K. Ludi: consumption behaviour in Zambia
In the long run, it is put forward that Zambian real PCE is influenced by real national
income (for want of better data on disposable income) and the real short-term lending
rates applicable at the time. In other words:
lnCt=β1lnYt+β2rt (1)
The obvious problem with using equation (1) as the final equation is the problem of non-
stationarity. Thus, the data is to be differenced according to its order of integration, and
the regression run again:
∆lnCt =β1 ∆lnYt +β2 ∆rt (2)
Equation (2) thus accounts for any short-term disturbances affecting real PCE in Zambia.
The data used to estimate equations (1) and (2) are extracted from the World Bank
Development Indicators (TSE). The period under analysis is from 1970 to 2001, and
unfortunately no later due to data unavailability. Table 1 below contains a list and
description of the variables used in the various steps of the Engle-Granger approach.
Table 1: Variables used and descriptions
Variable |
Description |
Unit |
Conversions / Workings |
LN_CONS_ZK |
Real private consumption |
Constant 1994 Zambian Kwatchas (millions) |
Natural logarithm. |
LN_GDP_ZK |
Real GDP at market prices |
Constant 1994 Zambian Kwatchas (millions) |
Natural logarithm. It was not possible |
LN_GDP_CAPITA |
Real GDP per capita |
Constant 1994 Zambian Kwatchas (millions) |
Natural logarithm. |
LN_INDTAX_ZK |
Net indirect taxes paid |
Constant 1994 Zambian Kwatchas (millions) |
Natural logarithm. |
LN_INV |
Gross domestic investment |
Constant 1994 Zambian Kwatchas (millions) |
Natural logarithm. |
LN_M3_ZK |
Real M3 money supply |
Constant 1994 Zambian |
Natural logarithm. |