Customer discrimination and employment outcomes for minority workers.

I. INTRODUCTION

Empirical studies generally show that, even after controlling for individual productivity characteristics such as education and experience, blacks have lower levels of employment and earnings than whites. The residual earnings gap between races is commonly attributed to labor market discrimination.(1)

Becker [1971] has identified three possible sources of this discrimination: the prejudice of employers, workers, and customers. In his model, employer and coworker discrimination will not persist over time in competitive labor markets; some economists (e.g., Nardinelli and Simon [1990]) therefore conclude that any labor market discrimination that does persist most likely results from consumer prejudice.

In addition to its possible importance in explaining residual gaps in employment and earnings between blacks and whites, the potential presence of customer discrimination in the labor market is of interest for at least two related reasons. First, there is considerable evidence that blacks have been disadvantaged by job suburbanization (for reviews of this evidence see Holzer [1991] or Ihlanfeldt [1992]). However, the reasons for this remain somewhat speculative. Kain [1968] has suggested that customer discrimination may account for the failure of inner-city blacks to follow jobs to the suburbs.(2)

Second, declining employment in manufacturing and the growth in services employment may have increased the proportion of jobs requiring face-to-face contact with consumers. If consumer discrimination exists, growth in consumer-contact may help to explain recent relative declines in the relative earnings and employment of blacks [Bound and Freeman 1992].

The theoretical effects of customer discrimination have been explored in some detail to date. Blacks could potentially avoid any adverse consequences associated with customer discrimination by segregating themselves into firms that sell only to nondiscriminatory customers [Cain 1986]. To the extent that employees in many sectors (such as manufacturing) have little contact with customers, or that discriminatory consumers do not mind contact with blacks in lower-status occupations (such as blue-collar jobs), there would be a relatively large set of jobs where the hiring and wages of blacks would be unaffected by the discriminatory tastes of white customers.

However, the extent to which blacks can avoid wage losses by segregating themselves in the workplace depends on several factors, such as the relative sizes of sectors in which they do and do not face discrimination and the production technologies of each, as well as the relative sizes of the black and white workforces. Thus, a relative "crowding" of blacks into jobs in the nondiscriminatory sectors may result in lower earnings for them, and perhaps for whites in that sector as well.(3) Furthermore, a variety of other product and labor market characteristics might reduce the earnings of both blacks and whites in these jobs. For instance, firms in predominantly black neighborhoods might have lower capital-labor ratios or less advanced technologies than those in white neighborhoods [Bates 1993]. Their customers are likely to have lower incomes as well, perhaps leading to lower prices and product market rents and therefore lower wages in these firms.

In contrast to previous theoretical work on customer discrimination, the empirical evidence to date can best be described as fragmentary (based on very specific industries and occupations) or indirect. For instance, Kahn and Shearer [1988] analyze customer effects by focusing on the salaries of professional basketball players, while Nardinelli and Simon [1990] do so using the values of baseball cards for white and black players. Other analyses of customer discrimination in specific labor markets include Ihlanfeldt and Young's [1994] work on fast-food restaurants in Atlanta and Neumark's [1996] analysis of customer discrimination by gender in restaurants in Philadelphia. Broader but more indirect evidence appears in Ihlanfeldt and Sjoquist [1991] and Kenney and Wissoker [1994], both of which use the racial composition of residents in subcounty areas as proxies for customer composition of firms located in those areas.

This paper provides some new evidence on consumer discrimination in the labor market obtained from a unique new survey of employers. In contrast to previous work, the data allow for a much more direct examination of the issue across a more representative sample of firms and newly filled jobs. We use explicit information on the race of customers and newly hired employees at establishments, on the degree of contact employees have with the customers, and detailed controls for job and establishment characteristics (including location) to estimate the effects of customer composition on employment. Estimated effects on employee wages by race are provided as well.

II. DATA AND ESTIMATION ISSUES

The data used in this paper are drawn from a new survey of employers that was administered between June 1992 and May 1994 to 800 employers in each of four large metropolitan areas: Atlanta, Boston, Detroit, and Los Angeles. The survey was administered over the phone to individuals responsible for hiring, and focused on the characteristics of overall employees, vacant jobs, and the most recently filled job and hired worker at each establishment. Other characteristics of the establishment, such as its size, presence of collective bargaining, and the demographic composition of its applicants and customers, were gauged as well.

The sample of firms surveyed was drawn from two sources: roughly 30 percent were generated by employees who were respondents in a household survey in the same four metropolitan areas; and the rest were generated by lists provided by Survey Sampling Inc. (SSI).(4) The latter sample was drawn ex ante to reflect the distribution of workers across establishment size categories in the labor force; such a sample includes many more large establishments than would be found in a truly random sample of all establishments.(5) The sample drawn from the household surveys implicitly reflects this distribution as well. Thus, both samples are already weighted by employer size, permitting analysis of either individual jobs (such as the one most recently filled) at these firms or overall employment without using additional weighting for establishment size. Sample weights, however, are still necessary when analyzing the data to adjust for some nonrandomness in the chosen sample of establishments and jobs.(6) But response rates to the survey among firms that passed the screening averaged 67 percent, and there is little evidence of selection bias or additional nonrandomness induced by response patterns in the data.(7)

The results that we present below were obtained from estimating equations with the following general form

[R.sub.jk] = [Alpha] + [Beta][CUS.sub.jk] + [Gamma][X.sub.j] + [Delta][X.sub.k] + [[Epsilon].sub.jk],

where R denotes race (white, black, or Hispanic) of the last hired worker; CUS represents variables for the percentages of the firm's customers who are black and Hispanic; and the X reflect a variety of control variables; while j and k denote the last job filled and the establishment, respectively. CUS is measured by responses to the survey questions, "What percentage of the customers at your firm are ____?", where the question was asked repeatedly for blacks, Hispanics, and Asians.(8)

While estimating equations for the racial composition of all current employees would be more consistent with the theoretical models discussed above, defining the dependent variable as the race of the last worker hired has a number of advantages. First, the [X.sub.j] variables can be used to control more fully for job characteristics (such as skill requirements) that might be correlated with both race of customers and employees, thereby limiting bias from unobserved heterogeneity across firms and jobs. Second, while the race of customers is likely to be endogenous with respect to the racial composition of employees, customers' race should be at least somewhat more exogenous with respect to the race of the last hired worker in the establishment.(9)

Some versions of equation (1) allow for separate effects of customer composition on white, black, and Hispanic employees; in these cases, variables measuring percent black and percent Hispanic among customers are included among the independent variables to allow for cross-group as well as own-group effects. Other equations that attempt to control more completely for unobserved heterogeneity focus only on blacks versus nonblacks among customers and employees. The customer variables are alternatively entered in continuous and categorical form (e.g., 0-24 percent, 25-49 percent, etc.) in all estimated equations, where the latter are used to capture nonlinearities in the customer composition effects.

One of the most attractive aspects of our data is that a wide range of controls is included among the [X.sub.j] and [X.sub.k] variables. The [X.sub.j] include one-digit occupation dummies and a variety of dummy variables for the hiring requirements of jobs and the cognitive/social tasks performed on these jobs. The hiring requirements are a set of characteristics that applicants need to have in order to be hired; these include college or high school degree, general or specific experience, references, and previous training. The tasks must each be performed on a daily basis in the job; these include reading or writing of paragraph-length material, arithmetic calculations, direct contact with customers, or computer use.(10) The [X.sub.k] include one-digit industry, establishment size, presence of collective bargaining, and geographic location both between and within the various metropolitan areas.

The controls for firm location within the MSA are particularly important, since intrametropolitan location is expected to be highly correlated with both customer composition and the presence of minorities in the pool of labor facing the firm. We therefore include several variables to measure this location: dummy variables for whether the establishment is located within the central city and for whether it is within a quarter mile of a public transit stop; and, most importantly, the average distance of the establishment to the locations of the white, black and Hispanic residents living within the metropolitan area.(11)

To more fully control for the supply of minority labor to any particular establishment, controls are included for the fractions of applicants for a job who are black and Hispanic. We also include the race of the survey respondent (who was responsible for hiring at the establishment), to control for possible employer prejudice in hiring that might exist independently of customer discrimination effects. But both the race of applicants and respondents might themselves be functions of the racial composition of customers; and both, along with the race of customers, might simply reflect the geographic location of the establishment within the metropolitan area.(12)

Therefore, results are presented for three specifications of each equation: one without any controls for location within the metropolitan area or race of applicants and respondents (but with controls for job and firm characteristics more generally); one …