tiểu luận kinh tế lượng DETERMINANTS OF LIFE EXPECTANCY AMONG DEVELOPING COUNTRIES IN 2 YEARS 2015 2016

Acknowledging of the importance of this indicator, this study attempts to examine the economic and health determinants of life expectancy among 123 developing countries for the period 20

FOREIGN TRADE UNIVERSITY FACULTY OF ECONOMICS AND INTERNATIONAL BUSINESS

Class ID: KTEE310(1-1920).1_LT

Instructor: Dr Nguyen Thuy Quynh

Table of Contents

ABSTRACT 1

INTRODUCTION 2

Chapter 1 OVERVIEW OF LIFE EXPECTANCY AND ITS DETERMINANTS 5

1.1 Definitions 5

1.1.1 Life expectancy and some related terms 5

1.1.2 Factors that affect life expectancy among developing countries 5

1.2 Review of economics theories about life expectancy 7

1.2.1 Development research 7

1.2.2 Research hypotheses 8

Chapter 2 MODEL SPECIFICATION OF THE IMPACTS OF 3 DIRECT FACTORS ON LIFE EXPECTANCY 9

2.1 Methodology 9

2.1.1 Method to derive the model 9

2.1.2 Method to collect and analyze the data 9

2.2 Theoretical model specification 10

2.2.1 Specification of the model 10

2.2.2 Explanation of the variables 11

2.3 Data description 12

2.3.1 Data sources 12

2.3.2 Descriptive statistics and interpretation 12

2.3.3 Correlation analysis 13

Chapter 3 ESTIMATED MODEL AND STATISTICAL INFERENCES OF THE IMPACTS OF 3 DIRECT FACTORS ON LIFE EXPECTANCY 15

3.1 Estimated model 15

3.1.1 Estimation result 15

3.1.2 Initial sample regression model 15

3.2 Testing problems of the model 15

3.2.1 Misspecification 15

3.2.2 Multi-collinearity 16

3.2.3 Heteroskedasticity 17

3.2.4 Autocorrection 18

3.2.5 Normal distribution of the disturbance 18

3.3 Testing research hypotheses 19

3.3.1 Testing the significance of an individual regression coefficients 19

3.3.2 Testing the overall significance of the model 20

3.4 Discussions 21

3.4.1 Meaning of the estimated coefficients 21

3.4.2 Coefficients of determination 22

3.4.3 Interpretation of estimation outputs 22

3.5 Recommendations 24

CONCLUSION 25

REFERENCES 27

APPENDIX 28

1 The STATA estimation outputs 28

2 Dataset of 123 developing countries during 2015 and 2016 30

INDIVIDUAL ASSESSMENT 38

ABSTRACT

Life expectancy is a key summary measure of the health and wellbeing of a population

A nation’s life expectancy reflects its social and economic conditions and the quality of its public health and healthcare infrastructure, among other factors Acknowledging of the importance of this indicator, this study attempts to examine the economic and health determinants of life expectancy among 123 developing countries for the period 2015-

2016 In this research, the panel data method together with multiple regression model are applied to draw conclusions about the relationship between life expectancy and selected economic and health factors All explanatory variables turned out to be statistically significant, which imply that our chosen factors, to be specific, GDP, death causes due to non-communicable diseases and disease prevention such as immunization for measles should be considered to play an vital role in the changes in life expectancy

Based on our research it has been suggested that these developing countries should formulate and implement appropriate economic and health policies and programs to improve the quality of life, as well as reduce the risk of dangerous diseases to make a huge change in the longevity of human

INTRODUCTION

1 Reasons for choosing the topic

Life expectancy at birth, widely used as an indicator of overall development of a country, has significantly increased over the last ten years in most of the developing countries in the world This has a particular indication for the developing world since they are striving earnestly for achieving socio-economic progress through investing on social sectors like health, education, environmental management, etc For its essential role in the economic growth of the developing countries, we wonder that: “By what ways can the governments of the developing countries can increase the average number of life expectancy in their country?”

In order to figure out how, the analysis of determinants of life expectancy cannot

be avoidable Although the determinants are different in different periods of time, we can conclude that the major determinants which have influences on life expectancy in developing countries include GDP per capita, cause of death by non-communicable diseases, immunization for measles In the report, we will use the econometric model to find out specifically whether they have negative or positive relationship And from the result, we might be able to find a way for solving the life expectancy problems

Having learnt the econometrics course, we realized the importance of practicing analyzing with social economic figures After having collected enough data from world bank and other sources and thanks to the guidance of our lecturer Dr Nguyen Thuy Quynh, we decided to choose our topic “Determinants of life expectancy among developing

countries in two years 2015 and 2016”

2 Research objectives

In building this research, we wish to answer these following questions:

• General question: In year 2015 and 2016, what factors affect the life expectancy?

• Specific questions:

✓ Which factors have the most influence on life expectancy from the beginning of 2015 to the end of 2016?

✓ To what extents can those factors affect life expectancy in two years 2015 and 2016?

✓ What recommendations should be made to increase the life expectancy among developing countries?

3 Objects and scope of the research

Objects:

This research was made to analyze the effects of 3 following factors (which are independent variables in the regression model built in the next chapter) on the life expectancy (the dependent variable):

✓ GDP per capita

✓ Cause of death by non-communicable diseases

✓ Immunization of measles

Scope of the research:

Research time: In two-year-time: from the beginning of 2015 till the end of 2016

Research space: 123 developing countries

4 Research findings

In establishing this research, we have gained so many things Firstly, the life expectancy plays a vital role in developing the economy As a result, the government in many developing countries have raised the expenditure on medical to higher level Secondly, there are several factors affecting the life expectancy, including the direct factors and indirect factors For instance, direct determinants are: health spending, food supply, education, …; indirect determinants are: economic misery, urbanization, environmental management and sustainability, and social safety nets In these factors, we have chosen

3 main factors as we have discussed above

5 Structure of the report

The report contains the following contents:

• Chapter 1: Overview of life expectancy and its determinants

• Chapter 2: Model specification of impacts of 3 direct factors on life expectancy

• Chapter 3: Estimated model and statistical inference of impacts of 3 direct factors

on life expectancy CONCLUSION

REFERENCES APPENDIX INDIVIDUAL ASSESSMENT

Due to the time restriction and limitations in database, as well as our only fundamental knowledge about macroeconomics in general and econometrics in particular, errors and omissions are probably inevitable We are enthusiasts who are willing to learn from our mistakes, therefore, we are eager to receive your feedbacks and recommendations to improve our report

Chapter 1 OVERVIEW OF LIFE EXPECTANCY AND ITS

DETERMINANTS

1.1 Definitions

1.1.1 Life expectancy and some related terms

It is a statistical measure of the average time an organism is expected to live,

based on the year of its birth, its current age and other demographic factors including

gender The most commonly-used measure is life expectancy at birth (LEB), which

can be defined in two ways Cohort LEB is the mean length of life of an actual birth cohort (all individuals born in a given year) and can be computed only for cohorts born many decades ago, so that all their members have died Period LEB is the mean length of life of a hypothetical cohort assumed to be exposed, from birth through death,

to the mortality rates observed at a given year

Roles of life expectancy in economic growth: Life expectancy is one of indicators

representing human development of a country and at some extent, it can reflect the nation

living standards According to Ranis et al (1999), economic growth and development is

a two-way relationship According to them, the first chain consists of economic growth benefiting human development, since economic growth is likely to lead families and individuals to use their heightened incomes, which implies better access to health services and other items, leading to higher life expectancy At the same time, the increased consumption in health, medical facilities is an important contributer to economic growth Lower mortality may increase income per capita by increasing the productivity of available resources (most notably human capital)

1.1.2 Factors that affect life expectancy among developing countries

1.1.2.1 Economic factor: Gross domestic product per capita (GDP per capita)

Gross Domestic Product (GDP) is the total monetary or market value of all the finished goods and services produced within a country's borders in a specific time period

As a broad measure of overall domestic production, it functions as a comprehensive scorecard of the country’s economic health

Effect of GDP on life expectancy: As GDP grows higher, health expenditure will

be of higher priority when government allocates its resources Health expenditure consists of all expenditures or costs for medical care, prevention, promotion, rehabilitation, community health activities, health administration and regulation and capital formation with the predominant objective of improving health in a country or region When the health expenditure increases, the average life expectancy of people living in that country is clearly improved The amount of health expenditure is one of most important indicators of development Due to a rapid increase in economic growth

in developing country, as a result of direct correlation between these two factors, the life expectancy also tends to increase

1.1.2.2 Health factors

a) Cause of death, by non-communicable diseases

Non-communicable disease is a disease that is not transmissible directly from one person to another NCDs included as indicators in the World Health Organization Global Action Plan for the Prevention and Control of NCDs 2013–2020 (i.e., cardiovascular disease, cancer, respiratory disease, and diabetes), comprised 70% of all deaths globally, and 80% of all premature NCD deaths in 2010 NCDs may be chronic or acute Most are non-infectious, although there are some non-communicable infectious diseases, such

as parasitic diseases in which the parasite's life cycle does not include direct host-to-host

transmission

Effect of non-communicable diseases on expectation of life: non-communicable

diseases (NCDs) constitute an overwhelming majority of global premature mortality As

we have found in the Central European journal of public health, in 2014, life expectancy

at birth was 76.87 years compared to 72.87 in 1996 The highest impact on life expectancy was recorded for ischaemic heart disease and PGLEs (both of which are non-communicable diseases) have increased For this figure, we can expect that when there

is an increase in the number of people having non-communicable disease, there will be

a extension in the average life expectancy

b) Immunization for measles:

Measles is a highly contagious viral disease despite the Measles has been estimated to account annually for as many as two million deaths to children worldwide

(WHO, 1985) However, with availability of a safe and effective vaccine, this disease is

no longer a tragedy Child immunization, measles, measures the percentage of children ages 12-23 months who receive the measles vaccination before 12 months A child is considered adequately immunized against measles after receiving one dose of vaccine

Effect of immunization, measles on life expectancy: Measles vaccine increases child

survival beyond protecting against measles All-cause mortality is significantly lower in children who received the measles vaccine after the third diphtheria-tetanus-pertussis (DTP) vaccination Vaccines dramatically reduced the incidence of infectious diseases that historically killed hundreds of millions and made a substantial contribution to life expectancy that during the last century increased from ∼47–80 years Obviously,

immunization of measles is positively correlated with life expectancy

1.2 Review of economics theories about life expectancy

1.2.1 Development research

The review of literature on life expectancy as a proxy for nation’s health status is useful in order to investigate the factors that affect it In this respect, this section is allocated to review of literature on determinants of nation’s life expectancy

In 2003, Erick Messias did a research about relationships between life expectancy

and gross domestic product (GDP), income disparity and illiteracy rate Simple and multiple linear regressions were performed to measure the association between income disparity, measured by the Gini coefficient, GDP per capita, and illiteracy rate The result was that the variable GDP per capita positively associated with life expectancy

In August 2014, Rino Rappuoli established a research about the impacts of

vaccines on longevity, wealth, science The result showed that children born in some developing countries in Asia during 1994–2013, vaccines have prevented 322 million illnesses, 21 million hospitalizations, and 732.000 premature deaths, saving $295 billion

in direct medical costs This result showed a positive impact of using vaccines on protecting health and increase the average life expectancy

Ruqiya Pervaiz and Özlem Ercantan (2018) investigated the relationship

between non-communicable diseases (NCDs) and age-standardized mortality (ASRM)

Data for ASRM of NCDs and premature mortality (before aged 70) in percentage for total NCDs in 2015 were obtained from the World Health Organization (WHO) Linear regression model was used for assessment of correlation between mortality and the rate

of having NCDs One-way ANOVA was used to test the difference in mean mortality of various groups of people having NCDs; P < 0.05 was considered significant

In September 2019, John Clements investigated in the relationship between GDP

per capita and life expectancy In his research, he established a regression model where life expectancy is a dependent variable and GDP per capita is independent variable, choosing significant level of 5% and estimating by OLS method The result is that when

an increase in GDP per capita will increase life expectancy and this exploration strongly suggest that investing in healthcare will have positive effects on a country’s economy

Summing up, all of the researches have shown the relationship between the independent variables which are GDP, non-communicable and measles immunization

However, none of these presented studies used OLS (ordinary least squares) method to

estimate both the single influences and the multi influences of these factors on the dependent variable-life expectancy Besides, this study represents an update for the mid 2010s, as well as an improvement on our previous studies, by including only developing countries and by including better estimates of relative factors through OLS method

1.2.2 Research hypotheses

i) GDP per capita, Cause of death caused by non-communicable diseases and measles immunization have significant impacts on life expectancy at birth

ii) GDP per capita affects positively life expectancy at birth

iii) Death caused by non-communicable diseases affects positively life expectancy at birth

iv) Immunization for measles affects positively life expectancy at birth

Chapter 2 MODEL SPECIFICATION OF THE IMPACTS

OF 3 DIRECT FACTORS ON LIFE EXPECTANCY

2.1 Methodology

2.1.1 Method to derive the model

Multiple Linear Regression is the statistical technique that we apply in our research to predict the outcome of a dependent variable by using several explanatory variables To be specific, in our research, we aim to demonstrate the statistically dependent relationship of GDP per capita, death caused by Non-communicable diseases and Immunization for measles rate on Life expectancy

2.1.2 Method to collect and analyze the data

• Collect the data

We use panel data method, which is the data on life expectancy among 123 developing countries in two years 2015 and 2016 The data were collected from World Bank, which has a very high level of precision

• Analyze the data

Our group has used Stata to analyze the dataset and interpret the correlation matrix between variables

We use the method of OLS, which was created by Carl Friedrich Gauss, a German mathematician Under certain basic assumptions, the OLS method has some very attractive statistical properties that have made it one of the most precise to estimate the unknown parameters Given basic assumptions of classical linear regression model (CLRM), the least-squares estimators, in the class of unbiased linear estimators, have minimum variance, that is, they are BLUE

✓ Efficient: minimum variance in the class of all such linear unbiased estimators

Besides, logarithmic transformation is a convenient means of transforming a highly skewed variable into a more normalized dataset In order to transform the distribution of the features to a more normally-shaped bell curve, we use log - log model to describe the relationships between our dependent and independent variables.

2.2 Theoretical model specification

2.2.1 Specification of the model

Base on related public researches and economic theories, the model used in this report

is constructed to examine the impacts of relevant factors including GDP per capita, death caused by Non-communicable diseases and Immunization for measles rate on Life expectancy at birth:

LE = f (GDP, NCOM, IMMU)

Where:

• LE: Life expectancy at birth (years)

• GDP: Gross Domestic Products per capita ($)

• NCOM: Cause of death, by non-communicable diseases (%)

• IMMU: Immunization for measles of children aged 12-23 months (%)

To display the relationship between the dependent variable, which is LE, and independent variables, GDP, NCOM, IMMU has the following form:

Population Regression Model:

(PRF) lnLE = 𝜷𝟏+ 𝜷𝟐 lnGDP + 𝜷𝟑 lnNCOM + 𝜷𝟒 lnIMMU + 𝒖𝒊

Sample Regression Model:

(SRF) lnLE = 𝜷 ̂ + 𝜷𝟏 ̂ lnGDP + 𝜷𝟐 ̂ lnNCOM + 𝜷𝟑 ̂ lnIMMU + 𝒖𝟒 ̂𝒊Where:

• lnLE: dependent variable

• lnGDP, lnNCOM and lnIMMU: independent variables

• 𝜷𝟏 is the intercept of the regression model

• 𝜷𝒊, i=2,4 is the slope coefficient of the independent variables

• 𝒖𝒊is the disturbance of the regression model

• 𝜷̂ is the estimator of the intercept 𝟏

• 𝜷̂, i=2,4 is the estimator of 𝜷𝒊 𝒊

• 𝒖̂ is the residual (the estimator of 𝒊 𝒖𝒊)

2.2.2 Explanation of the variables

lnGDP (X1) Natural logarithm of GDP per capita + Current $

lnNCOM (X2) Natural logarithm of Cause of death,

by non-communicable diseases + % of total

lnIMMU (X3) Natural logarithm of Immunization

% of children ages 12-23 months

Table 1 The meaning of independent variables

Life expectancy at birth is used as the dependent variable indicating that the number of

years a newborn infant would live if the existing conditions of mortality at the time of its birth remain to be the same throughout its lifespan

A proxy to measure life expectancy at birth in our dataset is “period life expectancy”

approach1 which only observe the mortality rates from birth through death among

individuals of different age groups at one particular period – commonly a year It does

not take into account how mortality rates are changing over time In order to estimate the average length of life, age-specific mortality rates are then needed

Economic factor of life expectancy is assumed to be represented by GDP per capita

Health determinants of life expectancy are explained via cause of death due to communicable diseases and immunization for measles rate

non-1 It is used by most international organizations, including the UN and the World Bank, when reporting ‘life expectancy’ figures

Reasons for expectations:

• Expectation value for GDP per capita: positive (>0)

An increase in GDP means an economic growth As a country develops, its people move out of poverty together with an increase in GDP will lead to the increase in living standards – higher real incomes and the ability to devote more resources to areas like health care and education Therefore, the life expectancy at birth rate will increase

• Expectation value for GDP per capita: positive (>0)

With the epidemiological transition, causes of death shifted from communicable to non-communicable diseases (NCDs) and life expectancy increased, as these NCD deaths occurred later in life Therefore, at least people could have higher chances of survival if being caught non-communicable disease instead of the massive loss due to infectious diseases, which as a result, expected lifespan will be increased

• Expectation value for GDP per capita: positive (>0)

Immunizations help protect people from getting infectious diseases or those that often have no medical treatments, which cause serious complications and even death

Besides, a small number of people susceptible to diseases, such as impaired immune systems may not be able to get timely vaccinations Better vaccines, especially for contagious disease like measles allow better control of existing diseases and reducing the burden of infectious diseases Hence, an improvement in immunization can increase life expectancy

2.3 Data description

2.3.1 Data sources

The dataset was collected from the official website of World Bank, including 246 observations of 123 developing countries in 2 separate years 2015 and 2016

2.3.2 Descriptive statistics and interpretation

Run the command “sum LE GDP NCOM IMMU” to interpret the dataset, the result obtained including the number of observations (Obs), the average value (Mean), the standard deviation (SD) as well as the minimum (Min) and maximum (Max) values of each variable as the table below:

Variable Obs Mean Std Dev Min Max

Table 2 Result from running SUM function of variables

From the table, we can indicate that:

• LE: the mean life expectancy among 123 developing countries in 2 years is

70.1397, standard deviation is 7.340437, the minimum figure is 50.881 and the maximum one is 83.6024

• GDP: the mean GDP per capita is 6602.815, standard deviation is 9487.788, the

minimum figure is 315.777 and the maximum one is 82081.6

• NCOM: the mean percentage of cause of death, by non-communicable diseases

is 65.8565, standard deviation is 22.21927, the minimum figure is 25.4 and the maximum one is 95.2

• IMMU: the mean percentage of immunization, measles is 86.76423, standard

deviation is 13.32825, the minimum figure is 37 and the maximum one is 99

Overall, the independent variables have a quite strong correlation relationship with the dependent one Especially, the variable “NCOM” has a very strong relationship (which

is 0.8987)

• The correlation coefficient between lnLE and lnGDP is 0.7856, which is positive

strongly positive Therefore, GDP has a positive effect on LE, a slight change in the GDP growth rate will lead to a remarkable change in the life expectancy

• The correlation coefficient between lnLE and lnNCOM is 0.8987, which is

strongly positive Therefore, NCOM has also a positive effect on LE, a slight change in the NCOM rate will lead to a considerable change in the life expectancy

• The correlation coefficient between lnLE and lnIMMU is 0.5854, which is

moderately positive Therefore, IMMU has a positive effect on LE, a slight change in the IMMU rate will lead to a major change in the life expectancy

• The correlation coefficient between lnGDP and lnNCOM is 0.7544, which is

strongly positive Therefore, GDP and NCOM have a positive effect on each other, a slight change in the GDP rate will lead to a major change in the NCOM rate

• The correlation coefficient between lnGDP and lnIMMU is 0.4667, which is

moderately positive Therefore, GDP and IMMU have a positive effect on each other, any change in the GDP rate will lead to a covariated change in the IMMU rate

• The correlation coefficient between lnNCOM and lnIMMU is 0.5433, which is

moderately positive Therefore, NCOM and IMMU have a positive effect on each other,

a slight change in the NCOM rate will lead to a major change in the IMMU rate

Chapter 3 ESTIMATED MODEL AND STATISTICAL INFERENCES OF THE IMPACTS OF 3 DIRECT FACTORS

ON LIFE EXPECTANCY

3.1 Estimated model

3.1.1 Estimation result

Run the command reg lnLE lnGDP lnNCOM lnIMMU to compute the estimation

result, the result obtained is as the below table:

3.1.2 Initial sample regression model

We have the Sample Regression Model:

lnLE = 𝜷̂𝟏 + 𝜷̂𝟐 lnGDP + 𝜷̂𝟑 lnNCOM + 𝜷̂𝟒 lnIMMU + 𝒖̂𝒊

According to the estimated result from Stata using the Ordinary Least Squares (OLS) method, we obtained the Sample Regression Function (SRF) as below:

the dependent variable thus is necessary So long as the influential omitted variables are exluded from the true model, OLS regression will produce unbiased estimates

Conducting Ramsey’s RESET test for omission of influential variables

ovtest (Choose Level of confident = 5%)

{ 𝐻0: 𝑀𝑜𝑑𝑒𝑙 ℎ𝑎𝑠 𝑛𝑜 𝑜𝑚𝑖𝑡𝑡𝑒𝑑 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠

𝐻1: 𝑀𝑜𝑑𝑒𝑙 ℎ𝑎𝑠 𝑜𝑚𝑖𝑡𝑡𝑒𝑑 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠F(3,239) = 0,76

to the input data can lead to large changes in the model, even resulting in changes of sign

of parameter estimates

Using Variance Inflation Factor (VIF) to test for multicollinearity by running vif function

in STATA, we get the result:

 VIF of 3 independent values are smaller than 10, multicollinearity does not occur between one and other independent variables

 Conclusion: According to the test result, the model does not face multicollinearity phenomenon

3.2.3 Heteroskedasticity

While studying classic linear regression model, one of the basic assumptions is that the variance of each Ui in the condition that the given value of X is unchanged,

which means: var (U i |X i ) = var (U j |X j ) = constant

However, in fact, because of the nature of health-economics relationship, the method of gathering and processing data, the hypothesis is violated causing heteroskedasticity, which leads to an inefficient OLS estimators Therefore, the testing is no longer reliable

• One way to detect heteroskedasticity is graphical examination of residuals Using

the command rvfplot, yline(0) in STATA we can get the graph between residuals

and fitted values of dependent variable lnLE

It can be seen that the variance of residuals is decreased with fitted values of dependent variable lnLE, which indicates that this dataset faces Heteroskedasticity problem

• Another way to check the model’s error is using the White test to explore relationship between squared residuals and all of the dependent variables

Conducting White test for detecting Heteroskedasticity

imtest,white function with 5% Level of confidence