Consumption Behaviour in Zambia: The Link to Poverty Alleviation?



K. Ludi: consumption behaviour in Zambia

11


the short-run can thus be explained by: the long run dynamics indicated in table 2; GDP
per capita from one period prior; a structural break dummy variable accounting for the
sharp and sudden decline in real PCE that occurred in 1992 due to the interest rate shock
mentioned earlier; short term treasury bill rates from the previous two periods; real
wealth from two periods prior as proxied by the M3 money supply; indirect taxes paid to
the government, where the burden usually falls entirely on the consumer; real lending
rates; real domestic investment from the previous period; and real government
expenditure from one period before. The lags involved make economic sense, as no
effects in an economy are instantaneous, but rather take time (lags) to cause adjustments.
The only ‘surprising’ independent variable that affects PCE in the short run is real
government consumption expenditure, which incorporates remuneration of public sector
employees. Since Zambia is largely a rural agricultural economy, the public sector
comprises most of the non-agricultural salary-earners, and thus contributes largely to the
‘higher’ income class that are larger contributors to PCE. The battery of usual diagnostic
tests on the residuals can be performed, with results in table 4.

Table 4: Selected diagnostic test results of the ECM of private consumption expenditure

Test

Test statistic

I P-value

Conclusion

Normality

Jarque-Bera

JB = 0.280

0.869

Residuals are normally distributed

Serial correlation

Breusch-Godfrey (2)

1.183

0.553

No serial correlation up to order 2

Heteroskedasticity

ARCH (1)

White (no cross terms)

0.918

14.467

0.338

0.634

No heteroskedasticity

No heteroskedasticity

Stability

Ramsey reset (1)

0.978

0.323

Stable regression

4.3. Engle-Yoo Third Step

The coefficients from the long-run cointegrating equation are biased and inaccurate due
to the fact that they are based on non-stationary time series, and thus cannot be
interpreted. Similarly, the t-statistics cannot be evaluated or used for inference, as they



More intriguing information

1. The Trade Effects of MERCOSUR and The Andean Community on U.S. Cotton Exports to CBI countries
2. Searching Threshold Inflation for India
3. The name is absent
4. THE WELFARE EFFECTS OF CONSUMING A CANCER PREVENTION DIET
5. An Incentive System for Salmonella Control in the Pork Supply Chain
6. The name is absent
7. How Offshoring Can Affect the Industries’ Skill Composition
8. The Clustering of Financial Services in London*
9. The East Asian banking sector—overweight?
10. The name is absent