1
Documen s de eball
The Cos O Owne ship In The
Go e nance O In e i m
Collabo a ions
Josep Rialp C iado
Vicen e Salas Fumás
Documen de eball nº 2000/1
Depa amen d'economia de l'emp esa
2
 Josep Rialp C iado and Vicen e Salas Fumás
Coo dinado documen s de eball:
D . Jo di López Sin as
h p://selene.uab.es/dep-economia-emp esa/documen .h m
e-mail: [email p o ec ed]
Telè on: 93 5812270
Edi a: Depa amen d'economia de l'emp esa
h p://selene.uab.es/dep-economia-emp esa/
Uni e si a Au ònoma de Ba celona
Facul a de Ciències Econòmiques i Emp esa ials
Edi ici B
08193-Bella e a (Ba celona), Spain
Tel. 93 5811209
Fax 93 5812555
3
Feb ua y, 2000
The Cos O Owne ship In The
Go e nance O In e i m Collabo a ions
Josep Rialp C iado
Depa amen d'economia de l'emp esa, Uni e si a Au ònoma de Ba celona
Vicen e Salas Fumás
Dp o. de Economía y Di ección de Emp esas, Uni e sidad de Za agoza
Documen de eball nº 2000/1
A Wo king Pape in he documen s de eball d'economia de l'emp esa se ies is in ended as a mean
whe eby a acul y esea che 's hough s and indings may be communica ed o in e es ed eade s
o hei commen s. Ne e heless, he ideas pu o wa d a e esponsibili y o he au ho .
Acco dingly a Wo king Pape should no be quo ed no e e ed o wi hou he w i en consen o
he au ho (s). Please, di ec you commen s and sugges ions o he au ho (s).
4
The Cos O Owne ship In The Go e nance O In e i m
Collabo a ions∗
Abs ac :
This pape in es iga es he selec ion o go e nance o ms in in e i m collabo a ions aking
in o accoun he p edic ions om ansac ion cos s and p ope y igh s heo ies. T ansac ion
cos s a gumen s a e o en used o jus i y he in oduc ion o hie a chical con ols in
collabo a ions, bu he owne ship dimension o going om “con ac s” o “hie a chies” has
been igno ed in he pas and wi h i he so called “cos s o owne ship”. The heo e ical
esul s, es ed wi h a sample o collabo a ions in which pa icipa e Spanish i ms, indica e
ha he cos o owne ship may o se he bene i s o hie a chical con ols and he e o e limi
hei di usion. E idence is also epo ed o possible complemen a i ies be ween epu a ion
e ec s and o ms o owne ship ha go oge he wi h hie a chical con ols (i.e. join
en u es), in con as wi h he gene ally assumed subs i u abili y be ween he wo.
Key Wo ds: in e i m collabo a ion, p ope y igh s, ansac ion cos , Spanish i ms.
Co espondence add ess:
Josep Rialp C iado
Dep. Economía de la Emp esa
Edi icio B - Campus U.A.B.
08193 - Bella e a (Ba celona) Spain
Phone 34 - 93 581 2258
Fax 34 - 93 581 2555
e-mail: [email p o ec ed].
∗ The au ho s wan o hank he assis an s o he Academy o Managemen Con e ence in Chicago, specially p o esso s
Gau am Ahuja and I an Mane , and also p o esso Es eban Ga cía Canal o hei commen s in ea lie d a s.
5
1. INTRODUCTION
Wha de e mines he bounda ies o he i m is an impo an ques ion in he heo y o economic
o ganiza ion, s ill unde e y li ely in ellec ual deba e (Holms om and Robe s, 1998). The
con en ional dis inc ion be ween i ms and ma ke s, hie a chies and p ices, as he mechanism
which sol e he coo dina ion and incen i e p oblems esul ing om di ision o labo and exchange,
no only is s ill unde he p ocess o academic cla i ica ion, bu i is challenged by he p oli e a ion
o in e media e o hyb id mechanisms o go e nance (Williamson, 1985, 1991; Imai and I ami,
1984; Powell, 1990). When a i m needs a good o se ice o ca y ou i s plans, i may buy i in he
ma ke , de elop i i sel , o join e o s wi h o he agen s, i ms, in wha is o en called in e i m
coope a ion. Now i ms do no exchange in an anonymous ma ke , nei he hey me ge in o a single
en i y, so he coope a i e ac ion is o en iewed as an example o “hyb ids” be ween he ma ke
and me ge o acquisi ions.
A close look a in e i m collabo a ions1 e eals ha hey di e in he go e nance s uc u e chosen
o egula e he collec i e ac ions: some imes such s uc u e elies mo e on ma ke p inciples
(con ac s, p ices); o he s i in ol es p inciples o en ound in he in e nal o ganiza ion o i ms
(hie a chy, plans, o de s). The e o e, he in ellec ual challenge is no only o explain why i ms o
why ma ke s o o ganize and go e n economic ansac ions, bu why i ms choose di e en
go e nance o ms o hei collabo a ions. This pape is w i en wi h he pu pose o con ibu ing o
answe he p e ious ques ion by in oducing in he heo e ical analysis he p oposi ions de i ed
om he heo y o p ope y igh s (G ossman and Ha , 1986; Ha and Moo e, 1990; Ha , 1995)
besides he widely used, so a , heo y o ansac ion cos s (Coase, 1937; Williamson, 1975; 1985,
1991). The p oposi ions a e es ed using da a abou in e i m collabo a ions in which pa icipa e
Spanish i ms.
The esea ch ques ion o his pape is no new. Fo example, Pisano (1989), Pisano and o he s
(1988), Osbo n and Baugh (1990), Gula i (1995) in es iga e why i ms choose equi y exchanges
(mino i y in e es s, join en u es) o suppo collabo a ions. Mo e ecen ly, Oxley (1997) u he
dis inguishes among bila e al and mul ila e al con ac , wi hin he collabo a ions which do no
in ol e equi y exchanges, while Gula i and Singh (1998) sepa a e collabo a ions unde mino i y
equi y in e es s and collabo a ions go e ned unde he o m o join en u es. In all hese pape s,
he basic unde lying amewo k is bo owed om T ansac ion Cos : as any collec i e ac ion,
in e i m collabo a ion in ol es coo dina ion and incen i es needs. Con ac s ha e limi a ions when

6
hey come o sol e hese needs, specially i in e dependencies a e in ense (Gula i and Singh, 1998),
and payo a e di icul o enume a e ex-an e and e i y ex-pos c ea ing app op iabili y haza ds
(Oxley, 1997). The e o e, in ce ain cases, i is mo e e icien (i has lowe ansac ion cos s) o
subs i u e o complemen con ac s wi h some o he p inciples and ules p ope o he hie a chy
(adminis a i e con ols, o mal coo dina ion uni s, au ho i y).
The p ope y igh s heo y has poin ed ou he limi a ions o ansac ion cos s o explain he
bounda ies o he i m wi h he a gumen ha he cos s o he i m (hie a chy), ha e no been
p ope ly expelled ou (Ha , 1995). In o he wo ds, ansac ion cos economics has emphasized he
cos s o using he ma ke o go e n exchange, bu i has no p o ided a igo ous explana ion o he
cos s o he i m as a go e nance s uc u e. To ou knowledge, he li e a u e on explaining he
choice o go e nance o ms o in e i m collabo a ions has igno ed so a he p ope y igh s
app oach and, he e o e, we belie e, i has igno ed some basic cos s o in oducing hie a chical
p inciples in he go e nance o collabo a ions. We spell in mo e de ail which a e such cos s in
sec ion wo o he pape , bu o mo i a e he exposi ion le us conside he case o join en u es.
One o he basic p edic ions o he p ope y igh s model is ha join owne ship (each pa y has
e o powe on how o alloca e he asse s) is ine icien when ei he sepa a e owne ship (each pa y
is he owne o some asse s) o single owne ship (all asse s belong o one pa y) a e also easible.
Howe e , a join en u e wi h wo pa ies, each holding 50 pe cen o he sha es, is a widely
di used o m o join owne ship in in e i m collabo a ions. To econcile he empi ical e idence
wi h he heo e ical p edic ion is qui e impo an , i we ake in o accoun he ecogni ion gi en by
he academic p o ession o he p ope y igh s app oach2, and he di usion o balanced join
en u e3.
Sec ion wo o he pape p esen s he heo e ical model and pos ula es he hypo hesis o be es ed.
Besides ex ending ansac ion cos o p ope ly accoun o he “cos s o owne ship”, he pape
combines esul s om he heo y o implici con ac s and he conside a ion o “ma ke powe ”
in e es s o expand he ac o s ha de e mine he choice o he go e nance o m. Sec ion h ee
con en s he desc ip ion o he da a, he a iables, he me hodologies used o es he hypo hesis
(o de ed logi and mul inomial logi models), and he empi ical esul s. The pape closes wi h
discussion o he esul s and conclusions.
7
2. THEORETICAL MODELS
2.1. T ansac ion cos and p ope y igh s
The basic p oposi ion o ansac ion cos s economics es ablishes ha “ ansac ions which di e in
hei a ibu es, a e aligned wi h go e nance s uc u es which di e in hei cos s and compe encies
in a disc imina o y way” (Williamson, 1991, p. 277). The applica ion o ansac ion cos heo y o
explain he go e nance o m in in e - i ms collabo a ions equi es o p ope ly iden i y he ele an
a ibu es in ol ed in he ansac ion, he al e na i e go e nance o ms conside ed and he cos s
unc ions esul ing om using a go e nance o m in each ansac ion.
The ansac ion
In his pape , he uni o analysis ( ansac ion) is wha we label as “in e i m collabo a ion”, de ined
as any non ansi o y ela ion among independen companies which in ol es exchange and/o
sha ing o esou ces and capabili ies o ob ain mu ually bene icial ou comes. Wi h his de ini ion,
we iden i y ansac ions whe e he e a e a leas wo i ms and bo h emain legally independen
en i ies (so he ansac ion does no ake place wi hin a single i m, as would be he case i i was
he esul o a me ge , acquisi ion o in e nal de elopmen ). Also we iden i y ansac ions whe e he
ela ion among he pa icipa ing i ms goes beyond an anonymous spo exchange o goods o
se ice a a ma ke p ede e mined p ice (so e en ha he ansac ion akes place in he domain o
he ma ke i in ol es long e m ela ions wi h iden i iable pa ies). In some cases he collabo a ion
in ol es exchange o an al eady exis ing good o se ice in a long- e m non-spo basis. In o he s,
collabo a ing i ms sha e esou ces o ca y ou ex-no o ac i i ies ei he c ea ing a he same ime a
new en i y ( i m) o wi hou c ea ing i .
The ansac ion a ibu e lis ed as de e minan s o he cos o go e nance, di e ac oss he au ho s.
Williamson (1985) selec ed h ee: he na u e o he in ol ed asse s, gene al o speci ic; he
a ailable in o ma ion, unce ain y and symme y; he equency unde which he ansac ion akes
place. Milg om and Robe s (1992), on he o he hand, indica e i e po en ially ele an a ibu es:
he speci ici y o in es men s, equency and du a ion be ween consecu i e ansac ions;
complexi y and unce ain y abou possible ac ions in he u u e; di icul y o measu ing he
pe o mance along he ansac ion; in e ela ions among ansac ion and/o pe sons in ol ed.
In he p esen pape we shall e e o he le el o “ ansac ion complexi y”, assumed o be an
inc easing unc ion o in e dependencies among ansac ing uni s, le el o asse speci ici y,
unce ain y abou u u e con ingencies and di icul ies o measu ing pe o mance. As in o he s
8
esea ch pape s, he a ibu es conside ed as sou ces o complexi y a e no di ec ly obse able, and
hey will ha e o be app oxima e by a iables ha can be obse ed.
The go e nance o ms
Imai and I ami (1984) made he impo an dis inc ion be ween whe e ansac ions ake place, in he
ma ke o wi hin i ms, and he p inciples used o go e n hem, o ganiza ion p inciples o ma ke
p inciples. In e i m collabo a ion akes place in he domain o he ma ke , as he wo o mo e i ms
in ol ed emain independen , bu some imes hey ely mos ly on o ganiza ion p incipals, such as
hie a chy, adminis a i e con ol, and o mal coo dina ion uni s. Mos o he p e ious wo ks on he
go e nance o collabo a ions ha e made he dis inc ion be ween go e nance o ms ha ely on
con ac s and o ms ha ely on hie a chical con ols. F om an empi ical pe spec i e, con ac s a e
u he di ided in unila e al o bila e al exchange (Oxley, 1997). Wi hin he hie a chy, di e ences
in deg ee ha e been a ibu ed o equi y (mino i y) exchange, and o he c ea ion o a new
independen i m -join en u e- (Gula i and Singh, 1998).
Implici in he no ion o hie a chy he e is he idea o au ho i y and ia (Williamson, 1991). The
li e a u e on in e i m collabo a ion has ex ensi ely desc ibed how join en u es, o example,
imply he in oduc ion o hie a chical p incipals, compa ed wi h he case o con ac s. Howe e , he
implica ions in e ms o alloca ing owne ship igh s ha e been igno ed. This is why we ex end he
choice o go e nance o conside he heo y o p ope y igh s, in oduced in G ossman and Ha
(1986) and summa ized in de ail in Ha (1995).
Agen s pa icipa ing in exchange and/o p oduc ion a e o en o ced o ely on incomple e
con ac s. This is due o he ac ha when he con ac is w i en down i is impossible o ex emely
cos ly o an icipa e all u u e con ingencies, o such con ingencies will be un e i iable by hi d
pa ies (mainly he cou s) when hey ac ually occu . The e o e, incomple e con ac s a e
ambiguous abou wha will happen in ce ain ci cums ances, so a mechanism has o be designed o
esol e such ambigui y. P esumably, a new ba gaining p ocess will be ini ia ed, bu e en in his
si ua ion i has o be an icipa ed wha will happen in case o disag eemen .
The p ope y igh s heo y associa es he ac ual esolu ion o he incomple e con ac s wi h he
alloca ion o “ esidual decision igh s”. To hold such igh s o e he non-human asse s o he
ansac ion means o ha e he igh s o decide on any non-an icipa ed con ingency. In his ega d, o
be he “owne ” o an asse is o hold he esidual decision igh s on he asse . Owne ship is an
ins i u ion o alloca e esidual decision igh s. In a wo ld o comple e con ac s, whe e e e y hing is
p ede e mined ex-an e, owne ship is i ele an . Bu when con ac s a e incomple e and he e a e
9
speci ici ies in he ansac ion, he alloca ion o decision igh s (owne ship) ma e s in e ms o
incen i es o in es mo e o less in he speci ic asse s, be o e he ansac ion akes place.
In he p ope y igh s amewo k, go e nance o ms di e acco ding o how esidual decision igh s
a e alloca ed. When asse s in bo h sides o he ansac ion belong o one o he pa ies we e e o
single owne ship because he decision igh s a e concen a ed in only one pa . This is he si ua ion
c ea ed when wo p e iously legally independen i ms me ge in o a single one. I each o he wo
o mo e ansac ing pa ies has e o powe on he use o he asse s, he go e nance is known as
join owne ship; a 50 – 50 equi y join en u e be ween wo i ms will be a o m o join owne ship.
I he esidual decision igh s a e alloca ed o a hi d pa y no di ec ly in ol ed in he ansac ion,
as i would be he case in collabo a ion whe e some hi d pa y a bi a ion is con empla ed in he
con ac o esol e dispu es, he go e nance o m is known as ila e al go e nance. Finally,
owne ship o he asse s can be dis ibu ed among ansac ing pa ies; sepa a ed owne ship.
A ypology.
Bo owing om ansac ion cos s economics and om he heo y o p ope y igh s, he go e nance
o ms a ailable o exchange and p oduc ion can be desc ibed as ollows; see Figu e 1
Spo Classical Con ac ing co esponds o pu e ma ke exchange whe e p ices a e he main (and
su icien ) in o ma ion o adjus esou ce alloca ion decision in an e icien way. Buye s and selle s
can subs i u e each o he a no cos and he e o e he iden i y o he ansac ing pa ies is i ele an
(anonymi y). Finally, he “quid” and he “quo” o he exchange occu a he same ime, so
exchange only akes place when he e is mu ual compliance.
Long Te m Classical Con ac is a go e nance o m close o he pu e ma ke go e nance as i
con inues o ely on p ices o guide adjus men s and coo dina e decisions. Con ac s a e ai ly
s anda dized and comple e bu he epea ed and long- e m pe spec i e o he ansac ion makes
ele an he iden i y o he pa ies. To subs i u e a pa ne has an impo an cos (a leas you ha e
he sea ch cos s, ying o ind he bes new pa ne ). A ime dis ance will gene ally exis be ween
he “quid” and he “quo” o he ansac ion, and he e o e he o mal and explici con ac plays a
mo e signi ican ole in egula ing he ansac ion.
Neoclassical Con ac ing is a go e nance o m whe e con ac s p o ide a amewo k unde which
p oduc ion o exchange will ake place, and some deg ee o incomple eness and ambigui y is
acknowledged. The owne ship o he asse s is sepa a ed among he ansac ing pa ies, as in he
p e ious wo cases, bu he con ac may include sa egua ds o he ansac ing pa ies in he o m
16
Figu e 2. Rela ionship be ween Go e nance Cos s and T ansac ion Complexi y.
Go e nance Cos s
1
*
X2
*
X
Go e nmen h ough Go e nmen h ough
Long Te m Classical Con ac Hie a chy
Go e nmen h ough
Neoclassical Con ac
Sou ce: Adap ed om Williamson (1991).
Figu e 2, oge he wi h he a ibu es o he T ansac ion which ha e been ecognized as sou ces o
complexi y, allow us o o mula e he ollowing hypo hesis:
H1a: The e exis wo o de ed complexi y h esholds in in e - i m collabo a ions, which
de e mine he espec i e choice o Long Te m Classical con ac ing, Neoclassical con ac ing
and Hie a chical a angemen s.
H1b: Complexi y is posi i ely associa ed wi h coo dina ion needs, as in e dependencies
inc ease, and also wi h incen i e p oblems, in o ma ion asymme ies, esou ce speci ici y and
measu emen cos .
So a , no conside a ion is made o he cos s o owne ship, as he subs i u ion o con ac s by
hie a chy has only conside ed he in oduc ion o adminis a i e and ope a ional mechanisms o
coo dina ion and con ol. As i has been men ioned abo e, o go om “con ac s” o “ i ms”
implies o change he alloca ion o owne ship igh s on non-human asse s and his has implica ions
o e iciency.
Long-Te m Classical Con ac
Neoclassical Con ac
Hie a chy
T ansac ion Complexi y

17
Mo eo e , join owne ship is only e icien , compa ed wi h all owne ship al e na i es, unde
special condi ions. So he p ope y H’x < N’x no longe holds in gene al: sepa a ed owne ship o
physical asse s, which is he case in Neoclassical con ac ing, may be mo e e icien in e ms o ex-
an e incen i es o in es in asse s speci ic o he ansac ion, han join owne ship.
F om he p ope y igh s pe spec i e, hen, hie a chy may be less a ac i e han con ac s a high
le els o complexi y (unce ain y and esou ce speci ici ies), especially when he speci ic asse s a e
human capi al no inco po a ed in o physical capi al in he p ocess o he ansac ion.
H2: When he e a e ex-an e in es men s in human capi al embodied on he pe sons making
he in es men , in e - i m collabo a ions which imply join owne ship o he non human
asse s a e less likely han hose which keep owne ship o asse s sepa a ed among he
collabo a ing pa ies.
Consequen ly, in he case o join en u es, we may ha e o weigh he po en ial bene i s o
in oducing adminis a i e sys ems, which help o sol e coo dina ion needs, wi h he di ec cos s o
hese sys ems AND wi h he indi ec cos s due o he choice o an ine icien go e nance o m om
he poin o iew o he alloca ion o owne ship igh s.
2.2. Collusion oppo uni ies and implici con ac s as de e minan s o he i ms bounda ies
Collusion in e es s
In e - i m collabo a ions may no only espond o e iciency conside a ions aimed o lowe
p oduc ion and ansac ion cos s, bu o acili a e coo dina ion among i ms in he p oduc ma ke
and alle ia e compe i i e p essu es.
I i ms a e i als in he p oduc ma ke he cos s and bene i s o collabo a ion a e mo e di icul o
e alua e since now hey may in e e e wi h he compe i i e p ocess. Veugele s and Kes eloo
(1996) each his conclusion o i ms collabo a ing in R&D ac i i ies. Reynolds and Snapp (1986),
on he o he hand, show ha i ms in he same p oduc ma ke may educe he in ensi y o
compe i ion and ob ain p o i s close o he monopoly solu ion by c ea ing equi y linkages among
hem (c oss-sha eholding and join en u es). The sea ch o mo e ma ke powe and he acili a e o
collusi e p ac ices may hen be addi ional easons why i ms collabo a e.
18
H3: Fo a gi en le el o ansac ion complexi y, in e - i m collabo a ions wi h equi y linkages
a e mo e likely when he e a e oppo uni ies o a enua e p oduc ma ke compe i ion among
collabo a ing i ms.
Implici con ac s
The economic analysis o in e - i m collabo a ions as a pa o he b oade ques ion o explaining
he bounda ies o he i m (ma ke , in eg a ion, hyb ids) has conside ed so a he selec ion among
explici con ac s (mo e o less comple e) o egula e he ansac ions. Howe e , collabo a ing
pa ies may also ely on implici con ac s, i.e. con ac s ha a e sus ained by mu ual p omises o
good ai h, non-oppo unis ic beha io and p io i y o collec i e in e es o e indi idual ones.
T ansac ing pa ies a e willing o pa icipa e in ansac ions go e ned by implici con ac s because
he e a e sa ings in cos s compa ed wi h he use o explici con ac s sus ained by he cou s and
legal p ocedu es, and because each one belie es ha he o he pa ies will comply wi h he
p omises o no ac ing oppo unis ically. The educ ion in cos s is ob ious i we ake in o accoun
ha con ac s do no ha e o be w i en down and ha no ex-pos e i ica ion by hi d pa will be
implemen ed (since he e is no o mal con ac , e i ica ion is excluded as a possibili y).
The c edibili y o p omises o “good beha io ” will be sus ained by he p esence o social no ms,
Coleman (1990), Kandel and Lazea (1992), and by he egois ic in e es o p ese ing a good
epu a ion. I one pa y abuses o he us o he o he , who accep ed o ansac unde he
go e nance o an implici con ac , i may be b oken a social no m which es ablishes wha a e
conside ed he desi able pa e ns o beha io om he poin o iew o he collec i i y. Social
p essu es in he o m o sen imen s o “guil ” and/o “shame” make p omises c edible as pa ies
know ha hose making he p omises will y o a oid such social sanc ions.
Repu a ion helps o sus ain implici con ac s, as pe sons build up belie es abou u u e beha io o
he o he s based upon obse ed beha io in he pas (K eps, 1990). I one agen has abused o he
us o he o he s in he pas , i is expec ed ha no one will wan o be in ol ed in implici
con ac s wi h such agen in he u u e. The po en ial sho - e m bene i o he abuse may be o se
by he u u e losses, as he abuse will no be able o pa icipa e o he ad an ages o implici
con ac s. In ac , some ansac ions may gene a e so high ansac ion cos s i hey ely on explici
con ac s ha hey can only be iable i implici con ac s a e easible. The e o e, he access o such
ansac ions will be es ic ed o hose whose epu a ion induces in o he s belie es o good ai h in
hei beha io .
19
In he in e - i m collabo a ion li e a u e, implici con ac s can be conside ed subs i u es o explici
con ac p o isions in Neoclassical con ac ing, and/o subs i u es o oice and o he sa egua d
mechanisms agains oppo unis ic beha io when Hie a chies subs i u e con ac s. The e o e, o a
gi en le el o ansac ion complexi y i is expec ed o obse e less hie a chical go e nance
s uc u es in collabo a ions whe e implici con ac s a e easible han in hose whe e hey a e no
(Gula i, 1995; Gula i and Singh, 1998; Oxley, 1997).
H4a: Fo a gi en le el o ansac ion complexi y, he likelihood o con ac ual sa egua ds and
hie a chical con ols in in e - i m collabo a ion is lowe when collabo a ions a e ca ied
o wa d unde condi ions which a o social no ms and epu a ion e ec s, as sus ainable o
implici con ac s.
No ice, howe e , ha , as indica ed abo e, he p ope y igh s heo y inds complemen a i ies
be ween implici con ac s and some o ms o hie a chical go e nance s uc u es conside ed in
in e - i m collabo a ions, such as join en u es, as a case o join owne ship. Recall ha he
e iciency o join owne ship as a o m o go e nance is eco e ed when epu a ion conside a ions
a e in oduced in o he analysis. The eason is ha , wi h join owne ship, i is possible o lowe he
pa ies payo s in he case o con ac abuse and, consequen ly, inc ease he u u e losses which a e
compa ed wi h sho - e m bene i s o de ec ing; Halonen (1994).
H4b: The likelihood o join owne ship in in e - i m collabo a ions wi h pe son-embodied
speci ic human capi al inc eases unde condi ions ha a o epu a ion e ec s.
3. EMPIRICAL ANALYSIS
The da a used in his wo k come om a de ailed s udy o he in e - i m collabo a ions announced in
Spanish leading economic newspape s and magazines du ing he h ee-yea pe iod o 1990 ill
1992. The condi ion o an announcemen o be included in he da abase was ha a leas one o he
pa icipa ing i ms had o be om Spain. When possible, he desc ip ion o he collabo a ion and o
he selec ed go e nance s uc u e was comple ed wi h in o ma ion coming om di e en sou ces.
Finally, a o al o 1148 e idences o collabo a ions we e iden i ied and comple ely desc ibed wi h
he a iables lis ed below.
Va iables
Dependen Va iable
The dependen a iable is he go e nance o m o he collabo a ion. The coding o his a iable
p oceeds as ollows:
20
i) Long Te m Classical Con ac = 0, when he collabo a ion is an Exchange Ag eemen
ii) Neoclassical Con ac = 1, when he collabo a ion is an Alliance
iii) Hie a chy = 2, when he collabo a ion in ol e a C oss-Sha eholding o a Join Ven u e.
An in e i m collabo a ion is iden i ied as an Exchange Ag eemen when i in ol es he supply o a
p oduc , se ice o know-how om one i m o he o he in a ecu en basis; when i ms sha e a
gi en acili y o ma ke ing, dis ibu ion o a e sale se ices; echnology licensing; subcon ac ing
o manu ac u e o R&D.
Fo example, since 1995 F eixene , he Spanish i m leade o he spa kling wine ma ke , is
collabo a ing wi h he Ge man dis ibu o Eckes. The ela ion is based on a dis ibu ion con ac
among bo h i ms. The ime pe iod ha was ini ially es ablished we e h ee yea s (a he end o his
pe iod bo h i ms could main ain he ela ion i hey we e in e es ed in). The con ac clauses
es ablish he supply o he p oduc (speci ying all hei cha ac e is ics) and some aspec s o he
ma ke ing p og am. Fo example, Eckes mus sell a minimum o bo les. I hey could no do i ,
F eixene would be able o cancel he con ac , compensa ing Eckes wi h an amoun equi alen o
one-yea p o i s8.
Collabo a ion is classi ied as an Alliance when i ms se up a p ojec ha implies he p oduc ion o
de elopmen o a non-exis ing p oduc o se ice. No new legally independen en i y is c ea ed and
he p ojec may ha e a limi ed o open ime ho izon. Fo example, Telmex, a Mexican i m, and
Tele ónica de eloped a p ojec wi h he pu pose o inally c ea ing a digi al sys em ope a ed
h ough a subma ine cable o op ic ibe . No ime limi was es ablished in he con ac .
The exchange o mino i y in e es , C oss-Sha eholdings, and he c ea ion o new en i ies, Join
en u es, a e he collabo a ions mo e easy o iden i y and p ope ly sepa a e om Alliances and
Exchange Ag eemen s. Fo example, Banco de San ande and Royal Bank o Sco land exchanged
equi y ( he o me bough 9.9 pe cen o sha es o he la e and Bank o Sco land bough 4.9 pe
cen o Banco San ande ) o do join in es men s in Eu ope and exchange expe ience and
echnology. On he o he hand, Ace inox, a Spanish company, and A mco, a US i m, join ly
c ea ed he i m No h Ame ican S ainless, an example o join en u e.
The codi ica ion o he a iables does no explici ly ecognize he di e en o ms o owne ship.
Howe e no ice ha Exchange Ag eemen s and Alliances a e cases o Sepa a e Owne ship as
collabo a ing i ms con ol he esidual decisions igh , i any, o he physical asse s in ol ed in he
collabo a ion. C oss-sha eholdings and Join en u es, on he o he hand, a e ea ed ini ially as
21
o ms o Join Owne ship. Ideally, i would be o in e es o know in which cases he pa ies ha e
indi idual e o powe on he asse s combined in he collabo a ion and/o he ac ual sha es hold o
each o hem in he join en u e, bu such in o ma ion was no a ailable. No ice, howe e , ha
Ge inge and He be (1989) and Killing (1983) epo empi ical e idence on he ac ha , in
gene al, he c ea ion o a join en u e inco po a es he igh o e o o e s a egic decisions by he
pa ne s. We ely on his e idence o assume ha his e o powe is p esen in mos o he join
en u es o ou sample, al hough his e idence has no been con as ed. To see he obus ness o he
esul s o his assump ion some sensi i i y analysis will be pe o med, o example, excluding
c oss-sha eholdings as a o m o join owne ship.
Explana o y a iables
In his pape we conside ha i is possible o assign a mo e o less deg ee o complexi y o he
obse ed ansac ion depending upon he unce ain y, he asse s speci ici y, in e dependencies,
numbe o pa ies, measu emen p oblems, e c. The assump ion pos ula es ha mo e complex
ansac ions p esen highe alues in hese unde lying a ibu es.
Mos o en, empi ical da a do no allow measu ing di ec ly he heo e ically pos ula ed a ibu es;
he e o e, we ha e o manage wi h he app op ia e p oxies. In ou pa icula case, he a ibu es o
he ansac ion ac ually obse ed a e: he numbe o collabo a ing pa ies; whe he he e a e
mul iple ac i i ies con empla ed in he ansac ion o only a single one; whe he he collabo a ion
in ol es R&D ac i i ies o no ; whe he he collabo a ion had o do wi h he expansion in he same
p oduc -ma ke whe e he i ms a e al eady ac i e o , o he con a y, in ol es di e si ica ion in o
new ones. We now ela e hese obse able a iable a ibu es wi h he sou ces o complexi y.
The numbe o independen pa ies in ol ed in he ansac ion will a ec he complexi y o such
ansac ion as he numbe o po en ial in e ac ions, and he e o e in e dependencies among hem,
will inc ease wi h such numbe . So, he numbe o pa ne s is conside ed o ha e posi i e
in luences on he coo dina ion needs. The p oblems o measu ing indi idual con ibu ions o he
collec i e ac ion, which acili a e ee iding beha io (Alchian and Demse z, 1972), will also likely
inc ease.
As he numbe o ac i i ies o be pe o med wi hin he in e - i m collabo a ion inc eases, i is likely
ha he numbe o in e dependencies will also inc ease. Collabo a ions wi h mul iple ac i i ies
(coded wi h 1) as compa ed wi h hose wi h a single ac i i y (0), ha e mo e coo dina ion needs.
Fu he mo e, he likely in e ela ions among hem will make mo e di icul o spell ou he

22
con ibu ions and pay-o o each pa ne and, he e o e, i will inc ease he isks o exp op ia ion,
compa ed wi h hose when only one ac i i y is pe o med. The e o e i is possible o associa e his
obse ed a ibu e wi h highe deg ees o ansac ion complexi y.
The wo o he a ibu es a ec ing he complexi y o he ansac ion a e he ealiza ion o echnology
and R&D ela ed ac i i ies (1), compa ed wi h he absence o such ac i i ies (0), and whe he he
pu pose o he collabo a ion is o di e si y in o new p oduc -ma ke (1), compa ed wi h expansion
in he cu en ones (0).
The p esence o R&D ela ed ac i i ies in he collabo a ion sugges ha he ansac ion in ol e
p oduc ion and ans e o knowledge, which is di icul o p o ec wi h con ac s. In his si ua ion,
ansac ion complexi y inc eases because (A ow, 1969; Williamson, 1985):
i) Fo ixing a p ice i is necessa y o know he p oduc , bu once he knowledge is acqui ed
he e is no poin in paying he p ice.
ii) The knowledge may be non-codi ied and o en ex-an e unce ain, so he ou comes o he
ansac ion a e di icul o enume a e and e i y.
iii) The gene a ion o knowledge o en in ol es sunk cos s and he e o e speci ic esou ces o
he ansac ion.
Finally, i will be assumed ha inno a ion h ough R&D ac i i ies is likely o ha e mo e pe son
embodied human capi al han o he ac i i ies. Inno a ion is highly labo in ensi e and he cos s
incu ed on i a e widely ecognized as sunk cos s (S igli z, 1987).
In e - i m collabo a ions which imply di e si ica ion in o new p oduc s and ma ke s will be
associa ed wi h mo e ansac ion complexi y han hose which only expand cu en p oduc s-
ma ke s o he pa ies, because i is expec ed highe unce ain y in he o me han in he la e .
Expansion will allow he ansac ing pa ies o d aw om p e ious expe iences a he ime o
enume a ing possible con ingencies o ake place along he ansac ion. Di e si ica ion, on he o he
hand, implies newness and less po en ial o an icipa ing dis up ing e en s and o e alua ing hei
likelihood o occu ence.
The condi ions ha a o social p essu e and epu a ion e ec s as sus ainable o implici con ac s,
a e p oxies by he a iables na ionali y o he collabo a ing pa and loca ion o he ac i i ies o be
ca ied ou . Social no ms a e mo e likely in cul u ally homogeneous social communi ies. Mo eo e ,
geog aphical p oximi y will acili a e he obse a ion o he beha io o he pa ne s and check
whe he such beha io is consis en o no wi h he p emises. Obse a ion o he beha io , o a
23
leas , o he consequences o i , is necessa y o apply social sanc ions such as shame, and o s a
he applica ion o economic penal ies by excluding he de ia ing pa om u u e collabo a ions.
The pa ne na ionali y a iable is di ided in i e ca ego ies: Spanish, Eu opean Union, Uni ed
S a es, Japan and O he coun ies. Implici con ac s a e expec ed o be mo e iable when pa ne s
a e Spanish o Eu opean (cul u al p oximi y). I may also be a gued ha highe socio-cul u al
dis ance among collabo a ing pa ne s may di icul collabo a ions ha in ol e s ong in e pe sonal
ela ions. So, ins ead o in oducing mo e hie a chical con ols, i ms choose o ely on mo e
s anda dized con ac s such as licensing. So, socio-cul u al dis ance may a o bo h, long- e m
con ac s and hie a chies.
Collusion oppo uni ies a e p oxied by a a iable ha e e s o he na u e o he ela ions among
collabo a ing i ms. A dis inc ion is made be ween ho izon al (1) and e ical (0) ela ion. In he
i s case, collabo a ing i ms belong o he same s age o he indus y’s alue chain and he e o e
a e likely o compe e in he same ma ke . Ve ical ela ions occu when i ms belong o consecu i e
s ages o he alue chain. The sea ch o ma ke powe , as a possible objec i e o some o m o
equi y linkages among i ms, can be expec ed unde ho izon al ela ions.
In o ma ion is also a ailable on he economic sec o whe e he collabo a ion akes place.
Manu ac u ing, Cons uc ion and Se ice sec o s a e widely ep esen ed. I is acknowledge ha
pa en p o ec ion di e ac oss indus ies (Oxley, 1997; Gula i and Singh, 1998), so he legal sys em
will p o ec in a di e en deg ee he in e es s o he pa s depending upon he economic sec o
whe e he collabo a ion is de eloped. Fu he mo e, acco ding o Ha igan (1985) apidly changing
echnological de elopmen induces he o ma ion o somewha mo e in o mal o ms o
collabo a ion such as non-equi y ag eemen s. As indus ies become ma u e, mo e o mal modes o
collabo a ion such as join en u es become he p e e ed ones. So, in choosing pa icula modes o
collabo a ion, he le el o echnological sophis ica ion o indus ies plays an impo an ole
(Hagedoo n and Na ula, 1996, p. 280). Osbo n and Baughn’s (1990) su ey sugges ha
echnological s abili y o indus ial sec o s is a c ucial ac o in explaining di e en pa e ns o
equi y and non-equi y pa ne ships. The es ima ion will con ol o sec o e ec s including sec o
dummies as explana o y a iables.
Finally, wo ime dummies, one o he yea 1990 and o he o he yea 1991 (so he omi ed yea
is 1992), a e included as con ol a iables o accoun o possible mac oeconomic ac o s common
o all collabo a ing i ms in a gi en yea . Table 1 summa izes he explana o y a iables used in he
analysis and p o ides basic desc ip i e s a is ics o each o hem.
24
Inse Table 1.
O e all, 26 pe cen o he collabo a ions a e Exchange Ag eemen s, 32 pe cen Alliances and 41
pe cen in ol e equi y in e es s (C oss-sha eholdings and join en u es). The dominan o m o
collabo a ion has he ollowing a ibu es: i in ol es ac i i ies di e en om R&D, i is c ea ed
wi h he pu pose o ma ke expansion, he pa ne is om ou side om Spain, he collabo a ion
akes place in a manu ac u ing sec o and he i ms in ol ed ha e an ho izon al ela ion.
Me hodology
To es he hypo heses we shall use wo me hodologies: o de ed logi and mul inomial logi choice
models.
Le xi be he le el o complexi y o collabo a ion i whe e:
Xi = βyi + εi(3)
The ec o yi includes all he a iables (p oxies) conside ed as sou ces o complexi y ( oge he wi h
possible con ol e ms); he ec o β includes he weigh s a ached o each sou ce o complexi y;
and εi is he e o e m.
The heo e ical model, syn hesized in condi ions (1) and (2), assumes ha he unobse able
a iable xi de e mines he choice o one o he o de ed go e nance o ms as i alls in one o he
ollowing disc e e in e als:
i) I xi < µ0, Long Te m Con ac (=0) is selec ed
ii) I µ0 < xi < µ1, Neoclassical Con ac (=1) is selec ed
iii) I µ1< xi, Hie a chy (=2) is selec ed (4)
The o de ed logi s a is ical analysis es ima es he ec o o pa ame e s β aking in o accoun he
obse ed cha ac e is ics o he collabo a ion yi and he mapping in (4). I assumes ha he
unde lying p obabili y dis ibu ion o εi is no mal.
Accep ance o hypo hesis 1 equi es ha he h esholds le els µ0 and µ1, wi h µ0 < µ1, a e
con i med by he s a is ical es ima ion and ha he a iables conside ed associa ed wi h complexi y
do in ac ha e a β coe icien posi i e and signi ican .
Ou model de i es he h esholds µ0 and µ1 assuming p ope ies (1) and (2) o he unde lying and
unobse able cos unc ions. The mul inomial logi choice model p o ides an s a is ical
25
me hodology which is based upon he compa ison o he u ili ies (nega i e cos s), associa ed wi h
each o he al e na i e go e nance o ms, gi en a ec o o ansac ion’s cha ac e is ics. In
p obabilis ic o m, he model is exp essed as ollows:
Pij = P(G = j / yi) = e αjyi / Σk eαkyi
Whe e Pij is he p obabili y ha collabo a ion i is go e ned by s uc u e j, j = 0, 1, 2; αj is a ec o
o coe icien s which de e mine he impac o he explana o y a iables on he p obabili y ha each
o he go e nance o ms will be selec ed. The a ia ion o αj ac oss go e nance o ms is consis en
wi h he assump ion ha he cos unc ions (u ili ies) a e di e en o each o hem, see Figu e 2.
To es ima e he coe icien αj he u ili y o one o he al e na i es is used as no maliza ion alue. In
ou case, he al e na i e will be he Long Te m Classical con ac ing. The e o e, he pa ame e s o
he o he al e na i es ha e o be in e p e ed in e e ence o he omi ed one. A pa icula alue o
one es ima ed coe icien αlj, indica es he ex en o which he a ibu e l o he ansac ion
con ibu es o he u ili y o go e nance al e na i e j, beyond he con ibu ion ha his a ibu e
would ha e in de e mining he u ili y o he base op ion, Long Te m Classical con ac ing.
The exis ence o he h esholds o complexi y pos ula ed in hypo hesis 1 is he esul o inc easing
ixed cos s and dec easing ma ginal cos s wi h espec o complexi y as go e nance o ms
app oxima e he hie a chy. I he base op ion o compa e wi h is he Long Te m Classical
con ac ing, hen we should expec ha he likelihood o choosing o he go e nance o ms di e en
om he base op ion a lowe le els o ansac ion complexi y dec eases as such go e nance o ms
a e close o he hie a chy. A he same ime, he likelihood o choosing an al e na i e go e nance
o m compa ed wi h ha o choosing he base op ion, as complexi y inc eases should inc ease
mo e in ensi ely as mo e hie a chical is he al e na i e o m. The e o e, we expec α1 < α2.
The mul inomial logi choice model allows o ano he way o es H1a, and also will be he
me hodology used o es he o he hypo heses as hey a e o mula ed in e ms o compa isons o
likelihood alues.
4. RESULTS
The i s es ima ion co esponds o he o de ed logi model; Table 2. The s a is ical package
LIMDEP 5 used o such es ima ion, ixes a h eshold alue o µ0 = 0, and he e o e, in Table 2
only he es ima ed alue o µ1, he second h eshold, is epo ed.
32
h eshold alue o he a iable which de e mines he choice o he go e nance o m. Howe e , such
h eshold is he co ne s one o he heo y, as ou Figu e 2 makes clea . Mo eo e , ou in e p e a ion
o he a iable which de e mines he choice o go e nance in e ms o complexi y o he ansac ion
as his complexi y alls wi hin ce ain limi s, h esholds alues, i is mo e consis en wi h he heo y
ha he in e p e a ion gi en in Oxley (1997, p. 401). Secondly, we es ablish a de ailed link be ween
he coe icien s es ima ed in he mul inomial logi model and he o de ed logi esul s. Such link is
possible because hese coe icien s can be ela ed o he di e ences in ixed cos s and in a iable
cos s o go e nance, ac oss go e nance o ms, as a unc ion o any a iable ha makes he
ansac ion mo e o less complex in he e ms es ablished in he pape .
As i is gene ally he case in p e ious li e a u e and hypo hesized in he heo y sec ion, he pape
also inds e idence ha he cul u al and social p oximi y o he pa ne s, measu ed by hei
geog aphical o igin, may in luence he choice o he go e nance o m. P oximi y a o s in gene al
Neoclassical con ac ing, i.e. he in e media e o m be ween Classical (Long-Te m) con ac s and
hie a chical con ols. The e o e, collabo a ing i ms may pe cei e Classical con ac s and he
p esence o equi y in e es s as go e nance o ms which p o ide mo e p o ec ion agains ex-pos
oppo unism han Neoclassical con ac s, and hey will be p e e ed when pa ne s a e mo e
“dis an ” so implici con ac s a e less easible.
This esul , oge he wi h o he s, o example, he ac ha equi y in e es s a e mo e likely han
Classical con ac s bu less likely han Neoclassical ones when i ms a e ho izon ally ela ed,
sugges s he ele ance o disagg ega ing con ac s in wo (o mo e) ca ego ies. Compa ing a single
ca ego y o con ac wi h go e nance o ms including equi y in e es s, we would ind ha equi y
in e es s a e mo e likely when collabo a ion is among ho izon ally ela ed i ms and, acco ding o
ou hypo hesis, conclude ha he e may be collusion in e es s behind in e i m collabo a ions.
When Classical and Neoclassical con ac s a e dis inguished, such conclusion does no eme ge.
Ou pape and he esul s ob ained, al hough encou aging in e ms o opening new a enues o
empi ical wo k on he bounda ies o he i m, ha e limi a ions due mainly o he na u e o he da a.
The in o ma ion abou he collabo a ions is qui e limi ed bo h in e ms o he ac ual ea u es o he
selec ed go e nance o m and in e ms o he a ibu es which a e iden i ied as sou ces o
complexi y in he ansac ion. We would ha e bene i ed om a mo e de ailed desc ip ion o he
e ms o he con ac s o make su e ha he coding o he dependen a iable was he co ec one.
This di icul y is o en ecognized in he li e a u e, Oxley (1997, p. 391), bu ne e heless i is
impo an o ecall i and in oduce some cau ion in he conclusions. In o ma ion was also limi ed
on he ype o echnological ac i i ies and in pa icula whe he hey eally in ol ed pe son

33
embodied speci ic human capi al. To ha e his de ailed in o ma ion, as well as he deg ee o
complemen a i y be ween human and non-human asse s in ol ed in he collabo a ion, i will be
c i ical o be able o p og ess in empi ical analysis o he p ope y igh s p edic ions. We hope ha
he p elimina y e idences p esen ed in his pape will encou age u u e e o s o collec hese da a
and e i y he obus ness o he esul s.
34
Table 1. Desc ip i e S a is ics o he Va iables.
Va iable Mean S anda d
De ia ion Minimum Maximum
Numbe o pa ne s (mean) 2.27 0.94 2 16
Coope a ion ange (%) 0.527 0.5 0 1
1. Mul iple ac i i ies
0. One ac i i y
Kind o ac i i y (%) 0.082 0.27 0 1
1. R+D
0. O he kind
Na u e (%) 0.609 0.49 0 1
1. Ho izon al
0. Ve ical
Pu pose (%) 0.07 0.25 0 1
1. Di e si ica ion
0. O he
Nacionali y (%) 0.652 0.48 0 1
1. In e na ional
0. Na ional
Sec o (%)
Manu ac u ing (dummy) 0.436 0.50 0 1
Gene al se ices (dummy) 0.265 0.44 0 1
Ene gy (dummy) 0.112 0.32 0 1
Cons uc ion (dummy) 0.069 0.25 0 1
Finance se ices (dummy) 0.118 0.32 0 1
Pa ne ’s coun y
Eu ope (dummy) 0.405 0.491 0 1
USA (dummy) 0.103 0.304 0 1
Japan (dummy) 0.034 0.183 0 1
O he (dummy) 0.456 0.498 0 1
To al collabo a ions 1148
35
Table 2. O de ed Logi Es ima ion o he Choice o Go e nance.
Model 1 Model 2
In e cep . -0.44775**
(0.19514) -0.76107**
(0.38380)
Nº pa ne s 0.11299
(0.78897E-01) 0.86489E-01
(0.84830E-01)
Mul iple 3.5044***
(0.16573) 3.5683***
(0.17632)
R+D 0.52309**
(0.25124) 0.55994**
(0.25094)
Di e si ica ion 2.1279***
(0.46526) 2.3986***
(0.47620)
Ho izon al 0.41458***
(0.14597)
Spanish 0.27627E-01
(0.22904)
Eu ope 0.37750*
(0.22256)
USA -0.40945
(0.28502)
Japan 0.23452
(0.38883)
Manu ac u ing -0.34588
(0.21862)
Cons uc ion 0.15029E-01
(0.24332)
Ene gy -0.81374***
(0.26971)
Gn al. Se ices -0.77157**
(0.30925)
Da a90 0.52688***
(0.15886)
Da a91 0.36143**
(0.16182)
MU(1) 2.5476***
(0.13570) 2.6503***
(0.13970)
Log. Likelihood -885.6349 -852.6389
Chi squa e. 709.5562*** 775.5482***
Pe cen age co ec classi ica ion 64.8% 66.63%
S anda d E o in pa en hesis.
*Signi icance < 0.1; ** Signi icance < 0.05; ***Signi icance < 0.01.
36
Table 3: Mul inomial Logi Es ima ion o he Choice o Go e nance.
Alliance Join en u e –C oss
sha eholdings Alliance Join en u e –C oss
sha eholdings
In e cep -2.4017***
(0.63206) -4.2256***
(0.73369) -2.3870***
(0.63723) -4.2234***
(0.73981)
Nº pa ne s 0.51583***
(0.19420) 0.51923**
(0.20334) 0.54326***
(0.19848) 0.54672***
(0.2075)
Mul iple 2.5029***
(0.31307) 5.1406***
(0.33287) 2.4637***
(0.31237) 5.1166***
(0.33196)
R+D 1.0269***
(0.27512) -1.3540*
(0.70703) 0.28192
(0.32978) -2.6932**
(1.1236)
Di e si ica ion 2.1697**
(1.0928) 4.3480***
(1.1347) 2.0183*
(1.1106) 4.1415***
(1.1422)
Ho izon al 1.312***
(0.21442) 0.63020**
(0.24799) 1.4202***
(0.21995) 0.70958***
(0.25066)
Spanish 1.1482***
(0.36132) 0.60872
(0.41303) 0.87867**
(0.36733) 0.37770
(0.41682)
Eu ope 0.64788*
(0.34792) 1.0296***
(0.39552) 0.65997*
(0.34714) 1.0509***
(0.39716)
USA 0.47468E-01
(0.42847) 0.17438
(0.50061) 0.85475E-01
(0.42740) 0.19522
(0.50204)
Japan 0.12250
(0.63599) 0.77240
(0.68677) 0.24290
(0.62801) 0.87603
(0.68857)
Manu ac u ing -1.6333***
(0.32403) -0.43285
(0.38995) -1.6766***
(0.32684) -0.47242
(0.39295)
Cons uc ion -0.28044
(0.35296) 0.92721E-01
(0.42069) -0.31080
(0.35632) 0.71995E-01
(0.42405)
Ene gy -0.64690
(0.39997) -1.4997***
(0.49656) -0.55857
(0.40407) -1.4160***
(0.50083)
Gn al. Se ices 1.1349
(0.82259) 0.55134
(0.87064) 1.1882
(0.82892) 0.60150
(0.87738)
Da a90 0.42867*
(0.23815) 0.96341***
(0.28187) 0.39667
(0.24174) 0.94437***
(0.28419)
Da a91 0.38318
(0.23856) 0.60610**
(0.28497) 0.35872
(0.24159) 0.58309**
(0.28676)
R&D – Spanish 2.4489***
(0.67906) 3.5564**
(1.4725)
Log. Likelihood -753.9903 -745.4062
Chi squa e 972.8455*** 990.0137***
Pe cen age co ec classi ica ion 71.95% 71.95%
S anda d e o in pa en hesis.
*Signi icance < 0.1; ** Signi icance < 0.05; ***Signi icance < 0.01.
37
NOTES:
1 In his pape he wo d “collabo a ion” is p e e ed o “coope a ion” because he la e sugges s he con e gence o
objec i es among i ms pa icipa ing in he exchange o p oduc ion and he e o e he absence o con lic s o in e es s.
The ac ha i ms collabo a e does no imply ha each one changes i s p ope objec i es and in e es s and assume
collec i e ones. The e m “alliance” also widely used o e e o any o m o collabo a ion, as in Gula i and Singh
(1998, p. 781), is used in his pape o iden i y a subse o collabo a ion o ms.
2 See o example he issue o he Jou nal o Economics S udies, Janua y (1999).
3 The e ha e been se e al pape s discussing he ad an ages and disad an ages o balanced join en u es e sus non-
balanced one’s (see Kogu (1988) o a good classi ica ion). Bu no e e ences ha e been made o he cos s and bene i s
o owne ship as conside ed in he p ope y igh app oach.
4 A de ailed compa ison be ween Classical and Neoclassical con ac ing is ound in Mcneil (1974, 1978). See also
Williamson (1991, p. 271-273; 1985, p. 70-72).
5 To igno e wo o he possible al e na i es, spo con ac ing and me ge o in e nal de elopmen , when explaining he
choice o go e nance o ms in collabo a ion, may in ac bias he esul s. In any case, he li e a u e on he de e minan s
o “ e ical in eg a ion” is e y la ge. Wo h o men ion because o i s p oximi y o in e i m collabo a ion, is he s udy
o e ical in eg a ion in he con ex o he heo y o he mul ina ional i m; Ga ignon and Ande son (1988) is an
excellen example o he wo k in his ield.
6 The exp ession L’x means he de i a i e o he unc ion L(x) wi h espec o he a iable x.
7 Linea i y o he a iable cos s unc ions is assumed o simplici y and wi hou loss o gene ali y.
8 “Compensa ion due o clien s” sa es comme ce agen s. I is supposed ha he agen c ea es some clien s o he
expo e and, o c ea ing hem, hey mus ecei e non-su ende compensa ion.
9 The esul is he same i numbe o pa ne s is ea ed as a dummy a iable wi h 0 o wo pa ne s and 1 o mo e han
wo.

38
BIBLIOGRAPHY
Alchian, A.A. and Demse z, H. (1972): “P oduc ion, In o ma ion Cos and Economic O ganiza ion”. Ame ican
Economic Re iew, Vol. 62, p. 777 – 795.
A ow, K. J. (1969): "The O ganiza ion o Economic Ac i i y: Issues Pe inen o he Choice o Ma ke e sus
Nonma ke Alloca ion" en The Analysis and E alua ion o Public Expedi u e: The PPB Sys em, Vol. 1, U.S. Join
Economic Commi ee, 91s Cong ess, 1s Session, 47-64. Washing on: U.S. Go e men P in ing O ice.
Bake , G; Gibbons, R. and Mu phy, K.J., (1997): “Rela ional Con ac s and he Theo y o he Fi m”. D a , MIT.
Coase, R. (1937): "The Na u e o he i m". Economica, nº 4, p. 386 - 405.
Coleman, J. (1990): Founda ions o Social Theo y. Camb idge, Mass: Ha a d Uni e si y P ess.
Ga ignon, H. and Ande son, E. (1988): “The Mul ina ional Co po a ion’s Deg ee o Con ol o e Fo eign Subsidia ies:
An empi ical Tes o a T ansac ion Cos s Explana ion”. Jou nal o Law, Economics and O ganiza ion, ol. 4, nº 2, Fall,
p. 305 – 336.
Ge inge , J.M. and He be , L.(1989): “The Impo ance o Con ols in In e na ional Join Ven u es”. Jou nal o
In e na ional Business S udies, Summe , p. 235 – 254.
G ossman, S. and Ha , O. (1986): “The Cos s and Bene i s o Owne ship: A Theo y o Ve ical and La e al
In eg a ion”. Jou nal o Poli ical Economy, 94, p. 691 - 719.
Gula i, R. (1995): "Does Familia i y B eed T us ? The Implica ions o Repea ed Ties o Con ac ual Choice in
Alliances". Academy o Managemen Jou nal, Vol. 38, nº 1, p. 85 - 112.
Gula i, R. and Singh, H. (1998): “The A chi ec u e o Coope a ion: Managing Coo dina ion Cos and App op ia ion
Conce ns in S a egic Alliances”. Adminis a i e Science qua e ly, 43, p. 781 – 814.
Hagedoo n, J. and Na ula, R. (1996): “Choosing O ganiza ional Modes o S a egic Technology Pa ne ing:
In e na ional and Sec o al di e ences”. Jou nal o In e na ional Business S udies , Second qua e , p. 265 – 284.
Halonen, M. (1994): “Repu a ion and Alloca ion o Owne ship”. STICERD Theo e ical Economics Discussion Pape .
London School o Economics.
Ha igan, K. R. (1985): S a egies o join en u es. Lexing on Mass: Lexing on books.
Ha , O. and Moo e, J. (1990): “P ope y igh s and he Na u e o he Fi m”. 98 Jou nal o Poli ical Economy 1119 –
1158.
Ha . O. (1995): Fi ms, Con ac s, and Financial S uc u e. Cla endon P ess. Ox o d.
Holms om, B. and Robe s J. (1998): “Bounda ies o he Fi m Re isi ed”. 12 Jou nal o Economic Pe spec i es, 73 –
94.
Holms om, B. (1999): “The Fi m as a Sub-economy”. Jou nal o Law, Economics and O ganiza ion, Vol. 15, nº 1, p.
74 – 101.
Imai, K. and I ami, H. (1984): "In e pene a ion o O ganiza ion and Ma ke ". In e na ional Jou nal o Indus ial
O ganiza ion, 2, p. 285 - 310.
Kandel, E. and Lazea , E. P. (1992): “Pee P essu e and Pa ne ship”. Jou nal o Poli ical Economy, Vol. 100, nº4, p.
801 -.817.
Killing, J. P. (1983): S a egic o Join Ven u e Success. London: C oom Helm.
Kogu , B (1988). “Join Ven u es: Theo e ical and Empi ical Pe spec i es”. S a egic Managemen Jou nal, Vol. 9, p.
319 – 332.
K eps, D. (1990): Co po a e Cul u e and Economic Theo y, in J. Al and K. Schepsle (eds.) Pe spec i e in Posi i e
Poli ical Economy. Camb idge. Camb idge Uni e si y P ess.
Mcneil, I. R. (1974): “The Many Fea u es o Con ac s”. Sou he n Cali o nia Law Re iew, 47, p. 691 – 816.
Mcneil, I. R. (1978): “Con ac s: Adjus men s o Long-Te m Economic Rela ions unde Classical, Neoclassical and
Rela ional Con ac Law”. No hwes e n Uni e si y Law Re iew, 72, p. 854 - 906
39
Milg om, P. and Robe s, J. (1992): Economics, O ganiza ion and Managemen . P en ice-Hall. Englewood Cli s, New
Je sey.
Osbo n, R. N. and Baughn, C. C. (1990): "Fo ms o In e o ganiza ional Go e nance o Mul ina ional Alliances".
Academy o Managemen Jou nal, Vol. 33, Nº 3, p. 503 - 519.
Oxley, J. E. (1997): “App op iabili y Haza ds and Go e nance in S a egic Alliances: a T ansac ion Cos App oach”.
Jou nal o Law, Economics and O ganiza ion, Vol. 13, nº 2, p. 387 – 409.
Pisano, G. P. (1989): "Using Equi y Pa icipa ion o Suppo Exchange: E idence om he Bio echnology Indus y".
Jou nal o Law, Economics and O ganiza ion, Vol. 5, Nº 1, p. 109-126.
Pisano, G. P.; Russo, M. V. and Teece, D. (1988): “Join Ven u es and Collabo a ing Ag eemen s in he
Telecommunica ion Equipmen Indus y” in In e na ional Collabo a i e Ven u es in U.S. Manu ac u ing, D. Mowe y
(ed.). Camb idge, MA.
Powell, W. (1990): “Nei he Ma ke no Hie a chy: Ne wo k Fo ms o O ganiza ion”. Resea ch in O ganiza ional
Beha io , Vol. 12, JAI P ess, p. 295-336.
Rajan, R. and Zingales, L. (1998): “Powe in a Theo y o he Fi m”. 113 Qua e ly Jou nal o Economics, p. 387 – 432.
Reynolds, R. and Snapp, B. (1986): "The Compe i i e E ec s o Pa ial Equi y In e es s and Join Ven u es". 4
In e na ional Jou nal o Indus ial O ganiza ion, p. 141 - 153.
S igli z, J. (1987): “Technological Change, Sunk Cos s and Compe i ion”. B ooking Pape s on Economic Ac i i y, 3, p.
883 – 937.
Veugele s, R. and Kes eloo , K. (1996): "Ba gained Sha es in Join Ven u es Among Asymme ic Pa ne s: Is he
Ma hew E ec Ca alyzing?". Jou nal o Economics, Vol. 64, Nº 1, p. 23 - 51.
Williamson, O. E. (1975): Ma ke s and Hie a chies, New Yo k, NY: The F ee P ess.
Williamson, O. E. (1985): The Economic Ins i u ions o Capi alism, New Yo k, NY: The F ee P ess.
Williamson, O. E. (1991): "Compa a i e Economic O ganiza ion. The Analysis o Disc e e S uc u al Al e na i es".
Adminis a i e Science Qua e ly, Vol. 36, nº 2, p. 269-296.
40
Issues:
95/1 P oduc i idad del abajo, e iciencia e hipó esis de con e gencia en la indus ia ex il-
con ección eu opea
Jo di López Sin as
95/2 El amaño de la emp esa y la emune ación de los máximos di ec i os
Ped o O ín Ángel
95/3 Mul iple-Sou cing and Speci ic In es men s
Miguel A. Ga cía-Ces ona
96/1 La es uc u a in e na de pues os y sala ios en la je a quía emp esa ial
Ped o O ín Ángel
96/2 E icien P i a iza ion Unde Incomple e Con ac s
Miguel A. Ga cía-Ces ona
Vicen e Salas-Fumás
96/3 Ins i u ional Imp in ing, Global Cul u al Models, and Pa e ns o O ganiza ional Lea ning:
E idence om Fi ms in he Middle-Range Coun ies
Mau o F. Guillén (The Wha on School, Uni e si y o Pennsyl ania)
96/4 The ela ionship be ween i m size and inno a ion ac i i y: a double decision app oach
Es e Ma ínez-Ros (Uni e si a Au ònoma de Ba celona)
José M. Labeaga (UNED & Uni e si a Pompeu Fab a)
96/5 An App oach o Asse -Liabili y Risk Con ol Th ough Asse -Liabili y Secu i ies
Joan Mon llo i Se a s
Ma ía-An onia Ta azón Rodón
97/1 P o ección de los adminis ado es an e el me cado de capi ales: e idencia empí ica en
España
Ra ael C espí i Clade a
97/2 De e minan s o Owne ship S uc u e: A Panel Da a App oach o he Spanish Case
Ra ael C espí i Clade a
97/3 The Spanish Law o Suspension o Paymen s: An Economic Analysis F om
Empi ical E idence
Es eban an Hemmen Almazo
98/1 Boa d Tu no e and Fi m Pe o mance in Spanish Companies
Ca les Gispe i Pellice
98/2 Lib e compe encia en e a egulación en la dis ibución de medicamen os:
eo ía y e idencia empí ica pa a el caso español
E a Jansson
41
98/3 Fi m’s Cu en Pe o mance and Inno a i e Beha io A e he Main De e minan s o
Sala ies in Small-Medium En e p ises
Jo di López Sin as y Es e Ma ínez Ros
98/4 On The De e minan s o Expo In e naliza ion: An Empi ical
Compa ison Be ween Ca alan and Spanish (Non-Ca alan) Expo ing Fi ms
Alex Rialp i C iado
98/5 Modelo de p e isión y análisis del equilib io inancie o en la emp esa
An onio Amo ós Mes es
99/1 A aluació dinàmica de la p oduc i i a dels hospi als i la se a descomposició
en can i ecnològic i can i en e iciència ècnica
Magda Solà
99/2 Block T ans e s: Implica ions o he Go e nance o Spanish Co po a ions
Ra ael C espí, and Ca les Gispe
99/3 The Asymme y o IBEX-35 Re u ns Wi h TAR Models
Mª Dolo es Má quez, and Césa Villazón
99/4 Sou ces and Implica ions o Asymme ic Compe i ion: An Empi ical S udy
Pila López Belbeze
99/5 El ap endizaje en los acue dos de colabo ación in e emp esa ial
Josep Rialp i C iado
00/1 The Cos o Owne ship in he Go e nance o In e i m Collabo a ions
Josep Rialp i C iado, i Vicen e Salas Fumás