THE DEMAND FOR CAR OWNERSHIP: A NOTE
| Date | 01 June 1972 |
| Published date | 01 June 1972 |
| Author | D. G. Rhys,M. J. Buxton |
| DOI | http://doi.org/10.1111/j.1467-9485.1972.tb00518.x |
THE DEMAND FOR CAR OWNERSHIP:
A NOTE
M. J.
BUXTON
and
D.
G.
RHYS
Existing statistical estimates of the aggregate demand for cars suggest that
the main explanatory variables are price and some measure of per capita
income. Whilst other factors, such as socio-economic indices of various kinds
and variables relating to credit conditions have been tried, none of these
have tended to add much to the explanatory power of the models involved.
In
fact price and income variables alone, in some cases, have been found
to explain up to 95 per cent.
of
the variation in demand.'
These main econometric studies have yielded values for the long-run
elasticity of demand in the range 1.1 to
42.
For instance Evans (1969)
suggests that
a
one per cent. change in income would lead
to
a 1.1 per cent.
change in the expenditure
on
cars (although he measures the short-run
elasticity as
2.2).
At
the other end of the scale, Suits (1961) estimates the
income elasticity of the demand for new cars as
42.
Similarly the empirical
studies place the long-run price-elasticity of the demand for cars between
-0.6 and -1.5, although Evans calculates a short-run coefficient
of
-3.1.
These studies however, were concerned with aggregate demand functions,
and were not intended to examine variations
in
demand response according
to area, or
on
a regional basis.
A
study by Bennett (1967) introduced the
rural or urban location
of
the family unit as an explanatory variable. He
found that the higher the ratio of rural to urban families the higher the level
of the demand for cars.
Our basic model followed upon the work of Sleeman (1961 and 1969) and
Tanner (1963), and consists
of
a
cross-section analysis, on a geographical
basis, of the level of car-ownership. Basically it uses car-ownership per
1
,OOO
population as the dependent variable and population density, average
income per capita, and an index
of
the age structure of the local population
as independent variables.
A
stepwise multiple regression programme' was
used to computer analyse the data which was for 1968 and 1969.
Car
ownership:
This is the average number
of
'
private cars and private
vans
'
licensed in the particular licensing authority in the relevant year per
1,OOO population (Ministry of Transport, 1969 and 1970).
Population
Density:
Estimated population per square mile in June of the
relevant year (General Register Office, 1969 and 1970).
For example, Chow (1960), expressing the stock of automobiles in new-car
equivalent units finds that price and disposable income variables together explain
85-90
per
cent.
of
the variance in the stock of cars. Using expected income instead
of
disposable income the function can explain 90-95 pre cent. of the variance.
Stepwise Regression Program (BMDO2R),
from
Biomedical
Computer
Programs,
University
of
California Press, 1970. 175
To continue reading
Request your trial