Effect of Workers’ Remittances on Private Savings Behavior in Pakistan: Empirical Results
The results of DF and ADF Unit Root tests for checking the stationary of the data are shown in Table1a. In order to scrutinize the integrating level of variables, standard tests like DF and ADF (Dickey & Fuller, 1979) are applied. The results of Dickey Fuller and Augmented Dickey Fuller tests show that the three variables are stationary at level (RRID, FDI and LTRM) because null hypothesis of the existence of unit root is rejected showing stationary of series I existed in long run among variables. If it lies under the lower critical bound, then null hypothesis will be accepted. Finally, if it lies within the critical bound values, the result seemed inconclusive.
When two variables has been co integrated in long run, then in the second step, the error correction mechanism will be used to examine the dynamics of the model in short run. This technique was first analyzed by (Sargan, 1964) and after this was famous by (Engle and Granger, 1987).
The basis of error correction model was Granger representation theorem. In error correction model, the short-term dynamics of the variables in the system were influenced by the deviation from the equilibrium. ECM model was best for co-integration as it included both long-run and short run information and ECMs were formulated in the terms of first differences, which eliminated the trends from variables and resolved the problem of spurious regression and it also measured the correction from disequilibrium of the previous period. The remaining variables (LRGDPC and LPS) are non-stationary at level but become stationary after taking their first difference.i.e. I.
First of all, at different lags on the first difference of each variable, F-Statistics is computed for the joint significance of variables in long run. When 4 lag is imposed, there is a strong evidence of existence of Co-integration among the variables because the F-Calculated is F = 5.1358, which is greater than the critical value of the F-Statistics of the upper level of the bound (3.646) calculated by (Pesaran, et al.) at the 5 percent significance level. It is concluded from F-statistics that there exists a long run relationship among the variables. payday loans no credit check
Given the existence of long run relationship among the variables, ARDL model is estimated to find the long run and short run dynamics of the variables in equation. The long run and short run results are reported in table 2 and table 3 respectively. The long run statistics shows that coefficient of RRID (Real deposit Rate) is -.0165. It showed that a 1 percent increase in real rate deposit rate tend to decrease the private savings by 1.65 percent. The coefficient of LRGDPC (log of real GDP Per Capita) is 2.62 which show positive effect of GDP per capita on private savings.
The coefficient of LTRM (Log of total worker remittances) is.619 which shows positive relationship between worker remittances and private savings. It showed that a 1 percent increase in worker remittances increase the private savings by.62 percent approximately. While coefficient of FDI (Foreign direct investment) is -.7293E-5 which reflects negative impact of foreign direct investment on private savings and are highly significant at 1 percent level.
Table 1a: DF & ADF Unit Root Test at Level
|Variables||Without trend||With trend||Order of Integration|
|LRGDPC||-0.925||-1.454||1 non stationary|
|LPS||-1.569||0.621||1 non stationary|
Table 1b: DF & ADF Unit Root Test at 1st Difference
|Variables||Without trend||With trend||Order of integration|
Table 2: Long Run Coefficients of ARDL Based on Schwarz Bayesian Criterion Dependent Variable LPS
Table 3: ARDL Error Correction Mechanism (Short run Dynamics) Dependent Variable DLPS