You are on page 1of 22

Race and Loss of Privilege: African American/White Differences in the Determinants of Job Layoffs from Upper-Tier Occupations Author(s):

George Wilson and Debra Branch McBrier Source: Sociological Forum, Vol. 20, No. 2 (Jun., 2005), pp. 301-321 Published by: Springer Stable URL: http://www.jstor.org/stable/4540896 . Accessed: 10/09/2013 05:07
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.

Springer is collaborating with JSTOR to digitize, preserve and extend access to Sociological Forum.

http://www.jstor.org

This content downloaded from 121.54.54.33 on Tue, 10 Sep 2013 05:07:58 AM All use subject to JSTOR Terms and Conditions

Forum,Vol.20, No. 2, June2005 (2005) Sociological DOI: 10.1007/s11206-005-4102-6

Race and Loss of Privilege:African American/WhiteDifferences in the Determinants of Job Layoffs From Upper-TierOccupations
George Wilson1,3 and Debra Branch McBrier2

Data from the Panel Study of Income Dynamics supports predictions from the minority vulnerability thesis concerning the determinants of job layoffs among African Americans and Whites who work in upper-middle-class occupations. Specifically, after controlling for seniority, layoffs for African Americans are relatively unstructured by traditional stratification-based causal factors, namely, background socioeconomic status, human-capital credentials, and job/labor market characteristics.Analyses also indicate that racial differences in the determinants of layoffs are more pronounced in nonservice-based than service-based firms in the private sector and in the private sector relative to the public sector.
KEY WORDS: job layoffs; workplace; racial inequality.

In the last decade and a half, a developing line of sociological research on racial stratification in the workplace has focused on African Americans and Whites who occupy relatively privileged positions in the American occupational structure (for reviews see Patillo-McCoy, 1999; Wilson et al., 1999). This research has shed light on the dynamics of inequality among a growing segment of the increasingly differentiated African American population. Sociologists have documented that in recent decades African Americans have increased their representation in upper-tier occupational categories-namely, those that include managers, executives,
of Sociology,Universityof Miami,CoralGables,Florida. 1Department of Sociology,Box 19599,Universityof Texasat Arlington,Arlington, Texas. 2Department 3To whomcorrespondence shouldbe addressedat Departmentof Sociology,Merrick Building, Universityof Miami,CoralGables,Florida33124;e-mail:gwilsonl@miami.edu.
301
0884-8971/05/0600-0301/0 2005 Springer Science+Business Media, Inc.

This content downloaded from 121.54.54.33 on Tue, 10 Sep 2013 05:07:58 AM All use subject to JSTOR Terms and Conditions

302

Wilson andMcBrier

twice the rate of Whites(Farley,1996; and professionals-at approximately and Williams,1989). Jaynes in upper-tier To date, studiesof racialstratification amongincumbents occupationshave explored several fundamentalissues. For example, one line of researchdemonstratesthat AfricanAmericanmanagersand executives tend to be channeled into raciallydelineated career and mobility in organizations andfirms socioeconomicrewards tracksthatyield marginal that are owned and managedby Whites (Collins, 1993;Durr and Logan, 1997;Moore, 1981).A second line of inquiryshows that AfricanAmerican andrewardmanagersand executiveshave restrictedaccessto higher-order generatingjob tasks, includingsupervisoryauthority,job autonomy,and substantivelycomplex work (Hyllegard,1996;Hyllegardand Lavin, 1992; Kluegel, 1978; Smith, 2002, 1997; Smith and Elliott, 2002; TomaskovicDevey, 1993;Wilson,1997).A thirdline of researchestablishesthatAfrican Americans, relative to Whites, receive fewer income and socioeconomic for investmentsin human-capital status "returns" credentialssuch as years of educationas well as experiencein the laborforce and with their present employer (Farley and Allen, 1987;Jaynes and Williams,1989;Son et al., 1989;Thomas,1995;Thomaset al., 1994). Issues that enhance our understandingof racial stratificationin upper-tieroccupations,however, have yet to be addressed.Specifically, we lack examinationsof how privilegedpositions are lost. For example, race-specificprocesses governing layoffs, which have reached unprecedented levels in relativelyprivilegedpositions duringthe 1980s and 1990s (see Fairlie and Kletzer, 1998;Neumark,2000) may constitute an aspect of racial inequality in the workplace. Recent anecdotal evidence from academictreatisesand media accountsindicatethat,with regardto layoffs, which are typically based on the discretionaryjudgment of employers (Cornfield,1983;McCuneet al., 1988), African Americans"do not have opportunitiesto demonstratetheir value to employers"(McCune et al., 1988:148)and have to work "twice as hard" to avoid "gettingthe axe" If accurate,these statementsindicatethat (Zwerlingand Silver, 1992:655). African Americans are relatively susceptible to "negative stratificationbased events" (Feathermanand Hauser, 1978) which, even if temporary, have long-lastingconsequences,such as delaying the timing (Fairlie and Kletzer, 1998; Leana and Feldman, 1995; Root, 1988) and reducing the level of socioeconomic rewards over the career (Golden, 1992; Mouw, 2000;Swinnertonand Wial, 1995). To enhance our understanding of racial inequalityat upper levels of the American occupationalstructure,we need to undertakeanalyses of how race affects layoffs, a criticalform of job loss. This study utilizes lonprobability sampleto assess gitudinaldata from a nationallyrepresentative

This content downloaded from 121.54.54.33 on Tue, 10 Sep 2013 05:07:58 AM All use subject to JSTOR Terms and Conditions

RacialDifferencesin Reasonsfor LayoffsFromUpper-Tier Occupations

303

predictionsfrom a theoreticalperspectivethat identifiesdifferencesin the reasonsfor layoffs among a cohort of AfricanAmericansand Whites who occupyprivilegedpositionsin the Americanoccupationalstructure. THEORY The MinorityVulnerability Thesis The "minorityvulnerabilitythesis" distills recent case studies and survey-basedanalyses that address race-based patterns of layoffs from upper-tieroccupationsin firms that are predominantlyowned and manof those layoffs determinants aged by Whites to elucidatethe race-specific et 1993; Cornfield, al., 1998; Cascio, 1993; Collins, 1983; Elvira (Boisjoly and Zatlick, 2002; Fernandez, 1975; Jaffee, 1986; Moore, 1990; Reskin, 2000;Spalter-Rothand Deitch, 1999;Schervish,1986;Zwerlingand Silver, 1992).The thrustof this perspectivederives from the institutionalcontext in which layoffs occur:while layoffs are necessaryafter such structurally based events as downsizingand firmreorganization, employersusuallydewill be off et cide who laid (see Boisjoly al., 1998;Fernandez,1981). the minorityvulnerability thesisdocumenthow emStudiescomprising by meritocratic ideologies make layployersin work settingscharacterized off decisions that reinforce existing patterns of racial exclusion. Accordingly, race-basedpatternsof layoffs are a manifestationof "modernracial prejudice"(Pettigrewand Martin,1987), which is characterizedas situational, ostensiblynonracial,and institutionalin nature.In general,the minority vulnerabilitythesis posits that dynamicsrangingfrom a perceived need to conformto existingnormsof racialexclusionin orderto maintaina base to cognitivedistortions stable workforceand a steady customer/client inherent in "self-servingattributionbias" (Pettigrew,1985) and "statistical discrimination" (Tomaskovic-Deveyand Skaggs, 1999; Wilson, 1997) in infrom arising stereotypesresult in layoffs that are not discriminatory excluderacialminoritiesfromtop-level tent but serve to disproportionately positions. One line of research in the minority vulnerabilitythesis posits that African Americansin upper-tieroccupationsare expendable,relative to Whites, because they are placed in jobs that generate relatively low levels of revenue and are consideredmarginalto long-termtrajectoriesfor the economic viabilityof firms.In this vein, Collins (1997) and Fernandez in politically (1981) establishthat AfricanAmericansare overrepresented on functions that focus induced slots involving"racialized" producing job servicesused by or concernedwith minoritygroupsas comparedto "mainstream" job functionsthat are relativelyprofitableand focus on consumers

This content downloaded from 121.54.54.33 on Tue, 10 Sep 2013 05:07:58 AM All use subject to JSTOR Terms and Conditions

304

WilsonandMcBrier

and clients that are not raciallydifferentiated.Significantly, workersperracialized functions are to be vulnerable forming likely disproportionately to downsizingand job cuts because of both market-driven considerations and reduced governmentconcernwith race-basedemploymentinequities in the post-1980period (Burstein,1985;Landry,1987;Collins,1993). A second line of studies that informsthe minorityvulnerabilitythesis maintainsthat AfricanAmericansare disproportionately vulnerableto whetheremployed layoffsresultingfrom evaluationprocesses.Specifically, in racializedor mainstream slots, AfricanAmericanslack opportunitiesto demonstrateto employersthe informalor "particulartistic" (Kluegel,1978) suchas loyalty,trustworthiness, characteristics andleadership potentialthat are crucial for employers'decisions regardinglayoffs. Similarly,African Americans,relativeto Whites,cannotutilizeformsof social/cultural capital that derivefromfactorssuchas privilegedclassbackground, educationalattainment,and a prestigiousjob title to influenceperformanceevaluations. For example,Collins(1997) and Wilson (1997) documentthat the reliance of African Americanson segregatedjob networksrendersthem susceptibias" (Pettigrew,1985) so that ble to cognitivebiases such as "attribution reaffirm are assessed on selective bases that they negativestereotypesabout work. their suitabilityfor, and productivity at, Moreover,Fernandez(1981) establishesthat AfricanAmericans'tendencyto workin raciallydelineated work groups,that is, entities formedto accomplishspecifically targetedorthem to form leaves informational a of stabias, ganizationalgoals, prone in whichtheircharacter andperformance tisticaldiscrimination evaluations are viewed as less crediblethan those of Whitepeers. Finally,Wilsonet al. (1999) maintainthat the tendency for African Americansto be subordinated to Whites in authorityhierarchieslimits their opportunitiesto meet criteriathat influencelayoff decisions. informal,performance-based thesis maintainsthat at privileged Overall,the minorityvulnerability levels of the American occupationalstructureAfrican Americans'placement in racially delineated jobs and the constraintson their ability to demonstratethe "rightstuff' for favorable performanceevaluations reof layoffs.Specifically, sults in a set of race-specific determinants layoffsfor African Americans,comparedto Whites, are broadlybased and generalized:they are relativelyunstructured stratificationby a rangeof traditional, based causalfactorssuch as backgroundstatus,human-capital credentials, The minorityvulnerability and job/labor-market characteristics. thesis, accordingly,maintainsthat AfricanAmericansare more likely to experience layoffs fromupper-tieroccupationsthanWhiteseven when the two groups have similarbackground socioeconomicstatuses,have accumulated similar and commitment such as educationalattainment credentials, human-capital to work, and have similarjob/labor-market characteristics, includingunion statusas well as economicsector of employment.

This content downloaded from 121.54.54.33 on Tue, 10 Sep 2013 05:07:58 AM All use subject to JSTOR Terms and Conditions

RacialDifferencesin Reasonsfor LayoffsFromUpper-Tier Occupations

305

to the MinorityVulnerability Thesis Qualifications to the scope of the minorityvulnerability thesis Finally,qualifications emergefromstudiesthat compriseit. First,the causaldynamicsresponsible for race-specific determinants of layoffs shouldvary acrossindustrialcatewithin the sector. the characteristics of firmsin gories private Specifically, the "newservice-sector relative to nonserviceeconomy"(Neumark,2000), based firmsin, for example,traditional and extractiveindusmanufacturing in across racial tries, producegreaterequality groups the determinantsof Most in this layoffs. noteworthy regardis that service-basedfirmstend to be younger than nonservice-based firms (Reid and Rubin, 2003). Significantly,a line of sociologicalresearchhas documentedthat "youngerfirms tend not to have the deeplyentrenched, basednormsof racial institutionally exclusiontypicalof older firms"(Rosenfeld,2000:71),which-according to the minorityvulnerability thesis-account for racistemploymentpractices at upperlevels of the occupational structure(Pierce,2001;Reid and Rubin, 2003;Rosenfeld,2000). Second, causal dynamicsresponsiblefor race-baseddeterminantsof layoffs should be more pronouncedin the private sector than the public sector. Another line of sociological research has documented that the rate of workforce reduction since the early 1990s has accelerated not only in the private sector but also in the traditional occupational "niche"of the AfricanAmericanmiddleclass-the publicsector-because of factors ranging from civil service reforms that have narrowed the scope of traditional forms of security (Condrey and Maranto, 2001; Villemez and Bridges, 1994; Walters, 2002) to the growing trends toward "privatization" and "outsourcing"of government-based job functions and services(Kuykendalland Facer,2002;Lawthler,2003;Neumark, 2000). Nevertheless,numeroussociologicalstudies have found that the private sector is relativelyinhospitableto African Americansprimarilybecause of the narrowerreach and less stringent enforcement of legislation enacted to ensure equal economic opportunitiesfor minorities in the post-1965 "civil rights" era (Burstein, 1985; Collins, 1993; Dobbin et al., 1993; Edsall and Edsall, 1991; Moore, 1981). Accordingly, sectoral differencesin racialinequalityregarding job layoffs from upper-tier are to be occupations projected relatively benign in the public sector, similar to the pattern found along a range of socioeconomic outcomes such as income (Hout, 1984; Pomer, 1986), rates of promotion (Baldi and McBrier, 1998; Hout, 1984; Wilson et al., 1999), socioeconomic status (Feathermanand Hauser, 1978), and the attainmentof job authority (Elliott and Smith,2001; Smith, 1997;Smith and Elliott, 2002;Wilson, 1997).

This content downloaded from 121.54.54.33 on Tue, 10 Sep 2013 05:07:58 AM All use subject to JSTOR Terms and Conditions

306

WilsonandMcBrier

DATA AND METHODS Data fromthe 1994-2001waves of the Panel Studyof IncomeDynamics (PSID) are utilized to assess racial differencesin the determinantsof job layoffs over 5 years of the work careeramongrelativelyoccupationally privilegedAfricanAmericanandWhitemales (see Hill, 1992,for a description of the PSID data set). In particular,the sample population consists of all non-self-employedWhite and African American "heads of households"between the ages of 18 and 50 yearswho workedpart-or full-timein employmentand were interviewedin 1994, 1995, or 1996 nonagricultural as well as in subsequentwaves until they either experiencedjob lay off or continued to be employed in the Managers/Administrators or Professional/Technicalcensus-basedoccupationalcategories.Those who either retired from the workforce,quit their jobs, or were fired from their initial jobs during the sample years were excluded from the sample.4The applicationof these criteriaresulted in a sample of 1068 Whites and 676 African Americans.5 The model used in the analysesis operationalizedas follows: Dependent Variable:Layoff is operationalizedas whether or not sample membershave been "laidoff from their "mainjob" over 5 years of the
work career (1 = yes; 0 = no).6

75% (N = 935) of samplememberswere interviewedin 1994.Those origi4Approximately that is, those who previouslywere nally interviewedin 1995and 1996represent"splitoffs," trackedin the PSID but were not headsof householdsor were interviewed for the firsttime We wouldhave liked to include by the ICPSRas heads to compensatefor sampleattrition. a range of forms of job displacement-such as firings,quits, and retirements-as separate outcomesandperformed a multinomial to assessthe determinants of each logisticregression amongAfricanAmericanandWhites.However,the smallnumberof firingsandretirements over the 5-yearperiod amongsamplemembersprecludedperforming a multinomial logit. While the numberof quitswas sufficientto includeas a categoryin the analysis, we did not do so for theoreticalreasons.Specifically, shows sociologicalresearchon job displacement that quit positionsfor a varietyof reasons,such as movinginto betterpositions,perceiving to resign.Suchdiversereasons and being "forced" discrimination, becomingself-employed, make it difficultto link quitstheoretically with processesof discrimination and inequalityin the labormarket. Latinosandotherminority 5Including groupsin the analysesis precluded by theirinclusionin the broad"other" categoryby the Institutefor SurveyResearch. 6One methodologicalissue worthy of mention concerns whether self-reportedincidence of layoffs is inflated because workers, especially those in a dominant group, are apt to report firings, which are individuallycaused, as layoffs, which are the result of extraindividual organizationalcircumstance.Schervish (1986) addressed this issue and found no statisticallysignificantdifferences across racial groups between self-reported incidence of layoffs and employer-reported incidence of layoffs in employee personnel files.

This content downloaded from 121.54.54.33 on Tue, 10 Sep 2013 05:07:58 AM All use subject to JSTOR Terms and Conditions

RacialDifferencesin Reasonsfor LayoffsFromUpper-Tier Occupations

307

IndependentVariables:The influence of several categories of factors in of job layoffs amongAfricanAmericansand assessingthe determinants Whites are included.
Background Socioeconomic Status: Background status is measured with

two variables.The first is mother's education,which is coded categorically and based on meaningful distinctions in levels of educational attainment,namely,high school completionor less, college degree, and postcollege degree. Higher values in the coding of this variableindicate successively greater levels of educational attainment. The second is whichis measuredby whetherboth parentswere in the familystructure,
household until the respondent reached the age of 16 (1 = yes, 0 = no).

Human-CapitalCredentials:Several human-capitalcredentials are assessed. The first is level of educationalattainment,representedby two dummyvariables:"college degree,"and "postcollegedegree."Respondents with a high school degree or less serve as the referencecategory. The rationalefor coding educationas a series of dummyvariablesis to allow for the possibilityof nonlinearrelationshipswith the dependent variable.7Second, consistentwith prior research,attendanceat work is used as an indicatorof job commitment(see Muelleret al., 1992).Specifas the numberof job absences ically,job commitmentis operationalized absences are reverse-codedso that in a Job year. among respondents commitment scores reflect (i.e., fewer absences) and greaterjob higher lower scoresreflectlesserjob commitment(i.e., greaterabsences).
Job/Labor-Market Characteristics: The influence of several job/labor-

is assessed.First,union status of job is measured marketcharacteristics


by 1 = yes, 0 = no. Second, sector of employment is measured as 1 =

public, 0 = private. In addition, this study assesses the influence of


were coded dif7Acrosspredictorand controlvariablesin the statistical model, respondents uponwhetheror not they werelaidoff duringthe sampleyears.For those ferentlydepending laid off, variableswere coded at the time they experienced job loss. On average,layoff was For those not laid off, we assignedvaluesfor 2.4 yearsafterthe initialinterview. experienced one-halfof the time they the variables2.5 years after their initialinterview,approximately errorbetween were trackedin the study.This coding scheme minimizesthe measurement those who experiencedlayoff and those who did not. The majorexceptionto this coding of job commitment. For workerswho were not laid off schemeis job absences,our indicator duringthe sampleyears,absenceswere coded as the numberof timesworkwas missedin the 1-yearperiodbetweentheirfourthandfifthyear of inclusionin the sample;for workerslaid off more thanone year afterthe initialinterview, job absenceswere coded as the numberof withinthe first timesworkwasmissedin the preceding year.Finally,if layoffwasexperienced of timesworkwasmissed job absenceswerecodedas the number yearof the initialinterview, withinthat 1-yearperiod.

This content downloaded from 121.54.54.33 on Tue, 10 Sep 2013 05:07:58 AM All use subject to JSTOR Terms and Conditions

308

WilsonandMcBrier

industrialcategories within the private sector. In particular,it adopts a view of the private sector in the 1990s that is contained in recent labor-marketanalyses,which divides industriesinto and "service"and "nonservice"segments (see Neumark, 2000; Reid and Rubin, 2003). Specifically,the categoriesof (a) finance,insurance,and real estate; (b) commerce,andpublicutilities;and (c) business,personal, transportation, entertainment,and professionalservices are classified as services and coded as 1; the categoriesof (a) mining;(b) durablegoods production; (c) nondurable goods production;(d) construction;(e) forestry; and (f) retail and wholesale trade constitute at least two recognized types of industries-manufacturing and extractive-that are classified as nonservicesand coded as 0. In order to parcel out the effects of "seniority,"two Control Variables: variablesare includedas statisticalcontrols:experiencein the workforce since age 18 (years)and lengthof time with presentemployers(months). In addition,the numberof hoursworkedper week is introducedto control for part-time/full-time status. Two additionalcontrols are occupation (1 = managers/administrators,0 = professional/technical) and gender (1 = female, 0 = male).

Finally,the dependentvariablein this study,job layoff,is dichotomous. multivariate Therefore,logisticregressionis the appropriate analytictechdeterminants to the of This uses study logistic regression analyze nique.8 The deterlayoffsseparatelyfor the AfricanAmericanand Whitesamples. minantsof layoffsbetweenthe two racialgroupsin the entiresampleas well as acrosssegmentsof the privatesector-manufacturingand service-and of the logit coeffithe publicsector are assessed.Overall,the interpretation cients is based on odds ratiosthat are constructed by computingthe antilog of each unstandardized regressioncoefficient.9
8We consideredusing proportionalhazardsmodeling.However,assessingissues of timing withinthe 5-yearperiod would detractfrom the overriding thrustof the minorityvulnerain the likelihoodof job layoff. racialdifferences bilitythesis,namely,identifying the exponent(i.e., antilog)of eachmetricregression by computing 9Oddsratiosarecalculated coefficientand convertingthe odds ratio into a percentagechange in the odds definedas follows:100exp(b)- 1. The odds ratiofor each regressioncoefficient(exp(b)) signifiesthat for each unit changein the independent variable,the odds are expectedto increase/decrease on whetherthe oddsratiois less thanor greaterthan1.0).by of a factorof exp(b) (depending is measuredby the differenceof (the values of the odds ratio).The size of the relationship The furtherthe value the odds ratio (in eitherdirection)from 1.0 (meaningno relationship). the predictorvariable,and an odds ratio of the odds ratio is from 1.0, the more influential while an odds ratiogreaterthan 1.0 signifiesa less than 1.0 indicatesa negativerelationship, positiverelationship.

This content downloaded from 121.54.54.33 on Tue, 10 Sep 2013 05:07:58 AM All use subject to JSTOR Terms and Conditions

RacialDifferencesin Reasonsfor LayoffsFromUpper-Tier Occupations TableI. Incidenceof Job Layoffsover 5 Years:PSID Sample AfricanAmericans
(N = 676)

309

Whites
(N = 1068) Diff.

laid off Percentage Entiresample Privatesector Nonservice Servicesector Publicsector statusof those laid off Meansocioeconomic Entiresample Privatesector Nonservice Servicesector Publicsector
*p < 0.01. ***p > 0.001. 0.05. **p <

31 41 37 15 0.64 0.62 0.63 0.65

16 17 22 10 0.65 0.64 0.65 0.66

6.22** 7.77*** 6.25** 4.56* 1.32 1.50 1.43 1.27

RESULTS Table I shows that racial differencesexist in the incidence of layoffs over 5 years of the work career(descriptivestatisticsfor all variablesin the model appearin Appendix 1). the percentageof layoffs for African Americans(31%) In particular, was nearly twice that for Whites (16%) in the entire sample. In addition, acrosssegmentsof racialdifferencesappearin the percentagesof dismissals firmsmore than twice the percentage the privatesector:in manufacturing of African Americans(41%) as Whites (17%) experiencedlayoffs, while in service firmsthe percentageof AfricanAmericansexperiencinglayoffs was less than twice that of Whites (37%vs. 22%).Further,the racialgap in layoffsover the 5-yearperiodis smallerin the publicthanthe privatesector: one-thirdmore AfricanAmericans(15%) than Whites (10%) experienced layoffsin the publicsector. Finally,Table I also indicatesthat the racial gaps in the incidence of layoffs exist among African Americans and Whites whose status is simiand professional/technical lar within the managers/administrators occupational categories. Specifically,among the entire sample as well as across segmentsof the privatesector and in the publicsector,the mean socioeconomic statusscores amongAfricanAmericansand Whiteswho are laid off are statistically indistinguishable.10
was measuredwith Duncan SocioeconomicIndex scores. Similaranalysesindicate 10Status that amongthe entiresampleas well as acrossboth economicsectorsthere are statistically differencesamongAfricanAmericansand Whiteswho were not laid off. indistinguishable Resultsare availableupon request.

This content downloaded from 121.54.54.33 on Tue, 10 Sep 2013 05:07:58 AM All use subject to JSTOR Terms and Conditions

310

WilsonandMcBrier

TableII. LogisticRegressionfor Determinants of Job LayoffsAmongAfricanAmericans andWhites:PSID Sample


African Americans (N = 676) Whites (N = 1068)

Maineffectscontrol

Coefficient.

Oddsratio 0.99 0.98 0.99 1.11 1.05 0.99 0.98 1.03 1.02 0.96 0.90 0.83 0.85 0.278 -2.158 -227.43

Coefficient -0.03* (.01) -0.02* (0.01) 0.01(0.01) 0.10(0.01) 0.07(0.02) -0.09 (0.03) -0.20** (0.05) -0.15** (0.05) -0.20***(0.06) -0.06***(0.03) -0.19** (.06) -0.20** (.05) -0.19** (.05)

Oddsratio 0.97 0.98 1.01 1.10 1.07 0.91 0.78 0.84 0.79 0.93 0.78 0.78 0.79

Timewith employer -0.01 (.02) -0.02* (0.01) Hoursworked -0.01 (0.01) Laborforce experience Managers/administrators 0.13*(0.03) Female 0.05(0.02) status Background -0.01 (0.01) Mother'seducation -02 (0.02) Familystructure Humancapital 0.03(0.02) College 0.02(0.01) Postcollege -0.04* (0.02) Job commitment Job/labor-market Union status -0.10 (0.03) Publicsector -0.15** (0.04) -0.14* (0.04) Service PseudotR2 0.141 Constant -2.138 -241.84 Log-likelihood Note. Standard errorsare in parentheses.
*p < 0.05. **p < 0.01. ***p > 0.001.

Analysis of the Entire Sample

Table II presents findingsfrom the logistic regressionanalyses conof job layoffsamongthe entire sampleof African cerningthe determinants Americans and Whites (see Appendix 2 for a log-likelihoodratio test of equalitybetween coefficientsacrossthe modelsfor AfricanAmericansand Whites). The findingssupportthe notionthatrace-baseddynamics drivejob layoffs along lines set forth by the minorityvulnerabilitythesis: layoffs for African Americans,relative to Whites, are relativelyunstructured by traditionalstratification-based causalfactors.Severalfindingsstandout. First, the model has more explanatorypower for Whites than African Americans.Specifically, morethantwicethe numberof predictor variablesare stafor Whites than for African Americans-7-3. Further, tisticallysignificant the significantvariablesexert greater effects among Whites than African Americans:all variablesfor Whites exert either robust (i.e., significantat the 0.001 level) or moderate(i.e., significantat the .01 level) effects, while the significantvariablesfor African Americansexert either moderate or modest (i.e., significantat the 0.05 level) effects. Finally, the explained

This content downloaded from 121.54.54.33 on Tue, 10 Sep 2013 05:07:58 AM All use subject to JSTOR Terms and Conditions

RacialDifferencesin Reasonsfor LayoffsFromUpper-Tier Occupations

311

variancein the model is more than twice as great for Whitesas for African Americans(0.287- 0.131). effects supportsthe notion that the Second,the directionof significant process of layoffs for African Americansis relativelybroad-based.In this vein, the lone significantvariablethat measuresbackgroundstatus across the two models-family structure-acts on Whites:coming from an intact family reduces the likelihood of experiencinglayoffs by 22%. Among the variables,unit increases in job commitmentinsulate both human-capital AfricanAmericansand Whites from layoffs:specifically, unit increasesreduce the likelihoodof experiencinglayoffs amongAfricanAmericansand Whites, respectively,by 4 and 7%. In addition,two other human-capital variablesaffect only on Whites:havinga college degree and a graduatedegree reducesthe likelihoodof experiencinglayoffs by, respectively,16 and 21%. Among the job/labor-market variables,employmentin both the public sector and a service firm affects both African Americansand Whites. workingin the public sector reduces the likelihood of layoffs Specifically, African Americans by 17% and Whites by 22%, while working among in a service-basedfirm reduces the likelihood of African Americansand Whites experiencing layoffsby, respectively,15 and 21%.However, a third variable has an effect only on Whites:incumbencyin a job/labor-market unionizedpositionservesto reducethe likelihoodof layoff by 22%.11 Analysesin the Privateand Public Sector Similarmultivariateanalysesto assess the determinantsof layoffs in Table II were performedfor AfricanAmericansandWhites acrossthe services and nonservicesindustrial segmentsof the privatesector (PanelA) as well as in the publicsector (Panel B). Results are presentedin Table III. Based on the same criteria used to assess the minority vulnerabilthesis in the analyses of the entire sample, racial differences in the ity determinantsof job layoffs emerge across segments of the private sector. The racial differences in the determinantsof layoffs are greater in the nonservices segment than the services segment of the private sector. First, there are greater racial differences in the explanatorypower than service-basedfirms. of the statisticalmodel among nonservice-based
to ensurethat resultswere not confounded were performed on model specification 11Checks or multicollinearity. The Cook-Weisbergtest of the assumptionof by heteroscedasticity In all instancesX2statistics wasperformed for allregression commonerrorvariance analyses. of .01 were obtainedand had corresponding p-valuesthat rangedfrom0.642to 0.721,indiIn addition,collinearity were performed, catinglow levels of heteroscedasticity. diagnostics indicesproducedno evidenceof multicollinearity. and conditional

This content downloaded from 121.54.54.33 on Tue, 10 Sep 2013 05:07:58 AM All use subject to JSTOR Terms and Conditions

Table III. Logistic Regression for Determinants of Job Layoffs: Within Segments of P African Americans (N = 215) Nonservice Whites (N = 463) African A Main effects Control Coeff. Odds ratio 0.99 0.98 0.99 1.15 1.03 0.99 1.03 1.04 1.06 0.97 0.99 Coeff. -0.02* (0.02) -0.02* (0.01) 0.01 (0.01) 0.09 (0.02) 0.08 (0.03) 0.01 (0.01) -0.15** (0.04) -0.21** (0.07) -0.21*** (0.07) -0.06*** (0.03) -0.24** (0.07) 0.311 -1.760 -206.83 -0.04* (0.02) -0.01(0.01) 0.01 (0.01) 0.10 (0.02) 0.06 (0.02) -0.13** (0.03) -0.25*** (0.05) -0.09 (0.02) -0.24*** (0.06) -0.05** (0.03) -0.20** (0.05) 0.228 -2.673 -2.798 Odds ratio 0.98 0.98 1.01 1.09 1.08 1.01 0.83 .79 0.78 0.94 0.75

Coef

Panel A: Private sector Time with employer -0.01 (0.01) -0.02* (0.01) Hours worked -0.01 (0.01) Labor force experience 0.15* (0.04) Managers/administrators Female 0.03 (0.01) Background status -0.01 (0.01) Mother's education 0.03 (0.01) Family structure Human capital 0.04 (0.02) College 0.06 (0.02) Postcollege Job commitment -0.03* (0.01) Job/labor-market -0.01 (0.01) Union Status 0.088 Pseudo R2 -1.856 Constant 206.15 Log-likelihood Panel B: Public Sector Time with employer 0.01 (0.01) -0.03* (0.02) Hours worked Labor force experience -0.01 (0.02) 0.10 (0.02) Managers/administrators 0.07 (0.01) Female Background status 0.01 (0.01) Mother's education -0.16** (0.04) Family structure Human capital 0.03 (0.01) College 0.05 (0.01) Postcollege Job commitment -0.06*** (0.02) Job/labor-market -0.18** (0.04) Union status 0.174 Pseudo R2 -2.642 Constant -2.845 Log-likelihood Note. Standard errors are in parentheses. 0.05. **p < 0.01. ***p < 0.001. *p <

-0.01 (0 -0.02* ( 0.01 (0 0.12 (0 0.05 (0

-0.01 (0 0.05 (0

0.03 (0 -0.13** -0.03* (

-0.09 (0 0.132 -1.922 -205.76

1.01 0.97 0.99 1.10 1.07 1.01 0.83 1.03 1.05 0.96 0.82

0.96 0.99 1.01 1.10 1.06 0.87 0.73 0.91 0.74 0.95 0.80

This content downloaded from 121.54.54.33 on Tue, 10 Sep 2013 05:07:58 AM All use subject to JSTOR Terms and Conditions

RacialDifferencesin Reasonsfor LayoffsFromUpper-Tier Occupations

313

the racialgap in the numberof significant predictorvariables Specifically, in nonservice firms (five White, one African American) is greater than in service firms (five White, two African American).Further,the magnitude of the statisticallysignificantvariablesdiffers in nonserviceand service firms:in the model for nonservice firms, all statisticallysignificant variables for Whites exert either moderate or robust effects, while the lone significantvariablefor African Americansexerts a modest effect. In the service model, significantvariablesfor Whites exert moderate or rovariablesexert either modbust effects;for AfricanAmericans,significant est or moderate effects. Finally, the racial gap in the variance explained (0.311 White, 0 .088 by the model is greater among manufacturing-based African American) than service-basedfirms (0.254 White, 0.132 African American). Second,the directionof effectsprovidessupportfor the notionthat difof layoffsacrosssegmentsin the privatesector ferencesin the determinants occur along lines predicted by the minorityvulnerabilitythesis. In mancredentialsufacturingfirms, one variable that measureshuman-capital in commitment unit increases racial across commitment-acts groups: job and Whites Americans African of both likelihood reduce the experiencing layoffs. However, a range of other variablesamong those that measure backgroundstatus, human-capital credentials,and job/labor-market acts only on Whites:the likelihood of experiencinglayoffs characteristics is reduced when coming from an intact family, having a college or graduate degree, as well as being an incumbentin a unionized position. In credentialsthe servicemodel, two variablesthat measurehuman-capital the likelihood of a commitment-and having college degree-reduce job and Whites. Americans African for both However, experiencinglayoffs three other variables-unionized status, mother'seducation,and growing up in an intactfamily-serve to reducethe likelihoodof layoffsonly among Whites. Table III also reveals relatively minor racial differences in the determinantsof layoffs in the public sector. First, the racial gap in the explanatorypower of variablesin the public sector is relatively small and, variablesis narsimilarly,the gap in the numberof statisticallysignificant in African three row (five White, American). Further, the public sector, the effects of significantvariables are relatively equal: all variables that are statisticallysignificantfor both racial groups exert either moderate of robust effects. Finally, the racial gap in the variance explained by the model is small in the public sector (0.228 White, 0.174 African American). Second, the directionof effects supportsthe notion that differencesin of layoffs in the publicsector occur along lines predicted the determinants

This content downloaded from 121.54.54.33 on Tue, 10 Sep 2013 05:07:58 AM All use subject to JSTOR Terms and Conditions

314

Wilsonand McBrier

in the public sector, one thesis. Specifically, by the minorityvulnerability variable that measures background status (family structure), one that measureshuman-capital credentials(job commitment),and one that measures job/labor-market characteristics (workingin a unionized position), African reduce the likelihood that both Americansand Whites will experience layoffs. However, two other variables,one measuringhumancapistatus tal (havinga graduatedegree) and the other measuringbackground (mother'seducation)serve to reduce the likelihoodof layoffs only among Whites. AdditionalAnalyses Finally,additionalanalysesundertaken(resultsnot reportedin tabular form) consisted of convertingthe logit coefficientsfor predictorvariables that measurebackground credentials,andjob/laborstatus,human-capital market characteristics in the statisticalmodel to predicted probabilities of being laid off. Findingsfrom the analysisof the entire sample provide thesis. Specifically,acrossvalues all supportfor the minorityvulnerability variables that measure backgroundsocioeconomic status, human-capital AfricanAmericanshave credentials,and job/labor-market characteristics, of greaterpredictedprobabilities experiencing layoffsthanWhites.12 Overall, the racial gap in the predicted probabilityof a layoff tends to be credentials-education, largestfor a variablethat measureshuman-capital for example, ranges between 54% (high school: 37% African Americans, 24% Whites) and 114% (postcollege: African Americans 30%, Whites 14%). In addition,the racialgap in predictedprobabilityof layoff is greater than service-basedfirms in the private sector and among manufacturing greater in the private than the public sector. Specifically, across values among virtuallyall variables,the racial gap in predictedprobability of layoff is between 10 and 20 percentage points higher in the manufacturingthan the service model. Finally, across values among virtually all variables,the racial gap in predicted probabilityof layoff is approximatelytwice as high in the service model as in the public-sector model.
12"Values" refers to the coding scheme of particular variables. The predicted probabilities were calculated using the following equation: exp(predictedy)/(1 + exp(predictedy)). Finally, "baseline probabilities" for African Americans and whites were calculated, that is, probabilities when all other variables in the statistical model are set to their race-specific means. Results of the predicted probability analyses are available from the authors upon request.

This content downloaded from 121.54.54.33 on Tue, 10 Sep 2013 05:07:58 AM All use subject to JSTOR Terms and Conditions

RacialDifferencesin Reasonsfor LayoffsFromUpper-Tier Occupations

315

DISCUSSION at privilegedlocations in Sociologicalresearchon racial stratification the American occupationalstructurehas paid scant attention to the way job dismissalsunfold. Analyses of data from the PSID supportpredictions from the minorityvulnerability thesis concerningthe determinantsof one of prominenttype dismissal-layoffs-among AfricanAmericanandWhite incumbentsin managers/administrators and professional/technical occupational categories across 5 years of the work career. Specifically,findings indicatethat, aftercontrollingfor seniority,AfricanAmericansare susceptible to layoffs on a relativelybroad and generalizedbasis that is unstructuredby traditional, stratification-based causalfactors,namely,background socioeconomic status, human-capitalcredentials, and job/labor-market characteristics. we believe these findingsbroadenour understanding of Significantly, "socialclosure"(Weber,1968),that is, the processby whichmajoritygroup membersrestrictthe access of minoritiesto coveted positionsin the workplace (Parkin,1971;Tomaskovic-Devey,1993). They suggest, specifically, that in the context of upper-tieroccupations,closureoperatesin diametrically opposite ways for "negative"(e.g., job dismissal,downward mobility) versus "positive"(e.g., job authority,upwardmobility)stratification-based outcomes. If job layoffs are indicative,African Americans are disadvantaged because they are vulnerableto negative outcomes on a broaderand more generalizedbasis than Whites.Conversely,a body of sociologicalresearch has documentedthat African Americansare handicappedbecause of the relativelynarrowand circumscribed mannerin whichthey reachpositive outcomes at upper levels of the occupationalstructure.The attainment of positive rewardsamongAfrican Americans,relativeto Whites,is "loadup" on human-capital crepremisedon havingto disproportionately dentialssuch as educationand relevantjob experienceto compensate,prito demonstrate crucial, marily,for the lack of opportunities job-relevantinformalcharacteristics (Kluegel,1978;Muelleret al., 1992;SmithandElliott, 2002;Wilson,1997;Wilsonet al., 1999). In addition, analyses indicate that dynamicsassociatedwith the minority vulnerabilitythesis do not operate uniformlyacross segments of the differentiatedprivatesector nor across the private and public sectors. of layoffsare more pronounced First,racialdifferencesin the determinants in nonservice-basedthan in service-basedfirms in the private sector. We speculate that this findingis due-at least in part-to the relative youth of servicefirms,whichlack the deeply rooted exclusionary normstypically and in older firms that foster found manufacturing perpetuatedetrimental, race-basedemploymentpractices.Second,racialdifferencesin the reasons

This content downloaded from 121.54.54.33 on Tue, 10 Sep 2013 05:07:58 AM All use subject to JSTOR Terms and Conditions

316

Wilsonand McBrier

for layoffs are pronouncedin the privatesector relativeto the public sector. We attributethis findingto factors that have been utilized to explain the relativelyinhospitablenatureof private-sector employmentfor African outcomes:for example, Americansalong a rangeof other workplace-based in the post-1965civilrightsera, equalopportunity lawsin the privatesector, as comparedto the publicsector,have been relativelynarrowin scope and have not been as stringently enforced. Overall, we believe these findingsrequire us to offer policy-related recommendationsto alleviate racial disparitiesin job layoffs from priviin leged occupations.We urge the enactmentof these recommendations the privatesector,whereinequalityin the sourcesof job layoffsis most pronounced.In this vein, we see as crucialthe implementation of a strategyto formalizeconditionsof employment.In particular, clearlydelineated and stringentlyenforced guidelines should be establishedthat facilitate both the placementof AfricanAmericansin "mainstream" jobs and integrated work groups to enhance their opportunitiesto work with majoritygroup customers/clients and to work more closely with majoritygroup employees. Further,strategiesshould be formalizedthat lessen the long-standing relianceof AfricanAmericanson segregatedformaland informaljob networks.Finally,employees shouldbe given greateropportunitiesto appeal internalreviewboardswithauthority to reconlayoffdecisions:establishing sider the bases of layoffs,to mandateorganizational changes,as well as to adopt and implementpolicy stancesrelatedto layoffsconstituteimportant in this formof job displacement. steps towardreducingracialdisparities In sum, this study representsa modest step in advancingour underof job layoff amongAfricanAmericansand standingof the underpinnings Whites. First,more refined analysesof layoffs should be undertaken.For example, intergroupcomparisonsshould focus on individualswho work in similar upper-tierjobs such as lawyers, medical doctors, accountants, and so forth. Second, researchshould more directlycapturethe influence of the allocation and evaluative dynamicsthat are posited as producing racial disparitiesin layoffs. An acknowledgedshortcomingin this study is that an interpretationof the process of layoffs that is based, in part, on the absenceof relationships between a range of objectivecharacteristicscredentials,and job/labormarnamely, backgroundstatus,human-capital ket characteristics-and layoffsmay be problematicbecauseother unmeasured variables may explain the lack of significance.Accordingly,data should be collected in specificorganizations, where the potential exists to observe first-hand the extent to whichthe practicesof employersstructure layoffs. When these directionsfor future researchare undertakenwe exof dynamicsunderlyinglayoffs-a crucialfrom of pect our understanding be job displacement-will vastlyimproved.

This content downloaded from 121.54.54.33 on Tue, 10 Sep 2013 05:07:58 AM All use subject to JSTOR Terms and Conditions

RacialDifferencesin Reasonsfor LayoffsFromUpper-Tier Occupations

317

APPENDIX 1
for PanelStudyof IncomeDynamicsSample DescriptiveStatistics AfricanAmericans Whites
Variable Dependent laid off (N = 676) n = 207 (N = 1068) n = 171 n = 79 n = 57 n = 35 M = 4.3, SD = 1.1 M = 35.3, SD = 6.8 M = 14.3, SD = 5.1 M = .52, SD = 0.20 M = .25, SD = 0.12 M = 11.5, SD= 2.3 Intact = 72% 3.02* 2.11 4.33* 1.47 1.56 4.83* 5.13" Diff.

entiresample Privatesector

Manufacturing Service Public sector

n = 88 n = 83 n = 36 M = 3.2, SD = 1.0 M = 32.3, SD = 7.1 M = 12.6, SD = 4.3 M = .46, SD = 0.18 M = .27, SD = 0.10 M = 9.8, SD = 2.2 Intact = 56%

control Independent
Time with employer Hours worked per week Labor force experience Managers/administrators Female

status Background Humancapital High school Collegedegree Postcollegedegree


Job commitment Mother's education Family structure

22% 35% 43%


M = 3.2, SD = 1.0 M = 0.38, SD = 0.11 M = 0.42, SD = 0.13 M = 0.42, SD = 0.09

21% 34% 45%


M = 3.4, SD = 2.4 M = .49, SD = 0.14 M = .35, SD = 0.12 M = .51, SD = 0.15

1.12 1.23 1.63


1.54 5.27* 4.21* 6.07**

Job/labor-market
Unionized Public sector Service

Note. M = mean;SD = standard deviation. *p < 0.10.**p < 0.01.***p< 0.001.

APPENDIX 2: LOG-LIKELIHOOD RATIO TESTS

The formulafor the log-likelihoodstatisticfollows:


G2= -2 log(Lo/Li) = -2(log lo - log L1)

Specifically,we compared a restricted,constrainedmodel, in which model African Americansand Whites are pooled, to a full, unconstrained with a complete set of interactiontermsfor race and all independentvariables. The log-likelihoodratio test determineswhetherthe model with the interactionterms,whichallowsall coefficientsto differby race,fits the data better than the simplermodel that constrainsthe coefficients. significantly Twice the differencebetween the log-likelihoodstatisticsfor the two models is distributedas a chi-squarestatistic,with degrees of freedomequal to termsomittedin the constrained model.For each the numberof interaction

This content downloaded from 121.54.54.33 on Tue, 10 Sep 2013 05:07:58 AM All use subject to JSTOR Terms and Conditions

318

Wilson and McBrier

of the logisticanalyses,likelihoodratio tests were performedthat examine the statisticaldifferencebetween three blocks of predictorvariables: those that measurebackground and status,human-capital credentials, job/labormarketcharacteristics. The resultsfrom all tests supportcomparing sets of predictorvariables in all tests performedfor the analysesby race, acrossmodels. Specifically, the full model with interactions for race fit the datasignificantly better than the simplermodel with no interactiontermsfor race.
LL-full LL-reduced DF-full DF-reduced Chisquare

of individual test of significance blocks:PSID) Sample Log-likelihood 4 2 14.41 status -303.41 -316.55 Background Humancapital -310.83 5 10 -318.22 15.73 Job/labor-market -309.71 3 6 -313.26 13.33 test of significance of individual blocks: Log-likelihood Manufacturing segmentof privatesector 4 2 status -307.41 -317.55 14.66 Background Humancapital 5 -311.23 -317.22 10 15.45 Job/labor-market 2 4 -308.32 -314.26 13.18 of of test individual blocks: Service of Log-likelihood significance segment privatesector 4 status -306.36 2 14.41 -315.32 Background Humancapital 5 -310.59 -316.59 10 15.43 Job/labor market 2 4 -309.95 -314.16 14.83

REFERENCES
Baldi, Stephane, and Debra McBrier

1998 "Do the determinantsof promotion differ for blacks and whites?"Work and Occupations 24:478-497.
Greg Duncan, and

1997 Black Corporate Executives: The MakingandBreakingof a BlackMiddle Class. Philadelphia,PA: Temple UniversityPress.
Condrey, Steven, and Robert Maranto

Boisjoly Johanne, Timothy Smeeding

2001 RadicalReform of the Civil Service. 1998 "Theshiftingincidenceof involuntary New York:LexingtonBooks. job losses from 1968 to 1992."Indus- Cornfield, Daniel trialRelations37:207-231. 1983 "Chancesof layoffs in a corporation: A case study." Administrative Science Burstein, Paul 1985 Discrimination, Jobs, and Politics. 28:352-364. Quarterly Universityof ChicagoPress. Dobbin, Frank, John Sutton, John Meyer, Chicago:
Cascio, Wayne and W. Richard Scott

1993 "Downsizing:What do we know? What have we learned?" Academy of Management Executive 7:95104.
Collins, Sharon

1993 "Equal opportunity law and the constructionof internal labor markets." AmericanSociologicalReview 99:396-427.
Durr, Marlese, and John Logan

1993 "Blacks on the bubble: The vulner- 1997 "Racial submarkets in government AfricanAmericanmanability of black executives in white employment: corporations." agers in New York State."SociologiSociologicalQuarterly 34:429-447. cal Forum58:74-96.

This content downloaded from 121.54.54.33 on Tue, 10 Sep 2013 05:07:58 AM All use subject to JSTOR Terms and Conditions

RacialDifferencesin Reasonsfor LayoffsFromUpper-Tier Occupations Edsall,Thomas,andMaryEdsall 1991 ChainReaction:The Impactof Race, Rights,and Taxes on AmericanPolitics.New York:W. W. Norton. Elliott,James,and RyanSmith 2001 "Ethnic matching of supervisorsto subordinate work groups: Findings on "bottom-up"ascription and social closure."SocialProblems48:258276. Zatlick Elvira,Marta,and Christopher 2002 "Racial composition and voluntary turnover in a multiracial setting." Academy of Management Meetings Best Papers Proceedings22:3856. Fairlie,Robert,andLoriKletzer 1998 "Jobs lost, jobs regained: An analysis of black/white differences in job displacement in the 1980s." IndustrialRelations 37:460477. Farley,Reynolds 1996 The New American Reality. New York:RussellSageFoundation. Farley,Reynolds,andWalterAllen 1987 The Quality of Life and the Color Line In America.New York: Russell SageFoundation. Featherman, David,and RobertHauser 1978 Opportunity and Change.New York: AcademicPress. John Fernandez, 1975 RacismandSexismin WhiteCorporations.New York:JohnWiley. 1981 Black Managers in White Corporations. Lexington, MA: Lexington Books. Golden,Miriam 1992 "The politics of job loss." American Journal of Political Science 36:408430. Hill, Martha 1992 User'sGuideto the PanelStudyof Income Dynamics.Ann Arbor:Institute for SurveyResearch. Hout, Michael 1984 "Occupational mobilityof blackmen, 1962-1973." American Sociological Review49:308-323. David Hyllegard, 1996 Changing the Odds:OpenAdmissions andthe Life Chancesof the Disadvantaged. New Haven, CT:Yale University Press.

319

David,andDavid Lavin Hyllegard, 1992 "Higher education and challenging work:Openadmissions andethnicand in job complexity." genderdifferences Forum7:239-260. Sociological Jaffe,David 1986 "Thepoliticaleconomyof job loss in the U.S., 1970-80." Social Problems 33:297-318. Jaynes,Gerald,andRobinWilliams 1989 A Common Destiny. Washington, DC: NationalAcademyPress. Kluegel,James 1978 "The causes and cost of racialexclusion from job authority."American Review43:285-301. Sociological and Rex Facer Christine, Kuykendall, 2002 "Public in state agencies: employment The eliminationof the meritsystem." Review of PublicPersonnelAdministration22:133-145. Bart Landry, 1987 The New Black MiddleClass.Berkeley and Los Angeles: University of California Press. Wendall Lawthler, 2003 "Privatizingpersonnel: Outsourcing public sector functions." In Steven Hays andRobertKearney(eds.),Public Personnel Administration: Problems and Prospects:196-208. Upper SaddleRiver,NJ:Prentice-Hall. Leana,Carrie,andDaniel Feldman 1995 "Findingnew jobs after a plant closing:Antecedentsand outcomesof the occurrenceand quality of reemployment." Human Relations 48:13811401. McCune, Joseph, Richard Beatty, and RaymondMontagno in manufactur1988 "Downsizing: Practices ing firms."Human Resources Management27:145-161. Moore,Joan 1981 "Minorities in the Americanclasssystem."Daedalus110:275-299. Moore,Thomas 1990 "Thenatureandunequalincidenceof costs."Social Probjob displacement lems 37:230-242. Mouw,Ted 2000 "Job relocation and the racial gap in unemployment in Detroit and AmericanSociChicago,1980-1990." ologicalReview65:730-753.

This content downloaded from 121.54.54.33 on Tue, 10 Sep 2013 05:07:58 AM All use subject to JSTOR Terms and Conditions

320 Mueller, Charles, Jean Wallace, and James Price

Wilson and McBrier

1992 "Employee commitment:Resolving some issues."Work and Occupations 19:211-236. Schervish, Paul 1986 The Structural Determinantsof UnNeumark, David andPower 2000 On the Job: Is Long-TermEmployVulnerability employment: ment a Thingof the Past?New York: in LaborMarkets. Ithaca,NY:Cornell RussellSageFoundation Press. Press. University
Parkin, Frank

In Mark Western and Janeen Baxter (eds.), Reconfigurationof Class and Gender: 55-76. Palo Alto, CA: StanfordUniversity Press.

1971 ClassInequalityand the PoliticalOrder.New York:Praeger.

at work: 1997 "Race,income,andauthority A cross-temporal of blackand analysis white men." Social Problems38:19Patillo-McCoy, Mary 1999 Black Picket Fences: Privilege and 37. Peril among the Black Middle Class. 2002 "Race, gender, and authority in the workplace."Annual Review of Chicago: University of Chicago Press. Sociology28:509-542.
Pettigrew, Thomas

Smith, Ryan

1985 "New black-white patterns: How 2002 "Does ethnic concentration influence best to conceptualize them." AnSoemployees access to authority?" nual Review of Sociology 11:329cial Forces81:255-279. 326. Son, Soo, Suzanne Model, and Gene Fischer 1989 "Polarization and progress in the Pettigrew, Thomas, and Joanne Martin 1987 "Shapingthe organizationalcontext black community:Earningsand stafor African American inclusion." tus gains for young black men in the Journalof SocialIssues43:41-78. era of affirmative action." Sociological Forum4:309-327. Pierce, Jennifer 2001 "Manufacturing consentin the "new" Spalter-Roth, Roberta, and Cynthia Deitch globalized economy."Contemporary 1999 "I don't feel right-sized;I feel out of-work-sized": Gender,race, ethnicSociology 30:445-446. Pomer, Marshall ity, and the unequal costs of dis1986 "Labor market structure, intragenplacement."Work and Occupations erational mobility, and discrimina26:446-482. tion: Black male advancement out Swinnerton, Kenneth, and Howard Wial of low-paying occupations, 1962- 1995 "Is job stability declining in the 1973."AmericanSociologicalReview U.S. economy?" Industrial and 51:650-659. Labor Relations Review 48:293304. Reid, Lesley, and Beth Rubin 2003 "Integrating economicdualismandla- Thomas, Melvin bor marketsegmentation: The effects 1995 "Race,class,andoccupation: An analof race,gender,andstructural location ysis of black and white earningsfor on earnings,1974-2000." Sociological professional and non-professional 44:405-432. males, 1940-1990." Research in Quarterly Race and Ethnic Relations 8:139Reskin, Barbara 2000 "The proximate causes of employ156. ment discrimination." Contemporary Thomas, Melvin, Cedric Herring, and Sociology29:319-328. Hayward Horton 1994 "Discrimination over the life-course: Root, Kenneth 1988 "Job loss: Whose fault, what remeA synthetic cohortanalysis of earnings dies?" Researchin Politics and Socidifferencesbetween black and white males, 1940-1990."Social Problems ety 3:65-84. 41:608-628. Rosenfeld, Rachel 2000 "Employment flexibilityin the U.S.: Tomaskovic-Devey, Donald Changing and maintaining gender, 1993 Gender and Racial Inequality at Work.Ithaca,NY: ILR Press. class, and ethnic work relations."

Smith, Ryan, and James Elliott

This content downloaded from 121.54.54.33 on Tue, 10 Sep 2013 05:07:58 AM All use subject to JSTOR Terms and Conditions

RacialDifferencesin Reasonsfor LayoffsFromUpper-Tier Occupations Tomaskovic-Devey, Donald, and Sheryl Skaggs test of the 1999 "An establishment-level statistical discrimination hypothesis." Work and Occupations26:422445. Villemez,Wayne,andWilliamBridges 1994 The Employment Relationship: Causes and Consequencesof Modern Personnel Administration. Ithaca, NY: Cornell University Press. John Walters, 2002 Life after Civil Service Reform:The Texas, Georgia, and Florida ExperiVA: Endowment for ences.Arlington, the Businessof Government.

321

Weber,Max 1968 Economy and Society. New York: VantagePress. Wilson,George 1997 "Pathwaysto power: Racial differences in the determinants of job auSocialProblems44:38-54. thority." and Wilson, George, Ian Sakura-Lemessy, JonathanWest 1999 "Reaching the top: Racialdifferences in mobility pathsto upper-tier occupations."WorkandOccupations 26:165186. Zwerling,CraigandHilarySilver 1992 "Race and job dismissalsin a federal AmericanSociological bureaucracy." Review57:651-660.

This content downloaded from 121.54.54.33 on Tue, 10 Sep 2013 05:07:58 AM All use subject to JSTOR Terms and Conditions

You might also like