1. Increasing Work and Earnings Among Families in Social Housing James A. Riccio Centre for Economic and Social Inclusion 2012 Welfare-to-Work Conference Birmingham, UK

2. MDRC Not-for-profit social policy research organization Based in NYC Rigorously evaluates (and sometimes helps design) innovative social policies Pioneered large-scale random assignment evaluations of social programs Mission: Build evidence to improve the lives of low-income families 2

3. Topics General theme: Is social housing a good “platform” for intervention to improve work outcomes for low-income families? Evidence from the US: Jobs-Plus: A place-based employment intervention for residents of public housing (social housing) Evidence from the UK: ERA (Employment Retention and Advancement demonstration): Results for social housing residents More evidence from the US: NYC Work Rewards demonstration for “Housing Choice Voucher” recipients (like Housing Benefit) Final reflections 3

4. National Jobs-Plus Demonstration Target group: Residents of public housing (estates) Place-based employment intervention Multi-component, “saturation” strategy 4

5. Origins of Jobs-Plus—poverty and place Build mixed-income communities “from within” Response to growing concentration of joblessness, underemployment, welfare receipt, and poverty in public housing and surrounding neighborhoods Public and private sponsors: US Dept. of Housing and Urban Development (HUD) The Rockefeller Foundation Other public and private funders 5

6. Jobs-Plus sites Diverse housing developments in 6 cities: Baltimore Chattanooga Dayton Los Angeles St. Paul Seattle Randomly allocated developments within each city to program and control groups Local partnerships and collaboration: Public housing authorities Welfare agencies Workforce agencies Residents Other service agencies Mandatory partners 6

7. Jobs-Plus model: Comprehensive approach 3 components: 1. Employment and training services Convenient on-site “job centers” 2. New rent rules to “make work pay” Rent rises less as earnings grow 3. Community support for work Neighbor-to-neighbor outreach (“Community coaches” share info about work) “Saturation-level” outreach and assistance - Aim to assistance all working-age residents 7

8. Job assistance and service brokering on-site, where people live Creates easier opportunities and many informal opportunities to meet with, advise, encourage, and assist residents: Often on a “drop-in” basis at the Jobs-Plus office In the neighborhood (e.g., on a corner; at events) At residents’ homes Staff closer to residents’ day-to-day lives More holistic understanding of family problems, support networks, and neighborhood conditions 1. Employment and training component 8

9. Making work “pay” for everyone through new rent rules With traditional “30-percent-of-income” rule, earnings are implicitly “taxed” at 30% New rent rules: “Flat/fixed” rents (with income- based rents as “safety net” if job loss) 2. Rent reform 9

10. Promote “neighbor-to-neighbor” support for work, with aid of resident outreach workers “Building captains,” “community coaches… share information on job and training opportunities, other services, and financial incentives encourage participation in Jobs-Plus and employment add legitimacy to the program and open up further informal avenues of helping other residents 3. Community support for work 10

11. Foster integration of housing services and employment assistance Housing authorities promote employment from time new residents move in—message “comes with tenancy” Housing authorities link employment assistance with efforts to head off evictions over nonpayment of rent Housing authorities must transcend traditional, nearly exclusive focus on housing management Community support for work (cont.) 11

12. 3 sites allowed “full test” of JP concept Dayton, Los Angeles, and St. Paul: Positive effects Offered and sustained the full Jobs-Plus “package” Strong housing authority commitment ~ 3 of 4 working-age residents took advantage of services, rent incentives Infused development with self-sufficiency message Seattle: Positive short-term positive effects Early exit from demo because of reconstruction Baltimore and Chattanooga: No effects Incomplete implementation 12

13. 1998 cohort = focus of impact analysis All nondisabled working-age residents Living in Jobs-Plus or comparison developments in October 1998 (random assignment of developments) Includes recent arrivals and longer-termers in 1998 Includes those who moved or stayed after 1998 Long-term comparative interrupted-time series analysis, using Unemployment Insurance wage records Sample size: Program Comparison Total 2,123 2,651 4,774 13

14. Pooled average quarterly earnings for the 1998 cohort (full implementation sites) Figure pooled 1 Mean Quarterly Earnings for the 1998 Able-Bodied Sample: 3 sites pooled Post-program period

15. Pooled average quarterly earnings for the 1998 cohort (full implementation sites) Figure pooled 1 Mean Quarterly Earnings for the 1998 Able-Bodied Sample: 3 sites pooled Post-program period

16. Pooled average quarterly earnings for the 1998 cohort (full implementation sites) Figure pooled 1 Mean Quarterly Earnings for the 1998 Able-Bodied Sample: 3 sites pooled Post-program period

17. Earnings impacts through 7 years (full implementation sites) Site Avg. per year (2000 - 2006) Cumulative (2000 - 2006) Change (%) Pooled$1,300$9,09916% Dayton$984$6,88814% Los Angeles$1,176$8,23315% St. Paul$1,883$13,18119% All results statistically significant 17

18. Earnings impacts for subgroups 18

19. Earnings impacts by age group (during program) 19

20. Replication efforts New York City Now serving 3 large public housing developments A key feature of Mayor Bloomberg’s new “Young Men’s Initiative”: Will include up to 8 new Jobs-Plus sites San Antonio, Texas Operating on 1 large public housing development Obama administration (HUD) Proposing federal expansion in new budget 20

21. UK ERA UK Employment Retention and Advancement Demonstration Impacts by housing status Focus today on ND25+ group (largest effects)

22. 22 Evaluation sponsor UK Department for Work and Pensions (DWP) US Research Partner MDRC UK research partners Policy Studies Institute (PSI) Institute for Fiscal Studies (IFS) National Institute for Economic & Social Research (NIESR) Office for National Statistics (ONS) Lessons traveled in both directions! Transatlantic collaboration

23. 23 Unemployed and entering New Deal welfare-to-work program: (1) New Deal Lone Parents (NDLP) (2)New Deal 25-Plus (ND25+) Working part-time (lone parents): (3) Working Tax Credit (WTC) group UK ERA Target groups

24. 24 Operates within Jobcentre Plus centers Separate staff In some offices, separate post-employment teams 33-month program: If unemployed at intake, get job placement through New Deal w-t-w program (~ 9 months) Post-employment for all who work: 24+ months of “in-work” job coaching and support from “Advancement Support Advisors” 24 months of financial incentives UK ERA model

25. 25 Retention bonus  £ 400 3 times/yr for for sustained full-time work 2 yrs (£ 2,400 total) Tuition assistance if combine training + work  £ 1,000 Tuition bonus if complete training  up to £ 1,000 In-work emergency funds  £ 300/worker Financial incentives

26. 26 6 districts (Total N = 16,000 people): 4 districts in England – (E. Midlands, London, NE England, NW England) 1 in Scotland 1 in Wales Spread across 58 local offices (where random assignment took place) The pilot sites

27. ND25+: Long-term unemployed at baseline Random assignment New Deal intake ERA Control ERA New Deal Welfare-to-work + advancement focus Regular New Deal Welfare-to-work In-work support + incentives JOB Random assignment PRE- employment POST- employment 27

28. Months after Random Assignment % Employed Program group employment rates WTC 61 % 28

29. “Outcomes” vs. “impacts” 29 Target Group Outcome Ranking Sustained Impacts for Full Sample? WTC lone parentsHighestNo NDLPMediumNo ND25+LowestYes

30. 30 ND25+: Long-term unemployed Cumulative employment outcomes for ND25+ control group Ever worked in 5 years Months worked in 5 years ERA Control ERA Control 30

31. 31 ND25+: Long-term unemployed Impacts on employment outcomes Ever worked In 5 years + 1.1 months** (+ 8%) + 2.2 pp* (+ 4%) Months worked in 5 years ERA Control ERA Control

32. Tax Year Control group ND25+: Long-term unemployed Impacts on earnings trends Earnings (£) 32

33. Tax Year Control group ND25+: Long-term unemployed Impacts on earnings trends Earnings (£) ERA group 5-year cumulative impact: £1,814 ** (+ 11%) ERA ended for all by October 2007 (earlier for some) 33

34. ND25+: Long-term unemployed Impacts on ever worked within 5 years after random assignment Family housingSocial housingPrivate housing ERA Control ERA Control ERA Control 34

35. 35 ND25+: Long-term unemployed Impacts on ever worked within 5 years after random assignment Family housingSocial housingPrivate housing - 1.9 ERA Control ERA Control ERA Control ††

36. 36 ND25+: Long-term unemployed Impacts on cumulative 5-year earnings, by housing status at baseline Family housingSocial housingPrivate housing ERA Control ERA Control ERA Control

37. 37 ND25+: Long-term unemployed Impacts on cumulative 5-year earnings, by housing status at baseline Family housingSocial housingPrivate housing ERA Control ERA Control ERA Control £703

38. New York City’s Work Rewards Demonstration Target group: Low-income households with “Housing Choice Vouchers” Alternative to public housing Rent subsidies to help families pay rent in private housing market Tenants pay 30% of their income for rent Many don’t work, or are underemployed 38

39. Evaluation goals Testing 2 interventions: Do the interventions improve economic security? 1.Family Self Sufficiency Program (FSS) Case management “Escrow savings accounts” (asset-building) 5-year program 2.FSS+ Workforce Incentives Incentives adapted from UK ERA Cash rewards for full-time work: $150/month Cash rewards for education/training completion Rewards available for 2 years 39

40. Evaluation design: 3-way random assignment Voucher- holders sign up Random Assignment FSS-only FSS + incentives Control group Recruited volunteers from housing roster N = 1,455 households; 1,603 adults 5 years of follow-up (ongoing) 40

41. Cumulative impacts on earnings within 2.5 years after entering program FSS-OnlyFSS + Incentives Impact ($) Change (%) Impact ($) Change (%) Full sample503 2.8 5513.1 41

42. Cumulative impacts on earnings within 2.5 years after entering program FSS-OnlyFSS + Incentives Impact ($) Change (%) Impact ($) Change (%) Full sample503 2.8 5513.1 Employment subgroups Not working at baseline1,658 24.33,102**45.4 Working at baseline-706-2.4-1,904-6.5 †† 42

43. Subgroup earnings trends 43 Note: Earnings include $0 for non-workers Subgroup: Not working at baseline Control group

44. Subgroup earnings trends 44 FSS + Incentives FSS-Only Note: Earnings include $0 for non-workers Subgroup: Not working at baseline Incentives end Control group

45. New federal study of FSS MDRC launching test of national Family Self- Sufficiency Program (FSS) in 15+ cities Opportunity to test whether subgroup patterns hold elsewhere. But no special workforce incentives (unless we can raise other funding) 45

46. Earlier US welfare-to-work studies Studies of mandatory welfare-to-work programs for lone parents in the US (from the 1990s, early 2000s) looked at housing subgroups General pattern: Effects of welfare-to-work programs were larger for welfare recipients living in public housing or with rent vouchers than for other welfare recipients without housing subsidies. 46

47. Final reflections Growing evidence: Employment interventions can work for public/social housing and other rent-assisted groups – Implement UK ERA for long-term unemployed in social housing? Encouragement for linking housing and work policies— and, hence, for inter-agency collaboration Strategies that combine services + incentives may work best Place-based interventions, such as Jobs-Plus, offer multiple avenues for engaging families However, strategies to promote work advancement remain elusive and call for more innovation and testing 47

48. Final reflections (continued) Building further evidence for policy US: – Longer-term impact testing of services, incentives, and rent reform for subsidized tenants – Study replication of Jobs-Plus in public housing UK: – Confirm impact of ERA for long-term unemployed in social housing under new welfare reform; test a refined model? – Test a place-based work intervention (like Jobs-Plus)? Both: – Continue search for effective advancement strategies 48