Transcript John Kirlin, Assistant Deputy Director for SNAP Research, Economic

Joint Contributions of SNAP and
Unemployment Insurance
to the Social Safety Net:
A Data Linkage Project
John A. Kirlin and Michael Wiseman
September 20, 2010
Presented at APDU 2010 Annual Conference
“Public Data 2010: Opportunities and Challenges for the New Decade”
Outline
• Background
• Study and research teams
• What is the policy issue?
• Study rationale
• Hypotheses and data needs
• Planned analyses and schedule
Background
• SNAP (formerly food stamps) is by far the
largest food and nutrition assistance
program in US
– Over 41 million participants in June 2010
– Over $5.5 billion in monthly benefits
• 12 months before “only” 34.9 million
participants receiving $4.7 billion
Background
• Despite its popularity, only 2 of every 3
eligible individuals participates (2008)
• After years of research on SNAP/food
stamp participation, we still do not
understand all that we would like, e.g.:
– Why don’t all eligibles participate?
– How do people end up on SNAP? How do
they leave?
Background
• In addition to the SNAP questions, what
happens to individuals after their UI
benefits end?
– Some find work
– Some retire (early retirement rising)
– Some go on public assistance
Background
• We are interested in the “some go on
public assistance.” It’s a big unknown.
– How many?
– Who?
– For how long?
• Need longitudinal UI data matched to
longitudinal SNAP data
The SNAP-UI Study
Joint Contributions of SNAP and
Unemployment Insurance
to the Social Safety Net:
A Data Linkage Project
Research Teams – States
1) Jacob France Institute – Maryland
• Andrew Young School – Georgia
• Chapin Hall Center – Illinois
• Upjohn Institute – Michigan
• Ray Marshall Center – Texas
2) California Institute of Public Policy –
California
3) University of Missouri – Florida
Seven States
Policy Issue
• Economy is in a deep and protracted recession
• Existing assistance programs (UI, SNAP, TANF)
have responded in different ways
• How can such programs improve their response
to economic hardship?
• Can information about individuals’ use of UI and
SNAP suggest ways to better serve unmet
needs?
The Economy
The worst recession since WWII
Real and Potential Gross Domestic Product (GDP), 1949-2010
16
Trillions of Chained 2009 Dollars
14
12
10
8
6
Current Recession
Loss in GDP =
$2,053.33 billion
2009 dollars
4
2
0
1949 1953 1957 1961 1965 1969 1973 1977 1981 1985 1989 1993 1997 2001 2005 2009
Source: Federal Reserve Bank of St. Louis
Potential GDP
Real GDP
Unemployment has risen dramatically
Unemployment Rate in the United States, Jan. 1976 - Aug. 2010
(seasonally adjusted)
12%
10%
8%
6%
4%
2%
Source: U.S. Bureau of Labor Statistics
Shaded areas represent peak-to-trough NBER recessions.
2010
2008
2006
2004
2002
2000
1998
1996
1994
1992
1990
1988
1986
1984
1982
1980
1978
1976
0%
The deficit of available jobs has increased
drastically since early 2007
Unemployed Persons Seeking Full-Time Work and Open Non-Farm Positions,
Jan. 2001 - Jun. 2010 (seasonally adjusted)
16
14
12
Millions
10
8
Jan. 2001:
Surplus of
252 thousand
jobs
Oct. 2009:
Deficit of
11.3 million
jobs
Jun. 2002:
Deficit of
4.4 million
jobs
Mar. 2007:
Deficit of 710
thousand jobs
6
4
2
0
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Unemployed Persons, Looking for Full-Time Work
Source: U.S. Bureau of Labor Statistics
Open Non-Farm Positions
The SNAP caseload has grown in response
to this recession (and every past recession)
Number of SNAP and UI Claims, SNAP Claims as a Percentage of Population,
and UI Claims as a Percentage of Labor Force, 1980-2009
40
12%
35
10%
30
8%
Millions
25
20
6%
15
4%
10
2%
5
0
0%
1980
1982
1984 1986
1988
1990
1992
1994 1996
1998 2000
2002
2004
2006
2008
Average Monthly SNAP Participants (left axis)
Number of Claims for Any UI Program (left axis)
Percentage of U.S. Population Receiving SNAP (right axis)
Percent of Labor Force Collecting UI Benefits (right axis)
Sources: National Bureau of Economic Research, USDA Food and Nutrition Service, U.S. Census Bureau, U.S.
Department of Labor.
Shaded areas represent peak-to-trough NBER recessions.
But relatively few SNAP households collect
UI and SNAP at the same time
SNAP Households with Unemployment Income,
1997 - 2008
7%
12
6%
10
5%
8
4%
6
3%
4
2%
2
1%
0
0%
Millions
14
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
Total Number of SNAP Households (left axis)
Sources: USDA Food and Nutrition Service and
U.S. Bureau of Labor Statistics
Percentage of SNAP Households with UI (right axis)
National Unemployment Rate (right axis)
Rationale for the Study
• Historically, the relationship between
SNAP and UI has been tenuous, but
– Do households sequence their UI and SNAP
receipt?
– Given the severity of the current recession, is
the relationship changing?
– Most important part of the relationship is
probably dynamic, not cross-sectional
Rationale for the Study
• We do not know about the dynamic
interaction between the two programs
• This project intends to go to the state level
to understand the past and evolving
relationship between SNAP and UI, in
order to improve SNAP administration and
access
Possible Scenarios:
No UI Benefits
Earnings → SNAP → Earnings
0
SNAP Benefits
Covered Wages
Possible Scenarios:
Non-overlapping Spells
Earnings → UI → SNAP → Earnings
0
SNAP Benefits
UI Benefits
Covered Wages
Possible Scenarios:
Overlapping Spells
Earnings →UI → UI/SNAP → SNAP
0%
SNAP Benefits
UI Benefits
Covered Wages
Why the SNAP/UI relation may be
changing
• Worst recession since Great Depression
• A large proportion of job losses in this recession
has been a result of layoffs, making local reemployment more difficult
• Changed eligibility rules--broad-based
categorical eligibility for SNAP essentially
eliminates asset limits for many potential
applicants
Research Hypotheses
1) Both the concurrent and sequential links
between UI and SNAP grow during recessions.
2) Over time, the likelihood of taking up SNAP in
conjunction with UI has increased, and the lag
between initiation of UI and SNAP take-up has
declined.
3) There is a large group of people who are
collecting UI and may be eligible for SNAP, but
who are not receiving SNAP benefits.
Data Needs
• Need access to longitudinal and matched
data on SNAP and UI participation to test
previous hypotheses
• Available data are limited
– Administrative data for SNAP and UI are
maintained at the state level; the federal
government only has access to a crosssectional sample of the SNAP data
Solution
• Turn to researchers who:
– Have access to state data
– Have experience processing and analyzing
large data sets
– Are knowledgeable of the issues
7 States and 7 Research Teams
These 7 states face diverse economic
situations
Unemployment Rates in the Seven Project States, 1976-2010
(seasonally adjusted)
17%
16%
15%
14%
13%
12%
11%
10%
9%
8%
7%
6%
5%
4%
3%
2%
1%
0%
California
Florida
Georgia
Illinois
Maryland
Michigan
Texas
Sources: U.S. Bureau of Labor Statistics and the National Bureau of
Economic Research.
Shaded areas represent peak-to-trough NBER recessions.
2010
2008
2006
2004
2002
2000
1998
1996
1994
1992
1990
1988
1986
1984
1982
1980
1978
1976
U.S.
Three Levels of Planned Analysis
• Level 1 covers the foundational question of the
project: What is the intersection of SNAP and
UI?
– Divided into two parts:
• Cross-sectional over 2006 - 2010
• New SNAP entrant cohorts over 2006-2010
– Each looking backward and forward at UI
wage coverage and claims receipt
Level 2 Analyses
• Level 2 repeats the Level 1 analyses for
specific subgroups of the SNAP population
– Individuals (age, gender, citizenship,
race/ethnicity, work/earnings history)
– Households (size, type, composition, spell
length, metro/rural, income sources, and
gross and net income amounts)
Level 3 Analyses
• Level 3 offers opportunities for researchers to
address issues that are particular to:
– Their unique data strengths
– The needs of the SNAP or UI programs in
their state
Challenges
• Getting permission from multiple state agencies
to have their data matched and used by outside
researchers
– Our teams have prior experience working with their
states so a level of trust already exists
• Getting consistent data for similar time periods
from all states
• Using consistent definitions, especially given:
– Weekly, monthly, quarterly time periods for data
Study Schedule
• Finish data prep by late fall
• Analyses over winter
• Individual reports to ERS in the spring/summer
• ERS summary report thereafter
Joint Contributions of SNAP and Unemployment
Insurance to the Social Safety Net:
A Data Linkage Project
Thank you