Review of Industrial Organization 16: 69–87, 2000.
© 2000 Kluwer Academic Publishers. Printed in the Netherlands.
Capacity Utilisation and Excess Capacity: Theory,
Evidence, and Policy
Imperial College Management School, University of London, U.K.
Abstract. This paper models capacity utilisation as a cyclical variable that reﬂects both the value
of precautionary capacity and the desire to hold strategic excess capacity. Business unit data from
the Proﬁt Impact on Marketing Strategy (PIMS) database of large, predominantly US, companies are
used. Separate estimation is carried out for a number of SIC industry groups. Panel data estimation in
ﬁrst difference instrumental variable form is employed. Unlike many previous studies the utilisation
variable is well determined. The results show signiﬁcance for precautionary and strategic effects in
particular industries. The paper discusses the reasons for the industry speciﬁcity of the results. Policy
implications are discussed.
JEL Classiﬁcation: E22; L1; L4; L60.
This paper sets out to explain variation in capacity utilisation and excess capacity
using data on business units of large manufacturing ﬁrms. Excess capacity may
arise simply because of demand shocks and for this reason we need to model it
conditional on cyclical variables. Excess capacity may also be planned and here
we distinguish two reasons. First capacity may be held as a buffer, i.e., for precau-
tionary reasons, particularly where demand is volatile or uncertain. Second, excess
capacity may be used as a means of deterring entry and of signalling a readiness
to respond to that threat with price cuts. In the paper an econometric model is
estimated with panel data to explore the relative importance of these reasons for
This research was partly ﬁnanced by the U.K. Economic and Social Research Council (Research
Grant R000232176) and was facilitated by a sabbatical at the Research School of Social Sciences,
Australian National University, Canberra. I am indebted to Steve Dowrick and to referees for helpful
comments. Paul Yip and Nazera Dakhil helped with software. Advice on accessing and interpret-
ing the PIMS data was kindly provided by the London ofﬁce of the Strategic Planning Institute,
in particular Keith Roberts, Tony Clayton and Iain Brown. The paper has been presented at the
Atlantic Economic Conference and the annual Conference of the European Association for Research
in Industrial Economics.
The author is a Reader in Economics at Imperial College Management School, University of
London, 53 Princes’ Gate, Exhibition Rd., London SW7 2PG, U.K. Telephone 0171-594-9123; Fax:
071-823-7685; E-mail: firstname.lastname@example.org