top of page

Be Notified of New Research Summaries -

It's Free!

Can Job Coaching Help Single Parents on Welfare Earn More for Decades?

Writer: Greg Thorson
Greg Thorson
3 minutes ago
6 min read


Can a few years of personal job coaching change a single parent's working life for decades? Bobonis et al. (forthcoming) study a 1990s Canadian experiment that randomly offered 892 long-term welfare recipients, nearly all single mothers, a bonus for taking full-time work, the same bonus plus intensive job coaching, or neither. Using 20 years of tax records, they find that coaching raised annual earnings by about $1,600 to $3,000 (Canadian), or 21 to 27 percent, throughout the period. Coached parents were also more likely to work full time and less likely to receive welfare for about a decade. The bonus alone produced no lasting earnings gains.


Why This Article Was Selected for The Policy Scientist

Work requirements and welfare-to-work programs are again at the center of American social policy, as recent federal legislation extends work requirements in Medicaid and the Supplemental Nutrition Assistance Program and states look for ways to help recipients meet them. A central unresolved question is whether the services attached to these programs produce durable gains or only short-lived ones. Reviews such as Card, Kluve, and Weber (2018) show that active labor market programs often raise employment for a few years, but almost no rigorous evidence follows participants for two decades. This study does, linking a three-arm randomized trial to Canadian tax and employer records with match rates near 90 percent even 20 years later. The authors include Philip Oreopoulos and W. Craig Riddell, who have published extensively on labor markets, education, and the Self-Sufficiency Project itself. The design cleanly isolates the effect of coaching, though the sample is small and drawn from one province in the 1990s. The article is forthcoming in American Economic Journal: Applied Economics, one of the leading journals in applied economics.


Full Citation and Link to Article

Bobonis, G. J., Bonikowska, A., Oreopoulos, P., Riddell, W. C., & Ryan, S. P. (Forthcoming). A helping hand goes a long way: Long-term effects of counseling and support for welfare-to-work program participants. American Economic Journal: Applied Economics.


Central Research Question

The study asks whether intensive, one-on-one employment support for long-term welfare recipients produces earnings and employment gains that last long after the services end. It estimates the added effect of combining job coaching with a generous, time-limited earnings supplement and traces that effect over 20 years. It also asks how such services work, whether through more job search, moves to better jobs and employers, more education, or changed beliefs about personal control, and whether the services paid for themselves.


Previous Literature

Reviews by Kahn (2012), Crépon and Van Den Berg (2016), and Card, Kluve, and Weber (2018) conclude that job search assistance and related services generally raise employment and earnings for two to three years, but few studies look beyond five years; Schochet (2021) finds modest long-term employment gains but no earnings effects from Job Corps. Earlier evaluations of the Self-Sufficiency Project by Michalopoulos et al. (2002) and Robins, Michalopoulos, and Foley (2008) documented short- and medium-term gains from the Plus services, while Card and Hyslop (2005) and Riddell and Riddell (2014) found that the supplement's effects faded once payments stopped. Research on caseworkers (Arni and Schiprowski 2019; Schiprowski 2020) and on firm pay premiums (Card, Heining, and Kline 2013) suggests that both who helps job seekers and where they end up working matter for earnings.


Data

Between November 1994 and March 1995, 892 single parents in New Brunswick who had received welfare in at least 11 of the previous 12 months were randomly assigned: 293 to the Plus group, 296 to the Regular group, and 303 to a control group. Plus and Regular members who left welfare for full-time work within 12 months could receive, for up to 36 months, a supplement equal to half the gap between their earnings and a target of $30,600 (1994 Canadian dollars). Only the Plus group was also offered voluntary services, including a personal job coach who helped with résumés, interviews, job leads, workplace conflicts, and requests for raises, before and after participants found work. More than 95 percent of participants were women, and more than 90 percent had no more than a high school education.


Surveys at 18, 36, and 54 months, with response rates of 97, 92, and 86 percent, measure job search, job changes, and wages. Statistics Canada tax and employer–employee records measure earnings, employment, employer pay, and welfare receipt from two years before to 20 years after random assignment. Match rates exceed 98 percent early on and remain 87 to 90 percent in years 17–20, with no significant differences across groups.


Methods

The study estimates intent-to-treat effects by comparing average outcomes across randomly assigned groups, pooling the tax data into four-year periods and clustering standard errors by person. The main comparison is Plus versus Regular: because both groups faced the same financial incentive, the difference isolates the added effect of offering employment services.


The design has limitations. The small sample limits precision and rules out subgroup analysis. Plus members differed on a few baseline traits, such as having more children, although a joint test does not reject balance. Because coaching also raised the share who qualified for the supplement, part of the effect runs through the financial incentive. The design cannot detect displacement of other job seekers, and full-time work is inferred from annual earnings thresholds rather than recorded hours.


Findings/Size Effects

Within 12 months, 51 percent of the Plus group had qualified for the supplement, compared with 34.5 percent of the Regular group, a 16.4 percentage-point increase (about 47.5 percent). Plus members were 16 percentage points (50 percent) more likely to join job search activities and 12.3 points more likely to leave a job voluntarily, often moving directly to another. Over the next three years, their hourly wages were 0.64 to 0.78 Canadian dollars higher, a 7.8 to 8.9 percent increase.


Relative to the Regular group, Plus members earned $1,638 more per year in years 1–4 (27 percent above the Regular group's $6,025), $2,010 more in years 5–8 (27 percent), $2,634 in years 9–12 (about 27 percent), $2,998 in years 13–16 (about 23 percent), and $2,816 in years 17–20 (about 21 percent), in 2010 Canadian dollars. Based on the reported standard errors, the first four estimates are significant at the 5 percent level or better and the last at the 10 percent level. The Regular group, by contrast, earned no more than the control group once the supplement ended.


Plus members were 6.4 percentage points more likely to work full time in years 1–4 (27 percent above a 24 percent base) and in years 5–8 (21 percent), and 7.4 points more likely in years 9–12 (19 percent above 38 percent), roughly 64 to 74 more full-time workers per 1,000 parents offered services. Later estimates are positive but not significant, and in the second decade the groups were equally likely to have any job, so later earnings gains reflect more hours and higher pay rather than more people working. Welfare receipt fell by 5.9 percentage points in years 1–4 (7 percent) and 11 points in years 5–8 (18 percent), with insignificant declines afterward.


The gains appear to come from better job matches: in years 5–8, Plus members worked at firms where pay at the 25th and 50th percentiles was 13.5 and 11.4 percent higher. Evidence of more education is minimal, and changes in participants' sense of control were not statistically significant.


The services cost $1,714 more per participant than the Regular program. Over 21 years, Plus members received $9,000 less in welfare and paid $3,700 more in taxes, for a net government gain of about $7,500 after other transfers, which is not statistically significant. Their cumulative after-tax income rose by a significant $22,900, about 8.5 percent. The estimated marginal value of public funds is infinite, meaning the program appears to pay for itself, but its 95 percent confidence interval runs from 0.20 to infinity.


Conclusion

The study provides rare experimental evidence that intensive, personal employment support for long-term welfare recipients can raise earnings for two decades. Coached participants found full-time work sooner, moved to better-paying employers, relied less on welfare for about a decade, and earned 21 to 27 percent more each year, while the financial incentive alone produced gains that vanished once payments stopped. This suggests that the content of welfare-to-work services, not only their incentives, shapes long-run outcomes.


The evidence comes from a small sample in one province in the mid-1990s, served by highly motivated caseworkers and paired with a work-first incentive, so effects may differ at scale, in other labor markets, or without a supplement. The fiscal estimates are imprecise, and broader labor-market effects are unmeasured. Future research could test intensive coaching at scale, as a standalone service, and with designs that identify which caseworker practices drive the gains.

Comments

Rated 0 out of 5 stars.
No ratings yet

Add a rating
Screenshot of Greg Thorson
  • Facebook
  • Twitter
  • LinkedIn


The Policy Scientist

Offering Concise Summaries*
of the
Most Recent, Impactful 
Public Policy Research

*Summaries Powered by ChatGPT

bottom of page