PROVIDE Project Technical Paper 2005:1
incorrectly reported capital amount. Respondents probably interpreted the question
incorrectly and reported the principal loan amount rather than the capital component, the
same problem as the one identified before. The summary below clearly shows how capital is
grossly over-reported in these 195 households. Since many of the interest components were
also nonsensical, it was decided to follow the same procedure as before by replacing the two
components with estimated values for capital and interest, using the 55-45 split.
February 2005
Variable | Obs Mean Std. Dev. Min Max
-------------+-----------------------------------------------------
P0303Q0501 | 195 1621.282 984.4212 178 7000
P0303Q050101 | 195 55917.74 77056.19 0 550000
P0303Q050102 | 195 2002.79 9342.135 0 120000
After all corrections have been made a total of 1897 households correctly report a monthly
instalment and the capital-interest breakdown. Incidentally, the capital-interest split remains
fairly close to the initial estimate, with an average of 55.2% of the monthly instalment going
towards capital and the remainder towards interest on the loan. The total instalment, however,
is considerably lower than before. This is as a result of the incorrect capital components that
have been deleted.
Variable |
-------------+
P0303Q0501 |
P0303Q050101 |
P0303Q050102 |
|
Obs |
Mean |
Std. Dev. |
Min |
Max |
|
1897 |
1691.131 |
4730.005 |
3 |
190000 |
|
1897 |
934.128 |
4535.202 |
0 |
190000 |
|
1897 |
757.0033 |
885.5662 |
0 |
22500 |
The only other major change made in the housing section is made in question 6. Since all
household expenditures in a SAM should be reported inclusive of VAT, the VAT component
in this question should be added pro-rata to the other components that make up total
payments for housing services. It is assumed that no VAT is payable on assessment rates and
taxes, as this expenditure type is already a local or municipal tax on property. All other
components, namely water, electricity, gas, sanitary services and refuse removal should be
inclusive of VAT. VAT is therefore added pro-rata to these components. Some households
(11 in total) reported only VAT. For these households VAT is added to the various
components in six equal shares.38
Finally, a check is performed to see what the impact of the changes made has been on the
reported total. Poswell (2003) points out that prior to any changes a discrepancy between
calculated and reported housing expenditure totals in about 1.5% of households was seen.
Having made the changes discussed above, it is found that this discrepancy has risen to
almost 2% of households (520 observations). The average difference over all households,
including those with zero difference, is about R6.43, as seen in the Stata output below.
38 Previously (see PROVIDE, 2003a) the VAT component of households that reported only VAT was added to
the variable for household indirect taxes. This is viewed as an unnecessary complication, hence the
change in the procedure.
44
© PROVIDE Project