ReportsEconomics

The impact of homelessness prevention programs on homelessness

See allHide authors and affiliations

Science  12 Aug 2016:
Vol. 353, Issue 6300, pp. 694-699
DOI: 10.1126/science.aag0833
  • Fig. 1 Fund availability rate by week.

    Shown is the fraction of eligible callers in our homogenous sample that were referred to financial assistance each week from 2010 to 2012. The homogeneous sample includes all eligible callers who were seeking rent assistance with need amounts between $301 and $900, were nonveterans, were not receiving housing subsidies, were not requesting more than 1 month’s rent, had a social security number, had a family income below twice the poverty line, and were not homeowners (n = 1431).

  • Fig. 2 ITT effects of fund availability on shelter admittance 1 to 12 months after the call.

    We estimated 12 separate regressions of shelter entry 1 to 12 months (mo) after the call on an availability-of-funds indicator, demographic characteristics of the caller, ZIP code–level characteristics, fund-specific restriction variables, and seasonal control variables. The solid line plots the coefficients on the availability-of-funds indicator in these regressions (e.g., −0.015 denotes a 1.5-percentage-point decline in shelter entry); estimates of the coefficients on other covariates are given in table S8. The dashed lines denote the 95% confidence interval (CI) calculated using standard errors clustered at the ZIP code level. The estimates for the effect 1 to 6 months after the call are based on the main sample (n = 4448). Beyond 6 months, we lose some information because we do not have data about shelter entry beyond 6 months for all callers. Thus, the sample size decreases with each month beyond 6 months, with the 12-month shelter entry estimates based on a sample of 3892 callers.

  • Table 1 Mean characteristics and differences in means between treatment and control groups.

    Shown are the mean characteristics for the control groups and the regression-adjusted difference between groups. To test balance, each characteristic was regressed on an availability-of-funds indicator, fund-specific restriction variables, and seasonal control variables, and standard errors were clustered at the ZIP code level. Caller characteristics come from HPCC records from 20 January 2010 to 4 December 2012 for eligible, first-time callers; ZIP code–level characteristics come from the 2010–2012 American Community Surveys and are standardized to have a mean of 0 and a variance of 1. Shelter inhabitancy in the past 18 months comes from HMIS data for homeless shelters in Chicago.

    Dependent variableHomogeneous subsampleMain sample
    Control group meanCoefficient on availability of fundsControl group meanCoefficient on availability of funds
    Female0.874–0.100**0.794–0.032*
    White, non-Hispanic0.0660.0220.0620.011
    Black, non-Hispanic0.906–0.0300.908–0.017
    Other, non-Hispanic0.0330.0020.0320.005
    Hispanic0.0830.0010.069–0.001
    Age39.258–1.07839.129–0.838*
    Number of adults in caller's household1.3410.0381.353–0.019
    Number of minors in caller's household1.357–0.0121.389–0.083
    Standardized percentage in ZIP code with high school degree–0.0610.0010.014–0.034
    Standardized labor force participation rate in ZIP code–0.0600.032–0.015–0.003
    Standardized unemployment rate in ZIP code0.0170.0050.0140.005
    Standardized median age in ZIP code–0.1630.150**–0.0060.018
    Standardized monthly housing cost in ZIP code (in thousands)0.034–0.0800.009–0.040
    Standardized median household income in ZIP code (in thousands)–0.0710.026–0.004–0.021
    Standardized fraction black in ZIP code0.046–0.0480.024–0.012
    Standardized fraction white in ZIP code–0.0420.027–0.014–0.010
    Standardized fraction other races in ZIP code–0.0400.076–0.0370.051
    Applying because of benefit loss0.1320.0230.115–0.010
    Applying because of inability to pay bills0.022–0.0150.013–0.008**
    Applying because exiting shared housing0.0440.0130.0890.012
    Applying to flee abuse0.016–0.0120.016–0.002
    Applying because of job loss0.407–0.0300.3010.012
    Monthly income (in thousands)1.1750.080**1.210–0.052**
    Receiving Supplemental Nutrition Assistance Program benefits0.637–0.0570.633–0.019
    Receiving child support0.061–0.0100.060–0.017**
    Receiving earned income0.6820.0080.650–0.004
    Receiving disability payments0.084–0.0260.096–0.013
    Receiving social security income0.106–0.0030.1370.000
    Receiving income from Temporary Assistance for Needy Families0.039–0.0010.0610.003
    Receiving unemployment payments0.2010.0290.1510.023
    Receiving other income sources0.067–0.0230.073–0.005
    Living situation: rent housing0.907–0.0100.818–0.015
    Living situation: shared housing0.0930.0100.1740.015
    Shelter inhabitancy in the past 18 months0.0270.0030.0290.008
    n182143118584448

    *P < 0.10; **P < 0.05; two-tailed t test of the difference between treatment and control groups.

    Supplementary Materials

    Navigate This Article