Complex Strategic Choices: A New Approach and Application to Foreign Firm Agglomeration
Date | 01 August 2017 |
DOI | http://doi.org/10.1002/gsj.1161 |
Published date | 01 August 2017 |
Complex Strategic Choices: A New Approach
and Application to Foreign Firm Agglomeration
Zheying Wu,
1
*Robert Salomon,
2
and Xavier Martin
3
1
Department of Business Administration, School of Management, Fudan
University, Shanghai, China
2
Management and Organizations Department, NYU Stern School of
Business, New York, New York
3
Department of Management, School of Economics and Management,
Tilburg University, Tilburg, The Netherlands
Research summary: In this study, we introduce a new statistical technique to the field of
strategy that accounts for complex interrelated decisions. The technique is meant for
situations in which one strategic decision depends critically upon another—e.g., where
companies must commit to one decision before making another. We demonstrate the value
of the technique in the context of companies choosing countries in which to invest, fol-
lowed by areas within those countries in which to locate their facilities. We then describe
the generic benefits of the statistical technique, offer guidance for how to use it, and pro-
vide the programming code so that others can implement it. We explain how the technique
is particularly well suited to problems in which companies must decide between multiple
options and the sample size is small relative to the number of initial options.
Managerial summary: Research in strategy and international business has increas-
ingly adopted two-stage models to account for endogeneity. Established approaches,
however, deal most effectively with problems characterized by binary first-stage mod-
els. We introduce an alternative that allows for multiple choices in the first stage. We
use the technique to examine the agglomeration patterns of foreign firms, while
accounting for country selection. We find that, conditional on country selection, foreign
firms tend to agglomerate in host countries characterized by collectivist cultures, politi-
cal uncertainty, and economic uncertainty. The results imply that host country charac-
teristics exert a strong influence on the agglomeration preferences of foreign firms. We
provide guidance on when to use the empirical estimation technique and furnish details
of the estimation procedure so that others can apply the technique to a broader class
of selection problems. Copyright © 2017 Strategic Management Society.
Scholars in strategic management and international
business increasingly conduct research using meth-
ods that address endogeneity and selection bias.
Indeed, there is now an established body of work
that studies the performance consequences associated
with binary strategic decisions, such as alliance ver-
sus acquisition or make versus buy (e.g., Leiblein,
Reuer, & Dalsace, 2002; Shaver, 1998). Typically,
scholars study such questions using a two-stage
approach developed by Heckman (1979) (for a sum-
mary, see Hamilton & Nickerson, 2003). Although
the Heckman approach accounts for endogeneity and
selection bias in situations where the firm selects
between two options, we lack effective techniques to
Keywords: agglomeration; selection/endogeneity; nonlinear
models
*Correspondence to: Zheying Wu, School of Management,
Fudan University, Guoshun Road 670, Shanghai 200433,
China. E-mail: zwu@fudan.edu.cn
Copyright © 2017 Strategic Management Society
Global Strategy Journal
Global Strategy Journal, 7: 286–311 (2017)
Published online in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/gsj.1161
address research questions that involve more com-
plex strategic choice sets.
By complex choice sets, we refer primarily to situa-
tions where a firm can choose among three or more
discrete alternatives at the outset. For instance, a firm
may face the choice among multiple alliance partners
or may have to choose among multiple competing
technologies. Unfortunately, the Heckman (1979)
approach cannot efficiently and robustly correct for
endogeneity and/or selection bias when the first-stage
strategic choice set includes more than two decisions
(Bourguignon, Fournier, & Gurgand, 2007). And
more recently developed alternatives include highly
restrictive assumptions (e.g., Lee, 1983) or are so
demanding computationally as to be unsuitable for
most datasets (e.g., Dubin & McFadden, 1984).
To address these methodological constraints, we
introduce a more advanced econometric approach
pioneered by Dahl (2002). Dahl builds a two-stage
model based on the Heckman (1979), Lee (1983),
and Dubin and McFadden (1984) approaches.
Dahl’s methodology improves on Heckman by
allowing for multiple options in the first stage. It
improves upon Lee and Dubin and McFadden by
relaxing restrictive constraints.
We apply Dahl’s (2002) technique to an important
question in the global strategy literature: how foreign
entrants choose their location within a host country
(Shaver & Flyer, 2000). In particular, we examine
whether foreign entrants decide to locate close to
domestic competitors—i.e., within an agglomeration—
or far away. Prior studies identify various factors that
can influence the agglomeration patterns of foreign
entrants in the host country (Chung & Song, 2004;
Nachum & Wymbs, 2005; Shaver & Flyer, 2000).
However, that literature has not addressed one key
challenge—accounting for the role of country selection
as an antecedent to the agglomeration decision. There-
fore, we do not know, for example, whether observed
patterns of foreign entrant agglomeration in a host
country are driven by country selection. It could be that
multinational firms choose countries precisely because
they expect to collocate with local firms, thereby con-
flating country selection with agglomeration.
In addition to adopting a novel technique (Dahl,
2002) that allows us to control for country selection as
an antecedent to agglomeration, we extend the agglom-
eration literature by examining the impact of host
country institutions. When we refer to host country
institutions, we draw from work in institutional eco-
nomics and institutional sociology (North, 1991;
W.R. Scott, 2013). We consider institutions at the
national level and define national institutions as
“humanly devised constraints that structure political,
economic, and social interaction…[and that] provide
the incentive structure of an economy…” (North,
1991: p. 97).
1
We understand less than we should about how
national institutions influence agglomeration and, in
particular, how host country institutions influence
the agglomeration patterns exhibited by multina-
tional firms. Given that host country institutions
have been demonstrated to influence a variety of for-
eign firm strategies (e.g., Henisz & Delios, 2001;
Peng, Wang, & Jiang, 2008; Salomon & Wu, 2012),
such an extension to the agglomeration literature
helps fill an existing gap in our understanding.
With all that in mind, we examine whether, condi-
tional on country selection, the host country institu-
tional environment affects the agglomeration decisions
of foreign entrants. Theoretically, we assert that in
addition to the traditional factors that drive agglomera-
tion, institutional factors matter. We argue that foreign
entrants are swayed by national culture and economic/
political uncertainty in the host country (Martin, Salo-
mon, & Wu, 2010). Empirically, we analyze the
agglomeration patterns of 180 foreign plant invest-
ments made by 61 firms across 22 host countries in the
global semiconductor industry from 1975 to 2004. To
account for how endogeneity in country selection influ-
ences agglomeration, we implement a variant of the
two-stage selection-correction procedure proposed by
Dahl (2002). Thereby, we shed light on patterns of
agglomeration exhibited by foreign entrants while
addressing country selection, a critical initial decision
for foreign firms (Davidson, 1980; Dunning, 1988;
Henisz & Delios, 2001).
This study makes several contributions to the
strategy, international business, and agglomeration
literatures. First, it offers the Dahl (2002) two-stage
selectionmethodasabroadersolutiontogeneric
selection problems that include three or more choices
in the first stage. Although the application here is to
1
We build our definition of national institutions in a manner
consistent with the prevailing international business literature.
We recognize, however, that the economic geography litera-
ture has adopted a definition of institutions that is narrower,
focusing on specific institutions that facilitate industrial activ-
ity at the local level. The focus tends to be more on local edu-
cational, market, financial, media, and research institutions
(e.g., Fan & Scott, 2003; Markusen, 1996; Rantisi, 2002;
Sternberg & Arndt, 2001). We believe the two views are com-
patible, though reconciling the disparate foci is outside the
scope of the current study.
Foreign Firm Agglomeration Patterns 287
Copyright © 2017 Strategic Management Society Global Strategy Journal, 7: 286–311 (2017)
DOI: 10.1002/gsj
To continue reading
Request your trial