# The authors thank Danielle Gao and Ryan Helwig for research assistance, and Felicity Skidmore for editing.
1 Workfare (publicly subsidized work) is supposed to be a stopgap for those unable to find private-sector work in the short term.
2 See Harrison and Sum (1979), Gordon (1972), Piore (1975), and Howell (1997).
3 This concept was introduced by former Labor Secretary John Dunlop (1979).
4 See Spriggs and Klein (1994) and Spriggs and Schmitt (1996).
5 All the poverty-level wage calculations in this paper refer to the poverty-level wage for a family of four. Using the poverty-level wage for a family of three does not change the qualitative or quantitative results.
6 Organization for Economic Cooperation and Development (OECD) (1997) defined the low-wage cutoff as two-thirds of the median wage. Interestingly, the OECD finds no evidence of higher mobility among low earners in the less-regulated United Kingdom and United States compared to other countries in its study.
7 See Topel (1993).
8 See Bernstein (1997).
9 This trend has since slowed.
10 See Mishel et al. (1999a), table 3.21.
11 This family would have been ineligible for health coverage under Medicaid in 1997. Subsequent program changes have now made the children in such a family eligible for Medicaid coverage, raising the family's living standard a bit higher (Currie and Yelowitz 1998).
12 This comparison involves crossing the coding change in the Current Population Survey (CPS) education variable. The education category that changed the most was "some college." Those who had completed 13 to 15 years of schooling in pre-1992 files were labeled "some college." The new coding differentiates between those with associate degrees and those with some college. Since these are percentages that together cover everyone, the coding change only introduces error to the extent that those with high school or less would have been classified differently under the two coding schemes. Evidence from the 1990 CPS, which includes both coding formats, suggests a coding-induced shift from high school to some college — making the changes shown in table 3 overestimates of the educational upgrading that took place over the period.
13 See appendix at the end of this chapter.
14 The economic returns and the workforce characteristics may in fact affect each other to some degree, but not enough to change the nature of the broad trends discussed in this paper.
15 Evidence for this is seen in the increase in the intercept term shown in appendix, table A1.
16 The combination of the findings for women are particularly worth noting. The net effect of declining economic returns to work and the negative structural factors would have led to a 4.9 percent increase in the share of the female workforce in the low-wage sector between 1979 and 1997. But their actual share in the low-wage sector fell over the period, by 1.8 percent. Thus, improvements made by women in education, occupation, and experience more than reversed the impact of the negative factors.
17 A recent example of the demand-shift argument was made by Johnson (1997). Institutional arguments can be found in Fortin and Lemieux (1997), Howell (1997), and Mishel et al. (1999a).
18 See Holzer (1996).
19 For example, the real wages of entry-level (one to five years' experience), college-educated workers fell by about 7 percent for both men and women during the 1989-97 period (Mishel et al. 1999a).
20 Of the 10 occupations projected to add the most jobs over the 1996-2006 period, 7 call for high school or less in terms of skill demands, and 5 are in the lowest pay category (Silvestri 1997, table 4).
21 See Mishel et al. (1999a).
22 See Blank and Blinder (1986) and Blank and Card (1993).
23 See Fortin and Lemieux (1997).
24 See Bernstein and Schmitt (1998). A policy related to the minimum wage is the living wage movement, which has been successfully passed in ordinances in numerous cities enforcing pay levels above the minimum for workers in firms with city contracts (Bernstein 1998).
25 As with immigration, various analysts have argued that the welfare-to-work component of welfare reform has the potential to further increase the supply of low-wage workers.
26 See Katz (1998).
27 Danziger and Acs (1997) do a similar decomposition.