The Effects of Welfare and IDA Program Rules on Asset Holdings

1
The Effects of Welfare and IDA Program Rules on
Asset Holdings

• Signe-Mary McKernan and Caroline Ratcliffe
• The Urban Institute
• Yunju Nam
• Washington University in Saint Louis
• UKCPR Welfare Reform Conference
• April 12, 2007
• Please do not quote any information from this
  presentation.
• All findings are preliminary and not for
  attribution.

2
Introduction

• Savings and assets can
• Cushion low-income families against sudden income
  losses and bolster long-term economic gains.
• Make a family ineligible for benefits from
  means-tested programs.
• Asset restrictions may have the unintended
  consequence of discouraging low-income families
  from saving.
• Federal and state governments have implemented
  programs and program rules to better promote
  savings.
• TANF and FSP asset restrictions relaxed.
• IDA programs supported.
• Few studies have examined the effect of program
  rules on saving and asset accumulation and
  existing research shows mixed results.

3
Research Questions

• This paper measures the effect of state program
  rules on asset holdings.
• AFDC/TANF, Food Stamps, Individual Development
  Accounts (IDA), EITC, and minimum wage
• What are the effects of specific TANF, Food
  Stamp, IDA, EITC, and minimum wage program rules
  on 
• Liquid asset holdings?
• Vehicle asset holdings?
• Net worth?

4
Background on State Program Rules

• AFDC/TANF
• Federal waivers to states and 1996 welfare reform
  (PRWORA).
• Increased limits on vehicle assets.
• Increased limits on liquid assets in unrestricted
  accounts.
• Created restricted account programs.
• Food Stamps
• Some steps to liberalize asset limits in 1990s,
  but key steps in 2000 through 2002.
• Increased vehicle asset limits expanded
  categorical eligibility.
• Individual Development Accounts (IDAs)
• Federal and state governments began adopting IDA
  programs during the 1990s. Matched saving
  accounts created to encourage asset accumulation.
  
• IDA programs differ with regard to match rates,
  maximum s available for match, eligible
  population (e.g., welfare recipients).

5
Literature

• Four studies examine the effect of AFDC/TANF
  asset limits on asset holdings and find mixed
  results.
• Studies focus on AFDC/TANF unrestricted
  (countable) asset limits and vehicle asset
  limits.
• Finding
• Liquid asset holdings and net worth
• Two studies find that relaxing AFDC/TANF program
  rules did not increase liquid assets or net worth
  (Hurst Ziliak 2006, Sullivan 2006).
• Two studies find that relaxing these program
  rules did increase liquid assets (Nam Kam 2006)
  or net worth (Powers 1998).
• Vehicle ownership and equity
• Two studies find that relaxing asset limits
  increased vehicle ownership (Hurst Ziliak 2006,
  Sullivan 2006), while a third study finds no
  effect of relaxed asset limits on vehicle
  ownership (Nam Kam 2006).

6
Literature

• Analyses of IDA account monitoring data suggest
  people in IDA programs save.
• Analyses of the American Dream Demonstration
  (ADD) program (designed to promote saving through
  IDAs) suggest that the program increased
  participants savings (Stegman Faris 2005).
• Mills et al. (2006) controls for selection into
  IDA program with a controlled field experiment.
• IDA program was found to
• Raise homeownership rates among blacks.
• Reduce financial assets and business ownership
  among blacks (possibly indicating the need to
  liquidate assets to afford down payments and
  housing transition costs).
• Raise business equity among whites.
• Have no statistically significant effect on net
  worth.

7
Contributions

• Examine comprehensive set of 16 program rules
  hypothesized to affect asset holding.
• AFDC/TANF, Food Stamps, Individual Development
  Accounts, earned income tax credit, and minimum
  wage
• Examine both unrestricted and restricted (IDA and
  new TANF) accounts.
• We find differences between these accounts.

8
Sample and Data

• Sample
• Low-education single mothers and low-education
  families
• Low-education High school degree only or less
• Age Working age, 18-54
• SIPP Data 1990, 1992, 1993, 1996, and 2001
  panels
• Data from 1991 through 2003
• Captures period of significant changes to program
  rules and both weak and strong economy.
• Asset data collected in SIPP topical modules (2-4
  times per panel)
• Annual data
• 10,487 low-education single mother person-years
• 53,215 low-education family person-years
• Supplemented with policy data and controls for
  the economy
• Variety of sources including Welfare Rules
  Database, FNS waiver database, Center for Social
  Development IDA data, etc.

9
Asset Measures

• Liquid asset holdings
• Have liquid asset holdings
• Value of liquid asset holdings
• Vehicle asset holdings
• Own a vehicle
• Vehicle equity
• Net worth
• Net worth including home value
• Net worth excluding home value

10
Asset Holdings
11
Sample Characteristics
12
State Program Rules
13
Number of States with Rule or Mean Value by Year
14
Asset Measures and Controls

• Asset Measures
• Liquid asset holdings
• Have liquid asset holdings
• Ln(value of liquid asset holdings)
• Vehicle asset holdings
• Own a vehicle
• Ln(vehicle equity)
• Net worth
• Net worth including home value
• Net worth excluding home value
• Demographic and Economic Controls
• Age, age squared, black, Hispanic, education less
  than high school, number of children in family,
  number of adults in family, metro area
• State unemployment rate, state per-capita income,
  state employment-population ratio

15
Empirical ModelHave Liquid Asset Holdings

• Linear Probability Model
• Yist 0/1 have liquid asset holdings
• WPst State AFDC/TANF policies
• FSPst State FSP policies
• IDAst State IDA policies
• EITCst State EITC policies
• MWst State minimum wage policies
• Xist Family characteristics
• Sst State economic conditions
• ?s State fixed effects
• ?t Year fixed effects
• ?ist Error term

16
Empirical Model

• Other outcomes
• Vehicle ownership linear probability model
• Ln(vehicle equity) and Ln(value of liquid assets)
  tobit model
• Net worth ordinary least squares
• All models use same set of explanatory variables,
  state and year fixed effects, and are weighted.

17
Outline of Results

• Liquid asset holdings
• Vehicle ownership
• Net worth

18
Liquid Asset Holding Results, 1991-2003AFDC/TANF
and FSP Rules

• Percentage holding liquid assets Mothers 31,
  Families 56

19
Liquid Asset Holding Results, 1991-2003IDA,
EITC, and Minimum Wage
20
Liquid Asset Holding Results, 1991-2003Demograph
ic and Economic Controls
21
Vehicle Asset Holding Results, 1991-2003
AFDC/TANF and FSP Rules
22
Vehicle Asset Holding Results, 1991-2003 IDA,
EITC, and Minimum Wage
23
Net Worth Results, 1991-2003 AFDC/TANF and FSP
Rules
24
Net Worth Results, 1991-2003 IDA, EITC, and
Minimum Wage
25
Summary of Results

• Overall, not many program rules are statistically
  significant. However, key findings hold in
  alternate specifications.
• State AFDC/TANF Program Rules
• More generous restricted asset account limits are
  associated with increased liquid asset holdings
• No evidence that more generous unrestricted asset
  account limits or AFDC/TANF vehicle exemptions
  are associated with increased asset holdings
• State FSP Rules
• More generous food stamp eligibility
  rulesexempting at least one vehicle and expanded
  categorical eligibilityare associated with
  increased vehicle ownership
• Expanded categorical eligibility is associated
  with increased net worth

26
Summary of Results (contd)

• IDA Program Rules
• More generous IDA rules affect asset holdings.
• Increases in the maximum match rate are
  associated with increases in vehicle ownership.
• Increases in the maximum amount qualified for
  match are associated with increases in liquid
  asset holdings.
• EITC/Minimum Wage
• State EITC program rules have a mixed
  relationship to asset holdings (EITC amount
  decreases, EITC refundable increases).
• State minimum wage is associated with increases
  in vehicle assets.

27
Conclusions

• Results suggest that some state program rules
  especially those aimed at asset building could
  affect low-education single mothers and
  families asset holdings.
• Not all state asset-related program rules have
  the same effect restrictions on withdrawals and
  incentives built into restricted asset account
  limits and IDA programs may provide better
  motivation to build assets.
• Non-asset related program rules (e.g., FSP
  categorical eligibility and minimum wage) may
  affect asset holding, suggesting that potential
  program interactions and indirect effects of
  program rules on non-target populations are
  potentially important.