1. See Averett and Hotchkiss (1995, 1996, and 1997), Hotchkiss (1991), and Farber and Levy (2000).
2. This phenomenon, identified by Averett and Hotchkiss (1996), has been the experience of women in the labor market.
3. For a discussion of labor market segmentation and dual labor market theory and references to this literature, see Kaufman and Hotchkiss (2000, Chapter 6).
4. The bivariate probit model with selection gives rise to the following likelihood function:
where <2 is the bivariate normal cumulative distribution function and $ is the univariate normal cumulative distribution function.
5. This method of calculating the marginal effect of a change in a dummy variable is referred to as a measure of discrete change and is described in greater detail by Long (1997, pp. 135-138).
6. Table C.7 (column 1) in Appendix C contains the marginal effects used to generate Figure 4.2.
7. This type of pooled, cross-sectional analysis has been applied by many researchers (for example, Card 1992; Gruber 1994 and 1996; Zveglich and Rodgers 1996; and Hamermesh and Trejo 2000). The technique, however, also has its critics (such as Heckman 1996). The primary criticism of this pooled, cross-sectional approach is that it is impossible to control for unobserved changes in the environment that occurred at the same time as the event of interest. Issues of potential concern in this regard are explored in Chapter 2.
8. Voluntary part-time is defined as (1994-2000) working less than 35 hours per week and did not want to work full-time, and (1981-1993) reason for working less than 35 hours per week coded as 07-15 (see Stratton 1994 for justification). Category reasons 07-15 are holiday, labor dispute, bad weather, own illness, on vacation, too busy with school or house, did not want to work full-time, full-time work week is less than 35 hours, or other.
9. Table C.7 (column 2) in Appendix C contains the marginal effects used to generate Figure 4.3.
10. These results are consistent with those of Schur (2002), who finds that part-time and contingent work grows among the disabled during tight labor markets (where demand is strong relative to supply), which would be expected if these arrangements are voluntary.
11. The musculoskeletal category includes disabilities of the special senses (e.g., hearing, seeing).
12. See Magill (1997) for a detailed discussion about how flexible and part-time work schedules can often serve as low- or no-cost solutions to the accommodating problem. However, indirect costs, such as having to hire additional workers to cover lost hours of a disabled part-time worker, should not be ignored.
13. The SIPP data showed the same differential decline for both disabled and nondis-abled workers as seen in the CPS data, but essentially inconsequential differences across disability status in any given year.
14. An additional distribution index, the IP Index (see Watts 1992), was also evaluated, and the conclusions were the same. While there is a direct mathematical relationship between these two indices, the IP index reflects the percentage of workers in the labor market that would have to switch occupation or industry in order to equalize the distribution of disabled and nondisabled workers across occupations or industries, while maintaining the original occupational structure. Since the conclusions were the same regardless of the index employed, the more familiar Duncan Index is detailed here.
15. The actual indices plotted in Figure 4.6 are contained in Table C.8 in Appendix C. The indices calculated for the distribution across industries are consistent with what Yelin and Cisternas (1996) calculated using the National Health Interview Survey; their data indicate that the relative distributions of disabled and nondis-abled workers have been fairly steady as far back as 1970. The occupational categories in their data were not comparable to those in the CPS, however.
16. Employment growth was calculated as the percentage change in the number of workers in an occupation from one time period to the next.
This Page Intentionally Left Blank
Was this article helpful?