Impact of International Trade on Economic Growth: A Study of Brazil and South Africa

1
International trade: Brazil and South Africa
By:
Course Title:
Tutor:
Date:

2
Chapter Three: Methodology
As evident, the study seeks to exhibit the impact of international trade to the country’s
economic growth. Notably, one of the measures applicable, in evaluating the economic growth of
a country is the Gross Domestic Product (GDP). Therefore, the study will use GDP and its
components to examine the economic growth. There are four components of the GDP, which
include investments (I), government expenditure (G), consumption (C), and net exports (NX)
thus GDP (Y) = I + C + G + NX. Notably, among the BRICS, both Brazil and South Africa are
categorized as developing countries hence the study will incorporate these countries to examine
if international trade has an impact on their respective economic growth.
The data used in the study was sourced from the world bank GDP database, which is
freely available to users. The GDP metadata has approximately 1600 variables, which include
GDP, year, imports, exports, land area, population, expenditure, manufacturing, and agricultures,
among others (World Bank, 2018). However, the study incorporated seven variables, which
include year, GDP, employment rates, investments, imports, exports, and expenses. Moreover,
the data has values from 1960 to 2018; however, some variables such as investment and
expenses have missing values. As a result, Brazil data used values from 1990 to 2018 whereas
South Africa used Data from 1972 to 2018. Moreover, the study omitted investments from the
South African data set due to numerous missing values. Notably, the study used the Statistical
Package for Social Sciences (SPSS) to perform data analysis.
Visualization
It is essential for a research to have a clear picture about the data set before any analysis.
As a result, the study incorporated two visualization tools, which include the line and scatter
graphs. The line graph was used to exhibit the trend of GDP over the years thus aid in

3
determining an increase or decrease (growth) of GDP. On the other side, the scatter plot was
used to exhibit the relationship between various variables. The plot incorporates both of the X-
axis, Y-axis, and a series of dots, whereby the dots exhibit the interaction between the variables
of interest. Consequently, it is important to showcase the descriptive statistics of the data set
since it aids in exhibiting the general characteristics of the variables (Manikandan, 2011).
Correlation
Similar to scatter plots correlation is a statistical tool that shows the level of
interdependencies between the variables of interest. For instance, the correlation will be used to
show the association between GDP, employment rates, exports, and investments. Notably, the
result values of correlation is referred to as the correlation coefficient ρx,y, which is lies between
limits of negative one (-1.0) and positive one (+1.0) (Hayes,2019. Consequently, in the event the
value is approximately zero (0), this shows low or absence of interdependence between the
variables of interest. However, if the result is positive, it shows that an increase in one variable
leads to an increase in the other variable of interest whereas a negative result shows that a
decrease in one case variable leads to an increase in the other variable of interest.
Regression Analysis
Besides, the study showcased the relationship between variables through the use of
regression method. Regression analysis technique is a predictive modeling procedure which
exhibits the relationship between the dependent and independent variables (Ray, 2015). The
dependent variable represents the main factor that a stud seeks to predict or understand whereas
independent variables are factors that study claims to have an impact on the main factor (Foley,
2018). Notably, there are various advantages of using the regression analysis, which include
showing the significant relationship and the strength of the impact of multiple between the

4
dependent and other independent variables. There are various types of regression techniques,
such as a ridge, multiple, stepwise, polynomial, linear, logistic, among others (Ray, 2015).
There are various factors considered when selecting any of the above models. The primal
step before selecting the model is the data exploration that aids in identifying the relationship
between the variables (Ray, 2015). Besides, it is essential to compare the goodness of fit for the
above methods using several measures, such as adjusted R-square, R square, and statistical
significance, among others. Notably, the most efficient way to evaluate the models is through
cross-validation whereby the dataset is divided into two groups (train and validate) and the mean
square difference between the observed and the predicted values will give the measure for
prediction accuracy (Ray, 2015). Moreover, the objective of the analysis may influence the
selection of the model. Furthermore, in cases whereby the variables in the dataset exhibits high
dimensionality and multicollinearity, it is recommended to use the Lasso and Ridge regression
models.
However, among the above methods, linear and multiple regressions are the most
common type of regression. The linear regression incorporates on one independent variable,
whereas the multiple involves more than one independent variables (Beers, 2019). Notably,
when using these techniques, the dependent variable should be continuous; however, the
explanatory variable can be either continuous or discrete. The general form of these types of
regression is given by;
Linear regression: Y = b0 + bX + e
Multiple regression: Y = b0 + b1X1 + b2X2 + ....... + bnXn + e
Whereby; Y is the dependent variable, Xs are the idependent variables, b0 is the
intercept, b (slope) is the coefficient of the explanatory variables, and e is the regression residual