Acausality
Analysis. The Journal of Developing Areas, Vol. 30(2), p. 167-182, 1996.
GALVAO, O. Efeitos Espaciais da Política de Comércio Exterior.
In: PIMES (ed.)., Desigualdades Regionais no Desenvolvimento Brasileiro,
Vol. 2. Recife: Pimes, UFPE, 1984.
GERSOVITZ, M.; PAXSON, C. The Economies of Africa and the Prices of
Their Exports. Princeton Studies in International Finance, No. 68. Princeton,
Department of Economics, Princeton University, 1990.
GOUVEA, R.; VASCONCELLOS, G. Avaliação das Estratégias
de Diversificação de Exportações com Base
na Abordagem da Formação de Carteiras de Títulos.
Revista Brasileira de Economia, Vol. 45 (1), p. 41-68, 1991.
GOUVEA, R. Exportações e Desequilíbrios Regionais.
Revista de Economia Política, Vol.12, No. 1, p. 121-131, 1995.
GOUVEA, R.; VASCONCELLOS, G. (1993). La Diversificaccion de Las Exportaciones
y La Eficiencia de La Cartera de Exportacion: Estudio Comparativo de
Los Paises del Sureste de Asia y de la America Latina. El Trimestre
Economico, Vol. LX (1), No. 237, p. 29-52, 1993.
Guimarães Neto, L. Desigualdades e Políticas Regionais
no Brasil. Planejamento e Políticas Publicas, No. 15, June 1997,
p. 41-95.
GUITERREZ, S.; FERRANTINO, M. Export Diversification Trends: Some Comparisons
for Latin America. The International Executive, Vol. 39(4), p. 465-477,
1997.
HIRSCHMAN, A. National Power and Structure of Foreign Trade. Berkeley
and Los Angeles, CA: University of California Press, 1945.
IPEA. Desafios do Desenvolvimento Regional e Urbano. In: IPEA (ed.).
O Brasil na Virada do Milênio. Rio de Janeiro: IPEA, 1997.
KLEBER, K. A Longa Espera pelo Apoio às Exportações.
Fórum de Lideres – Gazeta Mercantil, Ano III, No. 5, p.
6-22, 2000.
LABYS, W.; LORD, M. Portfolio Optimisation and the Design of Latin American
Export Diversification Policies. Journal of Development Studies, Vol.
9(4), p. 556-565, 1990.
LOVE, J. Concentration, Diversification, and Earnings Instability: Some
Evidences on Developing Countries. World Development, Vol. 11, No. 9,
p. 787-793, 1983.
LOVE, J. Export Instability and Recurrent and Development Expenditure
in a Sample of Developing Countries. The Journal of Developing Areas,
Vol. 24, p. 19-26, 1989.
MARKOWITZ, H. Portfolio Selection. The Journal of Finance, VII, p. 77-91,
1952.
MARKOWITZ, H. Portfolio Selection: Efficient Diversification of Investments.
New Haven: Yale University Press, 1959.
MURRAY, D. Export Earnings Instability: Price, Quantitiy, Supply, Demand?
Economic Development and Cultural Change, Vol. 1, p. 61-75, 1978.
MACBEAN, A.; NGUYEN, T. Commodity Policies. New York: Croom Helm, 1987.
MONTENEGRO, C.; SOTO, R. How distorted is Cuba’s Trade? Evidence
and Prediction from a Gravity Model. The Journal of International Trade
& Economic Development, Vol. 5(1), p. 45-68, 1996.
MURAT, M.; PIGLIARU, F. International Trade and Uneven Growth: A Model
with Intersectoral Spillovers of Knowledge. The Journal of International
Trade & Economic Development, 7:2, p. 221-236, 1998.
REDWOOD III, J. (1978). A Promoção das Exportações
e a Política de Desenvolvimento Regional. In: BAER, W. (ed.).
Dimensões do Desenvolvimento Brasileiro. Rio de Janeiro: Editora
Campus, 1978.
REDWOOD III, J. Implicit and Explicit Regional Policies in Brazil: The
Impact of the Public Sector on Spatial Development-Disparities since
the Second War. Unpublished Doctoral Dissertation, University of California
at Berkley, 1979.
REDWOOD III, J.; JATOBA, J. Industrialização e Desigualdades
Regionais. In: PIMES (ed.). Desigualdades Regionais no Desenvolvimento
Brasileiro. Recife: Pimes, 1984.
SOUZA, A. O Impacto das Exportações no Emprego Regional:
Um Aspecto Pouco Considerado na Política de Desenvolvimento do
Nordeste. Planejamento e Políticas Publicas, No. 16, December
1997, p. 233-263.
TYLER, W. A Incidência Regional de Políticas Não-Espaciais
de Incentivos no Brasil. Revista Brasileira de Economia, Vol. 38, No.
3, p. 183-204, 1984.
WOOD, A.; BERGE, K. Exporting Manufactures Human Resources, Natural
Resources and Trade Policy. The Journal of Development Studies, Vol.
34, No. 1, p. 35-59, 1997.
_______________
* Raul Gouvea is Wells Fargo Associate Professor of International
Business at the Anderson Schools of Management at the University of
New Mexico. Jana Hranaiova is Assistant Professor of Finance at the
Anderson Schools of Management and Sandia Federal Lecturer, University
of New Mexico, New Mexico. Sul Kassicieh is the Regent's Professor of
Management of Technology, Albert Franklin Black Professor of Entrepreneurship,
University of New Mexico.