Capturing Relatedness: Comprehensive Measures based on Secondary Data
| DOI | http://doi.org/10.1111/1467-8551.12124 |
| Published date | 01 January 2016 |
| Author | Elisabeth Nocker,Kurt Matzler,Christian Stadler,Harry P. Bowen |
| Date | 01 January 2016 |
British Journal of Management, Vol. 27, 197–213 (2016)
DOI: 10.1111/1467-8551.12124
Methodology Corner
Capturing Relatedness: Comprehensive
Measures based on Secondary Data
Elisabeth Nocker,1Harry P. Bowen,2Christian Stadler3and Kurt Matzler1
1Innsbruck University School of Management, Department of Strategic Management, Marketing and Tourism,
Universitaetsstr. 15, 6020 Innsbruck, Austria, 2McColl School of Business, Queens University of Charlotte,1900
Selwyn Avenue, Charlotte, NC 28274, USA, and 3Warwick Business School, University of Warwick, Coventry,
CV4 7AL, UK
Corresponding author email: Christian.Stadler@wbs.ac.uk
This paper presents new measures of technological and customer-side relatedness con-
structed from widely availablesecondary data. Relatedness is a concept central to predict-
ing the existence and nature of a relationship between corporate diversification and firm
performance. Yet,finding appropriate measures has been an ongoing struggle. The widely
used SIC-based entropy measure has low construct validity, and survey-based measures
are hard to replicate across firms and industries and overtime. The measures we develop
significantly outperform established measures in explainingvariation in firm performance
across firms and over time, and both sources of relatedness are found to be independent
and significant explanations of firm performance.
Introduction
A central theme in corporate strategy research is
the existence and nature of a relationship between
corporate diversification and firm performance
(Chatterjee and Wernerfelt, 1991). Building on the
seminal work of Andrews (1951), Anso (1957,
1958, 1965), Chandler (1962) and Gort (1962),
an immense amount of conceptual and empirical
work has been conducted on this topic (fora recent
review, see Hauschild and zu Knyphausen-Aufsess,
2013). At present, there is wide agreement that
firms who diversify into ‘related’ businesses (re-
lated diversifiers) outperform firms that diversify
into ‘unrelated’ businesses (unrelated diversifiers)
(Palich, Cardinal and Miller, 2000). While the
constructs of relatedness and related diversifica-
tion (Rumelt, 1974; Whittingtonand Mayer, 2000)
are central for understanding how diversification
strategy aects firm performance, divergence
exists between the theoretical constructs and their
operationalization (Robins and Wiersema, 1995).
Theoretically, the construct of relatedness de-
rives from the idea that a firm is able to realize
economies of scope if the industries in which it
operates are ‘similar’ (Markides and Williamson,
1994; Seth, 1990). ‘Similarity’ therefore implies
relatedness, in that the same resources, tech-
nologies, skills, knowledge and processes can be
deployed across similar industries so as to re-
alize economies of scope. In turn, economies
of scope can arise from sharing a common re-
source (Porter, 1985).1Such sharing is conceptual-
ized as multidimensional (Stimpert and Duhaime,
1997; Tanriverdi and Venkatraman, 2005; Pehrs-
son, 2006b), and can arise at any point along
a firm’s value chain. For example, sharing tech-
nological know-how (Markides and Williamson,
1994; Miller, 2004, 2006; Robins and Wiersema,
1995) or know-how regarding customers/markets
(Pehrsson, 2006b; Stimpert and Duhaime, 1997;
Tanriverdi and Venkatraman, 2005). Yet, most
1The term ‘economies of scope’ was first introduced by
Panzar and Willig (1977). Economies of scope are said to
exist when ‘joint production of two goods by one enter-
prise is less costly than the combined costs of production
of two specialty firms’ (Willig, 1979, p. 346).
© 2015 British Academy of Management. Published by John Wiley & Sons Ltd, 9600 Garsington Road, Oxford OX4
2DQ, UK and 350 Main Street, Malden, MA, 02148, USA.
198 E. Nocker et al.
measures (e.g. Miller, 2004, 2006; Neke and Hen-
ning, 2013; Robins and Wiersema, 1995), includ-
ing the widely used SIC-based entropy measure
(Jacquemin and Berry, 1979; Palepu, 1985) only fo-
cus on capturing similarities on the technology side
of the firm’s value chain. In empirical inquiry, this
narrow focus, by not capturing sources of related-
ness at other points along a firm’s valuechain, can
lead to biased inferences as to the true nature of
a relationship between diversification strategy and
firm performance.
For better alignment of the conceptualization
of relatedness with its empirical measurement, this
paper develops new firm-level measures of relat-
edness that capture the potential for economies
of scope on both the technological and cus-
tomer/market sides of a firm’s value chain. In de-
veloping our new measures,emphasis is placed not
only on alignment with the theoretical construct of
relatedness, but also ease of construction and ease
of replication over time. This is an important con-
tribution; ongoing attempts to developbetter mea-
sures of relatedness highlight that scholars con-
tinue to struggle with operationalizing relatedness
(e.g. Bryce and Winter, 2009; Farjoun, 1994 and
1998; Jacqemin and Berry, 1979; John and Harri-
son, 1999; Miller, 2004 and 2006; Neke and Hen-
ning, 2013; Palepu, 1985; Robins and Wiersema,
1995; Rumelt, 1974; Silverman, 1999).
As does the widely used entropy index
(Jacqemin and Berry, 1979; Palepu, 1985), our
measures assume that the realization of economies
of scope is enabled when a firm operates in similar
industries. Our measures capture the extent of
relatedness at the firm level as the sales weighted
average of the similarity between those industries
in which a firm is active. While consistent with
the commonly used SIC-based entropy index, our
approach to assessing industry similarity does not
rely on the judgment of those who created the
SIC system (or any other industry classification
systems such as the North American Industry
Classification System (NAICS)) as to the extent of
similarity (synergy potential) between any pair of
industries. Instead, we measure the actual distance
(similarity) between industries, based on a set of
industry variables that capture salient characteris-
tics of the content and process of a given industry
on both the technical and customer/market side
likely to support resource transfer and sharing.
This paper is not the first attempt to createeasily
replicable measures thatcapture relatedness across
the value chain. A commendable recent eort is by
Bryce and Winter (2009). While their measure is
easily replicable and generally applicable, a con-
cern with their approach is their assumption that
observed patterns of industries within a given cor-
porate portfolio already reflect the achievement of
relatedness. The assumption that activities within
a portfolio are a priori relatedis problematic, since
it renders the concepts of relatedness and diversifi-
cation (and its benefits) tautological (Neke and
Henning, 2013), and therefore sidesteps entirely
the issue of construct validity by simply assuming
it away.2We avoid such issues by not assuming that
a firm’s portfolioof industries is coherent. Instead,
we measure the synergy potential of a firm based
on the similarity of the characteristics of the in-
dustries in which a firm operates. This approach
has the additional advantage that distances, and
hence the similarity, between industries vary over
time, reflectingchanges in technologies and market
characteristics.
Overall, our intention is to oer a viablealterna-
tive to the SIC-based entropy measure. Despite its
often mentioned limitations, the entropy measure
of relatedness remains dominant, in part because
data to construct alternative measures are often
not widely available. This is not an issue faced by
our proposed measures.
In what follows, we first explain the theoretical
foundation of our measures, followed by a brief re-
view of established relatedness measures and then
the presentation of our new measures. Following
this, we present an analysis that compares our
new measures with the most widely used measures
in terms of their ability to explain variation in
firm performance. This analysis not only assesses
whether our measures significantly outperform
existing measures, it also indicates whether syner-
gies on the technological and customer sides of a
firm’s value chain are independent and statistically
significant sources of relatedness that aect firm
performance.
2Another concern with the Bryce and Winter (2009) mea-
sure is that it uses value added and not sales at the firm
level to weight the importance of industry pairings at the
firm level. This potentially confounds – or leaves unan-
swered – the underlying basis for the values of their in-
dex since value added can also include pure economic
profit and it also varies with industry characteristics (e.g.
capital-labour intensity).
© 2015 British Academy of Management.
To continue reading
Request your trial