Brazil-Argentina trade and its impacts in brazilian states*/El comercio Brasil-Argentina y su impacto en los estados brasileños

Abstract

This paper aims to evaluate a number of spatial aspects of Brazil's current commercial policy, emphasizing those relating to bilateral trade with Argentina, its main trading partner in Mercosur. A national computable general equilibrium model was developed and implemented (EFES-ARG) -integrated with an interstate trade model-, in order to evaluate the sectoral/regional impacts of different trade strategies towards Argentina. The analysis of the short-run regional aspects of Brazil- Argentina trade relations reveals a trend towards concentration of the level of economic activity in the states of the Brazilian south and southeast. [PUBLICATION ABSTRACT]

Headnote

ABSTRACT: This paper aims to evaluate a number of spatial aspects of Brazil's current commercial policy, emphasizing those relating to bilateral trade with Argentina, its main trading partner in Mercosur. A national computable general equilibrium model was developed and implemented (EFES-ARG) -integrated with an interstate trade model-, in order to evaluate the sectoral/regional impacts of different trade strategies towards Argentina. The analysis of the short-run regional aspects of Brazil- Argentina trade relations reveals a trend towards concentration of the level of economic activity in the states of the Brazilian south and southeast.

JEL classification: C68, F17, R13.

Key words: Computable general equilibrium models; Commercial policy; Interregional trade; Regional analysis.

El comercio Brasil-Argentina y su impacto en los estados brasileños

RESUMEN: Este trabajo trata de evaluar algunos impactos espaciales de la actual política comercial de Brasil, especialmente de las relaciones bilaterales de comercio con Argentina, su principal socio en el MERCOSUR. Para realizar esta tarea y evaluar los impactos regionales y sectoriales de las diferentes estrategias de comercio con Argentina, se ha desarrollado e implementado un modelo nacional de equilibrio general computacional (EFES_ARG), integrando con un modelo de comercio interestatal. El análisis de los impactos regionales a corto plazo del comercio bilateral Brasil-Argentina permite concluir que se observa una cierta tendencia a la concentra- ción de los niveles de actividad económica en los Estados Brasileños del Sur y del Sureste.

Clasificación JEL: C68, F17, R13.

Palabras clave: modelos de equilibrio general computables, política comercial, comercio interregional y análisis regional.

(ProQuest: ... denotes formulae omitted.)

1. Introduction

Since its establishment in December 1994 at the Ouro Preto meeting, Mercosur has endured at least three major crises (devaluation of the real in January 1999, global economic recession starting in the second half of 2000, and the recent Argentine crisis). It is hard to envisage favorable prospects for its reintegration and consolidation in the near future. The economy of Argentina, the main trading partner in the bloc, achieved large surpluses in its trade with Brazil, thanks to the steady appreciation of the real during 1995-1998. These trade surpluses enabled Argentina to offset the difficulties it was increasingly facing in its trade with the United States, the European Union and Southeast Asia, resulting from a lack of competitiveness in its main productive sectors. The devaluation of the real reversed Brazil-Argentina trade flows, generating an increase in Brazil's exports to its neighbor. In addition to losing this considerable trade surplus, Argentina faced other problems such as dwindling privatization-related capital inflows, deteriorating terms of trade, and mistrust surrounding the sustainability of the currency board (which involved free peso-dollar convertibility at a fixed parity), all of this against a backdrop of domestic political instability. At the present time, with the Argentine economy in a process of recovering from a state of near-collapse following the demise of convertibility, the future of Mercosur seems uncertain.

This paper aims to evaluate a number of spatial aspects of Brazil's current commercial policy, emphasizing those relating to bilateral trade with Argentina. Apart from this introduction and the final thoughts, the second section of the study provides a brief description of trade flows between the Brazilian states and other countries in Mercosur (highlighting trade with Argentina). Sections III and IV present the models used in our commercial policy simulations. Section V presents the results of simulations involving hypothetical developments of bilateral agreements with Argentina, and changes in the parameters of trade between the two countries, identifying their implications for Brazil at the subnational level.

2. Structure of trade: 1996-20011

To analyze the structure of Brazil's international trade with Argentina in 1996- 2001, it is first necessary to make a few brief comments on the external situation, current circumstances in Brazil and issues concerning the process of Mercosur consolidation. The period 1997-2000 was one of instability in the global economy, marked by economic crises in various parts of the world that had direct consequences on Brazil's trade flows: a) it became more difficult to finance the country's external trade; b) there was a decline in the purchasing power of several agriculturalgoods- importing countries; c) contagion caused a slowdown in the growth of Latin American economies; and d) the performance of EU countries also weakened.

During this period the situation in Brazil was characterized by very small increases in GDP in the middle years (0.2% in 1998 and 0.8% in 1999), flanked by higher growth rates in 1997 (3.3%) and 2000 (4.5%). In January 1999 the exchange rate regime was revamped2 and the real was put into a float, resulting in a steep devaluation of the local currency. The external and domestic situations had impacts on Brazil's trade balance. Following a period in which exports and imports both expanded (by 11 and 15%, respectively, in 1997 compared to the previous year), the country's external trade retreated in the two ensuing years, before recovering again in 2000 with exports rising by 15% and imports by 13% compared to 1999. This stronger performance was largely the result of changes in export promotion policy, the entry of foreign firms, and also the strategies of firms operating in external trade.3

With regard to Mercosur, despite progress made in the integration process, there are still a variety of factors that hinder the expansion of trade between the countries of the bloc. These include balance of payments instability in a number of countries, infrastructure shortcomings especially in the transport sector; domestic inequalities; technological differences between productive sectors; and the productive structures of the various countries4. Accordingly, steps need to be taken to reduce the differences that exist between member-countries, with a view to overcoming the persistent obstacles that prevent economic integration being brought to fruition.

As pointed out by Mendes [1997], analysis of trade structure based on aggregate trade flows identifies the short-run circumstantial elements of the process more accurately than structural ones, but does not show the behavior of the various economic sectors and types of products involved in trade. Accordingly, as a contribution to our understanding of these aspects, this section aims to study the composition of import and export baskets and the main sectors involved in Brazil-Argentina trade. To carry out this analysis we compare the periods 1996-1998 and 1999-2001. This division is justified because the first period saw the consolidation of the Real Plan, and in 1999 Brazil revamped its currency regime. These two elements are a priori highly relevant in understanding the behavior of Brazil's external trade.

Characteristics of Brazil-Argentina Trade

Table 1 shows the 10 leading export and import categories in Brazil's trade with Argentina. The 10 leading export products account for 61.53% of Brazil's total exports to Argentina. This demonstrates an aspect of concentration in Brazil's export product list. The same feature is also present on the import side, with the 10 leading products accounting for 78.76% of Brazil's total imports from its neighbor. The data shown in table 1 reveal the existence of intra-industry trade, since similar products are included in both baskets in the following chapters: other vehicles, nuclear reactors, plastics and products thereof, organic chemicals and mineral fuels.

A more detailed analysis of the regional and sectoral characteristics of Brazil's trade with Argentina and other Mercosur countries makes it possible to more precisely identify the different patterns in the macro regions, federal units and main productive sectors involved in trade with Mercosur. This analysis leads to inferences concerning: a) the general conditions of sectors in the context of the country's productive structure; b) the behavior of each sector as trade evolves; c) the share of each state and/or region in trade with these countries; and d) interaction between the evolution of trade and local productive structure (see Mendes [1997]).

In its trade with Argentina, Brazil's exports slipped from 12.46% (1996-1998) to 10.46% (1999-2001), partly as the result of a worsening of the Argentine crisis (see table 2). In terms of Brazil's macro regions, this trend shows through as a decline in Argentina's importance as a destination for exports from the southeast (from 15.20% to 11.57%), and from the northeast and center-west. Exports from the southern region of Brazil to Argentina held steady throughout the period. In contrast, trade between Brazil's northern region and Argentina expanded from 3.39% (1996-1998) to 8.17% (1999-2001).

Table 2 also reveals the importance of Argentina as a destination for merchandise exports from a number of Brazilian states (see figure 1). Of total exports from Amazonas, about 21% was sent to Argentina in the first period, rising to 29.74% in 1999- 2001. In the northeast region, the states of Bahia, Ceará and Sergipe claimed the largest proportion of trade with Argentina in both periods.

For the vast majority of federal units, exports to Argentina exceed the total exported to Paraguay and Uruguay. Table 2 shows the small proportion of exports from Brazil's macro regions and states that goes to these countries.

Brazil's external trade with Argentina displays great regional concentration, with the southeast and southern regions accounting for over 85% of the total exported in both periods (table 3). The southern region gained share at the expense of the southeast during the period under analysis. The state of São Paulo accounted for over 50% of exports to Argentina, followed by Rio Grande do Sul, Minas Gerais, Paraná, Santa Catarina, Bahia and Rio de Janeiro, which between them accounted for 39% of exports to that country.

This structural differentiation between regions and states stems from the historical concentration of economic activity in the southeast and south of the country. Another relevant point is that a proportion of exports from the southern and southeastern states may include products manufactured or originating in other states and regions of Brazil (i.e., re-exports). Thus the simple analysis of external trade cannot capture interstate trade, yet in many cases this generates more income for the state than international trade does.

3. The EFES-ARG model

A national computable general equilibrium model was developed and implemented (EFES-ARG), in order to evaluate the sectoral impact of different aspects of Brazil's commercial policy with Argentina. The model's structure represents an extension of the EFES model [Haddad and Domingues (2001)], which is a deterministic model, specified to generate annual projections for the Brazilian economy. It can also be used for comparative statics exercises in short-run simulations (with constant capital stock). The model identifies 42 sectors and 80 products, two products used as margin (commerce and transport services), three types of indirect tax, and five user groups (producers, investors, households, external sector and «other demands»). Its extension (EFES-ARG) pays special attention to the specification of international flows. The external sector was broken down into six different components representing specific trade blocs, namely Argentina, rest of Mercosur, NAFTA, rest of FTAA, EU, and rest of the world. This makes it possible to evaluate the effect of policies relating to changes in the structure and determinants of bilateral trade flows in the Brazilian economy5.

The mathematical structure of EFES-ARG is based on the MONASH model, developed for the Australian economy [Dixon and Parmenter (1996)]. EFES-ARG belongs to the Johansen class of models, which produce solutions on the basis of a system of linearized equations. Atypical result shows the percentage variation in the set of endogenous variables following implementation of a given economic policy, compared to their values in the absence of that policy in a given economic setting. The schematic presentation of Johansen solutions for these models is standard in the literature. Further details can be found in Dixon et al. (1982), Harrison and Pearson (1994, 1996), and Dixon and Parmenter (1996).

In this paper EFES-ARG was integrated with an interstate trade model such that the national results obtained were regionalized. The interstate model is presented in the next section.

4. Interstate trade model6

The development of the interstate trade model is based on Haddad, et al [1999] and was implemented for the first time in Haddad et al. (2002). Whereas that article dealt with trade flows between countries in a global economy, the present study focuses attention on interactions between states in a national economy. A matrix of interstate trade flows was constructed for 1997, based on data from the Conselho de Política Fazendária (Confaz, 1999) and IBGE (IBGE, 1999)7. Given production and final demand in each state, the following identity is established:

...[1]

where:

...[2]

...[3]

and:

CIF [arrow right] private consumption in state i

Ii [arrow right] investment in state i

Gig [arrow right] government spending in state i

Xi [arrow right] exports from state i

Mi [arrow right] imports by state i

X and M are composed of interstate domestic and external flows, i.e. they encompass both interstate and international flows. The components of domestic absorption are consumption, investment and government expenditure.

The trade flows X and M for each state can be broken down into two parts, domestic and external:

...[4]

...[5]

xij represents sales from state i to state j; Xi represents exports from state i to other countries; similarly, mij represents purchases by state i from state j, and Mi represents purchases made by state i abroad. By definition, the interstate flow matrices [xij] and [mij] are the same.

Substituting [4] and [5] in [1], gives:

...[6]

This enables us to obtain a matrix similar to the traditional input-output system, the rows of which contain the sales made by each state to all other states (interstate flows) together with final demand, representing the total distribution of the state's production. The columns represent the structure of expenditure in each state.

In this theoretical framework, the key assumption involves a fixed domestic import coefficient, similar to the technical coefficient of the input-output matrix:

... where Zj is total expenditure by state j

The coefficient tij measures the proportion of total expenditure by state j on imports from state i, and the diagonal element (tij for i = j) is null. As in input-output models, this proportion is assumed fixed regardless of the state's total expenditure. Accordingly, for each state there is an optimal amount of imports for any level of expenditure in a given period.

Based on this hypothesis, equation [6] can be written as follows:

...[7]

Where Fi is the final demand in state i.

This n-equation system can be written in matrix notation as follows:

...[8]

where:

T is the matrix of interstate import coefficients (nxn).

Z is the vector of total output (nx1).

F is the final demand vector (nx1).

Solving [8] gives the output of each state needed to satisfy the total demand for domestic production:

...[9]

In other words, given the exogenous components of domestic absorption and external demand, Z measures the output of each state needed to satisfy this final demand. (I - T)-1 is the Machlup-Goodwin domestic trade multiplier matrix, which captures the direct and indirect impacts of changes in final demand in a given state on the total production of all states, given the existing interstate trade structure.

In the same way input-output models operate, the effects of an increase in final demand can be observed through (I - T)-1. For example, assuming an increase in final demand in the state of São Paulo, and given that state's menu of domestic imports (tij for j = São Paulo), the first impact would be a direct rise in the state's import requirements and, hence, an increase in exports to São Paulo from other states. The income generated by São Paulo's purchases in other states generates an increase in production followed by further increases in expenditure. These effects have repercussions throughout the economy, whose total effect is given by the trade multiplier matriz (I - T)-1.

5. Regional aspects of Brazil-Argentina commercial policy

In this section, we use the EFES-ARG integrated to the interstate trade model to evaluate short-run aspects of Brazil-Argentina commercial policy.

5.1. Description of Simulations

The simulations performed in this section represent four Brazil-Argentina trade scenarios. They were developed on the basis of recent events in Argentina and aim to capture potential developments in trading relations between the two countries.

Simulation 1: 20% reduction in Brazilian exports to Argentina

The first simulation imposes a 20% reduction in Brazil's exports to Argentina. This scenario reflects the recession in the Argentine economy, and its direct repercussions on external demand from that country. The percentage fall stipulated was based on estimates made in the specialist press. As the impacts on the Brazilian economy are measured relative to this shock, its precise magnitude is unimportant. Analysis of the results focuses on the sectors and regions of the Brazilian economy that are relatively most affected.

Simulation 2: Liberalization of Brazil-Argentina trade flows

One of the measures discussed for responding to the Argentine crisis involves enhancing trade openness in Mercosur, in order to stimulate intra-bloc trade and help economic activity in Argentina to recover. Accordingly, the second simulation estimates the impact of full Brazil-Argentina trade liberalization, with abolition of all import tariffs on bilateral trade between the two countries. The EFES-ARG model simulates this by abolishing tariffs on imports from Argentina and imposing subsidies on Brazilian exports to that country such that a zero-tariff-equivalent reduction in their prices could be implemented. The latter seeks to capture the improved access for Brazilian exports to the Argentine market, as a result of tariffreduction. The results obtained from the simulations capture not only the macroeconomic impact on the Brazilian economy, but especially its sectoral and spatial implications.

Simulation 3: Liberalization of the automotive sector

As referred in section II, one of the key areas in Brazil-Argentina trade is the automotive sector. The sectoral regime in Brazil aims, among other things, to regulate trade flows between the two countries in the automotive product chain. Given the importance of that sector in the structure of the Brazilian economy, a specific simulation was carried out, imposing trade openness between Brazil and Argentina in the automotive sector alone. As this represents a subset of the shocks generated by the full openness simulation, its results can be viewed comparatively. In other words, the relative weight of automotive sector openness in the impact of full liberalization between Brazil and Argentina can be directly observed.

Simulation 4: Currency devaluation in Argentina

A final simulation was carried out to project the impact on the Brazilian economy of exchange-rate devaluation in Argentina. Currency devaluations are one of the most characteristic features of balance of payments crises, and a movement in this direction can already be discerned in Argentina. Insofar as exchange-rate devaluations directly make Argentine exports more competitive, a shiftimpact on Brazilian imports and production can be expected. This is simulated in the EFES-ARG model, via a shiftin the FOB price of Argentine imports in the Brazilian market, equivalent to the 20% devaluation of the Argentine peso against the Brazilian real. This scenario assumes the Brazilian monetary authorities do not retaliate.

5.2. Results

Table 4 shows the results of the four simulations for selected macroeconomic variables. The sectoral impacts of each simulation are presented in table 5, through a breakdown of GDP components. Table 6 then shows the regional impacts, in terms of variations in activity level. Table 7 summarizes the spatial impacts by analyzing the effect on the Williamson coefficient of variation, which measures regional inequality. The results of each simulation are discussed in turn below.

Simulation 1: 20% reduction in Brazilian exports to Argentina

This scenario implies a 0.105% fall in Brazil's real GDP. Although small, the impact is not negligible, bearing in mind the share of exports to Argentina in Brazil's total external trade. This drop in activity level entails a reduction in imports, although less than the fall in exports, probably resulting in a marginal trade deficit (table 4).

Given the structure of Brazil's exports to Argentina, the industrial sector is most affected, especially in branches with high technological content, such as machinery and transport equipment (table 5). This sectoral concentration, together with interstate trade relations, results in a negative impact concentrated in the states of the south and southeast (especially Minas and São Paulo), in addition to Bahia and Amazonas (table 6).

Simulation 2: Liberalization of Brazil-Argentina trade flows

Full trade liberalization between the two largest Mercosur partners boosts real GDP growth. In this scenario, exports expand by nearly three times the rise in imports, which suggests the potential for a marginal trade surplus (table 4). This result naturally depends on the import tariffestimate used.

As in the previous simulation, industry is the sector most affected, albeit positively this time. The machinery and transport equipment segments benefit most in this scenario. In addition, the rise in other branches of industry, and in service sectors (such as transport and communication), implies a major intra-industry and inter-sectoral impact (table 5). The spatial distribution of these positive impacts (table 6) indicates greater benefit for states with high levels of exports to Argentina, such as São Paulo, Amazonas and those of the southern region.

Simulation 3: Liberalization of the automotive sector

This simulation is a subset of the full openness simulation, so its results are qualitatively similar, including positive GDP growth and an improvement in the marginal trade balance. An interesting aspect of this exercise is its evaluation of the relative importance of automotive-sector openness in the framework of full Brazil-Argentina trade liberalization. The results given in table 4 show that about 30% of the impact of full Brazil-Argentina trade openness is provided by liberalization in the automotive sector. Table 5 suggests even greater dependence in the case of industry: nearly 44% of the positive impact of openness stems from the impact of liberalization in the automotive sector. The importance of liberalization in this sector can also be seen in the growth of directly related sectors, such as metallurgy, or indirectly related ones, such as construction, commerce, communications and financial institutions.

The regional effect shown in table 6 reveals that the impact is concentrated in the main producer states (São Paulo, Paraná and Minas Gerais), in the production of both automobiles and autoparts. The positive impact in Amazonas is explained partly by exports of electronics and autoparts from the duty free zone, and the high share of electronics shipped as inputs in producer states.

Simulation 4: Currency devaluation in Argentina

The simulation of exchange-rate devaluation in Argentina produces a negative impact on real GDP in Brazil, tending towards a marginal trade deficit, with exports growing nearly three times less than imports (table 4). This scenario entails a major negative impact on agriculture-livestock and industry, especially extractive activities and nonmetallic minerals (table 5). These are probably the sectors most affected by the increasing competitiveness of Argentine products in the Brazilian market, with potentially significant input substitution effects.

The spatial distribution of the results is shown in table 6. There is a major negative effect on states with strong agricultural and agricultural-related production, such as Santa Catarina, Rio Grande do Sul, Mato Grosso and Mato Grosso do Sul. The trend towards input substitution has major implications in states such as São Paulo, Bahia, Rio de Janeiro and Amazonas.

As the results suggest, the role of interstate trade in state economies needs to be highlighted. Regional interactions need to be studied to gain a better understanding of how regional economies are affected, in international markets and domestic ones, since for the smaller economies, in particular, the performance of the more developed regions is crucial. An inspection of table 8, which contains estimates of interstate and international export coefficients for all federal units of the country, reveals several important features of the Brazilian regional system. In all states to a greater or lesser extent, interstate sales outstrip exports abroad. In general, interstate flows are relatively more important for the less developed states8.

Apart from this, when one considers the Brazilian states' main trading partners, together with the degree of openness (exports plus imports, divided by GDP) vis-àvis specific partners, whether states or countries, the importance of interstate flows, mainly with São Paulo, becomes even clearer (table 9). If Brazil's states were independent countries, willing to grant most favored nation (MFN) status to some of their partners, the «countries» listed in table 9 would probably be the potential beneficiaries. Little attention would be paid to our Mercosur partner except in Espírito Santo and states in the southern region, where Argentina is one of the 10 main trading partners.

6. Conclusions

This analysis of the short-run regional aspects of Brazilian commercial policy, focusing on economic integration and bilateral relations with Argentina, reveals a trend towards concentration of the level of economic activity in the states of the Brazilian south and southeast. We can summarize this findings as follows:

a) For simulation 1 the states of south and southeast (mainly Minas Gerais and São Paulo), in addition to Bahia and Amazonas concentrated the negative impact;

b) For simulation 2 the positive aspects are at São Paulo, Amazonas and at those states located at the southern region.

c) For simulation 3 there is also a positive impact at São Paulo, Paraná and Minas Gerais.

d) For simulation 4 the major negative effects are at Santa Catarina, Rio Grande do Sul, Minas Gerais and Mato Grosso do Sul. The results draw attention to a phenomenon that has permeated the debate on the regional issue, namely the role of trade as an engine of growth.

An inspection of table 8, enables us to conclude that in all states to a greater or lesser extent, interstate sales outstrip exports abroad. In general, interstate flows are relatively more important for the less developed states.

It is necessary to emphasize that the importance of trade between the Brazilian states and Argentina is more concentrated in states in the southern region of Brazil. For the other states the results shows that interstate trade is more important.

The estimates reveal the importance of interstate trade flows in the states' economies. It is therefore necessary to make a more in-depth analysis of trade flows between the Brazilian states, potentially leading to generalizations regarding the type of trade involved, changes in its composition through time as the Brazilian economy develops, and the implications of these structural differences in the coordination and implementation of development policies.

Thus, the future for many regions of Brazil depends more on their linkages with other domestic markets than on their performance in international ones. So, there is room for public policy action, through interventions to modernize the country's transport network, improve integration between producer and consumer markets and thereby maximize the effects of Brazil's trade policy strategies.

Footnote

* This paper presents the results of the research project «Regional Aspects of Brazil's Trade Policy», financed by the Inter-American Development Bank, under grant # 1394/2002.

1 The authors gratefully acknowledge assistance provided by William Thomas in compiling the data used in this section.

2 Until January 1999, exchange-rate fluctuations were managed by the Central Bank in a currency band system allowing for variations of between 7 and 8% per year.

3 For a more detailed analysis of Brazil's trade balance in 1997-2000, see Piccinini and Puga [2001].

4 Allied to differences in productive structures, soil and climate differences stimulate specialization in the production of certain goods and services, resulting in different modes of integration in the intra-bloc trading process from country to country.

5 The basic structure of the model is presented in Annex 1.

6 This section is based on Haddad et al. (2002).

7 The structure of this matrix was studied in Domingues et al. (2002).

8 Exceptions include the states of Amapá, Maranhão and Pará, which have transport and communications systems predominately aimed at transporting mining exports.

References

References

Baer, W.; Haddad, E. and Hewings, G. J. D. (1998): «The Regional Impact of Neo-Liberal Policies in Brazil», Economia Aplicada, vol. 2, n.o 2.

Devlin, R. and French-Davis, R. (1998): «Towards an Evaluation of Regional Integration in Latin America in the 1990's», INTAL-ITD Working Paper # 2. Buenos Aires: BID/INTAL. Confaz (1999): Balança comercial interestadual de 1997. Brasília: Ministério da Fazenda.

Dixon, P.B.. and Parmenter, B.R. (1996): «Computable General Equilibrium Modelling for Policy Analysis and Forecasting», in Amman, H. M.; Kendrick, D. A. and Rust, J. (eds.), Handbook of Computational Economics, vol. 1, pp. 3-85. Elsevier, Amsterdam.

Dixon, P.B.; Parmenter, B.R.; Sutton, J. and Vincent, D. P. (1982): ORANI: A Multisectoral Model of The Australian Economy. North-Holland, Amsterdam.

Domingues, E.P. and Haddad, E. (2005): «A. Sensitivity analysis in computable general equilibrium models: An Application for the Regional Effects of the Free Trade Area of the Americas (FTAA)». Brazilian Review of Econometrics, vol. 25, n.o 1.

Domingues, E. P.; Hewings, G.J.D.; Haddad, E.A. and Perobelli, F. (2002): «Structural changes in the Brazilian interregional economic system, 1985-1997: holistic matrix interpretation», Australasian Journal of Regional Studies, vol. 8, n.o 1.

Domingues, E. P. and Lemos, M. B. (2004): «Regional impacts of trade liberalization strategies in Brazil », in European Regional Science Conference. Porto, Portugal.

Haddad, E.A. y Domingues, E. (2001): «EFES - Um Modelo Aplicado de Equilíbrio Geral para a Economia Brasileira: Projeções Setoriais para 1999-2004», Estudos Econômicos, vol. 31, n.o 1.

Haddad, E.A.; Domingues, E. and Perobelli, F. S. (2002): «Regional Effects of Economic Integration: The Case of Brazil», Journal of Policy Modeling, vol. 24.

Haddad, E.A., Hewings, G.J.D. and Sonis, M. (1999): «Trade and Interdependence in the Economic Growth Process: A Multiplier Analysis for Latin America», Economia Aplicada, vol. 3, n.° 2.

Harrison, W.J. and Pearson, K.R. (1994): «Computing Solutions for Large General Equilibrium Models Using GEMPACK», Preliminary Working Paper, n.° IP-64, IMPACT Project, Monash University, Clayton.

Harrison, W.J. and Pearson, K.R. (1996): «An Introduction to GEMPACK», GEMPACK User Documentation GPD-1, IMPACT Project and KPSOFT, 3rd edition. September, 1996.

Ibge (1997): Sistema de contas nacionais: Brasil. Rio de Janeiro: Fundação IBGE.

Mendes, C.C. (1997): «Efeitos do Mercosul no Brasil: Uma visão setorial e locacional do comércio», Discussion Paper n.° 510. Brasílea: IPEA. August, 1997.

MDIC. Ministry of Development, Industry and International Trade.

Piccinini, M.S. and Puga, F.P. (2001): «A balança Comercial Brasileira: Desempenho no período 1997/2000», Discussion Paper 90. Rio de Janeiro: BNDES. September, 2001.

Appendix

(ProQuest: Appendix omitted.)