CHAPTER FOUR
B. Vector Error correction Model for Investment
Inflation as an indicator of macroeconomic instability is also used in the long run analysis and the result showed that inflation deters investment significantly. That is, a percent increment in log of inflation deters investment (in log percentage) by nearly forty percent. It suggests that an instable macroeconomic environment is not conducive for investment. This may discourages entrepreneurs from putting their fund in the country so long as the inflation rate is higher (especially double digit inflation and beyond).
B. Vector Error correction Model for Investment
Since the variables in the investment equation are found to be cointegrated, we proceed to estimate the vector error correction model which represents both the long run and short run adjustments among the variables under study. The log changes in the relevant variables represent short run elasticity’s (alternatively, short run variation), while the error correction term (ECT) represents the speed of adjustment back to the long run relationship among the variables. A VECM is estimated beginning with the general over parameterized model. Then the VECM is subjected to a systematic reduction and diagnostic testing process until an acceptable
parsimonious model is obtained. In the process, all insignificant explanatory regressors with their corresponding lags are dropped until further reduction is rejected (Hendry, 1997).
In the short run dynamic equation, all weakly exogenous variables identified in the long run investment equation are entered in the right hand side of the model in their appropriate lagged difference form. In addition the error correction term with one period lag is also incorporated in the VECM.
Using the VECM specification (section 3.3.3), a short run dynamic equation is estimated for investment function. Dropping insignificant regressors from the specification (i.e. step-by-step elimination of insignificant regressors from the general VECM model) following the general to specific modeling strategy, a parsimonious result for investment is reported below.
The estimated coefficients of the VECM revealed that the signs of all variables are in line with the theoretical expectation. The result showed that investment is positively associated with both domestic (saving) and foreign (aid) capital. However, domestic saving promoted investment only in the short run; it remained an important source for financing investment and its positive influence is only a short run phenomenon. Foreign aid (lagged one period) also affects domestic investment positively and significantly in the short run.
Volatility of aid flow influenced investment negatively but found insignificant in the short run.
The result indicates that volatility of aid has a minimal effect on investment in the short run;
however, it has a deleterious effect in the long run since it makes long run development planning difficult and creates uncertain environment on investment activity. This pointed that the deleterious effect of uncertainty of aid on investment is only a long run phenomenon. The estimated short run investment equation also shows that debt servicing has a negative contribution. This indicates that debt servicing seriously affects capital formation activity but its impact is limited to the short run. The other variable used as a proxy for macroeconomic instability is inflation (regardless of the fact that it remained under control in the Derg period).
The result revealed that inflation works against investment in the study period in Ethiopia. Such effect is transmitted indirectly through the measures that are taken to put the pressure under
control, which in fact has a wide spread effect not only on investment but also on other macro-variables. Also inflation has a negative effect on investment through discouraging entrepreneurs which works through the increment in the cost of production. Finally, the coefficient of the error correcting term is found to be statistically significant. It points that 36.5 percent of the disequilibrium in the previous period is corrected in one year. Therefore, it takes 2.7 years to adjust for the disequilibrium to the long run path.
Table viii(F) Result for the Dynamic Investment Equation
variable Coeff. t-value
constant -2.249*** -5.8
DlA_1 0.389*** 3.68
DlS_1 0.212*** 3.46
DlDS -0.052** -2.36
DlDS_1 -0.159*** -5.45
DlUA -0.488 -1.24
DlINF -0.070*** -3.99
ECT_1 -0.365*** -5.78
Note: *** and **denotes significance at 1 % and5 % level respectively. The optimal lag length is determined at lag length of two using Akakie Information Criteria (AIC).
R^2=0.5634
F(7,29)= 5.35 [0.0005]***
Diagnostic Tests DW =1.6765
ARCH(1,2) test: Chi2(2)=0.206 [0.9022]
AR(1,2) test :F(2,27)=0.737[0.4880]
Hettest: F(1,35)=1.68[0.2032]
Normality test: Chi2(2)=0.668[0.716]
RESET test: F(3,26)=0.28[0.8414)
The goodness of fit of the model is quite acceptable-the independent variables explaining 56 percent of the variation in the dependent variable. The null hypothesis of the joint insignificance of the coefficients of all explanatory variables is rejected by the F-statistic. The different kinds of diagnostic tests performed on the model indicated no problem on the subject of regression analysis. All the tests failed to reject the null hypothesis at any conventional significance level.
That is, the null of constant variance (homoscedastic errors) is not rejected as given by the Breusch-Pagan test for heteroscedasticity. The Breusch-Godfrey LM test for autocorrelation also shows that there is no serial autocorrelation. Furthermore, the LM test for autoregressive conditional heteroscedasticity indicated that the null of no ARCH effects is not rejected. In addition, the Ramsey’s (1969) RESET test for model misspecification does not reject the null of no functional misspecification in the estimated investment equation. Lastly, the Jarque-Bera test for normality indicates that the errors are normally distributed since the null hypothesis of normally distributed error terms is not rejected at any conventional level. Thus, the various diagnostic tests conducted indicate that the overall fit of the model is acceptable enough statistically.