Prize in Economic Sciences in Memory of Alfred Nobel 1992 "for having extended the domain of microeconomic analysis to a wide range of human behaviour and interaction, including nonmarket behaviour"

Gary S. Becker

USA

University of Chicago

Chicago, IL, USA

b. 1930

Economic Discrimination Another example of Becker's unconventional application of the theory of rational, optimizing behavior is his analysis of discrimination on the basis of race, sex, etc. This was Becker's first significant research contribution, published in his book entitled, The Economics of Discrimination, 1957. Discrimination is defined as a situation where an economic agent is prepared to incur a cost in order to refrain from an economic transaction, or from entering into an economic contract, with someone who is characterized by traits other than his/her own with respect to race or sex. Becker demonstrates that such behavior, in purely analytical terms, acts as a "tax wedge" between social and private economic rates of return. The explanation is that the discriminating agent behaves as if the price of the good or service purchased from the discriminated agent were higher than the price actually paid, and the selling price to the discriminated agent is lower than the price actually obtained. Discrimination thus tends to be economically detrimental

not only to those who are discriminated against, but also to those who practice discrimination.

The Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel 1972 "for their pioneering contributions to general economic equilibrium theory and welfare theory"

John R. Hicks Kenneth J. Arrow

1/2 of the prize 1/2 of the prize

United Kingdom USA

All Souls College

Oxford, United Kingdom

Harvard University

Cambridge, MA, USA

b. 1904

d. 1989

b. 1921

The Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel 2006 "for his analysis of intertemporal tradeoffs in macroeconomic policy"

Photo: R. Talaie

Edmund S. Phelps

USA

Columbia University

New York, NY, USA

b. 1933

Although Edmund Phelps is best known for his work in macroeconomics, his contributions to labor economics and public finance also deserve mention. Among other contributions he initiated the literature on statistical discrimination, derived new results regarding the structure of optimal income taxation, and examined the properties of an optimal inflation tax. Phelps’s ideas concerning statistical discrimination were outlined in his monograph (1972a) and formalized in an article (1972b). Around the same time, Kenneth Arrow published equally influential papers on statistical discrimination

(Arrow, 1972a, 1972b, 1973). These studies by Phelps and Arrow – both commonly referred to as seminal contributions to the theory of statistical discrimination – emphasize that unequal treatment of equally productive workers can arise when employers have imperfect information about individual worker characteristics. When individual productivity is measured with error, it may be worthwhile to use group data – information on average productivities in the group to which an individual belongs – so as to improve predictions about an individual worker’s productivity. A consequence of such behavior is that individuals with the same characteristics may be treated differently

3198 J. G. Altonji and R. M. Blank

technology. The presence of multiple worker characteristics in the model may lead to a pattern of biases that would be hard to sort out a priori.

A related way to test for employer based discrimination is to examine profitability of firms. Hellerstein et al. (1997) use the Worker Establ ishment Characteristics database to test for sex discrimination by examining whether there exists a cross-sectional relationship between profitability of a firm and the sex composit ion of the workforce, using Becker ' s (1971) original argument that, under certain conditions, discriminatory firms will have lower profits than non-discriminatory ones. They also explore how market power affects the discrimination-profi tabil i ty relationship, and whether discriminatory firms are bought out or are weakened over time.

The cross-section results using plant level data (firm level data) imply that a 10 percen- tage point increase in the proportion of female employees raises the profit rate by 4.6% (3.7%). The effect of percent female is weakened by the addition of 4-digit industry controls but remains statistically significant. There is evidence that the effect is largest for firms in the highest quartile of market share. These cross-section (short run) results are consistent with Becker ' s discrimination model. 22 The results of the dynamic models are weaker. Firms estimated to be more discriminatory in 1990 generally do worse in 1995 and are more l ikely to change ownership, but the estimates are noisy and statistically insig- nificant. 23

This last paper is interesting but shares a major problem with Hellerstein et al. (1996), namely, the variation in worker composition, including percent female, is l ikely to be correlated with heterogeneity in the production technology and may be endogenous to the model. Overall, we find this set of papers very interesting. As a way to test for discrimina- tion, research that looks simultaneously at productivity and wages is likely to be more fruitful than further analyses of the "unexplained" wage differential.

4.4. Testing for statistical discrimination

The basic premise of the statistical discrimination literature is that employers assess the value of younger workers using only the limited information contained in resumes, recom- mendations, and personal interviews. Given lack of information about actual productivity, employers have an incentive to "statistically discriminate" among young workers on the bas i s of easily observable variables such as race or gender, if these provide clues to a worker ' s , labor force preparation. However, there is almost no empirical literature testing whether employers do in fact statistically discriminate on the basic of race or gender.

Altonji and~Pierret (1997) provide a test of statistical discrimination by firms. Speci-

22 Hersch (1991) finds ~hat charges of EEO violations lead to reductions in the stock value of firms. If the firms discriminate against blacks or women and the charges lead to greater employment of these groups, then profits would be expected to rise. The legal costs, settlement costs and surrounding negative publicity may more than offset this effect, however.

23 It would be interesting to examine whether establishments that become part of publicly traded firms are more likely to increase their use of women.

Ch. 48: Race and Gender in the Labor Market 3199

fically, they consider a situation in which (1) group membership s is negatively related to productivity; (2) the relationship between group membership and productivity does not vary with experience; and (3) firms learn over time. They show that if firms statistically discriminate on the basis of group membership in this situation, then the relationship between wages and group membership will not vary with experience. If, on the other hand, firms do not statistically discriminate, then the wage gap will widen with experience. They also investigate the consequences of adding to a wage equation a typically hard-to- observe characteristic z that is positively related to productivity and negatively related to minority group membership. They show that not only should the coefficient on z rise with time in the labor market as firms learn about productivity, but the coefficient on s should fall if statistical discrimination occurs when the worker is first hired.

Their argument is as follows. Let Yit be the log ot" the marginal revenue product of worker i with ti years of experience. Yit is determined by

Yit = rs + H(ti) + cq q + A z + 9qi, (4.4)

where s is 1 if the person a member of the minority group, q is a vector of information about the worker that is relevant to productivity and is observed by employers, and z is a vector of correlates of productivity that are not observed directly by employers but are available to the econometrician, such as income of an older sibling or a test score. H(ti) is the experience profile of productivity. The variable ~ consists of other determinants of productivity and is not directly observed by the employer or the econometrician. Let e be the error in the employer' s belief about the log of productivity of the worker at the time the worker enters the labor market.

Each period that a worker is in the labor market, firms observe a noisy signal of the productivity of the worker, ~. The vector It = { ~j ..... ~t } summarizes the worker' s perfor- mance history. This information, as well as q and s, are public, so competition leads firms to set the wage level equal to expected productivity given s, q, and It, if firms violate the law and use the information in s to set wages. In this case Altonji and Pierret show that the log wage level wt will be

wt = log[E(exp(yi~) I s, q, It)] = As + H*(t) + pq + E(e [ It), (4.5)

where H*(t) is equal to H(t) plus a term that accounts for the fact that the log of the expectation of productivity given s, q, and L will be influenced by change over time in uncertainty about e, and • and p depend on r and c~ ~ as well as the relationship of z and 97 to s and q. The coefficient on s does not change with experience if, as the derivation of Eq. (4.5) assumes, firms make full use of the information in s, because q is time invariant and e is independent of s.

Eq. (4.5) is the process that generates wages. Suppose the econometrician observes only s and z, and regresses wr on these variables. (In short, the econometrician does not observe q, which the employer knows, but does observe z.) Let the coefficients of the regression o1" wt on s and z in period t be bst and bzt. Then

3200

E(wt I s, Z, t) = bsts + bztz + H*( t ) .

Altonji and Pierret show that

b~; = bso + Ot ~ , ,

J. G. Altonji and R. M. Blank

(4.6)

(4.7a)

bzt = bzo + Ot q)z, (4.7b)

where qb,. and ~z are the coefficients of the regression of e on s and z and 0, surmnarizes how much the film knows about e at time t. Under plausible conditions, q5 < 0 and ~z > 0. For instance, this is true when s = 1 for blacks and 0 for whites and the variable z is AFQT, father's education, or the wage rate of an older sibling. Note also that 0t is 0 in period 0, because in this period employers know nothing about e, so E(e I I0) = 0. 0t rises toward 1 as firms learn about e and E(e I/ t) is e. Consequently, b,t falls with experience and bzt rises with experience. Or, stated another way, if employers statistically discrimi- nate, over time they will learn the true productivity of the worker and the wage of the worker will become more closely related to productivity-related variables (z) and less closely related to race.

On the other hand, if firms obey the law and do not make direct use of s, then the coefficient on s will rise with time. That is, the race differential will widen as experience accumulates. To see this note that in this case s behaves the same as a z variable, which is essentially unobserved (unused) by the firm. With learning, firms are acquiring additional information about performance that may legitimately be used to differentiate among workers. If race is negatively related to productivity, then the new information will lead to a decline in wages, so over time the impact of race should become larger and more negative.

Altonji and Pierret also show that, regardless of whether firms statistically discriminate, adding to the wage equation a z variable that is positively correlated with race will reduce the racial difference in the experience profile. The intuition is that part of the effect of the new information about productivity is absorbed by the z variable which reduces the impact of the race variable. They also consider the effect of on the job training in their models.

In their empirical study of young men from the NLSY, they find that the race gap does widen substantially with experience, in contrast to the prediction of a model in which firms fully statistically discriminate on the basis of race. They also find that adding father's education, the AFQT score, or the sibling wage rate to the model (z variables) reduces the degree to which-the race gap widens with experience. This second result is consistent with employer learning~about productivity and is predicted to hold regardless of whether firms statistically discriminate by race. Other results provide support for the hypothesis that firms do statistically discriminate on the basis of education. Over time, wages become more strongly correlated with hard-to-observe productivity related variables and less strongly correlated with easily observable variables such as education. The main limitation of Altonji and Pierret's analysis is that the effects of statistical discrimination on wage

EMPLOYER LEARNING 335

TABLE III THE EFFECTS OF STANDARDIZED AFQT, FATHER'S EDUCATION, SIBLING WAGE, AND

SCHOOLING ON WAGES

Dependent Variable: Log Wage; Experience Measure: Potential Experience. OLS estimates (standard errors)

Model: (1) (2) (3) (4)

(a) Education 0.0505 0.0832 0.0563 0.0780 (0.0118) (0.0151) (0.0120) (0.0155)

(b) Black -0.1333 -0.1296 0.0454 -0.0284 (0.0255) (0.0257) (0.0609) (0.0704)

(c) Standardized AFQT 0.0792 -0.0206 0.0789 0.0065 (0.0145) (0.0361) (0.0144) (0.0413)

(d) Log of sibling's wage 0.1602 0.0560 0.1617 0.0604 (0.0208) (0.0352) (0.0207) (0.0351)

(e) Father's educationJlO 0.0362 0.0154 0.0385 0.0295 (0.0356) (0.0963) (0.0354) (0.0968)

(f) Education * 0.0005 -0.0269 -0.0035 -0.0220 experience/10 (0.0093) (0.0123) (0.0094) (0.0128)

(g) Standardized AFQT 0.0843 0.0614 * experience/10 (0.0285) (0.0333)

(h) Log of sibling wage * 0.1194 0.1151 experience/10 (0.0393) (0.0393)

(i) Father's education * 0.0176 0.0055 experience/100 (0.0789) (0.0794)

(j) Black * experience/10 -0.1500 -0.0861 (0.0474) (0.0570)

R2 0.2991 0.3014 0.3002 0.3016

Experience is modeled with a cubic polynomial. All equations control for year effects, education inter- acted with a cubic time trend, Black interacted with a cubic time trend, AFQT interacted with a cubic time trend, father's education interacted with a cubic time trend, sibling wage interacted with a cubic time trend, two-digit occupation at first job, and urban residence. Also included are sibling's gender and dummy variables to control for whether father's education is missing and whether sibling's wage is missing, and interactions between these dummy variables and experience when experience interactions are included. Standard errors are White/Huber standard errors computed accounting for the fact that there are multiple observations for each worker. The sample size is 21,058 observations from 2976 individuals.

known. This test amounts to a t-test of whether the sum of the products of -cov (s,z)/var (s) and the coefficient on z x t for each z variable is equal to the coefficient on s X t. For whites, the sum of the products equals -.0021 and the coefficient on s x t is -.0020. For blacks, we obtain -.0042 and -.0049. In both cases we fail to reject the proposition.

1V.3. The Experience Profile of the Effects of AFQT and Education on Wages

As noted earlier, employer learning implies that awtlaAFQT is nondecreasing in t, i.e., a2w ItlAFQT, At ? 0, with a strict

1 2 3 4 5

-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

t/10

ln w

not statistically discriminating – effect of becomes more important over

time

initially there is little discrimination

red line — not taking into account (i.e. few controls)

blue line — taking into account AFQT, sibling wage and father’s education

1

ביקוש לעובדיםה- 5יחידה

33

יהיה ה ישיפוע קו הרגרסנמוכה אזי var(ε( טעות המדידההאומד הוא מדויק באופן יחסי ו אם ,לעומת זאת

לייחס לעובד אשר הצליח מעל הממוצע במבחן ם ישתמשו באינפורמציה של המבחן על מנת בידיוהמע, גבוה

מושלם בין ) כמעט(כלומר אנו מניחים כי יש קשר , כאשר הטעות שואפת לאפס. תפוקה ממוצעת גבוהה יותר

ii: כישורי העובד והיכלות המותנית של העובד היא תוצאות המבחן לבין ttqE =)|(.

אבל נניח . כעת נניח כי קיימות שתי קבוצות של עובדים עם אותה תחזית לגבי תוחלת התפוקה הצפויה שלהם

דוגמא טובה לכך היא העלייה . האיכות שלהם בצורה טובה אמידתגם כי באחת מן הקבוצות קיים קושי ל

היה קושי אמיתי להעריך את טיב התעודות על אשר ,90מ לשעבר בתחילת שנות ה "דולה ממדינות בריההג

ההכשרה המקצועית של העולים החדשים אשר רבים מהם היו רופאים ומהנדסים לעומת הותיקים אשר רכשו

שוק ות עשויות לסבול מאפליה סטטיסטית בשל מספרם המצומצם בדוגמא נוספ. את השכלתם בישראל

המעבידים יתנו חשיבות נמוכה יותר לתוצאות המבחנים , במקרים אלו. נשים בתפקידי ניהול, העבודה למשל

. באופן דומה) למשל, העולים(ויטו לשפוט את כל חברי קבוצה זו

הסיבה לכך שחברי קבוצה זו אשר הם למעשה בעלי יכולת גבוהה אינם יכולים לשכנע את המעבידים ביכולתם

י המעבידים בהתאם "באמצעות הציונים שהם משיגים או התעודות שהם מציגים ולכן יוערכו עהגבוהה

. לכישורים הממוצעים של הקבוצה

מתאפיינת בכך כי חברי הקבוצה מוערכים על בסיס היכולת הממוצעת של כל תכי אפליה סטטיסטי, יש לציין

.ה אומר כי הקבוצה כולה תסבול מהאפליהאך אין ז. חברי הקבוצה ולא על פי היכולת האמיתית שלהם

הם קיבלו שכר פחות או יותר זהה וזאת 90 ניתן לראות כי כאשר העולים הגיעו בתחילת שנות ה 1' בטבלה מס

לעומת זאת לאחר כמה שנים ניתן לראות כי קיים . מ לשעבר"ללא קשר לכמות ההשכלה שרכשו במדינות בריה

הוכיחו את עצמם בפני המעסיקים ועל , ולים אשר היו בעלי השכלה גבוההאשר בו הע" מניפה"כאן אלמנט של

כן זכו לקבל שכר גבוה יותר ושכרם עלה בצורה החדה ביותר לעומת זאת עולים אשר היו בעלי השכלה נמוכה

כלומר . יותר שכרם עלה בצורה שטוחה יותר ועולם אשר היו ללא השכלה שכרם עלה בצורה הנמוכה ביותר

איזו תעודה היא בעלת אמינות גבוה יותר ואיזו תעודה היא בעלת , להבחין בין התעודות השונותהשוק למד

. ועלתה אמינות המבחן או במקרה זה אמינות התעודה של ההשכלה ε" רעש"אמינות נמוכה יותר וכך ירד ה

, ת השהייה בישראלשכר חודשי לעולים חדשים לפי רמות הכנסה ושנו : 2' טבלה מס

ביקוש לעובדיםה- 5יחידה

34

.25גברים אשר עלו לאחר גיל

שנות לימוד16מעל שנות לימוד13-15 שנות לימוד12עד השכלה

שנים

בישראל

סטיית תקןשכר ממוצעסטיית תקןשכר ממוצעסטיית תקןשכר ממוצע

1 1688 681 1717 551 1875 1016

2 1922 612 2070 919 2249 1061

3 1994 795 2188 716 2376 1377

4 2014 715 2385 1244 3142 1764

5 2233 762 2503 1018 3495 2071

6 2299 821 2686 1340 3548 2132

7 2449 884 2905 1222 4232 2575

8 2380 693 3078 1501 3812 1948

9 2448 843 3161 1575 3531 2095

10 2925 1034 3647 2289 4516 2034

שיעור גידול

שנתי שכר

ממוצע

6.50% 8.86% 11.16%

.1991-2000סקר הכנסות של הלשכה המרכזית לסטטיסטיקה : מקור הנתונים

. על רכישת ההשכלהתהשפעת האפליה הסטטיסטי 5.6.2

נניח כי קיימו שתי קבוצות . נבחן כעת את רכישת ההשכלה וכיצד הפרטים מחליטים על רכישת ההשכלה

כל פרט מכל קבוצה צריך לבחור כמה . b ואילו קבוצת המיעוט היא קבוצה aא קבוצה הקבוצה העיקרית הי

. השכלה לרכוש

מעלה את התפוקה של הפרט )תואר וכדומה, שנת השכלה למשל(אנו יודעים כי תוספת של יחידת השכלה

העלייה בשכר היא בגודל של מקדם הרגרסיה שהראינו . כר שלווכתוצאה מכך היא צריה גם להגדיל את הש

Chapter 48

RACE AND GENDER IN THE LABOR M A R K E T

JOSEPH G. ALTONJI*

Institute Jbr Policy Research and Department of Economics, Northwestern University and NBER

REBECCA M. BLANK*

School of Public Policy, University of Michigan and NBER

Contents

Abst rac t 3144 JEL codes 3144 1 In t roduct ion 3144 2 A n overv iew of facts about race and gender in the labor marke t 3146

2.1 Trends and differences in labor market outcomes and background characteristics 3146 2.2 Methodologies for decomposing wage changes between groups 3153 2.3 Estimating simple models of wage determination 3156 2.4 Estimating simple models of labor force participation 3161

3 Theor ies o f race and gender d i f fe rences in labor marke t o u t co mes 3164 3.1 The impact of group differences in preferences and skills 3165 3.2 An introduction to theories of discrimination 3168 3.3 Taste-based discrimination 3170 3.4 Discrimination and occupational exclusion 3176 3.5 Statistical discrimination, worker incentives, and the consequences of affirmative action 3180

4 Direct ev idence on d i sc r imina t ion in the labor marke t 3191 4.1 Audit studies and sex blind hiring 3192 4.2 Discrimination in professional sports 3195 4.3 Directly estimating marginal product or profitability 3196 4.4 Testing for statistical discrimination 3198

5 P re -marke t h u m a n capi tal d i f ferences: educat ion and fami ly background 3201 5.1 Race differences in pre-market human capital 3201 5.2 Gender differences in pre-market human capital 3204

6 Exper ience , seniori ty, t raining and labor market search 3207 6.1 Race differences in experience, seniority, training and :nobility 3208 6.2 Gender differences in experience, seniority, training and mobility 3213

* We are grateful to the Russell Sage Foundation and Institute for Policy Research for research support, and to Rachel Dunifon, Todd Elder, Raymond Kang, Joshua Pinkston, and James Sullivan for excellent research assistance. We also thank Orley AshepXelter and David Card for their patience and encouragement and partici pants in the Handbook pre-conference for helpful suggestions. All errors and omissions are our responsibility.

Handbook of Labor Economics, Volume 3, Edited by O. Ashenfeher and D. Card © 1999 Elsevier Science B.V. All rights reserved.

3143

3144 J. G. Altonji and R. M. Blank

7 Job characteristics, taste differentials, and the gender wage gap 3220 7.1 Overview 3220 7.2 The occupational feminization of wages 3221 7.3 The impact of other job characteristics 3223

8 Beyond wages: gender differentials in fringe benefits 3224 9 Trends in race and gender differentials 3225

9.1 Methodologies for decomposing wage changes between groups over time 3225 9.2 Accounting for trends in the black/white wage differential 3234 9.3 Accounting for trends in the male/female wage differential 3240 9.4 The overlap between race and gender 3244

10 Policy issues relating to race and gender in the labor market 3244 10.1 The impact of anti-discrimination policy 3245 10.2 The role of policies that particularly affect women in the labor market 3247

11 Conclusion and comments on a research agenda 3249 References 3251

Abstract

This chapter summarizes recent research in economics that investigates differentials by race and gender in the labor market. We start with a statistical overview of the trends in labor market outcomes by race, gender and Hispanic origin, including some simple regressions on the determi- nants of wages and employment. This is followed in Section 3 by an extended review of current theories about discrimination in the labor market, including recent extensions of taste-based theories, theories of occupational exclusion, and theories of statistical discrimination. Section 4 discusses empirical research that provides direct evidence of discrimination in the labor market, beyond "unexplained gaps" in wage or employment regressions. The remainder of the chapter reviews the evidence on race and gender gaps, particularly wage gaps. Section 5 reviews research on the impact of pre-market human capital differences in education and family background that differ by race and gender. Section 6 reviews the impact of differences in both the levels and the returns to experience and seniority, with discussion of the role of training and labor market search and turnover on race and gender differentials. Section 7 reviews the role of job characteristics (particularly occupational characteristics) in the gender wage gap. Section 8 reviews the smaller literature on differences in fringe benefits by gender. Section 9 is an extensive discussion of the empirical work that accounts for changes in the trends in race and gender differentials over time. Of particular interest is the new research literature that investigates the impact of widening wage inequality on race and gender wage gaps. Section 10 reviews research that relates policy changes to race and gender differentials, including anti-discrimination policy. The chapter concludes with comments about a,future research agenda. © 1999 Elsevier Science B.V. All rights reserved.

JEL codes: J7; J15; J16

1. Introduct ion

Race and gender differentials in the labor market remain stubbornly persistent. Although

the black/white wage gap appeared to be converging rapidly during the 1960s and early

Ch. 48: Race and Gender in the Labor Market 3145

1970s, black/white male wages have now stagnated for almost two decades. The black/ white female wage gap has actually risen over the past 15 years. The Hispanic/white wage gap has risen among both males and females in recent years. In contrast, the gender wage gap showed no change in the 1960s and 1970s. Not until the late 1970s did it begin to converge steadily (although a significant gender gap still exists). Of course, these wage gaps are only the most visible form of differences in labor market outcomes by race and gender. Substantial differences in labor force participation, unemployment rates, occupa- tional location, non-wage compensation, job characteristics and job mobility all exist by both race and sex.

This chapter is designed to provide an introduction into the literature that analyzes these differences. As we shall show, there are significant differences in the discussion of race versus gender. Where appropriate, we deal with both issues simultaneously, but in many sections we deal with race and gender differences sequentially, both because the literature on the two is quite distinct and because the conceptual models behind race and gender differences are often dissimilar.

It is important to note that our use of the term "race" in this chapter is extremely limited. With only a few exceptions, we discuss black/white differences in labor market outcomes throughout this chapter. This reflects a major lack in the research literature. There is remarkably little empirical work on Hispanic/non-Hispanic white differences or on Hispanic/black differences in labor market outcomes. There is even less empirical work looking at other racial groups, such as Asian Americans or American Indians. In part, this reflects a lack of data on these groups. However, the widespread availability of Census data and an increase in the race/ethnic categories in a host of datasets makes this excuse increasingly inadequate. We strongly hope that future research will remedy this gap, investigating many of the issues that we discuss here for other labor market groups.

The chapter attempts to summarize some of the most important research areas relating to race and gender in the labor market. Of necessity, there are topics which we will cover inadequately or not at all. In Section 2 we provide a statistical overview of the differentials by race and gender in the labor market. Section 3 discusses theories about how race and gender differences in the labor market arise, with particular attention to new theoretical developments integrating costly search into models of discrimination.

In Section 4 we begin our review of the empirical literature by considering recent studies that provide what we consider to be direct evidence on the role of discrimination, a literature that is remarkably small. In Section 5 we examine the role of differences in human capital accumulation prior to labor force entry, touching on the recent literature on the role of race differences in basic skills, and the literature on the role of differences in the type of education that women receive on the gender gap in wages and occupational location. Section 6 considers the contribution of experience, seniority, training, and labor market search to race and gender differentials.

In Section 7 we consider the consequences of different job characteristics for the gender wage gap, including the effects of occupational location, the "feminization" of occupa- tions, and the impact of part-time and temporary jobs. This research is closely related to

3146 .L G. Altor~]i and R. M. Blank

the extended and controversial discussion about the extent to which these differences are related to taste differentials versus constraints in the types of jobs available to men and women. While most of the chapter focusses on wage differentials, and to a lesser degree, employment rate differentials, in Section 8 we discuss the much smaller literature on the race and gender differentials in fringe benefits.

Perhaps more high quality research has been devoted to the analysis of changes over time in race and gender differentials than any other topic in this chapter. This has been a very active area over the past 10 years, and the work has been closely connected to more general analyses of changes in wage structure and the rise in inequality. Section 9 begins with a presentation of the standard methodology for decomposing wage changes between groups and then turns to research on the effects of changes in the prices of observed and unobserved skills. Our emphasis is on recent methodological developments.

In Section 10 we consider the effect of labor market policy on labor market outcomes. We summarize the research evaluating the impact of anti-discrimination legislation, and also briefly review two areas where policy has had large impacts on female workers, namely, the impact of maternity leave benefits and the impact of comparable worth legislation. We close with a few comments on a future research agenda in Section 11.

2. An overview of facts about race and gender in the labor market

2.1. Trends and differences in labor market outcomes and background characterist ics

Race and gender differentials in the labor market have been persistent over time, although the nature and magnitude of those differences have changed, as this section discusses. We begin with a basic set of facts about gender, race, and Hispanic/white differences in labor market outcomes and in personal characteristics (such as human capital measures) that are likely to be related to labor market outcomes. We then provide some simple estimates of how differences in wages and employment are related to differences in characteristics and differences in labor market treatment given characteristics. One purpose of this analysis is to illustrate with the most recent data the basic regression techniques that have been used in hundreds of labor market studies of race and gender differences. We particularly discuss the difficulties that arise in differentiating between the effects of labor market discrimina- tion and the effects of race and gender differences in preferences and human capital.

Table 1 shows a current set of key labor market outcomes for all workers, for white, black, and Hispal~c male workers, and for white, black, and Hispanic female workers. It is based on tabulations of the C~iarrent Population Survey (CPS) data from March 1996.

Row 2 of Table 1 indicates that black and Hispanic men as well as white women earn about two-thirds of that earned by white male workers on an hourly basis. Black and Hispanic women earn even less than minority men, only slightly over half of what white males earn. Figs. 1 and 2 show median weekly earnings among full-time male and female

Ch. 48."

&

Race and Gender in the Labor Market

-4

C~

-d

<

~ ~ . ~

o

~ o ~

o'3 ,~- u'] ' ,o r--

~ c q

c~ o

~5~5

c5 u']

o

o

oo c'-I o

o~

~5~5

cq.o. o o

c3 oo

< & ~

k [ 3

• 0 o

~5 ~5

3147

<Z

C~

..=

©

%

E

b~ "5

5",

b c~

3148

800

J. G. Altonjiand R.M. Blank

700

6O0

o "o 500 I'..- (33 o) T- -

40O

300

1965

. . " , - . . . . . . . - " . . . . . . . , . - "

. . . . , - I " ' " " ". White ................. . . . . . . . . . • . . . . . . . . . . . . . . . . . . . ,

' - , . . . . . . . . . . . - . . . . . . . . . . . . . . . .

" " " . . . . . . . . .

Hispanic .........................

i i i i I i i i i I i r i r F i ~ ~ i t r ~ i i I i I i i I i i

1970 1975 1980 1985 1990 1995

Fig. 1. Median weekly earnings of full-time male workers. Source: Bureau of Labor Statistics.

800

700

600

o~

' o 500

ob

4OO

300

... White ....... , . - . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - . . . . . . . . . . . .

m " ' " ~ . . . . . . . . . . . . . _ . . . . . . - - - -

....... ~ i s ; a ~ " .................................................

T 1965 1970 1975 1980 1985 1990 1995

Fig. 2. Median weekly earnings of full-time female workers. Source: Bureau of Labor Statistics.

Ch. 48: Race and Gender in the Labor Market 3149

workers from 1967 to the present fbr whites and blacks and from 1986 to the present for Hispanics. 1

The wage trends in these two figures reveal that women, part icularly white women, have experienced an increase in their earnings relative to men. But after declining in the 1960s, wage gaps have widened among racial/ethnic groups for both men and women. Although black men ' s wages rose faster than white men 's in the 1960s and early 1970s, there has been little relative improvement (and even some deterioration) in the 25 years since then. Both white and black men show declines in their median weekly earnings over the last decade. Hispanic men show the strongest recent wage declines, but some of this is due to immigration, which has brought an increasing population of less-skil led Hispanic men into

the workforce. Among women, white women ' s wages have risen steadily since 1980, as Fig. 2 indi-

cates. Black women ' s wages almost reached parity with white women in the 1970s, but have diverged again in the last 15 years, as black women have experienced little wage growth. Hispanic women, like Hispanic men, are doing relat ively worse over the past decade, in part because of shifts in labor force composit ion due to immigration.

Annual earnings (shown in row 3 of Table 1) show an even larger differential than hourly wages, suggesting that weeks and hours worked are lower among minorities and females. Indeed, rows 4 and 5 confirm that white men not only earn more per hour, they also work more weeks per year and more hours per week. These differences are less among full-t ime/full-year workers as rows 8 and 9 indicate, but they are still substantial. Row 6 shows that women are part icularly l ikely to be working part-time.

Consistent with the weeks and hours data, rows 10-13 indicate that white men are more l ikely to ever be employed over the past year and to be employed at any point in time. Unemployment among white women has been as low or lower than among white men since the early 1980s. Blacks have about twice the unemployment rates of whites. Figs. 3 and 4 graph unemployment rates from 1955 to the present among men and women and between whites, blacks and Hispanics. Unemployment rates are quite cyclical among all groups of men, although black male unemployment is more cyclical than white male unemployment. The differential between black, white and Hispanic male unemployment rates is remarkably constant over much of this time period. W o m e n ' s unemployment has been less cyclical than men 's . As has occurred with their wages, the gap between black and Hispanic women ' s unemployment rates and white women ' s unemployment rates is higher over the 1980s and early 1990s than it was in the early 1970s.

Wages and unemployment rates are often affected by overall labor force participation rates, which have changed dramatically over time. Labor force participation rates by race and gender are shown in Fig. 5 from 1955 to the present. This chart clearly depicts the convergence in labor force participation among all groups. Men have experienced a steady

i Data for Figs. 1-5 are from the Bureau of Labor Statistics, tabulated from the Current Population Survey. Prior to 1972, the data for blacks includes all non-whites. Beginning in 1979, the data in Figs. 1 and 2 are for workers ages 25 and over.

3150

25

J. G. Altonji and R. M. Blank

. . o 1 5 c- (11 P 11)

10

20

",. "t.\ "t ;'.. .-, " i; '" ' , .."" : ........ '-. ,-"" '", .

1950 1955 1960 1965 1970 1975 1980 1985 1990 1995

Fig. 3. Male unemployment rates (annual averages). Source: Bureau of Labor Statistics.

decline in their labor force involvement, with the largest declines among black men. Women have shown dramatic increases in labor force participation over these years.

25

20

.~ 15 c

1 0

1950 1955 1960 1965 1970 1975 1 9 8 0 1985 1990 1995

Fig. 4. Female unemployment rates (almual averages). Source: Bureau of Labor Statistics.

Ch. 48: Race and Gender in the Labor Market

lOO

3151

o d.)

13_

9O

8 0

7 0

6O

5O

4 0

0 1 9 5 0

...................................................................................... wh_!t_e ===============================================================================================

. .J'" White w o m e n . , ............

. . - -

i F r i I i i r i r q i i i P i i i t I i i i i I r i i i F i i i i I , i i i I i i i i I i ,

1955 1960 1965 1970 1975 1980 1985 1990 1995

Fig. 5. Labor force participation rates, 25 54-year-olds. Source: Bureau of Labor Statistics.

White women have entered the labor market at a particularly high rate. While their rates of labor force participation used to be far lower than those of black women, they are now at parity. Hispanic women's labor force participation, although rising steadily, is still far below that of black and white women.

In delineating the causes of these labor market differences, labor economists look first at the substantial differences in the attributes that different workers bring with them to the workplace. Table 2 shows a set of key personal characteristics among all persons in 1996, and among the same six race/gender groups observed in Table 1.2 Educational differences among these groups are large, with race and ethnicity mattering much more than gender. Both male and female Hispanics have particularly low education levels. White women's educational levels are quite similar to white males (this was not true in earlier periods), while blacks have less education than whites but more than Hispanics. These differential investments in education may reflect different preferences and choices, and/or they may reflect "pre-market" discrimination. For instance, there is substantial evidence that blacks have been consistently denied access to suburban housing and crowded into inner city residential neighborhoods with substandard schools. Under these circumstances, blacks will receive a poorer public education and may leave school earlier.

Row 7 of Table 2 shows a "potential experience" calculation, based on calculating (age - years of education - 5) for each individual. This calculation assumes that people are working during all their adult years when they are not in school. Although this variable

2 The results in Table 2 would not be very different if the tabulations included all workers rather than all

persons.

3152

" a

©

C9

©

..~ZZ

d d ~ d d ~ d d ~ d ~ d

~5 d d ~ d d ~ d d

o, o

o ~ d d ~ ~ d

o o

~ d d ~ d L d

v

~ , ~ _ ~ ~ ~ .~ . ~ , ~

J. G. Altonji and R. M. Blank

~ Q O Q ~ O 0 0 0

d d & d d d d d d ~

o

N

N~

N

©

cq

Ch. 48: Race and Gender in the Labor Market 3153

is commonly used because many datasets lack information on actual experience, it is a particularly poor proxy for experience among women, who are more l ikely to leave the labor market during their child-bearing years. We return to this point below when we look at alternative data with information on actual experience.

Rows 8-10 of Table 2 indicate that the family and personal commitments of different workers also vary substantially. Whites are much more l ikely to be manied; Hispanics have more children to care for; and black females have greater child care responsibilities than black males. To the extent that family responsibilities influence labor market choices and create labor market constraints, these differences may be important in explaining differences in labor market outcomes.

Rows 11-20 of Table 2 indicate substantial variation in the geographic location of different groups. Blacks are more l ikely to be in the southern regions and Hispanics are more l ikely to be in the western regions. Minorities are also far more l ikely to be in major urban areas (a relatively recent shift for black Americans, who were traditionally more l ikely to be located in rural areas.) As Bound and Freeman (1992) and Bound and Holzer (1993, 1996) emphasize, to the extent that local labor markets differ and that labor is largely immobile in the short-run, 3 these differences in regional location will also shape labor market outcomes.

Table 3 looks at occupation and industry differences by race and gender. As others have observed, these differences are large. Black and Hispanic men are more l ikely to be in less skilled jobs. Women are generally more likely to be in clerical and service occupations or in professional services (which includes education). White women and Hispanic men are more l ikely to be in retail trade; blacks are more l ikely to be in public administration.

A key question is whether occupational and industry differences represent preferential choices or constraints. If one believes that firms discriminate in their propensity to hire into certain occupations, then occupational location is an outcome of discrimination rather than a choice-based characteristic. We discuss the research literature on this issue below. In the regressions reported in this chapter, we follow standard procedure and report regressions with and without controls for occupation, industry and job characteristics (public sector location or part-time work.) Regressions that do not control for these variables in any way probably underestimate the importance of background and choice-based characteristics on labor market outcomes. Regressions that fully control for these variables probably under- estimate the effect of labor market constraints. We allow readers to look at both outcomes.

2.2. Methodologies f o r decomposing wage changes be tween groups

One way to explore the wage differential between groups is to decompose it into "explained" and "unexplained" components. Assume that wages for individual i in group 1 at t ime t can be written as

Wji: = tgl:Xli: + tzli: (2.1)

3 Indeed, the more mobile is labor, the less local labor markets will differ.

3154

7Z

¢.~

g

3

o .=

o

3. G. Altonji and R. M. Blank

q q q q ~ q q ~ q q o o o o o ~ o o o o o o d d

~ 0 0 ~ 0 0 ~ 0 0 0 0 0

~ ~ ~ ~ ~.

~ 0 0 ~ 0 0 0 0 0 0 0 0 d d d d d d d d d d d d d

d

_~ ~ ~ ~ .o

~ ~ ~ o o

0 d o ~ d

. ~ ~

Ch. 48: Race and Gender in the Labor Market 3155

o o o o o o o o o o o o o

0 O O O 0 0 0 0 0 0 0 0 0 0

o ~ o o o o o o o ~ o d d d d d

0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 ~ 0 0 0 0 ~ 0 d O 0 d d d d d o d o d o

d d o d d d

q q q q q q o ~ o o o ~ o o o o o o o d d d d d d d d

b ~ ~' ~ ~ ~ 0=

c~

©

"a

r j

b

o=

3156 J. G. Altonji and R. M. Blank

and wages for individual j in group 2 at t ime t can be written as

w2j~ = ,~2tx2j~ + ~2j,, (2.2)

where/31t and/32t are defined so that E(ujit [ Xjit) = 0 and E(u2j r I X2j t ) = O. The difference in mean wages for year t can be written as 4

Wit - W2~ = (Xl, - X2t)/31t + (/31, -/32t)X2t, (2.3)

where Wut and Xut represent the mean wages and control characteristics for all individuals in group g in year t. The first term in this decomposi t ion represents the "explained" component, that due to average differences in background characteristics (such as educa- tion or experience) of workers from groups 1 and 2. It is the predicted gap between groups 1 and 2 using group 1 - typical ly white men - as the norm. The second term is the "unexplained" component, and represents differences in the estimated coefficients, i.e., differences in the returns to similar characteristics between groups 1 and 2. The share of the total wage differential due to the second component is often referred to as the "share due to discrimination." This is misleading terminology, however, because if any important control variables are omitted that are correlated with the included Xs, then the/3 coeffi- cients will be affected. The second component therefore captures both the effects of discrimination and unobserved group differences in productivity and tastes. It is also misleading to label only this second component as the result of discrimination, since discriminatory barriers in the labor market and elsewhere in the economy can affect the Xs, the characteristics of individuals in the labor market.

2.3. Est imating simple models o f wage determinatio,~

In this section we explore race and gender gaps in wages through a set of simple models of wage determination. Table 4 shows the differences in race and gender coefficients over time, across specifications and between all workers and full-time/full-year workers. Columns (1) and (4) report regressions of log hourly wages in 1979 and 1995 respectively on dummy variables for black, Hispanic and female, without including any further control variables. Columns (2) and (5) include controls for education, experience and regional location, a minimal set of personal characteristics that an individual brings to a job. Colurmas (3) and (6) add further controls for occupation, industry and job characteristics.

Part A of Table 4 focuses on all workers. As control variables are added to the model the negativ6 effect of race or gender on hourly wages becomes less significant. In 1995, black males received 21% lower hourly wages than white males if no control variables were included; they received 12% less once education, experience and region were controlled for, and they received 9% less when a full set of control variables were included. Among white women, there is only a small effect of adding controls for education and experience

4 Alternatively, the average wage difference can be decomposed as Eq. (2.3~): W~t - W2t = (Xtt - X 2 t ) ~ 2 t

+(/31~ -/32~)X~t. This 'alternative decomposition can produce quite different results from the first. Many authors report both results, or (occasionally) the average of the two.

Ch. 48: Race and Gender in the Labor Market

Table 4 Coefficients on race and gender in wage regressions ~

3157

1979 1995

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

Part (A) all workers (1) Black -0.143 -0.107 -0.061 0.207 -0.119 -0.089

(0.010) (0.010) (0.010) (0.012) (0.011) (0.011) (2) Hispanic -0.152 -0.053 0.040 0.379 -0.131 -0.102

(0.010) (0.010) (0.010) (0.010) (0.010) (0.009) (3) Female -0.436 -0.421 -0.348 -0.279 -0.272 0.221

(0.006) (0.005) (0.006) (0.007) (0.006) (0.007)

Controls" (4) Education, No Yes Yes No Yes Yes experience, and region (5) Occupation, No No Yes No No Yes industry and job characteristics b

Part (B) full-time-full year workers (6) Black -0.139 -0.115 -0.064 0.148 -0.102 -0.067

(0.012) (0.011) (0.011) (0.012) (0.011) (0.010) (7) Hispanic -0.184 -0.093 -0.076 -0.344 -0.139 -0.101

(0.012) (0.012) (0.011) (0.010) (0.010) (0.010) (8) Female -0.421 -0.399 -0.360 0.265 0.266 -0.241

(0.006) (0.006) (0.007) (0.007) (0.006) (0.007)

Controls (9) Education, No Yes Yes No Yes Yes experience, and region (10) Occupation, No No Yes No No Yes industry and job characteristics b

~' Source: Authors' regressions using tile Current Population Survey, March 1980 and March 1996. Standard errors are in parentheses.

b Job characteristics include public sector and part-time status.

( s u g g e s t i n g tha t t h e s e c h a r a c t e r i s t i c s a m o n g w h i t e w o m e n and w h i t e m e n are qu i t e s i m i l a r

as T a b l e 2 i n d i c a t e s ) , b u t c o n t r o l l i n g fo r o c c u p a t i o n a n d i n d u s t r y r e su l t s in subs t an t i a l l y

s m a l l e r n e g a t i v e e f fec t s .

Pa r t B o f T a b l e 4 l o o k s o n l y at f u l l - t i m e / f u l l - y e a r w o r k e r s . 5 T h e r e su l t s are s u r p r i s i n g l y

3158 J. G. Altonji and R. M. Blank

similar to those for all workers, both in the magnitude of the coefficients within any specification and in the change in coefficients over time and across specifications.

The results in Table 4 show that there are ongoing and significant race and gender differences in the labor market, even after controlling for occupational and industry loca- tion. The remaining negative effects faced by minority and female workers indicate that either we are omitting some key variables from this specification that are relevant to labor market productivity, and/or there are substantial "unexplained" constraints in labor market returns among minorities and women.

Table 5 uses the decomposition shown in Eq. (2.3) to decompose changes in log hourly wages in 1979 (part A) and 1995 (part B) for three groups: blacks versus whites, Hispanics versus whites, and females versus males. The top row of Table 5 shows the difference in log hourly wages between these three groups in 1979. The second and third rows decom- pose this into the share due to differences in characteristics and differences in coefficients. In the "Partial" specification, the only control variables are education, experience and region; the "Full" specification also controls for occupation, industry and job character- istics. Rows 4-10 show how much of the total difference in characteristics is due to specific sets of variables; rows 11-18 show how much of the total difference in coefficients can be ascribed to specific sets of coefficients. Part B repeats the same analysis for 1995. We report the detailed breakdowns because it is standard in the literature to do so, but it is important to emphasize the decompositions for subgroups of variables and the intercept term are not invariant to the scale of the variables. Variables such as education and experience have a natural scale but occupation and industry do not. For example, changing the omitted category for occupation will change the contribution of differences in the intercept and differences in occupation coefficients, as Oaxaca and Ransom (1999) discuss.

Two patterns are visible for all three groups in the table. First, as one moves from the partial to the full specification, the share of the wage differential explained by character- istics increases substantially. This is expected as we control more completely for job characteristics. Second, as one moves from 1979 to 1995, the share of the differential due to characteristics declines, indicating that over time these groups' characteristics are moving closer to those of white men. The exception to this is the Hispanic versus white comparison. The increasing importance over time of differences in characteristics is consistent with increased in-migration of Hispanics with poorer skill characteristics than native Hispanics.

Loo~n'g just at the 1995 results, it is clear that differentials in education and experience continue to negatively affect wages for black workers. The returns to education for blacks are actually stroriger than for whites, but the returns to experience are substantially lower, more than offsetting the ad'~antage in educational returns. One sees a similar pattern among Hispanics, although their mean characteristics remain further from those of whites, hence characteristic differences are more important.

5 Full-time/full-year workers work a minimum of 35 h/week and 48 weeks/year.

Ch. 48: Race and Gender in the Labor Market 3159

Table 5 Decomposition of race and gender wage differentials '

Specification Blacks vs whites Hispanics vs whites Females vs males

Partial Full Partial Full Partial Full

Part (A) 1979 (1) Log(hourly wage) difference -0.165 -0.126 0.457

Amount due to (2) Characteristics -0.063 (3) Coefficients -0.102

-0.108 -0.086 0.105 0.026 0.126 -0.061 -0.041 0.025 0.432 0.335

Differences due to characteristics (4) Education -0.023 (5) Experience -0.033 (6) Personal characteristics b 0.030 (7) City and region 0.026 (8) Occupation N/A (9) Industry N/A (10) Job characteristics ~ N/A

-0.017 0.002 0.001 -0.022 -0.011 -0.009 -0.024 -0.013 0.010

0.013 0.027 0.039 -0.049 N/A -0.025 -0.007 N/A 0.018

0.003 N/A 0.003

0.002 -0.024

0.004 -0.001 N/A N/A N/A

-0.001 -0.018 -0.002 -0.000

0.028 -0.060 -0.018

Differences due to parameters" (11) Education 0.080 (12) Experience - 0.100 (13) Personal characteristics t' 0.082 (14) City and region 0.002 (15) Occupation N/A (16) Industry N/A (17) Job characteristics c N/A (18) Intercept - 0.168

0.045 0.031 0.051 0.032 -0.153 -0.111 0.071 0.074 0.054 0.036 -0.057 -0.056 0.025 N/A 0.021

-0.016 N/A 0.013 0.008 N/A 0.005

-0.252 0.145 0.122

0.041 -0.612

0.019 -0.039 N/A N/A N/A

0.146

--0.031 0.410 0.014

-0.023 0.056 0.046 0.0t6

-0.009

Part (B) 1995 (19) Log(hourly wage) difference

-0.211 --0.305 0.286

Amount due to (20) Characteristics 0.082 (21) Coefficients -0.134

0.114 0.193 -0.226 -0.008 -0.076 -0.098 -0.112 -0.079 -0.279 -0.211

Differences due to characteristics (22) Education -0.028 (23) Experience -0.058 (24) Personal characteristics b -0.025 (25) City and region 0.030 (26) Occupation N/A (27) Industry N/A (28) Job characteristics ~ N/A

-0.013 -0.055 -0.024 -0.048 -0.185 -0.152 -0.020 0.010 0.008

0.020 0.038 0.033 -0.058 N/A -0.080

0.006 N/A 0.012 -0.000 N/A 0.001

0.000 -0.005 -0.002

0.001 N/A N/A N/A

0.001 -0.003

0.002 0.001 0.012

-0.036 -0.020

3160 J. G. Altonji and R. M. Blank

Table 5 (continued)

Specification Blacks vs whites Hispanics vs whites Females vs males

Partial Full Partial Full Partial Full

Differences due to parameters (29) Education 0.091 0.082 0.022 0.012 0.003 0.022 (30) Experience -0.197 -0.145 0.208 -0.025 -0.093 -0.023 (31) Personal characteristics b 0.055 0.047 0.031 0.025 0.019 0.014 (32) City and region 0.016 0.030 -0.036 -0.032 -0.037 -0.013 (33) Occupation N/A 0.005 N/A -0.058 N/A 0.060 (34) Industry N/A 0.032 N/A 0.046 N/A 0.004 (35) Job characteristics c N/A 0.009 N/A 0.033 N/A 0.014 (36) Intercept 0.100 -0.148 0.079 -0.081 -0.165 -0.237

~' Source: Authors' regressions using the Current Population Survey, March 1980 b Personal characteristics include sex and race when appropriate.

Job characteristics include public sector and part-time status.

and March 1996.

There are fewer differences be tween males and females in their background character-

istics, so that characteris t ics play only a small role in labor market differentials for w o m e n

in 1995. The returns to both educat ion and exper ience are sl ightly lower for women. A

large share of the coeff icient effect for w o m e n and blacks comes f rom a lower intercept

term. This is typical ly interpreted as ongoing d iscr iminatory constraints in the labor marke t

for these groups. It should be kept in m i n d that cohor t effects m a y bias es t imates o f the

return to exper ience in cross-sect ion regressions o f the type we report here. One wil l get a

low return to exper ience i f the recent cohorts have r ece ived better school ing or had more

full access to labor market opportunit ies. This migh t be impor tant for w o m e n and blacks.

Whi l e the CPS data provides a large national sample o f workers , it has serious l imits.

Mos t important ly, it lacks any measure o f ability, it has inadequate informat ion on past

labor market exper ience, and it is l imi ted in its f ami ly background characterist ics. To

invest igate the impor tance of these l imitations, we ran regress ions for blacks and w o m e n

using data f rom the Nat ional Longi tudinal Survey of Youth (NLSY) for 1994. The N L S Y

provides data on a cohort of workers ages 29 -37 in 1994, hence it is representat ive o f only

a l imi ted age group in the labor market . It is also a m u c h smal ler sample, without enough

observat ions on Hispanics to look separately at this group. The N L S Y has been col lec ted

annually since 1979, however , and has a much r icher set o f variables than the CPS. It

a l lows us to add three crucial sets o f var iables to our fo rmal estimates: actual years o f past

exper ience in the labor market ; the ind iv idua l ' s score on the A r m e d Forces Qual i fy ing

Tex t (AFQT) whlbh is typical ly used as a measure o f ability, 6 and a set of fami ly back-

c, An extended discussion about the appropriate interpretation of AFQT scores has occurred recently. This is not a measure of innate ability, but is clearly related to years of schooling. With controls for educalion in the model, one might interpret the AFQT results as a measure of how much an individual has learned, conditional upon years of schooling. Thus, it can represent poor school quality as well as differences in ability. Further discussion of this issue occurs in Section 5.

Ch. 48: Race and Gender in the Labor Market 3161

ground variables including father 's and mother 's education and father 's and mother 's employment status when the individual was an adolescent.

Table 6 shows the results of our NLSY regressions for 1994. Models 1 and 5 repeat the partial and full specifications used with CPS data. Models 2 and 6 add AFQT scores and family background. Models 3 and 7 also replace potential experience with actual experience. Models 4 and 8 add family characteristics and (for the regressions in rows 8 11) race or sex dummies where appropriate. Rows 6 and 7 show the coefficients on dummy variables for race and gender in these models. Rows 8-9 and 10-11 are decom- positions of wage differentials based on separate male/female regressions and white/

black regressions. For both the partial and the full specification, three patterns are apparent in Table 6.

First, the inclusion of A F Q T scores eliminates much of the black/white wage differential, as others have noted (Neal and Johnson, 1996). Second, the effect on the female/male wage differential of controlling for actual experience, AFQT scores, and family character- istics is relatively modest, lowering the unexplained wage differential only slightly. 7 Third, the decomposit ion of results in the NLSY is quite similar to that using CPS data. For women, virtually all of the wage difference is due to coefficient differences in the more complex models. For blacks, a much higher share is due to characteristic differences, particularly as more control variables are added to the model.

The results in Table 6 confirm that an improved specification can reduce the unex- plained effects for blacks and for women. In fact, for blacks, the inclusion of the AFQT scores virtually eliminates any remaining black/white differences. For women, however, even with a richer set of control variables in the model, a significant portion of the male/ female wage differential remains unexplained.

2.4. Estimating simple models of labor force participation

Not all of the concern about race and gender differences in the labor market rew)lves around wages. Differentials in labor force participation between these groups are also a concern. This has been particularly true as participation rates among less-skilled black men have declined, and as pol icy-makers have focused welfare reform efforts on increas- ing the labor force participation of less-skilled women. Fig. 5 indicates there have been dramatic trends in labor force participation over time.

Table 7 shows the results of estimating separate labor lbrce part icipation equations for blacks versus whites, Hispanics versus whites, and females versus males in 1979 (part A) and 1995 (part B), using data from the CPS. The first row shows relative labor force participation ratios. Rows 2 and 3 decompose a simple labor force participation regression for these groups into the share due to characteristics versus the share due to coefficients. This regression includes controls for education, potential experience, race and gender

7 Our measure of actual experience is relatively crude. Using more detailed controls for actual experience would probably have a bigger effect on the gender gap. See Section 6.2.1.

3162 J. G. Altonji and R. M. Blank

% "m

" 0

0 "m

0

©

~ o

8 ~ ~ o 8

8 o o ~ ~ Z

Z Z Z

8 ~ 8 o ©

~ 0 0 0 ~ Z Z Z

© c>

~ ~ ~~

d q d q i o 0 w w

i o o

d q d q

o ° d q

d q d q i o o w

~ c~ o . d o ~ l o d

-g

2

O 0 O 0 I I I I

o o c 5 o I I I I

o o I I

o ~ o c 5

I I

c S d I I

~ c5c~

" ~ o ~ ~ ~ ' ~

Z ~ o o

I I

o o

I I

~ [

I I

c~.

I

c ¢ " ' ~

Z (.) <.)

©

© ©

"U "~

Ch. 48: Race and Gender in the Labor Market 3163

Table 7 Decomposition of race and gender labor force participation differentials ~

Blacks vs whites Hispanics vs whites Females vs males

Part (A) 1979 (1) Labor force participation difference

-0.065 -0.047 -0.273

Amount due to

(2) Characteristics -0.046 0.052 -0.005 (3) Coefficients -0.019 0.006 -0.267

Differences due to characteristics (4) Education 0.011 -0.016 0.001 (5) Experience -0.014 -0.005 -0.002 (6) Personal Characteristics* -0.014 -0.025 -0.004 (7) City and Region -0.007 0.006 -0.000

Differences due to parameters (8) Education 0.042 0.025 0.052 (9) Experience 0.318 0.041 0.015 (10) Personal characteristics I' 0.112 -0.017 -0.209 (11) City and region -0.016 0.030 -0.014 (12) Intercept - 0.474 0.069 0.112

Part (B) 1995 (13) Labor force participation difference

-0.086 0.081 -0.156

Amount due to (14) Characteristics -0.048 0.077 0.008 (15) Coefficients - 0.037 0.004 - 0.148

Differences due to characteristics

(16) Education -0.007 -0.021 -0.009 (17) Experience -0.015 -0.032 -0.003 (18) Personal characteristics b -0.017 -0.015 -0.004 (19) City and region -0.009 -0.009 -0.003

Differences due to parameters (20) Education 0.077 0.046 0.041 (21) Experience 0.189 0.002 0.109 (22) Personal characteristics b 0.058 -0.070 -0.121 (23) City and region 0.062 -0.011 0.007 (24) Intercept -0.423 0.030 -0.170

a Source: Authors' regressions using Current Population Survey, March 1980 and March 1996. b Personal characteristics include marital status, no. of children less than 6, total no. of children, and sex and

race when appropriate.

3164 J. G. Altonji and R. M. Blank

(when appropriate), marital status, total number of children, number of children less than age 6 years, and SMSA and regional location.

Looking at the results for 1995 in Part B of Table 7, there are striking differences between blacks and Hispanics on the one hand and males and females on the other hand. Black and Hispanic differences in labor force participation are largely due to group differences in background characteristics. In contrast, male/female differences in labor force participation are entirely due to differences in coefficients. In particular, the coefficients on personal characteristics (children and marital status) are much more nega- tive for women than for men. Women as well as blacks continue to have a large unex- plained difference in the intercept term. In contrast, the effect of education and experience on labor force participation is actually higher for women than for men and for blacks and Hispanics than for whites.

The results in this section only briefly summarize some of the key differences in outcomes and background characteristics between female, black, Hispanic, white, and male workers. Among the key conclusions in this section: There are substantial differ- ences between male/female differentials in the labor market and black/white or Hispa- nic/white differentials. Male/female wage differentials remain greater than those of minority men versus white men and the decomposition of those differentials is differ- ent. There are fewer differences between blacks and Hispanics, although the aggregate category "Hispanic" includes workers from a very diverse set of backgrounds. Even controlling for occupation, industry, and job characteristics, there remain significant differentials between white males and other workers. Some of this may be due to incompletely specified models, as the inclusion of the AFQT scores for black men indicates. Some of it almost surely represents ongoing constraints in the labor market for women and minorities. Over time, minorities and women have acquired more education and experience than before, hence their human capital characteristics are less important in explaining their wage differentials in 1995 than 15 years earlier. But there remain significant unexplained differences in the coefficients that determine the returns to worker and job characteristics among black, Hispanic, and women work- ers. Below, we discuss research that investigates more causally complex questions about these differences.

3. Thegries of race and gender differences in labor market outcomes

In this section we discuss theoretical research on the sources of race and gender differences in labor market o~atcomes.We begin in Section 3.1 by reviewing the hypothesis that group differences in wages, occupations, and employment patterns are the consequence of preference and skill differences rather than discrimination. This "preferences/human capi- tal" hypothesis is the null hypothesis underlying most of the empirical research on race and gender differences. In this case, discrimination is assumed to be the residual difference that exists in labor market outcomes that cannot be explained by these factors. However,

3250 J. G. AItonji and R. M. Blank

emerge and persist. After more than a decade with almost no new theoretical research on discrimination, within the past few years, there has been a set of very good new papers that have improved existing models by incorporating costly search and differential labor market information. Building further on these models would be useful, as would theore- tical work that takes existing models and investigates the effects of various labor market policies. Particularly given the emerging debate about race-bl ind versus preferential poli- cies, we need better models by which to evaluate the impact of different approaches.

Second, most of the existing literature on race and gender focuses on black and white males or on males versus females. While these are important groups, we could learn much more about comparative labor market differentials by widening the research focus to include other groups. The recent wage and employment experiences of black women (which have deteriorated) are understudied. In addition, there is a major need for more research in economics on Hispanics and on Asian Americans with regard to their labor market involvements. In addition, because each of these populations (like the white population) are extremely heterogeneous, research on the relative experiences of various ethnic subgroups (such as Mexican Americans) can also be useful. Greater cross-group research can provide comparative information that helps us better understand the nature of racial, ethnic and gender-based differences in the labor market.

Third, despite major public and private resources devoted to anti-discrimination policy, the research literature on the results of these efforts is sparse. While we recognize the difficulties of studying nationally enacted legislation, in many cases there are differences over t ime or across regions in the implementat ion of such legislation, or there is variation in related state-specific legislation. Such research may require the collection of adminis- trative and outcome data at a sub-national level, which is always t ime-consuming and difficult, but it is l ikely to provide useful information, particularly in a world where existing anti-discrimination measures in education and in the labor market are at the center of a major public debate about the appropriate response to ongoing racial differentials.

Finally, we are struck by a few specific areas that appear ripe for more research. For instance, the impact of women ' s changing selectivity into the labor market on their wages has not been revisited in recent years. Much of the upsurge in female labor force participa- tion in recent years has been among non-married women or among women with pre-school children. This suggests that our older estimates of selectivity could be outdated, and impacts may vary among different groups of women workers.

Moving from issues of gender to issues of race, the growing interest in research on the impact of widening wage inequality on changes in the returns to unobserved skills opens up a number of new research topics. Most importantly, we need to find more effective ways to measure schoot quality and its determinants, if we want to test the hypothesis that education quality differentials are a major cause of the black/white wage gap. Similarly, we need more data that provides good measures of worker skills, to further understand the result that controlling for AFQT test scores eliminates the race differential; it is possible that firm-specific studies are one way to provide this. It would also be useful to know more about how less-skilled workers can overcome some of the negative wage effects they have

Ch. 48: Race and Gender in the Labor Market 3251

recent ly been exper ienc ing . F i rm-spec i f ic training p rograms , n e w m a n a g e m e n t techni-

ques, and/or new workp lace t echnolog ies may all be impor t an t w a y s by which current ly

l o w - w a g e workers can increase their product ivi ty .

Overal l , we are e n c o u r a g e d by the recent g rowth in bo th theore t ica l and empir ical

approaches to s tudying race and gender differentials in the labor force. Af te r a per iod

of hiatus, this is an area w h i c h is again genera t ing interest a m o n g top scholars. W e expec t

that fur ther good resea rch will be fo r thcoming in the years ahead.

References

Aaronson, D. (1998), "Using sibling data to estimate the impact of neighborhoods on children's educational outcomes", Journal of Human Resources 33 (4): 915-946.

Abbott, Michael G. and Charles M. Beach (1994), "Wage changes and job changes of Canadian women: evidence from the 1986-87 labour market activity survey", Journal of Human Resources 29 (2): 429-460.

Aigner, Dennis J. and Glenn G. Cain (1977), "Statistical theories of discrimination in labor markets", Industrial and Labor Relations Review, 30: 175-187.

Akerlof, George (1976), "The economics of caste and of the rate race and other woeful tales", Quarterly Journal of Economics 90: 599-617.

Akerloff, George (1980), "A theory of social custom, of which unemployment may be one consequence", Quarterly Journal of Economics 94: 749-776.

Altonji, Joseph G. (1993), "The demand for and return to education when education outcomes are uncertainty", Journal of Labor Economics 11 (1): 48-83.

Altonji, Joseph G. (1995), "The effects of high school curriculum on education and labor market outcomes", Journal of Human Resources 30 (3): 410-438.

Altonji, Joseph G. and Christina H. Paxson (1992), "Labor supply, hours constraints and job mobility", Journal of Human Resources 27 (2): 256-278.

Altonji, Joseph G. and Charles R. Pien'et (1997), "Employer learning and statistical discrimination", Working paper (NBER, Cambridge MA).

Altonji, J.G. and R.A. Shakotko (1987), "Do wages rise with job seniority?" Review of Economic Studies 54: 437-439.

Altonji, Joseph G. and James R. Spletzer (1991), "Worker characteristics, job characteristics and the receipt of on-the-job training", Industrial and Labor Relations Review 45 (1): 58-79.

Altonji, J.G. and N. Williams (1998), "The effects of labor market experience, job seniority, and job mobility on wage growth", Research in Labor Economics, in press.

Anderson, Deborah and David Shapiro (1996), "Racial differences in access to high-paying jobs and the wage gap between black and white women", Industrial and Labor Relations Review 49 (2): 273-286.

Arrow, Kenneth (1973), "The theory of discrimination", in: O.A. Ashenfelter and A. Rees, eds., Discrimination in labor markets (Princeton University Press, Princeton, NJ) pp. 3-33.

Ashraf, Javed (1996), "Is gender pay discrimination on the wane? Evidence from panel data (1968 1989)", Industrial and Labor Relations Review 49 (3): 537-546.

Baldwin, Marjorie and William G. Johnson (1992), "Estimating the effects of wage discrimination", Review of Economics and Statistics 74 (3): 446-455.

Barron, John M., Dan A. Black and Mark A. Loewenstein (1993), "Gender differences in training, capital and wages", Journal of Human Resources 28 (2): 343-64.

Benabou, Roland (1996), "Equity and efficiency in human capital investment: the local connection", Review of Economic Studies 63 (2): 237-264.

3252 J. G. Altonji and R. M. Blank

Bergmann, Barbara R. (1974), "Occupational segregation, wages and profits when employers discriminate by race or sex", Eastern Economic Journal 1 (1,2) 103-110.

Becket, Elizabeth and Cotton M. Lindsay (1994), "Sex differences in tenure profiles: effects of shared firm specific investment", Journal of Labor Economics 12 (1): 98-118.

Becket, Gary S. (1971) The economics of discrimination, 2nd edition (The University of Chicago Press, Chicago, IL).

Becker, Gary S. (1985), "Human capital, effort and the sexual division of labor", Journal of Labor Economics 3 (l, Suppl.): $33-$58.

Becker, Gary S. (1991) A treatise on the family, enlarged edition (Harvard University Press, Cambridge, MA). Belier, A.H. (1982), "Occupational segregation by sex: determinants and changes", Journal of Human Resources

17): 371-392. Benabou, Roland (1996), "Equity and efficiency in human capital investment: the local connection", Review of

Economic Studies 63 (2): 237-264. Berryman, Sue (1983), Who will do science? (Ford Foundation). Black, Dan A. (1995), "Discrimination in an equilibrium search model", Journal of Labor Economics 13(2): 309

334. Blank, Rebecca M. (1990a), "Are part-time jobs bad jobs?" in: Gary Burtless, ed., A future of lousy jobs

(Brookings Institution, Washington, DC). Blank, Rebecca M. (1990b), "Understanding part-time work", in: Laurie Bassi and David Crawford, eds.,

Research in labor economics, Vol. 11 (JAI Press, Greenwich, CN). Blank, Rebecca M. (1991), "The effects of double-blind versus single-blind reviewing: experimental evidence

from the American Economic Review", American Economic Review 81 (5): 1041-1047. Blank, Rebecca M. (1998), "Contingent work in a changing labor market", in: Richard Freeman and Peter

Gottschalk, eds., Generating jobs (Russell Sage Foundation, New York). Blau, Francine D. (1998), "Trends in the well-being of American women (1970-1995)", Journal of Economic

Literature 36 (1): 112-165. Blau, Francine D. and Andrea H. Belier (1988), "Trends in earnings differentials by gender (1971-81)", Indus-

trial and Labor Relations Review 41 (4): 513-529. Blau, Francine D. and .Andrea H. Beller (1992), "Black-white earnings over the 1970s and 1980s: gender

differences in trends", Review of Economics and Statistics 74 (2): 276-286. Blan, Francine D. and Lawrence M. Kahn (1992), "Race and gender pay differentials", in: David Lewin, Olivia S.

Mitchell and Peter D. Sherer, eds., Research frontiers in industrial relations and human resources (Industrial Relations Research Association, Madison, WI).

Blau, Francine D. and Lawrence M. Kahn (1997), "Swimming upstream: trends in the gender wage differential in the 1980s", Journal of Labor Economics 15 (1, part 1): 1-42.

Blau, Francine D., Marianne A. Ferber and Anne E. Winkler (1998), The economics of women, men and work (prentice Hall, Englewood Cliffs, NJ).

Bratsberg, Bernt and Dek Terrell (1998), "Experience, tenure and wage growth of young black and white men", Journal of Human Resources, in press.

Borjas, G.J. and S.G. Bronars (1989), "Consumer discrimination and self-employment", Journal of Political Ecor~my 97 (3): 581-606.

Brown, Charles (1982), "The federal attack on labor market discrimination: the mouse that roared?" Research in Labor Economics 5:33 68.

Charles (198~!~ Black-White earnings ratios since the Civil Rights Act of 1964: the importance of labor Brown, market dropouts", Quarterly Journal of Economics 99: 31-44.

Brown, Charles and Mary Corcoran (1997), "Sex based differences in school content and the male/female wage gap", Journal of Labor Economics 15 (3, part 1): 431-465.

Brown, James N. (1989), "Why do wages increase with tenure? On-the-job training and life-cycle wage growth observed within firms", American Economic Review 79 (5,6): 971-992.

Bound, John and Richard B. Freeman (1992), "What went wrong? The erosion of retative earnings and

Ch. 48: Race and Gender in the Labor Market 3253

employment aurong young black men in the 1980s", Quarterly Journal of Economics 107: 201- 232.

Bound, John and Harry J. Holzer (1993), "Industrial shifts, skill levels and the labor market for white and black males", The Review of Economics and Statistics 75 (3):387-396.

Bound, John and Harry J. Holzer (1996), "Demand shifts, population adjustments and labor market outcomes during the 1980s' , Working paper no. 5685 (NBER, Cambridge, MA).

Bowlus, Audra J. and Zvi Eckstein (1998), "Discrimination and skill differences in an equilibrium search model", Unpublished paper (University of Western Ontario).

Butler, Richard and James J. Heckinan (1977), "The government's impact on the labor market status of black Americans: a critical revew' , in: Leonard Hausman et al., eds., Equal rights and industrial relations (Industrial Relations Research Association, Madison, WI).

Cain, Glen G. (1986), "The economic analysis of labor market discrimination: a survey" in: O. Ashenfelter and R. Layard, eds., Handbook of labor economics, Vol. 1 (North-Holland, Amsterdam) pp. 693-785.

Card, David and Alan B. Krueger (1992), "School quality and black-white relative earnings: a direct assess- ment ' , Quarterly Journal of Economics 107: 151-200.

Card, David and Thomas Lemieux (1994), "Changing wage structure and black-white wage differentials", The American Economic Review 84: 29-33.

Card, David and Thomas Lemieux (1996), "Wage dispersion, returns to skill and black white wage differen- tials", Journal of Econometrics 74: 319-361.

Chauvin, Keith W. and Ronald A. Ash (1994), "Gender earnings differentials in total pay, base pay and contin gent pay", Industrial and Labor Relations Review 47 (4): 634-649.

Chay, Kenneth Y. (1998), "The impact of federal civil rights policy on black economic progress: evidence l?om the Equal Opportunity Act of 1972", Industrial and Labor Relations Review 51 (4): 608-632.

Chay, Kenneth Y. and Bo Honore (1998), "Estimation of semiparametric censored regression models: an application to changes in black-white earnings inequality during the 1960s ' , Journal of Human Resources 33 (1): 4-39.

Chay, Kenneth and David S. Lee (1997) "Changes in relative wages in the 1980s: returns to observed and unobserved skills and black-white wage differentials", Mimeo. (Industrial Relations Section, Princeton University).

Coate, Stephen and Glenn Loury (1993a), "Antidiscrimination enlbrcement and the problem of patronization". American Economic Review 83(2): 92 98.

Coate, Stephen and Glenn Loury (1993b), "Will affirmative-action policies eliminate negative stereotypes?" American Economic Review 83 (5): 1220-1240.

Coleman, Mary T. and John Pencavel (1993), "Trends in market work behavior of women since t940", Industrial and Labor Relations Review 46 (4): 653-676.

Cross, Harry, G. Kenny, J. Mell and W. Zimmerman (1990), Employer hiring practices: differential treatment of Hispanic and Anglo job seekers (Urban Institute Press, Washington, DC).

Crossley, Thomas F., Stephen R.G. Jones and Peter Kuhn (1994), "Gender differences in displacement cost". Journal of Human Resources 29 (2): 461-480.

Cunningham, James S. and Nadja Zalokar (1992), "The economic progress of black women (194/)-198(/): occupational distribution and relative wages", Industrial and Labor Relations Review 45 (3): 540-555.

D'Amico, Ronald and Nan L. Maxwell (1994), "The impact of post-school joblessness on male black-white wage differentials", Industrial Relations 33 (2): 184-205.

Darity, William A. and Samuel L. Myers (1994), "The black underclass: critical essays on race and unwanted- ness" (with Emmett D. Carson and William Sabol), Critical Studies in Black Life and Culture 27: 281.

Dickens, William, Thomas J. Kane and Charles Schultze (1996), "Does the bell curve ring true? A reconsidera-- t ion' , Unpublished manuscript (Brookings Institution, Washington, DC).

Donohue, John J. III and James Heckman (1991), "Continuous versus episodic change: the impact of civil rights policy on the economic status of blacks", Journal of Economic Literature 29: 1603-1643.

3254 ,I. G. Altonji and R. M. Blank

Doiron, Denise-J. and W.-Craig Riddell (1994), "The impact of unionization on male-female earnings differ- ences in Canada", Journal of Human Resources 29 (2): 504-534.

Dm'lauf, Stephen D. (1996), "A theory ofpersistent income inequality", Journal of Economic Growth I (1): 75-93. Ehrenberg, Ronald G. and Robert S. Smith (1987), "Comparable-worth wage adjustments and female employ-

ment in the state and local sector", Journal of Labor Economics 5 (1): 43-62. Eide, Eric and Jeff Grogger (1995), "Changes in college skills and the rise in the college wage premium", Journal

of Human Resources 30 (2): 280-310. Ellwood, David T. (1986), "The spatial mismatch hypothesis: are there jobs missing in the ghetto", in R. Freeman

and H. Holzer, eds., The black youth employment crisis (University of Chicago Press, Chicago, IL) pp. 147- 190.

England, Paula (1982), "The failure of human capital theory to explain occupational sex segregation", Journal of Human Resources 17 (3): 358-370.

Even, William E. and David A. Macpherson (1990), "The gender gap in pensions and wages", Review of Economics and Statistics 72 (2): 259-265.

Even, William E. and David A. Macpherson (1993), "The decline of private-sector unionism and the gender wage gap", Journal of Human Resources 28 (2): 279-296.

Even, William E. and David A. Macpherson (1994), "Gender differences in pensions", Journal of Human Resources 29 (2): 555-587.

Farber, H. and R. Gibbons (1996), "Learning and wage dynamics", Quarterly Journal of Economics: 1007-1047. Fernandez, Roberto M. (1994), "Race, space and job accessibility: evidence from a plant relocation", Economic

Geography 70 (4): 390-417. Fields, Judith and Edward N. Wolff (1995), "Interindustry wage differentials and the gender wage gap", Indus-

trial and Labor Relations Review 49 (1): 105-120. Filer, Randall K. (1993), "The usefulness of predicted values for prior work experience in analyzing labor market

outcomes for women", Journal of Human Resources 28 (3): 519-537. Fix, Michael and Raymond J. Strnyk, eds. (1992), Clear and convincing evidence: measurement of discrimination

in America (Urban Institute Press, Lanham, MD). Fosu, Augustin Kwasi (1992), "Occupational mobility of black women (1958-1981): the impact of post-1964

antidiscrimination measures", Industrial and Labor Relations Review 45 (2): 281-294. Gerhart, Barry and Nabil E1 Cheildl (1991), "Earnings and percentage female: a longitudinal study", Industrial

Relations 30 (1): 62-78. Granovetter, Mark (1995), Getting a job: a study of contacts and careers, 2nd edition (University of Chicago

Press, Chicago, IL). Gittleman, Maury B. and David R. Howell (1995), "Changes in the structure and quality of jobs in the United

States: effects by race and gender (1973-1990)", Industrial and Labor Relations Review 48 (3): 420- 440.

Grogger, Jeff (1996), "Does school quality explain the recent black/white wage trend? Journal of Labor Econom- ics 14: 231-253.

Goldberger, Arthur S. and Charles Manski (1995), "Review article: the bell curve by Hernstein and Murray", The Journal of Economic Literature 33: 762-776.

Goldin ~laudia (1990), Understand the gender gap: an economic history of American women (Oxford University Press, Oxford, UK).

Goldin, Claudia and Cecilia Rouse (1997), "Orchestrating impartiality: the impact of 'blind' auditions on female musicians", work, ing paper no. 5903 (NBER, Cambridge, MA).

Gronan, Reuben (1988), ",Sex-related wage differentials and women's interrupted labor careers - the chicken or the egg?" Journal of Labor Economics 6 (3): 277-301.

Groshen, Erica L. (1991), "The structure of the female/male wage differential: is it who you are, what you do, or where you work?" Journal of Hmnan Resources 26 (3): 457-472.

Gruber, Jonathon (1994), "The incidence of mandated maternity benefits", American Economic Review 84 (3): 622-641.

Ch. 48: Race and Gender in the Labor Market 3255

Hamermesh, Daniel S. and Jeff E. Biddle (1994), "Beauty and the labor market", American Economic Review 84 (5): 1174.

Hashimoto, Masinori (1981), "Firm-specific human capital as a shared investment", American Economic Review 71: 475482 .

Heckman, James J., (1995), "Lessons from 'the Bell Ccurve' ", Journal of Political Economy 5 (103): 1091-1120. Heckman, James J. and Brook S. Payner (1989), "Determining the impact of federal antidiscrimination policy on

the economic status of blacks: a study of South Carolina", American Economic Review 79 (1): 138-177. Heckinan, James J. and Peter Siegelman (1992), "The Urban Institute Audit Studies: their methods and findings",

in: M. Fix and R.J. Struyck, eds., Clear and convincing evidence: measurement of discrimination in America (Urban Institute Press, Lanham, MD).

Hellerstein, Judith K. and David Neumark (1999), "Sex, wages and productivity: an empirical analysis of Israeli firm-level data", International Economic Review 40 (1): 95-123.

Hellerstein, Judith K., David Neumark and Kenneth R. Troske (1996), "Wages, productivity and worker char actelistics: evidence from plant-level production functions and wage equations", Working paper no. 5626 (NBER, Cambridge, MA).

Hellerstein, Judith K., David Neumark and Kenneth R. Troske (1997), "Market forces and sex discrimination", Working paper no. 6312 (NBER, Cambridge, MA).

Herrnstein, Richard and Charles Murray (1994), The bell curve: intelligence and class structure in American life (Free Press, New York).

Hersch, Joni (1991), "Male-female differences in hourly wages: the role of human capital, working conditions and housework", Industrial and Labor Relations Review 44 (4): 746-759.

Hersch, Joni and Leslie S. Stratton (1997), "Housework, fixed effects and wages of married workers", Journal of Human Resources 32 (2): 285-307.

Hill, Elizabeth T. (1995), "Labor market effects of women's post-school-age training", Industrial and Labor Relations Review 49 (1): 138-149.

Hill, M. Anne and Mark R. Killingsworth (1989), Comparable worth: analysis and evidence (ILR Press, Ithaca, NY).

Holzer, Harry J., 1991), "The spatial Inismatch hypothesis: what has the evidence shown?" Urban Studies 28 ( 1): 105-122.

Holzer, Harry J. and Keith R. Ihlaudteldt (1999), "Customer discrimination and employment outcomes for minority workers", Quarterly Journal of Economics, in press.

Holzer, Harry and David Neumark (1996), "Are affirmative action hires less qualified? Evidence from employer- employee data on new hires", Working paper no. 5603 (NBER, Cambridge, MA).

Houseman, Susan N. (1997), "Temporary, part-time and contract employment in the united states: new evidence from an employer survey", Unpublished manuscript (W.E. Up.john Institute for Employment Research, Kalamazoo, MI).

Ihlanfeldt, Keith R. and David L. Sjoquist (1990), "Job accessibility and racial differences in youth employment rates", American Economic Review 80 (1): 267 276.

Jacobson, Louis, Robert LaLonde and Daniel Sullivan (1993), The costs of worker displacement (W.E. Up.john Institute for Employment Research, Kalamazoo, MI).

James, F. and S.W. DelCastillo (1991) "Measuring job discrimination by private employers against young black and Hispanic males seeking entry level work in the Denver metropolitan area", Unpublished paper (Univer- sity of Colorado, Denver, CO).

Jencks, Christopher and Susan E. Mayer (1990), "Residential segregation, job proximity and black job uppor tunities" in: L. Lynn and M. McGeary, eds., Inner-city poverty in the United States (National Academy Press, Washington, DC) pp. 187-222.

Johnson, George and Gary Solon (1986), "Estimates of the direct effects of comparable worth policy", American Economic Review 76 (5): 1117-1125.

Johnson, George E. and Frank P. Stafford (1995), "A model of occupational choice with disnibutions of relative abilities", Mimeo. (Department of Economics, University of Michigan, Ann Arbor, MI).

3256 J. G. Altonji and R. M. Blank

Johnson, George E. and Frank P. Stafford (1998), "Alternative approaches to occupational exclusion", in: I. Persson and C. Jonung, eds., Women 's work and wages (Routledge, London).

Jovanovic, Boyan (1979), "Job matching and the theory of turnover", Journal of Political Economy 87: 972-990. Juhn, Chinhui (1992), "The decline of male labor market participation: the role of" declining market opportu~

nities", Quarterly Journal of Economics 107: 79-121. Juhn, Chinhui (1997), Labor market dropouts, selection bias and trends in black and white wages (Department of

Economics, University of Houston). Juhn, Chinhui, Kevin M. Murphy and Brooks Pierce (1991 a), "Accounting for the slowdown in black-white wage

convergence" in: Marvin H. Kosters, ed., Workers and their wages (AEI Press, Washington, DC). Juhn, Chinhni, Kevin M. Murphy and Robert H. Topel (1991b), "Unemployment, nonemployment and wages:

why has the natural rate increased through time?" Brookings Papers on Economic Activity 2: 75-142. Kahn, Lawrence M. (1991), "Discrimination in professional sports: a survey of the literature" Industrial and

Labor Relations Review 44:395-418. Kahn, Lawrence M. and Peter D. Sherer (1988), "Racial differences in professional basketball players' compen-

sation", Journal of Labor Economics 6: 40-61. Kahn, Shulamit (1992), "Economic implications of public-sector comparable worth: the case of San Jose,

California", Industrial Relations 31 (2): 270 291. Kain, John F. (1968), "Housing segregation, negro employment and metropolitan decentralization", Quarterly

Journal of Economics 82: 175-197. Katz, Lawrence F. and Kevin M. Murphy (1992), "Changes in relative wages (1963 1987): supply and dem~md

factors", Quarterly Journal of Economics 107 (l): 35-78. Kidd, Michael P. and Michael Shannon (1996), "Does the level of occupational aggregation affect estimates of

the gender wage gap?" Industrial and Labor Relations Review 49 (2): 317-329. Killingsworth, Mark R. (1990), The economics of comparable worth (W.E. Upjohn Institute for Employment

Research, Kalamazoo, MI). Kim, Moon Kak and Solomon W. Polachek (1994), "Panel estimates of male-female earnings functions", Journal

of Human Resources 29 (2): 406-428. King, Mary C. (I 995), "Hmnan capital and black women's occupational mobility", Industrial Relations 34 (2):

282-298. Klerman, Jacob Alex and Arleen Leibowitz (1994), "The work-employment distinction among new mothers",

Journal of Human Resources 29 (2): 277-303. Korenman, Sanders and David Neumark (1992), "Maniage, motherhood and wages", Journal of Hmnan

Resources 27 (2): 233 255. Korenman, Sanders and Susan C. Turner (1996), "Employment contacts and minority-white wage differences",

Industrial Relations 35 (1): 106-122. Korenman, Sanders and Christopher Winship (1999), "A reaualysis of the bell curve", in: Kenneth Arrow,

Samuel Bowles and Stephen Durlauf, eds., Meritocracy and economic inequality (Princeton University Press, Princeton, NJ).

Lang, Kevin (1986), "A language theory of discrimination", Quarterly Journal of Economics 101: 363-382. Lang, Kevin (1993), "Language and economists' theories of discrimination", International Journal of the Sociol-

ogy "of Language 103: 165-183. Leonard, Jonathan (1984), "The impact of affirmative action on employment", Journal of Labor Economics 2 (4):

439-463. Lewis, Gregory (19¢6~, "Gender integration of occupations in the federal civil service: extent and effects on

male-female earnings"; Industrial and Labor Relations Review 49 (3): 472-483. Light, Audrey and Manuelita Ureta (1990), "Gender differences in wages and job turnover among continuously

employed workers", American Economic Review 80 (2): .293-298. Light, Audrey and Manuelita Ureta (1992), "Panel estimates of male and female job turnover behavior: can

female nonquitters be identified?" Journal of Labor Economics 10 (2): 156-181.

Ch. 48: Race and Gender in the Labor Market 3257

Light, Audrey and Manuelita Ureta (1995), "Early career work experience and gender wage differentials", Journal of Labor Economics 13 (1): 121-154.

Loury, Glenn C. (1977), "A dynamic theory of racial income differences" in: P.A. Wallace and A.M. LaMond, eds., Women, minorities and employment discrimination (D.C. Heath and Co, Lexington, MA).

Loury, Glenn C. (1981), "Intergenerational transfers and the distribution of earnings", Econometrica 49 (4): 843- 867.

Loprest, Pamela J. (1992), "Gender differences in wage growth and job mobility", American Economic Review, Papers and Proceedings 82 (2): 526-532.

Lundberg, Shelly J. (1991), "The enforcement of equal opportunity laws under imperfect information: affirmative action and alternatives", Quarterly Journal of Economics 106 (1): 309-326.

Lundberg, Shelly J, and Richard Startz (1983), "Private discrimination and social intervention in competive labor markets", American Economic Review 73: 340-347.

Lundberg, Shelly J. and Richard Startz (1996), "Inequality and race: models and policy", Unpublished paper.

Lundberg, Shelly J. and Richard Startz (1998), "On the persistence of racila inequality", Journal of Labor Economics 16 (2): 292-324.

Lynch, Lisa M. (1992), "Private sector training and the earnings of young workers", American Economic Review 82 (1): 299-312.

Macpherson, David A. and Barry T. Hirsch (1995), "Wages and gender composition: why do women's jobs pay less?" Journal of Labor Economics 13 (3): 426471 .

Maxwell, Nan (1994), "The effect on black-white wage differences of differences in the quantity and quality of education", Industrial and Labor Relations Review 47 (2): 249-264.

McCrate, Elaine and Laura Leete (1994), "Black-white wage differences among young women (1977-86)", Industrial Relations 33 (2): 168-183.

McIntyre, Shelby J., Dennis J. Moberg and Ban'y Z. Posner (1980), "Discrimination in recrnitment: an empirical analysis: comment", Industrial and Labor Relations Review 33 (4): 543-547.

Michael, Robert T., Heidi I. Hartmann and Brigid O'Fatxell, eds. (1989), Pay equity: empirical inquiries (National Academy Press for the National Research Council, Washington, DC).

Montgomery, James D. (1991), "Social networks and labor-market outcomes: toward an economic analysis", American Economic Review 81 (5): 1408-1418.

Nardinelli, Clark and Curtis Simon (1990), "Customer racial discrimination in the market for memorabilia: the case of baseball", The Quarterly Journal of Economics 105 (3): 575-595.

Neal, Derek A. and William R. Johnson (1996), "The role of premarket factors in black-white wage ditlercnces", Journal of Political Economy 104 (5): 869-895.

Neumark, David (1996), "Sex discrimination in restaurant tilting: an audit study", The Quarterly Journal of Economics 111 (3): 915-942.

Neumark, David and Sanders Korenman (1994), "Sources of bias in women's wage equations: results using sibling data", Journal of Human Resources 29 (2): 379-405.

Newman, Jerry M. (1978), "Discrilnination in recruitment: an empirical analysis", Industrial and Labor Relations Review 32 (1): 15-23.

Oaxaca, Ronald L. and Michael L. Ransonr (1999), "Identification in detailed wage decompositions", Review of Economic Statistics 81 (1): 154-157.

Olsen, Reed Nell and Edwin A. Sexton (1996), "Gender differences in the returns to and the acquisition of on-the- job training", Industrial Relations 35 (1): 59-77.

O'Neill, June (1990), "The role of human capital in earnings differences between black and white men", Journal of Economic Perspectives 4: 25-45.

O'Neill, June and Solomon Polachek (1993), "Why the gender gap in wages narrowed in the 1980s", Journal of Labor Economics 11 (1, Part 1): 205 228.

O'Neill, June, Michael Brian and James Cunningham (1989), "Effects of comparable worth policy: evidence from Washington State", American Economic Review Papers and Proceedings 79 (2): 305-309.

3258 J. G. Altonji and R. M. Blank

Oettinger, Gerald (1996), "Statistical discrimination and the early career evolution of the black-white wage gap", Journal of Labor Economics 14 (1): 52-78.

Orazem, Peter F. and J. Peter Mattila (1990), "The implementation process of comparable worth: winners and losers", Journal of Political Economy 98 (1): 134-153.

Paglin, Morton and Anthony M. Rufolo (1990), "Heterogenous fiuman capital, occupational choice and male- female earnings differences", Journal of Labor Economics 8 (1, part 1): 123-144.

Paulin, Elizabeth A. and Jennifer M. Mellor (1996), "Gender, race and promotions within a private-sector firm", Industrial Relations 35 (2): 276-295.

Phelps, Edmund S. (1972), "The statistical theory of racism and sexism", American Economic Review 62: 659- 661.

Polachek, S.W. (1978), "Sex differences in college major", Industrial and Labor Relations Review 31 (4):498- 508.

Polachek, Solomon (1990), "Occupational self selection: a human capital approach to sex differences in occupa- tional structure", Review of Economics and Statistics 58: 60-69.

Rodgers, William and William Spriggs (1996), "What does the AFQT really measure: race, wages, schooling and the AFQT score", Review of Black Political Economy 24 (4): 13-46.

Rosen, S. (1986), "The theory of equalizing differences" in: O. Ashenfelter and R. Layard, eds., Handbook of labor economics (North-Holland, Amserdam).

Royalty, Anne Beeson (1996), "The effects of job turnover on the training of men and women", Industrial and Labor Relations Review 49 (3): 506-521.

Ruhm, Christopher J. (1999), "The economic consequences of parental leave mandates: lessons from Europe", Quarterly Journal of Economics, in press.

Ruhm, Christopher J. and Jackqueline L. Teague (1997), "Parental leave policies in Europe and North America", in: Francine Blau and Ronald Ehrenberg, eds., Gender and family issues in the workplace (Russell Sage Foundation, New York).

Schumann, Paul L., Dennis A. Ahlburg and Christine Brown Mahoney (1994), "The effects of human capital and job characteristics on pay", Journal of Human Resources 29 (2): 481-503.

Segal, Lewis M. and Daniel G. Sullivan (1997a), "The temporary labor force", Economic Perspectives 19 (2): 2 - 20.

Segal, Lewis M. and Daniel G. Sullivan (1997b), "The growth of temporary services work", Journal of Economic Perspectives 11 (2): 117-136.

Sicherman, Nachum (1996), "Gender differences in departures from a large firm", Industrial and Labor Relations Review 49 (3): 484-505.

Smith, James P. (1993), "Affirmative action and the racial wage gap", American Economic Review 83 (2): 79- 84.

Smith, James P. and Michael P. Ward (1989), "Women in the labor market and in the family", Journal of Economic Perspectives 3 (1): 9-23.

Smith, James P. and Finis Welch (1984), "Affirmative action and labor markets", Journal of Labor Economics 2 (2): 269-302.

Smith, Jmnes P. and Finis Welch (1989), "Black economic progress after Myrdal", Journal of Economic Litera- ture,27: 519-564.

Solberg, Ei'ic and Teresa Laughlin (1995), "The gender pay gap, fringe benefits and occupational crowding", Industrial and Labor Relations Review 48 (4): 692-708.

Sorenson, Elaine 0 9~0), "The crowding hypothesis and comparable worth issue", Journal of Human Resources 25 (1): 55-89. ,.

Sorenson, Elaine (1994),'Comparable worth: is it a worth policy? (Princeton University Press, Princeton, N J). Thaler, R. and S. Rosen (1975), "The value of saving a life: evidence from the labor market" in: N. Terleckyj, ed.,

Household production and consumption (National Bureau of Economic Research, New York). Thomas, Duncan (1990), "Intra-household resource allocation: an inferential approach", Journal of Human

Resources 25 (4): 635-664.

Ch. 48: Race and Gender in the Labor Market 3259

Topel, Robert H. (1991), "Specific capital, mobility and wages: wages rise with job seniority", Jom'nal of Political Economy 99: 145-176.

Topel, Robert H. and Michael Ward (1992), "Job mobility and the careers of young men", Quarterly Journal of Economics 107 (92): 439-479.

Turner, Margery A., Michael Fix and Raymond J. Struyk (1991), Opportunities deified, opportunities diminished: racial discrimination in hiring (Urban Institute Press, Washington, DC).

Vella, Francis (1993), "Nonwage benefits in a simultaneous model of wages and hours; labor supply functions of young females", Journal of Labor Economics 11 (4): 704-723.

Veum, Jonathan R. (1996), "Gender and race differences in company training", Industrial Relations 35 (1): 32- 44.

Waldfogel, Jane (1996), "The impact of the fmnily and medical leave act on coverage, leave-taking, employment and earnings", Unpublished manuscript (Columbia University).

Waldfogel, Jane (1997), "Working mothers then and now: a cross-cohort analysis of the effects of maternity leave on women's pay", in: Francine Blau and Ronald Ehrenberg, eds., Gender and family issues in the workplace (Russell Sage Foundation, New York).

Welch, Finis (1990), "The employment of black men", Journal of Labor Economics 8 (2): $26-$75. Wellington, Alison J. (1993), "Changes in the male/female wage gap (1976-85)", Journal of Human Resources

28 (2): 383Mll . Winship, Christopher and Sander Korenman (1997), "Does staying in school make you smarter? The effect of

education on IQ in the Bell curve", in: Bernie Devlin, Stephen D. Feinberg, Daniel P. Resnick and Kathryn Roeder. eds., Intelligence, genes and success: scientists respond to the Bell Curve (Springer, New York).

Wolpin, Kenneth (1992), "The determinants of black-white differences in early employment careers: search, layoffs, quits and endogeneous wage growth", Journal of Political Economy 100 (3): 535-560.

Zax, Jeffrey S. and John F. Kain (1996), "Moving to the suburbs: do relocating companies leave their black employees behind?" Journal of Labor Economics 14 (3): 472-505.

?מי מפחד לשמוע מבטא ערבי במקום העבודהמאת עידו סולומון 17.12.2009 | 07:07

לאורך השנים אני מעסיק עובדים מהמגזר הערבי מתוך . "אחד המחקרים האחרונים שערך מכון המחקר גיאוקרטוגרפיה נבע מחוויה אישית<<גיליתי שזו , ברגע שניסיתי לשלב אותם בתפקידי רוחב ולא רק כסוקרים, לצערי. "מבעלי המכון, אבי דגני' אומר פרופ, "רצון ומודעות לנושא

. משימה בלתי אפשרית

זה בא לידי ביטוי במושגים . רובם אינם שולטים מספיק בעברית וניכר שהשפה אינה שגורה בפיהם. "אומר דגני, "המחסום העיקרי הוא השפה"אני לא . שאנשים פשוט לא אוהבים לשמוע, המזרחי או הערבי -שלא לדבר על המבטא שלהם , בהגייה לא נכונה של שמות, רבים שזרים להם

". יכול להושיב אותם מצדו השני של הטלפון כי בעסק שלי הם הפנים של החברה

? אופיס שלא דורשים התמודדות עם קהל-מה לגבי תפקידי בק

חות שכתב העידו מיד שהעברית אינה שפתו"הדו. ניסיתי לעשות שיתוף פעולה עם איש מקצוע ערבי מוכר ובעל מוניטין בתחום סקרי דעת הקהל"גם כמרצה באקדמיה נתקלתי לא אחת. היו אומרים שהמכון שלי לא רציני, אם לא כולם, חלק גדול מהם, השגורה ואם הייתי מגיש אותם ללקוחותי

בעבודה קצת . כשאתה מקבל עבודה סמינריונית אתה יכול לגלות סלחנות. אבל אינם שולטים בעברית, בסטודנטים ערבים שמסיימים תואר ". פחות

כבר שנים הוא עוסק בשאלה מה מבדל את ערביי . אלא שהדברים נאמרים מדם לבו, דבריו של דגני יכולים להישמע לעתים כבעלי גוון גזעניהוא אחד החסמים העיקריים שמעכב , לטענתו, מחסום השפה. תוך חיפוש אינטנסיבי אחר הפתרונות, ישראל ומונע את שילובם בחברה היהודית

. והוא מתחיל במערכת החינוך, את השתלבות הערבים בתעסוקה ובחברה כאחד

ואילו , לומדים בבתי הספר במגזר היהודי המפוקחים על ידי משרד החינוך בעברית, אז נקבעו בחוק שתי שפות רשמיות לישראל, 1948מאז אלף ילדים בני כל 360בישראל לומדים כיום בערבית , לפי נתוני המשרד. מדובר בהחלטה גורפת שאין עליה עוררין. המגזר הערבי לומד בערבית

שבהם נלמדים השיעורים, יוצאים מן הכלל הם ארבעה בתי ספר מעורבים. כשעברית נלמדת בבתי הספר הערביים בשיעורי שפה בלבד, הגילאים . בשתי השפות

ויוצאים לשוק העבודה 18אין פלא כי רבים מבני המגזר מגיעים לגיל , כך שאם בבית מדברים הילדים בערבית ובבית הספר לומדים בערבית . כשהשפה העברית שבה יידרשו להשתמש דלה לעתים במיוחד

אינו קורא מספיק עיתונות עברית ומתבגר בנסיבות של זרות ניכרת , רוב הנוער הערבי אינו מצויד ביכולת דיבור וקריאה רהוטה דיה בעברית"לקונית או -הדבר חשוב פחות במקצועות שהכתיבה בהם מקצועית. "טוען דגני, "לתרבות ולמה שמעסיק מדי יום את רוב אזרחי ישראל, להוויה . שבהם השתלבות ערבים אכן טובה יותר, כמו רפואה, טכנית

כגון מנהלי שיווק , גם ההגייה הערבית הכבדה המאפיינת את הדיבור שלהם בעברית היא מחסום אמיתי ולעתים פסיכולוגי במילוי תפקידים רבים", תרבותית וחברתית, שהיא אוטונומיה מוסלמית, ערבית-הבעיה מתחילה באוטונומיה החינוכית. עורכי דין וחוקרים, פרסומאים, ונציגי מכירות

הדבר אינו מתיישב עם . כמו אנגלית, האוטונומיה הזו הופכת את העברית לשפה זרה ואף מזניחה שפות אחרות. ושפתה העיקרית היא ערבית, מייצרת היבדלות, גם זו החרדית והיהודית, כל אוטונומיה חינוכית. אינטגרציה תעסוקתית והשתלבות אזרחית בחברה ובהוויה הישראלית

". שבהגדרה היא ערך הופכי לאינטגרציה

? אפליה של מעסיקים

כולם ערבים , מקרב הנשאלים 40%. ערך באחרונה דגני סקר מייצג בעניין,כדי לבדוק אם תזת מחסום השפה מקובלת גם על בני המגזר הערבי

Page 1 of 3Print

12/17/2009http://www.themarker.com/ibo/misc/printFriendly.jhtml?ElementId=%2Fibo%2Freposito...

. סברו שדווקא אפליה מצד מעסיקים יהודים היא המכשלה העיקרית העומדת בינם לבין השתלבות בשוק העבודה,ישראלים

מהנשאלים טענו כי האוכלוסייה הערבית אינה משולבת מספיק בהוויה 13%; תפסה אי השליטה המספקת בעברית) 14%(את המקום השני . אמרו כי החסם הוא חשש ביטחוני מצד המעסיקים 11%-ו; ובחברה הישראלית

מהנשאלים אמרו כי הם מסכימים במידה רבה או רבה מאוד כי האוכלוסייה הערבית תשתלב טוב יותר בשוק התעסוקה הישראלי אם בבתי 51%, בעיני דגני.מהם היו מעדיפים שעברית תילמד במסגרת בתי ספר משותפים ליהודים ולערבים 75%-וכ, הספר הערביים ילמדו בשפה העברית

שנאה קל יותר ללבות מאשר . יש במנהיגות הערבית מי שמלבים שנאה בקרב הציבור הערבי, למרבה הצער. "מדובר בהוכחה ניצחת לטענתו ". אומץ לאומי לעריכת תיקון תפישתי בקשר עם מערכת החינוך

לא יעלה על . נציע לבחון את דרך השתלבותם של המיעוטים היהודיים במדינות שונות", אומר דגני, לכל מי שיתנגדו לשינוי שיטת הלימוד לעבריתלו השכלתם העיקרית היתה נרכשת במערכת , הדעת שיהודי הפזורה היו מצליחים להתערות בעולם ולהגיע למעמד חשוב וגם להצלחה כלכלית

ולימודי העברית , לימודים נוספיםעל תרבותם היהודית שומרים יהודים אלה באמצעות . ששפתה עברית, דתית או מסורתית, חינוך יהודית נבדלת ". כשהשליטה בשפה המקומית מחויבת המציאות, הם לימודי שפה זרה

"דעות קדומות: "החסם

בראש ועדת החקירה הפרלמנטרית , בין היתר, שעמד, כ אחמד טיבי"עונה בתקיפות ח, "אני ממש לא מסכים עם תוצאות הסקר ומסקנותיו" . בנושא קליטת עובדים ערבים בשירות הציבורי

, החסמים העיקריים הם דעות קדומות. רוב הערבים דוברים עברית ברמה מספקת כדי להיכנס לשוק העבודה. החסמים העיקריים אינם השפה"טק שיעור הערבים נמוך -אפילו בהיי. "אומר טיבי, "העדפה של מעסיקים יהודים לעובדים יהודים ובעיקר יחס לא שוויוני כלפי מועמדים ערבים

איש ולא תמצא 900עובדים , למשל, בבנק ישראל. למרות שכל הצעירים בוגרי המוסדות האקדמיים הרלוונטיים לתחום הם דוברי עברית מצוינתמחשבים או חשבונאות שלא יודעים עברית מספיק , האם מישהו יכול להגיד שאין מספיק ערבים שהם אנשי כלכלה. אפילו עשרה ערבים ביניהם ? טוב בשביל להיכנס לבנק

אבל אני רוצה מערכת , אני לא יוצא נגד זה. "אומר טיבי, "הרעיון של לימודים מעורבים או לימודים בעברית בבתי ספר ערבים לא מקובל עלי", שהמורשת, זה חלק מהמאפיין של מיעוט לאומי. חינוך אוטונומית בשפה הערבית עם תקציבים ואפשרויות הניתנות למערכת החינוך בעברית

באותה מידה אני תומך בהמשך לימודי העברית ומבין שחשוב שיטיבו לדבר עברית ואני גאה . זו לא בהכרח היפרדות. השפה והנרטיב חשובים לוולא ראיתי שמישהו מתלמידים , יש מספיק מקרים של בתי ספר מעורבים. שילוב בחברה הוא לאו דווקא בבתי ספר מעורבים. להיות אחד כזה

". אלה נהפך למנהל אינטל

" מעטים שולטים ברמת שפת אם, כולם יודעים עברית"

טק-רוב לקוחותיו הן חברות ישראליות גדולות ובינוניות מענפי ההיי. עיר הולדתו, וב בשפרעם'מנהל את חברת ההשמה ערבג) 37(אמיר חסון >> חסון חושב כי השליטה בעברית היא מחסום בדרך לתעסוקה . והמגע שלו עם מחפשי עבודה אקדמאים ערבים הוא על בסיס יומי, והפיננסים

. שלטענתו אינם סובלניים מספיק, אך תולה את האשמה בכך במראיינים היהודים, הולמת

זה קורה גם במקצועות שבהם השפה . הוא ייתפש אוטומטית כמועמד פחות ראוי, כשמועמד ערבי בא להתראיין ורמת העברית שלו נמוכה" . אומר חסון, "העברית אינה מרכיב עיקרי כמו מהנדס תוכנה או מקצועות טכניים אחרים

הוא מודה כי שליטתו בעברית אינה מאפיינת את. והעברית שבפיו מצוינת, תעשייה וניהול ותואר שני במינהל עסקים, לחסון תואר ראשון בהנדסהכולם יודעים . "הוא אומר, "העברית לא מספיק טובה ולעתים הם בעלי מבטא חריף, בקרב רוב הערבים. "רוב בוגרי בתי הספר התיכוניים במגזר

". אף שנולדו וגדלו פה, אבל מעטים שולטים בעברית ברמת שפת אם, את הבסיס

רוב האנשים מתחילים קריירה בלי הכוונה . האוכלוסייה הערבית מנותקת מהעולם החדש ומעידן הידע. "אינה הבעיה היחידה, לטעמו, עבריתיותר , בכלל. אין בחברה הערבית כלים לפיתוח קריירה או אפשרויות נטוורקינג. וברוב המקרים בוחרים מקצועי שלא מתחברים אליו, ומודל לחיקוי

". אבל יותר חשוב לי שישקיע באנגלית, ידבר עברית ברמה כזאת או אחרת 8-בני בן ה. מדאיג אותי נושא רמת האנגלית

" המחסום העיקרי הוא אי שליטה מספקת באנגלית"

נולדה וגדלה בקנדה וחלק מחייה , זהר-אלמגור-בריטמן-בפירמת רואי החשבון דלויט) ERS(מנהלת תחום ניהול סיכונים , הנאדי סעיד>> -שנות קריירה בתחום ההיי 20-שהחליט לחזור לאחר כ, יליד נצרת, לפני כחמש שנים היא הגיעה לישראל בעקבות בעלה. ב"הבוגרים גרה בארה

. ל"טק בחו

היא , תוך כדי לימודיה באולפן עברית הגיעה לראיון עבודה בפירמת רואי החשבון בחיפה ואף שלדבריה הראיון התנהל בעברית עילגת מצדה, אך כיום היא שולטת באנגלית, הדבר טרם קרה. תוך הבטחה שכשיירד העומס במשרד תוכל לשוב ללימודי העברית, התקבלה מיד לעבודה

. אך מרגישה שהעברית שבפיה עדיין בסיסית ולכן מעדיפה לדבר עמי באנגלית, עברית ומעט גרמנית, ערבית, צרפתית

הייתי כמעט בטוחה שזו דרכם לדחות אותי ומאוד . "היא אומרת, "ולכן היה לי ברור שלא אתקבל, בראיון העבודה התעקשו לדבר אתי בעברית"ל "אבל רוב העבודה שלי היא מול חברות שסוחרות בחו, כיום אני נדרשת לכתוב ולקרוא בעברית. הופתעתי לקבל מהם טלפון עם תשובה חיובית

עבור מועמדים , לדעתי. גם הקולגות שלי בישראל שמחים על ההזדמנות לשפשף את האנגלית. מהזמן באנגלית 99%-והתקשורת מתנהלת בזה נכון , העברית אינה המחסום העיקרי אלא דווקא אי שליטה מספקת באנגלית וכישורי מחשב, במיוחד כאלה שלמדו בישראל, ערבים אקדמאים

". במיוחד עבור כאלה שמנסים לעבוד בחברות גלובליות

" ?את מי יביא המועמד הערבי, יביא את חבריו מהיחידה 8200-בחור מ"

Page 2 of 3Print

12/17/2009http://www.themarker.com/ibo/misc/printFriendly.jhtml?ElementId=%2Fibo%2Freposito...

ששימש עד לאחרונה יועץ בכיר בחברת הייעוץ מקינזי וכיום מנהל את ,אומר חביב חזאן,"עבור אקדמאים ערבים העברית אינה מהווה חסם">> . קרן אל בוואדר המשקיעה במגזר הערבי בישראל

או שהעברית שלו טובה מספיק או שהעברית לא רלוונטית להצטיינות : אם סטודנט ערבי מסיים את הטכניון בהצטיינות יש שתי אפשרויות"כמו הרתיעה של מעסיקים יהודים מהעסקת ערבים או המרחק , חמורים בהרבה, יש חסמים אחרים. אין אפשרות שלישית. במקצוע שבחר

, מעבר לכך. רוב המשרות נמצאות דווקא במרכז, בעוד שרוב האקדמאים הערבים גרים בצפון ולא מוכנים לשנות את מקום מגוריהם. הגיאוגרפילית שיווק בטבע או הדוד שהיה ראש חטיבה "אין לנו את האמא שהיתה סמנכ, לרוב הערבים אין היכרות עמוקה עם שוק התעסוקה הישראלי

". ?ואת מי יביא המועמד הערבי, ם יביא את חבריו מהיחידה"או ממר 8200-אז בחור מ. טק מדברים כל הזמן על חבר מביא חבר-בהיי. בבנק

עברית הוא למד כמו . לחביב תואר ראשון ושני במשפטים מהאוניברסיטה העברית ותואר שני במינהל עסקים מאוניברסיטת אימורי שבאטלנטה . אך לדבריו לא הסתפק בחשיפה לשפה רק בין כותלי בית הספר', רבים מחבריו החל בכיתה ג

כשאתה מגיע לאוניברסיטה ופוגש אנשים שעברית היא שפת אמם אתה , הייתי ילד סקרן אז נחשפתי לעברית גם דרך עיתונים וטלוויזיה ועדיין"ובטח לאחר שאתה מתחיל ' סטאז, אחרי ארבע שנות לימודים. גבוהה ולא מעודכנת, מבין שהעברית שרכשנו בבית הספר היא לעתים ספרותית

ובמקום לראות גיא פינס הם , הבעיה היא שלילדים של היום יש מספיק ערוצי טלוויזיה בערבית. גם הפער הזה מצטמצם, להשתפשף בעבודה ". יעדיפו את המקביל הערבי שלו

Page 3 of 3Print

12/17/2009http://www.themarker.com/ibo/misc/printFriendly.jhtml?ElementId=%2Fibo%2Freposito...

Why Do Arabs Earn Less than Jews in Israel?

By

Muhammad Asali Columbia University, New York

Abstract

This paper uses fourteen years of income data between 1990 and 2003 to measure wage

differentials between Israeli Arab and Jewish workers in Israel. The wage gap it discovers is

decomposed into components corresponding to human capital, occupational segregation, selectivity,

and a residual, which may reflect discrimination. The unadjusted hourly wage gap between Arab

and Jewish workers almost doubled from 40% in 1990 to 77% in 1999. By 2003, however, it had

declined to 56%. The general picture is for inequality in salaries to have exacerbated sharply during

the 90s, along with diminished opportunities in a skilled and integrated workforce for Arabs, before

easing to some degree in the first three years of this decade. This paper sets out to explore a range

of explanations for these trends.

An earlier version of this paper was supervised by Michael Beenstock and was the author's

MA thesis. The paper forms part of the Falk Institute Policy Focus Group on key economic

policy issues in Israel.

The Maurice Falk Institute for Economic Research in Israel

Jerusalem, March 2006 • Discussion Paper No. 06.03

Why Do Arabs Earn Less than Jews in Israel?

1 Introduction

Israel is a multicultural, multiethnic society. Its population brings together Western and

Eastern Jews, foreign- and locally-born citizens, and Arabs from a range of Muslim, Chris-

tian, and Druze religious and ethnic backgrounds. Arab Israeli citizens constitute about

20% of the total population;1 yet despite extensive studies of ethnic wage disparities in Is-

rael (e.g., Neuman and Silber (1996), Neuman and Oaxaca (1998), and Neuman and Oaxaca

(2004b)), very little attention has been paid speci�cally to the characteristics of this group

as workforce participants.

The objective of this paper is to measure and document the evolution of wage gaps

between Arabs and Jews in the Israeli labor market in the years 1990�2003, aiming to

characterize and evaluate the di¤erent mechanisms according to which these gaps may be

said to have developed.

The existence of an observable wage gap in itself, though, does not count as su¢ cient proof

that a certain labor market is marked by discrimination. In order to separate out the e¤ects

of discrimination from those of potentially unrelated factors, this study used a modi�ed form

0I owe special thanks to Dan O�Flaherty, Janet Currie, and Nachum Sicherman for their continuoussupport and invaluable comments. The helpful comments of an anonymous referee are also gratefully ac-knowledged. Michael Beenstock, Joseph Zeira, Lena Edlund, and Till von Wachter o¤ered helpful commentsand discussions. Thanks must also go to seminar participants at Columbia University, and to the Falk Insti-tute for Economic Research in Israel and the Israel Social Sciences Data Center, at the Hebrew Universityof Jerusalem, for providing this study�s data.

1Arabs mentioned here are citizens and residents of Israel. They live and work in Israel, have Israelicitizenship, and share the same national institutions with the Jewish citizens. Palestinians, living in theWest Bank and Gaza Strip, are not the subject of this paper.

1

of the Oaxaca-Blinder decomposition in order to disaggregate observed wage gaps into three

components: those owing to human capital di¤erences, to occupational segregation, and to

discrimination (or unexplained gap).

As well as recording the Jewish-Arab income gap in gross terms, this study also set out

to analyze patterns of change between the di¤erent wage gap components over the years of

the study�s data. It was found that the overall shape or trend of the changes between the

study�s explanatory categories for the Jewish-Arab wage di¤erential were robust to di¤erent

model speci�cations and underlying assumptions, even when speci�c levels of income gap

components varied in more unexpected ways.

The Jewish-Arab male hourly wage gap hovered at around 45% (of Arab hourly wage) in

the years 1990�1994, going on to peak at 77% in 1999. Since then, the hourly wage gap has

started to decrease, falling to a level of 56% by the end of 2003. The unexplained component

of the gap (regularly interpreted as discriminatory) accounted only for 5%�10%, or less in

other instances, of the overall wage gap in the years 1990�1991. However, since 1992, this

component has accounted for an increasing portion (about 25%�38%) of the overall wage

gap. Occupational segregation explained a portion of 35%�80% of the overall wage gap over

the entire period.

It is important to seek to understand these �uctuations in the Jewish-Arab income gap

against the background of the comprehensive changes undergone by the Israeli economy as

a whole between the years 1990�2003. 1990 saw the arrival in the country of some 200,000

immigrants, with 176,000 arriving in 1991. This massive in�ow of immigrants continued at a

2

yearly rate of about 77,000 until 1995; with numbers subsequently starting to fall to a point

where only 23,000 immigrants entered the country in 2003.2

Besides being boosted by this new pool of available and sometimes skilled labor, the Israeli

economy was also fed over the 90s by a large in�ux of temporary foreign workers. Those

workers were �imported�by the state as a solution to the immigration-driven demand for

construction workers and the shortage of Palestinian workers due to closures i.e., restrictions

on Palestinian workers�freedom of movement imposed by Israel. By 1995 there were 92,000

non-Israeli nationals working in the country (or about 4.7% of the total employed workforce).

This number only ceased rising subsequently in 2002 at 232,000 counted foreign employees

(or 10.2% of the total employed workforce).

The wider socio-political context of changes in Israel�s labor force composition over the

study period is of the Jewish-Arab peace process, set in train with the 1993 signing of the

Oslo accords but then dealt a fatal blow by the assassination of the Prime Minister Yitzhak

Rabin in 1995 and the onset of second Intifada in September 2000. Within pre-1967 Israel,

the so-called �2000-events�(referring to the killing of 13 Israeli-Arab rioters by Israeli soldiers)

also in�uenced the climate in which employment (hiring, �ring, and training decisions) were

made with respect to Jews and Arabs. It has to be supposed that events in this socio-political

register will have had a direct e¤ect on labor market outcomes, especially as these pertain

to Jewish-Arab (wage) disparities.

Given the inseparability of social, political, and economic factors, it is di¢ cult to come

2Israel�s total population (excluding the Gaza Strip and West Bank) was estimated at around 5 millionin the early 1990s and has steadily increased since then. In 2003 the estimated population was 6.7 million.See the Statistical Abstract of Israel, Central Bureau of Statistics (2005).

3

to conclusive explanations as to the changing pattern of income inequalities between Arabs

and Jews over the study period. While not claiming to be incontrovertible, this paper

o¤ers a series of interpretations of the phenomenon in terms of the changing demographic

composition of the workforce, changing perceptions of ethnicity on the part of both Arabs

and Jews, and shifting skills shortages in the broader Israeli economy and speci�cally in the

Arab population compared to the Jewish (Arabs�human capital de�cit). It is hoped that the

paper�s terms and explanations will stimulate further research examining the study�s issues

from di¤erent perspectives and in greater depth.

The paper is organized as follows: the next section describes the methodology used to

measure wage gaps and their decompositions. Section 3 features a detailed description of

the study data and de�nes the paper�s terms as regards its explanatory categories for the

causes of income gaps. Summary statistics for the relevant variables are found in section 3.1.

The main results of the study are in turn reported in section 4, after which section 5 spells

out the paper�s analysis of the changing shares in the Jewish-Arab income gap attributable

to the decomposed causal mechanisms. Policy implications are discussed in this section as

well.

2 Methodology

In an attempt to quantify the coe¢ cient of (wage) discrimination [Becker (1957)], Oaxaca

(1973) and Blinder (1973) proposed a simple, yet reliable, estimator based on a straightfor-

4

ward OLS estimation.

This paper adopts the Oaxaca-Blinder decomposition method in the context of an analy-

sis of wage gaps between Arab and Jewish workers, which are broken down into �rst, a

component representing more �neutral� di¤erentials in human capital, and second into a

component normatively representing the e¤ect of discrimination. My approach in e¤ect-

ing this decomposition broadly follows the more general treatment of wage decomposition

proposed by Neumark (1988) and Oaxaca and Ransom (1988, 1994).

Simply put, Oaxaca-Blinder decomposition entails running separate OLS wage regres-

sions, for each of the two groups under consideration, and comparing the means and es-

timated coe¢ cients of the variables from those regressions. A di¤erence in the average

productivity-related variables, weighted by the estimated coe¢ cients of the nondiscrimina-

tory wage standard, is attributed to the human capital (or explained) portion of the overall

wage gap. Any remaining (unexplained) wage gap, measured by the di¤erence in the esti-

mated coe¢ cients, evaluated at some level of average characteristics, may then be referred

to discrimination. We can represent this mathematically as follows:

Let the wage equation for any individual i in the ethnic group j (j =Arab, Jew) be

lnWij = X0ij�j + "ij (1)

where Wij is the hourly wage, Xij is a vector of worker characteristics, and "ij a zero-mean,

constant-variance error term. Then, the wage equations, estimated by OLS at the mean

point, will be

5

lnWj = �X 0j�̂j; for j = A; J (2)

where A and J subscripts stand for Arabs and Jews and upper bars signify averages of the

di¤erent variables. The regressors vector Xij in this paper includes variables for years of

schooling, experience, experience squared, a marital status dummy, a full-time employment

dummy, a large city dummy, one- and two-digits occupational dummies, and one- and two-

digits industrial dummies.3 Income survey for the year 1990 includes a categorial, rather

than a continuous, schooling variable. Consequently, analyzing the data for this year, I

use a set of schooling dummy variables, and use age and age-squared in lieu of experience

and experience squared. Nonetheless, the results of the decomposition and wage equations

remain meaningful and robust to comparison, so far as wage di¤erentials are concerned.

Writing the wage equation in 2 for each group and di¤erencing those, with mild arithmetic

transformation,4 yields the following wage gap decomposition:

lnWJ � lnWA = ln (1 +G) =��XJ � �XA

�0�̂�| {z }

Q

+ �X 0J

��̂J � �̂

��+ �X 0

A

��̂� � �̂A

�| {z }

D

(3)

where �̂�is the estimate of the nondiscriminatory wage coe¢ cients, and G is the gross (geo-

metric) wage gap. The �rst term in equation 3, Q, represents the human capital component

of the overall wage gap; D, the sum of the second and third terms, is the discriminatory

3The Income Surveys do not provide a direct measure of labor market experience. Consequently, I usepotential experience, de�ned as: Experience = Age� Schooling � 5:

4Namely adding and subtracting the term��XJ � �XA

�0�̂�:

6

component. The decomposition method performed in equation 3 is the general Oaxaca de-

composition, Oaxaca and Ransom (1994). If we assume that �̂�= �̂J or �̂

�= �̂A; then the

general Oaxaca decomposition reduces to the classical Oaxaca-Blinder decomposition.

I carry out these analyses under two di¤erent assumptions about the nondiscriminatory

wage standard ��. First, as in the classical Oaxaca-Blinder decomposition, I adopt the

estimated wage structure of the dominant group as the nondiscriminatory standard, i.e., �̂�=

�̂J : That is, in the absence of labor market discrimination, it is taken that the current Jewish

wage structure would apply to both Jews and Arabs. Alternatively, as in the general Oaxaca

decomposition, I assume that �̂�is equal to the characteristics-weighted wage coe¢ cients, as

this is shown to be equal to the estimated wage coe¢ cients from a simple pooled regression

that includes both Arabs and Jews (see Oaxaca and Ransom (1994)).

For the dataset in question, the average income varies widely across occupations (as well

as across ethnic groups). Therefore, even in the absence of unexplained wage gaps within

occupations, wage di¤erences could still exist according to di¤erent distributions of Arab and

Jewish workers across employment sectors. While controlling for occupational and industrial

a¢ liation in the wage regressions would eliminate inter-occupational wage gaps, it would also

have the e¤ect of underestimating the discriminatory component of the overall wage gap, to

the extent that occupational segregation, at least in part, derives from discrimination. This

could be through di¤erent barriers to entry o¤ered to representatives of the two groups. To

show this, let CAi�CJi�be the proportion of Arabs (Jews) employed in occupation i;5 andWA

i

5Note that this proportion is equal to the average of the corresponding dummy variable in the wageequation.

7

�W Ji

�the mean (log) hourly-wage of an Arab (Jewish) worker in occupation i (i = 1; 2; :::; I).

It then follows that: �W j =P

iCjiW

ji for j = A; J and

ln(G+ 1) =

IXi=1

�CJi W

Ji � CAi WA

i

�(4)

Therefore, even if we assume that there are no wage di¤erences between workers within each

occupation (i.e., W Ji = W

Ai = Wi 8i), wage gaps may still arise if the two populations have

di¤erent occupational distributions (i.e., CJi 6= CAi for some i), since, according to equation

4, the wage gap will be equal toP

i

�CJi � CAi

�Wi which will not be zero other than in the

cases CAi = CJi for all i (i.e., identical occupational distribution), or Wi = W for all i (i.e.,

equivalent average wages in all occupations for members of Arab and Jewish groups).

The above illustration indicates the possibility that including occupational dummies in

the wage regressions may result in underestimating the discriminatory component of the

overall wage gap, in this sense that the added variables would disguise labor market discrim-

ination as a human capital component. In seeking to account for i.e. decompose the wage

gap, this paper assesses the above possibility in two ways. First, I compare the estimated

coe¢ cient of the Jewish dummy in a pooled wage regression with and without occupational

dummies. Second, regression analysis includes occupational (and industrial) dummy vari-

ables in Oaxaca decompositions, comparing these results with those obtained without the

occupational dummies.

Oaxaca decomposition does not account for group di¤erences in group members�occu-

pational distribution. Di¤erent methods were suggested to measure the occupational segre-

8

gation component of the wage gap. For example, Brown et al. (1980) and Miller (1987) use

a multinomial logit model to estimate the distribution of one group across occupations, and

compare this with the observed distribution of the other group. Neuman and Silber (1996),

alternatively, estimate the occupational segregation component of a wage gap by comparing

each group�s share in a certain occupation with the share of that occupation in the total

employed labor force, and sum the di¤erences over all occupations.

This paper introduces a new, yet closely related, method to measure the component

of the wage gap attributable to occupational (and industrial) segregation. I include a set

of occupational dummies in the wage equations, and modify the Oaxaca decomposition by

dissociating the part explained by those dummies from the human capital component. Let

the estimated wage equation be

lnWj = �X 0j�̂j + C

0j ̂j;

where Cj is a vector of average occupational dummies for the group j = A; J: De�ning

Z 0j := [X0j; C

0j] and �̂ := [�̂

0; ̂0]0 I arrive at the following decomposition:

ln (1 +G) =��XJ � �XA

�0�̂�| {z }

Q

+h�Z 0J

��̂J � �̂

��+ �Z 0A

��̂� � �̂A

�i| {z }

D

+�CJ � CA

�0 ̂�| {z }

S

(5)

Equation 5 accommodates the e¤ect that di¤erent occupational distributions have on wage

gaps. The �rst two terms in this equation, Q and D, are the familiar human capital and

discrimination components. The last term, S, representing di¤erences in the occupational

9

distribution weighted by a nondiscriminatory norm, takes its place as the occupational seg-

regation component.6

Although estimating the individual contributions of sets of dummy variables to the un-

explained component of the wage gap may produce arbitrary results, since those depend

on the dropped category, it is true that the overall decomposition and estimated separate

contributions of dummy variables to the explained component are consistent and invariant

to any choice of the dropped category, Oaxaca and Ransom (1999). This fact makes possible

the isolation of the occupational distribution e¤ect from the human capital component. Note

that, in the absence of occupational segregation (i.e., CJ = CA), S = 0 and that the only

way in which occupations may a¤ect the decomposition is by adding the term C 0 ( ̂J � ̂A)

(the within-occupation di¤erences) to the discriminatory segment D. Also, if all occupations

share the same wage structure for Arabs and Jews (i.e., ̂A = ̂J), then the discriminatory

component due to within-occupation gaps disappears, although the segregation component,

manifesting di¤erent occupational distributions, will remain. Lastly, if all occupations pay

the same wage (as the dropped category, i.e., ̂ = 0), then equation 5 reduces exactly to equa-

tion 3, in which case the estimates of the human capital component and the discriminatory

component are not a¤ected by group di¤erences in occupational distribution.

In the appendix, section 6.2, I refer to the bias, in the wage gap decomposition, resulting

from possible self-selection into employment. It is possible for self-selection to have only

6Industrial dummies, when included in the wage equations, are treated in the same way as occupationaldummies. C will be a vector of not only occupational but also industrial dummies, and its respective vectorof coe¢ cients. The S component will represent both the occupational segregation and industrial segregationcomponents. Added together, S will be called the labor market segregation measure throughout the paper,for simplicity.

10

a mild e¤ect, if any, on the human capital or occupational segregation components of the

wage gap, since productivity-related variables are not a¤ected by the measurement method.

On the other hand, the unexplained component may be greatly a¤ected by the selectivity

correction, since this whole component relies on our estimators of wage regressions, which are

liable to change markedly on account of this correction. However, correcting for selectivity,

in all its variations and e¤ects, lies beyond the scope of this paper. While selection into

employment may result in bias in the decomposition estimates, it is at any rate not the

only source of disturbance. Selection into a speci�c occupation, for those who are already in

the labor market, represents another dimension of the problem. Although the occupational

segregation component is estimated consistently in this paper, without knowing the e¤ect

of occupational self-selection e¤ect we cannot identify that part of occupational segregation

which represents labor market discrimination (or barriers to entry in certain occupations for

members of the Arab ethnic group).

Two important points are worth emphasizing before we leave this section. First, the

paper refers to the unexplained component of the wage gap as �discriminatory.� However,

since the choice of explanatory variables can greatly a¤ect the results of decomposition,

the unintentional exclusion of certain relevant variables from the wage regressions may bias

�ndings related to the unexplained (or discriminatory) component.7 While it seems beyond

7The possibility of decomposition results being heavily in�uenced by the choice of regressors to beincluded was originally raised by Oaxaca:

It is clear that the magnitude of the estimated e¤ects of discrimination crucially depends uponthe choice of control variables for the wage regressions. A researcher�s choice of control variablesimplicitly reveals his or her attitude toward what constitutes discrimination in the labor market.Oaxaca (1973), p. 699.

11

reasonable doubt that some fraction of the unexplained component derives from labor market

discrimination, it is equally the case that some other fraction will be owing to other factors.

Such a fraction may merely signify the e¤ect of uncontrolled-for variables. Therefore, my

terminology refers to theD component as the unexplained component of the wage gap, rather

than the discriminatory component. In this sense, this component may be understood as an

upper bound for labor market discrimination.

Secondly, it has been noted that occupational segregation can be the result of labor

market discrimination.8 Now, this labor market segregation may re�ect barriers to entry into

well-paying jobs, but can only exist more neutrally as a manifestation of di¤erent preferences.

Some people may prefer to work in low-paying occupations. In this paper I do not further

decompose the occupational segregation component into self-selection on the one hand and

discrimination on the other. Therefore, the labor market segregation component as discussed

in this paper should be viewed as a compound of the two e¤ects. The issues of occupational

selection, and its e¤ect on measuring discrimination in the terms of economics, represent a

fertile area to be addressed in future research.

8As noted by Neuman and Silber (1996) (p.651, n.3), occupational segregation represents another dimen-sion of labor market discrimination. Segregation and barriers to entrance based solely on ethnic a¢ liation,other things being equal, should be viewed as discriminatory.Neumark (1988) expresses the same concern in other words:

The question of whether industry or occupation dummy variables should be included in regres-sions to estimate wage discrimination hinges on the extent to which the distribution of menand women across industries and occupations is itself a result of discrimination. p. 291.

12

3 Data

This paper�s data are drawn from the yearly income surveys conducted by the Central

Bureau of Statistics in Israel for the years 1990�2003. Income surveys are based on question-

naires conducted on household and individual levels and cover information on demographic,

personal, and labor market characteristics. The samples include Jewish and non-Jewish re-

spondents living exclusively in Israel (latterly including disputed East Jerusalem, but not

the Israeli occupied territories of the West Bank and Gaza Strip). Hence, all respondents in

the income survey, both Arabs and Jews, are residents and citizens of Israel.

Among the variables included in the analyses of the study data are: full-time employment

(in the form of a dummy which takes on the value 1 if the worker works at least 35 hours

a week, and zero otherwise); marital status (as a dummy which takes on the value zero if

the worker is single and 1 otherwise); and urban/non-urban location (a large city dummy

takes on the value 1 if the city of residence is Tel-Aviv, Haifa, or Jerusalem, and zero

otherwise). Occupational and industrial a¢ liation are coded according to the one- and two-

digits classi�cation variable.

In investigating the Jewish-Arab wage di¤erentials in the Israeli labor market I limit my

analysis to salaried, prime-aged (25�65) male workers. Worker�s hourly wage is calculated

by dividing monthly income by the product of hours worked per week and working weeks

per month. I deal with outliers, in terms of hourly wage, by dropping observations below

the 1st and above the 99th percentile of the log hourly wage distribution for each year. This

procedure is more robust and meaningful than dropping observations on a given currency

13

(New Israeli Shekel, NIS) cuto¤point, since the analyses involve di¤erent years, from 1990 to

2003, over which the currency value is not comparable. Moreover, this procedure circumvents

the problem of an a priori imposed NIS cuto¤ point by accommodating changes in the wage

distribution over period years (see Chandra (2000)).

Israel remains an immigrant society in the sense that incomers, including migrant workers,

constitute a large portion of its population. Therefore, it is important to include in any

analysis of labor market discrimination a treatment of the working participation in the

Israeli economy of foreign-born citizens. Since it is conceivable that labor markets outcomes

may re�ect, for instance, the displacement of resident workers by immigrants, it is correct

for analysis to consider incoming Jews�wages integrally with the other members of their

cohort. At the same time, though, it is also the case that for some Arabs and Jews, local

birth or longstanding assimilation can o¤er labor market participants advantages over recent

arrivals. To integrate these opposing considerations in my analyses, I exclude from certain

of my analyses recent immigrants (who arrived in the last 10�19 years).9 Hereafter, I refer

to the sample excluding these later immigrants as the reduced sample. As a benchmark, and

providing a basis for comparison, I carry out parallel analyses on the basis of the full sample,

excluding no worker or recent immigrant.

9I use an exclusion rule for immigrants based on a range of years (10�19) since arrival, rather than a�xed cuto¤ point, since the exact year of arrival is not always provided in the data. In the reduced samples,for each of the years 1990�1999, I excluded immigrants who arrived after 1980; and for each of the years2000�2003 I excluded immigrants who arrived after 1989.

14

3.1 Descriptive Statistics

Averages of the most important variables, for each investigated year, are given in the Tables

1 and 2. Table 1 describes the sample means after applying all the selection rules described

above but before excluding immigrants; in other words, it pertains to the full sample. Table

2 describes the mean characteristics of the reduced sample, that is, after excluding all the

newly arriving immigrants.10

In some of the years the omission of immigrants from the samples reduces (by very little)

the number of observations of Arabs. This may happen due to the non-one-to-one relation

between the �non-Jews,� as de�ned in the data, and Arabs. This non-coincidence makes

the interpretation of comparisons between the full and reduced samples more problematic,

though not to any great extent. It remains true that results from the reduced sample are

more robust and straightforward to interpret, since excluding new immigrants results in an

(almost) perfect match between the two de�nitions.

Another peculiarity of the data should be noted. In 1997 the Income Survey sample

design changed to cover the rural population as well as urban households. Respondents were

thereafter also drawn from East Jerusalem, while Kibbutzim, institutions, and groups of

Bedouins living outside localities are still absent from income surveys. The dataset for the

year 1997, serving as the linking year, has been produced in two versions: an old version

10The 1990 dataset includes a categorial, rather than a continuous, schooling variable. This means thatrelevant �gures are not provided in Tables 1 and 2. In this year, 44.5% of the Arab workers fell into the0�8 years of schooling group, as opposed to 14.6% of the Jewish workers (or 14.8% in the reduced sample).37.8% of the Arab workers had 9�12 years of schooling, versus 46.9%�47.9% of the Jewish workers. Only17.5% of the Arab workers had more than 12 years of schooling, while among the Jewish workers this �gurewas 37.3%�38.4%.

15

according to the old estimation method and sample design, and a new version based on a

wider population and deploying the new method of estimation. In Tables 1 and 2 (and in

all subsequent tables) I refer to the old 1997 version as 1997a, and to the new as 1997b. We

may note also that, after applying the new sample design in 1997, samples doubled in size.

4 Results

In Table 3, I provide estimates for the Jewish dummy coe¢ cient in the pooled wage equations,

from both the full and reduced sample. This is a measure of the overall wage gap� after

controlling for di¤erent variables. The dependent variable in all speci�cations is the logarithm

of individual hourly wage.

In column 1 the only regressor is the Jewish dummy, hence, the reported coe¢ cient mea-

sures simply the overall unadjusted (logarithmic) wage gap.11 The speci�cation in column

2 includes, beside the Jewish dummy, a years-of-schooling variable, a potential experience

variable, a squared potential experience variable, a marital status dummy (0 if the individual

is single and 1 otherwise), and a full-time employment dummy.12 Therefore, in speci�cation

2, the given coe¢ cient measures the adjusted wage gap, controlling for these variables insofar

11In this paper, the term �wage gap�designates the di¤erence in the average of logarithmic hourly wagebetween Jewish and Arab workers. However, it is important to note that this gap is no more than anapproximation to (and is less than) the geometric wage gap, which, in general, is again less than the observedhourly wage gap. For example, in the reduced sample of 2003 (see Table 3) the gross wage gap is 0.4056;however, this translates to a 0.5002 wage gap in geometric means. In Table 2, further, we see that the actualwage gap is 0.5559, or 55.59%, which is far higher than the initially reported 40.56% �gure.

12For the year 1990, since the data only provide a categorial schooling variable, I use a set of schoolingdummies replacing the years-of-schooling variable. Likewise, analysis could not rely on experience andexperience squared �gures, which were replaced by variables for age and age squared.

16

as they are deemed relevant to productivity. This coe¢ cient is not intended as a measure of

�discrimination�since, by construction, it imposes the assumption that the individual char-

acteristics of Arabs and Jews are similarly rewarded in the labor market. It is important to

recognize this as an (unlikely) assumption. The speci�cation in column 3 extends that in 2

by adding a set of one-digit occupational dummies.

Two facts are evident from Table 3. First, the overall wage gap, as measured by the

Jewish dummy coe¢ cient, is higher in the reduced sample than in the full sample, under

all the di¤erent speci�cations. This supports the claim that the relatively well-absorbed

immigrants in the (Israeli) labor market perform better than more recent arrivals. Hence,

excluding the recent immigrants� as in the reduced sample� yields a higher measure of the

inter-ethnic wage gap. At issue here are locally-born and comparatively assimilated Jews, as

against these cohorts plus Jewish immigrants. Second, it is apparent that introducing our

productivity-related variables into the wage equations greatly reduces the measured wage

gap (by 59%�98%).

It is also evident from Table 3 that wage gaps increased vastly and monotonically up

until 1999, when they started to decline. The new sample design, applied in 1997, resulted

in a higher wage gap in all speci�cations, as is evident from Table 3. The change in the

sample design, in that it began counting inhabitants of rural areas and East Jerusalem, was

expected to increase the measured wage gap, since it made available a comparison with a

greater proportion of less advantaged Arabs.

Tables 4�8 document the main �ndings of the paper. They present the overall wage

17

gap decompositions, according to the techniques discussed earlier in the paper, and under

di¤erent assumptions.13 For each sample, the results proved robust to di¤erent model speci-

�cations and assumptions as to the nondiscriminatory wage structure. Table 4 is predicated

on an equation between the estimated wage structure of Jewish workers and a nondiscrim-

inatory wage norm, i.e., �̂�= �̂J ; analyses in this table pertain to the full sample. Table 6

presents similar analyses as performed on the reduced sample. In Tables 5 and 7 analyses

are performed, for the full and reduced samples, under the assumption that �̂�= �̂pooled, i.e.,

that the nondiscriminatory wage norm is equal to the estimated wage structure taken from

the pooled regression, including both Arabs and Jews. All the estimates in those tables are

signi�cant at the 5% signi�cance level.

Results from both full and reduced samples are reported for the sake of completeness.

Nonetheless, I focus attention on results from the reduced sample, which, due to its special

composition, better serves the analyses of wage gaps, as yielding a less ambiguous picture of

di¤erences in income between (native or assimilated) Arab and Jewish labor market partic-

ipants. My description of data �ndings, and subsequent inferences, will be con�ned to the

assumption that �̂�= �̂J :

To summarize our �ndings, in the early to mid 1990s, the hourly wage gap hovered

between 40% and 50%.14 In 1995, the hourly wage gap recorded a level of 50% beneath

13Note that any di¤erence in G across di¤erent tables, when pertaining to the same sample, derivesfrom the absence of available values for the added explanatory variables. For example, in Table 6, we haveG = 0:4057 for the year 2003. However, when we add a richer set of variables, as in Table 8, G becomes0.4017 for the very same year.

14Figures are calculated from Table 2. For example, the hourly wage gap in 1990 is: 36:9=26:2 � 1 =0:408 4; or 40.8%.

18

which it has yet to dip. The hourly wage gap peaked in 1999 at a level of 77%; since then,

it has followed a steady path downwards, reaching a level of 56% in 2003.

While trends in the human capital and unexplained components of the gap are similar

under di¤erent assumptions as to the discriminatoriness of wage structure for the two groups,

it is worth noting that di¤erences in human capital explain a higher portion of the wage

gap when the pooled wage structure is assumed to be nondiscriminatory. Bearing this

in mind, my account of the evolution of wage gaps, and their decomposition, assumes an

approximation of nondiscriminatory wage structure to the wage structure manifested by the

dominant group (Jews). Further, I con�ne my attention to the reduced sample excluding

recently arriving immigrants.

Not controlling for occupational and industrial a¢ liation, productivity-related variables

could explain 0.244 of the gross wage gap in 1990 (0.269� see Table 6). That is, only a

minor gap of 0.025 remains unexplained when di¤erences in those variables are taken into

account. When one-digit occupational dummies are included in the wage regressions, this

portion declines to 0.018. When regressions are further extended to include more speci�c

occupational and industrial dummies (two-digits), and a large city dummy, as in Table 8,

this portion becomes -0.03, i.e., the factoring-in of the extended form of these productivity-

related variables not only explains the wage gap, but suggests that Arabs are favored, given

their characteristics and occupational choice. Nonetheless, within this extended analysis, a

new component emerges, that pertaining to labor market segregation. This accounted for a

wage gap of 0.05 and 0.134 in the short and general speci�cations respectively. As argued

19

in the Methodology section, this component may well represent a form of labor market

discrimination, manifested in barriers to entry for higher-paying professions; on the other

hand, it is likely that part of this component is attributable to the di¤erent self-selection of

groups into low paying occupations.

The same sets of analyses are then carried out for the other years of the study data, with

a similar pattern of results emerging. A greater measure of speci�city or detail in terms of

the productivity-related variables tends to explain a greater portion of the wage gap, and

therefore reduces its �unexplained�component. Given that axiomatically only part of the

unexplained wage gap measures discrimination in the labor market, it then becomes reason-

able to regard the unexplained wage gap in the most general or multivariate speci�cation

(as in Table 8) as a ceiling on the extent of within-occupation wage discrimination.

Over the whole period, there was a noticed convergence in some of the important productivity-

related variables, such as schooling, age, and marital status. The schooling gap between the

groups declined from 3.1 in 1991 to 2.4 in 2003; the di¤erence in the average age of the

workers declined from 5.6 in 1990 to 4.5 in 2003; and the di¤erence in the marriage rate

among workers declined from 10% (with a greater proportion of Jews married) to -2% (with

a greater proportion of Arabs married). Despite this convergence, human capital di¤erences

contributed a relatively unchanging portion to the wage gap (0.08�0.17). A higher weight (or

set of coe¢ cients) for the productivity-related variables, as assigned primarily to the dom-

inant group, can serve to reconcile these facts. Nonetheless, comparison on a year-to-year

basis tends to bear out the result that human capital di¤erences contributed a proportional

20

part to wage di¤erences.

The representation of labor market segregation becomes �ner in proportion to one�s

increasing speci�city about occupational a¢ liation. In other words, using the two-digit

occupational and industrial classi�cation, as in Table 8, works out as more precise than ac-

counting only for the one-digit occupational a¢ liation, as in Table 6. It is expected, though,

this greater measure of precision will lead to higher estimates of labor market segregation.

This is readily seen from a comparison of the results from Tables 6 and 8. Labor market

segregation contributed 0.13�0.25 to the wage gap over the period in question (and 0.15�0.20

in the second half of the period, from 1997�2003).

If we manage to control for all the observables, and assume that any unobservable wage-

relevant variables vary only to a negligible degree between consecutive years, then we would

be bound to attribute any change in the unexplained wage gap between years to labor market

discrimination. These assumptions seem reasonable for all years after 1992, judging from

the results of the general speci�cation, Table 8. In the years 1993�1997, the unexplained

wage gap was 0.04�0.10. In the second half of the sample (1997�2003), this component

contributed about 0.10�0.19 to the gross wage gap. The sharp change in the unexplained

wage gap between the old and new sample styles around the year 1997 seems to follow from

the change in the sample design. If anything, the (higher) �gures in the second half of

the period are, statistically, more accurate, since those samples are more representative and

statistically twice the size.

An explanation of this trend of apparently rising discrimination must seek socioeconomic

21

factors in seeking to rationalize �uctuations in the unexplained portion of the wage gap.

The last decade began with a major in�ux to Israel of Jews from the former Soviet Union,

with about half million immigrants arriving just in the �rst three years of the decade. (This

represents an increase of some 11% in the country�s population.) These immigrants were,

at the beginning of the decade, very educated, with about 14�15 years of schooling on

average. (See Locher (2004) for a description of the trend of immigrants�education.) In

one sense, the incorporation of such a large number of people might be seen as a good

stimulus to economic recovery in the sense of strengthening demand. The revival of the

construction industry stands as a clear indicator of this. Educated immigrants, on the other

hand, compete with locals for jobs. Supposing that some Jewish migrants won jobs at the

expense of skilled Arabs, we would expect the gross wage gap among skilled workers to be

high around this period. Further, since the cross-ethnic skilled group is homogenous in an

important dimension, schooling, we would expect the human capital component of the wage

gap to be very low i.e. for speci�c non-productivity-related di¤erences to explain the wage

gap in large measure. Consequently, we would expect the unexplained component of the

wage gap closely to follow the pattern of the overall gross gap.

Table 9 documents the decomposition of the wage gap over the whole period, for skilled

and unskilled labor. Results from Table 9 largely con�rm this prediction of higher gross wage

gaps whose most part is unexplained (or attributable to discrimination). The gross wage gap

among skilled workers reached, in 1992�1993, its highest level. (Due to the sample design

change in 1997, we can add 0.203=.319-.116 to the gross �gures before 1997 to make them

22

comparable to post-1997 �gures.) Moreover, the human capital component contributed, for

most of the time, less than 0.04 to the overall wage gap. The unexplained gap among skilled

workers reached its highest level of 0.222 in 1992, around the time of the in�ux of skilled

Jewish labor. Though measures of the unexplained portion were volatile afterwards, in many

years this value recorded a decline.

At the same time as the migration to Israel of large numbers of mostly Russian Jews,

the Israeli government imposed extremely stringent restrictions on the movement of the

Arab inhabitants (and workers) of Israel�s administered territories. A surging demand for

construction workers, following the mass immigration, was met with a sudden shortage of

Palestinian workers, whom the closures prohibited from reaching their workplaces. Israel

responded to this labor crisis by �importing�foreign workers. About 30,000 foreign workers

were employed in 1993; this �gure was on the rise till 2002, when the number of employed

foreign workers reached an unprecedented apogee of 232,000.15 Substituting Palestinian

workers, and unskilled Israeli Arabs, with foreign workers had the e¤ect of increasing the

wage gap immensely. The unexplained component of the wage gap increased, between 1992

and 1993, by 0.143, a drastic change for a single year. The replacement of Arab by foreign

employees would seem su¢ cient by itself to explain this yearly change in the �unexplained�

segment of the gap.

When it comes, however, to explaining wage disparities among skilled labor, labor market

segregation played a more important role that reduced opportunities for Arabs. This might

15Estimated numbers of foreign workers employed in Israel are taken from the Bank of Israel web site.They may be accessed online athttp://www.bankisrael.gov.il/series/export/html/?series=NA.EM_FRN.A&start=1990&end=2004

23

be attributed to the fact that a wider variety of occupations is open to skilled workers, while

unskilled workers have access to low-skilled occupations only.

The expectation before analysis was that those industries into which foreign workers

were imported, agriculture and construction, would register the greatest change in the wage

gap (and in the unexplained portion of the gap) across Israel�s economy. Table 10 presents

the wage gap and its decomposition over the whole investigated period, �rst for workers in

agriculture and construction only, then for workers in all remaining industries. The large

measure of volatility in the unexplained component of the wage gap among agriculture and

construction workers, at least in the �rst half of the 1990s, supports the assumption that

these industries were especially a¤ected by changing employment patterns, as the sectors

recorded an unexplained wage gap increase of 0.23 between the years 1992 and 1993. These

sectoral results, combined with the previous skilled-unskilled comparison, serve to explain

the huge change in the overall unexplained wage gap between those years as captured in the

general speci�cation (Table 8).

It is evident from Table 10 that the composition of agriculture and construction work-

ers (or similarly unskilled workers from Table 9) is not constant, but �uctuates with labor

market and wider socioeconomic circumstances. Figures also suggest that labor market seg-

regation obtains less among relatively unskilled workers in the agriculture and construction

sectors than among workers in other industries. This is not surprising, since the occupa-

tional classi�cation is narrower in these two industries than in the economy as a whole.

Lastly, in many cases, the unexplained wage gap among unskilled workers (agriculture and

24

construction workers) is higher than that among skilled workers. Workers�heterogeneity in

the former group (with some Jews, alongside foreign and Arab workers) possibly accounts

for this observation.

Pooling all the years from 1991�2003 together, analysis further logged the gross (unad-

justed) wage gap between Jewish and Arab workers in every main occupational group and

for six schooling categories. Results appear in Table 11. The overall (logarithmic) wage

gap, for all occupations and schooling categories, is 0.392. The gross wage gap for every

known occupational category comes in at a lower level than the general gap. The highest

gap, including all schooling categories, was among skilled workers in the construction and

industry sectors (0.303). The wage gap for this occupational group increases with education

(from a negative gap of 0.041 for the uneducated to a wage gap of 0.422 for the highly

educated workers� or those with 16+ years of schooling). A similar pattern is observed

among unskilled and service sector workers: the greater the years of schooling, the higher

the gross wage gap (excluding the zero schooling group). More schooling does apparently

bene�t Arab managers, although the most highly educated cohort of managers is excluded,

where the wage gap climbs very fast, from 0.064 to 0.441 after getting a Masters degree.

As expected, controlling for all the (available) productivity-related variables reduces

the measured gross wage gap. Table 12 reports duly adjusted wage gaps within each

occupational-schooling category. In contrast with the previous table (11), here I control

for schooling level within schooling category (when that is not constant), as well as for ex-

perience, experience squared, marital status, full-time employment, size of city of residence,

25

one-digit occupational a¢ liation (when relevant), one-digit industrial a¢ liation, a yearly

time e¤ect, and industry-year interaction terms. The overall adjusted wage gap, over the

whole fourteen year period, comes out at 0.15. The trend of changes in the wage gap for

managers and skilled workers in industry is similar to that manifested in the unadjusted

gap. In general, the wage gap declines with education in high-education occupations, such

as academic professionals and managers, and increases with education in low-education oc-

cupations, such as industry, construction, and service sectors.

Table 13 further speci�es groups�employment distribution in some important regards.

It reports the sum of observations, over the whole period (1991�2003), for each occupation-

schooling cell. Overall, the occupational a¢ liation of about 51% of the Arab workers is

industry and construction (which are among the least remunerative jobs); this contrasts

with 29% of the Jewish workers. 2.5% of the Arab workers are managers (the second highest

paying job-category), as opposed to 11% of the Jewish workers. Moreover, among workers

with 12 years of schooling, 57% of the Arabs work in industry and construction, as opposed

to 40% of the Jewish workers. More importantly, high-paying jobs are not open to Arabs

even when they match Jews� level of schooling: 7% of the Jewish workers are associate

professionals, 8% managers, and 16% clerical workers (the highest paying jobs) as opposed

to, respectively, 2.6%, 2.6%, and 8% of the Arab workers.

For sure, self-selection goes some way towards explaining di¤erences in groups�employ-

ment distribution. However, with the table bringing together fourteen years of data, the

possibility of group representatives choosing to perpetuate outcomes that they have expe-

26

rienced as unequal becomes increasingly remote. Barriers to entry then represents a more

likely explanation of occupational distribution� meaning that much of the labor market seg-

regation, as calculated in previous tables, should be associated with this second explanation

of persistent wage gaps. It is worth noting that �barriers to entry�does not imply in the

exclusion of Arabs from certain occupations (though discriminatory practices might occur),

but rather that Arab workers need a much higher level of credentials than their Jewish

counterparts to be admitted through the door to comparable jobs.

Both the overall, and the unexplained, wage gap peaked at levels of 0.506 and 0.192,

respectively, in 1999 (see Table 8). Under all assumptions, and all model speci�cations, 1999

saw the high point in these measures. A range of pecuniary (human capital di¤erentials)

and non-pecuniary (discriminatory) factors can be understood as combining to drive the

wage gap to its peak. However, if labor market discrimination, re�ecting other political

hostilities, is taken as part of the explanation of the wage gap, then it seems that the

subsequent decrease in the wage gap itself, as much as in its unexplained portion, represents

the triumph of (rational and pecuniary) economics over discrimination. A consideration of

historical factors in part bears out this hypothesis. In 2000, with the eruption of the second

Intifada, the number of Palestinian workers available for work in Israel dramatically declined

from 113,000 in 1999 to 30,000 in 2002. Concurrently, the pace of foreign workers entering

the country decelerated (from 28,000 new foreign workers in 2001 to -25,000 i.e. a net out�ow

in 2003).16 These facts led to a higher demand for unskilled Arab Israelis. The resulting,

16Source: Bank of Israel web site. See:http://www.bankisrael.gov.il/series/export/html/?series=NA.EM_FRN.A&start=1990&end=2004and http://www.bankisrael.gov.il/series/export/html/?series=NA.EM_TER.A&start=1990&end=2004

27

and theoretically predictable, higher wage for Arab workers is consistent with the continuous

decline in the gross, and unexplained, wage gap in, and after, the year 2000.

Knowledge of Hebrew, which is an important component of human capital, might have

direct bearing on the issue of wage gaps. Nonetheless, the Income Survey data do not

provide information about knowledge of Hebrew, thus I could not control for this variable

in my analyses. If knowledge of Hebrew has a signi�cant e¤ect on wage gaps, i.e., lack of

knowledge of Hebrew leads to lower wage, then the unexplained component of the wage gap,

as presented in this paper, might be overestimated.

Chiswick (1998) �nds that Hebrew, being the worker�s primary or sole language, increases

the worker�s earnings by 11�35%. He uses data about foreign-born men from the 1983 Census

of Israel. Note that he uses a variable about �usage�of Hebrew and not �knowledge�of Hebrew;

so it is not a direct assessment of the e¤ect of knowledge of Hebrew on earnings. Secondly,

he �nds that Hebrew-speaking usage increases with duration of residence. Israeli Arabs,

living in Israel since birth, are expected, accordingly, to have a high knowledge and usage of

Hebrew. Moreover, Hebrew language is a compulsory subject in Arab schools, taught from

the second grade until the twelfth grade. Put together, it is suggested that Israeli-Arabs

know the Hebrew language very well. While the e¤ect that knowledge of Hebrew have on

earnings is acknowledged, the di¤erence in Hebrew knowledge between Arab and Jewish

workers is likely to be negligible. So is its e¤ect on this study�s results.

In a study by Lecker (1997) it has been shown that the �usage of Hebrew�explained only

a tiny part of the wage gap (about 0.02). The comparison was made between Arab groups

28

in Israel: Christians-Muslims, Muslims-Druze, and Druze-Christians. The study was based

on data from the 1983 Census of Israel, when the average years of schooling for Arabs was

about 8.5 years. Given that the average years of schooling is much higher now (in this paper

it is 9.5�11.5 for Arabs), I postulate that the e¤ect which knowledge of Hebrew may have on

this study�s results is even tinier. Finally, Table 12 shows that the highest unexplained wage

gaps were among the highly educated workers (13 years of schooling and more). This may

suggest that �knowledge of Hebrew�is not the main reason behind the observed wage gaps,

and thus its inevitable absence from the analysis in this study is not likely to fundamentally

bias the results.

Inference in this paper is based on the implicit assumption that both Arab and Jewish

workers are employed in the Jewish sector. It is impossible to identify the ethnicity of the

employer using available data. Economic theory suggests that wages in the Arab sector

are lower than those in the Jewish sector, due to the relative rarity of capital in the Arab

sector. Consequently, the unexplained component of the wage gap may be overestimated.

This hypothesis is not testable given the data we use in this study. Nonetheless, in a simple

tabulation from the geographic sample of the 1995 Israeli Census I �nd that only 9.8% of

Arab respondents (14% of Arab males) work in their residence, when this residence is de�ned

as �Arab and other.� This is far from estimating the portion of Arab workers employed by

Arab employers, but it may suggest that our implicit assumption has small e¤ect, if any, on

this study�s results.

29

5 Concluding Remarks

This study measures and documents the Jewish-Arab male wage di¤erential, and its evo-

lution, in the Israeli labor market in the years 1990�2003. Many reasons can be adduced

in explanation of wage di¤erentials in the labor market� a statement that hardly runs the

risk of contradiction in an economy as subject to as great a variety of pressures as Israel�s.

Israel has experienced rapid and large-scale changes in the composition of its workforce over

the 1990s. The large in�ux of foreign workers, the Oslo peace process, massive inward mi-

gration, and the breakout of the second Intifada, while events in the political sphere, have

all had their impact on the country�s economy, precipitating shifts in the wage gap and its

constituent parts as analyzed.

In such a dynamic economy, it would be hard, if not impossible, to identify the e¤ect of

any speci�c factor on a given economic activity. Identifying the mechanisms for inter-ethnic

wage di¤erences is no exception. My approach to the large and persistent wage gap between

Arab and Jewish workers� who share the same citizenship and are agents of the same labor

market� must then be hedged about with a recognition of this caution. This paper sought

to illuminate this gap in terms of productivity-related di¤erentials, on the one hand, and

labor market segregation and, possibly, discrimination, on the other.

While we can predict the e¤ect of workers�supply shock, due to the ingress into the labor

market of foreign workers, on local wages, tensions in Israeli-Palestinian relations might

yield further unexpected consequences, at least according to the assumptions of classical

economics. Foreign workers are perceived as substitutes to Arab workers (both citizens of

30

Israel and Palestinians from the occupied territories); therefore, the massive in�ux of such

agents into the economy will reduce the demand for, and thus the wages of, Arab workers.

The wage gap will widen as a result. Moreover, since this change is not determined by

productivity-related variables, the increase in the wage gap will mostly be attributed to the

�unexplained�component, part of which measures discrimination in the labor market. The

results of decomposition, robust to di¤erent assumptions and model speci�cations, support

this prediction. In the years of mass labor migration to Israel, 1990�1995, the hourly wage

gap increased from 41% in 1990 to 50% in 1995. The unadjusted (logarithmic) hourly wage

gap stood at 0.26 in the beginning of the decade; it increased monotonically until 1999

reaching its peak level of 0.51, and, declining since, scored 0.41 in 2003.

While human capital di¤erences and brute labor market segregation more than explained

the wage di¤erential at the beginning of the 1990, they were incapable of explaining a wage

gap of 0.19 (an hourly wage gap of about 29%) in 1999. The period ended in 2003 leaving a

wage gap of 0.13 unexplained by human capital di¤erences or labor market segregation.

It remains worthwhile to insist on the unpredictability of correlations between shifts in

the country�s political climate and the direction of its Jewish-Arab wage gap. Evidently,

recurrent outbursts of hostility on the Israeli-Palestinian dimension will adversely change

the way in which Israeli Arabs are viewed within the Israeli economy, despite the fact that

they are not practically party to the struggle of Arabs in the Occupied Territories. A likely

e¤ect of this is higher wage discrimination as measured in the labor market� in other words,

a pure change in tastes. Nonetheless, at some point, as Israeli-Palestinian tension peaks, the

31

Israeli government will undertake drastic action (closures and curfews) curbing the ability of

Palestinian Arabs to participate in the Israeli economy. The shortage in Palestinian labor is

then compensated for by higher demand, resulting in higher wages, for Israeli Arabs or, in

other cases, foreign workers. When foreign workers substitute for Arabs, as was the case in

the 1990s, a possible scenario is a decline in Arab wages, and the inception and entrenchment

of inter-ethnic wage gaps. The opposite-case scenario happened in the 2000s, when foreign

workers did not replace local job-holders, and indeed aggregately left Israel.

An example of such a crisis point in Israeli-Palestinian relations almost ideally matching

the de�nition might be provided by breakout of the second Intifada in September, 2000.

Study results strongly bear out the framework of explanatory assumptions outlined above,

in the sense that the year 2000 saw the onset of a continuous decline in the wage gap, notably

in its unexplained portion.

Correcting the discriminatory wage gap, which comprises part or all of the unexplained

wage gap, may call for a¢ rmative action, laws against discrimination, and other politically

mandated action, which labor market participants themselves would be unlikely to initiate

spontaneously. It would not be wise, though, for intervention policy to start with this

unobservable, hard-to-measure part of the wage gap, which, in many cases, may lie beyond

the reach of amendment. Nevertheless, the importation of foreign workers is sure to harm

the local workforce, whether Arab or Jewish, and, more importantly, is liable to widen the

wage gap and exacerbate ill feeling between local Israeli citizens. From the point of view of

fostering good community relations, then, other means, such as employing the unemployed,

32

should be used in the case of labor shortages.

Wage gaps attributable to human capital di¤erences or occupational segregation can be

more straightforwardly bridged. A policy aiming to reduce these measures has the potential

to be very e¤ective, in view of the fact that more than 62% (81% on average) of the overall

wage gap was explained by the human capital and occupational segregation components

jointly. More training, equal schooling, and better welfare provision to the subordinate

group have the scope to cover much, if not all, of the human capital component of the

wage gap. Guaranteeing equal employment opportunities in all occupations would help to

eradicate the occupational segregation component of the wage gap.

Although my analysis subjoins evidence of the Arab-Palestinian substitution e¤ect, such

that increasingly fraught Israeli-Palestinian relations result in a lower measure of Jewish-Arab

labor market discrimination, regressions cannot prove that phenomenon incontrovertibly.

Causality cannot be established using annual data, since many related changes can happen

within time periods that are impossible to control for. Future research would be well advised

to utilize more frequent data intervals in seeking to reduce potential patterns of correlation

among variables.

While male labor force participation rate was stable among Jewish workers over the whole

period (1990�2003), for the share of Arab workers declining over the years of the study.

Trends of change in male labor force participation rate should have no e¤ect on results, so

long as similar trends are observed for Arabs and Jews; this was the case for the years 1990�

1998. The convergence in Jewish-Arab male wages since 1999, however, was accompanied by

33

a declining Arab-to-Jewish male participation-rate. This may suggest that the convergence

is neither due to less discrimination, nor to Arab-Palestinian substitution e¤ects, but is

rather a consequence of a �winners-and-losers�type-e¤ect known as the selective-withdrawal

hypothesis: less-skilled Arabs withdraw from the labor force, due to low o¤er wages, resulting

in lower observed wage gap. Again, this e¤ect may be a contributor to falls in the gap without

accounting for it to the exclusion of other factors. This hypothesis is not tested in the paper.

Furthermore, the omission of relevant productivity variables, selection issues, and model

misspeci�cation could all have skewed the results as obtained from the Oaxaca decomposi-

tion. The new added component, namely occupational segregation, can be a¤ected as well.

Moreover, the existence of discrimination� meaning that productivity-related work charac-

teristics are di¤erentially less-well rewarded for Arabs� could have led Arabs to underinvest

in those characteristics, with a further compounding e¤ect; this would mean, in turn, that the

degree of discrimination would be understated in the adjusted representation. This suggests

that we should be skeptical in supposing our results to capture discrimination as divested of

human capital di¤erences as these pertain to, for instance, willingness to work.

That the Jewish-Arab wage gap reached an alarming pitch is beyond question. It is like-

wise impossible to maintain anything other than that the immense observed labor market

wage gap is an indication of the existence of some labor market friction or failure (whether

segregation, disintegration, or discrimination). The appropriate way to measure this dis-

crimination is, however, a matter of controversy. In this paper, I used the Oaxaca-Blinder

decomposition, with some modi�cations, to measure the human capital, occupational seg-

34

regation, and unexplained (in part discriminatory) component of the wage gap. However,

other methods exist for measuring these components. The question of whether wage gaps

obtain due to pure discrimination, or to the erection of barriers to entry to high-paying pro-

fessions, has yet to be answered for Israel. Future research on the subject could thus bene�t

from a di¤erent focus to this work.

6 Appendix

6.1 Results of the Main Wage Regressions

For the sake of completeness and brevity, results of the wage regressions are provided in

this appendix, in Tables 14 and 15. However, due to limitations of space and time, I choose

to report only a representative part of the wage regressions: those pertaining to the most

general speci�cation (on which results from Table 8 are based).

The dependent variable is the logarithm of hourly wage, with explanatory variables for:

years of schooling, potential years of experience, squared potential experience, a marital

status dummy, a full-time employment dummy, a large city dummy (1 if city of residence

is Jerusalem, Tel-Aviv, or Haifa, and 0 otherwise), two-digits occupational dummies, and

two-digits industrial dummies. Notes beneath the table provide more information.

35

6.2 Selectivity-Corrected Wage Gap Decomposition

Self-selection into employment has been proved potentially to bias the wage regression coef-

�cients (see Heckman (1979) for more details). A consistent estimation of wage regressions

is needed to get consistent wage gap decomposition. Therefore, a correction for employ-

ment self-selection may be called for when grounds exist for supposing this e¤ect to induce

a bias in the wage equations. Neuman and Oaxaca (2004a) introduced di¤erent methods

for incorporating this correction into the Oaxaca decomposition. The authors suggest di¤er-

ent selectivity-corrected decompositions; however, each of these yields di¤erent results, and

their method does not help in choosing the �correct�one, as the authors acknowledge in the

paper.17 (See applications of their methods in Neuman and Oaxaca (1998, 2004b).)

Following Reimers (1983) I calculated the selectivity e¤ect as a whole and the corrected

gross wage gap, then decomposed this to a human capital component, an unexplained (or

discriminatory) component, and an occupational and industrial segregation component. The

decomposition in equation 5 is generalized as follows:

ln (1 +G) =��XJ � �XA

�0�̂�| {z }

Q

+h�Z 0J

��̂J � �̂

��+ �Z 0A

��̂� � �̂A

�i| {z }

D

+�CJ � CA

�0 ̂�| {z }

S

+��̂J �̂J � �̂A�̂A

�| {z }

Selection

where Q; D; and S are the familiar components as de�ned in the methodology section (2).

The last term measures the selectivity e¤ect, where �̂ is the coe¢ cient of the Inverse Mills

Ratio (�̂) in the modi�ed wage equations. See Neuman and Oaxaca (2004a) for discussion

17�None of what has been presented here authoritatively identi�es the �correct� decomposition... Thechoice of which selectivity corrected decomposition to use is largely judgmental because it inevitably re�ectsvalue judgments about what constitutes labor market inequity.�Neuman and Oaxaca (2004a), p. 8.

36

of selectivity-corrected wage decomposition and the relevant terminology.

The selectivity corrected wage equations are estimated by the Heckman two step proce-

dure and, when possible, by maximum likelihood estimation. The explanatory variables in

the selection equation are: dummy variables for schooling, to capture the �sheepskin�e¤ect,

age and age squared, a large city dummy, and year of marriage, or years since �rst marriage

(with a zero value for unmarried). This variable, arguably, a¤ects participants�decision to

join the labor market as an earlier year of marriage (or more years since marriage) may

lower the reservation wage. Parents�and husbands�commitments and responsibilities call

for an urgent income, even if this is lower. That is, a greater number of years since �rst mar-

riage may a¤ect (increase) the probability of a participant joining the labor market, without

a¤ecting the worker�s wage.

In Table 16, I report selectivity-corrected decomposition results without further decom-

posing the selectivity e¤ect into human capital and discrimination components. I believe

this would clearly re�ect the self-selection e¤ect in measuring the wage gap; besides, it has

a more intuitive interpretation. A negative �selectivity e¤ect�suggests that Arabs are posi-

tively selected into the labor market. Thus, accounting for the selectivity e¤ect, if negative,

will yield higher unexplained wage gaps. Comparing the D �gures from Table 16 with those

from Table 8 lends support to this interpretation. The unexplained wage gaps are much

higher when we account for selectivity into employment.

The table shows a negative selectivity e¤ect which, in turn, implies that Arabs are pos-

itively self-selected into employment, or at least to a greater tune than their Jewish coun-

37

terparts. Therefore, for the same human capital and occupational distribution di¤erences, a

higher portion of the wage gap will be unexplained. This may suggest that the uncorrected

wage decomposition may underestimate the unexplained, or discriminatory, wage gap.

The human capital and occupational and industrial segregation e¤ects are very similar

to those derived from the selectivity-uncorrected wage decompositions. This shifts the whole

selection e¤ect to the unexplained component of the wage gap, and to the discriminatory

part of that component.

This is by no means a complete analysis of the e¤ect of self selection on wage gap decom-

position. Selectivity-corrected results rely heavily on the choice of selection functional form

and variables, and assumptions about the distribution of the error terms in the estimated

equations. All these considerations lie beyond the scope of this paper. All the same, this

discussion has served to indicate a possible direction of bias in decomposing wage gaps while

neglecting the selection e¤ect. However, the selection equation, the functional form, and the

method of incorporating those into the decomposition, all need further analysis and research.

Results as presented in this appendix are suggestive only.

References

Becker, Gary S. 1957. �The Economics of Discrimination,�Chicago: University of Chicago

Press,.

Blinder, Alan S. 1973. �Wage Discrimination: Reduced Form and Structural Estimates,�

38

Journal of Human Resources, 8(4), 436�455.

Brown, Randall S., Moon, Marilyn, and Zoloth, Barbara S. 1980. �Incorporating Occupa-

tional Attainment in Studies of Male-Female Earnings Di¤erentials,�Journal of Human

Resources, 15(1), 3�28.

Central Bureau of Statistics, Israel. 1990�2003. Income Survey (Households and Individuals)

1990�2003. Hebrew University of Jerusalem: Israel Social Sciences Data Center.

Central Bureau of Statistics, Israel. 2005. Statistical Abstract of Israel. Jerusalem: Cenral

Bureau of Statistics.

Chandra, Amitabh. 2000. �Labor-Market Dropouts and the Racial Wage Gap: 1940-1990,�

The American Economic Review, 90(2), 333�338.

Chiswick, Barry R. 1998. �Hebrew Language Usage: Determinants and E¤ects on Earnings

Among Immigrants in Israel,�Journal of Population Economics, 11(2), 253�271.

Heckman, James J. 1979. �Sample Selection Bias as a Speci�cation Error,�Econometrica,

47(1), 153�162.

Lecker, Tikva. 1997. �Language Usage and Earnings Among Minorities: The Case of the

Arabs in Israel,�Journal of Socio-Economics, 26(5), 525�532.

Locher, Lilo. 2004. �Immigration from the Former Soviet Union to Israel: Who is Coming

When?,�European Economic Review, 48, 1243�1255.

39

Miller, Paul W. 1987. �The Wage E¤ect of the Occupational Segregation of Women in

Britain,�The Economic Journal, 97(388), 885�896.

Neuman, Shoshana, and Oaxaca, Ronald L. 1998. �Estimating Labour Market Discrimina-

tion with Selectivity Corrected Wage Equations: Methodological Considerations and an

Illustration from Israel,�Center for Economic Policy Research, 1915(June), 1�23.

Neuman, Shoshana, and Oaxaca, Ronald L. 2004a. �Wage Decompositions with Selectivity-

Corrected Wage Equations: A Methodological Note,� Journal of Economic Inequality,

2(1), 3�10.

Neuman, Shoshana, and Oaxaca, Ronald L. 2004b. �Wage Di¤erentials in the 1990s in Israel:

Endowments, Discrimination, and Selectivity,�CEPR Discussion Paper, 4709(Oct.), 1�

26.

Neuman, Shoshana, and Silber, Jacques G. 1996. �Wage Discrimination Across Ethnic

Groups: Evidence from Israel,�Economic Inquiry, 34(4), 648�661.

Neumark, David. 1988. �Employers�Discriminatory Behavior and the Estimation of Wage

Discrimination,�The Journal of Human Resources, 23(3), 279�295.

Oaxaca, Ronald. 1973. �Male-Female Wage Di¤erentials in Urban Labor Markets,� Inter-

national Economic Review, 14(3), 693�709.

Oaxaca, Ronald L., and Ransom, Michael R. 1988. �Searching for the E¤ect of Unionism on

the Wages of Union and Nonunion Workers,�Journal of Labor Research, 9(2), 139�148.

40

Oaxaca, Ronald L., and Ransom, Michael R. 1994. �On Discrimination and the Decompo-

sition of Wage Di¤erentials,�Journal of Econometrics, 61(1), 5�21.

Oaxaca, Ronald L., and Ransom, Michael R. 1999. �Identi�cation in Detailed Wage Decom-

positions,�Review of Economics and Statistics, 81(1), 154�157.

Reimers, Cordelia W. 1983. �Labor Market Discrimination Against Hispanic and Black

Men,�The Review of Economics and Statistics, 65(4), 570�579.

41

Table 1: Sample Means by Ethnic Group, Full SampleIncome Schooling Age Married Full-time N

Year Arab Jew Arab Jew Arab Jew Arab Jew Arab Jew Arab Jew1990 26.2 36.6 n/a n/a 36.03 41.57 .80 .90 .91 .93 169 27511991 26.0 35.1 9.58 12.69 35.45 41.12 .80 .88 .92 .93 207 29421992 26.7 37.1 9.72 12.70 37.74 41.60 .88 .89 .90 .94 182 28161993 24.7 34.6 10.02 12.89 35.65 41.19 .87 .88 .94 .94 172 25041994 26.8 36.0 10.62 13.09 36.25 41.12 .86 .87 .91 .93 202 27661995 29.0 40.4 10.42 13.17 36.80 41.09 .86 .86 .90 .94 465 29231996 26.3 38.2 10.47 13.23 36.82 41.11 .82 .85 .94 .92 519 29921997a 27.4 39.5 10.67 13.33 37.13 41.16 .85 .86 .94 .93 492 27861997b 27.5 41.4 10.65 13.42 36.76 41.08 .86 .86 .89 .92 900 52931998 28.1 42.5 10.99 13.64 36.92 41.14 .87 .86 .90 .92 954 53621999 26.5 43.4 11.11 13.73 36.74 41.20 .85 .86 .92 .93 951 53182000 28.2 44.6 11.42 13.59 36.56 41.02 .84 .84 .93 .93 892 53612001 31.5 47.3 11.80 13.77 36.74 41.04 .85 .85 .91 .92 946 53702002 30.9 44.9 11.42 13.81 36.47 41.09 .84 .83 .89 .91 892 55972003 30.4 43.6 11.46 13.83 36.55 40.94 .83 .82 .88 .91 887 5679NOTE.� The sample includes salaried, prime-aged (25�65), male workers. �Full-sample� includes all immi-grants, and �reduced-sample�excludes recently arriving immigrants. The variable �Income�is the real hourly-wage, in NIS of 2000 (CPI de�ated). The variable �Married�takes on the value 0 if the person is single, and1 otherwise. Year 1997a (1997b) refers to the Old (New) version of the Income Survey of 1997. The incomesurvey of 1990 includes categorial, rather than continuous, schooling variable. See text for details.

42

Table 2: Sample Means by Ethnic Group, Reduced SampleIncome Schooling Age Married Full-time N

Year Arab Jew Arab Jew Arab Jew Arab Jew Arab Jew Arab Jew1990 26.2 36.9 n/a n/a 36.03 41.64 .80 .90 .91 .93 169 26481991 26.1 36.5 9.49 12.55 35.51 41.31 .80 .88 .92 .93 202 26431992 27.3 39.5 9.55 12.59 37.79 41.72 .87 .89 .89 .94 174 24241993 25.1 37.5 9.73 12.69 35.74 41.32 .87 .87 .95 .93 159 20571994 27.6 38.8 10.42 12.97 36.28 41.02 .85 .86 .90 .94 185 22631995 29.1 43.6 10.12 13.03 36.74 41.14 .86 .85 .90 .93 437 23501996 27.2 41.7 10.26 13.14 36.81 41.05 .83 .84 .93 .92 478 23531997a 27.9 43.1 10.52 13.23 37.01 41.35 .84 .85 .93 .93 443 21601997b 27.9 45.2 10.65 13.37 36.76 41.22 .86 .86 .89 .92 900 41501998 29.1 45.8 10.84 13.51 36.72 41.11 .86 .85 .89 .92 884 42221999 26.8 47.3 10.95 13.67 36.62 41.17 .86 .85 .92 .92 880 40972000 28.8 48.2 11.23 13.56 36.41 40.97 .83 .84 .92 .92 778 42952001 32.0 51.7 11.49 13.74 36.44 41.05 .83 .84 .90 .92 820 42132002 30.9 48.7 11.42 13.82 36.47 41.10 .84 .82 .89 .91 892 43382003 30.4 47.3 11.45 13.80 36.55 41.00 .83 .81 .88 .90 884 4376NOTE.� See notes for Table 1.

43

Table 3: Jewish-Arab Wage Gap from Log Wage RegressionsFull Sample Reduced Sample

Year (1) (2) (3) (1) (2) (3)1990 :2588

(:0027):0299(:0025)

:0192(:0024)

:2690(:0028)

:0349(:0024)

:0250(:0024)

1991 :2185(:0028)

:0057(:0026)

�:0048(:0024)

:2528(:0028)

:0241(:0025)

:0158(:0024)

1992 :2634(:0029)

:0845(:0027)

:0660(:0026)

:3122(:0029)

:1039(:0026)

:0887(:0026)

1993 :2738(:0029)

:1173(:0028)

:0938(:0026)

:3473(:0030)

:1503(:0027)

:1454(:0026)

1994 :2749(:0028)

:1089(:0026)

:0699(:0024)

:3378(:0029)

:1340(:0025)

:1074(:0024)

1995 :2734(:0020)

:0902(:0019)

:0634(:0017)

:3582(:0021)

:1367(:0019)

:1245(:0018)

1996 :3014(:0019)

:1228(:0018)

:0865(:0017)

:3683(:0019)

:1468(:0018)

:1261(:0017)

1997a :3177(:0019)

:1306(:0018)

:1028(:0016)

:3901(:0020)

:1526(:0018)

:1402(:0017)

1997b :3648(:0018)

:1703(:0017)

:1451(:0016)

:4520(:0018)

:2340(:0017)

:2137(:0016)

1998 :3761(:0018)

:2062(:0017)

:1758(:0016)

:4335(:0018)

:2288(:0017)

:2100(:0016)

1999 :4249(:0017)

:2282(:0016)

:1893(:0015)

:5065(:0018)

:2744(:0016)

:2463(:0016)

2000 :4036(:0018)

:2243(:0017)

:1719(:0015)

:4691(:0019)

:2498(:0017)

:2079(:0017)

2001 :3648(:0018)

:1898(:0017)

:1746(:0019)

:4520(:0020)

:2241(:0017)

:1965(:0022)

2002 :3387(:0018)

:1408(:0017)

:1196(:0015)

:4216(:0018)

:2059(:0016)

:1796(:0016)

2003 :3246(:0017)

:1497(:0016)

:1321(:0015)

:4056(:0017)

:2111(:0016)

:1914(:0016)

NOTE.� Samples include salaried, prime-aged (25-65), male workers. �Full-sample�includesall immigrants, and �reduced-sample�excludes recently arriving immigrants.Reported gap is measured by the coe¢ cient of ethnic dummy variable (that takes on thevalue 1 if the worker is Jewish and 0 otherwise) in a pooled wage regression, that includesboth Arabs and Jews.The dependent variable in all speci�cations is the log of hourly wage. Speci�cation 1 includesonly Jewish dummy variable. Speci�cation 2 includes Jewish dummy variable, schooling,experience, experience squared, marital status, and full-time employment. Speci�cation 3includes all the variables in speci�cation 2 and a set of (one-digit) occupational dummies.Year 1997a (1997b) refers to the Old (New) version of the income survey; see text for details.Standard errors are in parentheses.

44

Table 4: Wage Gap Decomposition. �̂�= �̂J . Full Sample

(1) (2)Year G Q D G Q D S1990 .2596 .2383 .0213 .2596 .1958 .0126 .05121991 .2190 .2138 .0052 .2190 .1190 -.0075 .10751992 .2634 .1832 .0802 .2634 .0954 .0593 .10871993 .2739 .1549 .1190 .2739 .0552 .0965 .12221994 .2747 .1622 .1125 .2747 .0843 .0723 .11811995 .2740 .1801 .0939 .2740 .0730 .0698 .13121996 .3023 .1757 .1266 .3023 .0819 .0903 .13011997a .3184 .1807 .1377 .3184 .0843 .1096 .12451997b .3654 .1988 .1666 .3654 .1040 .1408 .12061998 .3754 .1761 .1993 .3754 .0838 .1701 .12151999 .4242 .1985 .2257 .4242 .1068 .1868 .13062000 .4047 .1819 .2228 .4047 .1019 .1693 .13352001 .3647 .1748 .1899 .3647 .0879 .1712 .10562002 .3385 .1986 .1399 .3385 .0931 .1160 .12942003 .3248 .1780 .1468 .3248 .0833 .1302 .1113NOTE.� Speci�cation 1 is based on wage equations with the regressors: schooling,experience, experience squared, marital status, working full-time. Speci�cation 2 isbased on wage equations with regressors as in 1 and a set of one-digit occupationaldummies. G refers to the gross wage gap (or ln(1+G)), Q refers to the human capitalcomponent of the wage gap, D refers to the unexplained (or discriminatory) componentof the wage gap, and S refers to the occupational segregation component of the wagegap.

45

Table 5: Wage Gap Decomposition. �̂�= �̂pooled. Full Sample

(1) (2)Year G Q D G Q D S1990 .2596 .2321 .0275 .2596 .1874 .0175 .05471991 .2190 .2137 .0053 .2190 .1167 -.0045 .10681992 .2634 .1832 .0802 .2634 .0951 .0618 .10651993 .2739 .1630 .1109 .2739 .0628 .0880 .12311994 .2747 .1707 .1040 .2747 .0872 .0662 .12131995 .2740 .1910 .0830 .2740 .0848 .0571 .13211996 .3023 .1906 .1117 .3023 .0913 .0766 .13441997a .3184 .1991 .1193 .3184 .0967 .0915 .13021997b .3654 .2119 .1535 .3654 .1152 .1285 .12171998 .3754 .1882 .1872 .3754 .0977 .1569 .12081999 .4242 .2169 .2073 .4242 .1202 .1681 .13592000 .4047 .1969 .2078 .4047 .1103 .1568 .13762001 .3647 .1873 .1774 .3647 .0935 .1601 .11112002 .3385 .2096 .1289 .3385 .1023 .1074 .12882003 .3248 .1876 .1372 .3248 .0963 .1180 .1105NOTE.� See notes for Table 4.

Table 6: Wage Gap Decomposition. �̂�= �̂J . Reduced Sample

(1) (2)Year G Q D G Q D S1990 .2690 .2443 .0248 .2683 .2037 .0179 .04671991 .2539 .2286 .0253 .2538 .1441 .0140 .09571992 .3125 .2165 .0959 .3099 .1382 .0810 .09071993 .3475 .1968 .1507 .3493 .1030 .1477 .09861994 .3376 .2028 .1349 .3374 .1347 .1077 .09501995 .3590 .2228 .1362 .3585 .1170 .1303 .11121996 .3685 .2170 .1515 .3674 .1244 .1324 .11061997a .3907 .2337 .1570 .3882 .1426 .1490 .09661997b .4524 .2251 .2273 .4524 .1383 .2089 .10521998 .4326 .2104 .2222 .4326 .1346 .2033 .09471999 .5060 .2336 .2724 .5060 .1504 .2456 .11002000 .4705 .2227 .2478 .4705 .1486 .2061 .11582001 .4519 .2248 .2271 .4519 .1761 .1971 .07872002 .4215 .2185 .2030 .4215 .1268 .1757 .11902003 .4057 .2011 .2046 .4057 .1195 .1887 .0975NOTE.� See notes for Table 4.

46

Table 7: Wage Gap Decomposition. �̂�= �̂pooled. Reduced Sample

(1) (2)Year G Q D G Q D S1990 .2698 .2379 .0319 .2698 .1958 .0227 .05131991 .2539 .2314 .0225 .2539 .1440 .0145 .09541992 .3125 .2146 .0979 .3125 .1417 .0821 .08871993 .3475 .2071 .1404 .3475 .1125 .1346 .10041994 .3376 .2108 .1268 .3376 .1378 .1004 .09941995 .3590 .2366 .1224 .3590 .1347 .1079 .11641996 .3685 .2382 .1303 .3685 .1446 .1080 .11591997a .3907 .2547 .1360 .3907 .1611 .1198 .10981997b .4524 .2464 .2060 .4524 .1537 .1822 .11651998 .4326 .2276 .2050 .4326 .1517 .1834 .09751999 .5060 .2631 .2429 .5060 .1664 .2107 .12892000 .4705 .2428 .2277 .4705 .1595 .1851 .12592001 .4519 .2477 .2042 .4519 .1875 .1741 .09032002 .4215 .2368 .1847 .4215 .1403 .1554 .12582003 .4057 .2166 .1891 .4057 .1371 .1641 .1045NOTE.� See notes for Table 4.

47

Table 8: Wage Gap Decomposition, General Speci�cationYear G Q D OS IS S1990 :2684

(:0019):1678(:0015)

�:0330(:0022)

:0913(:0016)

:0424(:0015)

:1337(:0017)

1991 :2536(:0019)

:1244(:0011)

�:0510(:0022)

:1026(:0017)

:0776(:0015)

:1802(:0018)

1992 :3090(:0024)

:1057(:0011)

�:0447(:0029)

:1654(:0020)

:0826(:0017)

:2480(:0021)

1993 :3496(:0023)

:0825(:0011)

:0778(:0026)

:1205(:0019)

:0688(:0013)

:1893(:0020)

1994 :3368(:0027)

:1131(:0011)

:0409(:0023)

:1133(:0016)

:0695(:0011)

:1828(:0018)

1995 :3579(:0018)

:0974(:0010)

:0756(:0018)

:1291(:0014)

:0556(:0009)

:1847(:0014)

1996 :3668(:0016)

:1168(:0009)

:0554(:0018)

:1250(:0015)

:0695(:0012)

:1945(:0015)

1997a :3873(:0017)

:1243(:0010)

:0959(:0017)

:1245(:0014)

:0426(:0011)

:1671(:0014)

1997b :4525(:0016)

:1293(:0009)

:1648(:0018)

:1151(:0012)

:0433(:0009)

:1584(:0013)

1998 :4323(:0017)

:1243(:0009)

:1559(:0019)

:1092(:0012)

:0429(:0009)

:1521(:0013)

1999 :5062(:0015)

:1399(:0009)

:1918(:0016)

:1147(:0010)

:0596(:0009)

:1744(:0011)

2000 :4707(:0016)

:1373(:0009)

:1641(:0017)

:1151(:0010)

:0542(:0008)

:1693(:0011)

2001 :4175(:0024)

:1284(:0013)

:1231(:0023)

:1000(:0016)

:0660(:0013)

:1660(:0017)

2002 :4202(:0016)

:1185(:0009)

:0991(:0017)

:1326(:0012)

:0700(:0010)

:2026(:0013)

2003 :4017(:0016)

:0979(:0009)

:1265(:0017)

:1077(:0012)

:0696(:0009)

:1773(:0012)

NOTE.� Reduced Samples (see text for details). �̂�= �̂J .

G is the gross (logarithmic) wage gap. Q is the human capital component of thewage gap. D is the unexplained component. OS is the occupational segregationcomponent. IS is the industrial segregation component. And S is the total labormarket segregation, i.e., S = OS + IS. Standard errors are in parentheses.Explanatory variables in the underlying wage regressions are: schooling, poten-tial experience, squared potential experience, marital status dummy, full-timeemployment dummy, large city dummy, two-digits occupational dummies, andtwo-digits industrial dummies. See Tables 14 and 15 for the individual wageregressions.

48

Table 9: Wage Gap Decomposition, by Skill GroupSkilled Unskilled

Year G Q D S G Q D S1990 :163 :033 �:025 :156 :200 :123 �:019 :0961991 :079 :026 �:011 :064 :159 :109 �:067 :1171992 :347 :023 :222 :101 :191 :082 �:062 :1711993 :174 :089 :021 :063 :250 :066 :080 :1041994 :095 :091 �:531 :535 :295 :132 :022 :1411995 :041 �:031 :025 :047 :288 :104 :070 :1141996 :085 :036 �:072 :122 :274 :093 :066 :1151997a :116 :029 �:037 :124 :318 :102 :130 :0861997b :319 :027 :193 :099 :339 :106 :143 :0901998 :287 :049 :157 :081 :313 :095 :152 :0661999 :361 :047 :172 :141 :364 :097 :193 :0742000 :261 :034 :073 :154 :365 :102 :171 :0912001 :175 :013 �:013 :175 :363 :115 :171 :0772002 :296 :047 :085 :164 :313 :084 :126 :1032003 :274 �:008 :157 :125 :297 :098 :089 :110

NOTE.� Data are from the Reduced Samples. �̂�= �̂J . Skilled worker is

de�ned as one with more than 12 years of schooling.G is the gross wage gap, Q is the human capital component of the wage gap, Dis the unexplained wage gap, S is the labor market, occupational and industrial,segregation component of the wage gap.All �gures are signi�cant at the 5% signi�cance level.The explanatory variables in the underlying wage regressions are: schooling,potential experience, squared potential experience, marital status dummy, full-time employment dummy, large-city dummy, two-digits occupational dummies,and two-digits industrial dummies.

49

Table 10: Wage Gap Decomposition, by Industrial A¢ liationAgriculture & Construction Other Industries

Year G Q D S G Q D S1990 :160 :140 �:036 :056 :239 :166 �:035 :1071991 :003� �:100 :108 �:005� :289 :121 :007 :1611992 :105 :024 �:161 :243 :327 :098 :028 :2011993 :317 :111 :072 :133 :304 :076 :053 :1751994 :227 :255 �:116 :088 :317 :104 :033 :1801995 :373 :079 :154 :140 :308 :097 :042 :1691996 :244 :027 :123 :094 :313 :111 :037 :1651997a :353 :185 :137 :031 :316 :100 :045 :1711997b :353 :169 :136 :047 :429 :120 :155 :1541998 :340 :189 :156 �:005� :408 :111 :148 :1481999 :364 :199 :091 :074 :499 :132 :189 :1782000 :469 :152 :339 �:022 :443 :128 :147 :1682001 :465 :145 :251 :069 :358 :090 :103 :1652002 :364 :144 :144 :076 :369 :100 :081 :1872003 :312 :174 :085 :053 :368 :076 :117 :175

NOTE.� Data are from the Reduced Samples. �̂�= �̂J . First panel pertains to workers

in the agriculture and construction industries only. Second panel pertains to workers inall industries, excluding agriculture and construction.The explanatory variables in the underlying wage regressions are: schooling, potentialexperience, squared potential experience, marital status dummy, full-time employmentdummy, large-city dummy, two-digits occupational dummies, and two-digits industrialdummies.G is the gross wage gap, Q is the human capital component of the wage gap, D is the un-explained wage gap, and S is the labor market, occupational and industrial, segregationcomponent of the wage gap.* Figure is not signi�cant at the 5% signi�cance level. All other (unstarred) �gures aresigni�cant at the 5% signi�cance level.

50

Table 11: Gross Wage Gap, by Occupation and SchoolingYears of Schooling

Occupation 0 1�6 7�11 12 13�15 16+ AllAcademic Professionals n=a n=a :978 :346 :276 :133 :170Associate Professionals & Technicians :341 �:276 :321 :277 :127 :027 :098Managers :417 :379 :251 :126 :064 :441 :271Clerical Workers :064 :053 :329 :110 :191 :340 :197Agents, Sales & Service Workers :249 :075 :136 :173 :238 :420 :247Skilled Agricultural Workers :054 :192 :094 :233 �:046 n=a :193Industry, Construction, & Other Skilled �:041 :033 :241 :301 :353 :422 :303Unskilled Workers :168 :073 :197 :204 :351 :307 :215Unknown :254 :560 :353 :281 :155 :223 :448

All :118 :074 :251 :297 :241 :202 :392NOTE.� Estimation is based on the pooled data of reduced samples for the years 1991�2003. Notethat 1990 is not included, since the schooling categories given in the data do not match those in thistable. Also, note that these data include both versions of the 1997 Income Survey. (See text for details.)Reported, in each cell, is the Jewish-Arab di¤erence in the average (logarithmic) real hourly-wage.(Approximately, this is the same as the unadjusted, geometric wage gap.)n/a: the �gure cannot be calculated due to lack of (mainly Arab) observations. See Table 13 for thenumber of observations in each cell, decomposed by ethnic group.

51

Table 12: Adjusted Wage Gap, by Occupation and SchoolingYears of Schooling

Occupation 0 1�6 7�11 12 13�15 16+ AllAcademic Professionals n=a n=a n=a :501

(:011):185(:005)

:072(:002)

:0959(:0018)

Associate Professionals & Technicians n=a n=a :214(:009)

:267(:005)

:139(:003)

�:110(:003)

:0430(:0020)

Managers n=a n=a :175(:008)

:108(:006)

:076(:006)

:353(:005)

:1843(:0030)

Clerical Workers n=a n=a :190(:004)

:046(:003)

:094(:004)

:327(:005)

:1289(:0019)

Agents, Sales & Service Workers :232(:005)

:119(:008)

:082(:002)

:152(:002)

:324(:004)

:473(:008)

:1695(:0014)

Skilled Agricultural Workers �:104(:013)

:207(:007)

�:031(:004)

:209(:005)

:320(:018)

n=a :1359(:0029)

Industry, Construction, & Other Skilled :043(:007)

�:007(:003)

:104(:001)

:202(:001)

:237(:003)

:296(:006)

:1485(:0007)

Unskilled Workers :240(:006)

:108(:005)

:095(:002)

:178(:003)

:234(:008)

�:061(:015)

:1359(:0016)

Unknown n=a :743(:009)

:105(:005)

:192(:004)

:158(:006)

:152(:006)

:2171(:0025)

All :1839(:0046)

:0668(:0023)

:1086(:0008)

:1715(:0009)

:1890(:0015)

:1250(:0014)

:150(:0005)

NOTE.� Estimation is based on the pooled data of reduced samples for the years 1991�2003. Note that1990 is not included, since the schooling categories given in the data do not match those in this table. Also,note that these data include both versions of the 1997 Income Survey. (See text for details.)Main entry is the coe¢ cient of the dummy variable �Jew� in a logarithmic, real hourly-wage regression.Explanatory variables are: years of schooling (where applicable), potential experience, squared potentialexperience, marital status dummy, full-time employment dummy, large-city dummy, one-digit occupa-tional dummies (where applicable), one-digit industrial dummies, time-e¤ect (year) dummies, and a set ofindustry-year interaction dummies. Standard errors are in parentheses.n/a: the �gure cannot be calculated due to lack of (mainly Arab) observations. See Table 13 for thenumber of observations in each cell, decomposed by ethnic group.

52

Table 13: Number of Observations, by Ethnic Group, Occupation, and SchoolingYears of Schooling

Occupation 0 1�6 7�11 12 13�15 16+ AllJewsAcademic Professionals 0 1 15 75 716 5018 5825Associate Professionals & Technicians 4 9 242 937 2597 1616 5405Managers 4 6 354 1085 1471 1997 4917Clerical Workers 4 33 1075 2204 1249 840 5405Agents, Sales & Service Workers 27 63 1325 2299 1348 649 5711Skilled Agricultural Workers 14 21 158 147 57 15 412Industry, Construction, & Other Skilled 75 219 5325 5406 1856 566 13447Unskilled Workers 51 123 1075 689 218 85 2241Unknown 5 21 347 686 715 765 2539

Total 184 496 9916 13528 10227 11551 45902

ArabsAcademic Professionals 0 0 5 8 82 492 587Associate Professionals & Technicians 1 2 24 52 194 203 476Managers 1 2 30 52 50 71 206Clerical Workers 1 8 68 163 99 57 396Agents, Sales & Service Workers 7 48 343 311 82 34 825Skilled Agricultural Workers 10 51 139 33 3 0 236Industry, Construction, & Other Skilled 32 323 2442 1153 151 47 4148Unskilled Workers 38 126 491 138 31 15 839Unknown 1 17 159 101 52 56 386

Total 91 577 3701 2011 744 975 100 8099NOTE.� Pooled data of the reduced samples for the years 1991�2003. These are observations withnon-missing income data; they are used in Tables 11 and 12. Note that these are the unweightednumbers of observations. In the analyses, throughout the paper, I use the weighted samples. TheCBS provides weights in the data to re�ect the real share of each group in the population.

53

Table 14: Wage Regressions, by Ethnic Group1990 1991 1992 1993 1994 1995 1996 1997a

JewsSchooling �� :0309

(:0002):0337(:0002)

:0236(:0002)

:0368(:0002)

:0293(:0002)

:0328(:0002)

:0379(:0003)

Experience :0659�(:0005)

:0343(:0002)

:0281(:0003)

:0391(:0002)

:0344(:0002)

:0397(:0002)

:0334(:0003)

:0318(:0002)

Exp2 �:0006�(:00001)

�:0005(:0000)

�:0003(:00000)

�:0005(:00000)

�:0004(:00000)

�:0005(:00000)

�:0004(:00000)

�:0004(:00000)

Marital :1391(:0023)

:1288(:0021)

:1297(:0023)

:1044(:0022)

:0767(:0021)

:0815(:0021)

:0494(:0021)

:1221(:0020)

Full time �:0797(:0026)

�:1223(:0027)

�:0668(:0030)

�:1119(:0028)

�:0134(:0028)

�:0208(:0028)

�:0706(:0025)

�:0543(:0027)

Large City �:0157(:0014)

�:0217(:0014)

�:0271(:0016)

:0099(:0015)

:0087(:0015)

:0056(:0015)

:0481(:0015)

:0187(:0015)

Occup/Ind. Yes Yes Yes Yes Yes Yes Yes YesIntercept :0419

(:0366)1:5836(:0299)

:6212(:0546)

2:6911(:0295)

1:5445(:0280)

2:5678(:0174)

2:2521(:0212)

2:8135(:0211)

R2 .4199 .4339 .4213 .4543 .4350 .4493 .4410 .4856

ArabsSchooling �� :0341

(:0006)�:0037(:0009)

:0072(:0008)

:0465(:0009)

:0312(:0005)

:0356(:0005)

:0232(:0004)

Experience �:0057�(:0010)

:0148(:0006)

:0356(:0009)

:0269(:0009)

:0441(:0008)

:0061(:0005)

:0217(:0005)

:0155(:0004)

Exp2 :0001�(:00001)

�:0001(:0000)

�:0004(:00001)

�:0004(:00001)

�:0007(:00001)

�:00001(:00001)

�:0002(:00001)

�:0002(:00001)

Marital :1993(:0027)

:0788(:0037)

�:0704(:0068)

�:0278(:0062)

:1651(:0053)

:1106(:0036)

:0070(:0030)

:0838(:0028)

Full time �:2381(:0048)

�:1584(:0119)

:0516(:0070)

�:1667(:0095)

�:1709(:0069)

�:1380(:0045)

�:1131(:0064)

�:1983(:0058)

Large City �:1478(:0035)

�:1416(:0039)

:1400(:0071)

:3555(:0096)

�:1293(:0084)

:0080(:0054)

:0514(:0043)

�:0259(:0030)

Occup/Ind. Yes Yes Yes Yes Yes Yes Yes YesIntercept 3:1458

(:0274)2:4205(:0417)

1:0158(:0248)

4:3781(:0599)

1:3049(:0476)

2:6434(:0328)

2:0947(:0309)

2:9169(:0291)

R2 .8141 .7217 .6237 .7115 .7996 .6303 .6246 .7272NOTE.� The dependent variable is the logarithm of hourly wage. The explanatory variables are:schooling, potential experience, squared potential experience, marital status dummy, full time em-ployment dummy, large city dummy, two-digits occupational dummies, and two-digits industrialdummies (shortly, Occup/Ind.). The wage regressions for Jewish workers included about 150�160explanatory variables. Those for Arab workers included 80�130 explanatory variables.* Income data for the year 1990 provides only categorial schooling variable. Thus, I used dummyvariables for schooling. Also, for this year, I used age and age squared instead of potential experienceand squared potential experience.

54

Table 15: Wage Regressions, Continued1997b 1998 1999 2000 2001 2002 2003

JewsSchooling :0424

(:0002):0428(:0002)

:0459(:0002)

:0491(:0002)

:0548(:0003)

:0410(:0002)

:0367(:0002)

Experience :0320(:0002)

:0383(:0002)

:0379(:0002)

:0317(:0002)

:0208(:0003)

:0293(:0002)

:0316(:0002)

Exp2 �:0004(:00000)

�:0005(:00000)

�:0005(:00000)

�:0004(:00000)

�:0002(:00001)

�:0003(:00000)

�:0004(:00000)

Marital :1269(:0021)

:1084(:0021)

:0762(:0020)

:0960(:0019)

:1293(:0025)

:1122(:0019)

:0829(:0018)

Full time :0261(:0025)

�:0151(:0024)

�:0362(:0024)

�:0321(:0023)

�:1628(:0031)

�:1242(:0022)

�:1410(:0021)

Large City �:0092(:0016)

:0341(:0016)

:0162(:0016)

:0170(:0015)

:0156(:0020)

�:0137(:0015)

:0042(:0016)

Occup/Ind. Yes Yes Yes Yes Yes Yes YesIntercept 2:5654

(:0093)2:5276(:0090)

3:1638(:0202)

2:2702(:0339)

2:7715(:0214)

2:9601(:0135)

2:8454(:0147)

R2 .4074 .4016 .4150 .4139 .4934 .4589 .4160

ArabsSchooling :0270

(:0004):0448(:0004)

:0353(:0004)

:0250(:0005)

:0332(:0007)

:0283(:0005)

:0247(:0004)

Experience :0167(:0004)

:0337(:0004)

:0245(:0004)

:0112(:0005)

:0288(:0007)

:0167(:0005)

:0153(:0005)

Exp2 �:0002(:00001)

�:0004(:00001)

�:0003(:00001)

�:00009(:00001)

�:0004(:00001)

�:0001(:00001)

�:0001(:00001)

Marital :0611(:0033)

:0353(:0037)

�:0356(:0031)

:0123(:0035)

�:0238(:0044)

:1516(:0036)

:1376(:0033)

Full time �:2305(:0043)

�:3390(:0043)

�:2583(:0041)

�:1699(:0054)

�:1765(:0062)

�:3409(:0041)

�:3625(:0038)

Large City �:3203(:0029)

�:2381(:0030)

�:1149(:0024)

:1055(:0058)

:0724(:0064)

�:1278(:0029)

�:0388(:0030)

Occup/Ind. Yes Yes Yes Yes Yes Yes YesIntercept 3:2491

(:0141)2:4980(:0157)

3:8367(:0397)

3:8219(:0302)

2:8340(:0448)

2:9015(:04611)

2:8629(:0533)

R2 .5696 .5453 .5484 .5720 .6823 .5452 .5363NOTE.� See notes for Table 14.

55

Table 16: Selectivity-Corrected Wage Gap Decomposition

Year G Q D S Selection eG1990 :2684 :1742

:1536�:2557:0322

:1702:1734

:1797�:0908

:0887:3592

1991 :2536 :1231:1236

�:0552:0336

:1933:1937

�:0077�:0973

:2613:3509

1992 :3090 :1038 :1349 :2445 �:1743 :48331993 :3496 :0812 :2448 :1864 �:1628 :51241994 :3368 :1165

:1167�:2047:0044

:1953:1933

:2297:0223

:1071:3145

1995 :3579 :1014:0986

:1309:0643

:1783:1799

�:0527:0150

:4106:3429

1996 :3668 :1233 :3379 :1932 �:2876 :65441997a :3873 :1267

:1276:2562:2295

:1628:1612

�:1585�:1311

:5458:5184

1997b :4525 :1278:1277

:5757:5534

:1531:1540

�:4041�:3826

:8566:8351

1998 :4323 :1250:1244

:4090:3626

:1444:1446

�:2461�:1993

:6784:6316

1999 :5062 :1417 :2932 :1643 �:0929 :59912000 :4707 :1414

:1393:3145:2072

:1625:1646

�:1477�:0405

:6184:5112

2001 :4175 :1330:1318

:2844:2336

:1574:1606

�:1574�:1085

:5749:5260

2002 :4202 :1160:1159

:0122:2060

:1979:1972

:0940�:0990

:3262:5192

2003 :4017 :1024:0996

:4068:5372

:1699:1746

�:2774�:4096

:6791:8113

NOTE.� Data are from the Reduced Samples. �̂�= �̂J .

G is the gross wage gap, Q is the human capital component of the wage gap,D is the unexplained component, S is the labor market, occupational andindustrial, segregation component, Selection is the component of the wagegap attributed to self selection into employment, and ~G is the selectivity-corrected gross wage gap. That is, the gross wage gap that would haveprevailed had there been no self selection into employment.Main entries are based on the two-steps Heckman procedure. Secondaryentries, if available, laid below the main ones, are based on maximum-likelihood estimation.The dependent variable in the wage equations is the logarithm of hourlywage; and the independent variables are schooling, experience, squaredexperience, marital status, full time employment, large city, two-digits oc-cupational and two-digits industrial dummies.The selection equation includes, as explanatory variables, a set of schoolingdummies (for the schooling stages of 0, 1�11, 12, 13�14, 15, and 16+), age,age squared, large city dummy, and year of marriage dummies (or numberof years since �rst marriage).

56