Data and methodology description
Through the study of a sensitive subject like teenage pregnancy and the extent of the
impact of this problem on education. Our study focused on three countries that have a high rate
teenage pregnancy in Sub-Sharan African countries (Congo, Mali, and Uganda) in a period of time
from 1968 to 2017. Data used for the study were downloaded from the World Bank dataset. We
have two models that show the impact of this issue on the school enrollment as well as economic
growth. Ordinary Least Squares regression (OLS) method used to predict outputs' values for
sample. The first model:
= 𝛽0 + 𝛽1 𝑡𝑒𝑒𝑛𝑎𝑔𝑒 𝑝𝑟𝑒𝑔𝑛𝑎𝑛𝑐𝑦 + 𝛽2 𝐺𝐷𝑃 𝑔𝑟𝑜𝑤𝑡ℎ 𝑝𝑒𝑟 𝑐𝑎𝑝𝑖𝑡𝑎
+ 𝛽3 𝑎𝑐𝑐𝑒𝑠𝑠 𝑡𝑜 𝑒𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑖𝑡𝑦 + 𝜀𝑡
Where the dependent variable Y is the rate of students who are in secondary education fulfills
the provision of basic education that started at the primary level, 𝛽1 teenage pregnancy rate is
our first independent variables in order to measure the impact on school enrollment, 𝛽2 GDP
growth per capita (proxy for poverty), 𝛽3 access to electricity is a good indicator in order to
capture the life standard.
The second model:
GDP per capita
= 𝛽0 + 𝑤 𝑙𝑛(𝑡𝑒𝑐ℎ𝑛𝑜𝑙𝑜𝑔𝑦) +
𝛽1 School enrollment + 𝛽2
investment + 𝛽3 Inflation + 𝛽5 urban population growth + 𝛽6 Exports +
𝛽7 Dum_Mali + 𝛽8 𝐷𝑢𝑚𝑚𝑦1 + 𝛽9 𝐷𝑢𝑚𝑚𝑦_2 + 𝜀𝑡
Through this model is related the first in terms of capture the impact of school enrollment on
GDP per capita which we assume that is a good representative economic growth. based on the
Solow model, A is the level of technology in each country and w is the weight. Through level of
technology, we found that each country has a different top trade partner. The table below show
the top trade partner in each country.
Top Partner Trade
We the dummy variables in order to measure a significant shift on GDP per capita in each
country. The inflation (proxy), Foreign investment, and Exports variables have a significant on
GDP; thus, we should take it into count. AR process was applied in the regressions.
Results and Discussion Model 2
GDP_PER_CAPITA_GROWTH SCHOOL_ENROLLMENT...FOREIGN_IN... INFLATION_... URBAN_POP...
The table above indicates the correlation among the variable that we use on our
regression in Model 2. It shows that there is a positive correlation between school enrolment and
GDP Per Capita Growth in Congo, Mali, and Uganda which does not mean causality. However,
our dependent variables are not highly correlated which mean they would capture the actual
Dependent Variable: GDP_PER_CAPITA_GROWTH
Method: Panel Least Squares
Sample (adjusted): 1981 2016
Periods included: 36
Cross-sections included: 3
Total panel (unbalanced) observations: 91
S.E. of regression
Sum squared resid
Mean dependent var
S.D. dependent var
Akaike info criterion
Dependent Variable: GDP Per Capita Growth
1- School Enrolment growth.
2- Foreign Investment % of GDP
3- Inflation % of GDP.
4- Urban population Growth.
5- Export % of GDP.
6- Dummy for Mali.
7- Dummy 2 is for the shift on GDP in 1990 for 1993 Congo
8- Dummy 1 is for the shift on GDP in 1987 for Mali
9- Level of technology weighed by top partner
10- lag 1,2,3 for GDP Per Capita Growth
In this model, we have attempting to see the impact of school enrollment on GDP per
capita which we assume that is a good representative economic growth. As school enrollment is
affected negatively by teenager pregnancy in the first model because teenager pregnancy
statically significant impact school enrolment. As consequences, the outcomes of education as
whole in these countries would be affected negatively due increase drop-out rate. However, in
term of measure the impact of these circumstances on economic growth and development, we
considered school enrollment on our measurement of economic growth to see the exactly
After testing the model from any violation, we found that some variables have problem
with unit root. Urban population Growth has problem with unit root in level, so we generate first
different for that variable. Also, we found problem on variable of Export and we generate first
In addition, we consider in this model autocorrelation which corrected by conducting tree
AR processes. Also, Multicollinearity have checked by using VIF of all the variables and the result
shows no Multicollinearity because the value of VIF were under 10.
In addition, the sample on this model after adjustment is consider the period between
1981 and 2016 which is good enough to see the long run impact on economic growth (GDP Per
Capita Growth). Also, the total observations are in total 91 which also is effective to see the actual
effect because it is more than 45 observations, and we will be able to find appropriate result.
However, the dependent variables mean is 0.487a and the stander deviation of dependent
variables is 4.643.
The regression result Variable
Urban Population Growth
Dummy (Mali country)
Dummy 1 (Mali shift in 1987)
Dummy 2 (Congo shift in 1990 to 1993)
Level of Technology
Δ GDP Per capita
* indicates significance at 1% level, ** significance at 5% level, *** significance at 10% level
After running the regression above, the result illustrates almost all the variables are
statistically insignificant except school enrolment, the dummy 1 variable for Mali shift in 1987,
and dummy 2, as well as the three AR processes which would capture the effect on GDP in
previous years on current GDP. In addition, the value of the Durbin-Watson statistic is 2.14 which
seem good and R-squared value is 0.63, which indicates that the variables could be a good fit
more than 0.5. However, the school enrolment is statistically significant at the level of 5% level
and have positive influence on GDP per capita growth, in other word a one percent change on
school enrolment is associated with an increase on GDP per capita by 6.1.
The dummy variable of the shift on GDP per capita growth of Mali in 1987 which has
positive impact on economic growth in these countries and the its probability is statically
significant. As well as, the dummy 2 which indicate shift on GDP Per Capita of Congo between
1990 and 1993 is statically significant. These two dummies created to see the real impact of
school enrolment on GDP Per Capita Growth by ignoring these years that may give us different
result. Moreover, the dummy that generated to capture the impact of Mali country on GDP Per
Captiva of these country is statically insignificant with positive sign. Which mean, there are no
impact of this dummy of GDP.
Moreover, when we look to the three AR processes that added to the model to solve the
problem of autocorrelation, they indicate different result. The first AR (1) is insignificant with
positive sing of coefficient. The second AR (2) processes has a statically significant impact on our
independent variable (GDP Per Capita Growth) at 1% level. Also, the last AR (3) processes is
statically significant at 10% level.
The regression result gives an evidence that as the school enrolment increase the GDP
Per Capita would increase too. Which mean, the educational outcomes in these countries are a
fundamental factor that will enhance economic development in long run. Investing in education
will result a benefit on economic growth. However, it is essential that these countries should
overcoming the violations that may impact education negatively such as teenage pregnancy and
poverty as well as access electricity and transportation.
Purchase answer to see full