Anda di halaman 1dari 35

Race, Gender, and Authority in the Workplace: Theory and Research Author(s): Ryan A.

Smith Reviewed work(s): Source: Annual Review of Sociology, Vol. 28 (2002), pp. 509-542 Published by: Annual Reviews Stable URL: http://www.jstor.org/stable/3069251 . Accessed: 25/09/2012 10:43
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Annual Reviews is collaborating with JSTOR to digitize, preserve and extend access to Annual Review of Sociology.

http://www.jstor.org

Copyright(

Annu. Rev. Sociol. 2002. 28:509-42 doi: 10.1146/annurev.soc.28.110601.141048 2002 by AnnualReviews. All rightsreserved

RACE, GENDER, AND AUTHORITY IN THE

WORKPLACE: Theory and Research


RutgersUniversity,School of Managementand LaborRelations,50 Labor CenterWay, New Brunswick,New Jersey 08901; e-mail: rasassocOl@aol.com

RyanA. Smith*

raceinequality, attainment, authority genderinequality Key Words job authority, to the studyof job authorU Abstract Thischapter surveys sociologicalapproaches andemergence as animportant ditheoretical measurement, foundations, ity,including The focus hereis mainlyon studiesof race andgender mensionof social inequality. of raceandgender andthe consequences in the determinants of authority differences in the overall advancements for income.Despite significant differencesin authority and remain race of minorities andworking status socioeconomic women, gender imporand of authority. Thispattern, whichis consistent to theirattainment tantimpediments andcross-temporal robustin state-level, studies,is sustained national,cross-national, andstructural locationwithinandbenet of anincumbent's human capitalinvestments for tweenseveraleconomicunits.Followinga reviewof thepredominant explanations injob authority is the conclusion thatthe mostpromising genderandracialdisparities in authority concernthe racial forpersistent racialandgenderdisparities explanations and the tendencyon the partof authority of the workplace and genderdemography bothexclusionary andinclusionary themselvesthrough elites to reproduce processes. of theseprocessesbased includeadditional delineation forfuture research Suggestions of menandwomenandstudiesthatsyntheon samplesof multiple racial/ethnic groups andqualitative to understanding the effects of employer size quantitative approaches andpractices on the authority attainment andemployeeattitudes/preferences process.

INTRODUCTION
of job authority as animportant dimensionof social inequalThe conceptualization treatisesof MaxWeberandKarlMarx.The ity may be tracedto the earlytheoretical in quantitative assessmentsof workinequalityis rooted most directuse of authority in Dahrendorf's(1959) critiqueand extension of Marx'stheoryof class relations andclass conflict.Threeconsecutive,andsometimesoverlapping, developmentsin the sociological studyof job authority may be discerned.First,buildingon the work of Karl Marx and Max Weber,researchersembarkedupon theory development, and the measurementof job authorityfor use in quantitative operationalization, address: P.O.Box 357,Maplewood, NJ07040. *Current
0360-0572/02/0811-0509$14.00 509

510

SMITH 1Next, the MarxandWeberlineages studiesof class analysisandstatusattainment. the andmeasurement of authority into two splintered subsequent conceptualization differentbutnot necessarilymutuallyexclusive arguments. Oneis theneo-Marxian argumentthatjob authorityis a categoricaldiscretephenomenonthatlends itself to the largerstudyof class analysis(Wright& Perrone1977, Lopreato1967, 1968, Wright 1993). The other argumentinvolves the view thatjob authorityis gradational or scalar,lending itself to the analysis of statusgroups-which is more in line with the Weberiantradition(Blau 1977, Halaby & Weakliem 1993, Halaby 1993). As a corollaryto this debate, some researcherschallenged Dahrendorf's originaldichotomousformulationof authority, arguingthatjob authoritycan also be conceptualizedas a polytomousvariable-that is, a variablewith threeor more hierarchical levels (Robinson 1979, Robinson & Kelley 1979). embedded in the debatesurrounding the relativeinfluenceof demandFinally, side versus supply-side explanationsof group inequality, the role of race and criteriain explainingauthority outcomes,has genderversusachievement-oriented dominatedauthority researchin recentyears. Overall,this researchechoes earlier claims that authorityis a unique and importantindicatorof workplacestratification comparableto, sometimesmore helpful than,or otherwisecomplementary of traditional indicatorsof socioeconomic statusin explainingbothgenderandracial inequalityat work. This chapterreviews the sociological literature on race, gender,andjob authorThe theoretical of foundations and ity. job authority its emergenceas an important indicatorof socioeconomic statusare discussed.Next, this chapterexamines how has been operationalized andmeasuredfor use in quantitative assessjob authority ments of workplaceinequality.It then reviews severalprominentexplanationsfor race and genderdifferencesin authority, followed by a summaryof the dominant literature thatdescribesthe causes and financialcosts of racial and genderexclusion fromjob authority. The most promisingexplanationsfor persistentracialand in genderdisparities authorityoutcomes concernthe racial and genderdemograelites to reproduce phy of the workplaceand the tendencyon the partof authority themselvesthroughboth exclusionaryand inclusionaryprocesses.Thereis a definite need for more studiesto further and explorethese phenomenavia quantitative to detailed information on decisions made on qualitativestrategiesdesigned yield the supply-sideand demand-sideof the authorityrelationship.

JOB AUTHORITY AND ITS MULTIPLEDIMENSIONS The Importance of Job Authority


thata comAccordingto Max Weber,authority may be definedas the "probability mandwith a given specific contentwill be obeyed by a given groupof persons...
1Adetailedreview of the class and statusattainment traditionsis beyond the scope of this chapter.See Kurz & Muller (1987) for the formerand Matras(1980) and Brieger (1995) for the latter.

ANDWORKPLACE AUTHORITY RACE, GENDER,

511

The importantdifference between power and authorityconsists in the fact that whereaspower is essentially tied to the personalityof individuals,authorityis always associatedwith social positions or roles... authorityis a legitimaterelation of dominationand subjection.In this sense, authoritycan be describedas legitimatepower"(quotedin Dahrendorf1959, p. 166). In the last 20 years,the studyof the relative distributionof legitimatepower (job authority)within the context of areaof sociological investigation.Job the workplacehas emergedas an important it is authority psychologically rewarding; bringsstatusboth inside and outside the workplace;it is related to job satisfaction,autonomy,class consciousness, class position, voting behavior,partyidentification,and political views (Kluegel 1978, Halaby 1979, Robinson & Kelley 1979, Wright 1979, Roos 1981, Spaeth 1985, Mueller & Parcel 1986, Jaffee 1989, Reskin & Ross 1992, Adler 1993, Reskin & Padavic 1994, Wilson 1997b, Smith 1999). In the sociological literature, however, job authorityis perhapsmost associated with income. As noted by Wrightet al. (1995, p. 407), "job authorityis one of the central ways in which the financial rewardsof work are allocated."Elsewhere,Halaby& Weakliem(1993, p. 17) deto be sociology's chief contribution to the studyof claredthe studyof job authority is a highly coveted workplaceresource. earningsinequality.In short,job authority As such, it comes as no surpriseto learnthatit is unequallydistributed by race and gender in Americansociety and cross-nationally. Explainingwhy this is the case is one of the main objectivesof this chapter.

Select Typesand Dimensions of JobAuthority


Job authorityhas been conceptualizedand subsequentlymeasuredin a variety of ways. Two majorclassificationsof authorityhave dominatedsociological literature:control over organizational resourcesand control over humanresources. Withinthese two possibilities, severaltypes of organizational have been authority identified. Ownershipin the form of control over the means of production,also known as control over the labor power of others (Wrightet al. 1982), is perhaps the ultimate form of authority,but researchershave traditionallyconceptualized ownership as something separate from authority-especially in postindustrial societies (Dahrendorf1959). Sanctioning authority or span of responsibilityincludes the ability to influence the pay or promotionsof others (Mueller et al. 1989, Wrightet al. 1995). Span of control-denotes the numberof people under directsupervision(Muelleret al. 1989). Decision-makingor managerialauthority relatesto organizational policy decisions, controloverproducts,services, budgets, or purchases(Rosenfeld et al. 1998). Also, hierarchicalauthorityposition refers to an individual'sformallocationwithinthe structure of organizational hierarchies (Kluegel 1978, Wrightet al. 1982, Speath 1985, Wrightet al. 1995). Finally,measuresof supervisoryauthorityestablishwhetheran individual"supervisesanyone on thejob," a query,however,thatfails to distinguishnominal supervisorystatus (relaying informationfrom superiorsto subordinates)from the exercise of real authority(Wrightet al. 1982, p. 714). Thus, there are multiple types and various dimensions of authority.Because of this, researchershave approachedthe study

512

SMITH of race andgenderdisparitiesin authority frommanydifferentangles-a fact that, as discussed below, paintsa multidimensional and often ambiguouspictureof the extent of racial and gendergaps in authority.

AND MEASUREMENT THEORY OF AUTHORITY


In his provocativebook, Class and Class Conflictin IndustrialSociety(1959), Ralf Dahrendorf developed a theorydetailingthe importanceof authority, specifically job authority,in postindustrialorganizations.He arguedthat differentialsin job the dynamicsof class relationsand class authoritywere criticalin understanding conflict in modem society. Even though it is beyond the scope of this chapterto discuss the debate surrounding the source of class conflict, it is important to pinfrom Marx. For Dahrendorf,class formation point Dahrendorf'schief departure and class conflict ensue not from opposing intereststhatexist in the relationsbetween those who own the means andproductionversusthose who do not, as Marx would contend. Instead, importantsocietal changes flowing from the onset and have led to the separationof those who own the developmentof industrialization meansof production fromthose who exercise controloverthe meansof production in the form of legitimateauthorityover both organizational resourcesand human resources (Lopreato 1967, 1968, Hazelrigg 1972, Fox et al. 1977, Robinson & & Cannon1987). Dahrendorf as the basic Kelley 1979, Vanneman posits authority determinant of class divisionin postindustrial society wherebyconflictensues over the mannerin which authorityis unequallydistributed in society (Lopreato1967, p. 281). Since conflict necessitates two opposing groups, Dahrendorfconceptualized authorityin strict dichotomous terms (i.e., those who exercise authority versusthose who are subjectto it). Because this construction has important implications for the mannerin which authorityis measuredin quantitative analyses of job inequality,it has met with widespreaddebate. Some have arguedthat a more accuratedepictionof the postindustrial workplacewould lead to an expandedconof on various based hierarchicaland multidimensional ceptualization authority Robinson 1979, Robinson & Kelley 1979). configurations (Lopreato1967, 1968, A brief review of this literature follows.

Dichotomous VersusPolytomous Authority


Presaging a later debate in authorityresearch on whether authorityshould be conceptualized and measured as a class or status variable (Wright & Perrone 1977, Halaby & Weakliem 1993), Van den Berghe (1963) deridedDahrendorf's tendency to view class conflict in solely dichotomousterms. According to Van den Berghe, "Reducingevery conflict situationto a dualist opposition involves strainingthe facts. Dahrendorf experiencesthe same difficultyas Marxin handling 'intermediate groups"'(p. 701). The debate over whetherjob authorityis better conceived as a dichotomous variable (having authority or not) or polytomous variable (no, low, and high

ANDWORKPLACE AUTHORITY GENDER, RACE,

513

authority,etc.) marks an importantchapterin the annals of authorityresearch. While some evidence supportsDahrendorf'soriginal two-categoryconstruction & Cannon1987), strongempirof authority & Jackman1983, Vanneman (Jackman differencesbetween incumbentsof low andhigh ical evidence points to important positions of authoritynot only in income (Lopreato 1968, Kluegel 1978, Smith 1997), but in subjectiveclass identificationand sociopoliticalattitudes(Robinson & Kelley 1979). Robinson & Kelley (1979), who conductedthe most extensive "Theauthority examinationof this issue, remarked: componentof the class system wouldhaveit, is, we suggest,best viewed not as a simpledichotomy,as Dahrendorf but as a continuum.... Ourcontinuousversion of authorityexplains significantly more variancein income than Dahrendorf'sdichotomy,indicatingthat there are income differenceswithin the commandclass" (p. 54). important When removed from larger sociological concerns of how best to characterize the stratification order,the mode of operationalizing workplaceauthorityhas largely dependedon the kinds of researchquestions being asked. In cases where researchersare simply interestedin the questionof who has authorityversus who does not, a simple dichotomous measure of authorityhas sufficed (Bridges & Miller 1979, p. 678, Hill & Morgan 1979, Wolf & Fligstein 1979a,b, Hill 1980, & Canon 1987). However,in othercases, the relativedisHalaby 1986, Vanneman tributionof authoritymay be measuredaccordingto the amountof authority(i.e., mean levels) thatan individualor grouphas, or perhapsone's location withinhierIn eithercase, both scalarand polytomousmeasures archicalauthoritystructures. of authorityare appropriate (Kluegel 1978, Robinson & Kelley 1979, Kalleberg & Griffin 1980, Wright et al. 1982, Spaeth 1985, Mueller et al. 1989). As discussed below, group differencesin both access to job authorityand its effects on workplaceoutcomes significantlyvary dependingon how authorityis measured. The debateover whetherto operationalizeauthorityas a dichotomousor polytomous variableis very much linked to theoreticaland empiricaldebatesover the use of class typologies versusstatusscales in social stratification studquantitative ies. A surveyof thisliterature revealsthreeimportant foci linkedto the development of authority research:the synthesisof opposingtheoriesof social stratification, the and measurementof job authorityfor predictionsof earnings operationalization inequality,and the assessmentof race and sex differencesin access to and income returnsfor authority. These threeareasof study are addressedin sequence below.

Authorityas Class Typologiesor Status Scales


The origins of the class versus statusdebate in authorityresearchmay be traced to Wright & Perrone's (1977) attemptto link Marxist theory with quantitative assessmentsof social inequality.Two sourcesof class differentiation characterized theirmodel: ownershipof the means of productionand the availabilityof control by respondentsover the labor power of others. Here the focus is on the latter. Respondentsfrom ISR's Surveyof WorkingConditions(1969) were asked, "Do you superviseanyoneas partof yourjob?"(p. 36). Individuals answeringyes to this

514

SMITH questionwere placed in the managercategory,which aptlydescribesan individual over subordinates in the workplacebutis also dominatedby the who has authority or "ambiguous" ownersof the means of production-the "intermediate" category describedby Vanden Berghe.While this is a difficultcategoryto classify within a an important Marxianframework traditional (Parcel& Mueller1983), it represents research. extension of Marx'straditional class model for use in quantitative As with Wrightandhis colleagues, Robinson& Kelley (1979) andKalleberg& researchas one dominatedby the status Griffin(1980) viewed social stratification attainment tradition,which proponentsof the conflict approachsaw as obscuring classes within hierarchian analysis of the inequalityflowing from contradictory cal work structures.Robinson & Kelley took themselves to be bridgingthe gap between the Blau & Duncan (1967) paradigm(e.g., status attainment/American theoriesoffered by KarlMarx and Ralf tradition)and the class conflict/European of class inequality tradition The former,which grewout of a Weberian Dahrendorf. (Blau 1977), stressed the importanceof individuallife chances as a function of marketrelations.Key indicatorslike education,training,and workexperiencenot amenableto only dominatedthis perspective,but such measureswere particularly along a continuousscale. operationalization in contrast,emphasizeddiscreteclassificationsinvolving The conflictapproach, individuals or groups located within the relational structureof ownership and researchpriorto controlof productionand was far less the subjectof quantitative thatauthority In fact, Dahrendorf's (1959, p. 171) argument Wright'sconstruction. does not permitthe constructionof a scale denotes the reluctanceon the part of Europeanconflicttheoriststo see beyonddiscreteclass categories-a myopiathat was as severe as American sociologists' inability to see beyond status scales. While there were deeper theoreticaland empirical rifts between American and Europeansociologists (e.g., differentways of conceptualizingandverifying social andmoreexplicitly,Robinson& Kelley's work, inequality),Wright'scontribution, served a deliberateunifying purpose.Before such efforts, it would be difficultto argue with Robinson & Kelley's assertion that "the Blau-Duncan and conflict andto ignore tend to deal only with theirschool's model of stratification traditions othermodels"(p. 42). Drawing on the strengths of the status attainmentand conflict traditions, Robinson & Kelley showed that a synthesis of Marx and Dahrendorfvariables producedstatisticallysignificanteffects on income in the United States and Great Britain.Specifically,once the synthesizedvariables(i.e., controlover the meansof areincludedin modelspredictingincome, and theexerciseof authority) production the amountof varianceexplainedincreasesby nearly50%over andabove the Blau & Duncanstatusmodel for men butnot women.Robinson& Kelley concludedthat two stratification systems existed-one based on Marx'smeans of productionand andtheotherbasedon statusas indicatedby educational Dahrendorf's job authority andoccupational enough,the two systems aresaidto position(p. 54). Interestingly be theoreticallydistinctand withoutoverlapin regardto the differentmechanisms by which class and statusare passed down from one generationto another.

ANDWORKPLACE AUTHORITY RACE, GENDER,

515

Going a step furtherin theirevaluation,Kalleberg& Griffin(1980) arguedthat conflict models actuallyperformbetterthanDuncan'ssocioeconomic index (SEI) in predictingincome. Like Robinson& Kelley, Kalleberg& Griffincapitalizedon threeemergingideas: Class as defined Wright'stypology.Theirstudyunderscored control is an and supervisory importantsource of income inequalby ownership distinct entities that must be conceptualized and status are class ity; occupational and empiricallyexamined independently,and Duncan's SEI does not adequately explain inequalityin job rewards. While the class typology versus statusscales debatewould continuefor another decade, some researchersfound the need to unify ratherthanchoose between the research(Spaeth1985). Althoughothereffortsto dominantstrainsof stratification of were alreadyunderway, Spaeth'sunique some this literature aspects synthesize on a control based "resource perspective"-the idea thatorganizational approach, over in control resources andcontroloverpersonnelhadimportant monetary power for implications earnings disparities-represented an importantadvancementin research. authority It is important to note the key distinctionbetween Spaeth'scontribution to the relativeto thatof Robinson& Kelley's (1979). Both attempted authorityliterature to unify importantaspects of the stratification literature,yet Speath's analysis is arguablymore extensive in that he assesses the effects of four types of stratification variableson income: Wright'sclass categories, Duncan's Socioeconomic Index,Robinson& Kelley's measuresof organizational level, andKluegel's scalar measureof supervisoryauthority.Robinson & Kelley, on the other hand, assess variableson income: class variables the effects of only two types of stratification andWright,andtraditional drawnfrom concepts developedby Marx,Dahrendorf, statusattainment variables(father'sandrespondent's educationandoccupation)as offeredby Blau & Duncan.The empiricalquestionaddressedby Spaeth,Robinson & Kelley, Kalleberg& Griffin,and laterHalaby & Weakliemis whethercontinuthanarediscrete/class ous/statusvariablesarebetterat predictingincome/earnings variables. In determiningwhich set of stratification variablesarebetterpredictorsof earnings, Spaethshows thatmeasuresof controloverpersonnelandmonetaryresources arenot only strongdeterminants of earnings,buttheyreduceto nonsignificancethe effects of both statusattainment variablesand generalmeasuresof authority(i.e., whetheror not a respondenthas supervisorystatus).Halaby & Weakliem(1993) arguedthat,when it came to predictingearningsinequality, Wright'sdiscreteclass categorieswere not as elegant and parsimoniousas continuous(scalar)measures of social status such as those offered by Robinson & Kelley (1979) and Spaeth (1985). At this juncture,authorityresearchunderwentan importantshift from an emphasis on whetherclass or status models were better at predictingincome to the practicalperils of linking general theory to the narrower question of variable measurement. In offeringa more simplifiedconceptualization andsubsequent measurement of thecontrolsourcesof earningsinequality, & Halaby WeakliemchallengedWright's

516

SMITH class model on the groundsthat it lacked conceptualand operationalcoherence and because it was not as parsimoniousas models offeredby others.They argued that "Wrightintroducedtwo different,even competing, class typologies, but no empiricalgroundsto choose between them"(p. 17). In response, Wright (1993) argued that, at best, the question of whether to use scales or typologies to predict earningsdependedon the researchquestions and the theoreticalparadigmon which they are based. Wright contended that the kinds of questions he was asking, rooted as they were in the theoretical bases of contradictory class locations,were particularly suitablefor hypothesizing aboutthe causal mechanismthatled to earningsinequalitybased on discreteclass on an eight-pointcontinuousscale, Halaby& Weakliem's categories.2Constructed authoritymeasure-Wright would contend-fundamentally obscuredthe underclass analysis, which, in Wright'stypology, included lying logic of contradictory the nominalcategoriesof capitalist,workers,andthe difficultto classify categories of managersand supervisors(p. 34). The Wright/Halaby exchangeis important beyond whatit reveals aboutthe diresearchhas developedin the last two decades. alecticalmannerin which authority On a deeper level, the debate highlights the chasm between theory development and theory testing. The gap between the two is often wide and only infrequently and ratherimperfectlybridged.While theories may be well developed, ways of accuratelytesting them may be less so. By the same token, variablesthat are indicatorsof some social phenomenamay be appropriately measured,but theirlink to well-developedtheoriesmay be tenuousat best. Nearly a decade afterthe Wright/Halaby exchange, the questionof whetherto use continuousscales versuscategoricaltypologies to trackthe processesthatlead to earningsinequalityin returnsto authoritymay very well remaina moot issue. have benefitedfrom this debate.Wright'srefrainthat"it Studentsof stratification all depends on the researchquestion"cannot be disputed, and Halaby's call for of key explanatoryvariables more parsimoniousand coherentoperationalizations across studiesis as poignanttoday as it was nearlya decade ago. Perhapswhat we debatesis thatany attemptto conceptualize learnedmost fromthe aforementioned
his critics,non-Marxian andMarxianalike,at 2Itis interesting to note thatWrightanticipated a worthyendeavor but the very inceptionof his comparative it, nonetheless, project thought Marxism' runs a numberof significantrisks. By to pursue. As he noted: "'Multivariate proposing to operationalizeand measure a range of centralconcepts within the Marxian tradition,it risks reducing the complexity of those concepts to a few simple empirical categories.By deploying these empiricalmeasuresin statisticalmodels, it risks losing the 'dialectical' and dynamic characterof the explanationsMarxistsgenerally advocate.And by suggesting that Marxianargumentscan be formulatedas 'testable hypotheses' within multivariate equations,it encouragesempiricistattackswhich may frequentlydo more to confuse thanclarifyissues. The underlyingassumptionof the Comparative Projecton Class andClass Consciousnesshas been thatthese risksareworthtaking"(Wright1989, Structure p. 16).

AUTHORITY RACE, GENDER,AND WORKPLACE

517

the stratification orderbasedon a single variableis boundto offer or operationalize a mere partialglimpse of society at any given point in time. Otis Dudley Duncan made the point this way: Thereis no such thingas a single index of socioeconomic statussuitablefor all purposesof social researchin a modem, complex society. Even in small and of the staticcommunitiesof the United States,it is a patentoversimplification facts to suppose that the whole populationmay be placed unambiguouslyin intervalsof a single scale of "class"or "status."Given the actualcomplexity variable of the stratification and multidimensionality structure, any particular a selected of a structure that be or index can at best reflect may strategic aspect from a certainpointof view. (Duncan1961, p. 139;quotedin Wolf & Fligstein 1979a, p. 104.) This suggests that prior syntheses of multiple conceptualizationsof society (class and status viewpoints), and multiple operationalizationsof stratification measuresand indices (dichotomous,polytomous, and scalar approaches),reprea more complete pictureof the stratificasent the firstbest steps to approximating tion order(Robinson& Kelley 1979, Spaeth 1985). In a roundaboutway, it would be the lack of such a synthetic approachthat would give rise to studies of race and gender differences in authorityat work. Importantresearchfrom the status attainmenttraditionwould produce findings that appearedincongruentwith what was known at the time about the distribution of minorities (mainly blacks) and women in the stratificationorder.When (education,occomparedto earlier time periods, key indicatorsof stratification cupational status, and income) showed significant improvementsfor blacks in absolute terms and relative to whites during the 1970s (Featherman& Hauser 1978, Hout 1984). Similarly, gender studies of status attainmentreportedthat men and women had similar levels of occupationalstatus and very few differences in the processes that led to occupationalattainment(Featherman & Hauser and 1976, Treiman& Terrell1975). Both sets of findingsseemed counterintuitive raised questions about whetherstatus attainmentindicatorswere indeed capable of fully capturingthe contours of race and gender inequality. Such suspicions gave rise to empirical assessments of different (i.e., smaller) units of analyses such as the characteristicsof the types of jobs that blacks and whites and men and women occupy. With this backdrop,the study of the "job"and specifically "job power" in the form of legitimate authorityemerged as an importantfocus in studies of race and gender stratification. Two broadresearchquestions would on race, gender,and authorityat work. Are therenet come to framethe literature Are there race and gender racial and gender differencesin authorityattainment? differencesin the amountof income received for occupying similar positions of authorityat work? When consideringthe sheer numberof studies attemptingto answer these two questions, it is clear that the study of ascriptivedifferences is the single most importantresearchfocus among authorityresearchersin the last two decades. Before summarizingthe main findings from this literature,a brief

518

SMITH discussionof selective theoriesof race andgenderdifferencesin authority is worth reviewing.

EXPLANATIONS FOR RACE AND GENDER DIFFERENCES IN JOB AUTHORITY


for theprocessesthatproduceandsustainraceandgenderdifferences Explanations in the determinants andconsequencesof job authority may be loosely categorized as eithersupply-sideor demand-sideexplanationsaimedat the level of individuals level theories), or jobs, (micro-level theories), society at large (macro-structural and industries (meso-level theories). organizations,occupations,

Micro Theories (IndividualLevel)


Early attemptsto explain race and gender differences in job authorityrelied on theories popularin occupationalsegregationand wage discriminationliterature (Hill 1980). Humancapital and status attainment explanations,drawnfrom neoclassical economics (Becker 1964, Thurow1969) and sociology (Blau & Duncan & Hauser1978), respectively,provided 1967, Sewell & Hauser1972, Featherman a set of assumptionsthatpointedto the behaviorsandcharacteristics of individuals in the form of investmentsin the attributes that were thoughtto lead to authority. Accordingto this view, women and minoritiesmay have less authoritythanwhite men because they have lower investmentsin factors such as training,education, and experience,or because they have less seniorityor intermittent laborforce attachment.Such factors may increase the likelihood of attainingauthoritywhile simultaneouslyserving as forces thatlegitimize the authoritystructure (Wright& Perrone1977). Moreover,both human capital and status attainmentapproachesassume that strategic decisions or aspirationsmay drive the career choices of individuals. Within the context of authorityresearch,these approachessuggest that women more so than men may opt out of contentionfor positions of authoritybecause they place less value on workplaceauthority(see England1992, p. 19 andReskin & Padavic 1994, pp. 77-78 for a discussion), or because women are more likely then men to assumefamily responsibilities(Wolf & Fligstein's 1979a, D'Amico's 1986, Jaffee 1989, Wrightet al. 1995, Hopcraft1996, Baxter 1997). Phrasessuch as compensatingdifferentials(Filer 1985, Jacobs & Steinberg1990) and mommy track(Ehrlich 1989) denote the idea that women'spreferences,either because of gender-rolesocialization(Reskin & Padavic 1994, p. 41) or rationalchoices, may promptthem to self-select themselves out of contentionfor positions of authority due to family responsibilities(Baron 1987, Wadman1992). However,despite its potentialimportance,the associationbetween family statusandjob authorityhas not been fully explored. Even though race and gender differences in human capital attributesdo not fully explain race and gender gaps in authority,studies show that investments in human capital significantly increase the odds of gaining access to authority

AUTHORITY ANDWORKPLACE GENDER, RACE,

519

(Kluegel 1978, Halaby 1979, Hill 1980, Jacobs 1992, Smith 1997, Wilson 1997b, Ross & Reskin 1992). However,the presence and magnitudeof such effects may depend on how authorityis measured.Some researchshows that when authority is measuredaccordingto the numberof subordinates supervised(spanof control), there is no effect of education on the authorityattainmentof blacks and whites authoritymeasuresshow thatunit increases (Muelleret al. 1989), buthierarchical in educationfacilitatethe movementof both black and white employees into both lower and higher levels of authority(Smith 1997, Wilson 1997b). In race and gender studies, most attentiongiven to the hypothesized effects of human capital variables on authorityassesses whether or not minorities and women arerequiredto have higher levels of humancapitalinvestmentsthantheir to reach similar levels of authority.Clear evidence supwhite male counterparts effects of humancapitalvariableson authorityare more The this contention. ports blacks thanwhites (Kluegel 1978, Muelleret al. 1989, Wilson pronouncedamong 1997b, Smith 2001). Womenin Sweden have to have more work experiencethan men to receive similarauthorityrewards(Hutlin 1996, 1998, p. 107), and women thanmen with similarschooling in the United States achievedmuch less authority andexperience(Wolf & Fligstein 1979a,b,Halaby 1979, McGuire& Reskin 1993, p. 494). for usheringworkhumancapitalinvestmentsarevitally important In summary, ers into positions of authority, regardlessof race or gender.The relativeeffects of while generallypositive, appearto vary dependingon such variableson authority, is measured.Moreover,the ratesof returnin authorthe mannerin which authority ity to humancapitalinvestmentsvarysignificantlyby race andgender,with blacks and women receiving less authoritythanwhites and men, respectively,for similar levels of humancapital investments.But as importantas humancapitalvariables their effects often are (especially at the entry ports of the authoritydistribution), variablesaretaken or aresignificantlydiminishedonce structural-related disappear into account(Kluegel 1978, Smith 1997).

MacroStructuralTheories (SocietalLevel)
of authority and Structural explanationsfor ascriptiveinequalityin the distribution an to individualto authority in earningsreturns counterweight represent important researcherssurmised level attainment models. Fromthe very beginning,authority of humancapitalendowments,minorities that apartfrom the unequaldistribution and women had differentialaccess to positions of authorityin the workplacebeof located in the most marginalizedstructures cause they were disproportionately indicatorstypically used to explain ascriptive the economy.3The macrostructural differences in authorityare region (Muelleret al. 1989, Smith 1997, 1999), city
3It is importantto note that both race and sex may be viewed as a dimension of social structure(Blau 1977). At the same time, positions within and between authorityhierarchies constitute the structuralentities within which race and sex groups are unequally distributed-ostensibly, as discussed below, because of the statusassociatedwith minority and female identity.

520

SMITH size (Smith 1999), industrialsector (Mueller et al. 1989), and employment sector (Wilson 1997b). Other structural predictorsof authorityhave been classified to union versus non-union status (Kluegel 1978, Mueller et al. 1989, according Smith2001), andmorerecently,race/ethnicconcentration of workersoperationalized at the level of the workplace,local industrialsectors, and local occupational groupings(Elliott& Smith2001, Smith& Elliott,forthcoming).Along these lines, minorities and women are more likely to have authorityin economic structures where they exist in large concentrationslike the South, the public sector, within coethnicenclaves,andin industriesor locales moresusceptibleto economic downturns(Muelleret al. 1989, McGuire& Reskin 1993, Wilson 1997b, Smith 1999, Smith & Elliott forthcoming). Additionally,it standsto reason that neo-institutionaldynamics influence the attainment of womenandminorities.Factorssuchas the outsideinfluence authority of periodicorganizational review,government policy, andboth the size and age of an organization may makeit moreor less amenableto the inclusionof women and minoritiesin its managerialand authority ranks(Baronet al. 1991, Marsdenet al. & Skaggs 1999). Both genderand 1996, Huffman1995, 1999, Tomaskovic-Devey racial equality at work appearto be more pronouncedin large organizations,in the public sector, in newer ratherthan older organizations,in organizationssubject to periodicreview, duringperiods of EEO enforcement,and in organizations thathave formalizedpersonnelpolicy (Baronet al. 1991, Wilson 1997b, Huffman 1999, 1995). However, save for a few studies that control for public sector emresearchhas not specificallytakeninto ploymentandestablishmentsize, authority accountthe directeffects of governmental outcomes.Instead, policies on authority areoften treatedas proxies for public sectoremploymentand largeestablishments antidiscrimination policies becausepublicsectoremployersandemployersof large organizationsare more susceptiblethanprivatesector and small organizationsto governmentinfluence. By and large, neither the separatenor the additive effects of structuraland human capital factors explain race and gender differences in authorityor ascriptive disparities in the economic rewardsthat flow from authority.For this reason, researchershave considered the role of employer discriminationin the authorityattainmentprocess. Two employer-sideexplanationsof race and gender discriminationin authorityhave dominatedthe literaturein recent years, including social closure/segregationand homosocial reproduction.These factors are loosely described below as meso-level theories of discriminationbecause laddersof majority-group positionedattheentryportsandpromotional gatekeepers or establishmentsare typically chargedwith the responsibility jobs/organizations of makingthe kindsof decisionsthatoften lead to the exclusion of some minorities and women.

Meso-Level Theories of Discrimination


arerootedin the idea that SOCIAL CLOSURE Meso-level theoriesof discrimination majoritygroupmemberswho occupy positions of authorityat workhave a vested

ANDWORKPLACE AUTHORITY RACE, GENDER,

521

interestin maintainingtheirhegemonyover such positions anddo so by excluding candidateswho differ from them in racial and gender identity.Exclusionarytheories allow for both conscious and not-so-conscious acts of discrimination.Max Weber's(1968) idea of "social closure,"its recent applicationto analyses of race and sex job/occupationalsegregation(Tomaskovic-Devey1993a,b,Reskin 1993), the concept of "statisticaldiscrimination," and class conflict (Dahrendorf1959, & Perrone 1977) Wright approaches,may each be classified as a conscious act of exclusion on the part of authorityelites. On the societal level, social closure underscoresthe idea thatpolitical and social elites preservepower and privileges for mobility to themselves or similarothers. Social cloby limiting opportunities sure may assume two additionalprocesses of exclusion: Minorities and women are likely to be segregatedinto the kinds of jobs, work settings, and industries that do not confer authority,but even when such factors are taken into account, to exercise authority they are still less likely than their white male counterparts (Tomaskovic-Devey1991, 1993, Reskin & Padavic 1994, p. 96, Elliott & Smith denotes the idea that gatekeep2001). In a similarvein, statisticaldiscrimination ers use race and gender as a proxy for likely productivitywhen they make hiring and promotiondecisions-especially when other personalinformationabout the candidateis lacking.
HOMOSOCIALREPRODUCTION The general lack of personal information and the

infrequentopportunitiesto build trust and mentorshiprelationshipsbetween auis also the basis of Rosabeth thorityelites and female and minoritysubordinates Moss Kanter'sidea of "homosocialreproduction" (see Reskin et al. 1999 for a unlike social the motivations behind homosocial reclosure, summary).Except, be bit a less conscious and overt. productionmay Although primarilyaimed at differences in Kanter's(1977) idea of attainment, explaining gender managerial homosocialreproduction has garnered a significantamountof attentionamongresearchersstudyingraceandgenderdifferencesin job authority, status, promotional andpromotionalaspirations(Kluegel 1978, Muelleret al. 1989, Baldi & McBrier 1997, Pfeffer & Davis-Blake 1987, Wilson 1997b, Smith 1999, Cassirer& Reskin involvedin makingdecisions to 2000). Kanterpointedto the inherentuncertainty into higher managerialranks.Such uncertainty promotesubordinates encourages authorityelites to develop managementenclaves composed of individuals who share a common set of social and demographiccharacteristics. the Importantly, chain of command,the more unstructured, nonrouhigher up the organizational In this context, shared tine, and subjectivearethe criteriafor authorityattainment. understanding, solidarity,commitment,and trustare betterfacilitatedin settings where gender homogeneity exists (Elliott & Smith 2001, p. 260). Beyond the have a tendencyto promotesubordiimplicationthatmen in positions of authority natemen,researchers have arguedthatthis idea partiallyexplainsracialdifferences in promotionrates(Baldi & McBrier1997) andauthority attainment (Muelleret al. 1989, Reskin et al. 1999, Smith & Elliott forthcoming,Elliott & Smith 2001). The effects of homosocialreproduction can apparently be mitigatedas the proincreases.Accordingto Kanter, portionof womenandminoritiesin anorganization

522

SMITH increasesso does the likelihoodof availablementors,political alas the proportion attainment lies, androle models-factors thatincreasethe likelihoodof managerial financialrewards.Dubbedthe strength-in-numbers and commensurate hypothesis (Jacobs 1992), this idea runs counterto predictionsdrawnfrom Blalock's (1967) analysisof racialsegregation.Accordingto Blalock, increasingnumbersof blacks heightenwhite resistance.Thus, withinthe conmoving into white neighborhoods text of organizational hypothesissuggests theory,the so-calledresistance-to-threat that as the proportionof women and minoritiesincreases in an organization,so does white, male resistance(Pfeffer & Davis-Blake 1987, Jacobs 1992). Adjudicating between the two hypotheseshas been difficultdue to the lack of firm-level analyses.

Bottom-Up Ascription
Recent research on race- and sex-based descriptionin the authorityattainment process stressesthe tendencyon the partof authorityelites to matchsubordinates on the basis of raceandsex (Tsui & O'Reilly 1989, Elliott & Smith andsupervisors and social closure denote top-downexhomosocial reproduction Whereas 2001). of on the elites, bottom-upascriptionhas been part authority clusionaryprocesses hierdescribedas an attenuating strategywhereby,at the bottomof organizational archies,employersactivelyseek to matchsubordinate/superordinate groupson the on basis of race andethnicityas a meansof reducingperceptionsof discrimination the partof subordinate racial/ethnicminorities(Elliott & Smith 2001). According to Elliott& Smith(2001, p. 261), "bottom-up pressuresforethnicmatching,regardto less of the source,push againsttop-downpressuresof homosocialreproduction influencewhen andwheremembersof particular minoritygroupsarelikely to gain at all, To the extentthatminoritieshave authority access to positionsof authority."4 at least one of threepatternsis nearly always present:Racial minoritiesare more racial minorities likely to have authorityover otherracial minorities;subordinate are more likely thanmajoritygroupmembersto have minoritysuperiors;and minoritiesare more likely thanmajoritygroupmembersto exercise authorityat the bottom of organizationalhierarchies-where, not surprisingly,racial minorities tend to be disproportionately represented.Not only is more researchneeded in this areato more precisely delineatethe mechanismsthatpropelthese processes, of the consequencesof such patternsfor but we also need a deeperunderstanding minorityperceptionsof workplacediscrimination.

groupsand theirchampiascriptionpartiallydenotes the idea that subordinate 4Bottom-up ons are not completely powerless and passive agents in the authorityattainmentprocess. groups(ortheiragents)pressingfor change Endogenousinfluenceon the partof subordinate fromwithinorganizations (i.e., diversitycouncils/teams)andexogenous forces operatingat action andEEO laws) may serve to mitigate local, state, andfederallevels (i.e., affirmative on the partof authorityelites. what would otherwisebe complete ascriptivereproduction More targetedresearchin this areais needed.

AUTHORITY RACE, GENDER,AND WORKPLACE

523

Demography Relational/Organizational
The racial and gender composition of organizationsand the impact that such composition has on organizationaloutcomes is an importantarea of investigation among sociologists (Baron & Bielby 1980, Pfeffer 1983).5 This approach in relationto othersinfluenceinterarguesthatone's demographiccharacteristics action between co-workers and supervisorsin a mannerthat mediates a variety outcomes at work (Tsui & O'Reilly of individual-leveland organizational-level Similarities 1989, Tsui et al. 1992, Tsui & Gutek 1999)-including job authority. or differences in the race/ethnicand gender characteristicsof one's co-workers and superiorsmay either enhance or decrease one's workplace experiences. A literature demography providessome justificationfor findingin the organizational why white men might want to exclude minorities and women from positions of authority.Studies show that heterogeneousgroup interactionincreases negative workplace experiences and decreased mobility chances for white men (Tsui & O'Reilly 1989, Tsui et al. 1992, Smith 2001), but white men benefit most from working in coethnic occupationalstructures(Smith & Elliott, forthcoming).In this context, the inclusionary/exclusionary propertiesof homosocial reproduction and less a matterof benign or may be seen as more a matterof self-preservation proximatediscrimination. The next section presents a review of the causes and economic consequences of race and genderdifferencesin job authority. Majorfindingsarepresentedseparatelyfor race andgendergroupsfor two non-ideologicalreasons.First,save for a few exceptions,raceandgenderstudiesof authority havedevelopedindependently of each other.Second, while there is some overlap,the factors that facilitate and constrainthe authorityexperiences of women relative to men are largely different from those that structurethe authorityoutcomes of racial minoritiesrelative to whites (Paracel& Mueller 1983, p. 158). I begin with a review of the major literature followed by a review of the research findingsfromthe race andauthority on genderand authority.

THE CAUSES AND CONSEQUENCES OF RACEDIFFERENCES IN AUTHORITY


Amid the backdropof the civil rights movement,studies documentingthe causes and economic consequences of racial differences in socioeconomic status became popular among status attainmentand social class researchers(Wright & Perrone 1977, Featherman& Hauser 1978, Kluegel 1978, Wright 1979, Hout 1984). Status attainmentand human capital approachesto understandingpatterns of socioeconomic progress noted that the early to mid-1970s ushered in a
5See Reskin et al. (1999) for a recentreview of this literature. Here the theoryis presented as a precursor to a discussion of relevantfindingsin the authorityliterature.

524

SMITH black mobility.The racialgap betweenblacks and whites periodof unprecedented indicatorsof attainment (education,occupationalstatus, alongkey individual-level and income) significantlynarrowedthroughout the 1970s (Farley& Allen 1987). Despite this progress,significantracialdifferencesin socioeconomic achievement were interestedin findingout why. In this remainedfirmlyintact,and researchers and consequencesof job context, the studyof race differencesin the determinants authorityfor income emerged as an importantarea of sociological investigation. Severalimportant researchquestionsframethe literature on raceandjob authority: (a) Whatis the extentof bothracial(Kluegel 1978, Parcel& Mueller1983, Mueller et al. 1989, G. Wilson 1997a, Smith 2001) and ethnic differences in authority (Englandet al. 1999, Smith2001, Elliott& Smith2001, Smith& Elliottforthcomdifferforblacksandwhites (Kluegel ing); (b) do theprocessesthatlead to authority 1978, Parcel& Mueller 1983, Muelleret al. 1989, G. Wilson 1997a, Smith2001); (c) what is the financial cost of racial exclusion from authority(Kluegel 1978, Parcel& Mueller 1983, Smith 1997, Wilson 1997b); and (d) to what extent does the racialdemography of the workplaceinfluencethe authorityattainment process of black, white, Asian, and Latinomen and women (Smith 2001, Elliott & Smith 2001, Smith & Elliott, forthcoming)?

Differencesin JobAuthority The Extentof Racial/Ethnic


to studiesof blacksand The raceandauthority literature has largelybeen restricted whites, perhapsbecause these groups were at the center of civil rights struggles extendedwhat duringthe 1960s and 1970s. The majorfindingsfromthis literature was knownaboutconsistentpatternsof racialdisparitiesalong key socioeconomic indicators.Regardlessof the data source or the measureof authorityunderconsideration,a clear and consistent finding may be gleaned from the literatureon race/ethnicityand authority. Namely, blacks areless likely thanwhites to exercise authorityin the workplace,and the authoritygap is not fully explained by race differencesin human capital investments,parentalbackground,or where blacks are located in the labor market(Kluegel 1978, Wright 1979, Parcel & Mueller 1983, Mueller et al. 1989, McGuire & Reskin 1993, Smith 1997, Smith 1999, Wilson 1997b). Moreover,the largestracialdisparityoccurs at higherlevels of occupationalstatuswherethe criteriafor promotionareoften vague relativeto lower occupationallevels (Kanter1977, Kluegel 1978, Smith 1999, Wilson 1997a). Because authoritymeasures vary from one study to another,a clear understanding of the overarchingmagnitudeof the racial/ethnicgap in job authority is difficult to ascertain.Studies based on continuousmeasuresof authoritymay of authority (Kluegel 1978), providean averageracialdifferencein the distribution the mean numberof subordinatesan individualsupervises (Muelleret al. 1989), of individualswho have say over the pay, promotion,hiring,and or the proportion or broad firing, supervisoryor managerialcontrol over others (Parcel& Mueller et Mueller al. 1989, Englandet al. 1999, Smith 2001). Still othermeasures, 1983, whethercontinuousor categorical,combine multipletypes of authorityindicators

AUTHORITY ANDWORKPLACE GENDER, RACE,

525

to form hierarchicaltrichotomies(high, low, no authority)(Wilson 1997a, Smith 1997, 1999, Smith & Elliott), or composite lineardichotomies(Parcel& Mueller 1983, McGuire& Reskin 1993). Regardlessof the way authorityis measured,one consistent finding seems clear: Minoritiesare less likely than whites to exercise of authorityare taken into authorityat work even when all known determinants account. Findings based on a Wisconsin sample show that the authorityrank of black men is abouthalf thatof white men (Kluegel 1978, p. 290), while nationalstudies report that white men's authorityscores range from 10% to 30% higher than those of blacks dependingon the authoritymeasureunderconsideration(Parcel & Mueller 1983, Mueller et al. 1989). Concerningthe more coveted positions of authority,such as the exercise of high authority(having say over the pay and promotionsof othersand say over hiringandfiring),white men arenearlytwice as likely as blacks andHispanicsto hold such positions (Smith2001, Smith & Elliott forthcoming),while Asian men appearto be as likely as white men to have such authority(Smith & Elliott forthcoming).Datafrom the 1993 wave of the National LongitudinalSurveyof Youthshow greaterparitybetween white men andLatino men when authorityis measuredas a dichotomousvariablebased on a whether or "supervisor" a respondent'scensus occupation included the title "manager" (Englandet al. 1999). Using this procedure,Englandand associates (1999) also then black showed thatboth white women andLatinawomen have more authority men, andthe gendergap in authority amongblacks was virtuallynon-existent(see Englandet al. 1999, Table4.1). also vary by time period.Analyses of Race differencesin authorityattainment repeatedcross sections of GeneralSocial Surveydata(1972-1994) and datafrom two waves of the Panel Study of Income Dynamics (1976 and 1985) show thatthe racial gap among men in accessing positions of high authority(Smith 1999) and to hire/fireandinfluencethe pay andpromotionsof subordinates (Wilson authority 1997b) has actuallyincreasedfrom the 1970s to mid-1990s-with the 1980s constitutingthe greatestera of inequalityamong the three decades in question. This patternis noteworthyin its own right, but it is particularlysignificantwhen one of black gains duringthe 1980s along otherkey considersthe relativedeterioration indicatorsof socioeconomic status including employment,earnings, managerial attainment,and broadoccupationalmobility (Jaynes & Williams 1989, Leonard 1990, Nkomo & Cox 1990, Burstein& Edwards1994, Bell & Nkomo 1994, Cancio et al. 1996, Bound & Freeman1992, Bound & Dresser 1999).6 No shortageof explanationsexists for the wideningracialgap in authority(and othersocioeconomic indicators)duringthe 1980s. In additionto majorshifts in the structure of the economy from a goods-producingto a service-oriented economy, the 1980s usheredin a period a patternthatled to increasedblack unemployment, action(Bobo & Smith 1994), a reduction of intensewhite oppositionto affirmative
6Thepace of occupationalsex integrationalso slowed duringthe 1980s when comparedto previousdecades (Reskin 1993; Adler 1993).

526

SMITH or complete eliminationof affirmative action practiceson the partof many firms (Kelly & Dobbin 1998), andnoticeablyweakerenforcementof equal employment laws (Leonard1990, pp. 58-59). While the associationbetweenthese opportunity factors and declines in black authorityattainmentis more than plausible, more research,in line with neo-institutional approaches(Tomaskovic-Devey& Skaggs influences 1999), is needed to drawa tighterlink between externalenvironmental and ascriptivedifferencesin authorityoutcomes.

RaceDifferencesin the Processesthat Leadto Authority


of authority determinants differfor blacks andwhites either Manyof the important because employersrewardthe credentialsof one groupmore thanthe other,or bechancesof one group cause the variablesthatimproveor detractfromthe authority have no observableimpacton the othergroup.In eithercase, when this occursin a framework,it providesstrongevidence that the races reach authority quantitative differentmeans.The findingthatthereareimportant differthroughdemonstrably ences betweenblacks andwhites in the processesthatusherworkersinto authority has not been contradictedsince it was firstobservedby Kluegel (1978). Race differencesin educationalattainment(Kluegel 1978) and in the amount but not dominant of authority returnsto some humancapitalfactorsare important Black women receive a lower returnthan sourcesof racialdisparitiesin authority. white women for similar investmentsin firm tenure (McGuire & Reskin 1993, factorsplay a moreimportant role thanhumancapitalfactors p. 495-96). Structural in differentiallydistributingthe races, but the combined effects of both human determinants still leave much of the racial gap in authority capital and structural Race differencesin the additiveeffects of humancapital unexplained(Smith 1999). of structural indicatorshelp to explainrace differencesin credentialsand a variety in whicheconomicstructures thedistribution of authority. Regardlessof themanner bottomof such structures, to be at the blacks and Latinos tend areoperationalized, which tends to decreasetheirchances of gaining authorityrelativeto whites. Assessments of whetherthe processes that lead to authoritydiffer for minorities and whites are often based on observationsof race/ethnicspecific regression results for minorities and whites. After establishing a basis for comparingcoefficients across race-specificmodels, differencesand similaritiesin the number of statisticallysignificant determinantsof authorityare observed. Since human indicatorssignificantlyimpact the authorityattainmentof capital and structural have concludedthatthe authorityattainminoritiesmore thanwhites, researchers ment process of minoritiesis more circumscribedand governedby a tighter set of rules than that of their white counterparts (Muelleret al. 1989, Wilson 1997a, Smith 2001).

The Consequencesof RaceDifferences in AuthorityAttainmentfor Income


The first studies to link job authorityto race and income sought to shed further light on the sources of black-white differences in socioeconomic status using

AUTHORITY RACE, GENDER,AND WORKPLACE

527

as an indicatorof earningsinequality(Wright positionswithin authoritystructures & Perrone 1977, Kluegel 1978). Wright'swork in this arearepresentsone of the firstexaminationsof race differencesregardingthe consequencesof authority(as class position) for group disparitiesin income (Wright& Perrone 1977; Wright 1978, 1979). An importantfinding in this researchrevealed no discernablerace differences in income returnsto education within class/authoritycategories-a thanraceas a factorthatdetermines findingthatsuggestedclass was moreimportant the life chances of blacks relativeto whites.7 wrinkleabsentin statusattainment studWright'sresearchaddedan important ies of race differencesin educationalreturnsto income. Accordingto Wright,the in their own right,operatethroughthe effects of race on income, while important (what positionsthatracialgroupsoccupy withinandbetweenworkplacestructures Because black men occuWrightrefersto as the "socialrelationsof production"). pied the lowest positions in the managerial hierarchy, they receivedcomparatively less economic payoff thantheirwhite male counterparts for becoming a manager (Wright& Perrone1977, p. 52). The dominanceof class effects over racialeffects on income in Wright'sresearchdid not mean thatrace was not an important factor that determinedthe life chance of blacks. Wrightcautionedagainstthis interpretationnoting thatit would be erroneousto assumethat"allracialdiscrimination is (1979, p. 197).8Instead,he speculatedthatracial reallydisguisedclass oppression" discrimination may in fact takeplace duringthe prelabormarketand promotional stages of the employmentprocess. In fact, we now know that race differencesin prelabormarketfactorsdo not accountfor much of the net disparitiesin authority or income. Instead,as much as we can discern,racialdisparitiesin income accrue because blacks receive comparativelylower returnsthan whites to their human capital investments,even when they occupy similar levels of authorityand are located in the same industries(Kluegel 1978, McGuire& Reskin 1993). McGuire & Reskin found thata full 55% of the earningsgap between white men and black men, and 62% of the earningsgap between black women and white men was accounted for by the comparativelylower returnsthatblacks receive for occupying similarauthority levels andindustrial locationsas whites andhavingsimilarlevels of humancapital(p. 499). However,databased on young cohortsof black, white, and Latino men and women suggest that authorityonly explains about2% of the pay differencesbetween ethnic and gender groups,but the authorsspeculatethat this findingmay be attributed to the way they measuredauthority(Englandet al. 1999, p. 166, fn 16). In addition,evidence suggests thatracial discrimination is more likely to take place at higher levels of the authoritystructure(Kluegel 1978). Kluegel's work showedthatblackmen receive a lower income return to authority thando whiteswith the income disparitymore pronouncedat higher authoritylevels. Kluegel
7As discussed below, Wrightalso examinedwhetherthere were interactionsbetween class andsex. Explicittestingof the significanceof raceand sex wouldlateremergeas a centralfocus of severalauthoritystudies (Reskin & Ross 1992; Wilson 1997a,b;Smith 1997, 1999). 8See Wright(1979, pp. 197-207) for a detaileddiscussion of racism and class domination.

528

SMITH calculated the cost of black men's exclusion from positions of authorityto be aboutone thirdof the total black/whiteincome gap. Longitudinaldata from the Panel Study of Income Dynamics (1975, 1976) furtherunderscoretwo previousfindingsbased on cross-sectionaldata:Race (and sex) differencesin earningsarepartlydue to differentialaccess to span of control effect and spanof responsibility;andclass and/orauthoritywill have a substantial characteristics on earningsnet of humancapitalandlabormarket (Parcel& Mueller 1983). Accordingto Parcel& Mueller,racialdisparitiesin earningsare a function of the differentlabormarketpositions blacks and whites occupy in additionto the fact thatblacks receive unequalrates of returnfor occupying similarpositions of authority. Studies of change over time in the authorityattainmentprocess and income returnsto authorityreveal that the racial gap in authorityand in the amountof income returnsto authorityor managerialpositions has either increased(Wilson 1997a,b)or remainedconstantover time (Jacobs 1992, p. 293; Smith 1997, 1999). As noted earlier,these patternsare associated with changes in the structureof the economy, increases in black unemployment,and growing opposition toward antidiscrimination legislation and social policies designed to improve economic life chances.

RacialDemographyand the AuthorityAttainmentProcess


Although researchershave long held that the racial and gender composition of outcomes(Baron& Bielby 1980, Pfeffer 1983, affectsorganizational organizations Pfeffer & Davis-Blake 1987), authorityanalystshave only recentlybegun to deal with this issue acrossrace, ethnic, andgendercategories.One studyhas examined whetherthe ethnic concentrationsof establishments,occupations,and industries influencethe authorityattainmentof white, black, Hispanic, and Asian men and women (Elliott & Smith 2001). Using data from the Multi City Survey of Urban Inequalityand 1990 decennial census occupationalcodes, Elliott & Smith 2001 find that racial/ethnicconcentrationamong roughly equivalentcoworkersat the level of work groups, industries,and occupationalsectors has little effect on the chances of minorities accessing positions of authority.However, concentration work groups in the form of racial/ethnicmatchingof supervisorsto subordinate exerteda strong and consistenteffect among all groups, implying that, for some to supervise dependsa greatdeal on the opportunity groups,authorityattainment largely coethnic work groups (cf. Mueller et al. 1989). Elsewhere, an extended analysis suggested that only white men benefit from homogenous occupational niche employment(Smith & Elliott forthcoming),but unlike other groups,white men's authorityattainmentis decreasedif they have a female supervisor(Smith 2001). racialdifferences literature documentsimportant To sum, the raceandauthority and whites. two most studied-blacks in the authorityattainmentof the groups in the to conclusions processes practices Major point systematic discriminatory

ANDWORKPLACE AUTHORITY RACE, GENDER,

529

that lead to authorityand in the amount of financialreturnsthat blacks receive for occupyingpositions of authoritysimilarto those of whites. Indeed,large-scale quantitativeassessments based on secondary survey data provide only a partial glimpse of the mechanismsthatgenerateand sustainracialinequalityin authority outcomes. We know very little aboutthe authorityof racialand ethnic groupsbeyond blacks andwhites. The few investigationsthatoffer analysesthatincludethe authority experiencesof LatinoandLatinaAmericans(Englandet al. 1999, Smith 2001) and of Latino and Asian Americans(Smith & Elliott forthcoming)reveal a with white andAsians on top andwith blacks and racial/ethnicauthority hierarchy Latinoson the bottom.Further clarificationof this hierarchyis needed along with extensive delineationof the economic consequences that it produces.Moreover, very little is known about trendsin the absolute and relative authorityoutcomes of racial/ethnicgroups.What little we do know is cause for some alarmbecause, of important antidiscrimination despite political pressureand the implementation legislation over the last 20 years, the racial gap in authorityand in the amount of income returnsto authorityhas either increased(Wilson 1997b) or remained constantover time (Smith 1997, 1999). Indirecttests of discrimination findconsistent supportfor the theorythatracial/ethnicdifferencesin the processes that lead to authorityare characterized by closer scrutinyof the formal labor marketcredentials of black men (Muelleret al. 1989, Wilson 1997, Smith 2001) and Latino men (Smith2001) comparedwith the scrutinyof theirwhite counterparts. Finally, a new line of authorityresearchthat emphasizes the impact of workplacediversity on authorityoutcomes shows thatonly white men benefit from homogeneous occupationalniche employment(Smith & Elliott forthcoming),but unlike other men's authority attainment is lessened if theyhave a female supervigroups,-white sor (Smith2001). Finally,employersappearto matchsupervisorsandsubordinates by race and ethnicity,which leads to lower perceptionsof discrimination(Elliott & Smith 2001). While an important datacoupled advance,additionalquantitative with qualitativeapproachesare needed to tap into the direct actions, sentiments, and motives of employersand workers.

THE CAUSES AND CONSEQUENCES OF GENDERDIFFERENCES AUTHORITY


The fact of important genderdifferencesin the processes thatlead to job authority and in the proportions of men and women who occupy positions of authorityhave been well documented(Wolf & Fligstein 1979a,b, Hill & Morgan 1979, Halaby 1979, Spaeth 1985, Jacobs 1992, McGuire & Reskin 1993). As with studies of race andauthority, genderandauthoritystudiesoffereda new way of assessing the role of ascriptionin determiningthe life chances of individualsand groups. The attainment and findingthatmen andwomenhadsimilarmeanlevels of occupational ascendedthe occupationalhierarchy throughsimilarprocesses (Treiman& Terrell & Hauser1976) causedmanyto questionwhetheroccupational 1975, Featherman

530

SMITH status measures of attainmentwere precise enough to capturethe full extent of gender inequality(Wolf & Fligstein 1979a,b, Hill 1980, D'Amico 1986). In this of jobs, and in particular the authorityassociatedwith context, the characteristics A jobs, became an importantunit of analysis in studies of gender stratification. literature revealsthatthe study summaryof findingsfromthe genderandauthority of social stratification has been advancedby authoritystudies in two important andin theprocessesthatlead ways. First,genderdifferencesin workplaceauthority, to authority,constitutean importantsource of gender inequalitythat is obscured when traditionalindicatorsof occupationalstatusare used to measureinequality. Second, earningsmodels that do not include measuresof job authorityare apt to the gendergap in earnings. significantlyunderestimate

GenderDifferencesin Authorityand Its Causes


Heraldedas a path-breaking contribution to the study of social stratification (Hill 1980, p. 110), Wolf & Fligstein (1979b) were among the first to examine the mechanismsthatgenerategenderdifferencesin authority. Their study,based on a sampleof men andwomenfromWisconsin,showedthatwomenhaveless authority then men (even net of humancapital, occupationalstatus, and self-employment status),and the processes thatlead to authoritydiffer for men and women in two respects:(a) Women receive lower authorityreturnsthan men for similarhuman capitalandoccupationalinvestments,and(b) the processesthatlead to supervisory authorityare more egalitarianthanthe processes that usher men and women into positionsthatgrantthemthe abilityto hire andfireandinfluencethe pay of others. was more The latterfindingconfirmedexpectationsthatemployerdiscrimination at higherthanat lower levels of the authority While Wolf & hierarchy. pronounced Fligstein's findings offered a new way of conceptualizinggender inequality,the conclusion that the behaviorsand policies of employers are most responsiblefor restrictingthe authoritychances of women was met with criticism (Bridges & Miller 1979). The chief critics, Bridges & Miller, agreed with the basic finding that women have less authoritythan men, but questioned the mannerin which Wolf & Fligstein arrivedat that result because their data were restrictedto high school graduates,age 35 (mid-life), who residedin Wisconsin.9 In a replicationusing nationaldata from the 1976 Panel Study of Income Dynamics, Bridges & Miller avoided those restrictionsand, in contrastto Wolf & Fligstein, found, among otherthings, thathaving childrenbelow age sixteen sigchancesof women. This resultprompted Bridges& nificantlylimitedthe authority of the authorityexperiMiller to offer a different(albeit untested)interpretation ences of women whereby the behaviorof women themselves, the desires of coare placed on parwith workers,and even prelabormarketforms of discrimination the actions of employers as plausible explanationsfor women's lower authority attainment.Even though authorityresearchershave embraced one explanation
9See Fligstein et al.'s (1981) responseto Bridges & Miller.

ANDWORKPLACE AUTHORITY RACE, GENDER,

531

or another,empiricallychoosing between these alternative explanationshas been more art than science. As noted earlier,absent direct informationfrom workers about their motivationsand direct informationabout what drives employer decisions, authorityresearchershave complied with what has become fashionablein all inferentialstatistics-that is, they have assumed facts that, while extremely plausible,provideincompleteevidence of discrimination.

GenderDifferencesin the Effectsof Human CapitalInvestmentson Authority


The humancapitalpredictionthatgenderdifferencesin authorityare a functionof differentialinvestmentsin education,workexperience,training,andhoursworked in authority research.By andlarge, individual per week is generallynot supported investmentsin humancapitalattributes chances of appearto enhancethe authority both men andwomen, butmen receive a muchhigherauthority return thanwomen for possessing similarlevels of humancapital.Forexample,educationhas a much strongereffect on the authoritychances of men than women (Wolf & Fligstein 1979a, Halaby 1979a, Hill 1980); each additionalyear of educationhas upwards of two to threetimes the effect for men as for women on authorityoutcomes (Hill & Morgan1979, p. 14, McGuire& Reskin 1993, p. 494). Importantly, the effects of humancapital variableson the authoritychances of women and men depend on whether observations are made at the lower or upper end of the authority At the low end, where an individualmay have the title of supervisorbut hierarchy. lack the ability to make decisions, educationand continuouswork experienceare for the authorityattainment more important of women thanmen (Jaffee 1989). At the high end, where supervisorshave decision-makingauthorityover the pay or promotionof workers,the effect of educationis strongerfor men (Hill & Morgan 1979). But accordingto Reskin & Ross's (1992, p. 356) studyof Illinois managers, there are no significantgender differences in human capital returnsto decisionmakingauthority. the role of family statusin structuring difBridges & Miller's findingregarding ferentialauthority outcomesbetweenmen andwomenmarkedan important precedent in authorityresearch.It stands to reason that the unequal division of labor within the householdmay preventwomen from seeking positions of authority. Innow popularized in phrasessuchas compensating deed,theneoclassicalarguments, differentials10(Filer 1985, Jacobs & Steinberg1990) and mommytrack(Ehrlich 1989), denotesthe idea thatwomen'spreferences,becauseof eithersocializationor
l?Inits most popularformulation,the idea of "compensatingdifferentials" was promoted by Filer (1985:427) to explain the sex gap in wages. Essentially,Filer arguedthat the sex gap in wages was a functionof the differential job choices women made relativeto menwith women opting for more pleasantandflexiblejobs, which supposedlypay less thanthe kinds of jobs valued by men, which, ceteris paribus,are less pleasantthen the jobs valued by women but bring with them higher wages and betterbenefits. Suffice it to say, there is little evidence to supportthis contention.

532

SMITH rationalchoices, may promptthem to self-select themselves out of contentionfor positions of authoritydue to family responsibilities(Baron 1987, Wadman1992). While there is some supportfor this proposition,the vast majorityof evidence arguesagainstit. On the one hand,researchbased on nationaland cross-national dataprovideslittle or no evidence thatwomen (relativeto men) chose not to seek authoritybecause of houseworkand/orfamily responsibilities(Wolf & Fligstein 1979a, D'Amico 1986, Jaffee 1989, Wright et al. 1995, Hopcraft 1996, Baxter 1997). On the otherhand,Wrightet al.'s (1995) comparative study of seven countries found some supportfor the self-selection hypothesisamong women living in Canada-but the other six countries(United States, United Kingdom, Australia, Sweden, Norway,Japan)showedno evidence of self-selection. D'Amico's (1986, p. 46) analysis of NationalLongitudinalData revealedthat women who received assistance at home with household chores were more likely to have say over the pay and promotionof others relative to women who did not, while other studies found that having childrenunderthe age of six has a negative impact on the earningsof female managersbut has no effect on male managers(Jacobs 1992, p. 296). Additionalresearchin this areais sorelyneededin orderto reconcilethese opposing findings and to sort out the causal orderingof the associationbetween In the meantime,it's worthemphasizingthatthere family statusandjob authority. is very little evidence thatwomen opt out of contentionfor authority positionsdue to family responsibilities-a findingconsistentwith genderstudiesof job segregation (Glass 1990) and wage differentials(Englandet al. 1988, Jacobs& Steinberg becausegenderdifferencesin such outcomesdependon whether 1990). Moreover, future observationsare made at the lower or upperend of the authorityhierarchy, assessments at various levels of authorityare required.In addition,family status measuresthatpoint to the merepresenceof childrenin the householdor the use of dichotomousmeasuresof maritalstatus (marriedor not) may obscure important genderdifferencesin the effects thatflow fromthe interactionof child andmarital status.

GenderDifferencesin the Effectsof Structural and Compositional Factorson Authority


of the economy, and The relativelocation of men and women within the structure within such structures,account for more of the their proportional representation of workers.With regard gendergap in authoritythanthe humancapitalattributes to structural effects, findings show that women are less likely than men to have locatedin the kinds of jobs (Halaby because they are disproportionately authority Roos 1981, Tomaskovic-Devey1993), occupations 1979a, Hill & Morgan 1979, (Robinson& Kelley 1979, Jaffee 1989,Reskin& Ross 1992), andeconomic sectors less likely thanthose of men to offer (Hill & Morgan1979) thatarecomparatively authority. Supportfor varioustheories concerningthe effects of the gender composition of economic units on men's and women's authorityoutcomes depends on two

AUTHORITY RACE, GENDER,AND WORKPLACE

533

underconsideration andwhetherdataaredrawnfrom factors,the level of authority nationalsamplesor samplesbased on state-leveldata.Forexample,findingsbased on analyses of national samples are more supportiveof Blalock's resistance-tothreathypothesis.A significantportionof the gendergap in authorityis explained by the presence of women in female-dominatedoccupations (Jaffee 1989)-an outcome thatvariesdependingon the level of authorityunderconsideration.That is, men in female-dominatedoccupationsappearto do betterthan women in the same occupationswhen it comes to attainingdecision-making authority.However, in the case of supervisoryauthority,bothmen andwomen areless likely to possess authorityas the female compositionof the occupationincreases.This means that at lower levels of authority,the femaleness of the occupationworks against the of both men and women (Jaffee 1989, p. 387). authorityattainment In contrast, there is sufficient evidence in support of Kanter's strength-innumbershypothesis. According to Reskin & Ross (1992), while men were far more likely than women to have final decision-makingauthorityat work, "the more female the census occupationin which managersworked,the more extensive their authority," (pp. 354-55). Furthersupportfor the strength-in-numbers hypothesiscomes fromin-depthstudiesof specializedjobs such as college administrators(Konrad& Pfeffer 1991, Kulis 1997), managerialand professionaljobs (Jolly et al. 1990), banktellers (Strober& Arnold 1987), andCaliforniastate-level establishments(Baronet al. 1991). In each case, the increasingpresenceof women led to increasesin the proportions of women in the organization. However,the relationshipdescribedhere is not necessarily a linear one throughoutall levels of economic units. Mountingevidence suggests thresholdeffects wherebythe presence of manywomen only enhancesthe inclusionof otherwomen at certainlevels of the organizational hierarchy(Cohen et al. 1998, Reskin & McBrier2000). Trendstudiesdocumentingchangeovertime in the net gendergapin managerial status are opposite from what has been learnedabouttrendsin the gender gap in supervisory positions(D'Amico 1986, Jacobs1992, Reskin& Padavic1994). With regardto managerialstatus, some studies show a clear narrowingof the gender has remainedconstantfor much gap, while the gendergap in supervisoryauthority of the 1970s and 1980s (Jacobs 1992, p. 295).

The Consequencesof GenderDifferences in Authorityfor Income


The exclusion of women from job authority,their restrictionto entry-level authority positions, and the comparativelylower income returnsthey receive for occupying levels of authoritysimilar to those of men contributeto their overall lower earnings.Assessing the role that authorityplays in explaining the gender gap in income is a trickypropositionbecause the magnitudeof genderdisparities is a function of whetherobservationsare made at the lower or higher end of the authoritydistribution.More generic measures of authority,such as supervisory status, only exhibit modest effects on the gender gap in income. In contrast,the

534

SMITH unequal distributionof men and women in authoritypositions that grant them control over monetaryresources and control over personneldoes producelarger income differencesbetween men and women (Spaeth 1979, Halaby 1979). Both Spaeth (1985) and McGuire & Reskin (1983) found that men receive twice the economic payoff thatwomen receive for possessing authoritythat allows them to controlmonetaryresourceseven when genderdifferencesin educationand experience are considered.In Spaeth's study, a one-unit increase along an authority scale representingcontrol over monetaryresourcesincreasedmen's net earnings by $383 (in 1981 dollars)comparedto $192 for women (p. 612). Notwithstanding the complexities, gender differencesin job authorityaccount for a large fraction of the pay gap among men and women with similaroccupations(Hill & Morgan 1979), jobs (Halaby 1979a), and equivalenthumancapitalinvestments(Parcel& Mueller 1983). Men and women who work in the same occupationsfor the same differencesaccountingfor employerreceive differentsalaries-with hierarchical 65% of the gap (Halaby 1979). Direct gender wage discriminationin returnsto human capital and hierarchicalrank ("unequalpay for equal work") and "rank segregation"in the form of restrictingwomen to low-paying jobs accounts for much of the gender gap in salaries, but the latter appearsto be more important (Halaby 1979a, Roos 1980). While the exercise of authorityat work enhances the earningsof most workers,one study shows that the earningsof white female heads of householdsare not increasedif they exercise authorityat work (Parcel& Mueller 1983, p. 195). To sum, men are more likely than women to have authority,and employer than women's attitudes behaviorsand organizational policies are more important Educationandjob tenure andbehaviorsin explainingthe gendergap in authority. of men thanwomen-especially attainment exerta strongereffect on the authority at high levels of authority. Family ties improvemen's, but not women's, chances and to the extent thatwomen occupy managerialpositions, they to gain authority, tend to be located at the bottomof the commandchain-largely supervisingother women and receiving lower earnings than men who occupy similar positions. In fact, gender differences in authorityattainmentaccount for much of the pay differences between men and women at high levels of authority(Halaby 1979, Hill 1980, McGuire& Reskin 1993) but have a modest effect at lower levels (i.e., supervisoryauthority)(Jacobs 1992, p. 296).

CONCLUSION
Job authorityis an importantdimension of socioeconomic status that remains a coveted workplaceresource.Its emergencemay be tracedto the foundingfathers as class categoriesor scalargraof modem sociology-with its operationalization the theoreticallineages of MarxandWeber,respectively.The dationsdemarcating patternof authority'semergenceas an importantindicatorof socioeconomic status marks a social scientific sequence of developmentfrom theory building to

AUTHORITY RACE, GENDER,AND WORKPLACE

535

measurementto hypothesistesting. At least threeoverlappinggenerationscan be discerned:first generation-theory builders;second generation-operationalization and measurement issues; and thirdgeneration-hypothesis testing, including assessments of the relative effects of ascriptionversus achievementand human determinants of authoritybased on samples stratified capitalversus the structural and race. Job is one of the few stratification variablesthe class, by gender, authority theoreticaland empiricallinkages of which remain strongregardlessof whether it is conceived as an indicatorof class or status,whetherit is operationalized as a discreteor gradational measure,or whetherit takes the form of a dependentor an independentvariable.Farfrom being the alphadimensionof social stratification, in all its manifestations as a measureof legitimatepower,job authority has proven to be quite a useful lens by which to observethe contoursof ascription-based work inequality.Throughsuch observations,the last 25 years of authorityresearchhas conclusions.First,raceandgenderdifferencesin authorproducedtwo overarching ity constituteone of the chief sources of race and genderinequalityin workplace outcomes. To paraphrase of Wrightet al. (1995, p. 407), the under-representation women, andI wouldadd,racialminorities,in positionsof authority, especiallyhigh is not simplyan instanceof gender(andracial)inequality;it levels of management, is probablya significantcause of inequality.Second, as arguablysociology's printo the study of earningsinequality(Halaby 1979), job authority cipal contribution is one of the primaryways in which the financialrewardsof work are distributed. Minorities and women receive a lower income returnthan do whites and men for occupying similarpositions of authority,and such disparitiesare more acute at high levels of authorityand among those who exercise control over monetary resourcesand controlover personnel. The debateover authorityas class or statusshapedthe subsequentoperationalizationandmeasurement for analysesof racialandgenderstratification of authority at work.The sheerquantityof researchon genderandauthority far exceeds the researchon race and authority, andtherehave been preciousfew attemptsto unravel the dual effects of race and gender on the authoritychances of minoritywomen (see McGuire& Reskin 1993, Englandet al. 1999). The study of ascriptivedifferences in authority,along with the processes that generatesuch differences,has been the primaryconcern of authorityresearchers over the last 20 years. In this line of research,two consistent findings transcend thanmen, authoritymeasures,data,and researchfoci: Womenhave less authority and minorities have less authoritythan whites. Moreover,when minorities and women do have authority, it is largelyat lower levels of authority andmainly when social, they supervisethe work of otherminoritiesand women. Despite important economic, political, and legislative achievementsin the last 20 years, these results remain robust and consistent in regional, national, and cross-nationalstudies at In fact,whatlittle we knowaboutchange single pointsin time andcross-temporally. over time in authority outcomesrevealsthatthe decadeof the 1980s represented an era of stagnationfor blacks and women-a patternthatis also evident along other statusindicators.

536

SMITH Finally, theories aimed at the micro, meso, and macro levels of analyses do This is partiallya not fully accountfor race and gender differencesin authority. reflectionof workplacediscriminationand the inherentlimitationsof measuring within a quantitative framework. discrimination

FUTURERESEARCH
researchportendnew and exciting areasof The collective limitationsof authority futureresearch.One promisingareaconcerns the mannerin which the organizaof groups. tional demographyof the workplaceimpacts the authorityattainment Not only does this researchrequirethe use of data based on samples of multiple racial/ethnicgroups of men and women, researcherswill also have to focus of the effects of employer on approachesthat seek to deepen our understanding on the authorityattainmentprocess. In this and employee attitudes/preferences case studies vein, Reskin (2000, p. 707-8) has called for in-depthorganizational Reacross studies. will allow in a manner that analyses comparative designed cent case studies by Bell & Nkomo (2001) and Thomas & Gabarro(1999) are particularlyilluminating,but they are limited to the work experiences of blacks and whites only. When consideringthe fact that the numberof Latinos will soon equal or eclipse thatof blacks, and firm-levelanalyses of Asian Americansare at a premium,broaderqualitativeanalyses are needed. The challenges posed by case studies analyses are often difficultto overcome. Despite promises of anonymity,employers remain very reluctantto release informationto researchersabout the racial or gender make-up of their authority can be used againstthemin a discriminastructure-especially if suchinformation leadersarereluctantto measure tion lawsuit.For the same reasons,organizational the effectivenessof theirdiversityprograms(Comer& Soliman 1996, p. 478-79). Additionalfutureresearchwould benefitfrom a longitudinallydesigned study. Very little is known about the mannerin which racial/ethnicand gender cohorts traversethe authorityattainment process over the course of theirwork lives. Most of what we know about change over time in authorityattainmentinvolves data from annual repeated cross-sections, which are importantbut not ideal for the study of change over time. It standsto reason that there are significantvariations in the authorityexperiences of differentracial, ethnic, and gender groups-and these patternsare likely to fluctuateover time. Within this context, the likely association between job authorityand change in the relative adherentsto neoinstitutionalfactors (i.e., state policies, organizationalage, the relative size of internalinterestgroups, and EEO enforcement)requiresmore detailed examinations (DiPrete & Grusky 1990, Baron et al. 1991, Tomaskovic-Devey& Skaggs 1999). Research on the dual impact of race and gender on the authoritychances of minority women is sorely needed, as is a clearer specificationof the impact of and the household division of laborin explainingthe gender gap family structure multilevel in authority. More sophisticatedapproachesmay considerquantitative,

ANDWORKPLACE AUTHORITY RACE, GENDER,

537

analysesat the individual,group,firm,andsocietal levels, coupledwith qualitative approachesthatutilize in-depthinterviewstappingemployer/employeeattitudes, preferences,expectations,and workplaceexperiences. cross-national studies researchcan benefitfromadditional In addition,authority a et al. Baxter & 1995, 1988, 1999) including compariWright Wright (Kalleberg son of Europeanand non-Europeancountries (Robinson 1984, Rosenfeld et al. & Rozgonyi 1986)-a precedentbegun nearly 20 years ago 1998, Tannenbaum but rarelyextended. researchthat Finally,thereare a varietyof paradigmsassociatedwith authority share common theoreticaland empirical insights. Extrapolatingfrom Haveman (2000), authoritystudies could be significantlyimprovedif researchersconsidered what has been learned from different paradigmsincluding: studies of organizationalecology, neo-institutionalresearch,resourcedependenceand social network research,human capital and social capital research,organizationaldeaimed at the level mography,and studies of social mobility and statusattainment in its own right, but of organizations.Each of these lines of researchis important all could be substantiallyimprovedthroughliberalcross-pollination. ACKNOWLEDGMENTS I am gratefulto LawrenceBobo, Barbara Reskin, andDorthySue Cobble for their this manuscript. on earlier versions of comments insightful The Annual Reviewof Sociologyis online at http://soc.annualreviews.org
LITERATURE CITED Adler M. 1993. Genderdifferencesin job autonomy: the consequencesof occupationalsex position.Sociol. Q. segregationandauthority 34:3:449-65 Baldi S, McBrier DB. 1997. Do the determinants of promotion differ for blacks and whites? Work.Occup. 24:478-97 Baron JN. 1987. Working partners: careercommittedmothersand theirhusbands.Bus. Horiz. Sept/Oct.,pp. 45-50 Baron JN, Bielby WT. 1980. Bringing the firms back in: stratification,segmentation, and the organizationof work. Am. Sociol. Rev.45:737-65 Baron JN, Mittman BS, Newman AE. 1991. Targets of opportunity:organizationaland of genderintegraenvironment determinants tion withinthe Californiacivil service, 19791985. Am.J. Sociol. 96:1362-401 Baxter J. 1997. Genderequality and participation in housework:a cross-nationalperspective. J. Compar. Fam. Stud.28:220-47 Baxter J, Wright EO. 1999. The glass ceiling hypothesis: a comparative study of the United States, Sweden and Australia. Unpubl.manuscr. Becker G. 1964. Human Capital. New York: ColumbiaUniv. Press Bell ELJ, Nkomo SM. 1994. Barriers to Work Place Advancement ExperiencedbyAfricanAmericans.Glass Ceiling Comm.Washington, DC: US Dep. Labor Bell ELJ, Nkomo SM. 2001. Our Separate Ways: Black and White Women and the Struggle for Professional Identity. Boston, MA: Harv.Bus. Sch. Press Blalock HM. 1967. Toward a Theoryof Minority GroupRelations. New York:Wiley

538

SMITH
efforts to manage diversity:Do they really work?J. Manage.Issues 8:470-83 D'Amico R. 1986. Authorityin the workplace: differencesamongmaturewomen.InMidlife at Work: a Fifteen YearPerspective, Women ed. LB Shaw,pp. 37-49. Lexington,MA: DC Heath DahrendorfR. 1959. Class and Class Conflict inIndustrialSociety.PaloAlto, CA: Stanford Univ. Press andtrend DipreteT, GruskyD. 1990. Structure in the process of stratification for American men and women.Am.J. Sociol. 96:107-43 DuncanOD. 1961. A socioeconomic index for all occupations.In Occupationsand Social Status,ed. A. Reiss, OD Duncan,CC North, p. 139. Glencoe, NY: Free Press Ehrlich E. 1989. The mommy track:juggling kids andcareersin corporate Americatakesa controversial turn.Bus. Week(Mar)20:12634 Elliott JR, Smith RA. 2001. Ethnic matching of supervisorsto subordinatework groups: findings on bottom-upascriptionand social closure. Soc. Prob. 48:258-76 Theories EnglandP. 1992. ComparableWorth: and Evidence.New York:Aldine de Gruyter England P, ChristopherK, Reid LL. 1999. Gender,race, ethnicity,and wages. In Latinas and AfricanAmerican Womenat Work: Race, Gender,and EconomicInequality,ed. I. Browne, pp. 139-182. New York:Russell Sage Found. FarleyR, Allen WR. 1987. The Color Line and the Quality of Life in America. New York: Russell Sage Found. FeathermanDF, Hauser H. 1976. Sexual inequalities and socioeconomic achievement in the U.S. 1962-1973. Am. Sociol. Rev. 41:462-83 Featherman DF, HauserH. 1978. Opportunity and Change.New York:Academic Press FilerR. 1985. Male-femalewage differencesin determiningoccupational structure.Indust. Lab.Relat. Rev. 38:426-37 Fligstein N, Sobel M, Wolf W. 1981. Response to Bridges and Miller. Am. Sociol. Rev.46:685-88

Blau PM. 1968. The hierarchyof authorityin Am. J. Sociol. 73:453-76 organizations. Blau PM. 1977. Inequalityand Heterogeneity: A PrimitiveTheoryof Social Structure. New York:Free Press Blau PM, DuncanOD. 1967. TheAmerican OcNew York:Wiley cupationalStructure. Bobo L, Smith RA. 1994. Anti-poverty,affirmative action, and racial attitudes.In Confronting Poverty:Prescriptionsfor Change, ed. S Danziger, G Sandefur, D Weinberg, Univ. pp. 365-95. Cambridge,MA: Harvard Press/ Russell Sage Found. Bound J, Dresser L. 1999. Losing ground:the erosionof relativeearningsof AfricanAmerican women during the 1980s. In Latinas andAfricanAmericanWomen at Work: Race, Gender, and Economic Inequality, ed. I. Browne, pp. 61-104. New York: Russell Sage Found. Bound J, Freeman RB. 1992. What went wrong?The erosion of relativeearningsand employmentamong young black men in the 1980s. Q. J. Econ. 107:201-32 Bridges WP,Miller B. 1979. Sex and authority in the workplace:a replicationand critique. Am. Sociol. Rev.46:677-83 andthe pheBriegerRL. 1995. Social structure nomenology of attainment.Annu. Rev. Sociol. 21:115-36 Burstein P, Edwards ME. 1994. The impact of employment discriminationlitigation on racialdisparityin earnings:evidence andunresolved issues. Law Soc. Rev. 28:79-111 Cancio S, Evans D, Maume D. 1996. Reconsidering the declining significance of race: racial differencesin early careerwages. Am. Sociol. Rev. 61:541-56 Cassirer N, ReskinB. 2000. Highhopes:organizational position, employmentexperiences, and women's and men's promotion aspirations. WorkOccup. 27:438-63 Cohen LE, BroschakJP, HavemanHA. 1998. And then there were more: the effect of organizational sex composition on the hiring andpromotionof managers. Am.Sociol. Rev. 63:711-27 Comer DR, Soliman CE. 1996. Organizational

AUTHORITY RACE, GENDER,AND WORKPLACE Fox W, Payne D, Priest T, Philliber W. 1977. Authority position, legitimacy of authority andacquiescenceto authority. Soc. structure, Forces 55:966-73 Glass J. 1990. The impactof occupationalsegregationon workingconditions.Soc. Forces 68(Mar):779-96 Halaby C. 1979. Job-specific sex differences in organizationalreward attainment:wage discriminationvs. rank discrimination.Soc. Forces 58:108-27 and workHalabyC. 1986. Workerattachment place authority.Am. Sociol. Rev. 51:63449 Halaby C. 1993. Reply to Wright.Am. Sociol. Rev. 58:35-36 Halaby C, Weakliem D. 1993. Ownership and authorityin the earningsfunction:nonnested tests of alternative specifications.Am. Sociol. Rev. 59:5-21 Haveman HA. 2000. The future of organizational sociology: forging ties among paradigms. Con. Soc. 29:476-86 Hazelrigg L. 1972. Class, property, and authority:Dahrendorf'scritiqueof Marx. Soc. Forces 50:473-87 Hill M. 1980. Authorityat work:How men and women differ.In TenYearsof thePSID, ed. G Duncan,J Cocoran.Ann Arbor:Univ. Mich. Press Hill M, MorganJ. 1979. Dimensions of occuFamilies, pation.In Five ThousandAmerican Vol. 7, ed. G Duncan,J Morgan.Ann Arbor, MI: Inst. Soc. Res., Univ. Mich. Hopcraft RL. 1996. The authorityattainment of women: competitivesectoreffects. Am. J. Econ. Sociol. 55:163-84 Hout M. 1984. Occupationalmobility of black men, 1962-1973. Am. Sociol. Rev. 49:30822 HuffmanML. 1995. Organizations, internallabor market policies, and gender inequality in workplacesupervisory Soc. Perauthority. spect. 38:381-97 Huffman ML. 1999. Who's in charge? Organizational influences on women's representation in managerialpositions. Soc. Sci. Q. 80:738-56

539

HutlinM. 1996. Genderdifferencesin authority attainment: the Swedish case. Unpubl.paper Hutlin M. 1998. Discriminationand the role of organizational leaders.Acta Sociol. 4:99113 Institutefor Social Research. 1969. Survey of workingconditions.Inst. Soc. Res., MI Jackman R. 1983. ClassAwareness M, Jackman in the United States. Berkeley: Univ. Calif. Press Jacobs J. 1992. Women's entry into management: trendsin earnings,authority,and values amongsalariedmanagers. Admin.Sci. Q. 37:282-301 JacobsJ, SteinbergR. 1990. Compensating differentials and the male-female wage gap. Soc. Forces 69:439-68 Jaffee D. 1989. Genderinequalityin the workplace autonomy and authority.Soc. Sci. Q. 70:375-90 Jaynes GD, Williams RM Jr. 1989. A Common Destiny: Blacks and AmericanSociety. DC: Natl. Acad. Press Washington, Grimm DL, JW,WozniakPR. 1990. PatJolly terns of sex desegregation in managerial andprofessionalspecialtyfields, 1950-1980. WorkOccup. 17:30-54 Kalleberg A. 1988. Comparativeperspectives on workstructures andinequality. Annu.Rev. Sociol. 14:203-25 KallebergA, GriffinL. 1980. Class, occupation and inequalityin job rewards.Am. J. Sociol. 85:731-68 KanterRM. 1977. Men and Women of the Corporation. New York:Basic Books beKelley E, Dobbin F 1998. How affirmative came diversity management:employer relaw, 1961 to sponse to anti-discrimination 1996. Am. Behav.Sci. 41(7):960-84 Kluegel J. 1978. The causes and cost of racial exclusionfromjob authority. Am.Sociol. Rev. 43:285-301 KonradA, Pfeffer J. 1991. Understanding the hiring of women and minorities.Soc. Educ. 64:141-57 Kulis S. 1997. Gendersegregationamong college and university employees. Soc. Educ. 70:151-73

540

SMITH
Annu.Rev. inequalityin work organizations. Sociol. 26:707-9 Reskin B, McBrier DB, Kmec JA. 1999. The determinants andconsequencesof workplace sex andrace composition.Annu.Rev.Sociol. 25:335-61 Reskin B, McBrierDB. 2000. Why not ascription? Organizations' employment of male and female managers.Am. Sociol. Rev. 65: 210-33 ReskinB, Ross C. 1992. Authorityandearnings thecontinuingsignificance amongmanagers: of sex. WorkOccup. 19:342-65 and Men at Reskin B, PadavicI. 1994. Women Work. ThousandOaks, CA: Pine Forge RobinsonRV. 1979. Ownership,authority,and occupationalprestige:a synthesisand crossnational study.PhD thesis. Yale Univ. Robinson RV. 1984. Reproducing class relations in industrial capitalism. Am. Sociol. Rev.49:182-96 Robinson RV, Kelley J. 1979. Class as conceived by Marx and Dahrendorf: effects on income inequality,class consciousness, and class conflict in the U.S. and GreatBritain. Am. Sociol. Rev.44:38-58 Roos P. 1981. Sex stratificationin the workplace: male-female differencesin economic returnsto occupation.Soc. Sci. Res. 10:19524 Rosenfeld RA, Van Buren ME, KallebergAL. 1998. Genderdifferencesin supervisoryauthority:variationamong advancedindustrialized democracies.Soc. Sci. Res. 27:23-49 Ross CE, Reskin BF 1992. Education,control at work, and job satisfaction.Soc. Sci. Res. 21:134-48 Sewell WH, HauserRM. 1972. Causesandconsequencesof highereducation:models of the statusattainment process.Am.J. Agric.Econ. 54:851-61 at SmithRA. 1997. Race, income, andauthority work:a cross-temporal analysisof black and white men, 1972-1994. Soc. Prob. 44:1937 SmithRA. 1999. Racialdifferencesin access to an analysis of change hierarchical authority: over time, 1972-1994. Sociol. Q. 40:367-96

KurzKW,MullerW.1987. Classmobilityin the industrialworld.Annu.Rev.Sociol. 13:41742 LeonardJS. 1990. The impact of affirmative actionregulationand equal employmentlaw on black employment. J. Econ. Perspect. 4(4):47-63 LopreatoJ. 1967. Authorityrelationsand class conflict.Soc. Forces 47:70-79 LopreatoJ. 1968. Class conflict and images of society.J. ConflictResol. 11:281-93 MarsdenPV, Cook CR, KnokeD. 1996. American organizationsand their environments: a descriptive review. In Organizations in America: Analyzing Their Structures and Human Resource Practices, ed. AL Kalleberg, D. Knoke, PV Marsden, JL Spaeth, pp. 45-66. ThousandOaks, CA: Sage Matras J. 1980. Comparativesocial mobility. Annu.Rev.Sociol. 6:401-31 McGuireG, Reskin B. 1993. Authorityhierarchies at work: the impact of race and sex. Gen. Soc. 7:487-506 MuellerCW,ParcelT. 1986. Ascription,dimensions of authorityand earnings:the case of supervisors.In Research in Social Stratification and Mobility, ed. RV. Robinson, pp. 199-22. Greenwich,CT:JAI Mueller CW, Parcel T, TanakaK. 1989. Particularismin authorityoutcomes:the case of supervisors.Soc. Sci. Res. 32:21-33 Nkomo SM, Cox T Jr. 1990. Factorsaffecting the upwardmobility of black managers in Rev.BlackPolit. privatesectororganizations. Econ. 18(3)39-57 Parcel T, Mueller C. 1983. Ascriptionand Labor Markets.New York:Academic Press In PfefferJ. 1983. Organizational demography. Researchin Organizational Behavior,ed. LL Cummings,BM Staw, pp. 299-359. Greenwich, CT:JAI Pfeffer J, Davis-Blake A. 1987. The effect of of women on salaries:the case of proportion Admin.Sci. Q. 32:1college administrators. 24 Reskin B. 1993. Sex segregationin the workplace. Annu.Rev.Sociol. 19:241-70 Reskin B. 2000. Getting it right: sex and race

AUTHORITY RACE, GENDER,AND WORKPLACE Smith RA. 2001. Particularism in control over monetaryresources at work: an analysis of racioethnicdifferencesin authority outcomes of black, white, and Latino men. Workand Occup. 28:447-68 Does ethnic SmithRA, ElliottJR.Forthcoming. concentration influenceemployees' access to authority?An examinationof race and genderin threemetroareas.Soc. Forces.In press Spaeth J. 1985. Job power and earnings.Am. Sociol. Rev. 50:603-17 ArnoldCL. 1987. The dynamicsof StroberMH, occupationalsegregation amongbanktellers. In Gender in the Workplace, ed. C. Brown, JA Pechman,pp. 107-48. Washington,DC: BrookingsInst. Tannenbaum AS, Rozgonyi T. 1986. Authority and reward in organizations: an international research.Surv.Res. Cent., Inst. for Soc. Res., Univ. Mich. Thomas DA, Gabarro JJ. 1999. Breaking Through:the Makingof MinorityExecutives in Corporate America.Boston, MA. Harvard Bus. Press ThurowL. 1970. Investment in HumanCapital. Belmont, CA: Wadsworth Tomaskovic-DeveyD. 1991. Labor process and the genderand race compositionofjobs. Presented to Res. Comm. 28th Int. Sociol. Ass., Columbus,OH Tomaskovic-DeveyD. 1993. Laborprocess inequalityandthe genderandracecomposition of jobs. Res. Soc. Strat.Mob. 12:215-47 Tomaskovic-DeveyD. 1993. Genderand Race Inequalityat Work:the Sources and Consequences of Job Segregation.ILR Press Tomaskovic-Devey D, Skaggs S. 1999. Degenderedjobs, organizational processes, and gender segregated employment. Res. Soc. Strat.Mob. 17:139-72 TreimanD, TerrellK. 1975. Sex and the process of status attainment:a comparison of working women and men. Am. Sociol. Rev. 40:174-200 Tsui AS, Egan TD, O'Reilly CA. 1992. Being different:relationaldemographyand organizational attachment. Admin.Sci. Q. 37:54979

541

Tsui AS, Gutek BA. 1999. DemographicDifCurrent Research ferences in Organizations: and FutureDirections.New York:Lexington Tsui AS, O'Reilly CA. II. 1989. Beyond simple demographiceffects: the importance of relational demography in supervisorsubordinatedyads. Acad. Manage. J. 32: 402-23 VanDen Berghe PL. 1963. Dialectic and functionalism:towarda theoreticalsynthesis.Am. Sociol. Rev. 28:5:695-705 VannemanR, Cannon LW. 1987. The American Perceptionof Class. Philadelphia:Temple Univ. Press Wadman MK. 1992. Mothers who take extendedtime off findtheircareerspay a heavy price. WallStreetJ. July 16:B1, B5 WardKB, Mueller,CW. 1985. Sex differences in earnings:the effects of industrialsector, authorityhierarchy,and humancapital variables. Work Occup. 12:437-63 Weber M. 1968. Economy and Society, ed. GuentherRoth. New York:Bedminster Wilson G. 1997a. Payoffs to power among males in the middle class: Has race declined in its significance? Sociol. Q. 38:4:60722 WilsonG. 1997b.Pathwaysto power:racialdifferencesin the determinants of job authority. Soc. Prob. 44(1):38-54 Wolf W, Fligstein N. 1979a. Sexual stratification:differencesin powerin the worksetting. Soc. Forces 58:94-107 Wolf W, Fligstein N. 1979b. Sex and authority in the workplace:the causes of sexual inequality.Am. Sociol. Rev.44:235-52 WrightEO. 1978. Race, class, and income inequality.Am.J. Sociol. 83:6:1368-97 WrightEO. 1979. Class Structureand Income Determination.New York.Academic Press WrightEO. 1989. The comparativeprojecton class structureand class consciousness: an overview.Acta Sociol. 32:3-22 WrightEO. 1993. Typologies, scales, andclass analysis: a comment on Halaby and Weakliem's "Ownership and authority in the earnings function." Am. Sociol. Rev. 58:3134

542

SMITH Am.So1982. The Americanclass structure. ciol. Rev.47:709-26 WrightEO, PerroneL. 1977. Marxistclass categories and income inequality.Am. Sociol. Rev.42:32-55

WrightEO, BaxterJ, BirkelundGE. 1995. The gender gap in workplaceauthority:a crossnational study. Am. Sociol. Rev. 60:40735 WrightEO, Costello C, Hachen D, SpragueJ.

Anda mungkin juga menyukai