{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

19 Roth - SellingWomen Short A ResearchNote on...

Info icon This preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: SellingWomen Short: A ResearchNote on GenderDifferences in Compensationon WallStreet* LOUISEMARIEROTH, University of Arizona Abstract a t Someresearch assuggestedhat,onceallformsof segregation re controlled,here h t is no o t b H s g gender ap in earnings. owever, therresearchuggestshatsubstantial arriers I togenderequality ersistevenwithinoccupations.suggest hatinstitutional orms t n p and marketforces that determinecompensation racticesare likely to produce p I r t a w different esults cross rofessions. hypothesizehatgenderinequality illpersist p on WallStreeteven whenmen and womenholdidentical ob titles.Usinga cohort j w p sampleof securities rofessionals ith highlysimilarhumancapitalcharacteristics, I find statistically ignificant enderdifferencesn 1997 earnings,controllingor i s g f human a I c b capital, ndsegregationyareaoffinance. offer backgroundharacteristics, t v possibleexplanationsfor ariationamongprofessions,mphasizinghe importance e w i p of institutional ractices ithinthesecuritiesndustry. Studies of gender inequality in earnings debate the importance of c a discrimination,once the effects of productivity-related haracteristics nd sex segregationhave been taken into account.Previousresearchhas demonstrated that human capitalvariablestogether account for 30% to 50% of the gender gap in pay (Blau & Ferber 1992; Jacobs 1989; Kilbourneet al. 1994; Marini 1989), while much of the remaining gap has long been credited to sex segregation of the labor force by occupation, organization,and specific job (Baron& Bielby 1980; Bird 1996; Blau & Ferber 1992; Dixon & Seron 1995; * I would like to thankMark Chaves,Dalton Conley,Jo Dixon, KathleenGerson,Doug Guthrie, Ruth Horowitz, Linda Molm, Calvin Morrill, GregPilling, Lynn Smith-Lovin,Henry Walker, and two anonymous reviewersfor comments on earlier drafts. This research was partially g supportedby a DoctoralDissertationImprovement rantfrom the National ScienceFoundation. t Please send all correspondenceo LouiseMarie Roth, Universityof Arizona, 433 Social Sciences A Building, Tucson, Z 85721. E-mail: [email protected] ) The Universityof North CarolinaPress SocialForces, ecember2003, 82(2):783-802 D 784/SocialForces822, December2003 England 1992; England& Farkas 1986; Guthrie & Roth 1999; Jacobs 1989; Marini1989;Reskin& Roos 1990;Tanneret al. 1999).Manyscholarsalso argue that residual discriminationagainst women contributes to the maintenance of gender inequality, even after human capital and segregation have been controlled (Bartlett& Miller 1988;Hagan& Kay 1995;Rosenberg,Perlstadt & Phillips 1993;Sokoloff 1988). However,some recent quantitativeresearchhas suggestedthat there is no gendergap in earningsfor recenthires within some o a & occupations, nce allformsof segregation recontrolled Morgan1998;Petersen ( Morgan 1995;Petersenet al. 1997). In other words, controllingfor cohort, for educational a a preparation, nd for segregation y organization ndjob, theremay b be few,if any,genderdifferencesn pay.Usingdataon securities rofessionals,his i t p article analyzes gender differences in compensation within an elite, maledominatedoccupationand examineswhethersuch differences an be attributed c to humancapitalor to segregation y job. b A notable example testing for gender inequality in a male-dominated i w ( occupations Morgan's1998)studyof multiplecohortsof engineers, hichfound no significant enderdifferencesn earnings ithinthe most recentcohorts.Her i w g b i findingsareprovocative, ecausethe sexcompositionof engineerings moremaledominatedthanthatof most otherprofessions, nd manyhavetheorized hat sex a t i discriminations most d likelyin contextsthat arenumerically ominatedby men (Gutek& Morash1982;Kanter1977;Yoder 1994).Accordingly, organ'sinding M f of statisticalquityin earnings mongrecentcohortsof engineers aises he question r t e a of whether similar processes may operate in other equally male-dominated t H professions. owever, organexplicitly ecognizedhatherfindingsmightnot be M r to c I generalizable otherprofessionalontexts(Morgan1998).In thisarticle, present rich and unique data that challengethe generalization f Morgan's indingsby o f revealinga substantialgender gap in compensationamong recent cohorts in anothermale-dominated profession. d a a c Theoretically,ifferences crossorganizationalnd professional ontextsseem D i a u likely. ifferent ndustries ndtypesof organizations se a varietyof methodsand mechanisms to determine pay rates. Market-drivendemands for particular professionalskills influence pay rates, and institutionalnorms affect starting salaries, romotions,raises,andbonuseswithinoccupations. hus,both markets T p and institutional orcesshapeinternaland external abormarkets nd the relative f l a rewards fferedto workers n variousoccupations. heseprocesses eadto a great o i T l variationin pay scalesacrossoccupations(Bridges&Villemez1991).The degree of i genderinequalityin differentprofessionsand differentorganizations s also r t t likelyto vary, ather hancorrespondingirectlyo productivity-related d inputssuch as human capitalcharacteristics. orexample,in stategovernment obs, Bridges F j and Nelson (1989) found gender inequality in pay that was based partly on institutionalnorms and D partlyon marketefficiencyconsiderations. iscretion, c subjectivity, ustom, and "bureaucratic olitics"stronglyinfluence pay scales, p W S / Selling omen hort 785 even when organizationsreferto marketmechanismsas a justificationfor pay & differentialsBridges Nelson 1989;Nelson & Bridges1999).Thus,the processes ( that might creategenderequityin some settingsand inequityin othersinteract with institutionalmechanismswithin organizationalcontexts. These contexts vary in many features, including criteria for performance evaluation and compensation, governancestructures,organizationalclimate, and flat versus hierarchicalfirm structures.Also, there may be considerablevariationacross organizationalfields in the quality and quantity of interpersonalinteractions both within and outside the employer organization,such as team structures, internal chains of accountability,allocation of responsibility,access to client V accounts,and the conventionsof client relationships. ariation y institutional b context is likely to produce different consequences for gender equality, regardlessof the proportion of females in an occupation. In this research, atafroma cohortsampleof WallStreetprofessionals ermit d p an assessment f genderinequality n a particularly ale-dominated rofessional i o m p contextthatdiffersinstitutionallyrom engineering. he cohortbegancareers n f T o WallStreetin the 1990s.The industry's igh demandfor skilledlaborwith master h of businessadministration MBA) credentials, ueledby the bull marketof the f ( mid- to late 1990s, could have createda more open opportunitystructurefor women at that time than in other historicalperiods (SIAResearch epartment D 1998). In this context,women on WallStreetmight be well positioned to gain incomeequality ith theirmalepeers.However,he institutional ontextof finance w t c is o unique in waysthat may contributeto the persistence f genderinequality. First,most of the highest-payingobs in financialprofessionstend to be very j client-intensive, and a majority of high-paying clients are male-dominated corporations.The nature and amount of client interactionmay contributeto differences in the gender gap in income, especially if client preferencesfor a activities dvantage aleworkers m homophily ndcertain ypesof client-entertaining t a (Blair-Loy2001b; Roth n.d. b). For example, in personal interviews, female r t b respondents eferredo stripclubsor "girlie ars," igarsmoking,and elk hunting c a as client-entertaining ctivitiesin which theirpresencewas unwelcomeand that harmedtheirabilityto developsolid clientrelationships.n this industry, omen I w in corporatefinanceor salesand tradingmay havedifficultiesestablishing ighh o t a profileclientrelationships r gainingaccessto lucrative ccounts,contributing o 2 manywomen'slower compensationvis-a-vistheir male peers (Blair-Loy 001b; Rothn.d.b). Second,securitiesfirmsare formallyflat,with a smallnumberof hierarchical w levels,in comparisonto hierarchical rganizations heresalaryfollowslevel and o s senioritywithin a bureaucratic tructure.As a result,Wall Streethas a rather I unusualinstitutionalizedompensation tructure.n an effortto reward mployees s c e on the basis of their performance, all Streetfirms compensatemost of their W employees with variable year-end bonuses that permit and justify wide D 2 786/SocialForces82:2, ecember 003 variationin total income among workersat the same level and in the samejob. Firm revenuesare divided among "groups" pecializingin particulartypes of s financialfunction or client industry.Within each group,employeesare ranked relativeto others at the same level on the basis of performanceevaluations, and managers allocate bonuses that are based on these rankings.1 These compensation practices permit differences in total compensation for d employeesin the samejob and also allow for much managerial iscretionover interemployee differences within each work group. In personal interviews, female securities professionals often claimed that managerial discretion produced gender-biased compensation outcomes. These industry-specific practicesmay also lead us to anticipatehow this context producesdramatically different results for recent female entrants in the securities industry than Morgan (1998) found among her sample of engineers.Thus, the gender gap in earningswithin occupationsmay not follow a patternof disappearance ver o time acrossindustriesbut mayvarywidelyacrossinstitutionalcontexts. In this article,I investigategenderinequalityin compensationamong elite W f w businessschoolgraduates ho entered allStreetsecurities irmsduringa threeb l yearperiodimmediately eforeWallStreet'songestbull marketin history(Wall c 1 Street 1 Journal, 5December 998,pp. C1,C25).I employa three-yearohortsample a thatwascarefully hosento controlforhumancapital, rganizationalrestige, nd c o p c marketconditions.In light of the institutionaland organizational ontextof the I w t securities ndustry, hypothesize hatgenderinequality ill persiston WallStreet i even when men and women hold identicaljobs. Research Design DATA w The datacome froma cohortof businessschoolgraduates ho begantheircareers in one of nine majorWallStreetsecuritiesfirms in the early 1990s.I randomly selected44 women and 32 men fromplacementreportsand alumniinformation for the years 1991-93 from five elite graduateprograms in finance.2Those i w w respondents ho remainedn the industry ouldhavepassedtheirfirstpromotion and shouldhavebeen at the vice presidentrankor higherat the time of the data collection.Onlythose respondents ho wereemployedand paidin the securities w w industryin 1997wereincluded,eliminating3 malerespondents ho had moved T into other industries.3 he remainingsamplecontains73 cases. In MBAprogramsnationally, omen represent pproximately 5-30%of all w 2 a S w ( graduates Guideto Business chools1996).Among the individuals hose names wereobtained hroughthebusiness choolsin thisstudy, omenconstituted 9.8% s w 1 t of the graduates ho enteredWallStreetsecuritiesfirms upon graduating. he w T W S / Selling omen hort 787 five graduateschools of business contained 621 graduateswho entered major Wall Streetfirms during the three years included in the sample.All graduates who startedtheir careersin nine majorWall Streetfirms as defined by Eccles and Crane (1988) were eligible for inclusion in the study.4These firms are the largest securities firms in a highly concentratedindustry,which has become increasingly consolidated over time (SIA Research Department 1998). To obtain comparable samples of men and women, I separated all eligible respondents'names by genderto createlists of 498 men and 123 women and selected separatesubsamplesusing a random numbers table. Many members of the cohort had changed firms within financialservices or had exited the profession.Using telephonedirectories, nternetsearches, nd I a the assistanceof alumni associations,I trackedpotentialrespondentswho had left the companieswherethey startedfinancialcareers.Among those randomly selectedfor an interview,10 refusedto participate(5 women and 5 men) and 29 were impossible to find (17 women and 12 men). These prospective respondentsmay have been impossibleto find becausethey had changedfirms or because they had moved to another city or country, both of which are common in this industry. Also, women were particularlydifficult to locate becausemany had marriedand changedtheir surnamessince business school or did not have a telephone number listed in their names. Thus, the response rateis 66%and the cooperationrateis 88%.Afterthe disqualification f 3 men o who had exited financialservicesbefore the end of 1997, the sample contains 44 women and 29 men. In-depth interviewsabout their careerhistories and total compensationfor 1997were conductedin 1998and 1999.5The interviews lasted an averageof one hour. While the sample size is modest, this business school cohort sample has several advantages.First, it controls for important human capital variables, market conditions, and organizational prestige. Elite MBA programs are mediating institutions that control for the most important educational f credential or a careeron WallStreetand,to a greatextent,for classbackground. In addition,these programsfilterfor previouseducationaland work experience T and for performanceon the GMAT. his sample of respondentswho entered w a small set of organizations ith comparableprestigewithin a single industry during a short period of time also largely controls for variation in market m conditions and organizational arketposition. These controls that were built into the samplingstrategy ermittedan assessmentof genderdifferences mong p a very comparablemen and women from a recent cohort of financial service professionalswhose labor was in high demand in the late 1990s. Second, the samplewas able to be drawnwithout the involvementof the firms themselves. Gaining the cooperation of even one of these highly secretive organizations a would havebeen extremelydifficult.Moreover, firm-generated amplewould s likely have contained less variation in areas of financial services, and 788/ SocialForces82:2, ecember2003 D respondentsmight have been less candid if they believed that their employer could have access to the data. COMPENSATION: THE DEPENDENT VARIABLE The outcome of interestfor this analysisis total compensationin 1997.Possibly W a morethanotherindustries, allStreetis drivenby compensation s a measureof success. As a baseline for this population, a survey of executive recruiters specializingin the securitiesindustry indicated that the median total annual compensationin 1997for the businessschool classof 1991was $635,000;for the classof 1992,$540,000; nd for the classof 1993,$430,000(Horowitz& Copulsky a 1998). It is revealingthat this survey indicated that "low range"Wall Street w i 3 professionals ho had graduatedn 1993,most of whom wereapproximately 0 e yearsold at the time of the survey, arned$300,000in 1997.This amountis more thantwicethe $128,521cutofffor the 95th percentile f householdincomein the o U.S.for the sameyear(U.S.Bureauof the Census1998),illustratinghe degreeto t which this population representsan economic elite. While some respondents h bemoanedthe emphasis n money,allWallStreetprofessionals ad a strongsense o that their compensation was an indicator of their success and their value as a o T t professionals. his orientation uggests hat earnings rean appropriate utcome s to analyzeand a proxymeasurefor successin the industry. In the currentstudy,respondents ereaskedto checkthe appropriate ox on w b an income scale for their total compensationin 1997. Five respondents(three womenandtwo t t a men) refused o disclose heircompensation ndareomittedfrom F the analysis. orall otherrespondents,ncomein dollarsis codedas the midpoint i of the rangeindicatedunless the respondentgave an exact income amount, in which case that number is used instead.6For respondentsearningmore than $1,000,000, income is coded as $1,100,000.7 The dependent variable in all m regression odels is the naturallogarithmof income in dollars. CONTROL VARIABLES h Controlvariables nclude i c background haracteristics, uman capitalother than the MBAdegree, ankin the industry, reaof finance,andwhetherthe respondent r a still workedfor one of the originalnine WallStreetfirmsat the time of the 1997 bonus payment.Measuresof backgroundcharacteristicsnclude family status i f a o w a dummyvariablesorwhether respondent asmarried r hadchildren t the time of the interview. ecause reviousresearch as revealed ifferent onsequences f B h d o p c f b parenthood or men and for women, I alsotestedfor interactions etweengender andparental tatus. owever, nderno circumstancesid the maineffectof having s u d H * children r the interactionerm(gender children) ecomestatisticallyignificant, o t b s and the interaction term was omitted in the final models.8 Race is coded dichotomously as white/nonwhite (nonwhite = 1). However, only seven W S / Selling omen hort 789 respondentswere not white, among whom one was AfricanAmericanand the i remainingsix were of Asiandescent.Becausethis variablewas nonsignificant n all models, it is omitted in the final models presentedin Table2. Threedummy variables capture rank within the industry, indicating whether respondents were below the vice president level, at the vice president level, or above the vice presidentlevel. Below the vice presidentlevel is the referencecategoryin multivariatemodels.9 The sample holds constantthe most importanteducationalhuman capital, h M sinceallrespondents avea prestigious BAdegreeandsimilar earsof experience y in the a S a h industry ftergraduating. everal dditional umancapitalcontrolvariables arealso included.I includea dummyvariable or havingan undergraduate ajor f m in economics,finance,or accounting ecausethesemajors ayofferskillsrelevant b m o f to careers n WallStreet.I testeda dummyvariable or whethera respondent ad h an undergraduate ajorin mathematicsor engineeringand a dummy variable m for workexperience n WallStreetpriorto completingthe MBA,eitheras a twoo o yearanalystin investmentbanking,as a salesassistant, r as a summerassociate. Thesevariables reincludedin model 1 but excludedfromthe fullmodelbecause a l e f of a consistent ackof significant ffects.Dummyvariablesorthe graduate chools s w t w of business hererespondents ompleted heirMBAdegrees erealsotested.None c of these variableswas significantin any models. Consistentlynonsignificant variableswere excludedfrom the final models. Includingthese variablesin the full model producedresultssimilarto those foundwhen theywereexcluded,and thesevariables id not affectthe d o magnitudeor significance f othercoefficients. An important ontrolvariables the naturalogarithm f the estimated umber c i o n l d of hoursworked erweek.Therespondent efinedhis or heraverage eeklyhours, p w whichactsas a measure f workeffortandworkcommitment. elf-reportedours o h S are expected to be inaccurateand somewhat inflated. However,one expects t T respondents o inflatetheirweeklyhoursin similarwaysacrossthe sample.'0 he naturallogarithmis usedbecausethe dependentvariables the naturallogarithm i i of i T earnings. hus,a valueof 1 would indicatethatearnings ncrease n proportion to hours.In this demanding rofessionone would expecta coefficient argerthan p l r 1, implyingthat thereareincreasing eturnsto hoursworked.ll Research on gender and labor markets suggests that sex segregation by function is also likely to lead to gender differencesin earnings.Consequently, dummy variableswere included for the variousfunctionsof financialservices: ...
View Full Document

{[ snackBarMessage ]}

What students are saying

  • Left Quote Icon

    As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

    Student Picture

    Kiran Temple University Fox School of Business ‘17, Course Hero Intern

  • Left Quote Icon

    I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

    Student Picture

    Dana University of Pennsylvania ‘17, Course Hero Intern

  • Left Quote Icon

    The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

    Student Picture

    Jill Tulane University ‘16, Course Hero Intern