Page 74 - rohana_journal_No_12-2020-final
P. 74
Research Journal of the University of Ruhuna, Sri Lanka- Rohana 12, 2020
The error correction coefficient, ∅, can be interpreted as the speed of adjustment to
long run equilibrium. This coefficient shows us how much of the adjustment to
equilibrium takes place in the current year. In order to ensure convergence towards
long run equilibrium sine of ∅ should be minus and significant. The results of the
short-run dynamic coefficients obtained from equation 03 are presented in Table 05.
Table 05: Results of Error Correction Model
Variable Coefficient Slandered T - Statistics P - Value
Error
LFDI(-1)) 0.3301 0.055 5.956 0.001
LCTR 1.1783 0.162 7.267 0.000
LCTR(-1)) -1.468 0.178 -8.233 0.000
LEXR 6.343 0.556 11.405 0.000
LFD 0.032 0.212 0.152 0.884
LGROT -1.281 0.105 -12.237 0.000
LGROT(-1)) -1.850 0.125 -14.784 0.000
LINF 1.502 0.088 17.160 0.000
LINFRA 2.090 0.197 10.589 0.000
LINFRA(-1)) 1.904 0.186 10.320 0.000
LOPEN 5.418 0.480 11.288 0.000
WAGIN) -1.279 0.319 -3.967 0.007
LWAGI(-1)) -3.488 0.405 -8.606 0.000
ECT(-1) -0.989 0.088 -11.243 0.000
Source: Author (2020)
It is worthy to note that the error correction term is negative and significant at 0%
level justifying the results of the cointegration model. The value of coefficient of
ECT is -0.9893, and this would mean that almost 98% of adjustment takes place
each year. This high value of the coefficient reveals that the speed of adjustment to
equilibrium is very fast after a shock.
The estimated model passes a diagnostic test against serial correlation (DW test and
LM test) and stability test of the cumulative sum of recursive residuals (CUSUM)
test. These test results suggests that the model is specified and valid for
interpretation of the results of the bound test for cointegration. Figure 05. depicts the
results of the CUSUM test and it clearly indicates that statistics fall within the bands
65