Sincronía Fall 2000

---

Latin American Exports and Economic Growth:The Empirical Evidence

 

 

Giovanni E. Reyes [1] (University of PittsburghGraduate School of Public and International Affairs)

---

 

 

 

1.  Introduction

 

      The research problem to be addressed here is concerning the degree at which, Latin American countries have established their international links in terms of exports, and how these exports have influenced economic growth.  Exports are studied as international links and variables of the current phenomenon of globalization.  The period under study is from 1960 to 1995.  Complete data from 1996 to 2000 is not available at the moment of publishing this article, and in addition, this last mentioned period had the impact of the international financial crisis originated in the Southeast Asia in the summer of 1997 and the effect of extraordinary high prices in oil during 1999 and 2000.  These events can give to the relationship of exports and economic growth an "abnormal" repercussion.

 

      One of the main assumptions of globalization is its fundamental feature of integration.  In economic terms, this statement will imply, for the purposes of this paper, the basic hypothesis: Latin American economies have become more integrated with their main trade partners in the international economic field; and this integration has had a positive effect on economic growth.  In this regard it is important to keep in mind that, Latin American nations are trying to achieve through the export leading policies, their main macroeconomic goals:  a) price stability; b) economic growth; c) low levels of unemployment; and d) positive levels in the balance of trade.

 

      In reference to research statements, this article will deal with the following assertion: exports from Latin American countries during the period 1960-1995 have been the main cause for the economic growth of these nations.  This paper has three level of analysis:  (a) Latin America as a whole region; (b) Latin American countries grouped in terms of size of economies, structure of exports, regions and positions; and (c) Latin American countries from an individual perspective.

 

 

2.  Basic model and indicators

 

      The fundamental assumption here is that with an increasing role in the export sector, economic growth increases within economies of Latin American nations.   Some studies tended to demonstrate that developing countries with a high level of exports can achieve better standards in terms of economic growth (Feder 1982; Balassa 1991; Yarbrough 1994).  In addition, it has been claimed that the role of exports has several other beneficial impacts other than economic growth alone. 

 

      A brief summary of the “trade optimist” arguments indicate that trade liberalization[2] generates rapid export and economic growth mainly because it:  (a) promotes competition, improved resource allocation, and economies of scale in areas where less developed countries -LDCs- have a comparative advantage, consequently costs of production are lowered;  (b)  generates pressures for increased efficiencies, product improvement, and technical change, thus raising factor productivity and further lowering costs of production;  (c)  accelerates overall economic growth, which raises profits and promotes greater saving and investment and thus furthers growth; and (d) attracts foreign capital and expertise, which are in scarce supply in LDCs.[3]  On the other hand, “trade pessimist” arguments focus their attention on three major elements:  (a) limited growth of world demand for primary exports;  (b)  secular deterioration in the terms of trade for primary producing nations; and (c) rise of the “new protectionism” against the exports of LDC manufactured and agricultural goods.  They declare that LDC exports grow slowly because of four main factors. 

 

      Those factors are: (a) a shift in developed countries from low technology, material-intensive goods to high-technology, skill-intensive products, which decreases the demand for Third World raw materials; (b) increased efficiency in industrial uses of raw materials and the substitution of synthetics for natural raw materials like rubber and cotton; (c) low income elasticity of demand for primary products and simple manufactured goods; (d) the rising productivity of agriculture in developed countries and their increasing protectionism for agriculture and labor-intensive developed-country industries.[4]

 

      Empirical evidence to evaluate the pessimistic and optimistic approaches in relation to trade show that when the world economy is expanding rapidly -this was the case during the years 1960 to 1973- the more open LDCs appear to perform better in both aggregate exports and economic growth than closed-economy nations.  However, when the world economy slowed down -particularly during the period 1973 to 1977- the more open economies of LDCs with the exception of the new industrialized countries in Southeast Asia, had a more difficult time in exports and economic growth.[5]  Singer and Gray argue that when world economic conditions were more unfavorable, such as the case for the period 1977 to 1983, high growth rates of export earning occur only when external demand is strong.[6]

 

      In any case, a model studied by Gershon Feder was the essential base in obtaining the results of this paper.[7]  This model had three stages in its development.  The first one is the classical approach, according to which economic growth is a result of the interaction of two economic elements:  capital and labor.[8]  The second stage includes exports and the openness of economies in addition to the capital and labor factors.[9]   Finally, the third stage includes all the above mentioned elements plus the set of dummy variables this study is adding:  (a)  structure of exports -oil, manufacturing, and agriculture/mining;  (b) Latin American regions -Mexico/Central America/Caribbean, Andean, and Southern Cone; (c)  positions -semiperiphery and periphery countries; and (d) size of economies -large economies, medium, and small ones.  As it was mentioned above, in addition to this groups of countries, this paper will study the Feder’s model on Latin American region as a whole, and to the individual countries.

 

      All elements from this multiple regressional model will be studied during four periods of time 1960-73; 1974-82; 1983-90; and 1991-95.  During each one of these periods Latin American countries faced different international economic conditions.  A summary of these international circumstances is presented in Table No.1. 

 

      Regarding the groups of Latin American countries, the dummy variables were studied in two different ways.  First they were considered alone within the macroeconomic model.  Second, all dummy variables were seen as a complete set to establish a more integrated interaction of them and thus arriving at more representative results.  Specific countries which formed different classifications regarding dummy variables are presented in each table as a reference.  The model used to establish whether or not exports were significant elements behind economic growth in Latin America is:[10]

 

aGNP = a (aL) + b (aI) + c ((aXrg) * (X/GNP)) + Svdv

 

      Where:

            aGNP        =     rate of annual growth of  gross national product

            aL          =     rate of annual growth of labor force 

            aI          =     rate of annual growth of investments

            aXrg        =     rate of annual growth of exports 

            X/GNP      =     the percent of export over the gross national product                   Svdv        =     specific variables presented as dummy variables. 

 

 

 

 

Table 1

Multiple Regression Analysis:

General Characterization of the Four Periods of Time Under Study

 

Sub period	Latin American Economic Characteristics			International Scenario
	Economic Growth	Inflation	Other
1960/73	moderate and high	low	- Agricultural exports - Beginning of intra-regional trade agreements	- Breton Woods institutions - Stability in the international trade and financial systems - Beginning of flexible exchange rates models (1973) - No more US$/gold standard (1971)
1974/82	moderate and high	low	- Generation of external debt	- Increases in oil prices (1973, 1979) - High international bank liquidity
1983/90	low	high	- Economic adjustment plans - More flexible exchange rates models - Promotion of export leading policies	- Increase in the US interest rate - Strong US dollar until 1985 - From 1985 less strong US dollar - Reductions of financial loans to Latin America
1991/95	moderate and high	low	- Economic adjustment plans - Promotion of exports - Reinforcement of intrarregional trade agreements	- Return of capital flows to Latin America - From 1990 to 1991 economic recession in the more developed countries - From 1992 economic expansion in the US and most of the Western European nations

Source: Cardozo, E. and Helwege, A.  Latin America’s economy.  (Cambridge, Massachusetts:  MIT, 1994), Economic Commission for Latin America and the Caribbean. Latin America:  the economic experience of the last 15 Years -1980-1995.  (Santiago, Chile:  CEPAL, 1996), Inter-American Development Bank.  Economic and social progress in Latin America 1996 report.  (Washington D.C.:  IDB, 1996), Jackson, J. The world trading system. (Cambridge, Mass.: MIT Press, 1994), Walther Ted.  The world economy.  (New York:  John Wiley & Sons, Inc. 1997).

 

 

 

3.  Results and discussion

 

      From Table 2 to Table 5 results are presented more on an individual basis regarding each dummy variable.  A more complete discussion of data obtained from statistical procedures is explained for Tables 6, 7 and 8 when all the dummy variables are considered simultaneously.

 

3.1.  Latin America as a region

 

      When we applied the model to aggregate data for the complete region during the complete period (1960-1995), it is evident that the two most important macroeconomic variables are labor and investment.  They appear to have a high level of significance, less than 1 percent margen of error.  Notwithstanding the coefficient of determination is low, only 27 percent, and therefore the model only explains 27 percent of the behavior of the dependent variable.  The F coefficient has high statistical significance with a value of 98.118 -see Table 2.

 

 

Table 2

Macroeconomic Analysis:  Results Considering all Latin American Countries

for Period 1960-1995

 

Characteristic	Period 1960-1995
Intercept	0.008
Labor	1.224**
Investment	0.132**
Exports (Exp/GNP)	0.002
R2	0.271
F	98.118**
No. of Observations	36

References:  R² = Coefficient of determination; F = Coefficient of Fisher, from ANOVA.

*  =  When coefficient has statistical significance at = 0.05

**  =  When coefficient has statistical significance at = 0.01

No. of observations:  years.

 

      From Table 3 we can see that again labor and investment are important variables for Latin American region.  In this case region 1, Mexico / Central America / Caribbean countries, have a significant statistical economic growth during the period 1960-1995.  This situation can be explained by two main factors:  (a) the United States is the “natural” market for the area, with a huge internal market[11]; and (b) the Caribbean Basin Initiative which began in 1984 was a program designed to encourage exports from this region to the United States market.  Another important feature from Table 3 is that the seventies appear as a decade with strong economic growth in the region, more than the economic growth from the sixties.

 

Table 3

Macroeconomic Analysis:  Results Considering all Latin American Countries

Size of Economies, Regions, and Structure of Exports

1960-1995

 

Characteristic	Period 1960-1995
Intercept	0.826
Labor	0.987**
Investment	0.122**
Exports (Exp/GNP)	0.003
X1	1.118
X2	0.435
R1	0.757*
R2	0.788
S1	0.222
S2	0.092
1960’s	0.779
1970’s	1.264**
1980’s	-1.086*
R2	0.318
F	30.395**
No. of Observations	36

References:  X1 = manufacturing exports; X2 = oil exports;  R1= Central American, Mexico and the Caribbean countries; R2 = Andean countries; S1 = large economies; S2 medium economies.-specific countries in references Tables 5.2 to 5.5.; 1960’s, 1970’s, 1980’s = decade of sixties, seventies and eighties, respectively.

R² = Coefficient of determination; F = Coefficient of Fisher, from ANOVA.

*  =  When coefficient has statistical significance at = 0.05

**  =  When coefficient has statistical significance at = 0.01

No. of observations:  years.

 

      Table 4 is presented only as an illustration rather than a specific statistical result, because of the limited number of observations, and the problems that this feature implies with the number of degrees of freedom.  In that table, we can see how the influence of labor, investments and exports have been developed during each decade.  Again the general trend is that labor and investment are important, even though labor appears as lacking statistical significance during the nineties.

 

 

Table 4

Macroeconomic Analysis:  Results Considering all Latin American Countries

and Decades 1960-1995

 

Characteristic	1960s	1970s	1980s	1990s
Intercept	1.292	0.790	0.789	2.913
Labor	0.856**	1.430**	1.006**	0.194
Investment	0.131**	0.088**	0.162**	0.102**
Exports (Exp/GNP)	0.001	0.001	0.001	0.001*
R2	0.342	0.153	0.299	0.269
F	37.482**	13.068**	30.839**	15.701**
No. of Observations	11	10	10	5

References: R² = Coefficient of determination; F = Coefficient of Fisher, from ANOVA.

*  =  When coefficient has statistical significance at = 0.05

**  =  When coefficient has statistical significance at = 0.01

No. of observations:  years.

 

 

3.2.  Groups of Latin American countries

 

3.2.1.  Considering size of economies

 

      According to Table 5 the variable regarding labor shows statistical significance in two periods 1960-73 and 1974-82.  Investments or capital factors always appear as a significant variable at 1 percent margin of error.  Another important point is to realize that during all the periods except for 1991-95 exports were significant factors for economic growth.  When economic conditions in the international economic scenario were more stable and the Bretton Wood institutions were working more in terms of its original design, especially during the period 1960-73, the large economies of Latin America seem to have significantly higher levels of economic growth. 

 

 

Table 5 

Macroeconomic Analysis:  Results Considering Size of Economies

 

Characteristic	Periods
	1960/73	1974/82	1983/90	1991/95
Labor	0.959**	1.584**	0.776	-0.226
Investment	0.121**	0.120**	0.137**	0.096**
Exports (Exp/GNP)	0.003**	0.002**	0.002**	0.009
S1	2.216**	-0.426	0.091	2.054*
S2	0.866	-1.558	0.479	3.846**
S3	0.987	-0.908	-1.038	2.775**
R2	0.328	0.277	0.340	0.326
F	24.660**	12.285**	14.616**	8.385**
No. of Observations	308	198	176	110

References:  S1 = Big economies: Argentina, Brazil and Mexico;  S2 = Medium economies:  Chile, Colombia, Peru and Venezuela;  S3 = Small economies: Bolivia, Costa Rica, Dominican Republic, Ecuador, El Salvador, Guatemala, Guyana, Haiti, Honduras, Jamaica, Nicaragua, Panama, Paraguay, Trinidad and Tobago, and Uruguay.  R² = Coefficient of determination; F = Coefficient of Fisher, from ANOVA.

*  =  When coefficient has statistical significance at = 0.05

**  =  When coefficient has statistical significance at = 0.01

No. of observations:  number of countries (22) times number of years.

 

      However, when international economic conditions were not so stable and yet Latin American countries continued to implement processes of economic adjustment (period 1991-95) small and middle size economies in the region appear to have higher levels of economic increase in Table 5.5.  The F coefficient emerges with high and significant values at 1 percent margin of error, and the coefficient of determination was only 34.  This last number reflects the limited explanation of the model used in terms of economic growth.

 

 

3.2.2.  Considering structure of exports

 

      Following results contained in Table 6 one realizes that labor, investments and exports had a similar impact in economic growth as shown in Table 5.  The labor factor was significant during the periods 1960-73, and 1974-82.  This result could be associated with the internal demand factor as an element for economic growth.   At the same time investments again emerge to have a high significant statistical values -1 percent margin of error- in all periods under study, and exports had a significant repercussion on economic growth in all periods but 1991-95.  Oil and agricultural export oriented countries had statistical positive significance with regards to economic growth in 1991-95.  F coefficient was highly significant, and the coefficient of determination could not rise beyond 35 percent.

 

 

 Table 6 

Macroeconomic Analysis:  Results Considering Structure of Exports

 

Characteristic	Periods
	1960/73	1974/82	1983/90	1991/95
Labor	0.956**	1.430**	0.709	-0.384
Investment	0.122**	0.119**	0.136**	0.100**
Exports (Exp/GNP)	0.003**	0.002**	0.002**	0.001
X1	-----	0.966	0.867	2.834
X2	1.056	0.319	-0.337	1.862*
X3	1.151*	-0.928	-0.593	3.486**
R2	0.315	0.285	0.327	0.336
F	27.907**	12.797**	13.788**	8.779**
No. of Observations	308	198	176	110

References: X1 = Manufacturing 1960/73 zero; 1974/95: Brazil; X2 = Oil 1960/73 Trinidad and Tobago and  Venezuela 1974/95 Ecuador, Mexico, Trinidad and Tobago, and Venezuela; X3 = Agriculture and Minining  1960/73 Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, Dominican Republic, Ecuador, El Salvador, Guatemala, Guyana, Haiti, Honduras, Jamaica, Mexico, Nicaragua, Panama, Paraguay, Peru and Uruguay 1974/95 Argentina, Bolivia, Chile, Colombia, Costa Rica, Dominican Republic, El Salvador, Guatemala, Guyana, Haiti, Honduras, Jamaica, Nicaragua, Panama, Paraguay, Peru, and Uruguay.

R² = Coefficient of determination; F = Coefficient of Fisher, from ANOVA.

*  =  When coefficient has statistical significance at = 0.05

**  =  When coefficient has statistical significance at = 0.01

No. of observations:  number of countries (22) times number of years.

 

3.2.3.  Considering regions

 

      Following similar results from Tables 5 and 6, Table 7, the evidence shows the pattern of impact on economic growth derived from the labor, investment and export factors:  (a) labor was significant during all periods except 1991-95;  (b) investments always appeared as significant elements at a 1 percent margin of error;  (c) exports were always significant except during 1991-95.  Based on their regional placement, Southern cone countries and Brazil had significant influence on economic growth during 1960-73.  They and the Andean countries emerge as having significant influence on economic expansion during 1991-95.  In this last feature, a factor behind this result could be the reinforced intratrade positions from the MERCOSUR nations -Argentina, Brazil, Paraguay and Uruguay.  They have incorporated Chile and Bolivia as partially integrated members since 1994.  Again the F coefficient appears to have high significant values at a 1 percent margin of error, and the coefficient of determination achieved its highest value during the periods 1983-90, and 1991-91 with 35 percent, showing the limited scope of explanation of economic growth using the exogenous variables of this model.

 

 

Table 7 

Macroeconomic Analysis:  Results Considering Regions

 

Characteristic	Periods
	1960/73	1974/82	1983/90	1991/95
Labor	1.020**	1.663**	1.052**	-0.011
Investment	0.122**	0.117**	0.137**	0.098**
Exports (Exp/GNP)	0.003**	0.002**	0.002**	0.001
R1	0.971	-1.311	-1.655	1.685
R2	0.816	-1.326	-1.654	3.432**
R3	1.193*	-0.477	0.561	2.908**
R2	0.316	0.277	0.356	0.353
F	23.291**	12.300**	15.707**	9.483**
No. of Observations	308	198	176	110

References: R1 = Central America, Mexico, Caribbean: Mexico, Guatemala, El Salvador, Honduras, Nicaragua, Costa Rica, Panama, Haiti, Dominican Republic, Jamaica, and Trinidad and Tobago; R2 = Andean Countries: Colombia, Venezuela, Guyana, Ecuador, Peru, and Bolivia; R3 = Southern Cone and Brazil, Chile, Argentina, Uruguay, Paraguay and Brazil

R² = Coefficient of determination; F = Coefficient of Fisher, from ANOVA.

*  =  When coefficient has statistical significance at = 0.05

**  =  When coefficient has statistical significance at = 0.01

No. of observations:  number of countries (22) times number of years.

 

 

 

3.2.4.  Considering Positions[12]

 

      Labor, investments and exports followed the same and consistent trend as it was shown in other tables: (a) labor was significant during all periods except 1991-95;  (b) investments always appeared as significant elements at a 1 percent margin of error;  (c) exports were always significant except during 1991-95.  The only periods in which semiperiphery and periphery countries did not have significant economic growth were during the generation of the regional international debt (1974-82) and during the era of implementing economic adjustment without economic growth (1983-1990).  As usual, the F coefficient appears to have high significant values at a 1 percent margin of error, and the coefficient of determination achieved its highest value during the period 1983-90 with 35 percent.

 

 

 

Table 8 

Macroeconomic Analysis:  Results Considering Positions

 

Characteristic	Periods
	1960/73	1974/82	1983/90	1991/95
Labor	0.885**	1.546**	0.852**	-0.240
Investment	0.122**	0.120**	0.136**	0.099**
Exports (Exp/GNP)	0.003**	0.002**	0.002**	0.001
P1	2.078**	-0.899	0.698	2.861**
P2	1.145*	-0.867	-1.306	2.867**
R2	0.324	0.273	0.356	0.302
F	29.102**	14.541**	18.925**	9.098**
No. of Observations	308	198	176	110

References: P1 = Semiperiphery: 1960/73  Argentina, Brasil, Mexico and Venezuela; 1974/95  Argentina, Brasil, Mexico, Venezuela, Chile and Colombia; P2 = Periphery: 1960/73  Bolivia, Chile, Colombia, Costa Rica, Dominican Republic, Ecuador, El Salvador, Guatemala, Guyana, Haiti, Honduras, Jamaica, Nicaragua, Panama, Paraguay, Peru, Trinidad and Tobago, and Uruguay; 1974/95  Bolivia, Costa Rica, Dominican Republic, Ecuador, El Salvador, Guatemala, Guyana, Haiti, Honduras, Jamaica, Nicaragua, Panama, Paraguay, Peru, Trinidad and Tobago, and Uruguay

R² = Coefficient of determination; F = Coefficient of Fisher, from ANOVA.

*  =  When coefficient has statistical significance at = 0.05

**  =  When coefficient has statistical significance at = 0.01

No. of observations:  number of countries (22) times number of years.

 

 

3.2.5.  Considering results with all dummy variables

 

      Table 9 shows the results when considering all the dummy variables as a complete set.  From these numbers one can infer that labor maintained its significance during all periods, except during 1991-95.  This situation appears to be related to the characteristic that economic adjustment programs diminished the impact of labor on the GNP.  In turn, this characteristic was associated with reduction in the internal effective demand, chiefly due to devaluations and depreciations of currencies.  One of the main results of this complete scenario was the increased levels of poverty in the region.

 

      Another important feature presented here is the highly significant degree of investments.  This variable as utilization of capital factors appears to have notable impact on economic growth in Latin American countries during the 36 years from 1960 to 1995.  Exports also had a heavy impact on economic expansion during all periods except 1991-95. 

 

 

 

      During this last mentioned period the weight of causality for economic growth appears to have changed to other factors that are not included in this multiple regressional model, as the highly statistical significance of the intercept is indicating.  The coefficient of determination shows that of all the exogenous variables, they can reach only 40 percent of the economic growth explanation.  This is evident, even though all independent variables under consideration have notable degree of causality (F value had a high significance, at a 1 percent margin of error).

 

      Pertinent inferences about dummy variables are derived from Tables 7 and 8 because colinearity was detected between semiperiphery positions and large economies.  All large economies emerge as having a semiperiphery condition. 

 

 Table 9

Macroeconomic Analysis:  Results Considering all Dummy Variables

Characteristic	Periods
	1960/73	1974/82	1983/90	1991/95
Intercept	0.926	0.310	0.458	3.848**
Labor	1.067**	1.732**	0.959*	-0.308
Investment	0.123**	0.110**	0.149**	0.099**
Exports (Exp/GNP)	0.001	-0.001	0.001	-0.001
X1	-----	1.090	0.231	-0.840
X2	0.734	1.530	0.415	-1.947*
P1	-0.727	-1.606	3.549*	0.068
R1	0.175	-1.793	-1.855*	-0.776
R2	0.219	-1.966	-2.522*	1.090
S1	1.840	0.650	-3.707*	-0.283
S2	-0.387	0.377	-0.932	0.112
R2	0.263	0.242	0.350	0.399
F	11.829**	5.983**	8.905**	6.581**
No. of Observations	308	198	176	110

References:  X1 = manufacturing exports; X2 = oil exports;  P1 = semiperiphery; R1= Cetral American, Mexico and the Caribbean countries; R2 = Andean countries; S1 = large economies; S2 medium economies.-specific countries in references Tables 5.2 to 5.5.

R² = Coefficient of determination; F = Coefficient of Fisher, from ANOVA.

*  =  When coefficient has statistical significance at = 0.05

**  =  When coefficient has statistical significance at = 0.01

No. of observations:  number of countries (22) times number of years.

 

 

 

3.2.6.  Considering results with all dummy variables except positions

 

      From Tables 10 and 11 it is evident that labor was not important on economic growth during the period 1991-95 only; that investments have always a significant role in economic expansion; and that exports do not appear to have important repercussions on the increase of regional production.  Even more, oil exports emerge as having a negative impact on economic growth during 1991-95.  One explanation for this condition can be found in the historically low price levels that oil has had, on average data, during the nineties. 

 

      Concerning Table 10 and in terms of regions, Central America / Mexico / Caribbean nations appear to have economic growth based on their geographical location.  Andean countries also experienced economic expansion affected by regional conditions.  One factor influencing this result is the implementation of the Caribbean Basin Initiative -CBI.  It began in 1984, and among its leading aims was to promote trade links with the United States based on foreign investments and non traditional exports.  The highest value of the coefficient of determination was about 40 percent, and the F value as usual shows high statistical significance. 

 

 

Table 10 

Macroeconomic Analysis: 

Results Considering all Dummy Variables Except Positions

Characteristic	Periods
	1960/73	1974/82	1983/90	1991/95
Intercept	1.015	0.178	0.742	3.856**
Labor	1.034**	1.726**	0.967*	-0.309
Investment	0.123**	0.111**	0.152**	0.099**
Exports (Exp/GNP)	-0.002	-0.001	0.001	-0.001
X1	-----	1.074	0.290	-0.838
X2	0.403	1.302	0.925	-1.937*
R1	0.206	-1.621	-2.225*	-0.781
R2	1.999	-1.667	-3.176**	1.078
S1	1.102	-0.788	-0.518	-0.221
S2	-0.463	-0.847	1.778	0.165
R2	0.262	0.239	0.334	0.399
F	13.307**	6.581**	9.275**	7.386**
No. of Observations	308	198	176	110

References: X1 = manufacturing exports; X2 = oil exports; R1= Cetral American, Mexico and the Caribbean countries; R2 = Andean countries; S1 = large economies; S2 medium economies.-specific countries in references Tables 5.2 to 5.5.

R² = Coefficient of determination; F = Coefficient of Fisher, from ANOVA.

*  =  When coefficient has statistical significance at = 0.05

**  =  When coefficient has statistical significance at = 0.01

No. of observations:  number of countries (22) times number of years.

 

3.2.7.  Considering results with all dummy variables except size of economies

 

      In reference to Table 11, besides the already mentioned situation of labor, investments and exports, as well as the oil producing countries during the nineties, Andean countries appear to have a more negative impact during the eighties, that is during the implementation of economic adjustment programs.  Here an important factor was that especially Argentina, Bolivia, and Brazil were having extraordinary high levels of  inflation.[13]  Finally and again, the highest value of the coefficient of determination was about 40 percent, and the F value as usual shows high statistical significance. 

 

Table 11 

Macroeconomic Analysis: 

Results Considering all Dummy Variables Except Size of Economies

 

Characteristic	Periods
	1960/73	1974/82	1983/90	1991/95
Intercept	1.144	0.344	0.061	3.808**
Labor	0.973**	1.727**	1.094*	-0.296
Investment	0.122**	0.110**	0.151**	0.099**
Exports (Exp/GNP)	0.001	-0.001	0.003	-0.001
X1	-----	1.261	-1.379	-1.062
X2	-0.157	1.531	0.154	-2.000*
P1	0.857	-1.148	1.598	0.016
R1	0.311	-1.799	-1.870	-0.776
R2	-0.031	-1.953	-2.142*	1.196
R2	0.257	0.242	0.333	0.398
F	14.845**	7.544**	10.436**	8.372**
No. of Observations	308	198	176	110

References: X1 = manufacturing exports; X2 = oil exports;  P1 = semiperiphery; R1= Cetral American, Mexico and the Caribbean countries; R2 = Andean countries;  -specific countries in references Tables 5.2 to 5.5.

R² = Coefficient of determination; F = Coefficient of Fisher, from ANOVA.

*  =  When coefficient has statistical significance at = 0.05

**  =  When coefficient has statistical significance at = 0.01

No. of observations:  number of countries (22) times number of years.

 

 

 

3.3.  Latin American countries on individual basis

 

      This study applied the econometric analysis to individual countries rather than including these countries as individual dummy variables with the other economic characteristics in order to avoid problems concerning colinearity.  When we apply the macroeconomic model, in terms of the main macroeconomic variables -labor, investment, and exports- to individual countries it is possible to define several important characteristics.  First, most of the countries appear to have investment as a crucial macroeconomic variable for economic growth during the period 1960-1995.  The exceptions are Chile, Colombia, Ecuador, El Salvador and Honduras.  Second, in the cases of El Salvador, Panama, Venezuela, and Mexico, the labor aspect appears to have high statistical significance.  Labor was also important in achieving economic growth in the cases of Brazil, Colombia, and Guyana.  Mexico and El Salvador are the only cases in which exports constitute an important variable for economic growth with statistical significance within the range of 5 percent margen of error.  Third, better levels for explaining economic growth were observed in the cases of Brazil, the Dominican Republic, Mexico and Panama.  In those cases the values for the coefficient of determination were:  64, 64, 77, and 60 percent respectively (see Table 12).

 

 

 

 

Table 12

Macroeconomic Analysis:  Results Considering all Latin American Countries

and Labor, Investment and Exports,

1960-1995

No.	Country	Interc.	Labor	Invest.	Exports	R2	F
1	Argentina	7.849	-4.532	0.087*	0.005	0.143	1.802
2	Bolivia	4.443	3.098	0.054*	-0.006	0.167	2.045
3	Brazil	3.172	2.813*	0.286**	-0.013	0.642	19.076**
4	Chile	5.716	-0.986	0.068	0.003	0.132	1.603
5	Colombia	1.859	0.946*	0.052	0.008	0.158	2.023
6	Costa Rica	10.542	-2.157	0.102**	-0.001	0.384	6.752
7	Dominican Rep.	6.115	-1.518	0.217**	-0.001	0.642	19.413**
8	Ecuador	2.846	2.902	0.021	-0.005	0.154	1.952
9	El Salvador	4.031	2.702**	0.005	0.003*	0.492	10.321**
10	Guatemala	6.635	-1.112	0.072*	-0.006	0.201	2.832*
11	Guyana	1.832	2.274*	0.072*	0.002	0.192	2.632*
12	Haiti	5.868	-3.173	0.092**	-0.001	0.312	4.843**
13	Honduras	2.448	0.201	0.093	0.001	0.142	1.883
14	Jamaica	1.731	-0.063	0.182**	0.005	0.442	8.551**
15	Mexico	1.401	1.742**	0.233**	0.001*	0.772	36.974**
16	Nicaragua	0.903	0.763	0.233**	0.002	0.462	9.312**
17	Panama	4.811	3.283**	0.152**	0.006	0.602	16.342**
18	Paraguay	0.126	1.262	0.223**	0.003	0.491	10.253**
19	Peru	4.170	2.433	0.301**	0.001	0.582	15.062**
20	Trinidad and Tobago	0.833	2.172	0.171**	0.003	0.233	3.261*
21	Uruguay	0.673	0.691	0.212**	0.001	0.542	12.536**
22	Venezuela	4.012	2.123**	0.132**	0.041	0.572	14.297**

References: R² = Coefficient of determination; F = Coefficient of Fisher, from ANOVA.

*  =  When coefficient has statistical significance at = 0.05

**  =  When coefficient has statistical significance at = 0.01

No. of observations for each country:  36 (years).

 

 

      When we applied analysis taking into account only the dummy variables for the years of each decade, we can see the rates of economic growth, not only for individual countries, but also for the region as a whole (see Table 13).  From these results the high levels of economic growth during the sixties and seventies are evident, as well as during the nineties.  The eighties appear having a kind of insignificant economic gain in terms of national production.  For some countries, nevertheless, the eighties were years that on the average represented some kind of economic improvement.  This was the case in Brazil, Chile, Colombia, Costa Rica, and Paraguay.  These nations had an important recovery concerning their national output especially by the end of the last decade.  On the other hand, important economic contraction can be seen particularly in the cases of Haiti and Nicaragua.

 

 

Table 13

Macroeconomic Analysis:  Results Considering all Latin American Countries

and 1960s, 1970s, 1980s and 1990s

1960-1995

 

No.	Country	1960s	1970s	1980s	1990s	R2	F
1	Argentina	3.908**	3.156*	0.433	3.533*	0.102	0.907
2	Bolivia	4.542**	4.037**	-0.228	4.056**	0.436	6.256**
3	Brazil	5.873**	8.549**	3.065**	1.162	0.369	4.665**
4	Chile	3.994*	3.061	4.479**	6.161**	0.042	0.354
5	Colombia	4.982**	5.886**	3.432**	4.367**	0.293	3.333*
6	Costa Rica	5.495**	6.345**	3.645**	4.187**	0.173	1.666
7	Dominican Rep.	3.587*	8.132**	2.933	2.169	0.184	1.787
8	Ecuador	3.765**	8.887**	2.488*	3.333*	0.369	4.512**
9	El Salvador	5.177**	4.557**	-0.343	5.098**	0.402	5.535**
10	Guatemala	4.833**	5.986**	1.178	4.097**	0.464	7.056**
11	Guyana	3.289*	1.995	-2.852	5.334**	0.274	3.032*
12	Haiti	0.232	3.687**	0.197	-2.123	0.302	3.512**
13	Honduras	3.865**	6.137**	2.632*	2.567	0.163	1.638
14	Jamaica	3.628**	0.828	0.987	1.532	0.082	0.786
15	Mexico	5.124**	6.473**	2.133*	1.598	0.284	3.132*
16	Nicaragua	6.566**	0.697	-1.388	0.835	0.187	1.788
17	Panama	7.448**	4.749**	0.166	5.533**	0.302	3.465**
18	Paraguay	4.032**	7.746**	4.187**	3.098*	0.231	2.435*
19	Peru	3.167*	3.964*	0.293	3.666	0.082	0.733
20	Trinidad and Tobago	4.421**	5.286**	-1.121	1.666	0.193	1.957
21	Uruguay	1.865	2.688*	0.632	2.834	0.042	0.399
22	Venezuela	5.287**	4.023**	-0.289	3.806*	0.258	2.766*

References: R² = Coefficient of determination; F = Coefficient of Fisher, from ANOVA.

*  =  When coefficient has statistical significance at = 0.05

**  =  When coefficient has statistical significance at = 0.01

No. of observations for each country:  36 (years).

 

      Finally, when we applied the model taking into account the main macroeconomic variables and the years of each decade (see Table 14)  it is important to realize that labor and investment keep their trend in terms of being crucial variables for economic growth.  It is also meaningful to see that the Latin American region has not recovered a statistical significant level of economic growth during the nineties.  Based on the coefficient of determination results, countries in which the model more significantly explained economic growth were Brazil (76 percent), Mexico (80 percent), and the Dominican Republic (75 percent).  In terms of decades, El Salvador, Guyana and Nicaragua appear to have important economic contraction within the region.

 

 

 

 

Table 14

Macroeconomic Analysis:  Results Considering all Latin American Countries

Labor, Investment, and Exports during 1960s, 1970s, 1980s and 1990s

1960-1995

 

No	Country	Labor	Invest	Exports	1960s	1970s	1980s	1990s	R2	F
1	Argentina	6.87	0.07	0.02	12.46	12.09	9.74	10.22	0.19	1.01**
2	Bolivia	1.65	0.02	0.01	0.24	-0.02	-3.96	-0.15	0.45	3.46
3	Brazil	3.44	0.27**	0.01	5.35	3.18	-3.51	2.45	0.76	13.26**
4	Chile	3.62	0.07	0.04	12.58	8.12	10.19	12.43	0.22	1.17
5	Colombia	0.38	0.05*	0.04	3.31	4.51	2.06	2.81	0.39	2.68**
6	Costa Rica	3.58	0.08**	0.02*	15.32	16.11**	13.66*	12.65*	0.57	5.55**
7	Dominican Rep.	1.95	0.21**	0.02	6.08	10.29**	6.41	5.22	0.75	12.61**
8	Ecuador	3.48	0.05	0.02	7.19	1.10	-6.78	-3.81	0.41	2.89*
9	El Salvador	1.78*	0.05	0.03	0.98	1.14	-3.71*	-0.22	0.58	5.72**
10	Guatemala	0.25	0.05*	0.03	5.48	5.93*	1.78	4.57	0.57	5.50**
11	Guyana	2.30	0.07*	0.01	-2.30	-2.03	-4.22*	3.20	0.38	2.62*
12	Haiti	1.32	0.07*	0.01	-1.88	0.35	-2.39	-4.81	0.42	3.08*
13	Honduras	2.99	0.07	0.04	13.44*	16.05*	12.47	10.30	0.35	2.27*
14	Jamaica	0.81	0.19**	0.05	3.49	3.24	0.90	3.13	0.51	4.33**
15	Mexico	2.19*	0.23**	0.08	3.74	1.84	-2.26	-2.80	0.80	17.22**
16	Nicaragua	5.92	0.20**	0.06	12.25	17.85	-15.83	-17.92	0.53	4.85**
17	Panama	1.30	0.14**	0.05	1.77	0.45	-1.09	-1.15	0.61	6.60**
18	Paraguay	0.59	0.19**	0.05	1.65	3.55	1.91	0.49	0.55	5.23**
19	Peru	2.17	0.30**	0.06	7.65	9.03	4.20	4.73	0.64	7.61**
20	Trinidad &Tob.	2.05	0.11	0.02	0.04	1.66	-3.24	-0.68	0.31	1.92
21	Uruguay	3.45	0.24**	0.08	2.51	3.28	0.02	-1.33	0.58	5.86**
22	Venezuela	1.85	0.12**	0.04*	2.39	2.97	-5.18	-1.64	0.66	8.04**

References: R² = Coefficient of determination; F = Coefficient of Fisher, from ANOVA.

*  =  When coefficient has statistical significance at = 0.05

**  =  When coefficient has statistical significance at = 0.01

      No. of observations for each country:  36 (years).

 

 

 

4.    Conclusions

 

      Taking into consideration that the main element of this paper is related to the question of whether or not Latin American economies have had exports as a significant force behind economic growth, and keeping in mind the macroeconomic model utilized here, it is possible to underline two main points. 

 

First, among the variables under study as factors for rates of economic growth -labor, investments, exports, and openness of economies- investment and labor have shown evidence as being dominant factors for economic growth in Latin America during the period 1960-1995.  Only in some specific cases of oil exporter countries, namely Mexico and Venezuela, and during limited periods (especially during the seventies), exports were having a notable effect on economic growth levels.  Labor, as a variable, appears as having a significant impact on economic growth until 1990.

 

      Second, characteristics for grouping Latin American countries and used here as dummy variables -(a)  structure of exports -oil, manufacturing, and agriculture/mining;  (b) Latin American regions -Mexico/Central America/Caribbean, Andean, and Southern Cone; (c)  positions -semiperiphery and periphery countries; and (d) size of economies -large economies, medium, and small ones- did not show to be factors with statistical significant influence on economic growth.  Nevertheless it is important to realize that a more serious and systematic effort for opening Latin American economies has been developed since the end of the eighties.

(Pittsburgh, October 2000)

 

 

 

 

Notes

 

[1] Professor, University College, University of Pittsburgh, with numerous publications in the areas of economics of development and political affairs; former Executive Director of the International Center for Pre-Investment and Development in Central American and the Caribbean; former representative for the International Coffee Organization in London, ex-consultant for the Economic Commission for Latin America and the Caribbean, the Inter-American Development Bank, and the United Nations Organization.

 

[1]   Trade liberalization is understood in the context of this paper as a set of policies aimed at including export promotion, currency devaluation, and removal of trade restrictions and some governmental controls.  See Lal, D.  and Rajapatirana, S.  “Foreign trade regimes and economic growth in developing countries”, World Bank Research Observer 2, no. 2 (July 1987);  Maizels, A.  Exports and economic growth of developing countries.  (London:  Cambridge University Press, 1988); and Todaro, M.  Economic development in the Third World.  (New York:  Longman, 1995).

 

[1]   See Eicher, C. and Witt, L.  Agriculture in economic development.  (New York:  McGraw-Hill, 1987), pp. 311-322, and Todaro, M.  op. cit. pp. 439-453.

 

[1]  See Krueger, A.  Trade and employment in developing countries.  (Chicago, Illinois:  University of Chicago Press, 1987), pp. 54-67.

 

[1] See Helleiner, G.  International trade and economic development.  (Harmondsworth, England: Penguin, 1990); and Stewart, F. Theory and reality in development.  (London:  Mcmillan, 1986), pp. 125-132, 143-154.

 

[1]  See Singer, H. and Gray, P. “Trade policy and growth of developing countries” in World Development 16, no. 3 (March 1988): 395-403.

 

[1] Feder, G. On exports and economic growth, in Journal of Development Economics 12 (1982) (New York: North-Holland Publishing Company, 1982) pp. 59-73.

 

[1] Chereny, H. et. al.  A uniform analysis of development patterns, Economic Development Report no. 148.  July 1980 (Cambridge Ma.: Harvard University Press).

 

[1]  See Feder, op.cit. p. 64.

 

[1]   Based on a macroecnomic multiple regressional model from Gershon Feder, see Feder, G. On exports and economic growth, in Journal of Development Economics 12 (1982) (New York: North-Holland Publishing Company, 1982) pp. 59-73.

 

[1]   By the year 2,000 the United States population will be 268 million, which is about 53 percent of the complete Latin American population.  In addition and in terms of the internal market demand in the United States it is important to take into account that they constitute 4 percent of the world’s population  with 22 percent of the worldwide wealth.  The internal demand in the United States accounts for almost 68 percent of the “driven force” behind their trend of economic growth since 1991 to the present.  See  Gwynne, R. and Kay, C.  (eds.)  Latin America transformed:  globalization and modernity.  (London, U.K.: Arnold Publs.  1999), pp. 98-120, 156-159;  Singer, D.  Whose millennium?.  (New York:  Monthly Review Press, 1999), pp. 184-186; and Porter, M.  The competitive advantage of nations.  (New York:  The Free Press, 1998), pp. 277-284, 543-546, and 719-721.

 

[1]   Positions are according to network analysis models.  Countries in the same position have a similar pattern of international trade relations with other nations.

 

[1]   In terms of annual percent of inflation:  Argentina had 4,924 in 1989; Bolivia 7,945 in 1985; and Brazil 983 in 1989.  See International Monetary Fund.  International Financial Statistics 1992.  (Washington D.C.:  IMF, 1993), pp. 34-53.

 

 

 

5.    Bibliography

 

Chereny, H. et. al.  A uniform analysis of development patterns, Economic Development Report no. 148.  July 1980 (Cambridge Ma.: Harvard University Press).

 

Eicher, C. and Witt, L.  Agriculture in economic development.  (New York:  McGraw-Hill, 1987).

 

Feder, G. On exports and economic growth, in Journal of Development Economics 12 (1982) (New York: North-Holland Publishing Company, 1982).

 

Gwynne, R. and Kay, C.  (eds.)  Latin America transformed:  globalization and modernity.  (London, U.K.: Arnold Publs.  1999).

 

Helleiner, G.  International trade and economic development.  (Harmondsworth, England: Penguin, 1990).

 

International Monetary Fund.  International Financial Statistics 1992.  (Washington D.C.:  IMF, 1993).

 

Krueger, A.  Trade and employment in developing countries.  (Chicago, Illinois:  University of Chicago Press, 1987).

 

Lal, D.  and Rajapatirana, S.  “Foreign trade regimes and economic growth in developing countries”, World Bank Research Observer 2, no. 2 (July 1987).

 

Maizels, A.  Exports and economic growth of developing countries.  (London:  Cambridge University Press, 1988)

 

Porter, M.  The competitive advantage of nations.  (New York:  The Free Press, 1998).

 

Singer, D.  Whose millennium?.  (New York:  Monthly Review Press, 1999).

 

Singer, H. and Gray, P. “Trade policy and growth of developing countries” in World Development 16, no. 3 (March 1988).

 

Stewart, F. Theory and reality in development.  (London:  Mcmillan, 1986).

 

Todaro, M.  Economic development in the Third World.  (New York:  Longman, 1995).

 

 

-------------------------0-------------------------