THE DEMAND FOR NEW CARS II—AN EMPIRICAL MODEL FOR THE UK
| Author | J. C. ODLING‐SMEE,A. G. ARMSTRONG |
| DOI | http://doi.org/10.1111/j.1468-0084.1979.mp41003002.x |
| Date | 01 August 1979 |
| Published date | 01 August 1979 |
THE DEMAND FOR NEW CARS TI-
AN EMPIRICAL MODEL FOR THE UK
By A. G. ARMSTRONG and J. C. ODLING-SMEE
1. INTRODUCTION
In an earlier article' we developed a general theoretical model of the demand
for new cars. Since such a large part of the demand for new cars comes from people
who already own cars2 the model emphasized the replacement decision. The
concept of a newness preference function was introduced in order to analyse the
nature of the decisions to sell the existing car and whether to replace it with a new
or used car. The main underlying factor in the demand for new cars is the need to
replace ageing cars. Past purchases of new cars are therefore an important
determinant of replacement demand. However, given the nature of the product,
replacement can be postponed or brought forward and we therefore considered
both normal and abnormal replacement demand. Various exogenous factors such
as income and credit conditions were postulated as determinants of both types of
replacement demand. These and other exogenous factors were also needed in the
model to explain the demand from people buying a car for the first time.
The analysis suggested that the shift to a new equilibrium level of demand
following a change in one of the exogenous variables involves two separate
effects. First, normal replacement demand is permanently altered. Secondly there
is a temporary change in abnormal replacement demand because people find
themselves in a sub-optimal position which they seek to correct.
The earlier article therefore argued that normal replacement demand should be
a function of past purchases of new cars and of the levels of exogenous variables;
abnormal replacement demand should be a function of changes in the levels of the
exogenous variables. New demand (i.e. demand from people not replacing existing
cars) was similarly shown to depend on both the levels of, and changes in, exogenous
variables. The total demand for new cars in period t, q, may be written as:
z_) (l)
where the w's are a set of weights relating to purchases in previous time-periods, z
is a vector of exogenous variables and n and a are functions signifying normal and
abnormal demand. The first two terms may be regarded as measuring the normal
demand for new cars and the last term abnormal demand (or, alternatively, changes
in the timing of normal purchases).
The aim in this paper is to fit this model to quarterly UK data over recent
years. A preliminary attempt at fitting this model was made a few years ago and
'A. G. Armstrong and J. C. Odling-Smee, The Demand for New Cars IA Theoretical Model of
Replacement Demand', BULLETIN, November 1978.
2 In the UK in 1970, 91 per cent of new car buyers were already car owners. ('The Motor
Transactions Survey, 1971', Economic Trends, No. 242, December 1973.)
193
194 BULLETIN
has been described elsewhere3. The estimates described there were rapidly out-
dated by the increase in oil prices in the autumn of 1973. The results given here are
based on estimates for a longer time-period extending to the end of 1976 thus
enabling due allowance to be made for the effect of petrol prices on demand. In
addition since 1973 the car market has been influenced simultaneously by a number
of 'new' factorsa high rate of general inflation, large increases in new car prices,
a widening gap between new and used car prices and increased economic
uncertainty. All these influences are now considered in the estimation of the model
in a way which was not possible nor necessary in the pre-1973 formulation.
The plan of this paper is as follows. In section 2 we examine the exogenous
factors operating on car demand, both new and replacement, and we select those
likely to be important. The main results of the equations in the model, including
the estimation of the polynomial distributed lag function which we used to obtain
the weights associated with past purchases, are discussed in section 3. The overall
explanatory power of the model is examined in section 4 and in section 5 we use the
model to forecast demand outside the estimation period. This is followed by a short
conclusion and an appendix describing the notation and giving notes on the
derivation of some variables.4
2. THE VARIABLES IN THE MODEL
2.1. Introduction
The central part of our model is replacement demand and the underlying level
of demand is thus related to the level of purchases in previous quarters as given by
the term wq_ in (1) above. We postpone until section 3 the discussion of the
estimation of the distributed lag function. Here, we consider the exogenous
variables operating on the timing and nature of replacement demand as well as
demand from new owners. We shall use both the levels and changes in the levels
of all these variables. The former will affect the long-run equilibrium demand and
in order to assess the short-term effects on demand of the move to a new equilibrium
following a shift in one of the exogenous variables it will be necessary to include
changes in the levels of the variables. Generally, one would expect to use the
change in the value of the variable from the previous quarter. However, in view
of the erratic quarterly pattern of some of the variables this is not the most
appropriate, particularly as a decision to buy a car may be taken after careful
consideration over a period of time reflecting both the magnitude of this category
of spending to the typical consumer and the fact that the purchase can be delayed.
In addition, the actual task of selling and buying the replacement car occupies a
certain length of time. We, therefore, use changes in the variables over the previous
year rather than the previous quarter.
A. G. Armstrong: The Demandfor New Cars: An Empirical Modelfor Short-term Forecasting (NEDO,
1974).We are grateful for the many helpful discussions we have had with members of the Society of
Motor Manufacturers and Traders Ltd (SMMT), the National Economic Development Office (NEDO),
and the Department of Industry at NEDO about earlier versions of this model, and the demand for new
cars more generally. The present article is our responsibility alone.
To continue reading
Request your trial