PROVIDE Project Technical Paper 2005:1
into a higher income bracket. The negative coefficient for variable logW suggests that less tax
is payable if more family members contribute to the pool of household income. This makes
economic sense as well. For example, if a single household member contributes R120,000 to
the household income, her marginal tax rate will be much higher than two members
contributing R60,000 each. Finally, the negative constant, which is equivalent to about -
R1,532 can be interpreted as a tax rebate (the actual tax rebate was about R3,800 in 2000).
The increase in the average reported tax rate was quite significant, increasing from R2,831 to
R4,041. However, given the large number of households reporting zero tax where they were
expected to have paid tax, this sharp increase was expected. The total expenditure levels were
adjusted accordingly.
February 2005
3.2.8. Mapping income and expenditure categories (mapexp.do and mapinc.do)
Since the IES 2000 was initially developed to provide information for the calculation of the
CPI, the expenditure items are not grouped in a way that is consistent with the commodity
categories used in the SAM. On the expenditure side each expenditure item has to be mapped
to a specific commodity group (see mapexp.do). These commodity groups are based on the
activity-based Standard Industrial Classification (SIC) codes of 1993. A total of 95
commodity groups are created based on the SIC codes, plus an additional commodity group
denoting domestic services (C1 - C96) (see Table 11).
Apart from the 96 commodity groups created, additional expenditure categories are also
created for inter-household transfers (variable hhtrans), the payment of income tax (variable
hhinctax), the payment of local or provincial ‘taxes’ (variable hhlocaltax) and savings
(variable hhsav). Some expenditure items, such as pocket money, costs relating to home
production, gambling expenses and other losses could not be mapped successfully and were
included in a variable called hhother. The value of home consumption plus input costs
(variable hhhphc) is also created separately, but can be incorporated into hhother if this
information is not needed separately. Variable hhother is netted out by allocating the
expenses pro-rata to the other expense categories (see discussion later).
A few important assumptions were made in order to map or create commodity group C89
(Financial Services Indirectly Measured, or FSIM). This is an expense category that cannot
be directly mapped from the IES 2000 data since it is an implicit expense incurred by
households that either borrow or invest money with a bank. A large part of the financial
service industry’s revenue comes from charging higher interest rates for loans than it pays for
deposits made at the bank. Thus, when households pay interest on loans an implicit finance
charge is included since the household pays a premium on the ‘base’ interest rate. Similarly,
when a household receives interest payments from a bank it actually receives less than the
‘base’ interest rate. These implicit finance charges appear nowhere on the expenditure side of
50
© PROVIDE Project