1. See, for example, Wilson (1987); Kasarda (1995); Holzer and Vroman (1992).
2. See Becker (1971). Discriminatory behavior can persist if employers are catering to the tastes of customers, or due to a variety of labor market imperfections. "Statistical" discrimination, based on imperfect employer information about the productivities of individuals within different groups, can persist as well under a variety of conditions.
3. Most of the results cited below appear in Holzer (1996), though some of the qualitative evidence also appears in Moss and Tilly (1995).
4. Zemsky (1997); Kirschenman and Neckerman (1991).
5. Pavetti (1995).
6. Burtless (1995); Pavetti (1997).
7. Maynard (1995).
8. Freeman (1992).
9. Holzer and Danziger (1998).
10. Blank (1995); Burtless (1995).
11. Neal and Johnson (1996).
12. Trejo (1997).
13. Bendick et. al. (1994); Fix and Struyk (1994).
14. See, for example, Kirschenman and Neckerman (1991).
15. Wilson (1996) emphasizes that black employers share many of these negative impressions of young black workers, generating some question about whether these perceptions really reflect discrimination.
16. Donohue and Siegelman (1991).
17. Bloch (1994).
18. Holzer (1998); Holzer and Ihlanfeldt (1998); Lane, this volume.
19. Holzer (1996).
20. Kirschenman and Neckerman (1991).
21. The data on Hispanics are primarily drawn from Los Angeles in Holzer (1996). Evidence of hiring discrimination against Hispanics appears in Kenney and Wissoker (1994).
22. Farley (1995).
23. Jargowsky (1997).
24. Hughes and Sternberg (1992); Raphael (1997).
25. Ihlanfeldt (1997).
26. Holzer (1991); Kain (1992).
27. Holzer et al. (1994); Holzer and Ihlanfeldt (1996).
28. Kain (1992); Stoll et al. (1998).
29. Ihlanfeldt (1996); Raphael et al. (1998).
30. Holzer et al. (1994).
31. Falcon and Melendez (1997).
32. Holzer (1987a); Browne and Hewitt (1996).
33. Wilson (1987); O'Regan and Quigley (1996).
34. Holzer (1986).
35. Juhn (1992).
36. See Freeman (1992). Perceptions of higher returns in illegal than legal work are based mostly on relative wages, and may not involve adjustments for the risks of incarceration or long-term prospects.
37. Holzer and LaLonde (1998).
38. For discussion of employer perspectives on these issues, see the paper by Lane in this volume.
39. Of these, only CET has been formally and successfully evaluated (Melendez (1996)), though evaluation of replication efforts are still under way.
40. Enterprise Zones and other approaches that stress economic development in or near low-income neighborhoods are another way of overcoming the geographic "mismatch" problem. While evaluations of earlier efforts indicated that they were not cost-effective means of generating employment for zone residents (e.g., Papke (1992)), the more recent "Empowerment Zone" projects of the Clinton administration may prove somewhat more successful, as the funds can be used for a much broader range of community and labor force development activities.
41. Bishop (1993).
42. Currently, only firms with 100 or more employees (and smaller ones with federal contracts) are required to file EEO-1 forms so that the racial composition of their establishments can be monitored. A different approach might involve the use of auditors, who can be targeted toward smaller and suburban establishments. Another approach could involve the use of real job applicants, who are supported and encouraged to apply for suburban jobs while carefully recording all establishments to which they apply.
43. Examples of these approaches include the National Supported Work Demonstrations, Youth Corps, and Youth build. See the paper by Barnow in this volume for more discussion of these issues.
44. See, for example, Cutler and Glaeser (1997).
45. Rosenbaum and Popkin (1991).
46. Katz et al. (1997).
47. Haar (1996).