RESIDENTIAL ELECTRICITY CONSUMPTION IN BOMBAY 85
of them have electricity consumption that is less than Rs. 13. Eighty-five of the chawl dwellers report their electricity expendi-
ture ranges between Rs. 13 to Rs. 105. Three percent of them have residential electricity consumption below Rs 13. The 76
of the dwellers of the flats below five floors have their residential electricity consumption within the range Rs. 13 to Rs. 105. Only
2 of them have electricity consumption below Rs. 13. Sixty-four percent of the dwellers of the flats above five floors have their
residential electricity consumption in the range of Rs. 13 to Rs. 105. Three percent of them have electricity consumption below
Rs. 13. The 76 of the bungalow occupants have electricity con- sumption in the range of Rs. 13 to Rs 105. Six percent of them
have electricity consumption below Rs. 13. The social pattern of the sample population is as follows: Maharashtrians—52,
Gujratis—15, South Indians—7, while the rest of 26 are from North India. Three percent of the Maharashtrians have elec-
tricity consumption below Rs 13. Eighty-five percent of them have residential electricity consumption that lies in the range of Rs. 13
to Rs. 105. Compared to this, 73 percent of the Gujratis have electricity consumption between Rs. 13 to Rs. 105 and only 1
percent have below Rs. 13. The South Indians have higher residen- tial electricity consumption. Seventy-eight percent of them have
it between Rs. 13 to Rs. 105, while only 1 percent of them have it below Rs. 13. The average monthly household income in Bom-
bay in Rs. 2550. The monthly expenditure on electricity of a household is only 3 percent of the total monthly income. The
mean household size as indicated by sample survey in Bombay is 5.24. As the household size increases, the electricity consumption
also increases. The mean number of rooms in Bombay is 2.2. The average age of the buildings in Bombay is 31 years, although some
structures are as old as 100 years.
The BMRDA survey reports the appliance characteristics of the households. The possession of various appliances of households in
Bombay is as follows. Sixty-nine percent of the households in Bom- bay have a television set. The VCR is possessed by 11 percent of the
households. Fifty percent of the Bombay households have a tape- recorder. Thirty-seven percent of the households have a refrig-
erator, while 61 percent of the households have an electric iron.
3. THE STRUCTURE OF ELECTRICITY DEMAND
The literature on the determinants of electricity consumption is vast. The simplest model of residential electricity demand is of
86 P. Tiwari
expressing the quantity of electricity consumption as a function of its current relative price and real income. Several restrictions
have been incorporated into the model in its wide application of studies of household electricity demand to improve its perfor-
mance and increase its flexibility see, e.g., Anderson, 1973; Bales- tra, 1967; Chern and Bouis, 1988; Chern et al., 1982; Houthakker
et al., 1974; Wills, 1981. Among the problems that are encoun- tered in the study of household electricity demand a major one
is the high degree of correlation between several explanatory variables in the demand equation with important implications for
its econometric estimation. The previous research bypasses this problem using systems of equations and appropriate restrictions
e.g., Garbacz, 1983; Halvorsen, 1975; Houthakker, 1980. Do- natos and Mergos 1991 used a single-regression equation even
under the conditions of multicollinearity using the “Ridge Regres- sion Technique” RR. The problem of heteroskedasticity also
persists besides the multicollinearity with the cross-sectional data. This paper proposes an alternate functional form of a single equa-
tion that avoids the problem of heteroskedasticity and multicollin- earity. The functional form takes care of the heteroskedasticity
problem, and it uses the RR technique of estimation to get over with the problem of multicollinearity.
The choice of variables depend on the determinants the residen- tial electricity consumption demand in Bombay. Other studies for
example, Hirst et al. 1981, find that the demand for comfort is a major determinant of monthly consumption: the key variables
are heating degree days and floor area a proxy for energy needed to heat the residence. Family size, income, and age of the house
were also the significant determinants of residential electricity consumption. Garbacz 1983a excludes dwelling size, focussing
on an index of appliance stock. Others have examined auxiliary heating Reilly and Shankle, 1988, preferred thermostat settings
Kushman and Anderson, 1986, or continuous occupancy Cap- per and Scott, 1982. The presence or absence of a particular
appliance has also been used in models of electricity consumption. For example, Parti and Parti 1980 use zero-one dummies and
interaction terms to determine the electricity consumption associ- ated with ownership of individual appliances.
The demand for residential electricity is derived from demand for services, such as lighting, heating, cooling, and cooking, which
are derived by using electric appliances. Therefore, the use of appliances and stock of appliances are major determinants of the
RESIDENTIAL ELECTRICITY CONSUMPTION IN BOMBAY 87
demand for the residential electricity. In the short run, the intensity with which consumers use electric appliances depends on their
income, housing unit structure, demographic characteristics, sea- sonal variations, weather, and electricity prices. The area covered
in the present study is Bombay, which does not have much varia- tions in its climatic conditions over the year. The residential elec-
tricity price structure, which is slab type, is the same throughout Bombay.
To formalize the discussion, the household’s problem is to max- imize utility 1:
U E;B,
1
where E is the electricity consumption and B represents other goods. The household is constrained by the existing stock of appli-
ances and their income. Utility maximization subject to these constraints generates a residential demand for energy that is the
function of income and prices, and may be written as 2:
E
i
5 F
Y
i
,P
i
,APINDX
i
,D
i
,H
i
. 2
Besides the price of electricity P and income Y, the other ex- planatory variables in the demand equation fall into three groups,
housing characteristics, demographic characteristics, and appliance holdings of the household.
We use expenditure on housing, in the form of expenditure on electricity, as a surrogate for consumption as the measurement
of the dependent variable related to the quantity of electricity consumed. The expenditure Elect
i
is, however, a product of unit price P and quantity consumed E; and the relation postulated
above becomes 3:
Elect
i
5 P
i
E
i
5 f
P
i
,Y,APINDX
i
,D
i
,H
i
, 3
where Elect
i
is the monthly expenditure on electricity by ith house- hold. The total annual income is used as proxy for permanent
income, consistent with permanent income hypothesis Friedman, 1957; Houthakker and Taylor, 1970. APINDX
i
is the appliance index, D
i
are the demographic variables, H
i
are house characteris- tics, and P
i
is the price of electricity per unit. The next issue is of functional forms of the demand function.
A survey of various linear and log-linear specifications indicates that log models have their own desirable properties such as re-
duced heteroskedasticity and reducing the influence of extreme electricity expenditures and incomes on parameter estimates.
88 P. Tiwari
4. VARIABLES CONSIDERED FOR ANALYSIS
