K. Ludi: consumption behaviour in Zambia
10
more economically realistic in explaining real PCE. The short-run regression that
explains real PCE in Zambia is therefore:
∆ ln Ct = ares _ co int t - 1 + β1 ∆ ln Y / capitat - 1 + β2 ∆ r _ tbillt - 2 + β3 ∆ ln M3t - 2 +
β4 ∆ ln Gt - 1 + β5 ∆ ln Tt + β6 ∆ r _ Iendrate + β7 ∆ ln Invt - 1 9 ( 4 )
The results of the ECM regression are in table 3.
Table 3: Regression results of the error correction model (dependent variable: D(LN_CONS_ZK)10)
Sample (adjusted): 1974 2001 Included observations: 28 after adjustments | ||||
Variable |
Coefficient |
Std. Error |
t-Statistic |
Prob. |
RES_COINT(-1) |
-0.421942 |
0.109743 |
-3.844812 |
0.0011 |
LN_GDP_CAPITA(-1) |
-0.028966 |
0.009264 |
-3.126732 |
0.0056 |
D(R_TBILL(-2)) |
-0.002131 |
0.000498 |
-4.282284 |
0.0004 |
D(LN_M3_ZK(-2)) |
-0.165104 |
0.078352 |
-2.107218 |
0.0486 |
D(LN_GCONS_ZK(-1)) |
0.132035 |
0.064002 |
2.062989 |
0.0531 |
D(LN_INDTAX_ZK) |
-0.251680 |
0.055979 |
-4.495959 |
0.0002 |
D(R_LENDRATE) |
-0.001372 |
0.000465 |
-2.950467 |
0.0082 |
D(LN_INV(-1)) |
0.072068 |
0.039942 |
1.804303 |
0.0871 |
D(DUM_92) |
-0.517581 |
0.065953 |
-7.847746 |
0.0000 |
R-squared |
0.863362 | |||
Adjusted R-squared |
0.805830 |
The coefficients of the independent variables may not be economically evaluated since
the variables have been differenced once. However, the variables may be statistically
evaluated. The t-statistics of all the independent variables are statistically significant at a
10 per cent level of significance. The adjusted-R2 value indicates that 81 per cent of the
short-run variation in real PCE can be explained by the model. Real PCE in Zambia in
10
Where: Ct = real PCE; Yt /capita = real GDP per capita; r_tbillt = real treasury bill yield rates; M3t
= real M3 money supply; Gt = real government consumption expenditure; Tt = real net indirect tax
revenue; r_lendratet = real short-term lending rate; and Invt = real gross domestic capital
formation, all at time t.
D(...) indicates that a variable has been differenced once to remove non-stationarity.