Employment rates, earnings, and child support collections were higher after referral to SHARE than before, and there were clear differences in outcomes among NCPs who took different paths after referral. Several factors probably contributed to these results. It appears that the NCPs in this study had reached a low point in their ability to work and/or pay child support during the quarter before referral, and that the increases observed in subsequent quarters reflect some natural recovery from this low point. It also seems likely that different types of NCPs responded differently to the initiative based on unobserved characteristics that were present before referral. In addition to these factors, however, it appears that some or all of the components of SHARE service of a summons, threat of incarceration, offer to renegotiate obligations and arrears, availability of WtW services, and ongoing monitoring of compliance may have played a role in the improvements in outcomes observed for NCPs who became actively engaged in the initiative. In general, these NCPs worked more, earned more, and paid more support six to nine quarters after referral to SHARE than at any point during the year preceding referral. Moreover, differences between the employment, earnings, and child support outcomes for NCPs who appeared at hearings and learned about SHARE and the outcomes for NCPs who never appeared insignificant prior to referral to SHARE become more marked and significant during the quarters after referral to the program.
Without a random assignment evaluation, we cannot establish definitively that SHARE is responsible for the observed increases, the extent to which it influenced these outcomes, or how it influenced them. However, the available evidence suggests that the intervention is promising. A more rigorous evaluation of SHARE or of similar initiatives could shed light on the effects of the program relative to other factors. An ideal future evaluation would use a controlled experiment that would determine outcomes for participants and also provide information on how targeted NCPs would have fared in the absence of the intervention that is, the program's "value-added." In such an evaluation, it would be important to examine how various components of the intervention contribute to observed changes in the outcomes of interest, and their relative importance in achieving the desired results overall and for various types of NCPs (for example, NCPs who had criminal records before referral, or those with poor employment histories).
Before embarking on such an evaluation, it might be useful to understand better why so many NCPs fail to appear at their court hearings and, once located, what strategies could more effectively help this relatively disadvantaged group of individuals. Implementing those strategies before conducting a more rigorous evaluation could help programs reach out more effectively to all individuals who may benefit from their services not just those who are easier to reach and thereby result in a more thorough test of the intervention.
13. This pattern of recovery from a pre-program low is typically referred to as "Ashenfelter's dip," for his observation that adult participants in job training programs often experience a dip in earnings prior to their decision to participate (Ashenfelter and Card 1985).
14. The tables in Appendix A contain data supporting all the findings and statistics discussed in this chapter. Many, but not all, of the findings and statistics also are presented in the tables and figures in this chapter. Some of the tables in Appendix A present more statistics than are discussed in the chapter.
15. There is no way to distinguish interceptions of IRS refunds from voluntary payments or wage garnishment in the child support payment data in this study.
16. Unfortunately, no data on child support obligations were available for this study.
17. Sample sizes increased substantially during the quarter of referral to SHARE and during the quarter after referral because data were available for more NCPs during those quarters. Pre-post referral results are not biased by this change in sample size; an analysis of all outcomes limiting the post-referral sample to NCPs for whom pre-referral data are available reveals patterns consistent with those reported here.
18. This jump in average earnings during the quarter of referral relative to the quarter just before referral is likely to be, in part, an artifact of our quarterly data analysis. That is, for NCPs whose date of referral to SHARE falls relatively early in the quarter, the quarter of referral is likely to include some post-referral earnings.
19. The only exposure these NCPs may have had to the initiative was receipt of the summons to appear in court, and 46 percent, at most, of the NCPs who did not appear at a hearing seem likely to have experienced that event. As discussed in Chapter II, the majority of NCPs who did not appear at a hearing could not be located (128 closed cases) or were incarcerated (24 closed cases). Service of the summons alone is likely to have had only a small influence on the NCPs who did receive the summons but failed to appear at a hearing.
20. Research suggests that the accumulation of arrears to unrealistic levels and the state's retention of child support to offset the custodial parents' welfare payments may motivate NCPs to evade the child support system by moving out of formal jobs and into the underground economy (see Miller and Knox 2001).