+ All Categories
Home > Documents > Race, Class and Income Inequlity

Race, Class and Income Inequlity

Date post: 14-Apr-2018
Category:
Upload: ruben-ananias
View: 216 times
Download: 0 times
Share this document with a friend

of 31

Transcript
  • 7/27/2019 Race, Class and Income Inequlity

    1/31

  • 7/27/2019 Race, Class and Income Inequlity

    2/31

    Race, Class, and Income Inequality'Erik Olin WrightUniversityof Wisconsin-Madison, Centeror Research nPolitics andSociety,and Institute or Researchon Poverty

    The basic thesis of this paper is that class, defined within the Marxisttradition as commonposition within the social relations of production,mediates racial differences in income returns to education. That is,class position is viewed as a determinant of the extent to which educa-tion can be transformed nto income, and thus it is hypothesized thatmuch of the commonly observedracial difference n returns to educationis a consequenceof the distribution of racialgroups into classcategories.The results of the study strongly confirm this perspective: the differ-ences in returns to education between black and white males largelydisappearwhen the regression equations are run within class positions.One of the most consistent findingsof researchon racial inequality is thatblack males receive considerablylower income returns to education thanwhite males. Weiss (1970, p. 154) found that, within specific age groups,black males received significantly lower returns to education than whitemales, whether educationwas measuredas years of schoolingor as achieve-ment level. Siegel (1965) found that, net of occupation and region of thecountry, the difference in expected incomes of black and white males in-creased monotonically with education: at less than elementary education,blacks earned$700 less than whites (net of occupationand region) in 1960;at the high school level this increased to $1,400, and at the college level, to$3,800. Duncan (1969) has shown that even after controlling for familybackground,number of siblings, and occupationalstatus black males stillreceive lower returns to education than white males.21 This research was supported by funds granted to the Institute for Research on Povertyat the University of Wisconsin-Madison by the Department of Health, Education, andWelfare, pursuant to the provisions of the Economic Opportunity Act of 1964. I would liketo express my gratitude to Arthur Stinchcombe, Barbara Heyns, Michael Reich, and LucaPerrone for their comments and criticisms on various aspects of this work and to AageSorenson for his skepticism about an earlier draft of this paper. I would also like to thankJames N. Morgan of the Institute for Social Research for making the data for this studyavailable. The opinions expressed herein are the responsibility solely of the author.2 To my knowledge, the only study which claims to present findings different from theseresults is the research of Ross Stolzenberg (1973, 1975). Stolzenberg estimates a rathercomplicated income-determination equation within 67 detailed occupational categories forboth black and white males. He then compares the partial derivatives of income withrespect to education for the equations and finds that in nearly one-half of the occupational? 1978 by The University of Chicago. 0002-9602/78/8306-0003$02.351368 AJS Volume 83 Number 6

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    3/31

    Race, Class, and Income InequalityNone of these studies, however,has controlled for class position as under-stood within the Marxisttradition,that is, the position of individualswithin

    the socialrelations of production.The underlyingpremise of a Marxistclassanalysis is that, while the diversedimensionsof social inequality cannot bereducedto class inequality, class relationsneverthelessplay a decisive rolein shapingother formsof inequality. In the study of incomeinequality, thisimplies that class relations organize the structure of income inequality inthe sense that classposition shapes the ways in which other causesinfluenceincome. If this notion is correct, then an analysis of racial differences inincome that ignores positions within the social relations of production isincomplete.More concretely, if it is true that the returns to educationvarysubstantially across class positions, and if it is true that black and whitemales are distributed quite differentlyacross class positions, then much ofthe racialdifference n returns to educationcouldin fact be a consequenceofthe class distributionof races.This paperwill exploresuch a possibility.OPERATIONALIZING CLASSBefore developing a series of explicit hypotheses about the interactions ofrace and class in the income-determinationprocess,it is necessary to discussbriefly how the Marxistnotion of class will be operationalized n this study.3When non-Marxistsuse the term "class,"it generallydesignatesa group ofpeople who share common "life chances" or marketpositions (Weber 1968,p. 927; Giddens 1973; Parkin 1971, pp. 18-23), common positions withinstatus hierarchies(Warner1960;Parsons 1970,p. 24), or commonpositionswithin authority or power structures (Dahrendorf 1959, p. 138; Lenskicategories the partial derivative is larger for blacks than for whites. Thus, he concludes,"Earlier findings suggesting high within-occupation racial differences in wage returns toschooling . . . were probably artifacts of the gross occupational classifications used. Thesepast findings appear to have been produced by the tendency of black men to be con-centrated in the lowest-paying detailed occupation categories within the major occupa-tional group in which they are employed" (1975, p. 314). The problem with this conclusionis that Stolzenberg uses a natural logarithmic transformation of income, whereas Siegel(1965) uses raw dollars. This means that Stolzenberg is estimating rates of returns toeducation rather than absolute returns. The absolute returns to education within the de-tailed occupational categories may still not have differed significantly between blacks andwhites, but Stolzenberg's results do not demonstrate this. I ran Stolzenberg's equationusing the data in the present study, calculated the partial derivatives for all blacks and allwhites, and discovered that the rates of return for all blacks were significantly greater thanfor all whites (Stolzenberg does not report the results for all blacks and all whites). Stolzen-berg's results thus indicate that these higher rates of return to education for black men as awhole can also be found within about one-half of the detailed occupations held by blackmen. His results do not indicate that the absolute returns for black and white men are thesame within detailed occupations.I For a more detailed discussion of this operationalization of class, see Wright (1976b, pp.137-39). For an extended discussion of the concept of class within the Marxist tradition,see Wright (1976a; 1976b, pp. 20-90) and Poulantzas (1975).

    1369

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    4/31

    American Journalof Sociology1966, p. 95). In contrast to these usages, Marxists define classes primarilyin terms of common structural positions within the social organization ofproduction.In contemporary American society, this means defining classesin terms of positions within capitalist social relations of production.For the purposes of the present analysis, position within the social rela-tions of productionwill be definedaccording to two basic criteria:whetheror not the individual owns his or her own means of production, and whetheror not the individual controls the labor power of others (i.e., supervisespeople on the job). These two criteriageneratefour basic class positions, asillustrated in figure 1. "Ownershipof the means of production" is opera-tionalized by the question, "Do you work for yourself or someone else?"For self-employed individuals, "control of the labor power of others" isoperationalizedin terms of having employees; for individualswho are notself-employed, this criterion is operationalized by the question, "Do yousupervise the work of others, or tell other employeeswhat work to do?"One further distinction will be made within this basic class schema.Clearly, some of the people who are placed in the manager/supervisorcategory are nominal supervisors. This would be the case, for example,for the head of a workteam who serves as the conduit for ordersfromabovebut who lacks any capacity whatsoever to invoke sanctions on his or herfellow workers. Proper managers are thus distinguished from nominalsupervisorsby the question,"Do you have any say in the pay andpromotionsof your subordinates?"It is important to be clearabout the logic underlyingthese classcategories.They are not simply proxies for occupations. "Occupation" designates thetechnical function performed by individuals within the labor process;"class" designates the social relationship within which those technicalfunctions are performed. While, of course, different class positions includedifferent mixes of occupations, every broad category of occupations is

    Ownership of the Means of ProductionYES NO

    Control YES Employers Managers/Supervisorsover thelabor powerof others NO Petty Workers

    Bourgeoisie

    FIG. 1.-Criteria for class position. In the upper left hand quadrant, the term "Em-ployers" is used rather than "Capitalists," since in the present study most of the individualsin this category employ fewer than 10 workers. See Wright (1976a, pp. 35-36) for a dis-cussion of small employers as a contradictory class location between the capitalist class andthe petty bourgeoisie.1370

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    5/31

    Race, Class, and Income Inequalityrepresentedwithin each class category (Wright 1976b, pp. 168-73; Wrightand Perrone1977).HYPOTHESESThe empiricalinvestigation will revolve around six hypotheses.41. Managersas a whole will receivemuchhigher returnsto educationthanworkers.-This basic result has alreadybeen established in an earlierstudy(Wright and Perrone 1975, 1977). The rationale underlyingthis hypothesisis based on an analysis of the specific position of managers within capitalistsocial relationsof production. Specifically,this analysis suggests that withinthe managerial category there will be a strong link between income andhierarchicalposition on the one hand andhierarchicalposition andeducationon the other. The high returnsto educationwithin the managerialcategoryare a result of this double link.First, examine the link between hierarchy and income. The behavior ofall employeeswithin a capitalist organization s controlledby a combinationof repressive sanctions and positive inducements. As one moves up themanagerialhierarchy, however, the balance shifts between these two modesof control.While repressivecontrolsmay be effective in creatingconformityto explicit rules, they are not terribly well suited to generating responsibleand enthusiastic job performance. Because the managerial hierarchyis oneof increasing responsibility (and, in a limited way, increasing power aswell), there will be a tendency for the behavior of higher managers to becontrolledmore exclusively through a structureof inducements.The resultis that managerial hierarchies will be characterized by a steep incomegradient attached to authority position (Tannenbaum et al. 1974, p. 107),even when the education of managersis held constant (Wright 1976b, pp.235-38).Second, examine the relationship between educational credentials andhierarchicalposition. In both the workingclass and the managerialcategory,education is in part a determinant of the value of the labor power of theindividual (orwhat non-Marxisteconomists typically call "humancapital").It is therefore to be expected that both workers and managerswould receivea positive income return to their education. However, among managers,educational credentials serve a second function. In addition to creatinggenuine skills, education also serves as an institutional mechanism forlegitimatinginequalities of power within capitalist organizations.In practice,4 All of the hypotheses that follow center on the relationship of the working class and themanagerial category to racial differences in 'returns to education. Since such a small per-centage of black males are either capitalists or petty bourgeois it is impossible, using thesample available for this study, to examine systematically the interactions of these classpositions with race.

    1371

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    6/31

    AmericanJournal of Sociologythis means that there will be a general tendency for people with lowercredentials not to be promoted above people with higher credentials, andthus there will be a tendency for managerialhierarchiesto be characterizedby rather steep educational credentialgradients (Tannenbaumet al. 1974,p. 112).The combination of this steep education gradient and steep incomegradientassociatedwith hierarchymeans that the managerialcategory as awhole will be characterizedby an especiallyhigh return to education.Thatis, in addition to the direct return to education resulting from increases inthe market value of laborpower (skills), which both managersand workersreceive, managersreceive an additional increment of income for education,stemming from the link between the legitimation function of educationwithin hierarchiesand the use of income as a control mechanism withinauthority hierarchies.(For a moredetailed discussion of this interpretation,see Wright and Perrone1977; Wright 1976b,pp. 105-10.)2. Black males will be moreconcentratedn the workingclass than whitemales.-While we will not explorethe actual mechanisms by which individ-uals are sorted into class positions, it is nevertheless predicted that oneoutcome of this sorting process is that blacks will be more heavily concen-trated in the workingclass than whites.3. Whenclass positionis ignored, blackmales will receive owerreturnstoeducation than white males.-This is the standard finding in sociologicalstudies of racial differences in returns to education. It will be formallytested in orderto show that the usual results hold for the data used in thepresent study.4. Within the workingclass, the returns to education or black and whitemaleswill be muchmoresimilar than or all blacksand all whites.-If at leastpart of the overalldifference n returns to education for blacks and whites isa consequence of the class distributionsof the two racial categories, then itwouldbe expectedthat within the workingclass itself the returnsshould bemuch more similar. While white workers may be in relatively privilegedworkingsituations as compared with black workers,neitherwhite nor blackworkers occupy positions of authority (by definition) and, thus, neitherreceive the legitimation increment of returns to education discussed inhypothesis 1.5. Within the supervisorcategory,the returns to education or black andwhite males will be more similar than for all blacks and all whites.-Theargumenthere is essentially the same as that presentedin hypothesis 4. Tothe extent that the overall racial differencesin returns to education are aconsequence of class distribution, within a single class position-in thiscase, the very bottom level of the managerialhierarchy-the returns forblacksandwhites should be muchmore similarthan forall blacks and whites.6. Within the managerialcategory,blackmales will havelower returnsto1372

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    7/31

    Race, Class, and Income Inequalityeducation hanwhite males.-The argument in hypothesis 1 concerningthehigh returns to education of managers hinges on the dual link betweeneducation and hierarchical position and between income and hierarchicalposition. If a particular category of managersis highly concentrated at thebottom of the authority structure, then this education-hierarchy egitima-tion mechanismwill tend to be attenuated. Although no data are availablein the present study concerning the hierarchicaldistribution of race withinthe manager category,white males can be expectedto be much moreevenlydistributed throughout the hierarchy than black males. If this is the case,then within the managerial category the returns to education for blackmales should be considerably smaller than returns for white males.DATAThe data for this study come from the eighth wave of the Panel Study ofIncome Dynamics (Institute for Social Research 1975) conducted by theInstitute for Social Research (ISR) at the University of Michigan. Whilein the original year of the panel study the sample was a randomsample of5,000 households, by the eighth year of the study, because of successivenonresponses, the sample was no longer genuinely random.5While thisprobably will not seriously affect the regressionresults, it 'certainly mayaffect hypothesis 2, concerningclass distribution of races. Two other datasets, the 1969 Survey of Working Conditions and the 1973 Quality ofEmployment Survey (both from ISR), will thus be used for hypothesis 2.6Throughout the analysis, the sampleswill be restrictedto active participantsin the labor force.EQUATIONSIn order to test the hypotheses about class andrace interactions with returnsto education, two regression equations will be estimated for each of therace-class groups being compared7i Two things need to be noted concerning nonrandomness of the sample in the Panel Studysf Income Dynamics. First, whenever an individual left the original household in the study(because of divorce, high school graduation, etc.), the "split-off" was also included in thesubsequent years of the panel. Thus, the sample is not particularly skewed on age dis-tribution. Second, a fairly complex system of weights has been devised to correct, at leastpartially, for nonrandomness in nonresponse. Thus, the regression results in the presentstudy are probably reasonably reliable in spite of the nonrandomness of the sample.6 A discussion of these data sets can be found in Wright (1976b, pp. 132-35).7 The significance of the slope differences between groups will be assessed using the con-ventional dummy-variable interaction model (Kmenta 1971, pp. 419-23). This meansthat, if we are comparing the education slopes for two groups, the t-test would be:

    t = (B11 - B12)/(S2Bll + V2S2B12)/(V1 + v2)1373

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    8/31

    AmericanJournal of SociologyIncome a + b1Education, (1)andIncome a + b1Education+ b2Occupational tatus + b3Age

    + b4Seniority+ b5Father'sStatus + b6Father'sEducation (2)+ b7ParentalEconomic Condition+ b8AnnualHours Worked.

    While the present study will not investigate the more complex structuralequationmodel which underlies the second equation, this equation is never-theless important in assessing the extent to which the class interactionsobserved in equation (1) may be consequences of the class distributions ofthe various control variables in equation (2). Thus, for example, if inhypothesis 1 the greater returns to education for managers were entirelydue to the occupational status distribution across class categories, then,when occupationalstatus is includedin the equation, the differentialreturnsto education between classes should be substantially reduced. All of thecontrol variables in equation (2) are to a greater or lesser extent eithercauses or consequencesof education, and all of them can also be plausiblythought to vary with class position. By controlling for these variables inequation (2), we will be able to see if the differences n education coefficientsbetween the various groupsbeing comparedcan be considereddirect conse-quencesof class and race.8

    where B11s the education coefficient for group 1, B12s the education coefficientfor group 2,S2B11 and S2B12 are the standard errorsof the coefficients for groups 1 and 2, respectively, andvPand v2are the degrees of freedom for groups 1 and 2, respectively. The denominator inthis t-test is the standard error of the education X class-dummy interaction term in theusual dummy-variable interaction model. It can be computed equally well from theseparate regression equations for the two groups using the above formula.8 A brief comment about the kind of information contained in these regression equationsmight clarify the logic of eq. (2). These equations describe characteristics of positions in asocial structure. While the data are tagged onto individuals, the equations themselves donot adequately describe individual income-attainment processes, but only the income-determination process for certain structural locations. Since we have no data on how indi-viduals are sorted into structural positions, and since education obviously may play animportant role in this sorting process, the education coefficients in the various equationsmay say very little about individual returns to education. What they describe is the returnto education which characterizes the position "worker," "manager," "black worker," etc.Now, positions within a social structure can be characterized along many dimensions.Marxists assert that the critical dimension of position within social structures is class rela-tions. Other social scientists have argued explicitly or implicitly that occupational status orthe social backgrounds selected into positions are the key dimensions. Equation (2) ineffect holds constant a variety of positional dimensions which have some claim to beingimportant determinants of the relationship between education and income. To the extentthat class and race positions differ on education coefficients even when all of these factorsare held constant, these coefficient differences can be considered direct, unmediatedconsequences of the class-race positions themselves. (See Wright and Perrone 1977, p. 38,for a further discussion of this issue.)1374

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    9/31

    Race, Class, and Income InequalityVARIABLESThe variables to be included in the analysis are measured as follows.

    1. Income is measured by total annual taxable income received by theindividual. In addition to wage and salary income, this variable includesincome from assets, interest, and other sources of unearned income. Bothregression equations were also estimated for three other income variables:annualearnings,an estimate of "permanent ncome" (an averageof incomeover the previous seven years), and imputed hourly wage (wage and salaryincome divided by total annual hours worked). In none of the comparisonsof racial differences n returns to educationdid the resultsdiffersignificantlyfor these alternative income variables (see Wright 1976b,pp. 328-39).2. Education s operationalizedby a quasi-credentialscale in which:0 = no schoolingor illiterate,1 = less than elementary school,2 = elementary school,3 = some high school,4 = completed high school,5 = high school plus some nonacademic training,6 = some college,7 = college degree, and8 = graduate training.

    3. OccupationalStatus is measured by the standard Duncan SEI scores.These scores were available only for the seventh wave of the panel study(1974), while the class position questions were asked only in the eighthwave. This means, in effect, that we have the individual's occupationalstatus score at the beginning of the year in which income was earned andhis class position at the end of the year (i.e., the beginning of the followingyear).94. Age is included in the regression both as a rough control for cohorteffects and as a rough measure of years of experiencein the labor market.5. Senioritydesignates the number of years the individual has worked forthe same employer, or the number of years an individual has been self-employedin the same business.6. Father'sStatus is measuredby the average SEI score for the father'sgross-census occupation category. While this is clearly a much weakervariable than a status score based on the three-digit occupation classifica-tion, it is the best available from the panel study data.I One other minor point concerning the status variable needs mentioning: about 6% of thesample represents "split-offs" in the 1975 year of the survey, that is, those who left ahousehold after the 1974 survey and set up a new family unit. For these people a three-digit occupation classification was not available, and thus in these cases the Duncan scoreis based on the average SEI value for the gross-census occupation categories.

    1375

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    10/31

    American Journalof Sociology7. Father's Education is measured by the same scale as respondent'seducation.8. Parental Economic Condition is a scale reflecting the respondent'ssubjective perceptionof parents' economic status in which:

    1 = parents generally poor,2 = parents generally about average, and3 = parents generally well-off.9. Annual Hours Worked s a product of the number of weeks worked inthe previous year and the averagenumberof hours worked per week.The means and standard deviations for each of these variables for each

    of the race and class groups included in this study are given in table 1.RESULTS

    Hypothesis1:Managersas a wholewill receivemuchhigherreturns oeduca-tionthanworkers.-Table 2 indicates that, in the simple regressionof incomeon education, workers receive $851 for each increment in education whilemanagers/supervisorsreceive $1,689. When the various control variablesinequation (2) are added, the returns for workersare $655, while the returnsfor managers/supervisorsare $1,169. In both cases, the difference n returnsis significant at the .001 level. This hypothesis is thus strongly confirmed.The theoretical rationale for this hypothesis centers on the relationshipsamong hierarchicalposition, education, and income. If this rationale iscorrect,then it would be expectedthat the high returns to educationamongmanagers/supervisorswould disappearif we examineda single hierarchicallevel within the managerial hierarchy. In a limited way, we can test thisproposition by examining separately the returns to education among meresupervisors (positions which involve no say over pay and promotions) andproper managers,since mere supervisorscan be considered the bottommostlevel of the hierarchy. As can be seen in table 2, mere supervisors differhardly at all from workersin returns to education (although they receivemore income at every level of education), whereasproper managersreceivemuch higher returns.Oneobviousobjectionto these resultsis that the truerelationshipbetweeneducationalcredentialsand incomemight be curvilinear.Thus, what appearto be slope differencesbetween workersand managers might in fact simplybe a consequenceof the two linear regressionsreflectingdifferentparts of asingle, nonlinear credential-income unction. Figure 2 presents the relation-ship between the linearregressions n table 2 and the mean incomesfor eacheducationallevel for workers, managers,and supervisors.Table 3 indicatesthe R2forequation (1) usingthe credentialscale and using a seriesof dummyvariables to represent the individual levels of the scale. Both the visual1376

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    11/31

    ~~nrz00 cE (O. ,O U)cr +CrzwNzCl) k , N N U

    .N w - U I m x co -) x (O, - om O m 40B U ) CN U ..s...*

    z ~ I

    Cl X I\_sy NsN

    0 Iz I

    X~~~~~~~~~~ _ - Cq t- m in cl in 00 'IC ooq oo t x Ct-4otbo

    I?4' Cd (O., o ssNo- (= o- (O. o s + m (O. c- o x Cr O. U-)~ Cd ..... ...........

    ClOz IO *_ J ot11m tf)\te. ZO % O @N

    Cd I r sSe?OO 0 O?

    H ol~ I ~j1 o24 12)it~~~~~~~~~ CN cI 00 m C) 00 m e m oo m +o CZ4 - 4) J 94

    Cl)

    uD~~~~~~~~~~~~~~~~~~~ 0

    4,o U)c)I I" Ocroo oo usoo'O r> r oo? Cd . . ..Nr O Co .On c r o r r A4cd -,1O m C' M', 6 !4 ,4 00 6) o0 -,1 O0 m m oo (= t- CD ,O oo

    * 0

    . . ~~~~~. . . . . . . . . .

    ! . :~~~~~~~~~~~~l . o00'O , C -~C - O-q 0 C) C C

    o !~~~~~~~~~~~cI,- -l) o t-- -; - - -; - - - -

    $n4 *n d. . ..

    1377

    ~~2124 .~~ .- >2 . .) 2122.) 2.) ~~~~~~~c.2.) C0 C~~~~~~2 ~~~~~~~~ ~ s4 MzC1377~~~~~~~~~~~~~~c

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    12/31

    CN CN

    (O" Ul) 00cl0 . . . C-q. 0

    t- 00C)0 1 . . . 00 U')0cdu 0 cq0UCd CN

    W 4,Cd cq oo0, cd cli41Cd

    4, 00Cdo . . .cn U)Z

    00 . . .0

    Cd

    P4z .0

    41Cd Cd4, . . . . . . 00

    00 C14m cl-i 00 cl-i cl-i. . . . . . . . . . . . . . .-- t- Ul)cl-i a, m 00 m -It =in t- a, 00 m o Ul) Ul)00 00P4 'Cod

    00a, 004.30 00

    0u.. . . ... . . .cd--j ), : w 40 Cd C/) :s cn cn > cn C/)w cn pq pq ce pq pq pqCd o ce

    cn

    13 78

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    13/31

    0CN

    Cdcd0 0 CdU .4 En

    > Ww WOC) C)d 0 v CN0 Cd

    041En o41 o 41cdo Cdr. ad u 0 Ew u 0

    u Cd CNcod

    . . . CN 00 0cd .1Cdcd 0

    0 04,'Cd -o >,, 41InCd CdCd0 In41u

    'In.M,. 1T4 0 Cd 0Cd

    In4 W . . .Cd

    P4cd t-oo C14Cd cc,4d

    0Cd C140000

    rA+-j - .

    En Cd 60 0 0-&-,0od 0 Cd0 00 >I" 4-10 0 s.,0

    ed,;: =$ 0cd q cdC) 'WII- 4.1Cd0r4En -M ..", cd1-4 Di Uw 0 EnIz -.- . w cd0 cd to cn .0M 0 Wd o 0

    ce0

    0

    1379

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    14/31

    AmericanJournalof Sociologyinspectionof figure2 and the closenessof the R2using the dummy variablesand a single scale in table 3 indicate that the relationship of credentialstoincome is reasonably linear within each class category. The differencesinslopes thus cannot be interpretedas artifacts of a single curvilinearrelation-ship between income and credentialsfor both classes.Hypothesis2: Blackmaleswill bemoreconcentratedn theworkingclassthanwhite males.-Table 4 gives the class distributionof black and white malesbasedon an averageof the 1969Surveyof WorkingConditionsand the 1973

    $26,000 -LINEAR REGRESSIONMEAN INCOME FOR EACH

    $24,000 - EDUCATION LEVEL /, MANAGERS

    $22,000 /$20,000 -

    $18,000 -

    o $16,000oZ / SUPERVISORS_ $14,000 // WORKERS

    < $12,000-j

    z $10,000zz$8,000

    $6,000 ,'

    $4,000

    $2,000

    l l l I . l l l0 1 2 3 4 5 6 7 8

    0 0 Ea) 0) a -C co-c E E E Efc,, s Ic E o0a a' a- 2 0 0 0 0Z 0t)LL L c/)I I I Z (10 ) (JH

    EDUCATIONALCREDENTIALSFIG. 2.-Relationship of income to education for workers, supervisors, and managers

    1380

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    15/31

    Race, Class, and Income InequalityTABLE 3

    COMPARISON OF EXPLAINED VARIANCE USINGEDUCATION SCALE AND EDUCATION DUMMIES(Annual Income = Dependent Variable)

    R2 UsingEducation R2 UsingCredential EducationScale DummiesEquation for:Workers ....... ........ .066 .069Managers/supervisors .. .135 .154Supervisors ............ .065 .065

    Managers . .155 .180Quality of Employment Survey and the distribution for the panel study.Because of the problemsof nonrandomness n the panel study, the averageof the two earliersurveys is undoubtedly a more accurate estimate of theactual class distribution of races. It is clear from these results that blackmales are indeed more concentratedin the working class than white males:61% of all black males compared with only 40% of white males fall withinthe workingclass.'0

    Hypothesis3: Whenclass positionis ignored,blackmales will receive owerreturns to education than white males.-Table 5 presents the regressionequationsforall blacksand all whites, for black and white workers,for blackand white supervisors,and for black and white managers. Figure 3 graphi-cally presents the results for the simple regressionof income on education.Table 6 presents the statistical tests for significance of the differencesinreturns to education for various race-class comparisons. In the simpleregression of income on education, white males receive $1,419 for eachincrement in education; black males receive only $860. When the controlsin equation (2) are added, the returns among white males decrease to $1,147and among black males to $614. Both of these differencesare significant atthe .001 level. Thus, as in most studiesof black-whitedifferences n returnsto education, black males as a whole do receive lower returns than whitemales as a whole.Hypothesis4: Withintheworkingclass, thereturns oeducationor blackandwhitemaleswill bemuchmore imilarthan or all blacksand all whites.-In thesimple regressionof income on education, the returnsfor all blackmales areless than one-half of the returns for all white males; within the working10 If anything, these figures underestimate the proportion of the black male population inthe working class, since unemployed persons are excluded from both the Survey of WorkingConditions and the Quality of Employment Survey. If one assumes that most unemployedblack males belong in the working class, then the actual proportion of black males who areworkers would probably be closer to 70% or 75%, and of white males closer to 45% or 50%.

    1381

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    16/31

    OC) C14 1 t- V') IC,4 t- I V') -4 1 : CCd

    Cd t- Cdoo "o -Cd

    Cdpq

    t- in Cd

    CdCd

    t-1-4 OC) 0cn C14 OC) Ch

    P40 cn

    CdP"u Cd0

    4-10

    ootw

    pq CdP40>4 Cd ONcn 0, clq 00cn

    Cd>4 Zu 0 PI

    >4 P40cnEcn Z H-4 cd In C'4 in CN

    0 En >4z 00 4

    P, 0EnCd

    00cdNO 4.1Cld Cd0Cld

    lu -04 0 EnW 414 Cd >0cld Cd

    U; En 0 rn NOEn S-4 bDS-4 S-4 -4 W S-4 cd0 > bID Cd 4.,0Cj rn 0 +J0 > 0- S-4 1214Cld Pjocn

    1382

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    17/31

    Z~~~~~~~~0 o o o

    ? d I0mC

    X11 N .,, mS1. ... om .. , m0

    0 11. I

    Y cd 0 cnI : : : ? ': : : ??: : : ' I

    Zd C' C N oo oo

    ms~~~~~ ~N o o inNT O+% cO c 00 TS~~~~~~~~~~0

    11 1 1~~~~~~~~~~~~C' d n Csn ) CZ

  • 7/27/2019 Race, Class and Income Inequlity

    18/31

    00

    Cd00 Cq

    cn 0mm t-Cd 0 10ou

    Cdcn

    cdP, dcd

    Cdu 00 0 . . .cd =1

    00 00 lzt

    Cdr.En0Cd Cd

    Cd

    00 00En Cd 00

    Enlt 0 00 C,4 "oCd 0

    rn t- 0 t- 44S-4 U-) rn >w -4 S-4> wga, 11 .. . . . .. . . .cd, C W,-., I, Cd C cd cd4., Cq 4- C4- -cn cn"n-, q pq 11-Sq pq 6,Pq PP &Pq , PP pq , PP-4Cd Cd

    pq pq

    1384

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    19/31

    It:$

    C13w-M cnw 0 . . . t- o,O :j CN 00 "t0 0

    0 :t. -0 . . 00 c --.(0 00 Cdtn

    Cd00 tn ON OCd 00d . . .cd

    0 ciC,)

    --loo0cd

    Cd WT, cdtn

    Cd00 CN m 0

    r. Cdo0cn cncn0

    00 Cd

    C-dr.0 0 cnCd 06d . . .ul o 75tncn0

    bzWmC"lm in O m

    Cd 0 00 00 CdtnInWIce

    00 U-) OCNCd.45od o 0

    rn rn 41Cd nCd aj od Cd4-1 tn

    C/)I.-I pq 4)q IU-11q pq 12Spq 4 &Pq pqCd4pq 0

    1385

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    20/31

    ALL WHITES

    $1 5,000ALL BLACKS

    Lu0z $10,000-j

    zz$5,000

    I I I IIIII0 1 2 3 4 5 6 7 8EDUCATION WHITE MANAGERS

    BLACK MANAGERS

    L' $1 5,000 - WHITE SUPERVISORWHITE WORKERS

    z ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~BAKSPRIO_ / ~~~- -- BACK WORKERVSO

    $5,000

    0 1 2 3 4 5 6 7 8EDUCATIONFIG. 3.-Returns to education for blacks and whites in different class positions

    1386

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    21/31

    Race, Class, and Income InequalityTABLE 6

    COMPARISON OF RETURNS TO EDUCATION FOR BLACKAND WHITE MALES ACROSS CLASS CATEGORIES ANDWITHIN CLASS CATEGORIES

    RETURNS TO EDUCATION

    Eq. (1) Eq. (2)All black and white males:Slope difference ............... 959 533t-value of difference ............ 8. 1* 3.6*Black slope as % white slope..... 49 54% difference in eq. (1) eliminatedby the controls in eq. (2) ... . ... 44Workers:Slope difference ............... 192 7t-value of difference ............ 1.8

  • 7/27/2019 Race, Class and Income Inequlity

    22/31

    American Journal of Sociologyin the multiple regression equation. The expectation was merely that,because of restrictions of blacks to lower levels of the authority hierarchy,the hierarchical-promotionmechanism would be blunted among blackmanagers, and thus the returns to education would be less among blackthan among white male managers.There was no a priori expectation thatthose returns would disappearentirely in equation (2).One possible clue to these results might be found in the occupationaldistribution among white and black managers (table 7). As would beexpected, black male managers are considerablymore concentrated thanwhite male managers among unskilled and semiskilledmanual occupations(38.5% comparedwith 13%, respectively). What is somewhatsurprising sthe much higherproportion of black managerswho are teachers comparedwith white managers (13% comparedwith 2%, respectively). Expressed ina differentway, nearly 60% of the black managersin professionalor techni-cal occupations are teachers as compared with less than 10% of whiteprofessional-technicalmanagers. Remember, these are proper managers,people who, unlike mere supervisors,state that they have some say in thepay and promotions of their subordinates. This implies that teacher-managers either occupy administrative positions within their educationalinstitutions or direct researchprojects in which they authorizethe pay andpromotions of research staff (all but one of the teacher-managerswerecollege or university teachers).If the regressionsin table 5 are rerun excludingteachers from the man-agerial category, the results conform much more to expectation (table 8).Black male managers still have significantly lowerreturns to education thando white male managers, but the returns are not nearly as small as returnsin the regressionsthat included teachers.I cannot offer a particularly coherentexplanationfor why the presenceof

    TABLE 7OCCUPATIONAL DISTRIBUTION AMONG MANAGERS

    FOR BLACK AND WHITE MALES (%)

    White Males Black MalesProfessional, technical, and kindred ..... 24. 6 22. 5

    Professionals.. ..................... 19.3 9.3Technicians........................ 3.0 0Teachers ............ .............. 2.3 13.2Managers and administrators ..... ..... 35.2 25.0Sales ............................... 4.7 .8Clerks .............................. 1.9 1.7Craftsmen and kindred....... ......... 20.5 11.5Operatives, laborers, and miscellaneous.. 13. 1 38.5

    Total ........................... o100 100

    1388

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    23/31

    Race, Class, and Income InequalityTABLE 8

    RETURNS TO EDUCATION FOR BLACKS AND WHITESWITHIN THE MANAGER/SUPERVISOR CATE-GORY, EXCLUDING TEACHERS

    TOTAL ANNUALINCOMEEq. (1) Eq. (2)

    Supervisors:White males 732 761(182) '208)Black males 506 934(217) (350)Difference 226 -173t-value .1.1

  • 7/27/2019 Race, Class and Income Inequlity

    24/31

    AmericanJournalof Sociologyselect people into the managerial/supervisorycategory and the mechanismswhich regulate the promotion patterns in managerialhierarchies.Racismwould affect the distribution of races within authority structures in twogeneralways. First, as in the sorting processfor occupations,various formsof racial discriminationaffect access to the mechanismswhich sort peopleinto the managerialhierarchy (educational credentials, connections, etc.).Since, as was argued earlier, people with lower credentials will tend not tobe promotedabove peoplewith highercredentials, the result will be a higherconcentrationof blacks at lower levels of the managerialstructure and ahigher concentration of blacks in the working class. Second, and perhapsmore important, because of the necessity to legitimate the social relationsof dominationembodiedin managerialhierarchies,racism will tend directlyto prevent the promotion of blacks above whites. Of course, this does notmean that blacks will never be promoted above whites. Particularlywhenstrong political struggles against racism occur, corporations and bureau-craciesmay see the imperativesof legitimation as requiringthe acceptanceof some blacks into "token" positions of authority within managerialstructures. But in the absence of such struggles, it would be expected thatthe logic of hierarchical domination within capitalist production relationsand the necessity of legitimating that domination would generate racistpatterns of recruitmentinto and promotionup managerialhierarchies.The above argument about recruitment and promotion presupposestheexistence of racism. Given the presenceof intense racist beliefs, it is easy toexplain why blacks will not be promoted above whites within hierarchies;but this begs the question about the existence of racism in the first place.It is beyond the scope of this paper to attempt a systematic account of theoriginsof racismand the socialprocesseswhichreproduce t in contemporaryAmericansociety. What I will do is very brieflyindicate the essential thrustof a classanalysisof the roleof racismin Americancapitalismand showv owvthe present study relates to that analysis.A common mistake made by Marxists in analyzing racism is to assumethat all forms of racial discriminationare unequivocally functional for thecapitalist class. This is similar to analysesof the capitalist state which arguethat every policy of the state is orchestratedby the capitalist class to serveits interests.Such"instrumentalist"views of the state andideologyminimizethe intensely contradictory character of capitalist society.11 Capitalismsimultaneously undermines and reproducesracism, and it is essential todisentangle these two tendencies if one is to genuinely understand therelationshipbetween class and race in contemporarycapitalism.One of the basic dynamicsof capitalist developmentstressedby Marx, as11 For a critique of instrumentalist views in Marxist theory, see Gold, Lo, and Wright(1975) and Esping-Anderson, Friedland, and Wright (1976).1390

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    25/31

    Race, Class, and Income Inequalitywell as many non-Marxist theorists, is the tendency for capitalism totransform all labor into the commodity labor power and to obliterate allqualitative distinctions between differentcategories of. abor. Fromthe pointof view of the capital accumulationprocess, the more labor becomesa purecommodity,regulated by pure marketprinciples unfettered by personal tiesand ascriptivebarriers, the more rapidlycan capitalism expand. In terms ofthe logic of accumulation developed by Marx in Capital, there will thereforebe systemic tendencies in capitalism to reduce racial discriminationin thelabor market and to treat black labor power as identical with any otherlabor power.But this is only one side of the story. Capitalismis not just a system ofcapital accumulation; it is also a class system in which workers struggleagainst capitalists, both over their condition as sellers of labor power andpotentially over the existence of the capitalist system itself. Whereas theessential dynamics of accumulation may lead to an underminingof racialdifferences in the labor market, the dynamics of class struggle tend tointensify racism. To the extent that the working class is divided along racialand ethnic lines, the collective power of the working class is reduced, andthus the capacity of workersto win demands against capital is decreased.The result will be an increasein the rate of exploitation of both white andblack workers,although the effects may well be more intense for blacksandother minoritiesthan for whites.12As a divide-and-conquer trategy, racismthus serves the interests of capitalists, both as individuals and as a class.13We thus have a basic contradiction: capitalism tends to undermine allqualitative distinctions between categoriesof labor, but the capitalist classneeds those qualitative divisions for its own reproduction as the dominant12 While Marxists have often claimed that racism hurts white as well as black workers,until recently there have not been any systematic empirical investigations of this proposi-tion. Two recent studies deal directly with this question. Reich (1971, 1973) shows that, inthe 50 largest SMSAs, the greater the racial inequality in median family earnings, thegreater the inequality of earnings among whites (as measured by the Gini coefficient forearnings among whites) and the weaker the level of unionization. Szymanski (1976) showsthat, for the 50 states, the greater the inequality between black and white median earnings,the lower the median earnings of white males and the greater the inequality among whites(again, measured by Gini coefficients). Both of these sets of results indicate that racism, individing the working class, leads to an increase in exploitation of all workers.13 The analysis of racism as a divide-and-conquer strategy has been perhaps the centraltheme in Marxist treatments of the subject. Marx emphasized this issue in his variousdiscussions of the "Irish Question." In 1870, for example, he wrote, "The English bour-geoisie has not only exploited the Irish poverty to keep down the working class in Englandby forced immigration of poor Irishmen, but it has also divided the proletariat into twohostile camps.... In all the big industrial centers in England there is profound antagonismbetween the Irish proletariat and the English proletariat. The average English workerhates the Irish worker as a competitor who lowers wages and the standard of life. . . . Thisantagonism among the proletarians of England is artificially nourished and supported bythe bourgeoisie. It knows that this scission is the true secret of maintaining its power"(Marx and Engels 1972, p. 162).

    1391

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    26/31

    AmericanJournal of Sociologyclass.14Both forces operate. The actual balance between the two dependsupon a variety of historical factors. For example, under conditions ofextremeshortagesof labor, obstacles to labor mobility in the labor marketare likely to be rather costly to individual capitalists, and thus it would beexpected that racial barrierswould be more rapidly eroded by imperativesof accumulation.On the other hand, when the supply of labor is relativelyabundant and when individual differences between laborers make littledifference o productivity (because of routinization, automation, etc.), thosestrictly economicimperatives are likely to be weaker. The extent to whichracial or ethnic divisions within the working class are being deepened oreroded in a given capitalist society cannot therefore be derived directlyfrom the abstract theory of capitalist economic development. It is only whensuch abstract theory is linked to specific political and ideological develop-ments that it becomes possible to assess the real dynamics of racism in agiven society.While the present study does not deal with this historical process, thedata in the research nevertheless can be related to both sides of this con-tradictory tendency within capitalism, that is, both to the perpetuation ofimportant racial divisions within the working class and to the commonsituation of all workers as workers,regardlessof race.The data presented in figure 3 and table 5 clearly indicate that, whileblack and white workers receive similar returns to education,black-workerincome is less than white-worker ncomeat every level of education.Onewayof assessingthis gap in income is to see what the expected difference n in-comebetwveen blackand white male workerwould be if they both had someintermediate value on the independentvariables included in the equation.In table9 this gap is calculated at levels of the independentvariableshalfwaybetween the means for each group in the comparison being made.15As canbe seen from this table, the income gaps between races are large and sta-tistically significantforboth regressionequationswithin each class category.14 Both Marxist theory and neoclassical economics recognize that the inherent economiclogic of capitalism is to reduce progressively economic divisions between races, in terms ofboth income and occupation. The difference (in these terms) between the two perspectivesis that neoclassical economics treats capitalism solely as an economic system and ignoresthe fundamental class antagonisms within that system. Thus, the political and ideologicalimperatives of controlling the working class play no role in the theory. Instead of seeing therelationship of capitalism to racism as an intrinsically contradictory process, neoclassicaleconomics typically treats racism as a problem of individual "tastes" for discrimination onthe part of employers and workers (see Becker 1971, pp. 13-18).15Thus, in the simple regression of income on education for the comparison of black andwhite workers, the income gap is assessed at a value of education equal to (Eblack worker +Ewhite worker)/2. The statistical significance of this income gap can be tested in a way exactlyanalogous to the test of slope differences, only in this case a t-test is performedon the differ-ences between constant terms adjusted to the appropriate values of the independentvariables (see Wright 1976b, pp. 55-57).1392

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    27/31

    Race, Class, and Income InequalityFurthermore,the addition of the various controls in the multiple regressionequations reduces the total difference n mean incomes between races withinclasses by no more than 50%, indicating that a substantial part of thedifference in mean incomes between races within classes should probablybe directly attributed to racial discrimination. In terms of Marxist theory,these results strongly suggest that black workers are exploited at a higherrate than white workers and that racism has generated real, material divi-sions between races within the working class.'6

    TABLE 9INCOME GAPSa BETWEEN RACES WITHIN CLASS CATEGORIES

    Mean Income(.. .......$) Eq. (1)b Eq. (2)bAll respondents:Gap in income .............. ......... 5,308 3, 698 1,868Black expected income as % of white ex-pected income ...................... 64 73 85% difference in means eliminated by con-trols............................... ... 30 65t-value of gap ........................ ... 16.4* 8.1*Workers:Gap in income .. ........... .. 2,870 2,203 1,428

    Black expected income as % of white ex-pected income 75 80 86% difference in means eliminated by con-trols .............................. ... 23 49t-value of gap ............. ............ ... 11. 9* 8.9*Supervisors:Gap in income .3, 872 2,896 2,140Black expected income as % of white ex-pected income 69 76 82% difference in means eliminated by con-trols............................... . .. 25 45t-value of gap ... . . ... 8.6* 6.8*Managers:Gap in income .... ................. . 5,051 3, 707 3, 011Black expected income as % of white ex-pected income ...................... 72 79 83% difference in means eliminated by con-trols .............................. ... 27 40t-value of gap . ..................... ... 8. 0* 6.8*

    a Income gaps represent he difference n expected incomes for two groups evaluated at a level of the inde-pendentvariables n the regression qual to the averageof theirrespectivemeanson the independentvariables.b Independent variables in eq. (1) = education only; independent variables in eq. (2) = education, age,seniority, background,occupational status, and annual hours worked.* Significantat the .001 level (one-tailed test).

    16 In order to interpret these results as indicating a higher rate of exploitation of blacklabor, it is necessary to assume that two workers who have the same values on all of thevariables in eq. (2) will have essentially similar complexities of labor (i.e., embodied laborin their own lab or power) and intensities of labor (i.e., pace of work within the laborprocess). Since eq. (2) contains annual hours worked, if we accept the above two assump-tions then two workers who have the same values on all the independent variables willproduce the same amount of total value in a year. Any difference in their incomes would1393

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    28/31

    AmericanJournal of SociologyUltimately, the political thrust of the Marxist theory of racismhinges onthe other side of the contradiction:that, in spite of the divisive character

    of racism and in spite of the material differencesbetween black and whiteworkerswhich racism generates, workers of all races nevertheless share afundamental class situation and thus share fundamental class interests.'7The central findingof this study-that black and white workershave verysimilar returns to education-reflects this commonclass situation. Table 5also indicatesthat black andwhite workershave very similarincomereturnsto occupationalstatus (whereaswhites taken as a whole have significantlygreater returns than blacks taken as a whole), which again reflects thecommonalityof their class position.Furthermore, t is easy to show that the income gaps between black andwhite workers, while significant, are much smaller than the gaps betweeneither black or white workersand small employers. Figure 4 indicates theexpected incomesof each race and class category assessedat the level of theindependentvariables of white employers.'8The unadjusted mean incomeof white workers is less than one-half that of white employers; the meanincome of black workers,on the other hand, is 75% of white workers'meanincome. In absolute dollaramounts, the mean white workers' ncomeis over$14,000 less than the mean white employers' income, whereas the mean

    black workers' income is only $2,900 less than the mean white workers'then reflect differences in the costs of reproduction of their labor power and thus differencesin rates of exploitation. Two factors could undermine this conclusion. First of all, thereproductive costs of labor power are not represented simply by wages, but by fringe bene-fits, state subsidies for education and other services, etc. To the extent that such additionalelements of the costs of reproducing labor power are themselves correlated with earnings,we are, if anything, underestimating the differences in rates of exploitation by looking ex-clusively at direct income. Second, if black workers as a whole are overqualified for the jobswhich they hold, wage differences could reflect in part social waste of potential surplusvalue rather than superexploitation of black labor in the technical sense of the concept. If acollege graduate works on an assembly line, he or she produces no more surplus value and isno more exploited than a high school dropout in the same production process. The com-plexity of the college graduate's potential labor is being socially wasted in such a situation.In the strategy used to compare black and white exploitation above, however, this situationwould appear as a greater rate of exploitation of the college graduate. There is no way, inthe present data, to differentiate between underemployment and more intense exploitation.For a much fuller discussion of the relationship between econometric models of income de-termination and the Marxist concept of exploitation, see Wright (1976b, pp. 120-31).17 "Fundamental" class interests refer to interests defined across modes of production (i.e.,interests in capitalism vs. interests in socialism), whereas "immediate" interests refer tointerests defined within a given mode of production. Black and white workers may well haveconflicting immediate interests under certain circumstances (as do many categories oflabor within the working class) and still share fundamental class interests. For a discussionof the importance of the distinction between fundamental and immediate interests, seeWright (1978, chap. 2).18 This procedure is basically similar to the familiar cross-substitution technique employedby Duncan (1969) and others.1394

    This content downloaded by the authorized user from 192.168.52.72 on Wed, 21 Nov 2012 20:07:51 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 Race, Class and Income Inequlity

    29/31

    Race, Class, and Income Inequality

    $ WHITE EMPLOYERS 100 %$25,000 Z

    $20,000 - t)-LJ| WHITE MANAGERS 74% < Z2i ~~~~~~~~64Lw


Recommended