英语论文网

ora(1990)

The coefficient of LRY is significant at 1% and is positive. It is of the expected sign since an increase in the GDP of USA imply that the income of USA is increasing and hence demand for Mauritian exports should increase. A 1% increase in real GDP of USA lead to a 0.895% increase in export volume of Mauritius or differently stated, a 10% increase in the real GDP of US leads to 8.95% increase in volume of exports. Therefore it is one of the important determinants of the volume of exports to the USA. The same result was obtained by and Senhadi and Montenegro(1999), Marquez and McNeilly(1988), Sachs and Warner(1995), Frankel and Romer(1999) and Edwards(1998) in their work.

LRER is another major determinant of export volume. positive and significant at 1%. It has a coefficient of 2.75 which implies that a 1% increase in the real exchange rate leads to a 2.75% increase in exports. This is because the exchange rate is expressed in terms of Rupees per US dollar. An increase in the LRER is indicative of a depreciation of the Mauritian Rupee vis-?-vis the US dollar, since $1 can buy more Mauritian Rupees. Therefore an increase in LRER means that the Rupees is cheaper vis-?-vis the US dollar and thus Mauritian exports will be cheaper for US which will cause an increase in its demand. This result support the results obtained by Himarios(1985) and Warner and Kreinin(1983)

The coefficient of LINVG is positive and significant at 1%. This means that a 1% increase in investment as a percentage of GDP leads to a 1.54% in volume of exports to the USA. It is another major determinant of export volume. This is because in the long run, an increase in investment is likely to increase productivity and output. Patnaik and Chandrasekhar(1996) obtained the same results.

The variable LINF has a negative coefficient and is significant at 2%. The results indicate that a 1% increase in inflation rate will bring about a 0.226% decrease in volume of exports. The results obtained are in line with those obtained by Kotan and Saygili(1999) and Gylfason(1998).

Finally the R2 is 0.8785 which implies that the explanatory variables jointly explain 87.8% of the variations in the dependent variable. The Durbin-Watson statistic of 1.3, which is close to 2, indicates that there is no autocorrelation and since it is higher than the R2, means that there is no evidence of spurious regression.

4.6.2 短期的方程——4.6.2 The Short run equation

According to the Representation Theorem developed in Engle and Granger(1987), if the variables are cointegrated, it can be shown that the ECM will be of the following form:

?LREXPt= ?0 + ?1?LREXPt-2 + ?2?LVt-2 + ?3?LRYt-2 + ?4?LRERt-2 + ?5?LINVGt-2 + ?6?LINFt-2 + ?7?LCUt-2+ ?8 Rt-1 + ut

Where Rt-1 is the lagged error correction term and is the residual from the cointegrating equation. If the variables in equation 1 have a cointegrating vector, then Rt is I(0) and represent the deviation from the equilibrium in period t. Hence the coefficient of Rt, ?7 indicates the adjustment towards equilibrium that takes place in a given period. It was shown in the Engle-Granger test for cointegration that the residuald from estimating equation 1 was I(0).

The ECM which give information on the short-run export functions are presented