G asho , Nils
A icle — Published Ve sion
Familia bu also adical? The mode a ing ole o egional
clus e s o amily i ms in he eme gence o adical
inno a ion
Re iew o Regional Resea ch
P o ided in Coope a ion wi h:
Sp inge Na u e
Sugges ed Ci a ion: G asho , Nils (2024) : Familia bu also adical? The mode a ing ole o egional
clus e s o amily i ms in he eme gence o adical inno a ion, Re iew o Regional Resea ch, ISSN
1613-9836, Sp inge , Be lin, Heidelbe g, Vol. 45, Iss. 1, pp. 17-49,
h ps://doi.o g/10.1007/s10037-023-00199-0
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/323282
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 p://c ea i ecommons.o g/licenses/by/4.0/
ORIGINAL PAPER
h ps://doi.o g/10.1007/s10037-023-00199-0
Re iew o Regional Resea ch (2025) 45:17–49
Familia bu also adical? The mode a ing ole o
egional clus e s o amily i ms in he eme gence o
adical inno a ion
Nils G asho 1
Accep ed: 8 Decembe 2023 / Published online: 19 Janua y 2024
© The Au ho (s) 2024
Abs ac Family i ms a e widely acknowledged o be he mos p edominan o m
o o ganiza ion and hold a g ea ele ance in mos economies. Ne e heless, despi e
hei popula i y, esea ch has hus a yielded inconsis en indings wi h ega d o
hei inno a i e pe o mance. This pape aims o add ess his esea ch gap by o-
cussing on a speci ic o m o inno a ion: adical inno a ion. I seeks o de e mine he
p opensi y o amily i ms o gene a e such inno a ions. Fu he mo e, by conside ing
he he e ogenei y be ween egions and i ms, his pape also in es iga es he po en-
ial mode a ing e ec s o being loca ed in a egional clus e and i m size. Based on
a ious da a sou ces, i is empi ically shown ha amily i ms a e on a e age less
capable o p oducing adical inno a ion han non- amily i ms. Howe e , he co e-
sponding egional con ex ma e s in his ega d. By being loca ed wi hin egional
clus e s, amily i ms can eap he bene i s o localiza ion ex e nali ies, leading o
p oduce mo e adical inno a ions han being loca ed ou side egional clus e s.
Keywo ds Radical inno a ion · Recombinan no el y · Family i ms · Regional
clus e s · Agglome a ion · Fi m size
1 In oduc ion
I is widely acknowledged ha amily i ms a e he mos p edominan o m o
o ganiza ion and ha hey ha e a g ea ele ance in mos economies (Basco and
Nils G asho
nils.g as[email p o ec ed]
1Facul y o Economics and Business Adminis a ion, F ied ich Schille Uni e si y Jena,
Ca l-Zeiss-S . 3, 07743 Jena, Ge many
K
18 N. G asho
Ba ke iˇ
ci¯
u ˙
e2016; Bjugg en e al. 2011; Cappelli e al. 2021). They accoun o
he dominan p opo ion o companies (be ween 65 o 80% o all Eu opean com-
panies) and a la ge p opo ion (on a e age be ween 40 and 50% o all jobs) o
Eu opean p i a e employmen (Eu opean Family Businesses 2021). As a esul , he
phenomenon o amily i m has no only cap u ed he a en ion o esea che s, bu
also eme ged as a subjec o conside able in e es among policymake s. Fo ins ance,
in he ini ial speech as he P esiden -elec U sula on de Leyen highligh ed ha :
“We should ne e o ge ha compe i i e sus ainabili y has always been a he hea
o ou social ma ke economy. We jus called i di e en ly. Think o he amily-
owned businesses all ac oss ou Union. They we e no buil solely on sha eholde
alue o he nex bonuses. They we e buil o las , o pass down gene a ions, o p o-
ide a ai li ing o employees. They we e buil on passion o quali y, adi ion and
inno a ion.” (Speech by P esiden -elec on de Leyen in he Eu opean Pa liamen
Plena y, 2019).1
Despi e hei popula i y and economic ele ance, when i comes o inno a ion,
one o he key ac o s o economic de elopmen (e.g. Ve spagen, 2005), esea ch
on amily i ms has so a only ound a he inconsis en esul s (Calab ò e al.,
2019). To u he esol e he inconsis encies in he esul s, ecen esea ch has high-
ligh ed he need o dis inguish be ween di e en ypes o inno a ion and o s udy
adical inno a ion in pa icula (Calab ò e al. 2019; Hu and Hughes 2020). In
con as o inc emen al inno a ions, adical inno a ions a ise om he syn hesis o
p e iously unconnec ed knowledge pieces (Fleming 2001;Ne ka 2003; Wei zman
1998).2The a ypical combina ion p ocesses also make adical inno a ions mo e
expensi e and mo e likely o ail han inc emen al inno a ions (Ay es 1988;Flem-
ing 2007). None heless, in he e en o success, hey can es ablish a comple ely
new echnological app oach ha leads o u he inc emen al ollow-up inno a ions
and he eby p o ide eno mous economic bene i s (Ahuja and Lampe 2001; A hu
2007). Fo his eason, adical inno a ions ha e a ac ed inc easing in e es om
policy make s (e.g. Sp inD3) and academics (e.g. Shkolnyko a and Kudic 2021),
who, gi en he dis inc i e cha ac e is ics o adical inno a ions, ha e ecen ly high-
ligh ed di e ences be ween di e en ypes o i ms, such as SMEs and la ge i ms, in
hei abili y o gene a e hese inno a ions (e.g. G asho and Kopka 2023). Howe e ,
despi e ecen calls (e.g. Hu and Hughes 2020) and some impo an excep ions (e.g.
Nie o e al. 2015; Schä e e al. 2017), he e has been limi ed esea ch on adical
inno a ion in he speci ic i m ype o amily i ms—especially om a quan i a i e
empi ical pe spec i e. In a i s s ep, his pape he e o e aims o con ibu e o he
ongoing discussion abou he ela ionship be ween amily i ms and inno a ion by
1The speech is also accessible unde : h ps://mul imedia.eu opa l.eu opa.eu/en/p esen a ion-by- he-
commission-p esiden -elec -o - he-college-o -commissione s-and- hei -p og amme-s a emen -by-u sula-
on-de -leyen-p esiden -elec -o - he-ec_I180740-V_ .
2These combina ions a e some imes also called ‘a ypical combina ions’ (e.g. Uzzi e al. 2013).
3In 2019, he Ge man go e nmen ounded he na ional agency “Agen u ü Sp unginno a ionen”
(Sp inD). Fo mo e in o ma ion, please see BMBF (2020).
K
Familia bu also adical? The mode a ing ole o egional clus e s o amily i ms in he... 19
empi ically in es iga ing he ex en o which amily i ms a e mo e likely o c ea e
adical inno a ions4 han non- amily i ms.
Beyond examining i m-speci ic di e ences ( amily s. non- amily i ms) in he
eme gence o adical inno a ion, he speci ic con ex migh also play a ole. Fol-
lowing Basco e al. (2021a), i is a gued ha con ex ual ac o s o he e ogenei y
a e o en o e looked when examining he inno a i e pe o mance o amily i ms.
No conside ing hese con ex ual in luences, howe e , can lead o po en ial misin-
e p e a ions (De Massis e al. 2012). In a second s ep, i is he e o e empi ically
in es iga ed unde which condi ions amily i ms can ac ually gene a e adical in-
no a ion. Based on ecen e o s o link amily businesses wi h he egional con ex
(e.g. Basco 2015; Basco e al. 2021b) and he cu en discussion abou he ole o
egional clus e s o he eme gence o adical inno a ion (e.g. G asho e al. 2019),
he po en ial mode a ing ole o egional clus e s is conside ed. In addi ion o he e-
gional con ex , ollowing he concep o he esou ce-based iew (e.g. Ba ney 1991)
and he sugges ions o De Massis e al. (2012), di e ences in e ms o i m size a e
also examined as a po en ial mode a ing a iable.
To empi ically in es iga e hese wo esea ch gaps, se e al da a sou ces a e com-
bined, pa icula ly i m-le el in o ma ion om he ORBIS da abase and in o ma ion
on in en ions om he PATSTAT da abase. The esul ing da abase includes 10,596
pa en -ac i e companies in Ge many be ween 2012 and 2020. Due o da a limi a-
ions ( ega ding he iden i ica ion o clus e s and amily i ms), he inal da a se
is pooled and a c oss-sec ional analysis is pe o med. Since he co esponding de-
penden a iable is a coun a iable, su e ing om o e -dispe sion, a ze o-in la ed
nega i e binomial eg ession app oach wi h obus s anda d e o s is applied.
By in es iga ing he wo unde lying esea ch ques ions in a quan i a i e way, his
a icle ex ends p e ious esea ch in egional and inno a ion s udies wi h espec o
a be e unde s anding o he e ogenous economic ac o s (in his case amily i ms)
in he con ex o egional clus e s and adical inno a ion, as well as in amily
business s udies wi h ega d o he ele ance o he egional con ex in s udying he
(inno a i e) pe o mance o amily i ms. Besides hese scien i ic con ibu ions, his
pape also o e s p ac ical insigh s o ( egional) policy make s o ully unde s and
he he e ogenei y o amily i ms and hus ha ness he po en ial o amily i ms in
c ea ing adical inno a ions.
The emainde o his pape is s uc u ed as ollows: The subsequen sec ion
in oduces he heo e ical backg ound on amily i ms, adical inno a ions, egional
clus e s and i m size, he eby deduc ing h ee hypo heses. In he hi d sec ion, he
me hodological app oach, he da abase and he co esponding a iables a e desc ibed
in de ail. The ea e , in he ou h sec ion, he main indings a e p esen ed and
discussed. The pape will end wi h concluding ema ks, including limi a ions and
p omising u u e esea ch endea ou s.
4Simila o Cas aldi e al. (2015), he e ms “inno a ion” and “in en ion” a e used in e changeably he e,
because he heo e ical amewo k o ecombinan inno a ion also uses he e m “inno a ion”. Bu , i is
highligh ed ha his s udy ocuses on echnological achie emen s a he han success ul comme cializa-
ion.
K
20 N. G asho
2 Theo e ical backg ound
2.1 Family i ms and adical inno a ions
Inno a ion is gene ally unde s ood o be he esul o ( e)combining exis ing knowl-
edge in a unique way o c ea e some hing new (A hu 2007; Basalla 1988, Cas aldi
e al. 2015). This common unde s anding o inno a ion has i s oo s in Schum-
pe e ’s idea o “Neue Kombina ionen” (Schumpe e 1934) and he ela ed wo k by
Wei zman (1998) in oducing he concep o “ ecombinan inno a ion”, which is
de ined as “(...) he way ha old ideas can be econ igu ed in new ways o make
new ideas.” (Wei zman 1998, p. 333). Ne e heless, he co esponding deg ee o
no el y can he eby be qui e di e en (e.g. Suwala 2017). On he one hand, in-
c emen al inno a ions can be cha ac e ized by a euse and e inemen o exis ing
combina ions, e e ing o exploi a i e sea ch p ocesses (Ma ch 1991;Mewes2019).
They de elop along well-de ined ajec o ies and a e he e o e he no m (Dosi 1982;
Schoenmake s and Duys e s 2010; Ve hoe en e al. 2016). On he o he hand, adical
inno a ions ely on an explo a i e sea ch o and de elopmen o comple ely new
combina ions o knowledge pieces ha ha e no been pu oge he be o e (Fleming
2001; Ma ch 1991;Mewes2019). Since he explo a ion o hese new and p e iously
unknown combina ions is accompanied by highe cos s and highe isks o ailu e
(in echnological as well as comme cial e ms) han inc emen al inno a ions (Ay es
1988;Fleming2007), hey a e ela i ely a e (Fleming 2001; Hesse and Fo nahl
2020). Ne e heless, i success ul, adical inno a ions can es ablish a comple ely
new echnological app oach (A hu 2007; Ve hoe en e al. 2016) leading o s ong
compe i i e ad an ages (e.g. Cas aldi e al. 2015) as well as o he c ea ion o en i e
new ma ke s and indus ies (e.g. G illi sch e al. 2018; Hende son and Cla k 1990;
Tushman and Ande son 1986).
Howe e , he abili y o gene a e adical inno a ion may di e be ween di e en
ypes o i ms (e.g. G asho and Kopka 2023). While esea ch on he ela ionship
be ween i m size and inno a ion in gene al has a long his o y (e.g. Cohen 2010),
pa icula ly in ecen yea s he e has been a g owing in e es in he speci ic case
o amily i ms and hei inno a i eness (Calab ò e al. 2019;U bina ie al.2017).
The economic ele ance o amily i ms has led esea che s o pay a en ion o he
ways in which amily i ms beha e di e en ly om hose o non- amily i ms, and
o he way in which amily-speci ic a ibu es con ibu e o amily i ms’ g ea e
p o i abili y and p oduc i i y (Cappelli e al. 2021; S ae and Thesma 2007)aswell
as hei inno a i eness (Nie o e al. 2015; Zybu a e al. 2021). In ac , schola s ha e
asc ibed some cha ac e is ics o amily i ms ha posi i ely a ec inno a ion, such
as hei long- e m o ien a ion (pa icula ly because hei amily’s o une, epu a ion
and u u e a e a s ake), he a he in o mal knowledge sha ing as well as s ewa dship
beha iou (Mille e al. 2008;Zah a2012, Zellwege 2007).5Bu , a he same ime,
schola s ha e also asc ibed some cha ac e is ics o amily i ms ha nega i ely a ec
inno a ion (Aiello e al. 2020; Hu and Hughes 2020). Fo example, hese include
5Al hough, hese cha ac e is ics may con ibu e in di e en deg ees o he eme gence o inc emen al and
adical inno a ions (Nie o e al., 2015).
K
Familia bu also adical? The mode a ing ole o egional clus e s o amily i ms in he... 21
he equen ly obse ed isk a e sion o amily i ms due o conce ns abou weal h
p ese a ion, pa en al al uism, which may a ou he hi ing o amily membe s
esul ing po en ially in a sho age o quali ied manage s, and he o e all a he low
inno a ion capabili ies (De Massis e al. 2014; Gómez-Mejía e al. 2007; Sciascia
e al. 2015; Si mon and Hi 2003).
Despi e his a ie y o a gumen s (going in bo h di ec ions), in he case o adical
inno a ions, amily i ms a e he e assumed o be less likely o achie e hem. In line
wi h he a gumen a ion by Nie o e al. (2015), i is easonable o assume ha pa ic-
ula ly he isk a e sion and he desi e o p ese e socio-emo ional weal h hinde he
engagemen in explo a i e sea ch p ocesses. Rega ding he la e aspec , con a y o
non- amily i ms, amily i ms a e con on ed wi h he ension be ween economic
and non-economic goals. These non-economic goals encompass among o he aspec s
he weal h p ese a ion o u u e gene a ions, cons an con ol o e he company,
a s ong amily iden i y and in a amily succession, as well as he p ese a ion o
binding social ies o clien s and supplie s (Filse e al. 2018; Gómez-Mejía e al.
2007; Nie o e al. 2015). Since adical inno a ions ely on he pionee ing ecombina-
ion o o me unconnec ed knowledge pieces, which is accompanied by unce ain y
and isk (Fleming 2001;Ne ka 2003) as well as po en ial dis up ion o p e ious
social ies (Nie o e al. 2015), i is likely ha amily i ms a e ela i ely eluc an o
in oducing adical inno a ions. Thus, he ollowing hypo hesis is p oposed:
Hypo hesis 1: Family i ms a e less likely o de elop adical inno a ions han non-
amily i ms.
2.2 Regional clus e s and adical inno a ions in amily i ms
While he geog aphical concen a ion o economic ac i i ies and he possible eco-
nomic e ec s ha e ascina ed esea che s om mul iples disciplines (G asho 2020),
a leas since Ma shall (1920), in he case o amily business esea ch he egional
con ex has o en been neglec ed (Basco 2015). Recen ly, howe e , he e ha e been
e o s o link amily businesses wi h he egional con ex (e.g. Basco e al. 2021b;
Basco and Suwala 2020). In his con ex , based on he no ion o he egional ami-
liness app oach6(Basco 2015), i has been a gued ha he ela i ely high local
embeddedness o amily i ms allows hem o be e exploi he p oximi y dimen-
sions o he co esponding egional con ex (Basco e al. 2021a, Boschma 2005).7
This goes in line wi h p e ious esea ch showing ha he sole loca ion in egional
clus e s is no su icien o ac ually bene i om po en ial localiza ion ex e nali ies,
6Regional amiliness is o iginally de inied as (...) he embeddedness o amily businesses in social, eco-
nomic, and p oduc i e s uc u es wi hin he spa ial con ex and he ype o connec ions ha eme ge and
in e ac wi h egional ac o s (i.e., angible and in angible ac o s) and egional p ocesses (e.g., spillo e s,
in o ma ion exchange, lea ning p ocesses, social in e ac ions, compe i ion dynamics, and ins i u ional dy-
namics) h ough egional p oximi y dimensions (i.e., ela ional, ins i u ional, o ganiza ional, social, and
cogni i e p oximi y) (Basco 2015, p. 260).
7As desc ibed by Basco and Suwala (2021), he e is a esea ch adi ion o add ess amily i ms in egional
s udies in he case o Indus ial dis ic s, which has only ecen ly been aking up again (e.g Cucculelli and
S o ai 2015; Pi ino e al. 2021).
K
22 N. G asho
such as knowledge spillo e s (G asho 2021; He as-Oli e e al. 2018). Th ough
hei s ong egional ies and engagemen , amily i ms a e mo e s ongly embed-
ded in he co esponding egional inno a ion sys em han non- amily i ms (Basco
e al. 2021a; Block and Spiegel 2013; Déniz and Suá ez 2005), making hem mo e
likely o bene i om localiza ion ex e nali ies, pa icula ly om he supply-side
ones (McCann and Fol a 2008). As a esul o hei long-s anding egional p esence
and hei social ela ionships (Basco e al. 2021a), amily i ms can c ea e a ce ain
( egional) epu a ion in e ms o us wo hiness, secu i y and s abili y making hem
mo e likely o a ac local alen s om he specialised labou pool (Hauswald e al.
2016). Ha ing such an access o he specialized labou pool can po en ially lead o
adically new ideas, as he expe ise o local human esou ces can challenge con-
en ional p ocesses and mindse s (Bekke s and F ei as 2008; G asho e al. 2019;
Zucke e al. 2002).8Mo eo e , he high deg ee o egional embeddedness and long-
e m o ien a ion ha e bo h he po en ial o educe he ansac ion and coo dina ion
cos s o coope a ion, e.g. wi h (local) supplie s, leading ul ima ely o mo e ( us ul)
ela ionships and mo e knowledge exchange (Block and Spiegel 2013). The access
o hese knowledge spillo e s can be used o en ich in-house knowledge and he eby
c ea e a he adical new knowledge (Dong e al. 2017; Faems e al. 2005). This holds
pa icula ly ue, since he p onounced social p oximi y may addi ionally allow o
o e come he challenges associa ed wi h coope a ing wi h cogni i e dis an pa ne s
(Adjei e al. 2019;Boschma2005). Fu he mo e, due o he a he close, us ul
and long- e m o ien ed ela ionships o amily i ms (Block and Spiegel 2013; Mille
e al. 2008) i is also likely ha pa icula aci knowledge is exchanged (Adjei e al.
2019), which is ele an o he c ea ion o adical inno a ions (Aud e sch 1998;
Masci elli 2000).
Howe e , i has also been sugges ed ha , o e ime, i ms loca ed wi hin clus e s
may ace (cogni i e) ine ia wi h espec o ma ke and echnological change, which
hampe s adical inno a ion (Hassink 2007; Poude and S John 1996; Schamp 2005).
Fu he mo e, when local ne wo ks hea ily depend on local ace- o- ace con ac s
and aci knowledge, hey become mo e suscep ible o lock-in si ua ions, he eby
pe pe ua ing he ine ia o i ms si ua ed wi hin clus e s (Boschma 2005; Ma in
and Sunley 2003). Mo eo e , in he case o indus ial dis ic s i has ecen ly been
shown ha he po en ial loca ional ad an ages do no necessa ily ma ch well wi h
he i m-speci ic ad an ages o amily i ms due o edundancies, bu ins ead may
e en dec ease he inancial pe o mance o amily i ms (Cucculelli and S o ai 2015;
Pi ino e al. 2021).
Ne e heless, i is impo an o no e ha indus ial dis ic s ep esen a special
o m o a egional clus e , wi h a pa icula emphasis on he social dimension, he eby
making edundancies wi h he i m-speci ic ad an ages o amily i ms mo e likely
(G asho and Fo nahl 2021; Pi ino e al. 2021). Addi ionally, his s udy ocuses on
inno a ion a he han inancial pe o mance, unlike Cucculelli and S o ai (2015)
as well as Pi ino e al. (2021), which a e no necessa ily in e wined (G asho
8Howe e , i has also been highligh ed ha amily i ms end o hi e employees wi h a simila cogni i e
backg ound, which would o cou se limi o e en o se he bene i s om he inco po a ed knowledge o
new employees (B inke ink 2018).
K
Familia bu also adical? The mode a ing ole o egional clus e s o amily i ms in he... 23
and Fo nahl 2021). Fo he conc e e esea ch con ex o his s udy, i is he e o e
easonable o assume ha amily i ms, which a e mo e locally embedded han
non- amily i ms, a e be e capable o exploi he localiza ion ex e nali ies wi hin
egional clus e s (Basco 2015; Basco e al. 2021a), which ha e been shown o a he
p omo e he eme gence o adical inno a ions (e.g. G asho e al. 2019). Thus, he
ollowing hypo hesis is p oposed:
Hypo hesis 2: Being loca ed in a egional clus e inc eases he likelihood o
amily i ms o c ea e adical inno a ions (compa ed o amily i ms loca ed ou side
egional clus e s).
2.3 Fi m size and adical inno a ions in amily i ms
Howe e , amily i ms a e o cou se no a homogenous g oup (Co be a and Sal a o
2004; Filse e al. 2018). In line wi h he esou ce-based iew9(RBV), hey di e in
e ms o hei esou ces10 and capabili ies (Ba ney 1991;We ne e al.2018). One
p ominen aspec ha has been equen ly conside ed in his con ex is he size o
i ms (Nie o e al., 2015). P io esea ch has shown ha i m size is a c ucial ac o
in d i ing o e all inno a ion in i ms (e.g. Cohen and Kleppe 1996).
Ne e heless, when i comes o adical inno a ions, he e a e a he ew indings
and hese a e gene ally inconsis en (Chandy and Tellis 2000). Fo example, in
hei ecen s udy abou he in luence o AI knowledge on he eme gence o adical
inno a ions in i ms, G asho and Kopka (2023) ound di e en esul s wi h espec
o he ole o i m size depending on he unde lying ea u es o AI echnologies
(AI applica ions s. AI echniques). In gene al, he e exis a gumen s in a ou and
agains a posi i e in luence o i m size on he eme gence o adical inno a ions.
While la ge i ms ha e, on he one hand, mo e ( inancial) esou ces and ( echnical)
capabili ies o ac ually de elop his ype o inno a ion, on he o he hand, hey a e
mo e likely o ace ine ia due hei complex in e nal s uc u e, making hem a he
in lexible o new ( a he adical) ideas (Chandy and Tellis 2000; Colombo e al.
2015).
In he case o amily i ms, i m size has also been shown o ma e o i m
pe o mance in gene al (e.g. Cucculelli and S o ai 2015) and i m inno a i eness in
pa icula (e.g. We ne e al. 2018). Howe e , as desc ibed in sec ion 2.1., amily
i ms ha e some unique cha ac e is ics ha could enhance o mi iga e he i m size
e ec compa ed o non- amily i ms. Recen empi ical s udies o ins ance show
ha small amily i ms a e mo e inno a i e han small non- amily i ms, while
9The Resou ce-Based View (RBV) assumes ha esou ces a e unequally dis ibu ed among i ms and
a e immobile, esul ing in a ying esou ce endowmen s and hei pe sis ence o e ime. As a esul o
his imbalance, i ms can po en ially achie e a esou ce-based compe i i e ad an age by le e aging hei
in e nal esou ce base. The e o e, he cen al concep o he RBV ocuses on how i ms can u ilize hei
esou ces o gain a compe i i e ad an age (Ba ney 1991; G asho and Kopka 2023;Newbe 2007).
10 Following he widely used de ini ion by Ba ney (1991), esou ces a e he e de ined as “(...) all asse s,
capabili ies, o ganiza ional p ocesses, i m a ibu es, in o ma ion, knowledge, e c. con olled by a i m
ha enable he i m o concei e o and implemen s a egies ha imp o e i s e iciency and e ec i eness.”
(Ba ney 1991, p. 101).
K
24 N. G asho
he opposi e holds ue o la ge i ms (We ne e al. 2018). Whe he his also
applies o adical inno a ions is s ill unclea . In gene al, i has been a gued ha
he endency o amily i ms o keep he con ol o he business in amily hands
(Aiello e al. 2020; Si mon and Hi 2003), migh imply ela i ely high agency cos s
(e.g. implemen a ion o an incen i e sys em) when hey a e la ge. Con a y, smalle
amily i ms do no ace hese agency cos s, as in luen ial managemen posi ions can
be illed wi h amily membe s, he eby p o iding hem wi h mo e inancial esou ces
ha can be in es ed in inno a i e ac i i ies (We ne e al. 2018). Mo eo e , in ligh
o he ela i ely high le el o owne ship concen a ion wi hin amily i ms i is also
likely ha a ela i ely la ge and complex in e nal s uc u e will slow down decision
making p ocesses, he eby en o cing he ine ia and igidi y o la ge amily i ms
(Aiello e al. 2020;We ne e al.2018). In addi ion, as amily i ms a e less likely o
use isky inancial capi al o a oid losing con ol, hey may ha e mo e p oblems han
(la ge) non- amily i ms in inancing hei g ow h and ela ed inno a ion ac i i ies
(Aiello e al. 2022; Gómez-Mejía e al. 2007; Ke s de V ies 1993).
Despi e he po en ial ad an ages in e ms o mo e ( inancial) esou ces and ( ech-
nical) capabili ies, in he speci ic case o amily i ms i is he e o e easonable o
assume ha he po en ial disad an ages in e ms o in lexibili y ou weigh, so ha
smalle amily i ms a e be e able o in oduce adical inno a ions han non- amily
i ms, while la ge amily i ms a e less able o do so. Thus, he ollowing hypo hesis
is p oposed:
Hypo hesis 3: Fi m size inhibi s he c ea ion o adical inno a ions in amily i ms
mo e han in he case o non- amily i ms.
3 Empi ical backg ound
3.1 Da a
The sample o he empi ical analysis is cons uc ed by using se e al da a sou ces. In
pa icula , simila o p e ious app oaches (e.g. G asho e al. 2020), his s udy com-
bines i m-le el in o ma ion (e.g. owne ship in o ma ion) om he ORBIS da abase,
o e ed by Bu eau an Dijk (B D), and in o ma ion on in en i e ac i i ies (e.g.
echnology classes) om he PATSTAT da abase.11 The esul ing da a se con ains
de ailed in o ma ion abou 10,596 ac i ely pa en ing (i.e. a leas one pa en iled)
o ganisa ions in Ge many be ween 2012 and 2020, o which 8.75% ac ually iled
adical pa en s.12 In pa icula , non- amily i ms a e ela i ely well ep esen ed, wi h
almos wice as many non- amily i ms (604) iling adical pa en s as amily i ms
(323).
11 To ma ch he pa en da a wi h he i m-le el da a om ORBIS, a unique pa en iden i ie is c ea ed based
on in o ma ion om PATSTAT.
12 Table 3in he appendix epo s he dis ibu ion ac oss indus ies.
K
Familia bu also adical? The mode a ing ole o egional clus e s o amily i ms in he... 31
Table 2 Reg ession esul s: Family i ms and adical inno a ion
Dependen a iable
ad_coun ad_coun ad_coun ad_coun
(1.) (2.) (3.) (4.)
Family_dummy – –0.521*** –0.746*** 0.264
– (0.158) (0.134) (0.358)
Clus e _dummy – – 0.535** –
– – (0.254) –
Family_dummy*
clus e _dummy
– – 0.757* –
– – (0.436) –
Log_size_mean –––0.392***
–––(0.044)
Family_dummy*
log_size_mean
– – – –0.066
–––(0.057)
Age_mean 0.012*** 0.012*** 0.009*** 0.001
(0.002) (0.002) (0.002) (0.001)
Independence_dummy 1.075*** 0.923*** 0.986*** 0.437**
(0.295) (0.298) (0.305) (0.209)
esea chin ensi eindus y 0.282* 0.251 0.181 –0.034
(0.151) (0.154) (0.130) (0.099)
Popdens_mean 0.0002** 0.0001* 0.0001 0.00004
(0.0001) (0.0001) (0.0001) (0.0001)
Sha e_academics_mean 0.054** 0.048** 0.055*** 0.022
(0.022) (0.020) (0.019) (0.016)
Cons an –2.536*** –2.113*** –2.272*** –2.983***
(0.811) (0.741) (0.723) (0.848)
In la e: pa _coun –0.249*** –0.024*** –0.231*** –0.212***
(0.032) (0.033) (0.032) (0.039)
Cons an 2.864*** 2.779*** 2.748*** 2.540***
(0.094) (0.098) (0.096) (0.105)
N 10,431 10,431 10,431 10,293
McFadden’s Adj R2 0.184 0.186 0.191 0.225
Log-likelihood –3854.557 –3840.852 –3813.604 –3621.056
Akaike In . C i .*N7727 7702 7661 7270
Robus s anda d e o s in pa en heses
*p< 0.10, ** p< 0.05, *** p<0.01
loca ing in highly u banized egions (a leas in Model 1 and 2) is ad an ageous o
he c ea ion o adical inno a ion by companies.
This can e en ually be explained by he la ge di e si y o di e en ac o s and
knowledge wi hin hese egions which o e s a a he la ge po en ial o knowledge
ecombina ion (Hesse and Fo nahl 2020). In addi ion, he p opo ion o a e age
sha e o pe sons wi h e ia y educa ion and/o wi h S&T occupa ion in he egion,
p oxying he egional abso p i e capaci y, seem o ma e o he eme gence o
adical inno a ion, which is line wi h p e ious s udies (e.g. Hesse 2020).
K
32 N. G asho
Model 2 in oduces he amily i m dummy a iable ( amily_dummy). As assumed
(see Hypo hesis 1), a signi ican nega i e in luence is ound. Family i ms a e he e-
o e on a e age less capable o c ea ing adical inno a ions han non- amily i ms.
This can be explained by he isk a e sion and he desi e o p ese e socio-emo-
ional weal h which bo h hinde he engagemen in explo a i e sea ch p ocesses and
he eby he c ea ion o adical inno a ions (Nie o e al. 2015). Hence, Hypo hesis 1
canno be ejec ed.
By ollowing an “in e ac ionis app oach” (Beugelsdijk 2007), he po en ial mod-
e a ing ole o he egional con ex is addi ionally in es iga ed in Model 3, whe e
a dummy a iable o i ms’ loca ion in a egional clus e is in oduced. As indica ed
by he signi ican posi i e in e ac ion e m (β= 0.757; p= 0.082), being loca ed in
a clus e inc eases he likelihood o amily i ms o c ea e adical inno a ions.25
Due o hei ac i e pa icipa ion and close connec ions wi hin hei egion, amily
i ms a e mo e s ongly embedded in he egional inno a ion sys em compa ed o
non- amily i ms (Basco e al. 2021a; Block and Spiegel 2013; Déniz and Suá ez
2005). As a esul , hey a e mo e likely o eap he bene i s o localiza ion ex-
e nali ies wi hin egional clus e s (Basco 2015; Basco e al. 2021a), which ha e
been shown o os e he eme gence o adical inno a ion (e.g. G asho e al. 2019).
Consequen ly, i can be esumed ha Hypo hesis 2 canno be ejec ed.
Since no all amily i ms belong o a homogenous g oup (e.g. Filse e al. 2018),
in Model 4, he mode a ing in luence o i m size is es ed. While he in luence o
i m size akes he assumed nega i e di ec ion, i is howe e insigni ican . Fi m size
he e o e does no signi ican ly inhibi he c ea ion o adical inno a ions in amily
i ms mo e han in non- amily i ms.26 Ins ead, in he case o non- amily i ms, we
e en ind e idence o a signi ican posi i e in luence o i m size, meaning ha
la ge non- amily i ms a e be e capable o gene a ing adical inno a ion. Indeed,
in bo h cases a signi ican posi i e coe icien is ound o i m size (see Fig. 3in he
Appendix). Fo non- amily i ms he a e age ma ginal e ec (AME) is 0.107, while
o amily i ms i is sligh ly lowe (AME is 0.073), al hough s ill highly signi ican .
Consequen ly, Hypo hesis 3 has o be ejec ed.
O e all, he esul s show ha , on a e age, amily businesses a e less likely o
p oduce adical inno a ion han non- amily businesses. Howe e , he co esponding
egional con ex ma e s in his con ex . Family i ms ha a e loca ed in egional
clus e s can c ea e mo e adical inno a ion, since hey can exploi he ad an ages o
localiza ion ex e nali ies h ough hei s ong egional embeddedness.
5Conclusion
Despi e he ela i ely ex ensi e esea ch on inno a ion in amily i ms (e.g. Calab ò
e al. 2019), i emains unclea whe he amily i ms ha e a g ea e likelihood o
25 In gene al, 9.77% o all amily i ms in he sample (co esponding o 568 i ms) a e loca ed in egional
clus e s. O hese amily i ms, 9.33% (co esponding o 53 amily i ms) c ea ed adical inno a ions,
compa ed o 5.15% (co esponding o 270 amily i ms) in he case o non-clus e ed amily i ms.
26 The co esponding p edic ed ma gins a e illus a ed in Fig. 2in he Appendix.
K
Familia bu also adical? The mode a ing ole o egional clus e s o amily i ms in he... 33
gene a e adical inno a ion and o wha ex en con ex ual a iables mode a e his
ela ionship. This pape add essed hese wo esea ch gaps by combining se e al da a
sou ces o empi ically examine he ela ionship be ween amily i ms and adical
inno a ion as well as he mode a ing e ec s o egional clus e s and i m size.
In summa y, he s udy p o ides h ee main esul s. Fi s , i shows ha , on a -
e age, amily i ms c ea e less adical inno a ion han non- amily i ms. Thei isk
a e sion and desi e o p ese e socio-emo ional weal h may ac in his con ex as
obs acles o engaging in explo a o y sea ch p ocesses, which ul ima ely hinde s he
c ea ion o adical inno a ion (Nie o e al. 2015). Second, he s udy shows ha
he speci ic con ex in egional clus e s can mode a e his ela ionship. On a e age,
being loca ed in a clus e a o s he eme gence o adical inno a ions in amily
businesses. Clus e s he e o e cons i u e a bene icial en i onmen (e.g. h ough hei
localiza ion ex e nali ies) o amily i ms. The po en ial subs i u ion e ec be ween
he a he edundan condi ions enhancing he ad an ages o indus ial dis ic s and
amily i ms shown in ecen s udies (e.g. Cucculelli and S o ai 2015; Pi ino e al.
2021) can he e o e no be con i med o he mo e gene al con ex o egional clus-
e s. Unlike in he case o indus ial dis ic s, egional clus e s do no necessa ily
build on he social and cul u al dimension (G asho and Fo nahl 2021). Hence, i
seems p omising o u u e esea ch o u he di e en ia e egional clus e s acco d-
ing o hei cha ac e is ics (e.g. size, sha e o SMEs, s eng h o social and cul u al
backg ound). Thi d, he s udy shows ha , unlike inno a ion in gene al (e.g. We ne
e al. 2018), he size o i ms does no signi ican ly hinde he c ea ion o adical
inno a ions in amily i ms mo e han in non- amily i ms. Ins ead, i is e ealed
ha in bo h cases la ge i ms a e mo e likely o ile new adical pa en s, which is
consis en wi h p e ious esea ch (e.g. G asho and Kopka 2023) ha explains his
by he ac ha la ge i ms can bene i om mo e in e nal R&D esou ces (O ega-
A gilés e al. 2009; Ramme and Schube 2016).
Ne e heless, his s udy does no come wi hou limi a ions, he eby o e ing ad-
di ional oppo uni ies o u he esea ch. Fi s o all, he unde lying da a base
o de e mining adical inno a ions and knowledge a ie y a e pa en s, which ha e
some d awbacks (e.g. G iliches 1990). Fu u e esea ch could he e o e use al e -
na i e, non-pa en -based da a (e.g. He as-Oli e e al. 2019). Addi ionally, u u e
esea ch could also in es iga e al e na i e pa en -based measu es o he eme gence
o adical inno a ion, e.g. backwa d ci a ions, and i s impac and di usion, e.g. o -
wa d ci a ions (Dahlin and Beh ens 2005; T aj enbe g e al. 1997). Mo eo e , due o
da a cons ain s (wi h espec o he iden i ica ion o clus e s and amily i ms), he
co esponding empi ical analysis is only based on pooled c oss-sec ional da a, which
aises po en ial conce ns o endogenei y (e.g. Block and Spiegel 2013; G asho e al.
2019). Fu u e esea ch may he e o e use panel da a in o de o de e mine dynamic
e ec s. In his con ex , i is also use ul o u u e s udies o addi ionally examine
o he ypes o egions (e.g. u al s. u ban egions) in o de o u he disen angle
he clus e e ec . Fu he mo e, simila o p e ious s udies (e.g. Block and Spiegel
2013), due o da a a ailabili y only in o ma ion on amily owne ship is used o
de e mine amily i ms. Fu u e s udies may conside he di ec in luence o amily
membe s in e ms o i m managemen o supe iso y boa d. In his con ex , i also
seems in e es ing o addi ionally conside he in luence o he indi idual cha ac e is-
K
34 N. G asho
ics o he owne s, especially be ween di e en gene a ions in leade ship (De Massis
e al. 2012). Las ly, he analysis is limi ed o he high- ech and polycen ic coun y
Ge many. The conside a ion o u he coun ies wi h di e en le els o economic
de elopmen and egional s uc u e could be aken up by u u e s udies o con ol
o po en ial coun y e ec s.
Ne e heless, despi e hese limi a ions all in all i can be esumed ha he esul s
abou he ela ionship be ween amily i ms and adical inno a ions con ibu e o
he amily i m, egional and inno a ion economics li e a u e. The a icle p o ides
insigh s abou he eme gence o adical inno a ion on he i m-le el by examining
he co esponding ole o amily i ms in Ge many and he mode a ing ole o
egional clus e s and i m size. In doing so, i ex ends ecen s udies in inno a ion
economics ha ha e examined adical inno a ion ac oss di e en ypes o i ms
(e.g. G asho and Kopka 2023). Fu he mo e, i con ibu es o he egional s udies
li e a u e by showing ha he clus e -speci ic bene i s do no acc ue o all ypes o
i ms, bu a he o speci ic ones (such as amily i ms), which adds o p e ious
s udies ha a emp o explain he he e ogenous i m-speci ic pe o mance e ec s o
being loca ed in egional clus e s (e.g. He as-Oli e e al. 2018). A he same ime, i
also con ibu es o he amily business s udies li e a u e by s essing he ele ance o
he egional con ex when examining he (inno a i e) pe o mance o amily i ms,
which suppo s ecen e o s o conside con ex ual he e ogenei y in amily business
esea ch (e.g. Basco e al. 2021a). In addi ion o he scien i ic con ibu ion, he esul s
also o e ele an policy implica ions. The esul s on he lowe a e age adical
pa en ac i i ies in amily businesses equi e special a en ion in he o m o policy
measu es ha should p ima ily add ess he isk a oidance o amily businesses, such
as echnical se ices and ad ice ha can p o ide an ex e nal pe spec i e and educe
he isk o po en ially misleading in es men s (Jones and G imshaw 2016; Shapi a
and You ie 2016). Fu he mo e, by inding e idence o a mode a ing in luence o
he loca ion o i ms in a egional clus e , his s udy addi ionally shows ha he
ela ionship be ween amily i ms and adical inno a ion is a mo e complex han
o en assumed. Hence, he e is a need o a mo e di e en ia ed iew on amily i ms
ha also e lec s his he e ogenei y (De Massis e al. 2012). By ollowing such an
app oach, amily i ms can also success ully become mo e adical in hei inno a ion
p ocess.
K
Familia bu also adical? The mode a ing ole o egional clus e s o amily i ms in he... 35
6 Appendix
Table 3 Indus ial dis ibu ion o sample and sha e o dependen a iable
NACE Ac i i y No.
o
i ms
Sha e o adi-
cal pa en s
(in %)
1 C op and animal p oduc ion, hun ing and ela ed se ice ac i i ies 10 2.27
3 Fishing and aquacul u e 1 0.00
5 Mining o coal and ligni e 1 0.00
6 Ex ac ion o c ude pe oleum and na u al gas 1 0.00
7 Mining o me al o es 1 0.00
8 O he mining and qua ying 16 4.61
10 Manu ac u e o ood p oduc s 92 3.48
11 Manu ac u e o be e ages 6 1.49
12 Manu ac u e o obacco p oduc s 1 12.50
13 Manu ac u e o ex iles 102 2.58
14 Manu ac u e o wea ing appa el 17 11.08
15 Manu ac u e o lea he and ela ed p oduc s 20 3.64
16 Manu ac u e o wood and o p oduc s o wood and co k. excep
u ni u e; manu ac u e o a icles o s aw and plai ing ma e ials
77 2.86
17 Manu ac u e o pape and pape p oduc s 76 2.47
18 P in ing and ep oduc ion o eco ded media 34 0.90
19 Manu ac u e o coke and e ined pe oleum p oduc s 8 8.62
20 Manu ac u e o chemicals and chemical p oduc s 248 2.01
21 Manu ac u e o basic pha maceu ical p oduc s and pha maceu ical
p epa a ions
107 1.35
22 Manu ac u e o ubbe and plas ic p oduc s 391 1.05
23 Manu ac u e o o he non-me allic mine al p oduc s 157 3.66
24 Manu ac u e o basic me als 113 2.90
25 Manu ac u e o ab ica ed me al p oduc s. excep machine y and
equipmen
741 2.61
26 Manu ac u e o compu e . elec onic and op ical p oduc s 707 2.23
27 Manu ac u e o elec ical equipmen 389 1.66
28 Manu ac u e o machine y and equipmen n.e.c. 1339 2.34
29 Manu ac u e o mo o ehicles. aile s and semi- aile s 152 1.64
30 Manu ac u e o o he anspo equipmen 79 4.60
31 Manu ac u e o u ni u e 83 1.01
32 O he manu ac u ing 385 2.10
33 Repai and ins alla ion o machine y and equipmen 58 2.26
35 Elec ici y. gas. s eam and ai condi ioning supply 59 5.43
K
36 N. G asho
Table 3 (Con inued)
NACE Ac i i y No.
o
i ms
Sha e o adi-
cal pa en s
(in %)
36 Wa e collec ion. ea men and supply 4 0.00
37 Sewe age 7 8.33
38 Was e collec ion. ea men and disposal ac i i ies; ma e ials eco -
e y
35 2.08
39 Remedia ion ac i i ies and o he was e managemen se ices 3 0.00
41 Cons uc ion o buildings 44 1.52
42 Ci il enginee ing 37 1.02
43 Specialised cons uc ion ac i i ies 278 2.22
45 Wholesale and e ail ade and epai o mo o ehicles and mo o -
cycles
82 1.99
46 Wholesale ade. excep o mo o ehicles and mo o cycles 1269 1.57
47 Re ail ade. excep o mo o ehicles and mo o cycles 275 2.40
49 Land anspo and anspo ia pipelines 29 3.92
50 Wa e anspo 3 16.67
51 Ai anspo 1 0.00
52 Wa ehousing and suppo ac i i ies o anspo a ion 37 4.08
53 Pos al and cou ie ac i i ies 3 3.29
55 Accommoda ion 1 0.00
56 Food and be e age se ice ac i i ies 11 0.00
58 Publishing ac i i ies 7 0.00
59 Mo ion pic u e. ideo and ele ision p og amme p oduc ion. sound
eco ding and music publishing ac i i ies
67.69
60 P og amming and b oadcas ing ac i i ies 1 0.00
61 Telecommunica ions 14 0.27
62 Compu e p og amming. consul ancy and ela ed ac i i ies 408 1.78
63 In o ma ion se ice ac i i ies 23 0.00
64 Financial se ice ac i i ies. excep insu ance and pension unding 232 2.05
66 Ac i i ies auxilia y o inancial se ices and insu ance ac i i ies 39 1.91
68 Real es a e ac i i ies 157 2.47
69 Legal and accoun ing ac i i ies 9 1.64
70 Ac i i ies o head o ices; managemen consul ancy ac i i ies 481 2.16
71 A chi ec u al and enginee ing ac i i ies; echnical es ing and anal-
ysis
583 3.03
72 Scien i ic esea ch and de elopmen 420 2.08
73 Ad e ising and ma ke esea ch 26 0.00
74 O he p o essional. scien i ic and echnical ac i i ies 94 3.13
75 Ve e ina y ac i i ies 2 0.00
77 Ren al and leasing ac i i ies 48 3.73
78 Employmen ac i i ies 8 0.00
79 T a el agency. ou ope a o and o he ese a ion se ice and
ela ed ac i i ies
70.00
K
Familia bu also adical? The mode a ing ole o egional clus e s o amily i ms in he... 37
Table 3 (Con inued)
NACE Ac i i y No.
o
i ms
Sha e o adi-
cal pa en s
(in %)
80 Secu i y and in es iga ion ac i i ies 1 0.00
81 Se ices o buildings and landscape ac i i ies 34 3.33
82 O ice adminis a i e. o ice suppo and o he business suppo
ac i i ies
185 4.04
84 Public adminis a ion and de ence; compulso y social secu i y 4 0.00
85 Educa ion 24 1.30
86 Human heal h ac i i ies 42 0.64
87 Residen ial ca e ac i i ies 3 0.00
88 Social wo k ac i i ies wi hou accommoda ion 7 18.18
90 C ea i e. a s and en e ainmen ac i i ies 3 0.00
92 Gambling and be ing ac i i ies 4 0.00
93 Spo s ac i i ies and amusemen and ec ea ion ac i i ies 9 0.00
94 Ac i i ies o membe ship o ganisa ions 13 1.79
95 Repai o compu e s and pe sonal and household goods 1 0.00
96 O he pe sonal se ice ac i i ies 111 1.99
K
38 N. G asho
Table 4 Pai wise co ela ion ma ix
Va iables (1) (2) (3) (4) (5) (6) (7) (8) (9)
(1) ad_coun 1.000 ––––––––
(2) amily_dummy –0.046*** 1.000 – – – – – – –
(3) clus e _dummy 0.050*** –0.035*** 1.000 – – – – – –
(4) age_mean 0.090*** –0.098*** 0.153*** 1.000 – – – – –
(5) independence_dummy 0.099*** –0.171*** 0.022** 0.150*** 1.000 – – – –
(6) esea chin ensi eindus y 0.045*** –0.096*** 0.061*** 0.071*** 0.031*** 1.000 – – –
(7) popdens_mean 0.018* –0.095*** –0.098*** –0.045*** 0.035*** –0.032*** 1.000 – –
(8) sha e_academics_mean 0.030*** –0.046*** –0.058*** –0.107*** 0.020** 0.010 0.307*** 1.000 –
(9) log_size_mean 0.169*** –0.381*** 0.164*** 0.383*** 0.148*** 0.181*** –0.019* –0.068*** 1.000
*** p < 0.01, ** p< 0.05, * p< 0.1
K
Familia bu also adical? The mode a ing ole o egional clus e s o amily i ms in he... 39
Table 5 Reg ession esul s (wi h mean numbe o adical and non- adical pa en s)
Dependen a iable
ad_mean ad_mean ad_mean ad_mean
(1.) (2.) (3.) (4.)
Family_dummy – –0.988*** –1.278*** –0.088
– (0.205) (0.147) (0.383)
Clus e _dummy – – 0.858*** –
– – (0.285) –
Family_dummy*
clus e _dummy
– – 0.894** –
– – (0.447) –
Log_size_mean –––0.685***
–––(0.042)
Family_dummy*
log_size_mean
– – – –0.006
–––(0.068)
Age_mean 0.015*** 0.014*** 0.012*** 0.001
(0.003) (0.003) (0.002) (0.001)
Independence_dummy 1.511*** 1.165*** 1.244*** 0.363
(0.336) (0.339) (0.002) (0.225)
Resea chin ensi eindus y 0.886*** 0.820*** 0.743*** 0.435***
(0.182) (0.179) (0.166) (0.141)
Popdens_mean 0.0002* 0.0001 0.0001* 0.00003
(0.0001) (0.0001) (0.0001) (0.0001)
Sha e_academics_mean 0.075*** 0.074*** 0.081*** 0.049***
(0.026) (0.026) (0.027) (0.018)
Cons an –7.444*** –6.963*** –7.277*** –9.336***
(0.966) (0.977) (1.005) (0.773)
In la e: pa _mean 0.0001 0.0002 0.0007 0.001**
(0.001) (0.001) (0.001) (0.001)
Cons an –24.907*** –23.219*** –23.475*** –26.534***
(0.274) (0.153) (0.077) (0.058)
N 10,431 10,431 10,431 10,293
McFadden’s Adj R2 0.157 0.174 0.193 0.378
Log-likelihood –1367.406 –1338.237 –1300.830 –992.307
Akaike In . C i .*N2753 2697 2636 2013
Robus s anda d e o s in pa en heses
*p< 0.10, ** p< 0.05, *** p<0.01
K
40 N. G asho
Table 6 Reg ession esul s ( o 1: 2012–2016; 2: 2016–2020)
(a) Time pe iod 1: 2012–2016
Dependen a iable
ad_coun ad_coun ad_coun ad_coun
(1.) (2.) (3.) (4.)
Family_dummy – –0.486*** –0.719*** 0.097
– (0.167) (0.138) (0.513)
Clus e _dummy – – 0.520* –
– – (0.270) –
Family_dummy*
clus e _dummy
– – 0.739* –
– – (0.442) –
Log_size_mean – – – 0.375***
– – – (0.063)
Family_dummy*
log_size_mean
– – – –0.035
– – – (0.074)
Age_mean 0.011*** 0.012*** 0.009*** –0.0002
(0.002) (0.002) (0.002) (0.001)
Independence_dummy 0.100*** 0.859*** 0.937*** 0.408
(0.300) (0.304) (0.320) (0.262)
Resea chin ensi eindus y 0.245 0.216 0.148 –0.054
(0.156) (0.156) (0.132) (0.116)
Popdens_mean 0.0002** 0.0002* 0.0001 0.00004
(0.0001) (0.0001) (0.0001) (0.0001)
Sha e_academics_mean 0.062*** 0.056*** 0.064*** 0.035*
(0.022) (0.021) (0.020) (0.020)
Cons an –2.978*** –2.609*** –2.796*** –3.489***
(0.836) (0.783) (0.767) (1.147)
In la e: pa _coun –0.384*** –0.375*** –0.367*** –0.253***
(0.059) (0.060) (0.061) (0.069)
Cons an 3.287*** 3.204*** 3.167*** 2.809***
(0.119) (0.124) (0.122) (0.151)
N 10,431 10,431 10,431 6,812
McFadden’s Adj R2 0.212 0.214 0.218 0.242
Log-likelihood –2926.274 –2917.165 –2896.779 –2234.260
Akaike In . C i .*N5871 5854 5828 4497
K
Familia bu also adical? The mode a ing ole o egional clus e s o amily i ms in he... 47
Ebe T, Gebaue JE, B enne T, Bleido n W, Gosling SD, Po e J, Ren ow PJ (2019) A e egional
di e ences in pe sonali y and hei co ela es obus ? Applying spa ial analysis echniques o examine
egional a ia ion in pe sonali y ac oss he US and Ge many. Wo king Pape s on Inno a ion and
Space, No. 05.19. Philipps Uni e si y, Ma bu g
Eu opean Clus e Obse a o y (2018) Me hodology—indica o s. h p://www.clus e obse a o y.eu/index.
h ml#! iew=abou obse a o y;u l=/abou obse a o y/me hodology/indica o s/. Accessed 16 July
2023
Eu opean Commission (2008) The concep o clus e s and clus e policies and hei ole o compe i-
i eness and inno a ion: Main s a is ical esul s and lessons lea ned. Eu ope INNOVA / PROINNO
Eu ope pape N° 9.
Eu opean Family Businesses (2021) Abou Eu opean amily businesses. h ps://eu opean amilybusinesses
.eu/abou -eu opean- amily-businesses/. Accessed 15 No 2021
Faems D, Van Looy B, Debacke e K (2005) In e o ganiza ional collabo a ion and inno a ion: owa d
a po olio app oach. J P od Inno Manag 22(3):238–250
Filse M, De Massis A, Gas J, K aus S, Niemand T (2018) T acing he oo s o inno a i eness in
amily SMEs: he e ec o amily unc ionali y and socioemo ional weal h. J P od Inno Manag
35(4):609–628
Fleming L (2001) Recombinan unce ain y in echnological sea ch. Manage Sci 47(1):117–132
Fleming L (2007) B eak h oughs and he long ail o inno a ions. MIT Sloan Manage Re 49(1):1–11
Geh ke B, F ie sch R, Neuhäusle P, Ramme C, Leidmann M (2013) Neuabg enzung o schungsin ensi e
Indus ien und Gü e : NIW/ISI/ZEW-Lis en 2012 (No. 8-2013) (S udien zum deu schen Inno a ion-
ssys em)
Gómez-Mejía LR, Haynes KT, Núñez-Nickel M, Jacobson KJL, Moyano-Fuen es J (2007) Socioemo ional
weal h and business isks in amily-con olled i ms: e idence om Spanish oli e oil mills. Adm
Sci Q 52(1):106–137
G asho N (2020) Fi m-speci ic clus e e ec s: a me a-analysis. Pap Reg Sci 99(5):1237–1260
G asho N (2021) Spill o e o Spill ou ?—A mul ile el analysis o he clus e and i m pe o mance
ela ionship. Ind Inno 28(10):1298–1331
G asho N, B enne T (2021) Radical inno a ions in clus e s: he ole o clus e in e nal and ex e nal ela-
ionships. In: Fo nahl D, G asho N (eds) The globaliza ion o egional clus e s: be ween localiza ion
and in e na ionaliza ion. Edwa d Elga Publishing, Chel enham, pp 189–209
G asho N, Fo nahl D (2021) “To be o no o be” loca ed in a clus e ?—A desc ip i e me a-analysis o he
i m-speci ic clus e e ec . Ann Reg Sci 67(3):541–591
G asho N, Kopka A (2023) A i icial in elligence and adical inno a ion: an oppo uni y o all compa-
nies? Small Bus Econ 61(2):771–797
G asho N, Hesse K, Fo nahl D (2019) Radical o no ? The ole o clus e s in he eme gence o adical
inno a ions. Eu Plan S ud 27(10):1904–1923
G asho N, Kopka A, Wessendo C, Fo nahl D (2020) Indus y 4.0 and clus e s: complemen a ies o
subs i u es in i m’s knowledge c ea ion? Compe Re 31(1):83–105
G eene WH (1994) Accoun ing o excess ze os and sample selec ion in poisson and nega i e binomial
eg ession models. S e n School o Business wo king pape no. 94–10
G iliches Z (1990) Pa en s a is ics as economic indica o s: a su ey. J Econ Li 28(4):1661–1707
G illi sch M, Asheim B, T ippl M (2018) Un ela ed knowledge combina ions: he unexplo ed po en ial o
egional indus ial pa h de elopmen . Camb J Reg Econ Soc 11(2):257–274
Hassink R (2007) The s eng h o weak lock-ins: he enewal o he Wes müns e land ex ile indus y.
En i on Plan A 39(5):1147–1165
Hauswald H, Hack A, Kelle manns FW, Pa zel H (2016) A ac ing new alen o amily i ms: who is
a ac ed and unde wha condi ions? En ep eneu sh Theo y P ac 40(5):963–989
Hende son RM, Cla k KB (1990) A chi ec u al inno a ion: he econ igu a ion o exis ing p oduc ech-
nologies and he ailu e o es ablished i ms. Adm Sci Q 35(1):9–30
He io SR, Le in hal D, Ma ch JG (1985) Lea ning om expe ience in o ganiza ions. Am Econ Re
75(2):298–302
He as-Oli e JL, Sempe e-Ripoll F, Rojas Al a ado R, Es elles-Miguel S (2018) Agglome a ions and
i m pe o mance: who bene i s and how much? Reg S ud 52(3):338–349
He as-Oli e JL, Sempe e-Ripoll F, Es elles-Miguel S, Rojas-Al a ado R (2019) Radical s inc emen al
inno a ion in Ma shallian Indus ial Dis ic s in he Valencian Region: wha p e ails? Eu Plan S ud
27(10):1924–1939
Hesse K (2020) Rela ed o whom? The impac o ela edness o egional on ie i ms on adical inno a-
ions. B emen Pape s on Economics & Inno a ion. Uni e si ä B emen
K
48 N. G asho
Hesse K, Fo nahl D (2020) Essen ial ing edien s o adical inno a ions? The ole o (un-) ela ed a ie y
and ex e nal linkages in Ge many. Pap Reg Sci 99(5):1165–1183
Hu Q, Hughes M (2020) Radical inno a ion in amily i ms: a sys ema ic analysis and esea ch agenda. In
J En ep eneu ial Beha Res 26(6):1199–1234
Jones BF (2009) The bu den o knowledge and he “dea h o he enaissance man”: Is inno a ion ge ing
ha de ? Re iew o Economic S udies 76(1):283–317
Jones B, G imshaw D (2016) The impac o skill o ma ion policies on inno a ion. In: Edle J, Cunning-
ham P, Gök A, Shapi a P (eds) Handbook o inno a ion policy impac . Edwa d Elga Publishing,
pp 108–128
Ke s de V ies MF (1993) The dynamics o amily con olled i ms: The good and he bad news. O gan Dyn
21(3):59–71
Lee C, Lee K, Pennings JM (2001) In e nal capabili ies, ex e nal ne wo ks, and pe o mance: a s udy on
echnology-based en u es. S a eg Manag J 22(6-7):615–640
Le i B, Ma ch JG (1988) O ganiza ional lea ning. Annu Re Sociol 14(1):319–338
Majumda SK (1997) The impac o size and age on i m-le el pe o mance: some e idence om India.
Re Ind O gan 12:231–241
Ma ch JG (1991) Explo a ion and exploi a ion in o ganiza ional lea ning. O gan Sci 2(1):71–87
Ma shall A (1920) P inciples o economics, 8 h edn. Macmillan, London
Ma in R, Sunley P (2003) Decons uc ing clus e s: chao ic concep o policy panacea? J Econ Geog
3(1):5–35
Masci elli R (2000) F om expe ience: Ha nessing aci knowledge o achie e b eak h ough inno a ion.
J P od Inno Manag 17(3):179–193
McCann BT, Fol a TB (2008) Loca ion ma e s: whe e we ha e been and whe e we migh go in agglome -
a ion esea ch. J Manage 34(3):532–565
Mewes L (2019) Scaling o a ypical knowledge combina ions in Ame ican me opoli an a eas om 1836
o 2010. Econ Geog 95(4):341–361
Mewes L, B oekel T (2020) Subsidized o change? The impac o R&D policy on egional echnological
di e si ica ion. Ann Reg Sci 65:221–252
Mille D, Le B e on-Mille I, Scholnick B (2008) S ewa dship s. s agna ion: an empi ical compa ison o
small amily and non- amily businesses. J Managemen S udies 45(1):50–78
Ne ka A (2003) Old is gold? The alue o empo al explo a ion in he c ea ion o new knowledge. Manage
Sci 49(2):211–229
Newbe SL (2007) Empi ical esea ch on he esou ce-based iew o he i m: an assessmen and sugges-
ions o u u e esea ch. S a eg Manag J 28(2):121–146
Nie o MJ, San ama ia L, Fe nandez Z (2015) Unde s anding he inno a ion beha io o amily i ms. Small
Bus Econ 53(2):382–399
O ega-A gilés R, Vi a elli M, Voig P (2009) R&D in SMEs: a pa adox? Small Bus Econ 33(1):3–11.
h ps://doi.o g/10.1007/s11187-009-9187-5
Pa k M, Leahey E, Funk RJ (2023) Pape s and pa en s a e becoming less dis up i e o e ime. Na u e
613(7942):138–144
Pi ino D, Visin in F, Minichilli A, Compagno C (2021) Family in ol emen in go e nance and i m pe -
o mance in indus ial dis ic s. The mode a ing ole o he indus y’s echnological pa adigm. En-
ep eneu sh Reg De 33(7-8):514–531
Poude R, S John CH (1996) Ho spo s and blind spo s: Geog aphical clus e s o i ms and inno a ion.
Acad Manage Re 21(4):1192–1225
Ramme C, Schube T (2016) Concen a ion on he ew? R&D and inno a ion in Ge man i ms be ween
2001 and 2013. F aunho e ISI Discussion Pape s-Inno a ion Sys ems and Policy Analysis (No. 54)
Rigby DL, B own WM (2015) Who bene i s om agglome a ion? Reg S ud 49(1):28–43
Schä e D, S ephan A, Mosque a JS (2017) Family owne ship: does i ma e o unding and success o
co po a e inno a ions? Small Bus Econ 48:931–951
Scha e ME, S illman S (2010) x o e id: s a a module o calcula e es s o o e iden i ying es ic ions a e
x eg, x i eg, x i eg2 and x h aylo . h p://ideas. epec.o g/c/boc/bocode/s456779.h ml. Accessed 11
Jan 2021
Schamp EW (2005) Decline o he dis ic , enewal o i ms: an e olu iona y app oach o oo wea p o-
duc ion in he Pi masens a ea, Ge many. En i on Plan A 37(4):617–634
Schoenmake s W, Duys e s G (2010) The echnological o igins o adical in en ions. Res Policy
39(8):1051–1059
Scholl T, B enne T (2016) De ec ing spa ial clus e ing using a i m-le el clus e index. Reg S ud
50(6):1054–1068
K
Familia bu also adical? The mode a ing ole o egional clus e s o amily i ms in he... 49
Schumpe e JA (1934) The heo y o economic de elopmen . Ha a d Uni e si y P ess, Camb idge
Sciascia S, No dq is M, Mazzola P, De Massis A (2015) Family owne ship and R&D in ensi y in small-
and medium-sized i ms. J P od Inno Manag 32(3):349–360
Shapi a P, You ie J (2016) The impac o echnology and inno a ion ad iso y se ices. In: Edle J, Cun-
ningham P, Gök A, Shapi a P (eds) Handbook o Inno a ion Policy Impac . Edwa d Elga Publishing,
pp 161–195
Shkolnyko a M, Kudic M (2021) Who bene i s om SMEs’ adical inno a ions?—empi ical e idence
om Ge man bio echnology. Small Bus Econ. h ps://doi.o g/10.1007/s11187-021-00464-x
Si mon DG, Hi MA (2003) Managing esou ces: linking unique esou ces, managemen , and weal h
c ea ion in amily i ms. En ep eneu sh Theo y P ac 27(4):339–358
S ae D, Thesma D (2007) Pe o mance and beha io o amily i ms: e idence om he F ench s ock
ma ke . J Eu Econ Assoc 5(4):709–751
Suwala L (2017) On c ea i i y: om concep ual ideas owa ds a sys emic unde s anding. In: B ydges T,
H acs BJ, Jansson J, Pasquinelli C, Powe D, Sjöholm J (eds) Eu opean Colloquium on Cul u e,
C ea i i y and Economy (CCE). Wo king Pape Compendium. Kunnskaps e ke , Lillehamme ,
pp 82–111
T aj enbe g M, Hende son R, Ja e A (1997) Uni e si y e sus co po a e pa en s: A window on he basic-
ness o in en ion. Econ Inno New Technol 5(1):19–50
Tushman ML, Ande son P (1986) Technological discon inui ies and o ganiza ional en i onmen s. Adm
Sci Q 31(3):439–465
U bina i A, F anzò S, De Massis A, F a ini F (2017) Inno a ion in amily i ms: a e iew o p io s udies
and a amewo k o u u e esea ch. In: B em A, Via do E (eds) Re olu ion o inno a ion manage-
men . Palg a e Macmillan, London, pp 213–246
Uzzi B, Mukhe jee S, S inge M, Jones B (2013) A ypical combina ions and scien i ic impac . Science
342(6157):468–472
Ve hoe en D, Bakke J, Veugele s R (2016) Measu ing echnological no el y wi h pa en -based indica o s.
Res Policy 45(3):707–723
Ve spagen B (2005) Inno a ion and economic g ow h. In: Fage be g J, Mowe y D, Nelson R (eds) The
Ox o d handbook o inno a ion. Ox o d Uni e si y P ess, Ox o d
Wei zman ML (1998) Recombinan g ow h. Q J Econ 113(2):331–360
We ne A, Sch öde C, Chlos a S (2018) D i ing ac o s o inno a ion in amily and non- amily SMEs.
Small Bus Econ 50(1):201–218
Woold idge JM (2002) Econome ic analysis o c oss sec ion and panel da a. MIT P ess, Camb idge
Zah a SA (2012) O ganiza ional lea ning and en ep eneu ship in amily i ms: explo ing he mode a ing
e ec o owne ship and cohesion. Small Bus Econ 38(1):51–65
Zellwege TM (2007) Time ho izon, cos s o equi y capi al, and gene ic in es men s a egies o i ms.
Fam Bus Re 20(1):1–15
Zucke LG, Da by MR, To e o M (2002) Labo mobili y om academe o comme ce. J Labo Econ
20(3):629–660
Zybu a J, Zybu a N, Ah ens JP, Woywode M (2021) Inno a ion in he pos -succession phase o am-
ily i ms: Family CEO successo s and leade ship cons ella ions as esou ces. J Fam Bus S a eg
12(2):100336
Publishe ’s No e Sp inge Na u e emains neu al wi h ega d o ju isdic ional claims in published maps
and ins i u ional a ilia ions.
K
