Brazil and China: business cycles and international trade [1] XVI

advertisement

Brazil and China: business cycles and international trade 1

XVI ENCONTRO DE ECONOMIA DA REGIÃO SUL

Área 5: Economia Internacional

Julimar da Silva Bichara – Universidad Autónoma de Madrid (Spain). E-mail: julimar.dasilva@uam.es

Sandro Eduardo Monsueto da Silva – FACE/UFG (Brazil). E-mail: monsueto@face.ufg.br

André Moreira Cunha – UFRGS and CNPq (Brazil). E-mail: andre.cunha@ufrgs.br

Marcos Tadeu Caputi Lélis –Unisinos and APEX-Brasil (Brazil). E-mail: mcaputi@uol.com.br

.

Abstract: The recent and expressive economic interaction between Latin America and Asia, particularly between Brazil and China, has attracted the attention of the academic world that seeks to understand the effects of this approach in terms of business cycles convergence, economic structure and development trajectory. This paper contributes to this debate presenting fresh evidence about the type and quality of this relationship. Our results reveal a robust relationship between business cycle dynamics and trade. China’s pattern of relationship shows evidence of productive specialisations tendency as a result of international trade. Brazil, however, presents a completely different relationship. We explore some normative implications of our results and future research possibilities as well.

Key Words: Business cycles convergence; Trade intensity; Brazil; China

JEL: F1; F15

Resumo: A crescente interação entre as economias latino-americanas e asiáticas, particularmente a China, tem estimulado o interesse dos pesquisadores em verificar as conexões potencialmente existentes entre a sincronização dos ciclos econômicos e a intensidade das relações comerciais. O presente artigo procura contribuir nessa linha de pesquisas e oferece evidências sobre a intensidade e qualidade desta relação. Enfatiza-se a divergência de resultados entre Brasil e China, cujas implicações normativas são também são exploradas.

Palavras-chave: ciclos de negócios; intensidade do comércio; Brasil; China.

1 This research was made possibe by the support of BNDES/ANPEC through the Programa de Fomento à Pesquisa em

Desenvolvimento Econômico – edital 2011. The opinions expressed in this research are of exclusive responsibility of its authors.

Business Cycles Convergence and International Trade: Brazil and China in a changing world

Introduction

The rise of China to the condition of great power, reinforcing the displacement process of significant share of the world’s production, trade and investment to the Asia-Pacific region, is one of the main issues in the recent evolution of world economy. In 2011, the Asia-Pacific countries accounted for 55% of world population, 35% of world income and 30% of world exports. The

Asian dynamism in the globalization period inaugurated in the last quarter of the twentieth century can be represented in the following way: if in 1980 Latin America had 11% of the world’s GDP measured in purchasing power parity, Asia (excluding Japan) accounted for 9%. Three decades later, Latin America accounted for 8.5% of world’s GDP, while Asia had risen up to 28%. While

Asian countries grew at an average rate of 7% yearly, Latin-American and African countries experienced rather lower rates, around 2% to 3% 2 . From the FDI’s absorption optic, the Asian countries (excluding Hong Kong) more than doubled their share between 1980 and 2010. In 1980,

Asian countries (excluding Japan) accounted for 4.4% of the value added produced by the manufacturing sector in global scale, well below the share of Latin-American countries, which was of 6.7%. In 2010, however, the Asian share rose up to 27.2% while the Latin-American share decreased to 5% (Unctad, 2012). The more developed economies (G7) and the rest of the world also experienced a decrease in their relative importance to manufactured production in this period, respectively: from 61% to 47% and from 28% to 21% 3 .

The globalization era implied on the incorporation of countries that previously had limited financial and trade links with the rest of the world. Countries of the former soviet bloc and other peripheral countries became, simultaneously, production sites and markets of consumer goods, services, and capital goods. In this context, Asian economies are among the ones that most emphasized the production and exportation of manufactured goods, which tends to contribute to their superior performance on GDP growth and productivity (Palma, 2011; Rodrik and McMillan,

2011).

These modifications on trade and production have had effects on the trajectory of business cycles between economies (Berge, 2012), in a dynamic where emerging economies achieve prominence (Canuto and Giugale, 2010; Timmer et al ., 2011. To the purposes of this article, it is noteworthy that the recent literature points to the growing interaction between Latin-American and

Asian economies, particularly with China (Calderón, 2008; Cesa-Bianchi et al . 2011; Cepal, 2011;

IADB, 2012; Rosales e Kuwayama, 2012; Lélis, Cunha e Lima, 2012). Accordingly, it is important to understand the evolution of the Brazilian business cycles and its main partners, including China, with the objective of verifying whether there is a tendency of growing cyclic convergence and whether trade is a relevant variable to explain this phenomenon, such as the pointed literature suggests. Furthermore, the political implications of that tied trade link will be evaluated in order to assess the political risks and alternatives to the Brazilian sustainable development path.

In order to achieve this objective, we took as references the methodology done by Frenkel and Rose (1998), and Calderón (2008). Following this authors we estimate, firstly, the correlation of business cycles through correlation indicators, and, then, their correlation with trade intensity for a

50 countries sample. Afterwards, we evaluate the specific cases of Brazil and China. The analysis of the business cycles is in the years from 1960 to 2010 and the trade model and cycle is estimated for the years from 1995 to 2010. Based on our evidence we explore some normative implications, evaluating alternatives of trade policies.

The results of this article contribute to the understanding of the relations between business cycle and trade and the consequences in terms of trade policy, particularly in a time of increasing

2 IMF (2012).

3 Data base source: United Nations National Accounts Main Aggregates Database

(http://unstats.un.org/unsd/snaama/introduction.asp, last accesses on March 03, 2012.). G7 – USA, UK, Japan, Germany, France,

Italy and Canada; Latin America - Argentina, Brazil, Chile, Colombia, Mexico, Peru e Venezuela; Asia - China, Hong Kong, South

Korea, Malaysia, Indonesia, India, Thailand, Philippines and Singapore; ROW – rest of the world.

world economy transformation. The comparison between Brazil and China provides new insights and opens the perspective for new researches of the growing relation between these economies.

This paper is structured in seven sections, apart from this introduction. The next section brings a brief review of the theoretical and empiric literature. Subsequently, we explain the methodology, the empirical strategy and the main findings. After the result analysis, we present some political implications and final remarks.

2. Revision of the Literature: trade and business cycle convergence

The recent literature points out that, linked to the intense internationalization process, China has become a new and increasingly important trade partner for Latin American countries. Their close trade resulted in a new phenomenon for Brazil and other Latin-American economies: business cycles are converging with the Chinese business cycles (Calderón, 2008; Cesa-Bianchi et al ., 2011;

IADB, 2012). This synchronization may be explained by trade and is related to a pattern of relations in which Latin American countries tend to specialize in the production and exportation of natural resources.

The researches that assess the relation between international trade and economic growth have gained more room in the international economy literature after the increase of intensity of international integration. One of the significant reasons of this increasing growth of interest on the topic is the increase of the international trade flows in the last few decades, at least until the recent financial crisis, and its consequences in terms of business cycle synchronization and integration stability. This literature, formulated during the peak of the globalization process and, therefore, before the financial crisis, pointed out that the main determinants of the international trade are: (i) the proliferation and deepening of regional integration agreements; (ii) the structural reforms process, such as the trade opening through the reduction of tariff and non-tariff barriers to trade;

(iii) the rise of emerging economies, particularly of the so-called BRIC (Brazil, Russia, India and

China) as great traders; (iv) the increase in demand of consumer goods in developed countries, particularly in the United States, between 1980 and 2000; and (v) the trade promotion efforts of some emerging countries, such as Brazil, due to the signature of new trade agreements and to the performance in WTO.

The expansion in trade flows would increase the channels of transmission of economic impulses among countries, which would generate a higher synchrony of business cycles. For instance, Frankel and Rose (1998) provide empiric evidences that higher trade intensity raises the business cycles’ correlation among developed countries. Moreover, Calderón et al. (2007), Shing and Wang (2004) and Rana (2007) broaden the analysis to developing countries. According to this literature, however, the effect of trade over the cycles’ correlation is not given; it depends on the dynamism of the trade flows, particularly on the kind of trade and the level of industrial development.

This debate has its sources in the Theory of Optimum Currency Area (TAMO) which has

Peter Kenen (1969) as one of its precursors, along with Mundell (1961) and McKinnon (1963), in which they recover its importance, in the 1990s, with the analysis of the costs and benefits of

European Monetary Union (EMU). In this debate, Eichengreen (1992), Krugman (1993) and

Frankel and Rose (1998) introduced their hypothesis. At the bottom line, this debate is centred on the estimation of the relationship among integration, industrial specialisation, business cycles synchronisation and integration costs. If the integration, which implies in more intensive trade, leads to specialisation, the cycles are more idiosyncratic, increasing the costs of monetary union, i.e., of losing autonomy in monetary policy.

Kenen (1969) argues that the industrial structure of the countries is a fundamental factor to the success of the Monetary Union, because similar industrial structures will produce business cycle synchronization. On the other hand, industrial specialisation leads to business cycle divergence and increases the costs of the Monetary Union. Eichengreen (1992) and Krugman (1993) add to that perspective the main findings of Ricardian theory of comparative advantage, where (i) each country exports the good in which it has comparative advantages; and (ii) each country to expand its

production of the good it exports, with labour being reallocated to it from the import-competing industry.

Krugman (1993), using the Ricardian international trade theory, argues that the integration, through the trade flow, will produce an industrial specialisation in European countries enrolled in

EMU, explained by the comparative advantages and transport cost. Krugman develops a quite simple theory to show the industrial specialisation motivated by integration. In that debate,

Krugman compares the United States with the European Union (EU), in order to show that the

United States had a higher level of regional specialisation than the EU due to free trade and factor mobility. Finally, Krugman (1993) concludes that “if Europe moves toward US levels of regional specialization and factor mobility, disparities in economic growth rates among countries and regions can be expected to increase substantially” (p. 254).

The same argument has been used by Eichengreen (1992). He argues that “because the

American market is so integrated, U.S. regions specialize more completely than EC countries in manufactures in which they have comparative advantage.” (p .14). Moreover, Eichengreen (1992) concludes regarding to the EU that “regional specialization will increase with the completion of

1992 Program, amplifying region-specific shocks.” (p.15). In terms of that debate, it implies that the integration (international trade) will produce industrial specialisation, which will produce business cycle dissociation. In the case of EU, it also means higher cost of EMU, creating conflicts between objectives of economic policy, especially in terms of fiscal and monetary policy.

In this context, Shin and Wang (2004) pointed out that the channels through which the trade impulses are propagated to the economic growth are: (i) inter-industry trade; (ii) intra-industry trade; (iii) demand spillovers; and (iv) political coordination. Therefore, if the kind of trade shock that is predominant among countries is derived from a demand shock, an increase in the business cycles’ correlation may be expected. A positive shock in a given country may provoke an increase in the demand of importations and, consequently, encompass a demand spillover effect, producing mutual economic growth. The extent of this impact depends on the level of trade integration of the economies or on the economical interdependency levels, or even on the trade intensity levels.

However, if the dominant shock is specific-industry, i.e., occurs exclusively in a given sector of economic activity, the effect on the business cycle may be negative, as it may lead to a trade specialization and, therefore, to a productive specialization. The derived trade relation will be of the Heckscher-Ohlin or Ricardian kind, since the trade will be inter-industry, as a result of the production specialization. Therefore, the inter-industrial trade and economic cycle’s convergence relation may be negative.

On the other hand, if trade is predominantly intra-industry, such as in developed countries, the effect of trade growth on the business cycle’s correlation does not necessarily take to asymmetric effects, since the specialization pattern is industrial; on the contrary, it may lead to an increase in the business cycle’s convergence. Therefore, as Frankel and Rose (1998) pointed out, the effect of trade over business cycle’s convergence is ambiguous and depends on the kind of trade shock and on the structure of the trade practiced by countries. Therefore, only an empirical analysis may seek clearer evidences on the subjacent relations of this phenomenon.

Frankel and Rose (1998) developed an innovative econometric strategy to evaluate the relationship between trade intensity and business cycle, using a sample of 21 industrialized countries. They have a hypothesis that this relation is endogenous, in a way that there would be a variable capable of determining, at the same time, trade intensity and cycles, such as monetary union. Therefore, the econometric analysis should consider this inconvenient in order to avoid using biased and inefficient estimators. In order to outline this potential problem, they use an instrumental variable (IV) model; the instruments are from the traditional gravitational equation, i.e., distance, border and language. The results of their estimations indicate a positive relation between trade intensity and business cycle.

Calderón et al . (2007), follow the previous authors, widens the geographical scope to 147 countries, including developed and developing countries, with the objective of assessing whether there are or not differences in the relation between trade and business cycles in those economies. As

expected, due to the difference in trade and specialization economic structure, the relation is heterogeneous, being more intense among developed countries than among developing countries or even between developed and developing countries.

Shin and Wang (2004) and Rana (2007) emphasized the Asian experience. The former evaluates the trade intensity effect over the correlation of business cycles of 10 Asian economies during 1976 to 1997. These authors introduce a methodological modification since they consider the instrumental variables used in previous researches are not adequate to explain the inter-industry trade. Nevertheless, they estimate that the intra-industry trade is the main determinant of the business cycle’s convergence and that inter-industry trade has a lower business cycle’s correlation, as it is expected by the theory. However, their research suggests that an increase in trade per se does not promote a higher economic convergence, as the effect of the demand shock may also be negative – a conclusion that is against to what the theory predicts.

Rana (2007) replicates Shing and Wang (2004) considering the IV model, as proposed by

Frankel and Rose (1998), but using data from 1993 to 2004. The results are equivalent to those of

Shing and Wang (2004), strengthening the hypothesis that intra-industry trade is what promotes a higher cycles’ convergence in the Asian countries’ case and that the growth of trade per se does not promote a higher symmetry on the evolution of economies.

The recent literature is converging to the perception that the economic convergence: (i) is reducing in the past few decades and (ii) depends on the development degree of countries. Imbs and

Wacziarg (2003) show that the business cycle’s correlation is related to the pattern of industrial specialisation. Kose et al.

(2008) and Calderón et al. (2007) suggest that this relation varies significantly according to the industrial development level, being lower among least developed countries. And, together with Blonigen (2012), they show that the effect of international trade over business cycle has diminished along with time, even though the relations continue statistically positive and significant. Some explanations for such a loss of power are derived from the financial flows’ importance growth and its determinant effect of international business cycle 4 .

3. Methodological Strategy

In order to achieve the proposed objective of analysing the relation between trade and economic growth in the Brazilian and Chinese economies, and that of their main trade partners, we will use the methodology proposed by Frankel and Rose (1998). Insofar, we selected a sample of 50 countries (Appendix) which represent the main trade partners. Such methodology involves: (i) the calculation of the business cycles correlation; and (ii) the econometric model to estimate the bilateral trade effects on the business cycles. In order to calculate the business cycle’s correlation we used two alternative techniques, the Hodrick-Prescott (HP) and the Baxter and King (BK)

5

. The

GDP data, in constant prices on local currency, are from the World Bank (2012), for the 1960-2010, in natural logarithms. Alternatively, we used the industrial product. The cycle’s correlation, for each methodology, was obtained from successive 15 year time-frames (1960 – 1974; 1961 – 1971; ... ;

1996 – 2010), building a correlation matrix among all analysed countries, along 37 periods of 15 years each.

In order to calculate the bilateral trade and the trade intensity, for total trade and manufacturing trade, we used UN Comtrade data, for each one of the 50 considered countries.

Thereby, it was possible to build a 50 X 49 trade intensity indicators matrix for each of the analysed years, from 1995 to 2010, period in which the database provides more homogeneous information for the sample countries.

The bilateral trade intensity between a reference country i and another country j, in time t, was calculated as from two proxies proposed by Frankel and Rose (1998). The first is based exclusively in the international trade data:

4 Berge (2012) goes on the opposite direction. He concludes that “…degree to which business cycles are now internationally synchronized has increased significantly over the past 20 years, due in part to closer economic ties between countries.” (p.30)

5 Baxter and King (1999).

TT ijt

= (X ijt

+ M ijt

) / (X it

+ X jt

+ M it

+ M jt

) (1)

Where: Xijt refer to the exports from country i to country j, during period t; Xit and Mit refer to total exports and imports from country i. Therefore, this trade intensity indicator is normalised by the total trade.

The second proxy normalised by the total bilateral trade through both the countries’ GDP:

TY ijt

= (X ijt

+ M ijt

) / (Y it

+ Y jt

) (2)

Where: Yit is the nominal GDP level of country i in period t, with data extracted from WDI-World

Bank. Additionally, to the second data set of bilateral trade (UN Comtrade), these indicators are adapted to consider GDP and manufactured products trade.

Afterwards, we estimate the relation between business cycles convergence and trade intensity using the model proposed by Frankel and Rose (1998), i.e.:

Corr (v,s) ijt

= á + â IT ijt

+ å ijt

(3)

Where: Corr (v, s) ijt

denotes the business cycles’ correlation between country i and j, during time t for a proxy of economic activity v, which in our case, corresponds to real GDP and industrial production calculated by filters, which indicates one of the alternative methodologies of cycles estimation, of Hodrick-Prescott (HP) or Baxter-King (BK). IT refers to trade intensity, which may de normalised by total trade and GDP as equations (1) or (2). Finally, á and â are the regression coefficients to be estimated. We estimate 8 different versions of the model derived from the two definitions of business cycle, from the two economic activity indicators and from the two measures of trade intensity. Additionally, these 8 versions are also estimated for the sample total, and subsequently for China and Brazil, considering the trade flow with other countries of the sample, accounting for a total of 24 estimations.

According to the theory, the value and signal of â will reveal how trade affects business cycles’ convergence among all the sample countries, in the case of the first model, and for Brazil and China and their main trade partners. If â is negative, it indicates that the industrial specialization effect dominates in the trade and, therefore, there is higher independence between the economies.

On the other hand, of â is positive, the intra-industry effect would be dominant in the trade relations, which would generate a higher business cycle’s correlation between countries. The value of â, as expected, reflects the importance of this phenomenon.

In order to avoid problems derived from a biased estimation, due to the probable endogenous character between economic activity and trade, we use an IV model, such as Frankel and Rose (1998), Calderón et al

. (2007) and Calderón (2008). Therefore, it is necessary to find the determinants of trade intensity to use them as instrumental variables. This work reproduces the strategy used by Frankel and Rose (1998) and uses bilateral trade intensity determinants, a simple gravitational equation, with the independent variables being: distance, language and border.

Thereby, before presenting the estimation results, we present a critical appraisement on the gravitational model, on its theoretical basis and on the main evidences that exist on its literature.

4. The Gravitational Model

The gravitational model takes into account the basic idea that the trade volume between two countries is directly proportional to the product of its size and inversely proportional to the distance between them, forming the analogy with the gravitational attraction models between two masses. Its use owes basically to the researches done by Tinbergen (1962), Poyhonen (1963) and Linnermann

(1966), who delimitated the basic variables for the estimation of the model, using factors that indicate the potential supply and demand as a proxy of the economy’s size, or as elements that explain the attraction between countries, and using literally the distance to explain the resistance to trade, or as approximation of transport costs. In its simpler specification, the bilateral trade flow between two countries (import, export and both) may be determined through the equation:

LNX ij

 

1

 

2

LN ( Y i

* Y j

)

 

3

LDist ij

  ij

(4)

Where: LNX ij is the bilateral trade logarithm of export from country i to country j; Y i

and Y j

are respectively the GDP of the countries i and j and LDist ij

is the distance logarithm between both countries. In general, we expect a positive signal for â

2

and negative for â

3.

Differently from most of the international trade empiric models, the use of gravitational equations for the analysis of trade flows precedes a theoretical formalization, without a clear correspondence among the main theoretical models and the variables used in their equations.

However, many researchers proved that it is possible to derive an equation based on pre-existing models. Dearforff (1998), for instance, shows that the variables and the signals of the gravitational equation may be deducted out of the Hecksher-Ohlin (H-O) model, without the product differentiation hypothesis. Eaton and Kortum (2002) developed an approach based on the Ricardian model of trade and homogeneous goods. Krugman (1980) derives the gravitational equation of a monopolistic competition model with economies of scale, product differentiation and with transport costs. In this model, when two economies decide to trade goods, the increasing returns generate gains with bilateral trade, even though both would have the same taste, technology or factor endowments.

From the Krugman model (1980) emerges the so-called home market effect, which says that large countries tend to be exporters if their domestic market attracts new firms. According to this theoretical approach, what determines the export structure are the dimensions of the domestic markets; in other words, countries do not export only goods for which they have comparative advantages, but also those ones that the domestic market allows to produce in a larger quantity due to the increasing returns to scale in imperfect competition conditions. Jordán and Parré (2006) suggest that the larger countries of the continent (Argentina, Brazil and Mexico) have a higher intensity of economies of scale. They conclude that larger dimension markets tolerate a disproportional incoming of companies, which would lead to more intense competition and to the reduction of price levels.

Despite the diversity of theories that may be used to interpret the gravitational model, the results found by the empiric researches show a consensus on the existence of a positive correlation between the bilateral trade and the size of the partner economy and a negative correlation between the trade and the distance that separates both countries. A deeper revision of the theoretical approaches developed about the gravitational equation may be seen on Harrigan (2002).

Apart from the classic variables defined on the first empiric studies (GDP, per capita GDP, and distance between markets), the subsequent analysis have used other variables in order to assess the attraction or repulsion factors of international trade, such as representative dummies and common borders, volume of tariff barriers to trade, exchange rate, and the existence or not of trade agreements (Kimura and Lee, 2004; WTO, 2013).

As already stated, the full sample has 50 countries and the data from 1960 to 2010 is for business cycle and that from 1995 to 2010 is for bilateral trade. However, the data both for trade and GDP are incomplete for some countries, especially the information of manufacturing GDP and trade. Therefore, the number of observations may vary between models. The models were applied in robust versions, using the Stata 11

6

. The next sections present the results of the gravitational model’s initial estimations, seeking evidences of the validity of the used instruments, and, subsequently, of the international trade effect on the business cycles’ correlation of the countries in the sample.

5. Results of the Gravitational Model

This subsection presents the results of three sets of gravitational models

7

: (i) to the whole sample; (ii) to Brazil; and (iii) to China. In the cases (ii) and (iii) we consider the relation of the analysed country and its relations with its main partners. The results are shown on the following tables

8

. We made estimations through the ordinary least squares (OLS) model and through the

6 The full outputs are not shown due to lack of space. They may be requested to the authors, if needed.

7 We use the CEPII dataset for the traditional gravity variables. Available at: .cepii.fr/CEPII/en/bdd_modele/presentation.asp?id=8

8 Due to lack of space we did not show here the full data of Stata. They may be requested to the authors, if needed.

random effects (RE) model in order to analyse the validity of the models

9

. In a general way, the signals and significances do not have noteworthy changes in the two estimation models, showing the consistency of the gravitational model.

The results evidence some differences among the three versions of the estimated gravitational models. Firstly, we should point out that the model used for the total sample shows that the estimated coefficients are robust, statistically significant and show the expected signal, especially in the OLS model. In general, therefore, neighbour countries tend to develop a higher degree of trade than with the other countries. The same conclusion was reached in the “language” variable, in other words, a common language between countries is also favourable for trade. Finally, the “distance” variable, which also shows a coincidental result with the theory, shows that the higher the distance between two countries, the higher the transport cost and the lower the expected trade intensity is. This variable is more important for explaining trade, as it is possible to observe in

Table 1, through the value of the estimated coefficients; moreover, its effect is even higher in the sum of the positive effects derived from a common border and language. This reveals that distance, as expected by the theory, introduces significant comparative disadvantages due to the transport costs.

Table 1 – Total of the Sample: Gravitational model

Border

Distance

Language

Constant

TT

0,612*

(0,03)

-1,124*

(0,01)

0,136*

(0,03)

2,995*

TY

OLS Model

0,434*

(0,03)

-1,140*

(0,01)

0,062**

(0,02)

-4,417*

TTM

0,674*

(0,04)

-1,210*

(0,01)

0,156*

(0,03)

3,447*

TYM

0,308*

(0,04)

-1,294*

(0,01)

0,055***

(0,03)

-1,908*

TT

0,596*

(0,13)

-1,132*

(0,04)

0,123

(0,10)

3,075*

Random Effects Model

TY

0,423*

(0,12)

-1,149*

(0,04)

0,046

(0,10)

-4,327*

TTM

0,657*

(0,14)

-1,219*

(0,04)

0,152

(0,10)

3,529*

TYM

0,297**

(0,14)

-1,306*

(0,04)

0,058

(0,10)

-1,811*

Obs.

R

F

2

X 2

Adjusted

Prob>F/X 2

(0,09)

38322

0,2528

7115,32

0,00

(0,08)

38316

0,2527

6998,78

0,00

(0,09)

38230

0,2596

7519,59

0,00

(0,09)

38020

0,2635

7057,26

0,00

(0,33)

38322

1483,95

0,00

(0,32)

38316

1498,73

0,00

(0,34)

38230

1559,98

0,00

(0,34)

38020

1609,17

0,00

Source: Research Results. Standard errors inside parenthesis. TT: Trade Intensity pondered by total trade; TY: Trade Intensity pondered by GDP;

TTM: Manufactured goods trade intensity pondered by total manufactured goods trade; TYM: Manufactures goods trade intensity pondered by manufactured goods GDP. *** p<0.10, ** p<0.05, * p<0.01

For the Brazilian case (Table 2) and for the Chinese case (Table 3) there are a few significant differences. In a general way, as in Frankel and Rose (1998), the trade intensity normalised by total trade (TT) shows better results than when it is normalised by GDP (TY), especially in the Brazilian case. The main difference is related to the language. It is expected that the same language increases trade. In the Brazilian case, however, the estimated coefficient has a negative signal, differently from what is expected in theory. The opposite happens in China, for whose estimated coefficient is positive and significant in the 8 specifications of the model.

The result for Brazil is related to the fact that there are very few countries that speak

Portuguese – in the 50 countries sample, only Portugal shares the same language as Brazil. In this way, the used specification would be equivalent to the introduction of a dummy to Portugal. The estimation reveals, therefore, that the trade relation between Brazil and Portugal is, on average, relatively smaller than that with other Brazilian trade partners.

9 It is not possible to estimate the fix effects model because the distance and language variables are constant.

Table 2 - Brazil: Gravitational Model

Border

Distance

Language

TT

0,598*

(0,15)

-0,290*

(0,09)

0,058

(0,05)

TY

OLS Model

TTM

0,549*

(0,17)

-0,086

(0,10)

-0,060

(0,08)

0,817*

(0,14)

-0,469*

(0,09)

-0,066

(0,06)

TYM

0,658*

(0,15)

-0,294*

(0,09)

-0,133***

(0,07)

TT

0,597

(0,56)

-0,291

(0,33)

0,058

(0,17)

Random Effects Model

TY TTM

0,548

(0,62)

-0,087

(0,36)

-0,059

(0,19)

0,823

(0,53)

-0,462

(0,32)

-0,068

(0,17)

TYM

0,657

(0,56)

-0,295

(0,33)

-0,133

(0,20)

Constant -2,965*

(0,81)

776

-13,013*

(0,90)

776

-1,747**

(0,79)

776

-10,111*

(0,83)

776

-2,956

(3,10)

776

-13,000*

(3,36)

776

-1,808

(3,00)

776

-10,101*

(3,06)

776 obs.

R

F

2

X 2

Adjusted

Prob>F/X 2

0,1808

52,60

0,00

0,0585

19,83

0,00

0,2445

97,15

0,00

0,1020

50,73

0,00

11,10

0,01

5,46

0,14

20,39

0,00

12,75

0,01

Source: Research Results. Standard errors inside parenthesis. TT: Trade Intensity pondered by total trade; TY: Trade Intensity pondered by GDP;

TTM: Manufactured goods trade intensity pondered by total manufactured goods trade; TYM: Manufactures goods trade intensity pondered by manufactured goods GDP. *** p<0.10, ** p<0.05, * p<0.01

Table 3 - China: Gravitational Model

Border

Distance

Language

Constant

TT

0,291**

(0,14)

-1,439*

(0,08)

0,591*

(0,07)

7,204*

TY

OLS Model

0,239

(0,16)

-1,431*

(0,08)

0,814*

(0,08)

-0,386

TTM

0,076

(0,16)

-1,590*

(0,08)

0,557*

(0,09)

8,437*

TYM

0,042

(0,18)

-1,592*

(0,08)

0,699*

(0,10)

1,985*

TT

0,284

(0,47)

-1,445*

(0,28)

0,582*

(0,21)

7,262*

Random Effects Model

TY

0,228

(0,50)

-1,440*

(0,25)

0,802*

(0,20)

-0,302

TTM

0,075

(0,54)

-1,591*

(0,28)

0,556**

(0,24)

8,449*

TYM

0,039

(0,57)

-1,595*

(0,27)

0,695*

(0,25)

2,009 obs.

R 2 Adjusted

F

X 2

Prob>F/X 2

(0,68)

776

0,3221

204,74

(0,66)

776

0,3296

203,31

(0,71)

776

0,3378

199,86

(0,72)

776

0,3240

186,71

(2,44)

776

61,44

(2,22)

776

77,80

(2,46)

776

66,26

(2,35)

776

70,38

0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00

Source: Research Results. Standard errors inside parenthesis. TT: Trade Intensity pondered by total trade; TY: Trade Intensity pondered by GDP;

TTM: Manufactured goods trade intensity pondered by total manufactured goods trade; TYM: Manufactures goods trade intensity pondered by manufactured goods GDP. *** p<0.10, ** p<0.05, * p<0.01

The coefficients estimated for the others variables of the model, i.e., distance and language, are consistent with the theory. Therefore, the gravitational models for the total sample, for Brazil and for China seem to represent a good instrument for the model that estimates the relation between economic convergence and trade intensity. The coefficients estimated for this model (β, through instrumental variable) are analysed subsequently.

6. Business Cycle and Trade: evidences to Brazil and China

This analysis is divided in two econometric exercises. We begin with the calculation of the correlation of business cycle, which afterwards will be used as a dependent variable in the estimation of the model that relates trade and business cycle. In order to calculate the business cycles’ convergence of the 50 countries sample against Brazil and China we use GDP data and industrial product data in natural log. From the Hodrick-Prescott (HP) and Baxter-King (BK) filters we calculate the business cycles and estimate the correlation coefficients for pairs of countries, in

15 year periods, estimated successively (1960 – 1974; 1961 – 1971; ... ; 1996 – 2010), totalizing a database with 37 periods. Therefore, we present 4 evolution series of the business cycles’ correlation coefficient between 1974 and 2010.

For the correlation of business cycles and trade model, represented in equation (3), eight specifications were estimated for the total sample and for the two analysed countries, the different trade intensity proxies and the kind of filter to calculate the business cycles, accounting for a total of

24 estimations that are represented on tables 4 to 6. Subsequently, the results of these two exercises are analysed, starting with the correlation of business cycles.

Out of the 50 countries in the sample, the analysis of the dynamic of business cycle may be divided in developed and emerging countries, and in this last group the Latin American countries deserve special attention. A constant point in this analysis is the decrease of economic convergence degree of Brazil and China with developed countries, which would represent a higher economic independence, particularly in the last few decades. It is noteworthy, moreover, the considerable reduction of the economic dependence of these two countries with the United States, even though that China presented a higher level of co-movement of its economy. Finally, we observe significant heterogeneity on other industrialized countries, while Brazil and China present a growing correlation level in recent years with Germany, France, Italy, United Kingdom and Japan.

The evolution of the business cycles correlation between South American countries with

Brazil and China is rather homogeneous. With Mexico, however, the results showed a very volatile evolution, both with Brazil and with China. Since the mid 1990s and the beginning of the new century, the correlation was negative and became positive in the last few years. We should still point out that the coefficients are superior for Brazil, reflecting a higher degree of interdependence with Brazil relatively comparing with China.

In the case of emerging Asian economies, on the other hand, the evolution of the correlation coefficient, in all considered cases, showed a significant growth, particularly during the late 1990s, reflecting an increasing convergence between the Brazilian and Chinese economies. The convergence with China is a little more homogeneous than with Brazil, as we should expect.

In short, some results may be highlighted on what concerns the convergence of business cycles in the Brazilian and Chinese cases with selected countries: on the one hand, a growing convergence among developing countries, in which there is a high interdependence between Brazil and Latin American countries and between China and Asian countries; on the other hand, there is significant heterogeneity with developed countries, in which there has been growing convergence in recent years with European countries and with Japan; there has been divergence or economic independence with the United States; finally, a high and sustainable correlation between Brazil and

China, only inferior to that with some Latin American countries.

For the model of business cycle and trade correlation, represented in equation (3), the eight specifications estimated are presented in tables 4 to 6, which show firstly how the model with instrumental variables is consistent and efficient and, moreover, the used instruments are statistically valid and consistent.

Table 4 – Estimated coefficients for â by VI. Trade intensity effects on the business cycle (total of the sample)

Correlation Variable

GDP

GDP

Manufacture GDP

Manufacture GDP

Filter

BK

HP

BK

HP

TT

9,503*

(0,00)

11,837*

(0,00)

TY

9,603*

(0,00)

11,889*

(0,00)

TTM

3,187*

(0,00)

3,801*

(0,00)

TTY

3,144*

(0,00)

3,717*

(0,00)

Source: Research Results. Standard errors inside parenthesis. Coefficients multiplied by 100. TT: Trade Intensity pondered by total trade; TY: Trade

Intensity pondered by GDP; TTM: Manufactured goods trade intensity pondered by total manufactured goods trade; TYM: Manufactures goods trade intensity pondered by manufactured goods GDP. BK: Baxter-King; HP:Hodrick-Prescott. *** p<0.10, ** p<0.05, * p<0.01

For the total sample of 50 countries (Table 4), βs estimated are in all cases positive and statistically significant, which implies that the higher the trade intensity, the higher the business cycles’ convergence between selected countries. Moreover, there are rather important differences in the estimations of β between the equations that use the correlation of GDP and manufacturing sector as a dependent variable. This suggests that the trade effect is higher when measured relatively to the total of the economy, than when measured in proportion to the total trade of manufacturing sector.

Moreover, the differences between the filter methodologies are minimal, confirming the consistency of both models.

The analysis for the Brazilian and Chinese cases, shows certain divergence of the trade effect on economic convergence. For Brazil (table 5), the estimated βs are positive and statistically significant in the model that uses the total trade intensity pondered by GDP. However, when we evaluate the economic convergence using manufacturing sector, the estimated coefficients are negative and even statistically significant in the case of the BK filter. In the first case the result is evidence that the Brazilian trade is intra-industry and has dynamic effects on the economic growth.

However, the second result shows the opposite, revealing that Brazilian manufacturing exports have a negative relation (BK model) or null (HP model) with the business cycles evolution of Brazilian trade partners.

Table 5 – Estimated coefficients for â by VI. Trade intensity effects on the business cycle (Brazil)

Correlation Variable

GDP

GDP

Manufacturing GDP

Manufacturing GDP

Filter

BK

HP

BK

HP

TT

11,178*

(0,03)

25,807*

(0,04)

TY

21,466*

(0,06)

46,466*

(0,08)

TTM

-9,942*

(0,03)

-2,217

(0,03)

TTY

-11,953*

(0,04)

-0,999

(0,05)

Source: Research Results. Standard errors inside parenthesis. Coefficients multiplied by 100. TT: Trade Intensity pondered by total trade; TY: Trade

Intensity pondered by GDP; TTM: Manufactured goods trade intensity pondered by total manufactured goods trade; TYM: Manufactures goods trade intensity pondered by manufactured goods GDP. BK: Baxter-King; HP:Hodrick-Prescott. *** p<0.10, ** p<0.05, * p<0.01

On the Chinese case (Table 6), the estimated βs are all negative and statistically significant, revealing that, on average, the trade does not have positive effects on the business cycle convergence of China and the main trade partner. According to the previously analysed theory, this result would be the product of inter-industry trade pattern, in which shocks are sector specific and, therefore, do not have dynamic effects over the economy. In other words, the trade intensity is not generating a business cycle convergence with China, but a productive specialization, which would implicate that economic growth dynamic of the Chinese economy would not depend on the trade flow and, more important, is not generating dynamic effects over the trade partners.

Table 6 – Estimated coefficients for â by VI. Trade intensity effects on the business cycle

(China)

Correlation Variable

GDP

GDP

Manufacturing GDP

Manufacturing GDP

Filter

BK

HP

BK

HP

TT

-7,083*

(0,01)

-8,434*

(0,02)

TY

-6,658*

(0,01)

-7,762*

(0,02)

TTM

-3,346*

(0,01)

-4,589*

(0,01)

TTY

-3,279*

(0,01)

-4,501*

(0,01)

Source: Research Results. Standard errors inside parenthesis. Coefficients multiplied by 100. TT: Trade Intensity pondered by total trade; TY: Trade

Intensity pondered by GDP; TTM: Manufactured goods trade intensity pondered by total manufactured goods trade; TYM: Manufactures goods trade intensity pondered by manufactured goods GDP. BK: Baxter-King; HP:Hodrick-Prescott. *** p<0.10, ** p<0.05, * p<0.01

The next section explores some normative implications of our results, emphasizing the

Brazilian perspective.

7. Political implications and recent trends in trade policy

Our results suggest that the Chinese trade relationship, in all cases, is inter-trade kind, and would lead, according to the theoretical and empirical literature, to a productive specialisation and, consequently to an economic dependence on its trade partners, particularly on developing economies such as Latin American countries and Brazil. Under this perspective it becomes crucial to research alternatives in trade and industrial policies to face this new international economic trend.

There is no doubt that the Chinese strategy in the international economy and its consequences over the Brazilian and Latin American economies become a big challenge to the development of these countries.

The economic literature does not offer a common point of view in terms of political and economic alternatives to deal with this challenge. It is possible to find in free trade an alternative and also in trade intervention policies. In this scenario, Helleiner (1992: 1-2), highlights the situations that theoretically claim a governmental intervention through trade policies:

1.

Economic efficiency: The principal strands of the arguments for government intervention that are based upon economic efficiency objectives are: (a) the ‘infant-industry’ argument;

(b) ‘externalities’ arguments, including those favouring the entire infant manufacturing sector, because of its purported positive effects elsewhere in the economy; and (c) the purported need to offset the negative incentives for industry created by already existing product or factor market ‘distortions’;

2.

Risk aversion and/or national security;

3.

Internal income redistribution, either in pursuit of distributional equity, in response to powerful interest groups and rent-seekers, or to meet government revenue requirements;

4.

Macroeconomic objectives, notably full employment and capacity utilization, and balance in external payments;

5.

Improvement in international terms of trade.

Moreover, the trade policy interventions would be justified by trade policy strategy arguments. In presence of externalities, imperfect competition and learning by doing, scale economies, the benefits of trade would be divided among home and foreign producers, in such situation trade policy can take on a strategic role (Helleiner, 1992; Krugman, 1986; Rodrik, 1988).

The government has tools which firms do not have, and intervention can help firm commit to strategies that would otherwise be credible. As a result, this commitment can tilt the international playing-field in favour of home firm, allowing home firm to grab a large scale of world market and profit. In certain conditions, a subsidy or tariff can shift enough additional rents to the home country actually to pay for itself, making the home country as a whole better off (Rodrik, 1993).

In this case, the direct application of the restrictive trade policy could have effects not only on the import tariffs but also on the export and production subsidies. The international literature argues that the use of subsidies (to exports or to production) is the most efficient and least costly mechanism to promote a specific sector (Rodrik, 1993). This mechanism has the advantage of being discretional, led by the government in order to achieve its industrial policy objectives. Moreover, such trade policy would influence the structure of production and consumption of tradable goods and services, and thus the level and the composition of imports and exports (Helleiner, 1992).

Furthermore, as UNCTAD (2010) suggests, in this field the WTO legislation may be rather ambiguous and subject to controversies, especially when the government uses its public expenses to stimulate the national producers or participates as a shareholder of enterprises, what actually is the case in most of the Chinese multinational companies (Maseiro and Caseiro, 2012).

In this microeconomic perspective, a significant determinant of the success of trade policy is related, according to Rodrik (1993), to the government’s capacity to act in a discretional and selective way in the definition of the industrial sectors, without being affected by third party interests. However, this point represents a clear challenge to Brazil, while, on the other hand, the

Chinese government has enough autonomy to drive its discretional power and to define in a stable and absolutist way the priority sectors; the Brazilian government, as a democratic country, has to deal with the democratic political system, subject to different lobby pressures in the definition of the economic policies, and thus has fewer autonomy to define its industrial policy

10

.

10 This is, clearly, the case of other democracies, as Rodrik (1993, 35) has highlighted about the USA policy trade definition:

“´proliferation of interest groups and diffusion of influences´ among executive agencies, ´multiple committee referrals´ in Congress, and the possibility of judicial review and overturn have made U.S. trade policies capricious and unpredictable”.

However, the traditional economic literature, even though recognizes the theoretical benefits of the strategic trade policy, distrusts the government’s intervention capacity, alleging that it is unable to act in favour of the common good, being an instrument of third party interest groups

(Grossman,1987; Krueger, 1974; Krugman 1987). In other words, it highlights the government’s incapacity to drive a successful trade policy, being influenced by strong lobbies, in a way that the trade policy ends up being a wealth channelling from the society as a whole to specific sectors that hold strong political power.

This matter is interesting in the Brazilian case because the country has winner and loser sectors in the trade relations with China. On the one hand, the agribusiness and the natural resources sector has been feeling positive impacts while, on the other hand, part of the industrial sector, mainly that of medium and small companies has been feeling negative impacts due to the growing

Sino-Brazilian trade relation. Without losing the focus on the theoretical perspective, it is also necessary to analyse the evolution of the main trends of the trade policy in the new century. From an empirical approach, a first trend that deserves attention is the growing commercial tension between developing countries and China

11

. This is a significant matter in current trade relations and, in the future, may constitute an authentic trade war, what would be nefarious for the development trajectory of Latin American countries. An UNCTAD (2010) research shows that the number of protectionist trade measures has grown recently and that a considerable part of the formal claims are against China

12

.

According to UNCTAD (2010), two main responses to the new challenge that China poses have been perceived in the international economy. On the one hand, the growing of protectionist measures – traditional tariff and non-tariff measures, which are harder to estimate and to catalog.

On the other hand, the number of protectionist strategies channelled through trade integration agreements has also grown. On what regards the so-called traditional protectionist measures, which increased especially after the crisis, the more commonly used are: the increase in import tariffs, the tightening of regulations regarding standards and certification; antidumping and safeguards measures; import licensing systems for specific products; and subsidies to encourage consumers to buy specific products

13

.

Moreover, new ways of trade protectionism strategies have started to be used by developing and developed countries, being known as “new trade restrictive measures (NTMs), such as bailing out of ailing firms and “buy-local” principles in government procurement, fell in the areas where the WTO rules provided only an ambiguous legal framework” (Unctad, 2010: xi). In this sense, it would be taken as evidence that the WTO rules are not enough to deal with the new challenges represented by the rise of new and big player, as China, and the trade relations in time of economic crisis.

As a consequence, WTO is increasingly loosing space in the international arena to regional or sectorial agreements between countries and between companies and countries. As a way to make up for trade rules that were missing under the multilateral trade framework, the number of regional or bilateral free trade agreements (FTAs) have increased over the last decade: in 2009 there were

171 FTAs and in the period between 2000 and June 2009 this number was increased by further 105 new agreements (UNCTAD, 2010).

According to the UNCTAD (2010), the recent rise in FTAs has three main determinants or objectives:

11 Many researches show that a undervalued renminbi not only causes problems to the industrial sectors intensive in work force of developed countries but also to developing countries, such as the developing countries in Asia or Brazil, what may cause growing trade tensions among developing countries (Eichengreen et. al., 2004; Greenaway et al., 2008; Jenkis and Barbosa, 2012).

12 According to the Unctad (2010), there were 120 antidumping (AD) investigations in the latter half of 2008, which was an increase of 35 over the first half of the same year. In recent years, Chinese-made products have been the major target of AD measures implemented around the world. According to the WTO, over 400 AD investigations into Chinese products were set in train from

2002 to 2008, representing 27 per cent of the total number of AD investigations worldwide. Similarly, there were 297 AD measures applied against Chinese goods, again accounting for 27 per cent of the total (Unctad, 2010). India and USA have been active in bringing in measures against Chinese-made goods since the financial crisis; India uses particularly AD measures, and USA uses countervailing duties (CVD).

13 For more details see UNCTAD (2010).

First, because sensitive sectors can be excluded, partners can be selected, and

“customization” of the contents is possible.

Second, Bergsten’s concept of “competitive liberalization”: reciprocity is the key to competitive liberalization, the idea works better in an FTA framework than in a multilateral framework. In FTAs, market opening by a partner or partners is more tangible and immediate than multilateral liberalization, especially when big markets like China and India are involved.

Third, as a way to attract foreign direct investment (FDI).

Dani Rodrik (2012), among others, has highlighted the negative consequences of this protectionist trend in international trade and suggested that the solutions should be developed within the framework of the WTO, with general measures that are a consensus among all members. This is a consequence of the last 15 years of uneven international trade, which have resulted in two clear winners: China and Germany. Both the countries have an economic policy strongly characterized by beggar-thy-neighbour , because they create benefits to its society while imposing costs to other countries. The international trade policy should, therefore, be regulated in a multilateral framework, to reduce the risks of a trade war that could encompass negative consequences to the future development. However, as it was shown before, it seems that international institutions have been unable to solve problems of this nature, especially in periods of economic crisis. Thus, this becomes the bigger challenge to WTO: to build sustainable rules that put international economic relations in a long-term equilibrium.

In this context of growing protectionist measures in international trade, of incapacity of the international institutions to offer sustainable solutions, and of strong Chinese growth as the main world exporter, achieving a greater market-share in the whole world, what could be the alternative economic policies for Brazil? The Brazilian trade policy should have as a main goal put the country in a sustainable economic development strategy, which encompasses the consolidation of a strong and competitive industrial sector. Insofar, Brazil has two alternatives: (i) macroeconomic policy, through the management of a more competitive exchange rate and (ii) microeconomic policy through industrial and trade policies focused on the industrial sectors that are able to develop international comparative advantages and, therefore, channel to the country part of the benefits associated to these sectors.

An undervalued exchange rate has positive effects over the competitiveness of the domestic industry, creating a relative growth of the profits, favouring investments, economic growth, and development. However, the exchange rate devaluation presents, at least, two new macroeconomic challenges: (i) inflation and (ii) another challenge of temporal consistency (the capacity of the government to keep a foreign exchange policy of devaluated domestic currency requires a sufficient amount of foreign currency reserves to discourage any speculative attack against the exchange rate regime). At the bottom line, the policy would be to reproduce the Chinese model of macroeconomic management grounded in the management of the exchange rate, artificially kept competitive, and that of capital flows’ control.

On the other hand, through the microeconomic point of view, the results presented in this paper of growing risks over the sustainability of Brazilian development, would also justify the adoption of an array of trade policy measures. Especially those related to subsidizing exports or specific industrial sectors. However the big challenge is to define which are these specific sectors, since in the current socio-political context in Brazil the government does not have enough autonomy to drive an industrial pro-sector trade policy without posing externalities to the natural resource intensive sector.

The strengthening of BNDES, increasing its presence in the industrial sector, becoming a share-holder of big Brazilian enterprises, may grant the government some possibility to act discretionally in the industrial sector, benefiting those sectors that hold the capacity to created positive externalities to the Brazilian economy as a whole.

Another alternative for Brazil would be to negotiate specific trade agreements with China or with countries hold a significant percentage of Brazilian exports, in order to increase the intra-

industrial trade flows. This kind of policy represents a relevant challenge for Brazilian international relations because Chinese influence in the Latin American region seems to reduce the incentives to regional integration (Jenkins and Barbosa, 2012), especially after the free trade agreements signed between China with Peru and Chile.

Maybe the best trade policy alternative for Brazil would be a combination of all the available instruments, such as: acting together with WTO to seek multilateral solutions to the beggar-thy-neighbour behaviour and, therefore, avoiding a trade war whose results are unforeseeable for developing countries such as Brazil; using trade policy and BNDES to promote the development of industrial sectors with the capacity to create positive externalities; and use the foreign exchange policy as a way to promote industrialization. Evidently, other traditional industrial policy measures should not be kept aside, such as investment in R&D and human capital, without bringing fiscal imbalance or price instability.

8. Conclusions

Our results suggest that the economic convergence is increasing among developing countries, in which we observe an even higher intensity between Brazil and other Latin American countries with China and Asian countries; on the other hand, there is a considerable heterogeneity in the developed countries’ case, in which we show an increasing convergence with European countries and Japan, while the relation with the United States is that of divergence or economic independence; finally, there is an intense and sustainable business cycle correlation between Brazil and China, which is only surpassed by that between Brazil and some Latin American countries.

We also present evidence that business cycle convergence is positively affected by trade intensity among the selected countries. Moreover, the trade effect is higher when measured relatively from the total of the economy than when measured from manufacturing trade.

For the Brazilian case we perceived two apparently contradictory results. When we take total trade as a reference, the result indicates dynamic effects on the growth of the productive economic activity. However, when we consider the trade of manufactured goods, the result indicates the opposite, revealing that the export of manufactured goods has a negative relation (BK model) or null (HP model) with the evolution of the business cycle of the Brazilian trade partners. In the

Chinese case, the estimated βs are negative and significant, revealing that on average trade does not have positive effects on the correlation of cycles in China with the other countries of the sample.

This means that trade intensity is generating a productive specialization, what implies that the growth dynamic of the Chinese economy does not depend on trade flows and, more importantly, is not generating dynamic effects on the trade partners.

Such results are relevant since the previous literature pointed out the growing interdependence between trade and business cycle, in general, and between the more dynamic emerging economies, particularly China, and the aggregate of the world economy. Special attention should be given to the fact that Latin American countries, among them Brazil should be highlighted, are increasingly tied to the Asian economies, given the strong trade link. There are important normative implications coming from the results of this research and from the perspective of new researches it embeds. If the business cycles’ convergence is in fact tied to trade and if according to

Cepal (2011), Palma (2011), Cesa-Bianchi ( et al.

2011), IADB (2012), Lélis, Cunha and Lima

(2012), Rosales and Kuwayama (2012), Rodrik and McMillan (2011), among others, the Latin

American economies in general, and Brazil in specific, have a density and diversification lost in its productive structures and international trade, therefore the specialisation pattern may imply on less dynamic development trajectories.

8. References

BAXTER, M.; KING, R. G. Measuring Business Cycle: Approximate Band-Pass Filters form

Economic Time Series. The Review of Economics and Statistics , 81 (4): 575-93, 1999.

BERGE, T. J. Has Globalization increased the Synchronicity of International Business cycles?

Federal Reserve Bank of Kansas City Economic Review , 97 (3): 5-39, 2012.

BLONIGEN, B. et al. Comovement in GDP trends and cycles among trading partners . NBER

Working Paper 18032, Cambridge, MA – May, 2012

CALDERÓN, C. Trade, specialization, and cycle synchronization: explaining output co-movement between Latin America, China, and India. In. LEDERMAN, D., OLARREAGA, M. e

PERRY, G. (Ed.), China`s and India`s Challege to Latin America . Banco Mundial, 2008.

CALDERON, C., CHONG, A., STEIN, E. Trade Intensity and Business Cycle Synchronization: are development countries any different? Journal of international economics , 71(1): 1-21, 2007.

CALDERON, C., CHONG, A., STEIN, E. Trade intensity and business cycle synchronization: are developing countries different. Journal of International Economics , 71 (1), 2-21, 2007.

CANUTO, O.; GIUGALE, M. (Editors). The Day After Tomorrow : a handbook on the future of economic policy in the developing world. Washington, DC: The World Bank, 2010.

CEPAL. La República Popular China y América Latina y el Caribe . Hacia una nueva fase en el vínculo económico y comercial, Junio. Santiago de Chile: Comisión Económica para

América Latina, 2011. Disponível em http://eclac.org.cl

(acesso em outubro de 2011)

CESA-BIANCHI, A.; PESARAN, M. H., REBUCCI, A. XU, T.

China’s Emergence in the World

Economy and Business Cycles in Latin America . IDB WORKING PAPER SERIES No.

IDB-WP-266, September. Inter-American Development Bank, 2011.

COMMANDEUR, J. J. F.; KOOPMAN, S. J. Practical Econometrics: An Introduction to State

Space Time Series Analysis. Oxford: Oxford University Press, 2007.

DEARDOFF, A.V. Determinants of bilateral trade: does gravity work in a neoclassical world. In:

Frankel, J.A. (ed), The Regionalization of the World Economy . London: Chicago University

Press, 1998.

EATON, J., KORTUM, S. Tecnology, geography, and trade. Econometrica , v.70, n.5, pp.1741-

1779, 2002.

EICHENGREEN, L B.Should the Maastricht Treaty Be Saved? Princeton Studies in International

Finance , No. 74, International Finance Section, Princeton Univ, December, 1992.

FRANKEL, J. e ROSE, A. The Endogeneity of the Optimum Currency Area Criteria. The

Economic Journal 108, 1009–1025, 1998.

HARRIGAN, J. Specialization and the Volume of Trade : do the data obey the laws? NBER

Working Paper 8675, 2002.

IADB. The World of Forking Paths : Latin America and the Caribbean facing global economic risks.

Washington, DC: Inter American Development Bank, 2012.

IMBS, J., WACZIARG, R.. Stages of Diversification. American Economic Review , 93(1), 63-86,

2003.

IMF. World Economic Outlook , September. Washington, DC, International Monetary Fund, 2012..

JENKINS, R.; BARBOSA, A. F. 2012. ‘Fear for Manufacturing? China and the Future of Industry in Brazil and Latin America’, The China Quarterly, 2012, pp 59-81.

JORDÁN, J., PARRÉ, J. L. Dinâmica das exportações da América Latina: economias de escala ou dumping recíproco?

Economia Aplicada , v.10, n.4, pp.589-607, 2006.

KIMURA, F., LEE, H.H. The Gravity Equation in International Trade in Services . European Trade

Study Group Conference, University of Nottingham, 2004.

KOSE, M. A., CHRISTOPHER, O. E., WHITEMAN, C. Understanding the evolution of world business cycles. Journal of International Economics , 75 (1), 110-130, 2008.

KENEN, Peter B. (1969). “The Theory of Optimum Currency Areas: An Eclectic View.” In

Monetary Problems of the International Economy , ed. Robert A. Mundell and Alexander K.

Swoboda, 41-60. Chicago: University of Chicago Press.

KRUGMAN, P. Scale economies, product differentiation and the pattern of trade. The American

Economic Review , v.70, n.5, pp. 950-959, 1980.

KRUGMAN, P. Lessons of Massachussets for EMU, in Torres, F. and Giovazzi, F. Adjustment and growth in the European Monetary System , University Press, Cambridge, 1993.

LEAMER, E. E.; STERN, R. M. Quantitative International Economics . Piscataway; Transaction

Publishers, 1970.

LÉLIS, M. T., CUNHA, A.M., SANTOS, C. El desempeño de las exportaciones de China y el

Brasil hacia América Latina, 1994-2009.

Revista de la Cepal , N. 106, Abril, p. 57-77, 2012.

LINNERMANN, H. An Econometric Study of International Trade Flows.

Amsterdam, North-

Holland, 1966.

MCKINNON, R. Optimum Currency Areas. American Economic Review vol. 53 (September) pp.717-725, 1963.

MEMEDOVIC, O., LAPADRE, L. Industrial Development and dynamics of international specialization patterns . United Nations Industrial Development Organization – Working

Paper 23/2009, Viena, 2009.

MUNDELL, R. A Theory of Optimal Currency Areas, American Economic Review (September) vol.51 pp.657-65, 1961.

PALMA, G. Why has productivity growth stagnated in most Latin American countries since the neo-liberal reforms?

Cambridge Working Papers in Economics (CWPE) 1030, July, 2011.

Available at http://www.econ.cam.ac.uk/dae/repec/cam/pdf/cwpe1030.pdf

. Accessed in

09/01/2011.

POYHONEN, P. A tentative model for the volume of trade between countries. Weltwirtschaftliches

Archiv , v.90, pp.93-99, 1963.

RODRIK, D., MCMILLAN, M. Globalization, Structural Change, and Productivity Growth .

NBER Working Paper No. 17143, June, 2011.

ROSALES, O., KUWAYAMA, M.

China y América Latina y el Caribe Hacia una relación económica y comercial estratégica . Santiago de Chile, CEPAL, 2012.

TIMMER, H., DAILAMI, M., IRVING, J., HAUSWALD, R., MASSON, P. Global Development

Horizons 2011 . Multipolarity: The New Global Economy. Washington, DC: The World

Bank, 2012.

TINBERGEN, J. Shaping the Word Economy : suggestions for an international economy policy.

New York: Twentieth Century Fund, 1962.

TYSZYNSKI, H. World Trade in Manufactured Commodities, 1899-1950. The Manchester School .

V. 19, nº 3, p. 272-304, 1951.

UNCTAD. Trade and Development Report, 2010 . Geneva: United Nations Conference on Trade and Development, 2010.

UNCTAD. World Investment Report 2012 . Geneva: United Nations Conference on Trade and

Development, 2012.

WORLD BANK. World Development Indicators , 2012. Available at: http://data.worldbank.org/indicator, access on 01/07/2012.

WTO. A Practical Guide to Trade Policy Analysis . World Trade Organization, 2013.

Appendix - Sample

Algeria

Name

Argentina

Australia

Belgium

Bolivia

Brazil

Canada

Chile

China

Colombia

Costa Rica

Ecuador

Egypt

Finland

France

Germany

India

Indonesia

Iran

Italy

Japan

Kuwait

Malaysia

Mexico

Morocco

Netherlands

Nigeria

Norway

Panama

Paraguay

Peru

Philippines

Poland

Portugal

Russia

Saudi Arabia

Singapore

South Africa

South Korea

Spain

Sweden

Switzerland

Thailand

Turkey

Ukraine

United Arab Emirates

United Kingdom

United States

Uruguay

Venezuela

DEU

IND

IDN

IRN

ITA

JPN

KWT

MYS

MEX

MAR

NLD

NGA

NOR

PAN

PRY

PER

PHL

POL

PRT

Code ISO

DZA

ARG

AUS

BEL

BOL

BRA

CAN

CHL

CHN

COL

CRI

ECU

EGY

FIN

FRA

RUS

SAU

SGP

ZAF

KOR

ESP

SWE

CHE

THA

TUR

UKR

ARE

GBR

USA

URY

VEN

Download