Financial vs Education Sector in Developing a Nation - Research Proposal Example

Stimulus: The last one Table of Contents Stimulus: The last one Table of Contents 2 Financial versus Education Sector in a developing nation 3 DataCollection 3 Methodology of Research 4 Data Analysis 6 Conclusion 8 Reference 9 Appendix 10 Financial versus Education Sector in a developing nation Financial sector stability is essential to ensure development of a nation. It assures a high leverage in its domestic as well as international transactions. On the other hand, equal attention must be bestowed upon the education sector of a nation as well for a progress in the quality of human resource within the nation. But the dilemma becomes intense when the sum allotted for improvisation purposes is fixed and needs to be spent wisely. As far as Hirschman’s unbalanced growth strategy is concerned, a big push is essential to shove an economy towards the path of development. But the direction of big push, or rather the sector chosen for the big push must be a significant one for a developing nation and this decision is made by considering the trickle-down effect of a change. In simpler words, the sector that has the largest trickle-down effect is chosen for the big push. The question that might arise at this point is that, how can the trickle-down effect be measured. Given that an inflow of funds in both the sectors can lead to an economic growth that again will be mirrored in the nation’s changing GDP, any development will solely be reflected in the rate at which the GDP of the nation is changing. Hence, the main objective of this paper is to find out how far the GDP of an average developing economy changes for changes in either secondary education enrolments in the nation or improvising the financial position of the nation. The variable found to have the more significant effect upon the dependent variable or rather the change in the GDP growth rate of the nation, will be the one where the government or rather the Finance Ministry will choose for investment and ensure a path of perpetual economic growth. Data Collection In order to carry on with the analysis, a sample of 50 nations have been collected, a list of which has been provided in the appendix to the chapter. Data on a number of macroeconomic variables had been gathered to formulate an economic model presented in the subsequent section. These variables are – per capita GDP (constant prices, following chain series) for the years 1990 and 2005 (ypc90 and ypc05, respectively), consumer price index (CPI) for the years 1985 and 1990 (cpi85 and cpi90), government share of real GDP per capita for the year 1990 (govgdp), ratio of sum of exports and imports to GDP for the year 1990 (open), ratio of private credit by deposit money banks and other financial institutions to GDP in 1990 (credit) and percentage of secondary school age population enrolled at secondary school in 1990 (seced). The steps that would follow this data collection process have been summarized in the following section. Methodology of Research The data been collected would be manipulated so as to form some other function of the variables and thus simplify the model formation, which will help to further carry on with the analysis. The economic model so formed along with the definitions of the variables being used in it is presented below in the form of a regression model. dlypci = 1 + 2lypc90i + 3lsecedi + 4govgdpi + 5openi + 6infli + 7crediti + ui Where, lypc90 = natural logarithm of ‘ypc90’, dlypc = difference between natural logarithms of ‘ypc05’ and ‘ypc90’, lseced = natural logarithm of ‘seced’, infl = difference between the natural logarithms of cpi85 and cpi90, govgdp, credit and infl have already been defined above. The estimation of the regression model would be followed by hypothesis testing, which is supposed to find out the significance of the coefficients being estimated. Hypothesis testing again need not be in the context of all the coefficients included in the above study and rather would be conducted only for three of them, apart from a joint significance testing of all the variables. The relevant hypotheses to be tested are – H01: H01:2=3=4=5=6=7=0 against H11:j0 for at least one j(2...7), using a significance level of 0.05. This one needs to be tested with the help of F-statistic, obtained from the ANOVA table. H02:2=0 against H12:20 using a significance level of 0.05. H03:3=0 against H13:3>0 using a significance level of 0.05. H04:7=0 against H14:7>0 using a significance level of 0.1. The remaining three hypotheses have to be tested with the help of Student’s t-statistic for each individual estimated coefficient. The procedure for hypothesis testing has been briefed as under – If the p-value or the level of significance of the estimated statistic exceeds the specified level of significance (0.05 in the first three cases and 0.01 in the last one), the respective null hypothesis cannot be rejected at the given level of significance and If the p-value or the level of significance of the estimated statistic is lower than the specified level of significance (0.05 in the first three cases and 0.01 in the last one), the respective null hypothesis can be rejected at the given level of significance. After the episode of hypothesis testing is over, the next step will be to conduct diagnostic checks to ensure that the regressed model is a relevant one; technically speaking, converge to the rules of a Classical Linear Regression Model framework. There are basically three assumptions which if satisfied can guarantee that the CLRM rules have not been hampered, namely – linearity, normality and homoscedasticity. To find out whether the assumptions are satisfied or not, the predicted residuals of the model are subjected to a number of tests, described below. If the residuals emerge successful in all of them, the model is considered to be complying with the CLRM assumptions. Among a large number of statistical tests, three have been chosen for three different purposes. Linearity check is done with the help of plotting a scatter diagram of the predicted residuals against the fitted values of the dependent variable. Normality check is popularly done with the help of estimated skewness and kurtosis checks. The relevant null and alternate hypotheses in this context will be, H05: Normality and H15: Non-Normality respectively. Homoscedasticity check is made with the help of White’s test. The relevant null and alternate hypotheses in this context will be, H06: Homoscedasticity and H16: Heteroscedasticity respectively. The statistics for all the above mathematical tests follow chi-square distribution, so that the results of the estimated statistics come in the form of estimated χ2 only. The rules to accept or reject the relevant null hypotheses are similar to those specified before. The level of significance with which to compare the level of significance of the estimated statistics is considered to be 0.05. In case there is a discrepancy in the estimated values, other measures are taken. But, as such an instance does not arise in the present study further research into the area is omitted. All the calculations in the paper have been conducted with the help of the statistical software STATA. Data Analysis The model being constructed previously has been estimated with the help of 50 data points corresponding to 50 different nations in the world. dlypci = 1.91 – 0.20lypc90i + 0.25lsecedi + 0.01govgdpi + 0.00openi + 0.11infli + 0.21crediti + ui (2.88) (2.67) (2.59) (1.60) (0.10) (1.08) (1.52) Details about the regression results have been presented in the appendix to the chapter. Consulting those details, it is found that, Estimated F-statistic = 1.77. At 6, 43 degrees of freedom, the probability value of the estimated statistic is, 0.1289 > 0.05. Thus, the relevant null hypothesis (H01) cannot be rejected at 5% level of significance. This implies that the variables taken together cannot explain variations in the regressed model significantly. However, nothing can be concluded about the significance of the estimated coefficients as such, since it might be that, there is need for more explanatory variables to make the F-statistic significant. Again, the p-values of the estimated t-statistics for the variables, ‘lypc90’, ‘lseced’ and ‘credit’ imply the following. P-value of estimated β2 = 0.011 < 0.05, P-value of estimated β3 = ½ x 0.013 = 0.0065 (for one-tailed test) < 0.05 and P-value of estimated β7 = ½ x 0.136 = 0.068 (for one-tailed test) > 0.01 Thus, the null hypothesis H02 and H03 cannot be rejected at 5% level of significance, but, the null hypothesis, H01 can be rejected at 1% level of significance. Hence, it is evident that the variables ‘lypc90’ and ‘lseced’ can play significant roles in explaining variations in the dependent variable ‘dlypc’, while, the variable ‘credit’ plays no such role. Thus, the contribution of the first two variables can be considered to be a significant one. However, nothing could be asserted as far as diagnostic checks are conducted. The results of the checks (details provided in the appendix), have been presented below – Linearity of the residuals can be confirmed after witnessing the trend in the movements of the estimated residuals. Normality of the residuals can be confirmed after studying the chi-square statistic corresponding to the Skewness and Kurtosis tests. Estimated chi-square statistic = 0.07 and the corresponding p-value of the estimated statistic = 0.9638. The corresponding null hypothesis, aimed at checking normality, H05 thus cannot be rejected at 5% level of significance. It implies that the assumption of normality of the estimated residuals remain undisturbed in the present case. Homoscedasticity of the residuals can be assured studying the respective estimated chi-square statistic of the estimated White’s Test, which is equal to 15.91. The p-value or the level of significance corresponding to this estimated statistic is, 0.9547, which is obviously greater than 0.05, the specified level of significance. Thus, the null hypothesis in this context cannot be ruled out at 5% level of significance. Hence, the assumption of homoscedasticity of the predicted residuals needed to satisfy the CLRM rules remains undisturbed. Since all three assumptions in the present context are found to be satisfied by the predicted residuals, so it can be said that the regression model so estimated is rather a sound one and complies with CLRM. Since the assumptions are already satisfied by the predicted residuals, there is no need for any further measures to remove discrepancies. Conclusion The present study has empirically found that it would be wiser for the Finance Ministry of the nation to infuse its resources in the secondary education segment of the nation. This is because, the sector is found to be affecting the dependent variable, change in GDP growth rate, more significantly than an infusion of the same sum of money into the financial sector of the nation. In fact, the exact figure of the estimated coefficient corresponding to ‘lseced’ is approximately equal to 0.25, implying that the change in the growth rate of GDP will be increased by 0.25% for 1% rise in the flow of funds to secondary education segment of the nation. On the other hand, though the statistic is equal to 0.21 in case of ‘credit’, the effect of the latter is an insignificant one, suggesting that the government should choose the former sector for improvisation, rather than the latter sector. Reference Heston, A., Summers, R. and Aten, B. (August 2009) Penn World Table Version 6.3, Center for International Comparisons of Production, Income and Prices at the University of Pennsylvania. Appendix Regression Results Normality Results Homoscedasticity Results Read More