Content provided in partnership with

Business Services Industry

More on the cost of being other than white and male: measurement of race, ethnic, and gender effects on yearly earnings

American Journal of Economics and Sociology, The, Jan, 1998 by Garey C. Durden, Patricia E. Gaynor

Introduction

This study is designed to measure the costs, in reduced yearly earnings, of being black or Hispanic (Mexican-American, Puerto-Rican), being female, or being both female and black or Hispanic. Following Corcoran and Duncan (1979) and the study in this journal by Jeremiah Cotton (1988), we are particularly interested in determining whether gender-based costs continue to exceed those associated with race or Hispanic origin and, if so, by how much. Isolation of the relative cost of gender is an important first step because such knowledge will provide a firmer basis for studies of the causes of inequality and the corresponding application of corrective policies. If earnings costs by gender are worse than race or ethnic-based costs, then gender-specific policy tailoring may be necessary.(1)

Most Popular Articles in News: The Ten Best Laptop bags; Tata plans cheapest-ever car for Indian market; GLOBALIZATION AND THE DEVELOPMENT OF UNDERDEVELOPMENT OF THE THIRD WORLD; Corn is good for you; Corn is not only a tasty treat, but also a cereal that ...; THE 50 BEST STYLISH HANDBAGS TO CARRY; More »

For purposes of this study the pure cost of being black or Hispanic is determined by male-male comparisons - that is, the amount by which white male estimated earnings exceed those for black or Hispanic males. Pure gender cost is the amount by which white male estimated earnings exceed those of white females. If earnings costs for black and Hispanic females exceed pure gender costs, the extra burden is assumed to be the negative result of being neither white nor male.

To make the necessary comparisons, we use Cotton's (1988) approach to estimate yearly earnings for each group, then determine total earnings costs for each (white males earnings-reference group earnings), and decompose total costs into skill-based and residual components. The former adjusts for differences in individual characteristics such as education, industry, and occupation type. The latter is assumed to result, at least in part, from discriminatory labor-market practices. The Cotton estimation procedure extends the traditional Oaxaca (1973) approach to account for changes in labor market demand and supply if discrimination were eliminated. The model is based in standard human-capital theory and ordinary least squares is the regression method.

Previous Work

Many studies make male-male and female-female earnings cost comparisons [for example Bloch and Smith (1977); Gwartney and Long (1978)] and find that white males generally earn more than minority males and white females earn more than minority counterparts (Asian groups are exceptions). Surprisingly, relatively few have specifically examined the extent to which gender costs might exceed those stemming from racial or ethnic characteristics.

There are at least two important exceptions, one by Corcoran and Duncan (CD, 1979) using 1975 hourly earnings data from the Panel Study of Income Dynamics, and another by Cotton (1988) using 1979 census data. CD found that the estimated hourly earnings for white males in 1975 to be $5.60, while black males earned $4.31, a race-based cost of $1.29. Earnings of white females were $3.61, leaving a pure gender cost of $1.99. The ratio of gender to race is $1.99/$1.29 = 1.53, suggesting that total gender cost is 1.53 times greater than total race costs. Black females earn only $3.17, a relative cost of $2.43 ($2.43/$1.29 = 1.91), indicating that, for this group, the gender-race mixed cost is 1.91 times the cost of race alone. When CD decomposed the total cost for each group, the discrimination-based residual portion for black males was $0.62, while that for white and black females was $1.17 and $1.62, with corresponding gender/race ratios of 1.88 and 2.60. This suggests the possibility that gender-based discrimination is much more in evidence than that associated with race alone.

Cotton (1988) made the same sorts of comparisons as CD, using the logs of hourly earnings and other variables from the 1979 census to estimate human-capital based equations. While briefly discussed here, the details from Cotton are summarized in Table One.(2) He found the total earnings (race) cost for black males to be .174, the gender cost to be .348 (white females differential) and the race-gender (black females) cost to be .461. The corresponding ratios are: pure gender cost relative to race = .348/.174 = 2.0; gender-race mixed cost relative to race alone, .461/.174 = 2.65, and these results are similar to those obtained by CD using very different data.

Cotton estimated the unexplained, discrimination-based portion of the difference between black and white males to be .045, only 25.9 percent of the total race cost. The unexplained pure gender cost is relatively much greater, .217 (for white females, 62.4 percent of the total cost) and the unexplained gender-race cost is greater still,. 272 (for black females, 58.9 percent of the total cost). Corresponding ratios are gender to race, .217/.046 = 4.82 and gender-race mixture to race, .272/.046 = 6.04. The larger ratios for the unexplained residuals found here (as compared to those from the CD study) are due to the fact that Cotton's estimations included proxy variables for industry and occupation. This tends to increase the explained portion of the total differential for males relative to females.

Business Services Industry

More on the cost of being other than white and male: measurement of race, ethnic, and gender effects on yearly earnings

American Journal of Economics and Sociology, The, Jan, 1998 by Garey C. Durden, Patricia E. Gaynor

Find Research Guides for:

Find Featured Titles for: Business

Most Popular Articles in News

Most Popular Publications in News