Vlas, C is ina O.
A icle
Racially di e se leade ship and sus ainable alliance
po olios
Adminis a i e Sciences
P o ided in Coope a ion wi h:
MDPI – Mul idisciplina y Digi al Publishing Ins i u e, Basel
Sugges ed Ci a ion: Vlas, C is ina O. (2024) : Racially di e se leade ship and sus ainable alliance
po olios, Adminis a i e Sciences, ISSN 2076-3387, MDPI, Basel, Vol. 14, Iss. 11, pp. 1-14,
h ps://doi.o g/10.3390/admsci14110279
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/321086
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h ps://c ea i ecommons.o g/licenses/by/4.0/
Ci a ion: Vlas, C is ina O. 2024.
Racially Di e se Leade ship and
Sus ainable Alliance Po olios.
Adminis a i e Sciences 14: 279.
h ps://doi.o g/10.3390/
admsci14110279
Recei ed: 12 July 2024
Re ised: 18 Oc obe 2024
Accep ed: 22 Oc obe 2024
Published: 28 Oc obe 2024
Copy igh : © 2024 by he au ho .
Licensee MDPI, Basel, Swi ze land.
This a icle is an open access a icle
dis ibu ed unde he e ms and
condi ions o he C ea i e Commons
A ibu ion (CC BY) license (h ps://
c ea i ecommons.o g/licenses/by/
4.0/).
adminis a i e
sciences
A icle
Racially Di e se Leade ship and Sus ainable Alliance Po olios
C is ina O. Vlas
Depa men o Managemen , Pompea College o Business, Uni e si y o New Ha en,
Wes Ha en, CT 06516, USA; [email p o ec ed]
Abs ac : In my esea ch, he e ec s ha he acial di e si y o i ms’ leade ship has in deciding
he sus ainable composi ion o i ms’ alliance po olios is in es iga ed, de ined as he dis ibu ion
o explo a o y, exploi a i e, and mixed alliances. G ounded in social ca ego iza ion, in o ma ion
elabo a ion, and social con ac mechanisms, acially homogeneous leade ship has a J-shaped ela ion-
ship wi h sus ainable alliance po olio composi ion. Ve y acially homogeneous o he e ogeneous
leade ship leads i ms owa ds main aining mo e explo a o y alliances in hei po olio as opposed
o mode a ely di e se leade ship, which p e e s he sa e y o exploi a i e alliances. Fu he , I explo e
how acially homogeneous leade ship di e s om acially he e ogeneous leade ship in ha he
o me has a highe p opensi y o main ain mo e explo a o y alliance po olios compa ed o he
la e . A wo-s age analysis on a panel o 128 pha maceu ical and so wa e i ms, accompanied
by esponse su ace analysis, yields suppo o ou heo izing. This s udy encou ages schola s o
u he in es iga e he di e en weigh s ha social ca ego iza ion, in o ma ion elabo a ion, and social
con ac exe cise on leade ship di e si y and how hey a e elemen al in i ms’ sus ainable alliance
decision-making.
Keywo ds: acial di e si y; leade ship; social ca ego iza ion; explo a ion; exploi a ion; alliance
po olios; sus ainabili y
1. In oduc ion
The decisions ega ding i ms’ sus ainable alliance po olio composi ions ha e e-
cei ed abundan sus ainable alliance in e es (Hoehn-Weiss and Ka im 2014;Me cal and
Benn 2013;Mou i e al. 2012;Sakai and Kimu a 2024;S e ne and La ie 2015). Alliances
a e c i ically impo an in science- and enginee ing-based indus ies (e.g., pha maceu icals
and so wa e), whe e i ms ely hea ily on expe imen a ion and coope a ion and whe e
a ac ing ex e nal esou ces is c ucial (Koza and Lewin 1998;Lau sen and Sal e 2014;Li
2013;Ro hae mel and Deeds 2004). Sus ainable alliance esea ch claims ha he alliance
decision is g ounded in i ms’ esou ce and capabili y needs and ul ima ely es s wi h
he decision make (Ahuja e al. 2008;Das and Teng 1998;La ie and Rosenkop 2006;
Tsang 1998). In he li e a u e, he composi ion o i ms’ alliance po olios—de ined he e as
he dis ibu ion o explo a o y, exploi a i e, and mixed alliances wi hin a i m’s alliance
po olio—is mainly guided by he i m’s esou ce and capabili y needs (Ahuja 2000;Ahuja
e al. 2008;Baum e al. 2000;Ha yson e al. 2008;La ie 2006). I emains unexplained why
some pha maceu ical and so wa e i ms main ain alliance po olios wi h a composi ion
ha does no seem o ul ill a esou ce o capabili y gap (La ie 2006;Wassme 2010). To
explain he de e minan s o such alliance decisions, his s udy ocuses on leade ship as he
guiding elemen in he alliance po olio composi ion.
P e iously, schola s ha e mos ly conside ed only he op managemen eam (TMT) as
he decisi e ac o in alliance decisions (Cha e jee and Hamb ick 2011;Hamb ick and Ma-
son 1984;Liu e al. 2022;Ha yson e al. 2008;Jiang e al. 2016;Koza and Lewin 1998;Ozcan
and Eisenha d 2009;Schilke and Goe zen 2010), and ew s udies ha e shown ha as a coop-
e a i e pa ne ship, he alliance in ol es a high deg ee o coope a ion and unde s anding
beyond wha he TMT can p o ide (Das and Teng 1998). While he alliance decision-making
Adm. Sci. 2024,14, 279. h ps://doi.o g/10.3390/admsci14110279 h ps://www.mdpi.com/jou nal/admsci
Adm. Sci. 2024,14, 279 2 o 14
p ocess may di e among i ms, he ecen dec ease in he ea ly e mina ion a e o sus ain-
able alliances (Das and Rahman 2010;Das and Teng 1998) can be explained by he p o ound
unde s anding and aci knowledge ha he uppe managemen as a g oup, beyond he
TMT, can ha e. Following his line o easoning, his s udy conside s uppe managemen
as he leade ship g oup (e.g., chie in o ma ion o ice , di ec o o esea ch) in ol ed in
alliance decision-making. In pa icula , acial di e si y a he uppe managemen le el
dic a es how, why, and wha in o ma ion is sha ed wi h an alliance pa ne (Gillespie e al.
2017;Jiang e al. 2016;Sakai and Kimu a 2024). Thei a i ude owa d in o ma ion sha ing,
collabo a ion, and isk ul ima ely guides i ms’ alliance p e e ences, deeply impac ing he
alliance po olio composi ion.
The bu den o alliance decisions is ca ied by i ms’ leade ship and he chie execu i e
o ice (CEO), wi h he leade ship’s di e si y con olling i ms’ decision-making (Gillespie
e al. 2017;Hamb ick 1994;Hamb ick and Mason 1984;Richa d e al. 2004;Zhao e al. 2023).
Di e se leade ship eams a e mo e likely o explo e uncon en ional ideas and challenge
g oup hink, which imp o es a i m’s pe o mance, pa icula ly in unce ain and complex
en i onmen s (Richa d e al. 2004). Racial di e si y in leade ship has also been ound o
in oduce complexi y in o decision-making, as leade s mus na iga e di e ences in cul u al
backg ounds and pe spec i es. While his can lead o iche , mo e in o med decisions, i
can also complica e he p ocess by equi ing addi ional e o o achie e a consensus. This
complexi y is impo an o conside in esea ch on leade ship decision-making because
i a ec s how leade s balance inclusi i y wi h e iciency (Williams and O’Reilly 1998b).
Demog aphic and mo i a ional cha ac e is ics o bo h leade ship and CEOs ha e been
ound o shape i ms’ mo i a ion and a i udes behind s a egic decisions (Richa d e al.
2004). Conside ing he mixed indings o p io li e a u e, his s udy ocuses on he ma ch-
ing/misma ching le els o demog aphic cha ac e is ics and how his phenomenon a ec s
sus ainable alliance decisions. Thus, I ask: How does he acial di e si y o he leade ship
in luence i ms’ sus ainable alliance po olio composi ion?
P e ious li e a u e showed ha when uppe managemen makes decisions, i is in lu-
enced by he demog aphic and psychological cha ac e is ics o i s membe s (Hamb ick and
Mason 1984;Liu e al. 2022). Wi h he p opo ion o acial mino i ies in leade ship g ow-
ing, I conside ha acial di e si y holds signi ican po en ial o in luence he sus ainable
composi ion o i ms’ alliance po olios. Despi e he abundan esea ch on acial di e si y
(And e ski e al. 2014;Richa d e al. 2007), he esea ch linking leade ship acial di e si y
and i ms’ sus ainable alliance po olio composi ion is limi ed. To some deg ee, acial
di e si y has been linked o he i ms’ abili y o compe e and inno a e (And e ski e al.
2014;Cox 1993;Richa d e al. 2004;Vlas e al. 2022a). Racially di e se g oups ha e been
shown o s imula e social con ac , enla ge esou ce access, and di e si y i ms’ pe spec i es
(Blau 1977). Leade ship is esponsible o alloca ing esou ces o alliances and helping he
CEO make alliance decisions. I expec ha a ious con igu a ions o social ca ego iza ion,
in o ma ion elabo a ion, and social con ac p ocesses de e mine how di e en le els o
leade ship’s acial di e si y a o o inhibi he i ms’ pe cep ion o and p opensi y o
isk- aking and oppo unism in alliances.
This s udy endea o s o aise schola s’ a en ion o he po en ial signi icance ha
he leade ship’s acial di e si y may ha e in deciding he composi ion o i ms’ alliance
po olios. This s udy’s hypo heses a e es ed on a panel o pha maceu ical and so wa e
i ms o e a ime ame o six yea s (2006–2011 inclusi e). This s udy claims impo an
con ibu ions o exis ing heo y and esea ch. By building on he ca ego iza ion–elabo a ion
model (CEM) ( an Knippenbe g e al. 2004) and social con ac mechanisms (Blau 1977), his
s udy in es iga es a possible J-shaped e ec o he leade ship’s acial di e si y on i ms’
sus ainable alliance po olio composi ion, whe e acially homogeneous leade ship display-
ing he highes p opensi y o main aining explo a o y alliance po olios is compa ed o
ex eme- o-mode a e acially he e ogeneous leade ship.
Adm. Sci. 2024,14, 279 3 o 14
2. Theo e ical Pe spec i es on Sus ainable Alliance Po olios and Leade ship Di e si y
As a widely dispe sed o ganiza ional o m in he las decade (La ie 2007), alliances
imply ei he a sho - e m, en -seeking ela ionship (exploi a i e) (Li 2013;Sakakiba a
2002), a long- e m, us -based ela ionship (explo a o y) (Jiang e al. 2016;Zana one e al.
2015;Sakai and Kimu a 2024), o a combina ion o bo h (mixed) (La ie 2007;La ie and
Rosenkop 2006). On one side, explo a o y alliances equi e pa ne i ms o be open, o
us , and o ully coope a e o longe pe iods o ime (Inkpen and Tsang 2007). On he
o he side, exploi a i e alliances imply a mo e supe icial ela ionship, usually o a sho e
pe iod o ime and o a speci ic en -seeking pu pose such as licensing o ma ke ing (Jiang
e al. 2016;Lau sen and Sal e 2014).
As he main d i e o sus ainabili y in alliance decision-making, leade ship di e si y
has been mos ly o e looked in s a egic managemen esea ch. While di e si y esea ch
has been abundan , how acial di e si y in e ac s wi h a ious i m-le el ou comes is
complica ed (De D eu e al. 2008;Robe son e al. 2013;Zhou and Rosini 2015). I is impo -
an o unde s and how acial di e si y ela es o o ganiza ional beha io and ou comes
because hese p ocesses shape a i m’s beha io and explain a i m’s choices (Woo en 2008).
Al hough no esea ch o da e has in es iga ed he e ec s o leade ship’s acial di e si y on
i ms’ sus ainable alliance po olio composi ion, he e has been some esea ch showing
ha s a egic choice is a unc ion o leade s’ cha ac e is ics (Golden and Zajac 2001). The
ca ego iza ion–elabo a ion model ( an Knippenbe g e al. 2004) iden i ies wo mechanisms
wi h he po en ial o a ec he composi ion o i ms’ alliance po olios. The i s one is
g ounded in he social con ac heo y (Blau 1977), and he second one d aws simul ane-
ously om he social ca ego iza ion pe spec i es (Taj el 1981;Tu ne e al. 1987) and he
in o ma ion-p ocessing models (Mannix and Neale 2005;Williams and O’Reilly 1998a).
Rega ding he social con ac mechanism, Blau’s heo y o he e ogenei y a gues ha
low and high deg ees o acial di e ences help social in e ac ion among people, while
medium le els o di e si y impede i (Blau 1977). Acco dingly, homogeneous and he e o-
geneous leade ship has ewe ba ie s o o e come in making decisions o eaching an
ag eemen on a s a egic di ec ion compa ed o mode a ely di e se leade ship. The heo y
claims ha g oups in which membe s ha e mo e oppo uni ies o socialize ace lowe
cul u al ba ie s wi h ega d o ac ion and hus de elop ela ions among hem. Membe s o
homogeneous leade ship g oups display simila no ms and p econcep ions (DiS e ano and
Mazne ski 2000), ha e low cul u al ba ie s, and de elop mo e cohesi e g oups in e ms
o communica ion and posi i e social ela ions (Richa d e al. 2004;Vlas e al. 2022b).
In leade ship wi h low acial di e si y, membe s a e likely o sha e uni ied iews
ha hey de elop oge he as a esul o sha ing pe cep ions and de eloping posi i e
eelings o inclusion in he g oup (Dahlin e al. 2005;Vlas e al. 2022a). As he e ogenei y
inc eases, subg oups a e likely o o m, and he ba ie s o social communica ion wi h
membe s o o he subg oups inc ease (Ea ley and Mosakowski 2000). These communica ion
ba ie s impede he low o in o ma ion be ween di e en subg oups (Alexande e al. 1995;
Wie sema and Bi d 1993). The seg ega ion be ween subg oups inc eases each g oup’s ine ia
and dep i es each indi idual subg oup o in o ma ion di e si y. Wi h u he inc eases
in he e ogenei y, g oups become di e se enough o encou age open communica ion. A
e y high le els o di e si y, he e ec o acial mino i ies is minimized. The low o
in o ma ional esou ces is imp o ed, and e en i di e se g oups do no sha e common
pe cep ions and a achmen s o any pa icula con ex , hei social con ac acili a es he
sha ing o in o ma ion among hem.
Beyond he social con ac p ocesses, social ca ego iza ion and in o ma ion elabo a ion
p ocesses a ise in wo k g oups. The ca ego iza ion–elabo a ion model (CEM) ( an Knip-
penbe g e al. 2004) p oposes ha people ca ego ize hemsel es and exchange in o ma ion
depending on how hey ca ego ize o he s in o an in-g oup/ou -g oup (Chen e al. 2002).
The social ca ego iza ion mechanism e e s o people’s endencies o ca ego ize simila
o he s as pa o he in-g oup and dissimila o he s as pa o he ou -g oup (Chen e al. 2002;
Taj el and Tu ne 1986;Tu ne e al. 1987). The in o ma ion-p ocessing mechanism e e s
Adm. Sci. 2024,14, 279 4 o 14
o he exchange and in eg a ion o o he s’ pe spec i es (Ha ison and Klein 2007;Thomas
and Ely 1996; an Knippenbe g e al. 2004;Vlas e al. 2022a,2022b). As de ined by an
Knippenbe g e al. (2004, p. 1011), in o ma ion elabo a ion is “ he exchange o in o ma ion
and pe spec i es, indi idual-le el p ocessing o in o ma ion and pe spec i es, he p ocess
o eeding back he esul s o his indi idual-le el p ocessing in o he g oup, and discussion
and in eg a ion o i s implica ions.” Acco ding o CEM, social ca ego iza ion may ha m
decision-making, while in o ma ion elabo a ion may bene i i . Low di e si y is expec ed
o enhance social in eg a ion h ough low ca ego iza ion bu unde mines decision-making
h ough a lack o access o in o ma ion. A he same ime, high di e si y is expec ed o
weaken social in eg a ion h ough high ca ego iza ion bu bene i decision-making h ough
expanded access o di e se in o ma ion.
3. Hypo heses Fo mula ion
To he deg ee ha leade ship shapes s a egic decision-making (Finkels ein 2009;Ham-
b ick 1994;Hamb ick and Mason 1984;Liu e al. 2022), acially di e se leade ship may ha e
a signi ican impac on he o ma ion o sus ainable alliance po olios. In his hypo hesis,
I claim ha acially homogeneous o he e ogeneous leade ship ha does no su e om
social ca ego iza ion ends o keep mo e explo a o y alliances in he alliance po olios com-
pa ed o a mode a ely di e se leade ship ha su e s om social ca ego iza ion inhibi ing
in o ma ion elabo a ion (Taj el 1981; an Knippenbe g e al. 2004).
To decide in a o o an explo a o y alliance (e.g., an R&D alliance) implies a willing-
ness o engage in a long- e m ela ionship based on us (Sakai and Kimu a 2024). Such
endea o s equi e he abili y and willingness o de elop in e pe sonal communica ion
wi h alliance pa ne s (La ie 2007;La ie e al. 2010;Zaha i and La ie 2013). Homogeneous
and he e ogeneous leade ship a e be e equipped o en e such us -based ela ionships
compa ed o mode a ely di e se leade ship due o a highe abili y o os e social con ac
(Blau 1977). In mode a ely di e se leade ship g oups whe e ca ego iza ion and social
compa ison p ocesses occu , he o ma ion o subg oups aises social ba ie s and hin-
de s communica ion (Ely and Thomas 2001;Taj el and Tu ne 1986). Addi ionally, social
ca ego iza ion p ocesses ha domina e a mode a e le els o di e si y do no allow o
much in o ma ion elabo a ion (Tu ne e al. 1987). The seg ega ion esul ing om social
ca ego iza ion canno lead o explo a o y-seeking beha io because ca ego iza ion nei he
s imula es cons uc i e deba e no os e s in o ma ion elabo a ion (Ba kema and Sh y ko
2007;Ha ison and Klein 2007;Jehn 1997;Richa d e al. 2017;Thomas and Ely 1996;Vlas
e al. 2022a). Cons uc i e deba e is essen ial in eaching a consensus, and mode a ely
di e se leade ship su e ing om seg ega ion canno each i .
Acco ding o Blau’s heo y o he e ogenei y, oppo uni ies o social con ac a e mo e
nume ous in homogeneous and he e ogeneous g oups compa ed o mode a ely di e se
g oups. They a e be e a ans e ing in o ma ion, hus minimizing he chances ha
social ca ego iza ion p ocesses ake place and helping he de elopmen o posi i e social
associa ions among leade ship membe s (Lau and Mu nighan 2005). In homogeneous
leade ship, o example, manage s do no ace cul u al ba ie s o social ac ion and de elop
deep solida i y wi hin hei acial g oup (K. Williams and O’Reilly 1998b). They ela e o
o he simila manage s in he g oup mo e easily, ans e knowledge wi hou inhibi ion,
and de elop s onge in e dependencies (Richa d e al. 2007). These, in u n, make uppe
managemen , as a g oup, mo e e ec i e in making di icul and isky decisions, such as
choosing in a o o a long- e m R&D alliance (Dahlin e al. 2005).
Hypo hesis 1. Leade ship’s acial di e si y displays a U-shaped ela ionship wi h i ms’ sus ainable
alliance po olio composi ion, such ha homogeneous and he e ogeneous leade ship ha e a highe
p ocli i y o mo e explo a o y alliance po olios compa ed o mode a ely di e se leade ship.
In acially homogeneous leade ship, manage s iden i y wi h he social g oup and see
each o he as being pa o he same social ca ego y (Hogg and Te y 2000;Taj el and Tu ne
Adm. Sci. 2024,14, 279 5 o 14
1986;Tu ne e al. 1987; an Knippenbe g e al. 2004). Because membe s o acially homoge-
neous leade ship g oups iden i y as pa o he same in-g oup, hey do no de elop eelings
o disc imina ion agains o he s in he way ha mode a ely di e se o e y he e ogeneous
leade ship do (Ea ley and Mosakowski 2000;Lau and Mu nighan 2005;Richa d e al. 2004).
E en i , in acially homogeneous leade ship, he access o a wide a ay o iewpoin s and
di e se in o ma ion is limi ed, membe s de elop inc eased us among membe s o he in-
g oup, leading o unanimi y in decisions. The lack o in e g oup bias makes homogeneous
leade ship ans e in o ma ion as e , ha e mo e posi i e-a ec i e e alua i e eac ions
o o he s’ beha io , and ha e a highe p opensi y o collabo a e (Ha ison and Klein 2007;
Thomas and Ely 1996). Compa ed o e y he e ogeneous o mode a ely di e se leade ship,
homogeneous leade ship g oups, as a esul o low-ca ego iza ion p ocesses and e y good
social con ac , a e mo e open o engaging in long- e m us -based ela ionships, such as
explo a o y alliances, and less open o engaging in sho - e m exploi a i e alliances (Jiang
e al. 2016).
On he o he side o he spec um, acially he e ogeneous leade ship su e s om
highe ca ego iza ion (Vlas e al. 2022a). A he same ime, acially he e ogeneous lead-
e ship bene i s om he highes elabo a ion de i ed om e y good access o a wide
pool o pe spec i es, skills, and p e e ences (Jackson 1992;Taj el and Tu ne 1986;Tu ne
e al. 1987). Access o an ex ensi e ange o pe spec i es makes hem mo e capable o
iden i ying ex e nal oppo uni ies and making no el s a egic decisions (Ba kema and
Sh y ko 2007;Boeke 1997;Sakai and Kimu a 2024). Howe e , due o high social ca ego-
iza ion, in o ma ion elabo a ion p ocesses a e inhibi ed o some deg ee. E en i acially
he e ogeneous leade ship is likely o pe cei e less isk in en e ing alliances ha in ol e
in ensi e knowledge sha ing—such as R&D alliances—access o a di e si y o pe spec i es
makes hese leade ship g oups mo e awa e o issues possibly ou o hei con ol. Social
ca ego iza ion p ocesses a e also likely o nega i ely a ec he e ogeneous leade ship’s
collabo a ion capabili y, decision-making, and us in he alliance pa ne .
O e all, based on he social ca ego iza ion, in o ma ion elabo a ion, and social con ac
mechanisms, I claim a J-shaped ela ionship be ween he acial di e si y o he leade ship
and he i ms’ sus ainable alliance po olio composi ions.
Hypo hesis 2. Racially homogeneous leade ship has a highe p ocli i y o explo a o y alliance
po olios compa ed o acially he e ogeneous leade ship.
4. Me hods
4.1. Sample and Da a
This s udy commenced by iden i ying US-based science and enginee ing i ms in
he pha maceu ical (SIC 2833, 2834, 2835, and 2836) and so wa e (SIC 7372, 7373, 7374,
and 7375) indus ies. This s udy used Dun and B ads ee , which is conside ed he mos
exhaus i e da abase o US-based i ms (Hmieleski and Ba on 2008;Kallebe g e al. 1990).
To a oid possible inconsis encies in oduced by he Sa banes–Oxley Ac o 2002, I e ie ed
da a o he pe iod a e he in oduc ion o his ac , lea ing a one-yea delay o he new
egula ion o se le. The e o e, he inclusi e yea s o 2006–2011 a e co e ed, wi h some da a
aced back o 2003.
I collec ed alliance da a om he SDC Pla inum da abase (La ie e al. 2010;S e ne and
La ie 2015). In o ma ion on he leade ship’s acial and gende di e si y and con ols was
ex ac ed om he Ins i u ional Sha eholde Se ices (ISS) da abase. Fi m-le el inancial
con ols we e collec ed om he Compus a /CRSP da abase, pa en - ela ed con ols om
he US Pa en and T adema k O ice (USPTO), and alliance-speci ic con ols om he
SDC da abase.
4.2. Analy ical S a egy
The inal da ase is a panel o 128 i ms analyzed be ween 2006 and 2011, including
he alliance and pa en da a aced back o 2003. A one-yea lag o all p edic o and con ol
Adm. Sci. 2024,14, 279 6 o 14
a iables ela i e o he dependen a iable helps mi iga e he po en ial in e dependence
among he mul iple i m obse a ions o e a numbe o yea s. A Hausman es yielded he
ixed-e ec s model as (Hausman 1978). G aphical ep esen a ion in he o m o a esponse
su ace analysis allowed me o be e in e p e and unde s and he ela ionships be ween
he di e en con igu a ions o he p edic o a iable (leade ship’s acial di e si y) and he
ou come a iable (sus ainable alliance po olio composi ion).
To es he hypo heses, I es ima ed a panel ixed-e ec s model wi h yea con ols. I
add essed possible endogenei y wi h a Heckman wo-s age model (Heckman 1979). Fi ms’
endencies o explo e alliances may be in luenced by he inhe en bene i s o his ac i i y,
i espec i e o he leade ship’s acial di e si y. To accoun o his endogenei y e ec , I i s
an a p obi model o he i ms’ explo a o y p opensi y. The p edic ed alues we e used o
compu e he explo a ion in e se Mills a io (λExplo a ion).
To accoun o he explo a ion o sel -selec ion bias, I inco po a ed he in e se Mills
a io as a con ol in he second-s age model. Po en ial in e dependence among obse a ions
was managed by he one-yea lag be ween all p edic o and con ol a iables and he
dependen a iable. To a oid an inc ease in mul icollinea i y, I sequen ially added a iables.
All models’ indi idual VIFs a e well below he ecommended h eshold o 10, sugges ing
ha mul icollinea i y is no signi ican .
Fo he pu pose o unning he panel ixed-e ec s eg ession analysis, I i s mean-
cen e ed he p edic o s and hen used hie a chical eg ession by sequen ially adding
a iables o a oid an inc ease in mul icollinea i y (Bashshu e al. 2011;Richa d e al. 2017).
The polynomial unc ion akes he ollowing o m:
SAPC =b0+b1×RD +b2×RD2+b3×K+e (1)
whe e SAPC deno es he sus ainable alliance po olio composi ion, RD deno es he leade -
ship acial di e si y, K is a ec o o he con ol a iables, and e deno es he e o e m. I
cons uc ed wo models: a base model wi h con ols only and a model wi h bo h i s -o de
and second-o de e ms. To es he i s hypo hesis, I used he model ha included he
i s - and second-o de e ms o he p edic o . A po en ial U-shape e ec was deno ed by
he signi ican nega i e i s -o de e m alongside a posi i e second-o de e m (Edwa ds
and Pa y 1993). To es he second hypo hesis, I engaged in ma hema ical and g aphical
ep esen a ions, as displayed in he ollowing.
4.3. Va iable Desc ip ion
4.3.1. Sus ainable Alliance Po olio Composi ion
The dependen a iable is he i ms’ sus ainable alliance po olio composi ion. This
ep esen s he deg ee o explo a ion unde aken by a i m h ough i s en i e alliance pool.
I cap u ed he i ms’ alliance po olios by pooling all alliances o med in he las i e
yea s (Kogu 1988). Fi ms can o m alliances ei he o explo e and gain access o pa ne s’
knowledge o o exploi and le e age i ms’ exis en knowledge (Koza and Lewin 1998;
La ie and Rosenkop 2006;Ro hae mel and Deeds 2004). I expec ed ha he i ms’ alliance
po olios would be composed o a a ie y o alliance ypes, such as licensing, ma ke ing,
o R&D alliances.
Following Ro hae mel and Deeds (2004), I de ined he alliances ha in ol e join R&D
ac i i ies as explo a o y, alliances ha in ol e join ma ke ing, licensing, esale, o p o-
duc ion ac i i ies as exploi a i e, and alliances ha combine bo h ac i i ies as mixed. Ou
concep ualiza ion assumes ha explo a ion and exploi a ion a e wo sepa a e indica o s
o ac i i ies ha inhibi each o he because hey use esou ces om i ms’ limi ed pools
o esou ces (Uo ila e al. 2009). Thus, when coding he alliance desc ip ions men ioned
in he SDC da abase, he explo a ion alliances we e coded as 1, exploi a ion alliances as
0, and mixed alliances as 0.5 (Hess and Ro hae mel 2011;Ro hae mel and Deeds 2004).
A e summing up all he alliances o med by each i m o e he las i e yea s (Kogu
1988), a a io e lec ing he pe cen age o explo a ion in he i ms’ alliance po olio was
compu ed. Fo example, a i m wi h one R&D alliance and ou licensing alliances will ha e
Adm. Sci. 2024,14, 279 7 o 14
an alliance po olio composi ion index o 0.2, e lec ing ha he i m’s alliance po olio is
20% explo a o y and he emaining 80% is exploi a i e. This index se es as ou dependen
a iable and anges om 0 o 1, wi h alues close o 1 e lec ing he i m’s explo a o y
endea o s and alues close o 0 e lec ing he i m’s exploi a i e endea o s.
4.3.2. Leade ship’s Racial Di e si y
The independen a iable is ope a ionalized using Blau’s he e ogenei y index (Blau
1977). Blau’s index is a commonly used measu e o ca ego ical a iables such as ace
because i cap u es quali a i e dis inc ions o di e si y as a ie y (Ha ison and Klein
2007). I has been widely used o measu e he di e si y o managemen eams (And e ski
e al. 2014;Bunde son and Su cli e 2002;Robe son e al. 2013;Richa d e al. 2007) and
is ecommended by esea che s as a measu e ha a ibu es equal weigh s o all acial
ca ego ies wi hou skewing he dis ibu ion in a o o any ca ego y (Ha ison and Klein
2007;Richa d e al. 2007). In he sample, leade ship’s acial di e si y encompasses i e acial
ca ego ies (e.g., Caucasian, A ican-Ame ican, Hispanic, Asian, and Ame ican-Indian),
and hus, he index heo e ically anges om 0 o 1, wi h an index o 0 e lec ing acial
homogenei y (only one acial ca ego y is ep esen ed) and an index o 1 e lec ing acial
he e ogenei y (all acial ca ego ies equally ep esen ed). The index is calcula ed a he
uppe managemen le el as ollows:
RD =1−∑5
i=1pi2(2)
whe e
pi
ep esen s he p opo ion o membe s o he leade ship in each acial ca ego y, i.
The minimum alue o he leade ship’s acial di e si y in ou sample is 0, and he maximum
is 0.67, wi h a mean o 0.34. The a e age g oup anges om 3 o 32, wi h a mean o 9.73.
4.3.3. Con ols
To minimize he possible al e na i e explana ions, I included con ols o he i ms’
cha ac e is ics, alliance- ela ed a iables, and di e si y- ela ed a iables, as hey may
signi ican ly co ela e wi h a leade ship’s acial di e si y. Speci ically, I con olled o he
i m’s p e ious pe o mance as e lec ed in he i m’s sales g ow h, which is logged because
he a iable is highly skewed. I con olled o he i m’s size as e lec ed in he i m’s o al
asse s (in billions o dolla s) because success ul o la ge i ms ha e a highe p opensi y o
en e explo a o y alliances, il ing he alliance po olio composi ion o a mo e explo a o y
end (Shan e al. 1994). I con olled o he i m alliance expe ience measu ed wi h he o al
numbe o alliances o med in he las 5 yea s (Kogu 1988) because mo e expe ienced i ms
a e mo e capable o de eloping and implemen ing knowledge om hei alliance pa ne s.
I con olled o i m sol ency because i cap u es he a ailable inancial slack esou ces ha
migh make i easie o explo e alliances. Gi en ha leade ship’s gende di e si y has been
ound o sha e a ia ions wi h acial di e si y (Richa d e al. 2003), his was included as a
con ol. I used he same Blau index ha I used o compu e leade ship’s acial di e si y o
measu e leade ship gende di e si y, as well, and included he in e se Mills a io compu ed
in he i s s age (λExplo a ion) o con ol o po en ial endogenei y.
The i s s age o he p obi model ha p o ided he in e se Mills a io eg essed he
i ms’ p obabili y o explo a ion in alliances on i m pe o mance (sales p oduc i i y), i m
age (logged), i m R&D in ensi y, pa en ing expe ience, join en u e expe ience, acquisi ion
expe ience, leade ship g oup’s a e age age, leade ship g oup’s size, CEO duali y, yea 2008
e ec , and yea e ec s. The pa en ing expe ience, join en u e expe ience, and acquisi ion
expe ience a iables we e compu ed as dummies, wi h hem aking a alue o 1 i he i m
pa en ed o engaged in a join en u e/acquisi ion in ha espec i e yea o 0 o he wise.
The yea 2008 e ec a iable akes a alue o 1 i he alliance e en yea comes a e 2008
and a alue o 0 i he alliance e en yea p ecedes 2008.
Finally, in he second s age, he ime e ec s wi h he alliance e en yea and any emain-
ing he e ogenei y wi h ixed-e ec s eg ession we e con olled o .
Adm. Sci. 2024,14, 279 8 o 14
5. Findings
Table 1de ails he means, s anda d de ia ions, and minimum and maximum o all
a iables. Table 2p esen s he Pea son co ela ions be ween a iables.
Table 1. Desc ip i e s a is ics.
Va iables Mean s.d. Min Max
1 Sus ainable alliance po olio composi ion 0.504 0.368 0 1
2 Leade ship acial di e si y 0.342 0.154 0 0.672
3 P e ious pe o mance (ln) 0.394 3.443 −8.70 9.254
4 Fi m size (mil $) 20.97 46.46 0.005 275.64
5 Alliance expe ience 9.117 14.07 1 152
6 Fo m sol ency ( housands) 0.295 3.903 0 69.66
7 Leade ship gende di e si y 0.358 0.124 0.142 0.5
8λExplo a ion 1.356 0.327 0.678 2.19
9 Explo a ion p opensi y 0.574 0.495 0 1
10 Fi m p oduc i i y 426.1 320.9 1.176 1987.35
11 Fi m age (ln) 3.148 0.756 1.098 4.905
12 R&D in ensi y a0.691 0.462 0 1
13 Pa en ing expe ience a0.879 0.326 0 1
14 Join en u e expe ience a0.105 0.307 0 1
15 Acquisi ion expe ience a0.347 0.476 0 1
16 Leade ship g oup a e age age a0.537 0.499 0 1
17 Leade ship g oup size a0.368 0.483 0 1
18 CEO duali y 0.374 0.485 0 1
19 2008 e ec a0.673 0.469 0 1
Numbe o obse a ions = 331. adummy a iable.
Table 2. Pea son co ela ions.
Va iables 1 2 3 4 5 6 7 8
1 Sus ainable alliance po olio composi ion
-
2
Leade ship acial di e si y 0.28 -
3
P e ious pe o mance (ln) −0.45 −0.20 -
4
Fi m size (mil $) −0.06 −0.00 0.18 -
5
Alliance expe ience 0.11 −0.09 0.00 0.42 -
6
Fi m sol ency (000) −0.07 −0.02 −0.02 −0.03 −0.04 -
7
Leade ship gende di e si y 0.39 0.58 −0.35 0.01 −0.05 −0.04 -
8
λExplo a ion −0.45 −0.18 0.64 −0.12 −0.29 0.03 −0.28 -
p< 0.05 o co ela ions in bold; wo- ailed es .
Table 3p esen s he Heckman i s -s age esul s, whe e he i ms’ explo a o y p open-
si y is eg essed on i m pe o mance (sales p oduc i i y), age (logged), R&D in ensi y,
pa en ing expe ience, join en u e expe ience, acquisi ion expe ience, leade ship a e -
age age, leade ship size, CEO duali y, 2008 e ec , and yea e ec s. The andom e ec s
p obi eg ession epo s a highly signi ican Wald chi-squa e o 32.12 (p< 0.001) and a
log-likelihood o −801.20.
Table 4shows he esul s o he panel ixed-e ec s eg ession. I s a wi h he baseline
model wi h he con ol a iables only. Table 4was used o es hypo hesis 1, which dis-
cusses he e ec o acially di e se leade ship on sus ainable alliance po olio composi ion.
Hypo hesis 1 p oposes ha , as a esul o high-elabo a ion and low-ca ego iza ion p ocesses
allowing o social con ac oppo uni ies, acially homogeneous and he e ogeneous lead-
e ship end o keep mo e explo a o y alliance po olios compa ed o mode a ely acially
di e se leade ship. Model 2 con i ms a cu ilinea e ec (
βRD
=
−
2.414, p= 0.015 and
βRD squa ed
= 3.287, p= 0.019). The model shows an imp o emen in he baseline model
R-squa e o 2.78%. I conclude ha hypo hesis 1 is suppo ed.
