Top management team composition and financial performance: examining the role of gender diversity

129 2, XXII, 2019 Business Adminis a ion and Managemen DOI: 10.15240/ ul/001/2019-2-009 In oduc ion The issue o emale ep esen a ion in op managemen eams is a g owing a ea o in e es among p ac i ione s, policymake s, and esea che s a ound he wo ld (Ahmadi, Nakaa, & Bou i, 2018). In pa icula , he e has been an ongoing deba e abou whe he gende di e si y in op managemen eams ela es o company pe o mance, mo e speci i cally i nancial pe o mance (Fa ag & Mallin, 2017; He nandez-Nicolás e al., 2015). The exis ing body o li e a u e sugges s ha op managemen eam gende di e si y has bo h posi i e and nega i e e ec s on company pe o mance. Addi ionally, he esul s o empi ical s udies on his issue emain mixed (e.g., Ali & Kon ad, 2017; Ege & Ind ucho á, 2014; Julizae ma & So i, 2012). On he one hand, he e is e idence ha gende di e si y in op managemen is associa ed wi h highe i nancial pe o mance (e.g., Campbell & Mínguez-Ve a, 2008; Pe yman, Fe nando, & T ipa hy, 2016). On he o he hand, he esul s gained by ano he su ey epo nega i e o no ela ionship be ween gende di e si y in op managemen eams and company i nancial pe o mance (e.g., Adams & Fe ei a, 2009; Chapple & Humph ey, 2014; Wang & Cli , 2009). Mos esea ch su eys in es iga ing he e ec o op managemen eam cha ac e is ics on company pe o mance implici ly o explici ly d aw on he uppe echelon heo y (Plöckinge e al., 2016). The uppe echelon heo y emphasizes he c i ical ole o op managemen eams on o ganiza ional s a egic decisions and pe o mance (Hamb ick & Mason, 1984). The heo y u he s a es ha op manage s’ obse able and unobse able backg ound cha ac e is ics play a signi i can ole in o ganiza ional decision-making and ul ima ely in o ganiza ional ou comes (Hamb ick, 2007). Thus, o explain o ganiza ional ou comes and e en ually i nancial ou comes, i is necessa y o conside he cha ac e is ics o he op managemen eam. This su ey is he e o e aimed o iden i y he possible link be ween company i nancial pe o mance and op manage ial eam composi ion based on gende ep esen a ion. Fu he mo e, p e ious empi ical su eys and exis ing li e a u e explo e he link be ween gende di e si y in op managemen eams and i nancial ou comes p edominan ly in la ge o ganiza ions, while only a ew s udies a e conduc ed in small- and medium-sized en e p ises (Dale-Olsen, Schone, & Ve ne , 2013; Mínguez-Ve a & Ma in, 2011; Wa d & Fo ke , 2015). Howe e , small- and medium- sized en e p ises a e conside ed o be c ucial engines o economic g ow h h oughou he wo ld. Small- and medium-sized en e p ises ep esen o e 95% o en e p ises in mos OECD coun ies, 99% in he EU, and 99.8% in he Czech Republic (Eu os a , 2017). Hence, some au ho s (e.g., F ancoeu , Labelle, & Sinclai - Desgagné, 2007) call o an examina ion o his issue in small- and medium-sized o ganiza ions ins ead o only in la ge ones. We espond o his call by ocusing ou s udy on small- and medium-sized en e p ises. Wi h ega d o deba es abou women’s ep esen a ion in op managemen eams, he impo ance o a ce ain numbe o pe cen age o women in he op managemen eam (TMT) is o en s essed. This aises ques ions such as, wha is he numbe o pe cen age o women in op managemen eams in l uencing i nancial pe o mance? Mo e speci i cally, wha p opo ion o women in op managemen is likely o cons i u e a c i ical mass ha con ibu es o be e i nancial pe o mance? As s a ed by Cha les, Redo and Zopounidis (2015), mos s udies ha e ypically ocused on he impac o he p esence o women on TOP MANAGEMENT TEAM COMPOSITION AND FINANCIAL PERFORMANCE: EXAMINING THE ROLE OF GENDER DIVERSITY Dana Ege o á, Ma a Nosko á EM_2_2019.indd 129EM_2_2019.indd 129 19.6.2019 15:11:1819.6.2019 15:11:18 130 2019, XXII, 2 Business Adminis a ion and Managemen i nancial pe o mance; howe e , based on c i ical mass heo y, he me e p esence o women on boa ds may no be su i cien o b ing a posi i e impac . Fu he mo e, shi ing he deba e om he p esence o women in op managemen eams o he size o he mino i y g oup ( he numbe o women in TMT) enabled an unde s anding o he impo ance o op managemen eam p ocesses (To chia, Calab o, & Huse, 2011). Fu he mo e, au ho s Kon ad, K ame , and E ku (2008) call o shi esea ch solely ocused on women’s p esence in op managemen eams o he p opo ion o emale op managemen membe ship exp essed by he numbe /pe cen age o women in op managemen eams. Simila ly, Ki sch (2017) sugges s paying close a en ion o h eshold e ec s, such as he shi om ze o o one-woman di ec o s o a c i ical mass o women di ec o s. In doing so, unlike p io su eys in e ms o measu emen s, his s udy uses he numbe /pe cen age o women in op managemen as he independen a iable and c i ical mass is in es iga ed. Thus, he p esen s udy con ibu es o he deba e on he ela ionship be ween op managemen gende di e si y and i nancial pe o mance by u he explo ing he c i ical mass heo y. Unde he abo e-men ioned amewo k, he pu pose o he p esen s udy is o explo e he ela ionship be ween gende di e si y in op managemen eams and he o ganiza ional i nancial pe o mance in small- and medium- sized en e p ises in he con ex o he Czech Republic. Speci i cally, we in es iga e wha p opo ion o women (wha cons i u es c i ical mass) in op managemen eams is likely o in l uence company i nancial pe o mance. The heo e ical amewo k guiding his s udy includes he uppe echelon heo y and he c i ical mass heo y. The pape is o ganized as ollows. The i s sec ion p esen s he heo e ical amewo k ega ding he uppe echelon heo y and he c i ical mass heo y. An o e iew o exis ing empi ical esea ch (o he p e ious i ndings) is also included. Nex , he esea ch design is desc ibed, which con ains a desc ip ion o he da a and indica o s measu emen . Nex , we p esen and discuss he esul s. Finally, he conclusions and limi a ions o he s udy a e p o ided. 1. Theo y Backg ound and Resea ch Ques ions 1.1 Uppe Echelon Theo y As p e iously men ioned, ou esea ch in es iga es he link be ween gende di e si y a op managemen and i nancial pe o mance based upon uppe echelon employed by Hamb ick and Mason (1984). The heo y inspi ed many s udies ocused on op managemen eams and helped o unde s and he causal link be ween op managemen cha ac e is ics and o ganiza ional pe o mance. Empi ical s udies ha ollowed om Hamb ick and Mason’s (1984) uppe echelon heo y sugges ha indeed, TMT ma e s o o ganiza ional ou comes (Oppong, 2014). Plöckinge e al. (2016) no ed ha exis ing empi ical s udies in i nancial esea ch ha e explici ly (e.g., Nielsen & Nielsen, 2013; Ops up & Villadsen, 2014) o implici ly (e.g., Campbell & Mínguez-Ve a, 2008; Conyon & He, 2017; Richa d, Ki by, & Chadwick, 2013) d awn on he main assump ion o he uppe echelon heo y. Simila ly, Nishii, Go e, and Ra e (2007) a gue ha se e al s udies analyse he e ec o uppe echelon cha ac e is ics and o ganiza ional pe o mance. The uppe echelons heo y emphasizes he dominan ole o op managemen eams and he c i ical impac hey ha e on o ganiza ional ou comes. Mo e speci i cally, he heo y s a es ha he o ganiza ion’s ou comes a e a e lec ion o i s op manage s (Cannella, 2001; Hamb ick & Mason, 1984). The co e o uppe echelons heo y is based on he ollowing wo in e connec ed pa s: (1) execu i es ac on he basis o hei pe sonalized in e p e a ions o si ua ions hey ace and (2) his pe sonalized cons ual is a unc ion o hei expe iences, alues, and pe sonali ies. The heo y also sugges s ha demog aphic cha ac e is ics o execu i es could be used as alid, al hough incomple e and imp ecise, indica o s o p oxies o mo e elusi e psychological p ope ies and execu i e disposi ions (Cannella, 2001; Hamb ick, 2007). Thus, i we wan o unde s and why o ganiza ions do he hings hey do and pe o m he way hey do, emphasis should be placed on he demog aphic cha ac e is ics o hei op execu i es (Ca pen e , Gele kanycz, & Sande s, 2004; Hamb ick, 2007). Nex , he impo an ea u e o uppe echelons heo y is a ocus on he op managemen eam a he han s ic ly on single EM_2_2019.indd 130EM_2_2019.indd 130 19.6.2019 15:11:1919.6.2019 15:11:19 131 2, XXII, 2019 Business Adminis a ion and Managemen chie execu i es (Finkels ein & Hamb ick, 1990). Hamb ick (2007) no ed ha ocus on execu i e g oups, a he han on indi iduals, o en b ings be e explana ions o o ganiza ional ou comes. Thus, as a gued by Finkels ein and Hamb ick (1990), i we wan o unde s and o ganiza ional ou comes, we need o look no jus a he CEO, bu a he a he en i e op managemen eam. An issue ha needs o be also conside ed is op managemen eam di e si y (TMT), which is conside ed as one o he undamen al uppe echelons cons uc s (Hamb ick, Cho, & Chen, 1996; Hamb ick, Humph ey, & Gup a, 2015). P e ious s udies ha e epo ed mixed and o en con adic o y esul s o TMT di e si y on pe o mance anging om a posi i e ela ionship be ween TMT di e si y and i nancial pe o mance h ough o an insigni i can o nega i e associa ion be ween TMT di e si y and i nancial pe o mance (Dwye , Richa d, & Chadwick, 2003; Pos & By on, 2015; Robe son & Hyeon, 2007). A possible explana ion o he con l ic ing esul s may be due o he ac ha , TMT di e si y is de i ned mos ly as a gene al cons uc (Nielsen, 2010). As no ed by Nielsen and Nielsen (2013), dealing wi h di e si y as a gene ic unidimensional concep does no e l ec i s complex na u e and impac . Fu he mo e, no all aspec s o di e si y in l uence eam decision-making and company pe o mance in he same way (Nielsen, 2010). The e o e, he au ho sugges s b eaking down he cons uc o di e si y in o single a ibu es and conside a ious di e si y a ibu es independen ly. Acco ding o K ishnan and Pa k (2005), gende is a mo e complex demog aphic a iable han o he a iables and an impo an measu e o di e si y. We he e o e ocus he esea ch on gende in op managemen eams and in es iga e whe he i enhances company i nancial pe o mance. Wi h ega d o he abo e, we a gue ha he uppe echelon heo y p o ides a sui able heo e ical amewo k om which we can examine how managemen execu i es and demog aphic di e si y, speci i cally gende di e si y, a e ela ed o i nancial pe o mance in small and medium-sized en e p ises. The e o e, unde he uppe echelon heo y amewo k, we posi he ollowing esea ch ques ion: Q1: Does gende di e si y in op managemen eams posi i ely a ec i nancial pe o mance in small- and medium-sized en e p ises? 1.2 C i ical Mass Theo y Mos p e ious s udies in es iga ing he ela ionship be ween gende di e si y in op managemen and i m pe o mance ha e ocused on he impac o he p esence o women on i nancial pe o mance; howe e , based on he c i ical mass heo y, he me e p esence o women on boa ds may no be su i cien enough o b ing signi i can changes o op managemen eams and imp o e co po a e go e nance (Cha les, Redo , & Zopounidis, 2015). Acco ding o he c i ical mass heo y, he p opo ion o socially and cul u ally di e se people in a g oup is conside ed as c i ical in shaping g oup dynamics and p ocesses (Kan e , 1977). On he basis o a ying p opo ional composi ions, Kan e (1977) iden i i ed he ollowing ou ypes o g oups: uni o m, skewed, i led, and balanced. Uni o m includes only one ype o people; all membe s sha e he same ( isible) cha ac e is ics. The ypological a io o his g oup is 100:1. Wi h espec o gende , all membe s o he g oup a e ei he male o emale. Skewed wi h la ge dominance o one ype o e ano he , he membe s o he mino i y (called okens) become symbolic ep esen a i es o hei g oup. This mino i y g oup is no able o o e come he p oblems o okenism; he a io o a dominan g oup o a mino i y g oup becomes 85:15. Wi h espec o gende , he male domina ed skewed g oup consis s o up o 20% women. Til ed, wi h less ex eme dis ibu ions, mino i y membe s can o m alliances; he a io o his g oup is 65:35 (up o 40%). Wi h espec o gende , a male domina ed il ed g oup consis s o 20 o 40% women. When balanced, he e a e oppo uni ies o elimina e he majo i y and gende -based di e ences in in e ac ion; he a io o his g oup is app oxima ely 60:40 down o 50:50 (mo e han 40%). Wi h espec o gende ep esen a ion, his g oup has 40 o 60% women (Kan e , 1977; Low, Robe s, & Whi ing, 2015; To chia, Calab o, & Huse, 2011). Thus, by shi ing he mino i y size g oup, i is unlikely ha he impac o his subg oup becomes p onounced on g oup in e ac ion and p ocesses (B oome, Conley, & Kimbe ly, 2011). A he same ime, he c i ical mass heo y p edic s ha when he size o he mino i y g oup eaches a ce ain h eshold o c i ical mass, he in l uence o he g oup g ows (Kon ad, K ame , & E ku , 2008). By analogy, i should be said ha when women each a ce ain h eshold, hei in l uence in op managemen eams EM_2_2019.indd 131EM_2_2019.indd 131 19.6.2019 15:11:1919.6.2019 15:11:19 132 2019, XXII, 2 Business Adminis a ion and Managemen g ows. On he o he hand, he heo y does no sugges which numbe is likely o cons i u e a c i ical mass (To chia, Calab o, & Huse, 2011). The e o e, s udies d awing on he c i ical mass heo y ha e been conduc ed o in es iga e which numbe is likely o cons i u e a c i ical mass o women in op managemen eams. Fo example, esea ch by Kon ad and K ame (2006) showed ha a c i ical mass is o med wi h h ee o mo e women. Simila ly, Liu, Wei, and Xie (2014) ound ha a c i ical mass o h ee o mo e women can cause a signi i can change in manage ial eams and imp o e co po a e go e nance. These au ho s also iden i i ed ha boa ds wi h h ee o mo e emale di ec o s ha e a much s onge impac on company pe o mance han boa ds wi h wo o ewe women. Fu he mo e, using a da a se o 151 lis ed Ge man i ms, Joecks, Pull, and Ve e (2013) p o ided e idence o he c i ical mass o emale ep esen a i es on he boa d o be eached a a sha e o app oxima ely 30%. On he o he hand, he esul s o he s udy pe o med by K ama ic and Mile ic (2017) show ha when a c i ical mass o 20-40% women on he managemen boa d has been eached, bank pe o mance imp o es. In con as , he s udy conduc ed by Lo ie-Lussie and Rin e (2002) did no suppo he c i ical mass heo y. Addi ionally, he esul s ob ained by Cha les, Redo and Zopounidis (2015) epo ed ha he e a e no signi i can in e ences in i m pe o mance be ween i ms wi h and wi hou a c i ical mass o women on hei boa ds. O e all, hese empi ical esul s p o ide oppo uni y o deepe in es iga ion o c i ical mass. The e o e, unde he c i ical mass heo y amewo k, we posi he ollowing esea ch ques ion: Q2: Wha pe cen age o women in op managemen eams cons i u es a c i ical mass ha posi i ely in l uences i nancial pe o mance in small- and medium-sized en e p ises? 2. Da a and Me hodology 2.1 Da a The da ase o his s udy is based on he 823 en e p ises o he manu ac u ing indus y in he Czech Republic. The manu ac u ing indus y was selec ed because i is he indus y wi h he la ges p oduc ion in he Czech Republic and hus is he mos impo an indus y o he Czech economy. Only en e p ises wi h 50 o 250 employees we e selec ed o ensu e ha only small and middle-sized en e p ises would be pa o he analysis. Addi ionally, o include only p o i -based en e p ises, hei legal o ms we e de e mined as ollows: indi idual en ep eneu s, pa ne ships, companies wi h limi ed liabili y, and public limi ed companies. F om a o al o 823 analysed companies, almos 64% a e companies wi h limi ed liabili y, and 34% a e public limi ed companies. The es consis o pa ne ships and indi idual en ep eneu s (see Tab. 1). In his s udy we do no conside he ime as a a iable, hus he da a included only one yea (2015). To analyse he ela ionship be ween gende di e si y o senio managemen and he i nancial pe o mance o he en e p ises, he ollowing da a we e ob ained o each analysed company: name, numbe o yea s in business, numbe o employees, egion, o al asse s, equi y, EBIT, EBT, numbe o senio manage s, and numbe o women in senio managemen . These da a we e mainly ob ained om he Bisnode Magnusweb da abase, which con ains comp ehensi e da a on all Czech economic en i ies, bo h co po a ions and indi iduals. A i s , 865 en e p ises we e ound by he da abase. Howe e , some we e excluded om he da a se due o missing da a o oo ex eme alues. Some missing da a we e collec ed om annual epo s o en e p ise websi es. Company wi h Limi ed Liabili y Public Limi ed Company Pa ne ship Indi idual En ep eneu s To al 528 285 5 5 823 64.16% 34.63% 0.61% 0.61% 100% Sou ce: own Tab. 1: Legal o ms o analysed en e p ises EM_2_2019.indd 132EM_2_2019.indd 132 19.6.2019 15:11:1919.6.2019 15:11:19 133 2, XXII, 2019 Business Adminis a ion and Managemen 2.2 Indica o Desc ip ions The en e p ise´s pe o mance is measu ed in se e al ways, mos ly by using i nancial indica o s. Resea ch examining gende di e si y and i s in l uence on he pe o mance o en e p ises mos ly used ma ke -based measu es whe e Tobin’s Q is among he mos common (Ma ino a, Plan enga, & Reme y, 2016; Regue a-Al a ado, de Fuen es, & La a ga, 2017; Rod íguez-Domínguez, Ga cía- Sánchez, & Gallego-Ál a ez, 2012; Ple ze e al., 2015). Accoun ing-based measu es, such as e u n on asse s (ROA), e u n on equi y (ROE) o e u n on sales (ROS) (Dale-Olsen, Schone, & Ve ne , 2013; Sumed ea, 2016) we e also used. In his s udy, we examine se e al ypes o en e p ises, mos o which a e no lis ed public limi ed companies. Thus, accoun ing- based measu es such as e u n on asse s (ROA) as he a io o he en e p ise’s p o i and i s o al asse s (EBIT/ o al asse ) as well as e u n on equi y (ROE) as he a io o he en e p ise’s ne p o i and capi al in es ed by he owne (EBT/Equi y) ha e been used as he mos app op ia e en e p ise pe o mance measu emen s. Ano he eason is he ac ha manage s a e able o in l uence he size o asse , i s use, and hus, he i nal alue o EBIT (ea nings be o e in e es and axes). Simila ly, o ROE, EBT (ea nings be o e axes) can also be in l uenced by managemen decisions, and he indica o shows how success ully he en e p ise is managed. To p o e he ela ionship be ween i nancial pe o mance and he numbe o women in senio managemen eams, he indica o PWOMAN was compiled ( he pe cen age o women in he op managemen eam o he en e p ise). In addi ion o his indica o , he ollowing indica o s we e also selec ed o examine o he ela ionships o PWOMAN: age o en e p ise (AGE), o al asse s o he en e p ise (SIZE), and numbe o employees (EMP). 2.3 Me hods S udies dealing wi h gende di e si y in TMT use se e al esea ch me hods. The mos commonly used a e he ollowing: panel da a analysis (e.g., Dale-Olsen, Schone, & Ve ne , 2013; Pe yman, Fe nando, & T ipa hy, 2016), me a-analysis (Ple ze e al., 2015), c ea ion o he eg ession model (Ahmadi, Nakaa, & Bou i, 2018; Regue a-Al a ado, de Fuen es, & La a ga, 2017) o calcula ion o basic co ela ion and desc ip i e s a is ics (Julizae ma & So i, 2012; Sumed ea, 2016). To desc ibe basic cha ac e is ics o he sample, he da a we e analysed by using simple s a is ical ools such as he mean, median and s anda d de ia ion. To e eal basic ela ionships be ween indica o s, he co ela ion ma ix based on he calcula ion o Spea man’s co ela ion coe i cien s was used. Namely, he k-means me hod by MacQueen (1967), Lloyd (1982) was used. This me hod di ides n obse a ions in o k clus e s, whe e each obse a ion belongs o he clus e wi h he nea es a e age. Thus, he calcula ion is pe o med by se e al epea ed i e a i e s eps. Fi s , i is necessa y o de e mine he numbe o clus e s. I he au ho does no ha e any ex e nal in o ma ion ha could help o de e mine he numbe in ad ance, i is ad isable o a emp he calcula ion se e al imes wi h di e en numbe s o clus e s and hen de e mine he op imal numbe . A e di iding da a in o clus e s, i is necessa y o calcula e he alue o he cen oid (cen e o g a i y), which ep esen s he a e age o he da a alues in a gi en clus e . Then, alues o indi idual da a a e compa ed o he alues o hei cen oids (using Euclidean dis ance), and i he e is a sho e dis ance o he cen oid o a di e en clus e , new e-alloca ion o he da a is made. A e ha , new calcula ions o cen oid alues mus be comple ed, and he whole p ocess is epea ed un il da a shi s s op. Dis ibu ion o da a in o he clus e s was pe o med using he p og am STATISTICA 12 S a So CR, s. . o. 3. Resul s and Discussion 3.1 Desc ip i e S a is ics Tab. 2 p esen s he basic s a is ics o ou sample. F om he desc ip i e s a is ics, we obse e ha he mean pe cen age o women in op managemen eams (PWOMAN) o he whole sample is app oxima ely 10.14%. This is close o he alues ob ained by Julizae ma and So i (2012), who epo ed a mean o 10.6% (women p opo ion in boa d) using a sample o 280 en e p ises in 2008 and 2009. Howe e , he p opo ion o women in he boa ds is usually lowe , o example Campbell and Mínguez- Ve a (2008) epo a alue o 3.2%, Ma ino a, Plan enga, and Reme y (2016) a alue o 5.4%, Pe yman, Fe nando and T ipa hy (2016) a alue o 4.8%, o Regue a-Al a ado, de EM_2_2019.indd 133EM_2_2019.indd 133 19.6.2019 15:11:1919.6.2019 15:11:19 134 2019, XXII, 2 Business Adminis a ion and Managemen Fuen es and La a ga (2017) a alue o 6.92%. In addi ion, in e ms o he pe cen age o emales in op managemen eams, he e we e 232 en e p ises wi h a leas one woman in he op managemen eam. In o al, he e we e 423 women in op managemen eams in hese 232 en e p ises, and he a e age pe cen age o emales in op managemen eams in his g oup o en e p ises was 34.5%. In e ms o age, he a e age alue was 20.34, wi h he maximum being 29, and he median being 22 yea s. This alue is in line wi h he mean alue o 22.58 yea s in business epo ed by Julizae ma and So i (2012) amongs Malaysian i ms. I may seem ha en e p ises in he Czech Republic a e ela i ely young o example, in compa ison o he mean alue o 56.77 yea s epo ed by Sumed ea (2016) amongs Romanian companies, his is caused by he ans o ma ion o Czech economy in ea ly 1990 (a e he collapse o communis egime) when many s a e en e p ises changed hei o m, and e en mo e en e p ises we e ounded. Thus, he oldes en e p ises included in he analysis we e ounded in 1988 and he younges in 2014. In e ms o size, he e a e la ge di e ences in he alues o asse s wi h a s anda d de ia ion o 321.520 million (Czech c owns). The numbe o employees was de e mined in ad ance wi h a minimum o 50 and maximum o 249 simila ly o Mínguez-Ve a and Ma in (2011). The Bisnode Magnusweb Da abase wo ks only wi h in e als and hei cen es; hus, in Tab. 2, he minimum is 75 employees, and he maximum is 225. The en e p ises wi h 100-199 employees a e dominan (388 en e p ises), he second a e hose wi h 50-99 employees (339 en e p ises), and he leas ep esen ed g oup a e en e p ises wi h 200-249 employees (96 en e p ises). As a as i nancial indica o s a e conce ned, he a e age ROA is 8.33% wi h a median o 6.92% and he maximum ROA is 79.47%, while he minimum is -44.26%. Fo compa ison, Ple ze e al. (2015), who analysed 10 s udies whe e ROA was he main indica o o i nancial pe o mance, epo ed ha i s alue is mos ly lowe han 8.24%. The a e age ROE is 14.94% wi h a median 12.14%; he maximum alue is 172.17%, while he minimum is -139.96%. Value epo ed in his s udy is sligh ly lowe han 19.81% epo ed by Ahmadi, Nakaa, and Bou i (2018), bu in line wi h al eady men ioned Ple ze e al. (2015), who ound ou s udies wi h he ollowing ROE alues: 15.27%, 14.91%, 9.92% and 6.73%. The p esen alues indica e ha he e a e bo h success ul and less success ul en e p ises in ou sample. 3.2 Co ela ion Analysis To in es iga e he possible link be ween en e p ises’ i nancial pe o mance and op managemen eam composi ion based on gende ep esen a ion, Spea man’s co ela ion coe i cien s o all indica o s we e i s calcula ed (see Tab. 3). Co ela ions ha a e signi i can a he le el o p < 0.5 a e ma ked wi h an as e isk (*). A s a is ically signi i can posi i e co ela ion possesses ROA and ROE, ROE and size, as well as size and numbe o employees. A s a is ically signi i can nega i e co ela ion can be ound be ween ROE and age, while o he co ela ions a e insigni i can . Due o he cons uc ion o ROA and ROE, i is no su p ise ha he e is a posi i e co ela ion obse ed ( he same i nding epo ed Gallego- Ál a ez, Ga cía-Sánchez, & Rod íguez- Domínguez, 2010). Addi ionally, he same signi i can posi i e co ela ion applies o ROA Mean Median S anda d de ia ion Min Max PWOMAN 10.14% 0 18.62% 0 100% AGE 20.34 22 5.38 3 29 SIZE 269,075,255 173,155,000 321,519,624 19,244,000 4,119,226,000 EMP 127.84 150 49.85 75 225 ROA 8.33% 6.92% 10.40% -44.26% 79.47% ROE 14.94% 12.14% 23.79% -139.96% 172.17% Sou ce: own Tab. 2: Desc ip i e s a is ics EM_2_2019.indd 134EM_2_2019.indd 134 19.6.2019 15:11:1919.6.2019 15:11:19 135 2, XXII, 2019 Business Adminis a ion and Managemen and size, which is consis en wi h he esul s o Sumed ea (2016). A posi i e signi i can co ela ion was also ound be ween he numbe o employees and size o he en e p ise, which means ha he highe he numbe o employees, he highe he asse s o he en e p ise. EMP indica o is no usually pa o hese analysis, howe e signi i can co ela ion be ween size (exp essed by o al asse s) and boa d size can be o en obse ed (e.g., Julizae ma & So i, 2012; Rod íguez-Domínguez, Ga cía- Sánchez, & Gallego-Ál a ez, 2012; Sumed ea, 2016). A signi i can nega i e co ela ion can be obse ed be ween age and ROE, which indica es a g ea e abili y o young en e p ises o app ecia e in es ed capi al. This esul is simila o he one ob ained by Low, Robe s, and Whi ing (2015). The indica o e l ec ing he p esence o women in op managemen eams shows a posi i e co ela ion wi h i nancial indica o s ROA and ROE as well as size and numbe o employees, bu only he co ela ion be ween he p esence o women in TMT and size appea s o be signi i can . In con as , he p esence o women in TMT is sligh ly nega i ely ela ed o age. A possible explana ion o his is he ac ha in younge i ms, women ha e an easie en y in o op managemen eams. The co ela ion be ween he p esence o women in TMT and ROA, is conside ed as he mos impo an co ela ion in his s udy, howe e i is insigni i can . The same esul was obse ed by e.g., Gallego-Ál a ez, Ga cía-Sánchez and Rod íguez-Domínguez (2010), Julizae ma and So i (2012) o Richa d, Ki byand, and Chadwick (2013). As Julizae ma and So i (2012) sugges , i can be caused by ew emale di ec o s in he sample. Thus, o he analysis has o be pe o med. To examine he link be ween en e p ises’ i nancial pe o mance and he p esence o women in op managemen eams mo e closely, i was necessa y o look in o he alues o indica o s o he g oups o en e p ises wi h one o mo e women in he op managemen eam and en e p ises wi h no women in he op managemen eam. Thus, we calcula ed he a e age alue o all analysed indica o s a e g ouping he en e p ises in o he ollowing 2 g oups: one wi h no women in he op managemen eam and he second wi h one o mo e women in he op managemen eam. AGE EMP SIZE PWOMAN ROE ROA AGE 1.000 0.064 0.026 -0.045 * -0.130 -0.055 EMP 0.064 1.000 * 0.440 0.064 -0.016 -0.012 SIZE 0.026 * 0.440 1.000 * 0.092 0.015 * 0.071 PWOMAN -0.045 0.064 * 0.092 1.000 0.031 0.004 ROE * -0.130 -0.016 0.015 0.031 1.000 * 0.860 ROA -0.055 -0.012 * 0.071 0.004 * 0.860 1.000 Sou ce: own No e: * signi i can a p < 0.5. Numbe o en e p ises ROA ROE SIZE EMP AGE All en e p ises 823 8.33% 14.94% 269,075,255 127.84 20.34 En e p ises wi h one o mo e women a TMT 232 8.74% 17.28% 335,515,789 134.75 19.95 En e p ises wi h no women a TMT 591 8.17% 14.02% 242,993,692 125.17 20.49 Sou ce: own Tab. 3: Spea man´s co ela ion coe i cien s ma ix Tab. 4: A e age alues o analysed indica o s o en e p ises wi h o wi hou gende di e si y in he op managemen eam EM_2_2019.indd 135EM_2_2019.indd 135 19.6.2019 15:11:1919.6.2019 15:11:19 136 2019, XXII, 2 Business Adminis a ion and Managemen F om he i gu es in Tab. 4, i is e iden ha en e p ises wi h one o mo e women in he op managemen eam p esen a highe le el o i nancial pe o mance as measu ed by ROA and ROE han hose wi h no women in he op managemen eam. Mo e accu a ely, en e p ises wi h one o mo e women ou pe o m en e p ises ha do no ha e any emale di ec o s by a mean o 0.57% in ROA and 3.26% in ROE. The g ea e di e ence in ROE may indica e be e use o equi y in en e p ises wi h manage ial gende di e si y. These indings a e consis en wi h he i ndings o exis ing s udies (Ahmadi, Nakaa, & Bou i, 2018; Julizae ma & So i, 2012; Liu, Wei, & Xie, 2014; Low, Robe s, & Whi ing, 2015) showing ha companies wi h women in op manage ial eams pe o m be e han hose wi hou women in he manage ial eam. Despi e he ac ha he ela ionship be ween he pe cen age o emales in TMT and ROA is small, he o e all i ndings suppo he idea ha he p esence o women in managemen would bene i i nancial pe o mance. In addi ion, he a e age alues indica e ha he g oup o en e p ises wi h gende di e si y in he op managemen eam ha e mo e asse s and mo e employees. Mo eo e , his g oup o en e p ises is sligh ly younge han he en e p ises wi hou women in he op managemen eam. The same i nding was obse ed by Sumed ea (2016), who p o ed ha olde companies ha e mo e masculine managemen eams. 3.3 Clus e Analysis To u he cla i y he ela ionships be ween all indica o s, clus e analysis was conduc ed o he g oup o 232 en e p ises wi h op managemen eam di e si y. As men ioned abo e, clus e s a e compiled whe e each obse a ion belongs o he clus e wi h he nea es a e age, and he clus e s a e concu en ly se as a apa om each o he as possible. I was decided ha he e would be h ee clus e s (see Tab. 5). In clus e numbe one, he e a e en e p ises ha di e conside ably om each o he in e ms o size, age, and he highes a e age alue o ROE (in compa ison o he lowes alues o PWOMAN). Howe e , since he i s clus e con ains only ou en e p ises, hey should be seen as a g oup o excep ions ha o med an indi idual clus e . I hese ou en e p ises we e included in o he clus e s, hey could conside ably dis o hei a e age alues. The o he wo clus e s a e mo e signi i can ; clus e 2 wi h 43 en e p ises and clus e 3 wi h 185. The second clus e con ains en e p ises wi h he highes alues o asse s and he highes ROA. The hi d clus e con ains en e p ises wi h he lowes alue o asse s, he lowes numbe o employees, he highes po ion o women in op managemen eams, and he highes ROE. When analysing a e age alues o ROE, ROA, and PWOMAN using he esul s o clus e s 2 and 3, a posi i e ela ionship can be seen be ween PWOMAN and ROE. The highe he PWOMAN, he highe he ROE. On he o he hand, ROA is highe in he clus e wi h he lowe PWOMAN, which denies he p e iously obse ed esul s. Howe e , in clus e 3, ROA shows i s a e age alue (compa e wi h Tab. 2). Thus, he cause o he illogical esul lies in clus e 2, whe e he e a e only 43 en e p ises. Wi h his low numbe o en e p ises, he a e age can be easily dis o ed by highe alues. Namely, he e Clus e 1 Clus e 2 Clus e 3 Numbe o en e p ises 4 43 185 AGE 18.75 20.58 19.83 EMP 188 161 128 SIZE 281,652,500 735,050,100 189,007,600 PWOMAN 25.89% 30.97% 37.37% ROE 18.88% 15.66% 17.62% ROA 8.14% 10.35% 8.38% Sou ce: own Tab. 5: A e age cha ac e is ics o indica o s in indi idual clus e s EM_2_2019.indd 136EM_2_2019.indd 136 19.6.2019 15:11:1919.6.2019 15:11:19 137 2, XXII, 2019 Business Adminis a ion and Managemen a e h ee en e p ises wi h conside ably high ROA (64%, 29.7% and 25.2%) alues. These en e p ises we e likely included by STATISTICA in o clus e 2 mainly due o hei high alue o asse s. I hese h ee en e p ises we e excluded om his clus e , he a e age ROA dec eases o 8.3%, while he ROE dec eases only o 14.07% and PWOMAN o 30%. Addi ionally, his analysis con i ms a posi i e ela ionship be ween PWOMAN and ROA. To iden i y he c i ical mass ( a io o women in op managemen eams), Tab. 6 was compiled. I con ains da a om 232 en e p ises wi h op managemen eam di e si y. Fi s , he esul s show ha he en e p ises wi h a p opo ion o women in TMT om 10% o 20% ha e lowe median ROE and ROA han he en e p ises wi h no women in TMT. These i ndings a e consis en wi h Kan e ’s (1977) heo y o okenism. Con a y o Kan e ’s p opo ion o oken g oup (up o 20% o mino i y g oup), we i nd ha ROE and ROA medians a e also lowe in he g oup o en e p ises wi h p opo ions om 21% o 30% compa ed o he g oup o en e p ises wi hou gende di e si y in TMT. A possible explana ion o his is he issue o he mode a ing e ec o cul u al alues, mo e speci i cally o conse a i e alues, which a e s ill dominan in ou con ex . Mo eo e , s e eo yping iews agains women, especially in manage ial and execu i e posi ions, is also a possible mode a ing ac o . Su p isingly, a simila i nding was also iden i i ed o he g oup o en e p ises wi h he p opo ion om 41% o 50%. These i ndings a e simila o hose in he g oup wi h he p opo ion om 21% o 30%. One in e p e a ion o his is he g owing h ea o male dominance due o inc easing gende equali y and he eby de ensi e beha iou (Blalock, 1967; Yode , 1991). Nex , he mos posi i e impac o ROE and ROA was ound in he g oup o en e p ises wi h he p opo ion o 31% o 40% emales in TMT. The e o e, ou conclusion is ha c i ical mass o emale ep esen a i es in op managemen eams is eached a a sha e o app oxima ely 31% o 40% o ou sample. Conclusion In his s udy, he uppe echelon heo y and he c i ical mass heo y a e combined o p o ide heo e ical insigh and empi ical analysis o gende di e si y in op managemen eams on company i nancial pe o mance. Unde he uppe echelon heo y amewo k, which p oposes posi i e ela ionships be ween a ie y o TMT demog aphic cha ac e is ic and company pe o mance (Ca pen e , Gele kanycz, & Sande s, 2004; Hamb ick & Mason, 1984), we in es iga ed whe he emale pa icipa ion in op managemen eams posi i ely in l uences company i nancial pe o mance. Based on he c i ical mass heo y p oposing ha women a e unlikely o ha e a posi i e impac in op managemen eams un il hey ise up om a ew okens in o a conside able mino i y in op managemen , we ied o de e mine wha pe cen age o women in op managemen eams cons i u e c i ical mass. Unlike p e ious s udies (Ali, Ng, & Kulik, 2014; Low, Robe s, & Whi ing, 2015; Richa d, Ki by, & Chadwick, 2013; Wa d & Fo ke , 2015), bo h heo ies we e applied o only he con ex o small and medium en e p ises in he manu ac u ing sec o . The s udy ound ha , on a e age, en e p ises wi h women in op managemen PWOMAN (%) Numbe o en e p ises Median ROE Min ROE Max ROE Median ROA Min ROA Max ROA 0591 11.95 -139.96 125.96 8.17 -44.26 79.47 10-20 53 9.60 -24.77 103.01 5.62 -20.81 51.94 21-30 44 11.24 -21.54 48.81 6.00 -7.46 38.47 31-40 71 14.71 -35.39 172.17 8.67 -16.72 55.67 41-50 35 11.26 -23.85 75.68 6.46 -11.86 64.04 51-100 29 12.84 -0.49 71.02 7.80 0.16 37.19 Sou ce: own Tab. 6: Values o i nancial pe o mance indica o s g ouped acco ding o PWOMAN EM_2_2019.indd 137EM_2_2019.indd 137 19.6.2019 15:11:2019.6.2019 15:11:20