Discussion - comment on Peter Cappelli, in this issue - Special Issue: Earnings Inequality

This paper by Harry Holzer and Keith Ihlanfeldt is an important contribution to the study of the "spatial mismatch hypothesis." Much of our knowledge has come from worker surveys, and their employer survey gives a different perspective, with better information about types of jobs, skill demands, and hiring requirements than most worker surveys offer. I have a number of quibbles about some measures, but these strong findings are remarkable. Indeed, better measures might give even stronger results.

The analyses show that employer proximity to black residences and to public transit increases the likelihood that employers will hire black employees. This is a strong finding, and it holds up after many controls. Task requirements of math or computer skills reduce black hiring, although talking to customers and reading/writing tasks do not. While two hiring requirements affect black hiring (college diploma and vocational training), others have no effect.

Recruiting method, which Holzer (1987) has studied elsewhere, is discussed only briefly here. Holzer and Ihlanfeldt stress that employers using help-wanted signs, walk-ins, and employee referrals have strong negative distance influences on black hiring, especially for applicants with high school education or less. They also find that school referrals have a strong effect (about twice its standard error), a point we will return to later.

The authors find that employers' relative distance from blacks increases wages. However, these effects may largely reflect firm attributes, for they nearly vanish after firm attributes are controlled in the total model, and they are severely reduced in the equation for blacks' wages. The authors say that the estimated distance effects "remain at least marginally significant," but the table suggests that after controls, few coefficients are twice their standard error, and then mostly in one city - Atlanta. At best, the wage effects seem ambiguous.

As is always the case with nonexperimental data, one must wonder about some of the findings: in this case, whether the black workers hired in white suburbs are somehow distinctive individuals in ways not captured by controls. The controls are as good as one finds in most studies, but such doubts can never be completely dispelled by survey methods.

Here I can provide some supportive evidence. In Chicago's Gautreaux residential mobility program, low-income blacks who applied to the program were randomly assigned to move to city or suburban areas, creating a quasi-experimental design. We studied city moves to predominantly black, low-income areas, and suburban moves to predominantly white, middle-income areas, scattered over 120 towns in the six counties around Chicago. The study of 300 adults found significantly higher employment among those in the suburbs than among those in the city, but earnings and hours per week were not different, even after extensive controls (Popkin, Rosenbaum, and Meaden 1993). In a study of 107 children, suburban movers were more likely to graduate from high school, attend college, and attend four-year colleges. Among those who did not attend college, suburban movers were more likely to get jobs and to get jobs with better pay and benefits (Kaufman and Rosenbaum 1992). The outcome differences between city and suburban movers were large for adults and even larger for children (Rosenbaum 1995; Rosenbaum et al. 1991; and see my Tables 1 to 3).

The adult findings in that quasi-experimental study support Holzer and Ihlanfeldt's findings that reduced distance can improve employment, but their findings also suggest that reduced distance may not raise adults' wages. To improve adults' wages, I suspect that programs must not only get people to the job, they must also provide people with the qualifications employers seek and with signals of those qualifications that employers trust. This suggests an interaction term - that applicants with trusted signals of qualifications will get larger pay gains from mobility programs than other applicants. People who lack qualifications, or lack trusted signals of them, will only get the same kinds of low-paid jobs in the suburbs that are available in the city.

Table 1
Percent of Respondents Employed Post-Move, Classified by Pre-Move
Employment, for Movers to City and to Suburbs

                                            Pre-Move
                           Employed        Unemployed       Total

City Post-Move

Employed                       42               13             55
                            (64.6%)          (30.2%)
Unemployed                     23               30             53
                            (35.4%)          (69.8%)
Total                          65               43            108

Suburb Post-Move

Employed                      106               37            143
                            (73.6%)          (46.2%)
Unemployed                     38               43             81
                            (26.4%)          (53.8%)
Total                         144               80            224

8 Numbers in parentheses are column percentages.
Source: Popkin, Rosenbaum, and Meaden (1993).
Table 2
City and Suburban Comparison on Wages and Hours Worked

                       Pre-Move      Post-Move
                         Mean           Mean         t     p

City Movers             N = 55          42
Hourly wages             $5.04         $6.20        6.52    .00
Hours/Week               33.27         31.92        -.60.    55

Suburban Movers         N = 143        106
Hourly wages             $4.96         $6.00        6.50    .00
Hours/Week               33.62         33.39        -.60    .55

Source: Popkin, Rosenbaum, and Meaden (1993).

To understand why, we must consider what qualifications employers want. According to most employer surveys, the employer's highest priority is the worker's personality. While scholars argue that employers ought to stress academic skills, one cannot ignore employers' stated concerns. In interviews with 51 employers, we found that employers' concerns about workers' personalities do not arise from subtle preferences but from terrible experiences with workers who do not come to work, do not do their share, disobey supervisors, harass or fight co-workers, and damage property (Rosenbaum and Binder 1994). Unfortunately, some employers feel such behaviors are associated with urban blacks and urban public schools, so mobility programs must do more than transport black workers to an employer's door.

What signals do employers use to infer applicants' qualifications? Many studies show that employers do not use high school diplomas, grades, or references in hiring (Bishop 1989; Rosenbaum et al. 1990; Rosenbaum 1996), even though cognitive skills predict wages six years after graduation (Murnane, Willett, and Levy 1994). Holzer and Ihlandfeldt also find that many hiring requirements have no effect. The employers we interviewed suggest an explanation: They believe that high school diplomas and good recommendations say little about applicants' work habits and do not even guarantee eighth-grade math and reading skills.

Instead of using such indicators, employers devise a variety of procedures that they believe screen out bad risks, but unfortunately their procedures are likely to be both ineffective and discriminatory. Employers report that they expect good workers to have a firm handshake, traditional hair styles, certain clothing styles, and so on (Rosenbaum and Binder 1994). Some of their "tests" entail conflicting demands, like those of the employer who expected applicants to speak assertively but not be insolent to supervisors.

This kind of desperate quasi-rational grasping for signals about applicants sounds a lot like statistical discrimination. This is both bad news and good news. It is bad news, because it indicates racial bias. But while bias based on prejudice can be reduced only by attitude change or coercive policies, statistical discrimination can be reduced simply by giving employers better signals than the discriminatory ones they are now using. For instance, a study of 185 employers finds that employers' hiring practices "do not discriminate against all black applicants, but simply against those they perceive as lower-class" and lacking in certain skills (Neckerman and Kirschenman 1990, p. 20). That study also finds that employers who use skill tests to discern applicants' abilities are more likely to hire blacks than those who do not use such tests. Apparently, if employers are reassured about applicants' skills, they are more likely to hire blacks.

Table 3
Youths' Education and Job Outcomes: City-Suburban Comparison

Percent                                    City    Suburb   Sig.(4)

Number of youths                            39       68
Drop out of school                          20        5      (*)
College track                               24       40      (**)
Attend college                              21       54      (***)
Attend four-year college                     4       27      (**)
Employed full-time (if not in college)      41       75      (****)
Pay under $3.50/hour                        43        9      (****)
Pay over $6.50/hour                          5       21      (****)
Job benefits                                23       55      (****)

a Significance of chi-square or t-test: * p [less than] .10, ** p
[less than] .05, *** p [less than] .025, **** p [less than] .005.

Source: Rosenbaum (1995); Rosenbaum et al. (1991).

These considerations have important implications for transportation programs, since transportation programs may have difficulty providing such information, for two reasons. First, residential location, which is usually considered a distance factor, is also a signal. It is among the ad hoc procedures employers use for assessing applicants. Many employers consider a housing-project address, a central-city address, or attendance at a city public school as signals of poor workers. Transportation programs will not fix these residential barriers.

Second, transportation programs do not help applicants present dependable information about themselves to employers. Indeed, they may move people away from the informal networks that could signal their positive attributes. Studies by Granovetter (1995) and Holzer (1987) find that informal contacts improve hiring. Holzer and Ihlanfeldt show that school referrals affect hiring. Bishop (1993) finds that references from vocational teachers and previous supervisors (particularly ones known by an employer) have significant positive effects on worker productivity. In contrast, more anonymous recommendations, from previous personnel offices and public employment agencies, have negative effects on productivity. (See also Kariya and Rosenbaum 1995.)

In a detailed qualitative study of 51 employers, we found that some employers use teacher contacts as a way of getting trusted information about students' work habits. Moreover, these contacts are particularly important for minorities. If a trusted teacher recommends a black to be as good as previously recommended whites, then employers are willing to take a chance that they would not have taken otherwise (Rosenbaum and Miller 1995; Rosenbaum and Jones 1995).

In another study, analyses of the High School and Beyond data find that school help is an important source of first jobs for some students. We find that females and minorities are more likely to get their first jobs from school help than are white males. We also find that while white males get the largest wage benefits from school help, black males also get significant wage benefits that they would not have gotten without that help (Rosenbaum, Roy, and Kariya 1995).

Thus, while Holzer and Ihlanfeldt advocate both transportation and residential mobility programs, our analysis suggests some difficulties with transportation programs. While they can make distant employers more available, they do not counteract employers' use of urban addresses as negative signals and they do not necessarily get trustworthy information about workers to employers. In contrast, residential integration gives blacks "non-stigmatized addresses," and it may help residents get informal signals from their church, neighbors, or schools that employers may trust.

Of course, the strongest findings in my studies have been for children. The biggest gains from residential mobility appear in the second generation. I do not know of another program for low-income black youth that doubles the rates of college attendance, employment, good pay, and job benefits. The employment gains came in part from informal contacts that teenagers made with local employers. Obviously, children's gains cannot come from adult transportation programs, unless we also provide school busing.

In sum, I conclude that reducing distance barriers may be necessary, but not sufficient. For urban blacks to get better-paid jobs, they must be able to present credentials that reassure employers about their qualifications. Mobility programs will be most effective at raising wages if they can certify participants. They must reassure employers that these urban blacks differ from employers' stereotypes, which now create their statistical discrimination. Mobility programs that also provide certification of workers' academic skills, school attendance and behavior, previous work experience, or previous volunteering experience will have greater effectiveness at overcoming employers' statistical discrimination. Transportation programs may be able to do this, but they may have greater difficulties than residential mobility programs. To the extent that informal networks are employers' most trusted source of information (as our studies imply), residential integration is more likely than transportation programs to help blacks get their qualifications communicated through such informal networks.

Interestingly, housing programs can also make use of signals to overcome statistical discrimination by landlords and neighbors. The Gautreaux program used some selection criteria to reassure landlords that participants had good rent payment records and did not destroy their apartments. These were not stringent selection criteria. They eliminated only one-third of applicants, but they helped persuade landlords to take participants. Cincinnati's HOME program used similar selection criteria to win landlord support. Unfortunately, a federal demonstration program to replicate Gautreaux, Moving to Opportunity, did not clearly state such assurances, and the city of Baltimore panicked over nightmarish visions of felony criminals sweeping through the suburbs. Failure to deal with statistical discrimination can undermine the effectiveness of housing mobility programs, just as it undermines employment.

References

Bishop, John. 1989. "Why the Apathy in American High Schools?" Educational Researcher, vol. 18, no. 1, pp. 6-10.

-----. 1993. "Improving Job Matches in the U.S. Labor Market." Brookings Papers: Microeconomics, pp. 335-400. Washington, D.C.: The Brookings Institution.

Granovetter, Mark. 1995. Getting a Job. 2nd Ed. Chicago, IL: University of Chicago Press.

Holzer, Harry. 1987. "Informal Job Search and Black Youth Unemployment." The American Economic Review, vol. 77, no. 3 (June), pp. 446-52.

Kariya, Takehiko and James Rosenbaum. 1995. "Institutional Linkages between Education and Work as Quasi-Internal Labor Markets." Research in Social Stratification and Mobility, vol. 14, pp. 101-36.

Kaufman, Julie and James Rosenbaum. 1992. "The Education and Employment of Low-Income Black Youth in White Suburbs." Educational Evaluation and Policy Analysis, vol. 14, no. 3 (Fall), pp. 229-40.

Murnane, Richard J., John B. Willett, and Frank Levy. 1994. "The Growing Importance of Cognitive Skills in Wage Determination." Unpublished paper, Graduate School of Education, Harvard University.

Neckerman, Kathryn M. and J. Kirschenman. 1990. "Hiring Strategies, Racial Bias, and Inner-City Workers." Paper presented at the annual meeting of the American Sociological Association, Washington, D.C., August 11.

Popkin, Susan, James Rosenbaum, and Patricia Meaden. 1993. "Labor Market Experiences of Low-Income Black Women in Middle-Class Suburbs." Journal of Policy Analysis and Management, vol. 12, no. 3, pp. 556-73.

Rosenbaum, James. 1995. "Housing Mobility Strategies for Changing the Geography of Opportunity." Housing Policy Debate, vol. 6, no. 1, pp. 231-70.

-----. 1996. "Policy Uses of Research on the High School-to-Work Transition." Sociology of Education (forthcoming).

Rosenbaum, James and Amy Binder. 1994. "Do Employers Really Need More Educated Youth?" Presented at the annual meeting of the American Sociological Association. Los Angeles, August 7.

Rosenbaum, James and Stephanie Jones. 1995. "Creating Linkages in the High School-to-Work Transition." In Restructuring Schools, ed. Maureen Hallinan, pp. 235-58. New York: Plenum.

Rosenbaum, James and Takehiko Kariya. 1991. "Do School Achievements Affect the Early Jobs of High School Graduates? - Results from the High School and Beyond Surveys in the United States and Japan." Sociology of Education, vol. 64, no. 2 (April), pp. 78-95.

Rosenbaum, James, Takehiko Kariya, Richard Settersten, and Tony Maier. 1990. "Market and Network Theories of the Transition from High School to Work: Their Application to Industrialized Societies." Annual Review of Sociology, vol. 16, pp. 263-99.

Rosenbaum, James and Shazia R. Miller. 1995. "Moving In, Up, or Out: Tournaments and Other Institutional Signals of Career Attainment." In Boundaryless Careers, ed. M. Arthur and D. Rousseau. Oxford, U.K.: Oxford University Press.

Rosenbaum, James, Susan Popkin, Julie Kaufman, and Jennifer Rusin. 1991. "Social Integration of Low-Income Black Adults in Middle-Class White Suburbs." Social Problems, vol. 38, no. 4 (November), pp. 448-61.

Rosenbaum, James, Kevin Roy, and Takehiko Kariya. 1995. "Do High Schools Help Some Students Enter the Labor Market?" Paper presented at the annual meeting of the American Sociological Association, Washington, D.C., August.

James E. Rosenbaum, Professor of Sociology, Education, and Social Policy, Northwestern University.

COPYRIGHT 1996 Federal Reserve Bank of Boston
COPYRIGHT 2004 Gale Group