Stephen A. WandnerSenior EconomistU.S. Department of Labor

Michael WisemanResearch Professor George Washington University

What can the European Social Fund learn from the American experience with operation of training programs under the Workforce Investment Act?

This question has many facets WIA attempts to increase effectiveness and

efficiency through performance “bonuses” This paper (in progress!) expands the

perspective to include the TANF and SNAP High Performance Bonuses

Federal-state program, operated by the states through local “one-stop career centers”

Supported by formula-based grants to states

Beyond the formula allocations, “High Performance Bonuses” are paid states on the basis of measures of achievement

Bonuses are small and, over time, have gotten smaller

Performance targets are set through negotiations between the (6) federal government regional offices and the states

Final targets depend on negotiation techniques and skills of regions and states

Performance targets cover many program outcomes

States have some control over measurement implementation

Regional DOL office strategies, capabilities, and enthusiasm differ

States vary in analytical capabilities for response to regional targeting proposals.

Measures adopted offer numerous opportunities for “gaming” by states

Results have been uneven, regionally and by state

Connection of HPB accomplishment to actual achievement challenged by research

▶ Negotiation may be useful, but attention must be paid to leveling playing field

▶ Should start with control for variation in characteristics of clients served, economic environment

▶ Addressing problems of motivation essentialBottom line: No evidence of significant

positive effects Watch developments with reauthorization

TANF famously replaced Aid to Families with Dependent Children beginning in 2007

Enabling legislation called for “high performance”

Employment targets—job entry, job retention, earnings gain—have “face validity”

Goals multiplied over time, as did winners Like WIA, stakes were small Program died, unmourned, in 2005

Began with data available to states, but it was clear procedures were not uniform and states lacked access to some data

Introduced a new resource, the “National Directory of New Hires”

Ultimately major measures wholly computed at federal level, with substantial lag between reference “performance year” and award

Federal computation not always intelligible, reliable

What to measureHow to measureControl for contextStrategic responseMissing feedback

Give thought to the objectivesBe cautious about statistical inference

Plan for improvement Institutional development may be an important by-product of performance assessment effort

SNAP is a national negative income tax operated outside of the tax system. Does not purport to be adequate for minimum subsistence

Delivered by electronic benefits transfer (EBT) and collected when recipients purchase food

Arguably the nation’s most important means-tested benefit

Plays a significant role in economic stimulus

SNAP is operated by states Benefits are wholly federally funded;

administration costs are split between states and federal government

Incentive problems addressed by sample-based quality control system

States liable for cost of errors, but attempts made to reduce emphasis on penalty and shift to rewards

Result (2002) was --

Based on QC audit, other sources Four bonus categories:o Payment accuracyoNegative error rateoApplication processing timelinesso Program access

Only $48 million (total state administrative costs were about $3 billion in FY 2007)

Awards delivered by September of following year

Category Definition

State Average

(Un-weighted)

Awards Made Best State

Best State Score

Total Awards

($Millions)

Payment Accuracy Sum of erroneous under- and over-payments as proportion of total benefits (%)

5.0% 8 Florida 0.8% $7.2

Payment Accuracy Improvement

Change in Payment Accuracy Measure, FY 2007-FY 2008 (Δ %)

3 Georgia -5.6% $4.1 "*"

Negative Error Rate Proportion of applications or cases denied, suspended, or terminated in error

11.0% 4 Nebraska 0.0% $0.7

Negative Error Rate Improvement

Change in Negative Error Measure, FY 2007-FY 2008 (Δ %; negative identifies error decline)

0.02% 2 Oklahoma -6.5% $2.3

Proportion of approved applicants given benefit access within target time (30 days for normal cases, 7 days for cases qualified for expedited processing)

87.8% 6 Montana 98.0% $0.3 $6.0

Program Access Ratio of average monthly number of SNAP participants over calendar year to number of persons in families with incomes less that 125 percent of the federal poverty standard (%)

58.6% 4 Missouri 90.0% $2.6

Program Access Improvement

Change in Program Access measure, CY 2007-CY2008 (Δ %)

3.8% 4 Maryland 10.0% $1.4

Total $48.0*

Application Processing Timeliness

State Award

($Millions)

Source: FNS; definitions paraphrased.Georgia won awards in both level and improvement categories

Food Stamp Program High Performance Bonuses, FY 2008

$24.0

$6.0

$12.0

Assessing sample-based penalties

Program access measuresChange versus levelsTechnical assistanceNet effect

Link to ground-level operations Audit the outcomes Take care with statistics Link to better practice

But don’t get carried away: Task is relatively simple: Deliver a well-

defined benefit to a target population each month

Outcome immediate Broad political support

Take in the museumsKeep watchGet back to the “Open Method”Start on the groundReward process