The correlation plot indicates that life expectancy at birth, expected years of schooling, mean years of schooling, and GNI per capita exhibit strong positive correlations with the HDI. Conversely, GDP has only a weak positive correlation. This suggests that the HDI is more influenced by components like education and health (life expectancy and schooling measures) than GDP. While GDP contributes, its impact on HDI is not as dominant as other dimensions.
Show the code
data <-read_excel("HDR23-24_Statistical_Annex_HDI_Table.xlsx")data$`Human Development Index (HDI)`<-as.numeric(data$`Human Development Index (HDI)`)data$`Expected years of schooling`<-as.numeric(data$`Expected years of schooling`)data$`Mean years of schooling`<-as.numeric(data$`Mean years of schooling`)data$`Gross national income (GNI) per capita`<-as.numeric(data$`Gross national income (GNI) per capita`)data$`GNI per capita rank minus HDI rank`<-as.numeric(data$`GNI per capita rank minus HDI rank`)data$`HDI rank`<-as.numeric(data$`HDI rank`)# Get GDP (current US$) for the latest available yeargdp_data <-WDI(indicator ="NY.GDP.MKTP.CD", country ="all", start =2022, end =2022)gdp_data <- gdp_data %>%select(country, NY.GDP.MKTP.CD)colnames(gdp_data) <-c("Country", "GDP")master <- gdp_data %>%inner_join(data, by ="Country")master <-na.omit(master)corrplot::corrplot(cor(master[,c(2:7)]))
The scatter plot between HDI and GDP shows a weak positive relationship. There is noticeable dispersion around the regression line, indicating variability in HDI that is not well explained by GDP alone. GDP alone does not fully account for variations in HDI across states or countries. Other factors, likely the educational and health dimensions, play a significant role.
Show the code
# Scatterplot of HDI vs. GDPggplot(master, aes(x = GDP, y =`Human Development Index (HDI)`)) +geom_point(color ="blue", size =2) +geom_smooth(method ="lm", color ="red") +labs(title ="HDI vs. GDP per Capita",x ="Gross Domestic Product per Capita",y ="Human Development Index (HDI)") +theme_minimal()
The rank comparison reveals a strong linear relationship between GDP and HDI rankings, with most points lying close to the 45-degree line. While GDP rank strongly aligns with HDI rank, this does not imply causation. The correlation might be driven by the inclusion of GNI (closely related to GDP) in the HDI calculation.
Show the code
# Rank comparisonmaster <- master %>%mutate(GDP_Rank =rank(-GDP),HDI_Rank =rank(-`Human Development Index (HDI)`),Rank_Difference = GDP_Rank - HDI_Rank)# Plot ranks comparisonggplot(master, aes(x = GDP_Rank, y = HDI_Rank)) +geom_point(color ="purple", size =2) +geom_abline(slope =1, intercept =0, linetype ="dashed", color ="gray") +labs(title ="Rank Comparison: GDP vs. HDI",x ="Rank by GDP per Capita",y ="Rank by HDI") +theme_minimal()
The regression model indicates a statistically significant positive relationship between GDP and HDI (p-value = 0.013). However, the coefficient of determination (R2) is very low (3.77%), suggesting that GDP explains only a small fraction of the variance in HDI. The residuals indicate that the model leaves substantial unexplained variability in HDI. While GDP has some predictive power for HDI, its influence is limited, and most of HDI’s variability stems from other factors.
The intercept represents the value of HDI when GDP is zero. In this case, the intercept of 0.7261 suggests that when GDP is zero, the model predicts an HDI of approximately 0.7261. Although this doesn’t have a practical interpretation (since GDP cannot be zero), it is still part of the regression equation.
The coefficient is extremely small, which suggests that while there is a relationship between GDP and HDI, the effect of a one-unit increase in GDP on HDI is very small. In other words, even a large change in GDP (in actual terms) only leads to a very slight increase in HDI. Given the small value of the coefficient, it’s important to note that the magnitude of change in HDI due to GDP is likely negligible in practical terms.
Show the code
# Linear regression analysismodel <-lm(`Human Development Index (HDI)`~ GDP,data = master)stargazer::stargazer(model, type ="text")
The rank differences between GDP and HDI mostly fall between -50 and 50. This relatively narrow range suggests that GDP and HDI rankings are generally consistent. However, the differences highlight that HDI is not merely a disguised version of GDP; other components influence its ranking.
Show the code
# Histogram of rank differencesggplot(master, aes(x = Rank_Difference)) +geom_histogram(binwidth =1, fill ="orange", color ="black", alpha =0.7) +labs(title ="Distribution of Rank Differences",x ="GNI Rank - HDI Rank",y ="Frequency") +theme_minimal()
Based on the analysis HDI is not purely GDP disguised. While GDP has some influence on HDI (as evidenced by the rank comparison and regression analysis), its contribution is relatively minor compared to other dimensions like education and health. The strong correlations of HDI with life expectancy, expected years of schooling, and mean years of schooling underscore the multidimensional nature of HDI. The weak relationship between GDP and HDI in scatter plots and regression analysis further supports the notion that HDI captures more than just economic output. The HDI incorporates multiple dimensions of human development and is not simply a rebranded version of GDP. Although GDP (via GNI) is a component, the index gives significant weight to health and education, distinguishing it as a broader measure of development.
