How Could Age Dependency Ration Influence The Savings And Following Capital Outflow.
The aim of my work it to analyze how the age dependency ration influences the saving. According to the definition of the World Bank, Age dependency ratio is the ratio of dependents-people younger than 15 or older than 64--to the working-age population--those ages 15-64. So the questions of my research is to find out how the quantity of pre-working-aged children and retired influences the saving per capita. Would people save more in their working age in order to save money for retirement period or to invest -or they will earn and spend. Another issue is to see if countries with huge level of ...view middle of the document...
It means that people try to save some money to spend it after they get retired. The main aim of his model was to explain how the saving and consumption are formed. But the weak part of his research could be the inflation and the expectance age of retired. A.Tylecote, the British economist, in his research showed that these effects of inflation and some another economical indicators could remarkably influence the demand function and could provoke economics cyclic oscillations taking the changing in demographical structure of the society into account. Then I read the book , written by Obstfeld M. and K. Rogoff. In their intertemporal account balance models of opening economies and closed economies and two period economies they gained that indicators such as real interest rate, substitution effect, income effect and welfare effect influence savings in the most extent.
In this way , thanks to the empirical analyses I want to see whether the influence of ADR on saving is significant or not. Also I want to see the individual effect of every country on savings and is it significant or not. In addition I will define whether ADR in general is necessary or we could use only retired people or only under labour age children in regression model.
Firstly , I will start by considering the regression with not fixed effect . I defined the panel structure of variables and use the command : Xtreg.
The regression with no fixed effect differ from fixed effect that in equation , ui- normal distributed variable, not correlated with εit . We can see the estimation of ADR_GEN and another variables, but not exactly to each country. According to that regression, the coefficients by real_ir and gdp_pc are significant. Sigma u- it is the standard deviation of coefficients by dammi- variables by countries. Dispersion coefficients is higher than dispersion residues. It means that influence of individual country is high . When the real interest rate will rise by 1 that will lead to the fall in savings by -0,14 and if ADR_GEN rised - fall of savings by -0,09, and meanwhile the GDP increase provokes slight rise. The correlation between regressors and between regressor and residues could give us the proposal that the estimates are biased and because of that it is more efficient to use fixed effect model. In the fixed effect model corr(ui, Xi) might be not equal to 0, because here ui is not unbiased error .here it is included in the regressive function.
where Zi - not observed variables , which didn't change for i-country. Every country will have each own meaning- will not change in dynamic(no t-index)
So we can estimate this model : . In this case, every country has the same β1, but they each of them has its own constant variable.
I make the fixed-effect regression , where I fix effect for each country
2-93 - it is a country code, where reference country is Afghanistan by alphabet.
Moreover, it should be noted that the dependence from the...