PATTERNS OF ACQUISITION OF CONSUMER DURABLES*
| Date | 01 May 1974 |
| Published date | 01 May 1974 |
| Author | J. J HEBDEN,J. F. PICKERING |
| DOI | http://doi.org/10.1111/j.1468-0084.1974.mp36002001.x |
OXFORD BULLETIN
of
ECONOMICS and STATISTICS
PATTERNS OF ACQUISITION OF CONSUMER DURABLES*
By J. J. HEBDEN and J. F. PICKERING
I. INTRODUCTION
In the analysis of household expenditures, distinction is often made between
consumable goods (foodstuffs, etc.) and durables (motor cars, furniture, electrical
appliances) since the two groups are held to be subject to different sorts of influence
on actual purchase behaviour. In particular, purchases of durables are normally
considered to be discretionary, that is their timing is not fixed and the period over
which consumption of the services or utilities provided by the durable extends
through time, whereas for consumables purchase and consumption are assumed to
occur at the same point in time. Further, whereas for nutritional and other
reasons many food purchases are made continuously and together, purchases of
durables are generally made singly and at discrete periods of time. The process
of acquiring durables is therefore cumulative and the actual stock of durables
owned by households is likely to depend upon many factors including income,
wealth, social class grouping, family size, duration of marriage, willingness to use
credit, etc.
The question to which we wish to address ourselves in this paper is whether it is
possible, by observing the outcome of past purchasing behaviour by individual
consumers, to establish that there is, within tolerable limits, a unique order in which
a particular set of durables tends to be acquired by all members of the population
or a particular sub-group. If such an acquisition pattern can be identified this is not
without interest since it indicates that even if consumer durable buying is dis-
cretionary with respect to timing it is less discretionary with respect to the order of
acquisition of additional durables. If this is the case, then knowledge of the order of
* This paper derives from a study on consumer purchasing of durable goods which is
financed by the SSRC and is under the direction of Dr. J. F. Pickering. Thanks are due to
Mr. B. C. Isherwood for valuable assistance in the processing of the data, and to Mrs. I. Crockett
for her assistance in the calculations. Professor Graham Pyatt made some valuable comments
on an earlier, extended version of the paper which appeared under the title 'Priority patterns
for consumer durables' as University of Sussex Economics Seminar Paper 73103.
67
Volume 36 May 1974 No. 2
68 BULLETIN
acquisition has some value for forecasting' and also in identifying particular
market segments from which new purchasers of a specific product are most likely to
be drawn.
Earlier researchers have adopted this technique for limited ranges of goods
owned in Britain (Pyatt; Brown, Buck and Pyatt), Israel (Paroush [1] and [2]), the
USA (McFall and Paroush [2]), and have found it possible to establish a reasonably
unique order of acquisition. In his recent work Paroush [2] showed that the onler
of acquisition was independent of the level of income but that the level of income
has a crucial role in determining the stage which the consumer has reached in the
acquisition process. He therefore claimed that one could predict the increase in
demand for each good on the basis of the proportion of households at each income
level together with a forecast of the growth and distribution of incomes. The
majority of studies have tended to use either the Guttman coefficient of repro-
ducibility (Goodman; Guttman [1]; Sagi) or the point correlation matrix (Guttrnan
[2]) but the pioneering work of Pyatt has been based upon the calculation of the
probabilities of owning each product at each point in the accumulation of a speciied
set of goods. In 11is study of electrical applicance ownership he showed that
differences were observable in the estimated acquisition patterns for different
social classes (Pyatt).
While Pyatt's ultimate interest was in the calculation of a priority pattern
taking into account also respondent indications of the items they hope to acquire
next, our concern here is with investigating in more detail certain aspects of
observable acquisition patterns. Consequently we shall use Pyatt's basic statistical
technique but will report some important statistical extensions to his published
methodology (see the Appendix to this paper). We shall also offer results for a
wider range of durable goods and will be able to relate these results to our (lis-
cussion of some methodological considerations that seem important, particularly
questions of the choice of sets of durables for analysis and the problems of dealing
with items that appear to become inferior goods as an acquisition pattern proceeds.
Before we describe the calculation of an acquisition pattern and the result; we
have obtained, it is necessary to comment on some of the limitations, as we see
them, of this approach to demand forecasting. The data from which the patt rns
are calculated tends to be information on ownership levels in different households
at the time of the survey and the historic order of acquisition of the durables
concerned is not studied. It is inferred that if a particular cross-sectional pricrity
pattern is identified this represents the order in which the items were acquired
through time in the past and in which acquisition is expected to take place in the
future. This does, of course, raise all the usual questions about the extent to which
cross-sectional data can be used for time-series prediction and offers no behavioural
explanation as to why items are acquired in the observed order. In this particular
context the assumptions of unchanging tastes and relative prices are likely to
prove unrealistic. The likely effects of the introduction of a new durable cannot be
forecast with this technique, either to predict the level of sales that might be
1Though its practical limitations in forecasting seem to us formidable, see the next para-
graph and the concluding remarks.
To continue reading
Request your trial