Econometrics report factors affecting the human development index (hdi) of countries around the world in the period 2015 2019

Table of content 1. Introduction 4 2. Overview of the topic 5 2.1. Theoretical framework 5 2.1.1 Definition of HDI 5 2.1.2. Formula of HDI 6 2.2. Literature Review 6 3. Research methodology 9 3.1. Methodology 9 3.2. Empirical Model 9 3.2.1. Model Specification 9 3.2.2. Variables Description 10 3.3. Data 11 3.3.1. Source of data 11 3.3.2. Descriptive statistics and variables interpretation 12 3.3.3. Correlation matrix between variables 12 4. Quantitative analysis 13 4.1. Choosing the Estimated Model 13 4.1.1. Breusch Pagan test 13 4.1.2 Hausman test 14 4.1.3. Testing for multicollinearity 17 4.1.4. Testing for heteroskedasticity 18 4.1.5 Testing for serial autocorrelation 19 4.2. Fixing the model 20 References 24 Data 27   1. Introduction Human development has been a topic of interest for scholars and organizations worldwide for many years. In the 1980s and 1990s, the United Nations Development Programme (UNDP) recommended human development as a necessary theory for socioeconomic development and presented it in their annual report. The UNDP believes that the true purpose of a nation is its people, and the goal of development should be to create a favorable environment that enables people to enjoy long, healthy, and creative lives. development is a humancentered perspective that emphasizes investing in growing human potential in areas such as education, healthcare, and skills. This investment enables humans to work creatively and effectively, without borders. It also means ensuring that economic growth is distributed widely and fairly.

FOREIGN TRADE UNIVERSITY FACULTY OF INTERNATIONAL ECONOMICS

Ha Noi, June 2023

Table of content

3.3.2 Descriptive statistics and variables interpretation

3.3.3 Correlation matrix between variables

4 Quantitative analysis

4.1 Choosing the Estimated Model

4.1.1 Breusch - Pagan test

4.1.2 Hausman test

4.1.3 Testing for multicollinearity

4.1.4 Testing for heteroskedasticity

4.1.5 Testing for serial auto-correlation

4.2 Fixing the model

References

Data

1 Introduction

Human development has been a topic of interest for scholars and organizationsworldwide for many years In the 1980s and 1990s, the United Nations DevelopmentProgramme (UNDP) recommended human development as a necessary theory forsocioeconomic development and presented it in their annual report The UNDP believesthat the true purpose of a nation is its people, and the goal of development should be tocreate a favorable environment that enables people to enjoy long, healthy, and creativelives development is a human-centered perspective that emphasizes investing in growinghuman potential in areas such as education, healthcare, and skills This investmentenables humans to work creatively and effectively, without borders It also meansensuring that economic growth is distributed widely and fairly

Recognizing the value of the HDI, our research group identified the need for astudy on the factors affecting the HDI Our goal is to comprehensively and fully evaluatethe factors affecting human development in relation to the development of the economy.Therefore, our research team chose the topic “Factors affecting the human developmentindex (HDI) of countries around the world in the period 2015 - 2019” for our report

The general purpose of this study is to comprehensively and fully evaluate thefactors affecting human development in relation to the development of the economy byemploying 1 dependent variable (HDI) and 8 independent variables (FER, InGNI, GINI,MYS, InLEB, GOV, POVER, Developed)

Meanwhile, this study specifically aims to: (1) Collecting and processing data onvariables directly related to the HDI is essential to understanding the factors that affecthuman development progress in different countries This involves gathering data onindicators such as MYS - Mean Years of Schooling Index, GOV - Governmentexpenditure on education, …; (2) Building a suitable and impeccable model to accuratelyanalyzing the data collected; (3) Introducing novelty and overcoming some limitations of

a few previous studies on the HDI; (4) Offering recommendations and solutions toseveral HDI-related problems is the final objective of the research

The research subjects of the report are countries around the world, includingdeveloping and developed countries, namely: Albania, Armenia, Austria, Belgium,Bulgaria, Belarus, Bolivia, Brazil, Columbia, Costa Rica, Cyprus, Czechia, Denmark,

Dominican Republic, Ecuador, Spain, Estonia, Finland, France, United Kingdom,Georgia, Greece, Croatia, Hungary, Indonesia, Ireland, Italy, Kyrgyz Republic, Lithuania,Luxembourg, Latvia, Malta, Netherlands, peru, Portugal paraguay, Romania, RussianFederation, Serbia, Slovenia, Sweden, Thailand, Ukraine, Uruguay, United States Thestudy will centralize on factors influencing the HDI index of these countries, in order tofind lessons learned and provide policy recommendations In this report, we focus on theperiod from 2015-2019, with the subject of countries over the world.

In this report we will use a quantitative method Regression method: we uses theregression method for panel data to evaluate the estimated factors that affect the HDIindex of 45 countries during the period 2015-2019 The model results serve as the basisfor recommendation

The report include 4 main chapters:

Chapter 1: Introduction

Chapter 2: Overview of the topic

Chapter 3: Research Methodology

Chapter 4: Quantitative Analysis

2 Overview of the topic

2.1 Theoretical framework

2.1.1 Definition of HDI

The Human Development Index (HDI) is a comparative and quantitative index ofincome level, literacy rate, life expectancy, and other factors of countries worldwide HDIprovides a general overview of a nation's development This index was developed in 1990

by Pakistani economist Mahbub ul Haq and Indian economist Amartya Sen

According to Mahbub ul Haq, who laid the foundation for the HumanDevelopment Reports, "The basic purpose of development is to expand people's choices

In principle, these choices can be infinite and can change over time, but they ofteninclude: greater access to knowledge, better nutrition and healthcare, a guaranteedstandard of living, protection from violence and crime, leisure time, freedom in culturaland political life, and participation in community activities The goal of development is to

create an environment that enables individuals to enjoy a healthy, long, and creative life"(UNDP, 2010).

The HDI index attracts the attention of policymakers, the media, and governmental organizations because it not only focuses on purely economic statistics butalso highlights achievements in human development This index emphasizes that it is thehuman beings and their capabilities, rather than just economic growth, that constitute thehighest criteria for assessing a country's level of development

non-2.1.2 Formula of HDI

The Human Development Index is a composite index that measures a country'sachievements in three dimensions of human development: health, education, and decentstandard of living Health is measured by the life expectancy at birth (I a) The educationlevel index (I e) is a composite of two indicators: the average years of schooling (I my) andthe expected years of schooling (ley) The standard of living index (I¿) is measured bythe Gross National Income (GNI) per capita, calculated in USD using purchasing powerparity (PPP USD) (UNDP, 2010)

In terms of formula, we have:

HDI=√3 I a ∗I e ∗I my

HDI ranges from 0 to 1 The more it approaches 1, the higher human developmentlevel reaches On the other hand, when the index approaches 0, the trend of humandevelopment tends to decrease According to Human Development Report (2021), everycountry is categorized into 4 groups in regard of HDI index:

● HDI <0.55: countries with low HDI

2.2 Literature Review

UNDP (1990) provided an index to measure human development - the HumanDevelopment Index (HDI) It must be clearly stated that the concept of humandevelopment is much broader than the measurement of human development As society

continues to develop and data collection capabilities improve, it has been recognized thatthe previous indices do not fully capture human development Therefore, in 2010, UNDPmade some adjustments in the calculation of the HDI While still measuring the threedimensions of economy, education, and health, the HDI (since 2010) underwent changes

in the measurement indicators, the value computation of component indices, and theformula for calculating the HDI

In terms of indicators, in the field of education, the literacy rate index wasreplaced by the average years of schooling for adults The gross enrollment index acrossall levels was recalculated based on expected years of schooling (UNDP, 2010) Thechange in calculating the education index helps better reflect the development ofeducation, as the previous index focused too much on the literacy rate and failed tocapture human capital, especially in developed countries (Herrero, Martínez, & Villar,2012) In the economic field, GNI index is used instead of GDP because it moreaccurately reflects a country's real income For the life expectancy index, the previouscalculation had a maximum observed value of 85 and a minimum observed value of 25(years) In the new calculation method, the minimum observed value has been adjusted to

20 (years) Regarding the changes in the calculation formula, the HDI since 2010 is nolonger computed as the simple average of the education, income, and health componentindices, as in the previous calculation Instead, the new HDI value is calculated as thegeometric mean of these three component indices This new calculation method isconsidered to better reflect human development According to this method, if there isuneven development across the three domains of economy, education, and health (highdisparity between component indices), it will reduce the value of the HDI

In the research paper "Review of HDI Critiques and Potential Improvements" byMilorad Kovacevic (UNDP) in 2010, the following factors influencing the HDI wereidentified: life expectancy at birth, fertility rate, adult literacy rate, average GDP percapita, and the level of development Quantitative analysis results showed that all thesevariables were statistically significant and had an impact on the HDI

A cross sectional study conducted by Grzech, Patel and Walker (2016) using datafrom 188 countries found that life expectancy and education gives the most impact onHDI values implying these elements are the main contributors of HDI improvement.Byidentifying more relatable and specific variables, this study is especially unique in that it

provides governments with a list of factors they can effectively target to improve theirHDI rankings.

The study "Examining the Relationship Between Human Development Index andSocio-Economic Variables: A Panel Data Analysis" by Jalil and Kamaruddin (2018)investigates the relationship of the three dimensions of human development depicted asthe socioeconomic variables on human development index The selected socio-economicvariables are mean years of schooling, expected years of schooling, life expectancy andhealth expenditure, GDP, GDP per capita With the short-panel balanced data and panelfixed effects model, the study showed a significant positive effect on the HDI of fourfactors

Çaglayan-Akay and Van (2017) investigate factors that influence the level ofeconomic development in 130 countries by using Bayesian ordered probit model In theirstudy, seven selected independent variables such as internet users, rural population, GDP,life expectancy at birth, health expenditure, and share of expected years of schooling areused to investigate the impact of these variables towards the HDI Different results werefound in both short term and long term periods For short term period, the variables thathad affected the HDI positively are all those selected independent variables However, inthe long term, it is found that only GDP, health expenditure, the share of expected years

of schooling and internet users have positive impact towards the HDI while lifeexpectancy at birth and rural population showed a contrary result The result has shownthat the expected years of schooling is one of the selected variables that have positiveimpact for both short and long run periods towards HDI

In 2016, Smit Shah, in his report "Determinants of Human Development Index: ACross-Country Empirical Analysis", considered seven independent variables: lifeexpectancy at birth, adult literacy rate (aged 15 and above), average GDP per capita,fertility rate, income inequality coefficient Gini, CO2 emissions, and inflation rate tostudy the factors influencing the HDI The research results indicated that only theinflation rate did not have an impact on the HDI, while the remaining six variables wereinfluential

Based on the findings of these studies, as well as previous publications by UNDPand other studies, our group selected these variables below to analyze their impact on the

HDI of each country: GNI per capita, mean years of schooling, expected years of schooling, Gini coefficient, life expectancy at birth and level of development.

To be more precise, every variable in this model is expected to have a positiveimpact on HDI, which is the dependent variable in this model We hypothesize that:H1: GNI per capita have a positive effect on HDI

H2: Mean years of schooling have a positive effect on HDI

H3: Expected years of schooling have a positive effect on HDI

H4: Gini Index have a negative effect on HDI

H5: Life expectancy at birth have a positive effect on HDI

H6: Level of development have a positive effect on HDI

3.2 Empirical Model

3.2.1 Model Specification

Based on inherited previous studies and combined with the ability to collectresearch data for research solving the variables, we estimate the determinants ofeconomic growth with fixed effects regression model

HDI i =β0+β1.Gini¿+β2.¿¿+ β3 MYS¿+ β4.EYS¿+β5.GNIPC¿+β6.Developed¿+u¿

In which:

HDI: dependent variable

Gini i , ¿i , MYS i, EYS i , GNIPC i , Developed i: independent variables

Developed i: dummy variable; developed i=1: developed countries; developed i=0: developingcountries

3.2.2 Variables Description

From our theoretical base, there are three key dimensions measuring humandevelopment Therefore, in our model, the six indicators will be an independent factor ofthe regression model

Table 1: Definition of variables used in the model

sign Dependent variable

Gini Gini Index

Gini index measuring incomeinequality Higher Gini leads to biggap in living standard, therebyinverse relationship with HDI

expectancy

Average life expectancy of

MYS Mean Years

of Schooling

Average years of schooling ofabove-25-year-old residents in acountry

ExpectedYears ofSchooling

Expected years of schooling ofresidents illustrating theimprovement in education level in

be relatively positive to HDI

● Developed = 1:

developedcountries

● Developed = 0:

developingcountries

HDI, MYS, EYS originates from UNDP meanwhile the statistics of Gini, LE, GNIPC are

received from World Bank

3.3.2 Descriptive statistics and variables interpretation

In order to estimate the model determined the dependent relationship between theindicators of HDI in 45 countries around the globe, we decided to choose the researchsample in the 5-year period from 2015 to 2019 with a total of 225 observations

A statistical description of the indicators used in the regression model can be seen

in Table 2 below

Table 2: Descriptive statistics

Source: The authors (2023)

3.3.3 Correlation matrix between variables

Table 3 below clearly illustrates the correlation, which is the degree of linearassociation between the dependent variable and the independent variables

Table 3: Correlation matrix between variables

LE 0.8487 -0.3947 1

GNIPC 0.9073 -0.4286 0.7581 0.5838 0.6337 1

Source: The authors (2023)

In general, the degree of correlation between independent variables is not high,ranging from 0.3102 to 0.7581 in absolute value, therefore we can predict that the modeldoes not suffer from multicollinearity

4 Quantitative analysis

4.1 Choosing the Estimated Model

In this part, we use Stata and dataset in order to examine and find out the mostsuitable choice among three models: Pooled OLS model (POLS) Fixed Effects model(FE) and Random Effects model (RE)

4.1.1 Breusch - Pagan test

We apply Breusch - Pagan test to choose between FE/RE or POLS for the substantialdistinction over units by using the commands xtreg and xttest0:

xtreg HDI Gini LE MYS EYS GNI i.Developed, re

Breusch and Pagan Lagrangian multiplier test for random effects

HDI[STT,t] = Xb + u[STT] + e[STT,t]

H0:cov(ai,Xit)=0

H1:cov(ai,Xit)≠0

(where H0 are null hypothesis and H1is alternative hypothesis)

We run the following commands:

xtreg HDI Gini LE MYS EYS GNI i.Developed, fe

Random-effects GLS regression  Number of obs  =  225

xtreg HDI Gini LE MYS EYS GNI i.Developed, re

(V_b-V_B is not positive definite)

As a result from the Hausman test, with 5% level of significance,

we have: Prob > chi2 = 0.0000 < 0.05 ⇒ Does not reject Ho

Conclusion: We choose the Fixed Effects model (FE) to estimate the data.

4.1.3 Testing for multicollinearity

Statistical consequences of multicollinearity include difficulties in testingindividual regression coefficients due to inflated standard errors Thus, you may beunable to declare an X variable significant even though (by itself) it has a strong

relationship with Y We use command VIF in order to test whether the model hasmulticollinearity or not.

reg GGI TGDP EPI GPDG

From the above table, we have: Mean VIF = 1.31 > 10

Therefore, we can conclude that the model has no multicollinearity.

4.1.4 Testing for heteroskedasticity

One of the assumptions made about residuals/errors in OLS regression is that theerrors have the same but unknown variance This is known as constant variance orhomoscedasticity When this assumption is violated, the problem is known asheteroscedasticity It has been shown that models involving a wide range of values aremore prone to heteroskedasticity because the differences between the smallest and largestvalues are so significant

With Ho: the model has homoscedasticity

H1: the model has heteroscedasticity

First, we run the regression model:

xtreg HDI Gini LE MYS EYS GNI i.Developed, fe

Conclusion: The Random Effects Model suffers from heteroskedasticity

4.1.5 Testing for serial auto-correlation

We run command xtserial to test if the model is suffering from auto-correlation or not

xtserial HDI Gini LE MYS EYS GNI

Wooldridge test for autocorrelation in panel data 

H0: no first-order autocorrelation 

With a 5% level of significance, we have Prob>F = 0.0000 < 0.05 Therefore the FEmodel suffers from auto-correlation.

1.6 Testing for cross section correlation

xtreg HDI Gini LE MYS EYS GNI i.Developed, fe

Pesaran's test of cross sectional independence = 1.948, Pr = 0.0514

Average absolute value of the off-diagonal elements = 0.574

As we can see, the CD test strongly rejects the null hypothesis of no cross-sectional dependence

Conclusion: The model suffers from cross-sectional dependence.