Term Paper

Child Mortality in Bangladesh: Female employment &

Health Expenditure

Course Title: Basic Econometrics

Course Code: ECO465

Submitted to:
Syed Abul Basher
Professor
Department of Economics
East West University

Submitted by:

Fahim Shakil : 2017-1-30-051

Md Easin Arafath : 2017-1-30-043
Ishrat Jahan Dia : 2014-3-30-026

Submission Date: 16.05.202

1|Page
 Introduction

Infant mortality rate (the probability of dying between birth and age one) is a very important
indicator to measure the progress of child health. In Bangladesh, the decline in infant mortality
(64/1000 in 2000 to 25/1000 in 2018) is a great achievement as a low-income country. To achieve
SDGs, reduction in infant mortality rate will contribute significantly to improve the health status
of children. However, Bangladesh is still far from reaching its targeted mortality rate.
Nevertheless, recently in developing countries like us, a progress has been made to improve child
survival and health. However, socioeconomic inequality still same which is a resistant factor of
mortality in children.

In this paper we have focused to improve the socioeconomic conditions of the households. Female
employment is an important factor which have positive impact on infant mortality as it pushes to
increase the female education rate. Besides, it increases the family income and create opportunity
to spend money for children’s health care. Another important factor is expenditure on health care.
They combinedly ensures the use of primary healthcare services such as immunization and family
planning, and better breastfeeding and complementary feeding practices. By taking into
consideration these factors, the progress for child health and survival can be achieved. Various
socio-economic and demographic, nutritional and other factors cause infant mortality. Our study
shows that reduction in infant mortality is strongly associated with female employment and health
expenditure.

2|Page
 Literature Review

We went through four research paper (name in the reference) on the cause and determinants of
infant mortality in developing countries and current condition in Bangladesh. Those studies show
that in spite of various effective intervention programs, the infant mortality rate is still high in
Bangladesh. The risk of dying in the first month of life is still very high. The most significant
predictors of neonatal, post-neonatal, infant and child mortality were residence, parent’s education,
type of toilet facility, wealth status, watching TV, months of breastfeeding, and birth interval.
Despite the improvement of medical technology, child mortality remained alarmingly high,
indicating that demographic, socioeconomic, household and environmental conditions must be
improved to substantially reduce child mortality in this population. In this paper our study is
limited to estimate the effect of female employment and health expenditure on infant mortality
rate.

Model Specification

A model was created to show the effect of female employment, health expenditure on the infant
mortality in Bangladesh. We collected data from 2000 to 2018 to test our model

Infant Mortality = f (Female Employment, Health Expenditure)

The linear regression model is:

Infant Mortality = 𝜷𝟎 + 𝜷𝟏 (Female Employment) + 𝜷𝟐 (Health Expenditure) + Et
Where:
𝜷𝟎 : The constant term, interception point of dependent & independent variables
𝜷𝟏 : Coefficient of female employment
𝜷𝟐 : Coefficient of health expendeture
t: The time trend.
E: The random error term assumed to be normally, identically and independently
distributed

3|Page
 Data

As we are working on time series Data, we took 19 years (2000-2018) of data for regression
analysis. We took them from the website of world development indicators (WDI).

Dependent variable: Infant mortality is the independent variable here. Infant mortality is the
death of young children under the age of 1. This death toll is measured by the infant mortality
rate, which is the number of deaths of children under one year of age per 1000 live births.

Infant Female Employment to Health expenditure to GDP

Year
Mortality Population ratio % ratio %
2000 63.9 26.06399918 2.00194679
2001 60.9 26.13299942 1.98674463
2002 58.1 26.16600037 2.15190127
2003 55.4 26.1779995 2.19150465
2004 52.9 26.03700066 2.21404936
2005 50.4 26.0882 2.27546357
2006 48 26.09700012 2.32201219
2007 45.5 26.51199913 2.38769203
2008 43.2 26.92200089 2.34595705
2009 41 27.18899918 2.40338314
2010 38.9 28.52799988 2.49591507
2011 36.8 28.70599937 2.56593749
2012 34.9 28.8560009 2.57301107
2013 33 28.95800018 2.50334758
2014 31.3 29.45800018 2.49601118
2015 29.6 29.95299911 2.46311463
2016 28 30.45499992 2.31177993
2017 26.5 33.44100189 2.27424875
2018 25.1 33.70999908 2.42735

Year I_Mortality% F_E2P% HE_GDP%

2000 63.9 26.06399918 2.00194679
2001 60.9 26.13299942 1.98674463
2002 58.1 26.16600037 2.15190127
2003 55.4 26.1779995 2.19150465

4|Page
 2004 52.9 26.03700066 2.21404936
2005 50.4 26.0882 2.27546357
2006 48 26.09700012 2.32201219
2007 45.5 26.51199913 2.38769203
2008 43.2 26.92200089 2.34595705
2009 41 27.18899918 2.40338314
2010 38.9 28.52799988 2.49591507
2011 36.8 28.70599937 2.56593749
2012 34.9 28.8560009 2.57301107
2013 33 28.95800018 2.50334758
2014 31.3 29.45800018 2.49601118
2015 29.6 29.95299911 2.46311463
2016 28 30.45499992 2.31177993
2017 26.5 33.44100189 2.27424875
2018 25.1 33.70999908 2.42735
Independent variable: Female employment to population ratio is on independent variable here.
Another independent variable is percentage of budget spent for health care. Both of them are in
percentage form.

5|Page
 Methodology

To determine the effect of female employment and health expenditure on infant mortality rate in
Bangladesh with empirical data, ordinary least square (OLS) regression is used to estimate in the
following way.

IMR = β0 + β1(FE) + β2(HE) + Et

Where IMR is the number death of infants (age of less than 1 year) per one thousand live birth,
FE is percentage of female employment, HE is percentage of GDP allotted for health and Et is
error term.
We used STATA software for regression analysis.

Descriptive Statistics

Variable Obs Mean Std. Dev. Min Max

I_Mortality 19 42.28421 12.18443 25.1 63.9

F_E2P 19 28.18164 2.415887 26.037 33.71
HE_GDP 19 2.336388 .1718386 1.986745 2.573011

Average infant mortality rate is 42 among one thousand live birth in Bangladesh which is quite
high comparing to many other developing countries. Female Employment is 28% which indicate

6|Page
a good ratio comparing the others countries. However only 2.34 percent of our budget is spent for
health which is very poor.

HE_GDP%
3
2.5
2
1.5
1
0.5
0
2000200220042006200820102012201420162018

Empirical Results

Our model is significant as p value is less than 0.05. Even individually each variable is significant is p
value is less that 0.0 5in each case. We can explain 94% of variance in dependent variable through our
model.

7|Page
The negative value of the coefficients indicates increasing the independent variables will decrease the
value of dependents variable. If 1 unit of female employment increase, it will decrease 3 unit of mortality.
Again, increase in 1 unit of health expenditure will decrease mortality by 33 unit. If there is zero female
employment zero health expenditure, there will be 214 infant death per 1000 live birth.

Test for Heteroskedasticity:

5
0
Residuals
-5
-10

20 30 40 50 60
Fitted values

. estat hettest

Breusch-Pagan / Cook-Weisberg test for heteroskedasticity

Ho: Constant variance
Variables: fitted values of I_Mortality

chi2(1) = 1.59
Prob > chi2 = 0.2074

From this result we can conclude that there is no homoscedasticity as p value is greater than 0.05.
Heteroscedasticity occurs when there is a circumstance in which the variability of a variable is unequal
across the range of values of a second variable that predicts it. So, our model is free from
heteroscedasticity.

8|Page
Multicollinearity Test: Multicollinearity refers to a situation in which two or more explanatory variables
in a multiple regression model are highly linearly related. We have perfect multicollinearity if, for
example as in the equation above, the correlation between two independent variables is equal to 1 or −1

. pwcorr . vif
(Year ignored because string variable)
Variable VIF 1/VIF
I_Mort~y F_E2P HE_GDP
F_E2P 1.25 0.798363
I_Mortality 1.0000 HE_GDP 1.25 0.798363
F_E2P -0.8723 1.0000
HE_GDP -0.7703 0.4490 1.0000 Mean VIF 1.25

From the test result we can say that our model is free from multicollinearity problem as VIF values are
less than 10.

Test for Autocorrelation: Autocorrelation, also known as serial correlation, is the correlation of a signal
with a delayed copy of itself as a function of delay. Informally, it is the similarity between observations as
a function of the time lag between them. For time series data we need to do Durbin-Watson test.
5

Durbin-Watson d-statistic( 3, 19) = 1.132701

0
Residuals
-5
-10

0 5 10 15 20
trend

As our calculated value is between the lower and upper limit, the result is inconclusive. This is a
limitation of our study as our sample size is small due to unavailability of data.

9|Page
 Conclusion

There are many factors for infant death in developing countries. Lack of education, low living
standard, lack of proper health services. From our study we can conclude that if the females are
educated and employed with simultaneous proper expenditure on health can significantly reduce
the rate of infant mortality rate of Bangladesh. Others studies also concluded the similar forecast.

10 | P a g e
 Reference

1. Infant Mortality Situation in Bangladesh in 2007: A District Level Analysis (A Research paper
done by Mazhabul Ahmed, Sabir Bairagi and Uttam Deb)
2. A comprehensive analysis on child mortality and its determinants in Bangladesh using frailty
models (by Jhidur Rahman Khan and Nabil Awan)
3. Determinants of Infant And Child Mortality in Bangladesh: Time Trends and Comparisons across
South Asia (Bangladesh Journal of Medical Science)
4. BBS, SID
5. Risk factors of infant mortality in Bangladesh (by Joyti Vijay, Kamalesh Kumar Patel)

11 | P a g e