Ceris-CNR, W.P. N° 10/1997

on whether they are diversified in more than one country. The latter clearly implies overlapping
between multinationality and diversification. In fact, entering a secondary industry in a foreign
country affects the extent of both multinationality and diversification.
More precisely, a firm can produce one product in one country, more than one product in
one country (the firm is locally diversified), one product in more than one country (a single-product
multinational), more than one product in more than one country (a diversified multinational)3. As a
result, the firm’s corporate structure can be roughly described by its output shares in its primary
industry, whether in the home country (HC-PI) or abroad (FC-PI), and in secondary industries,
whether produced in the domestic market (HC-SI) or in foreign countries (FC-SI). However, by
comparing the firm’s output shares’ distributions in two points in time we cannot observe entry and
exit decisions, as indeed we are observing how the firm’s output has been relocated across countries
and across industries. This, of course, can occur via an increase (reduction) in the quantity produced
in other industries/countries by already existing plants or because the firm has opened (closed) a new
plant or, alternatively, it has taken over (divested) an already existing plant. Obviously, only the
latter can strictly be defined as entry (exit).
In order to measure entries and exits, more refined data are required than aggregate output
shares. What is needed, for any individual firm, is a matrix with industries in the rows and countries in
the columns, where each cell reports output in a given country and in a given industry at the finest
possible level. In other words, information must include estimates of its domestic and foreign
production in all industries in which it operates.
A data base with such characteristics was constructed for a set of 313 leading European
firms for 1987. It was assembled as a part of a wide ranging study of the structure of European
Union Manufacturing and is fully discussed in Davies, Lyons et al. (1996). Its salient features for
present purposes are that it includes all the five leading producers (at the EU level) in each of the 100
3-digit NACE manufacturing industries, observed for 1987, and that each firm aggregate EU
production in each industry is disaggregated into separate figures for each member state in which it

3

Other combinations can be obviously envisaged and Davies, Rondi and Sembenelli (1997) show that this is
sometime the case with real world firms.

7