The Effects of Housing Assistance on Labor Supply

Evidence from a Voucher Lottery

Brian A. Jacob, Jens Ludwig

Budget line shifting up?

  • Standard Static Labor Supply Model

    • In-kind / Cash transfer?

    • How it constrains consumption?

    • Complement or substitute to leisure?

      (Murray, 1980; Leonesio, 1988; Shroder, 2002; Moffitt, 2002)

    • Maintain min. level of consumption?

      (Schone, 1992)

Increase labor supply?

  • Reduce the price of complements to work?

    • Residential stability

    • Neighborhood quality

    • Proximity to job

Selection-bias in prior studies

  • Welfare to Work (Gubits et al., 2006)

    • TANF recipients only (Temporary Assistance for Needy Families)

    • Eligibility determined by local housing agencies

      • “critical to their labor market success”

Lottery design

  • Chicago Housing Authority Corporation, Inc. (CHAC) 1997

  • Eligibility: Income (~$14,000), Family size

  • Random positions on the waiting list

  • Subsidy:

    Fair Market Rent — Family’s contribution(30% adj. income)

  • Max. FRM: 50th percentile of local private market rent (~$12,000)

Ambiguous prediction from theories

Lottery results

  • 82,607 applied, 18,110 received by 2003

  • Top 35,000 were notified the number and told would receive in 3yr

  • Sample: 90 % applicants living in private-market housing

    • Treatment: received voucher (1~18,110)

    • Control: told would not get (35,000~) 30,662

    • Exclusion: expecting to get but not getting (18,110~35,000)

  • Low take-up: half of the treatment did not use it

Subsidy size

  • Baseline: Income ~$14,000 Rent ~$8,160

  • ~Subsidy $8,265 = FMR $12,000 — Contribution $3,735

    • Out-of-pocket rent dropped to half (free cash?)

Data

  • Pre-lottery

    • CHAC application (household head)

    • Illinois Dept of Human Service (other family members)

  • Voucher utilization

    • Dept of Housing and Urban development (HUD)

Outcome measures

  • Labor participation

    • Illinois Unemployment Insurance (quarterly earning)
  • Social assistance

    • Illinois Dept of Human Service (TANF, AFDC, Foodstamp, Medicaid)
  • Residential location

    • National Change of Address, National credit bureau check (10% random sample of total applicant)

Effect of receiving an offer

  • Intend To Treat effect (50% compliers)

\[ y_{it} = \alpha + \beta_1(PostOffer_{it}) + \beta_2(PreOffer_{it}) + X\Gamma + \gamma_t + \epsilon_{it}. \]

    • Landlords were not willing to deal with paperwork

    • Limited time to lease an unit

Effect of using a voucher

  • Assume vouchers do not affect who choose not to take it

    • Two stage least square

\[ Leased_{it} = \alpha + \theta_1(PostOffer_{it}) + \theta_2(PreOffer_{it}) + X\Gamma + \gamma_t + \epsilon_{it}. \]

\[ y_{it} = \alpha + \pi_1(Leased_{it}) + X\Gamma + \gamma_t + \epsilon_{it}. \]

    • Include those who used it then gave it up (out of eligibility)

    • Include other voucher allocations (8% of control HHs)

Offers induced voucher utilization

  • Disadvantaged HHs are more likely to lease up

    • Female, unemployed, TANF recipients

Vouchers reduced labor participation

  • Larger impacts among

    • Women (lower baseline employment among men)

    • Non-social benefits recipients (full marginal tax rate)

Changes in effects over time

  • Before and after receiving an offer

\[ y_{it} = \alpha + \delta_k(\sum_{k} D_{it}^k) + X\Gamma + \gamma_t + \epsilon_{it}. \]

    • Time-sensitive behavioral responses

    • Fraction of HHs that have leased up

    • Changes in the economy condition

Impacts grew larger over time

  • Reject “disruption” account

  • Anticipation effect increased labor participation

    • after being on the waiting list, before receiving an offer

Residential stability & quality

  • Working more in a nicer place? Maybe not

Wait till the offer

  • Offer changed the timing of moves, not the likelihood of moves

Conclusion

  • Not work enhancing

  • Increase in pre-voucher labor supply

    • Consistent with life-cycle labor supply model

    • Work more when the expected wage rate is high

  • Defer moving until receiving offers

    • Mobility and living quality unchanged

Policy implication

  • Generaliable to other in-kind transfers

  • Leaky bucket

    • Deadweight loss due to reduction in labor supply ~$660
  • Perception of gains or losses? (Kahneman & Tversky, 1979)

  • Anticipatory utility? (Loewenstein, 1985)