Review of Industrial Organization 24: 267–283, 2004.
© 2004 Kluwer Academic Publishers. Printed in the Netherlands.
Geroski’s Stylized Facts and Mobility of Large
German Manufacturing Firms
UWE CANTNER and JENS J. KRÜGER
Friedrich-Schiller University Jena, Department of Economics, Carl-Zeiss-Strasse 3, D-07743 Jena,
E-mails: firstname.lastname@example.org, email@example.com
Abstract. Paul Geroski has established two stylized facts with respect to the prevalence of differ-
ential changes (mobility) of indicators of economic and technological performance: technological
indicators show a larger amount of mobility than do economic indicators. We assess the two stylized
facts for a sample of 392 large German ﬁrms observed over the period 1981–1993 and assigned to
eleven manufacturing industries. We analyze these data with two novel methods, Salter curves and
mobility indices based on fractile Markov chains. Our analysis supports Geroski’s two stylized facts
in the case of large German manufacturing ﬁrms taking account of sectoral differences.
Key words: Distribution dynamics, Markov chains, mobility measurement, productivity
JEL classiﬁcations: L11, O30, C10, C23
In his survey on large company performance Paul Geroski (1998) reviews the lit-
erature on the subject and summarizes the results in six stylized facts. The two
most remarkable of these stylized facts are concerned with the overall amount of
differential changes among ﬁrms with respect to economic and technological per-
formance indicators. These are stylized fact #3 which states that heterogeneities in
economic performance between ﬁrms persist into the long run more or less regard-
less of how performance is measured and stylized fact #5 which claims that most
ﬁrms are irregular and erratic innovators when innovations are measured by counts
of major innovations. If we term the amount of differential changes as mobility
these stylized facts say that economic performance indicators are characterized by
low mobility whereas technological performance indicators are much more mobile.
Geroski’s stylized facts are supported by an empirical analysis of a balanced
panel of 280 large quoted UK companies over the period 1972–1982 (without
distinguishing the ﬁrms with respect to their industries) and are also based on the
recent literature on industrial dynamics. Empirical studies like those of Geroski and
Toker (1996) for market shares or Jensen and McGuckin (1997) for relative labor
productivity found considerable persistence of these measures. In contrast to these