financial incentives and welfare reform in the united states

Financial Incentives and Welfare Reform in the United States / 129

Philip K. RobinsCharles MichalopoulosElsie Pan

Financial Incentives andWelfare Reform in theUnited States

Journal of Policy Analysis and Management, Vol. 20, No. 1, 129–149 (2001)© 2001 by the Association for Public Policy Analysis and ManagementPublished by John Wiley & Sons, Inc.

Abstract

This paper uses a microsimulation model to ask whether welfare recipients in theUnited States would work full-time if offered an earnings supplement that was con-ditioned on full-time employment. The simulations suggest that the earnings supple-ment would increase full-time employment, with little additional cash transfer costto the government. In contrast, financial incentives currently being used by many ofthe states are increasing employment and income, but are encouraging primarilypart-time employment. Encouraging full-time employment is particularly impor-tant in light of new time limits on welfare receipt. Faced with a loss of welfarebenefits, many recipients may find that part-time earnings do not allow them to beeconomically self-sufficient. © 2001 by the Association of Public Policy Analysisand Management.

INTRODUCTION

On August 22, 1996, President Clinton signed the Personal Responsibility and WorkOpportunity Reconciliation Act (PRWORA), which radically altered the structure ofthe welfare system in the United States. Among other things, PRWORA replaced theAid to Families with Dependent Children (AFDC) program, a federal entitlement,with the Temporary Aid to Needy Families (TANF) program, a system of block grantsto states.

One of the primary goals of TANF is to move welfare recipients into work and economicself-sufficiency. While states are given much flexibility in how to achieve this goal, thefederal government imposes some guidelines in the form of work participationrequirements and time limits on length of welfare receipt. The focus of this paper is onalternative financial incentive schemes that are being used, or could be used, to helpstates meet the work participation requirements. In particular, this paper reports theeffects of using a microsimulation model to investigate whether an earnings supplementconditioned on full-time work would encourage more people to work than the “enhancedearnings disregards” currently being used or tested by many states. The proposedincentive scheme, based on a full-time work requirement, is patterned after anexperimental program being tested on a group of welfare recipients in Canada.

Manuscript received February 4, 2000; revise and resubmit recommended May 8, 2000; revised June 28, 2000; acceptedAugust 10, 2000.

130 / Financial Incentives and Welfare Reform in the United States

BACKGROUND

The federal PRWORA legislation stipulates that 25 percent of the single-parentcaseload in a particular state be participating in work activities by fiscal year1997. The minimum participation requirement has been increasing by 5 percenteach year until FY 2002, when it will reach 50 percent. States that fail to meet thework participation requirements will not receive the full value of the federal TANFblock grant.

The federal legislation defines an “allowable work activity” as unsubsidizedemployment, subsidized private sector or public sector employment, on-the-jobtraining, job-search assistance for as long as six weeks, community serviceprograms, vocational education training for as long as one year, and educationfor persons who have not yet completed high school. The legislation emphasizeswork activities and places caps on the number of persons who can be placed ineducational activities. Reducing the caseload can also count toward theparticipation requirement.

States have considerable latitude in penalizing household heads who fail tocomply with the work activity requirements. The state, at its own discretion, canreduce and eventually terminate benefits. States can exempt certain people fromthe requirements, such as single parents of young children, but must meet federalrequirements for the percentage of their caseload participating in work activities.

The work requirement provisions of PRWORA make it crucial for states to findeffective ways of moving welfare recipients into work. Many studies have shownthat a significant portion of the caseload spends more than 60 months receivingbenefits (the maximum time limit specified under PRWORA, although many stateshave opted for shorter time limits). Bane and Ellwood (1994), for example, estimatethat the median length of total welfare receipt (not necessarily a continuous spell)is about 48 months. Pavetti (1995) estimates that 76 percent of the welfare caseloadat any point in time (which is dominated by long-term recipients) will eventuallyreceive welfare for at least 60 months. Pavetti finds that among those who receivedwelfare for 60 months or more, 63 percent lacked a high school diploma (or GED)at the time they started collecting welfare, 39 percent had no work experience, 53percent were under 25 years of age, 58 percent had never been married, and 52percent had a child under the age of one year. Clearly, in the absence of effectiveactions by the states, many individuals are likely to be in financial despair when thetime limit is reached.

The wide latitude given states to implement the 1996 legislation has led to manyinnovative welfare-to-work programs throughout the country. To stimulate work byhousehold heads, states have designed programs that provide both services andfinancial incentives (see, for example, United States General Accounting Office,1998). Until now, most of the emphasis has been on services, particularly servicesaimed at preparing welfare recipients for immediate employment (such as job-searchassistance). Less attention has been paid to financial incentives, although moststates have modified their benefit formula to provide financial incentives to work.Before 1996 (and since 1982), welfare benefits in most states were taxed at the rateof 100 percent (after four months of earnings). Such a high tax rate provided apowerful disincentive to work. Beginning in the early 1990s, some states were grantedwaivers to the AFDC program rules and several of these states introduced “enhanceddisregards,” which increased the amount of earnings not counted in the calculationof welfare benefits. Since PRWORA, establishment of enhanced disregards hasaccelerated. According to Gallagher et al. (1998), between January 1992 and October


1997, 41 states adopted some form of enhanced disregards, 11 of them prior toAugust 1996.1

The earnings disregards being used by the states under TANF have two components,a flat component and a variable component. The flat component is a fixed dollaramount of exempt earnings. The variable component is a percentage of earnings(either fixed or varying with the level of earnings, time spent on welfare, or caseloadstatus). Before TANF (from 1981 to 1996), the AFDC program had a flat disregard of$120 for the first 12 months of earnings and $90 thereafter. The variable disregardwas one third of earnings for the first four months of earnings and zero thereafter,thus creating a “tax” (or benefit reduction) rate of 100 percent on earnings.2 AfterTANF, many states adopted very liberal (or enhanced) disregards. For example,Connecticut currently disregards all earnings below the poverty level, so that theeffective benefit reduction rate is zero for these families. Other states, such as Nevada,disregard all earnings initially, but then phase in decreasing disregards over time. Asubstantial number of the states disregard between 20 percent and 50 percent ofearnings and have a flat disregard between $100 and $200 per month.3

During the period in which states were incorporating financial incentives into theirwelfare benefit formulas, work effort among welfare recipients increased dramatically.From 1993 to 1997, employment among single mothers on welfare rose by 14percentage points. According to Blank, Card, and Robins (2000), welfare mothersaccounted for close to half the rise in work by all single mothers over this period. AsBlank, Card, and Robins explain, these rises are especially notable in view of therapid decline in welfare use over the same period, which might have been expected toshift the pool of remaining welfare participants toward a more disadvantaged andless work-ready population. Even with this potential selection effect, however, workeffort among welfare recipients rose.

Of course, the work effort of welfare recipients was probably affected by otherchanges that occurred during this period. One important change was a substantialexpansion of the earned income tax credit (EITC). Blank, Card, and Robins showthat while some of the rise in employment among welfare recipients is undoubtedlydue to the expansion of the EITC, some of it is also probably due to the adoption ofenhanced welfare disregards.

Despite substantial increases in work effort among welfare recipients in recent years,most welfare recipients remain out of work or are working too few hours to beeconomically self-sufficient. Given the time limits, it is crucial that recipients becomeemployed full-time before exiting welfare.4

ENCOURAGING FULL-TIME WORK BY WELFARE RECIPIENTS—THE SELF-SUFFICIENCY PROJECT

Few of the states’ enhanced disregards are structured to encourage full-time work.The same is true for the EITC. The reason is that the financial rewards from working

1 Enhanced earnings disregards are not a new policy. From 1967 until 1981, the federal AFDC programprovided a modest earnings disregard for welfare recipients. Some of the earnings disregards introducedsince the early 1990s are similar to the pre-1982 disregard.2 From 1967 to 1981, the AFDC program had a flat disregard of $30 and a variable disregard of one-thirdthroughout the duration of a welfare spell.3 A more detailed description of the earnings disregards used in the states is presented in Robins,Michalopoulos, and Pan (2000).4 While encouraging full-time work is desirable for achieving economic self-sufficiency, there are costs todoing so. For example, working full-time reduces the time available for child-rearing, which could haveundesirable consequences for children. Furthermore, restricting government aid to full-time workers couldintroduce problems of equity.


can be achieved at low levels of work effort as well as at high levels of work effort. Forexample, in 1998, the EITC for a family with two children increased with earnings upto $9390 per year and was constant between earnings of $9390 and $12,260, afterwhich it was phased out. Thus, a person receiving a wage of $6 per hour would havereceived the maximum EITC subsidy of $3756 (more than $300 per month) by workingfull-time (e.g., 30 hours per week). However, a substantial subsidy can also have beenreceived for part-time work, and the full subsidy can be received at part-time work ifthe person is earning much more than $6 per hour. Similarly, welfare recipients inmost states can benefit from enhanced disregards at less than full-time work as wellas at full-time work.

The fact that full-time work is relatively infrequent among welfare recipients andthat full-time employment is needed to achieve economic self-sufficiency suggestsa possible need for restructuring financial incentives to encourage full-time work.A unique social experiment being conducted in two provinces in Canada (BritishColumbia and New Brunswick) is testing a financial incentive program that rewardswelfare recipients only if they work full-time.5 The program is called the Self-Sufficiency Project, or SSP. Under SSP, which began in late 1992, long-term single-parent welfare recipients (those receiving welfare benefits for at least a year),6 whotake a full-time job within one year, are eligible to receive an earnings supplementfor as long as three years. The SSP supplement is quite generous, and in certaincases, can double a person’s earnings. For example, in New Brunswick, someoneearning $10,000 per year (say working 40 hours per week for 50 weeks at $5 perhour) would receive supplement payments totaling $10,000 per year.7 As long asthe recipient continues to work full-time, the supplement can be received for aslong as three calendar years.

SSP has been remarkably successful during its early years of implementation. Inthe fifth quarter after the program began, full-time employment of the programgroup was more than double the full-time employment of the control group, 29percent versus 14 percent (see Card and Robins, 1998; Lin et al., 1998; Michalopouloset al., 2000). SSP achieved this result primarily by moving people from non-employment to full-time employment, but also by switching a significant numberof people from part-time to full-time employment. Although SSP increasedgovernment transfer payments by about $55 per month (net of taxes), each $1 thegovernment spent on additional transfer payments brought more than $2 ofincreased earnings and led to more than $3 of additional income for program groupmembers. SSP also reduced poverty (the fraction of the program group having afamily income below the low-income threshold) by 12 percentage points andincreased spending on food, clothing, and shelter.8

5 The welfare system in Canada is called Income Assistance. Canada has no food stamp program, so cashbenefits in Canada are generally higher than in the United States.6 The restriction to long-term recipients is intended to minimize “entry effects” (people applying for wel-fare in order to receive the supplement) and “windfall effects” (benefits accruing to recipients who wouldhave left welfare and worked full-time anyway in the absence of the earnings supplement). As indicated inCard, Robins, and Lin (1998), this provision effectively limited entry and windfall effects in the studysample.7 Formally, the SSP subsidy is given by 0.5(E* – E), where E* is target earnings ($30,000 in New Brunswickand $37,000 in British Columbia, both in Canadian dollars, when SSP began) and E is actual earnings. Thesubsidy, available only to people who work 30 hours per week or more, has been adjusted upward slightlyfor inflation since 1992.8 A companion experiment, conducted on a group of new applicants for welfare in British Columbia, didnot lead to any net increase in government cash transfer payments and had similar effects on employment,income, and poverty (Michalopoulos, Robins, and Card, 1999).


The success of SSP in Canada raises the intriguing question of whether such aprogram would generate similar effects in the United States.9 Although the welfaresystems in Canada and the United States are similar, differences make it difficult todraw comparisons. This has been especially true since PRWORA was enacted. Asindicated, the U.S. system now imposes a time limit on welfare receipt and stronglyemphasizes placing recipients in work activities before the time limit is reached. TheCanadian welfare system does not currently have time limits, although it emphasizespromoting economic self-sufficiency through work, as evidenced by the Canadiangovernment’s willingness to test SSP.

In designing SSP, researchers at the Manpower Demonstration ResearchCorporation (MDRC) used a microsimulation model to predict the effects ofalternative program models. As described in Greenberg et al. (1995a) andMichalopoulos (1999), the model performed extremely well in predicting the eventualeffects of the SSP program tested. Given its proven accurate predictive ability, themodel is used in this paper to estimate the effects of an SSP-type financial incentiveprogram in the United States.

Implementing SSP in the United States

Because welfare reform efforts are already underway in the United States, and becausethe EITC has been expanded significantly since 1994, the effects of an SSP financialincentive superimposed on the old AFDC system is not of particular policy relevance.Examined here is what would have happened if states had coupled the non-financialcomponents of their welfare-to-work programs with an SSP-type earnings supplementinstead of with enhanced earnings disregards.10

To answer this, data are used from three welfare-to-work programs currentlyoperating in the United States that are similar to TANF programs and that are beingevaluated using an experimental design. These programs are the Portland, Oregon,Job Opportunities and Basic Skills Training program (PJOBS) being evaluated aspart of the National Evaluation for Welfare-to-Work Strategies, the Florida FamilyTransition Program (FTP), and the Minnesota Family Investment Program (MFIP).The features of these three welfare-to-work programs are defined in Table 1.

In Table 1, the financial incentive features of the programs are distinguished fromtheir other features. As the table indicates, over the study periods, two of the threeprograms (MFIP and FTP) had enhanced disregards. In PJOBS, however, the pre-TANF disregards were used in the experimental welfare-to-work study. Subsequentto the study period, Oregon introduced a 50 percent variable disregard into its TANFprogram. All three programs use intensive case management and mandatoryemployment-focused services as outlined in the TANF legislation. Florida has instituteda shorter intermediate time limit (24 months or 36 months, depending on how job-ready a recipient is) than required by the federal legislation. MFIP dismantled thefood stamp program and turned food stamps, general assistance, and AFDC into onewelfare program. Having one welfare program makes MFIP more similar to theCanadian Income Assistance (IA) program.

9 Although the effects of the SSP program have been reported in a number of studies (Berlin, 2000; Blank,Card, and Robins, 2000; Card and Robins, 1998; Card, Robins, and Lin, 1998; Lin et al., 1998; Michalopouloset al., 2000; and Michalopoulos, Robins, and Card, 1999), none of these studies attempt to generalize thefindings to the U.S. welfare system.10 Perhaps an equally interesting question, not pursued in this paper but examined in Robins, Michalopoulos,and Pan (2000), concerns the effects of an SSP-type earnings supplement program in addition to en-hanced earnings disregards.


Follow-up survey data are available on program and control group members foreach of these three experimental programs. In PJOBS and FTP, follow-up survey dataare available for two years. In MFIP, follow-up survey data are available for threeyears. The microsimulation analysis conducted for this paper uses data from thesecond follow-up year for all three studies. By this time, members of the programgroup in each of the studies would have had some chance to respond to the TANF-like provisions they faced. Furthermore, the second follow-up year falls between 1995and 1997 for all three studies; therefore, members of the program group were probablyalso responding to the expanded EITC. Using the survey data from each program andthe microsimulation model, it is possible to estimate what additional effects, if any,an SSP-type program would have. It is also possible to use the data andmicrosimulation model to estimate the effects of the enhanced earnings disregardsadopted by the study states in response to TANF.

Program(Study Period)

Minnesota (MFIP)(1994 - 1995)

Florida (FTP)(1994 - mid 1998)

Portland, Oregon(PJOBS)(1993 - mid 1996)

Financial Incentives

1. Benefits increased by20% for workers andare reduced by 62%with earnings

2. Benefit may notexceed benefit fornon-workers

1. Disregard of $200plus of remainingearnings

1. Disregard of $30 plus1/3 of remainingearnings, first 4months; no disregardafter 4 months (pre-TANF rules)

Other Features

1. Mandatory employment-focusedactivities

2. Direct child care payments toproviders

3. Food Stamps cash-out4. 52% of long-term recipient full MFIP

program group in urban countiesemployed in quarter 7 (N=676)

5. 42% of long-term recipient MFIPfinancial incentives only programgroup in urban counties employed inquarter 7 (N=681)

6. 38% of long-term recipient controlgroup in urban counties employed inquarter 7 (N=687)

1. 24 month time limit on benefits2. Intensive case management3. Enhanced employment and training

services4. Parental responsibility mandates5. 53% of program group employed in

last quarter of year 2 (N=1,405)6. 45% of control group employed in

last quarter of year 2 (N=1,410)

1. Strongly mandatory2. Integrated case management3. Employment focused4. 46% of program group employed in

last quarter of year 2 (N=3,529)5. 35% of control group employed in

last quarter of year 2 (N=2,018)

Table 1. Features of the welfare-to-work programs in three states.


If adopted in the United States, an SSP program could be operated as a separateprogram, as it is in Canada. Berlin (2000) suggests that if a separate program werecreated, it might make sense to stop the TANF time limit for people working full-time.11 Thus, the time spent working full-time would not count against the 60 month(or less) TANF time limit. Of course, the SSP program could also have a time limit (asit does in Canada), which would limit its cost.12 With a time limit, the spirit of theTANF legislation would be maintained, but in a separate context that uses financialincentives to encourage full-time work.

THE SIMULATION MODEL

An SSP financial incentive can affect people’s decisions about work only if theychoose to work full-time to receive the incentive. Thus, the simulation mustinclude some indication of how a person would respond to different financialincentives. In this paper, the simulation model is based on a traditionalmicroeconomic model of the income or leisure choice, extended to include thechoice to receive welfare, and incorporating an estimate of the stigma associatedwith welfare. The simulation model is the same one Greenberg et al. (1995b)used and is based on the economic model Moffitt (1983) originally developed.

This economic model assumes that welfare recipients will choose how muchto work and whether to receive welfare to maximize their economic well-being.Economic well-being is a combination of income, time spent in non-workactivities, and the stigma associated with receiving welfare. In other words, awelfare recipient will work full-time to receive the SSP supplement only if doingso improves economic well-being. In making this decision, the recipient mustweigh the additional income resulting from the subsidy and earnings with thereduced time spent in activities outside of work (such as child rearing and leisure-time activities) and the loss of economic well-being through the stigma ofreceiving welfare.

An important part of the behavioral model is “welfare stigma,” which is basedon the notion that receiving welfare reduces economic well-being (all else equal).Although stigma is often ignored in economic models of behavior, Moffitt (1983)notes that “(t)he existence of stigma has been amply documented in thesociological literature.” Incorporating stigma into the behavioral model is keybecause it accounts for the fact that many people who are eligible for welfarebenefits do not receive them. In the absence of stigma, welfare benefits increaseincome without increasing work effort. The fact that many people decline thispure income gain is inconsistent with traditional economic models of well-being,but may be explained by the existence of stigma. There are, of course, otherreasons why people who are eligible for welfare do not receive it. They might beunaware that they are eligible, for example, or they might judge the costs toohigh of applying for welfare and complying with its rules. However, these factorsaffect the use of welfare in ways similar to stigma. Therefore, the use of the

11 As of June 2000, three states (Illinois, Rhode Island, and Maryland) have passed legislation that stopsthe time limit for families who work a minimum number of hours.12 TANF costs would be reduced in the short-run as persons shifted from TANF to SSP. Of course, thesereduced welfare costs would be offset by SSP costs. For the SSP program in Canada, the cost of SSPslightly exceeded the reduced welfare cost for long-term recipients, but was about the same as the reducedwelfare cost for new applicants (net of the additional income taxes resulting from the additional full-timework). Part of the additional SSP cost was “windfall,” resulting from SSP benefits being paid to personswho would have left welfare and worked full-time anyway in the absence of the SSP-type financial incen-tive.


term stigma in the simulation model implicitly incorporates these other reasonsfor not receiving welfare.13

The parameters of the economic model are not directly estimated in this paper.Instead, they are updates of Moffitt’s estimates and account for conditions during theperiod of the study samples and incorporate other relevant information from thelabor supply literature.14 In applying Moffitt’s model in the simulation analysis, themodel was expanded to allow individual variation in welfare stigma and individualvariation in one of the key parameters of the economic model. This is done in the firststep of the simulation, which compares outcomes reported by each member of thesample on the follow-up survey with the outcomes the base model predicted usingthe earnings disregard that the person actually faced. If the individual’s surveyresponses differ from the base model’s predictions of that person’s behavior, than thestigma and utility parameters are altered to make the two agree. This process andother alterations to Moffitt’s model are described in a technical appendix availablefrom the authors. Through the two sources of individual variation, the model istherefore “calibrated” to the samples used, thus updating Moffitt’s model so that itmatches the behavior of these samples.15

The simulations use data for people from the three welfare-to-work programstwo years after random assignment who met the same eligibility criteria as requiredin the Canadian SSP program. These consist of people who were on welfare in thelast month of the two-year follow-up period and had been on welfare for 11 of the12 previous months. Thus, the simulations estimate what would have happened if,at the end of the two-year follow-up period, the experimental welfare-to-workprograms had offered an SSP-type financial incentive instead of the enhanceddisregards they did offer.

The second step in estimating the effects of an SSP-type financial incentive (afteradjusting the parameters of the behavioral model) is to simulate labor supply andwelfare use under the various financial incentive schemes. Estimated effects are thencalculated by comparing the simulated outcomes under the various incentive schemes.For example, to determine the difference between outcomes under an SSP incentiveand outcomes under an enhanced disregard, outcomes are first simulated under eachincentive. The difference between the two sets of predicted outcomes indicates howmuch different behavior would be under the two financial incentives.

To make these comparisons, outcomes were simulated with the financial incentivesexisting under the following environments:

13 While the model incorporates the effects of welfare stigma, it ignores other factors that may be impor-tant. First, the model does not explicitly account for certain work-related costs such as transportation andchild-care expenses. Second, the model assumes that labor demand is perfectly elastic and that increasesin labor supply do not exert downward pressure on wage rates and that jobs are available at the individual’smarket wage.14 The model uses estimates of key parameters (primarily income and wage effects) that are well within therange obtained in other studies. Greenberg et al. (1995a, b) discuss the sensitivity of the simulation model’spredictions to the parameters used.15 A number of individuals reported earning so much that they should not have been eligible for welfareeven though administrative records indicated that they were receiving welfare. To reconcile earnings andwelfare benefits when doing the calibration, we reduced hours worked to a level that was consistent withwelfare payments reported on the administrative records. This meant that a number of individuals whoreported working full-time were assumed to have worked part-time and a number of people who reportedworking were assumed to be not working. This “correction” affected the Portland sample the most be-cause the 100 percent tax rate on earnings for welfare recipients under the AFDC disregards being used inPortland at the time meant that people lost their eligibility for welfare at relatively low earnings levels.


• AFDC. The base against which the other outcomes are compared, its environmentincludes no enhanced earnings disregard as part of the AFDC program (pre-TANF environment).

• TANF. This environment includes the TANF enhanced disregard currently beingused in the three states. The enhanced disregards actually implemented inOregon and Minnesota differ from those used in the PJOBS and MFIPexperimental studies. As indicated above, PJOBS did not use an enhanceddisregard, but Oregon consequently instituted an enhanced disregard as partof its TANF program. The enhanced disregard used in the Florida TANF programis the same one used in the FTP experimental study.

• SSP. In this environment, sample members in each of the three states are offeredan SSP-type financial incentive. The simulated outcomes reflect what mighthave happened if states had offered an SSP-type financial incentive at the endof the two-year follow-up period.

Comparing the simulated decisions under the AFDC disregard with the simulateddecisions under the TANF disregard and the SSP-type financial incentive providesestimates of the effects of each financial incentive relative to the pre-TANF AFDCprogram.16 Since each of the effects is estimated using a common base, comparingthe effects across the various alternatives indicates the relative effectiveness of eachof the financial incentive programs.

Simulated SSP Financial Incentive Schemes

Both SSP financial incentive schemes examined in this paper are patterned after theprograms tested in British Columbia and New Brunswick. To be included in thesimulation, a welfare recipient must have been receiving AFDC at the end of the two-year follow-up period as well as 11 of the 12 previous months. Once eligible, thewelfare recipient qualifies for the earnings supplement if he or she takes a full-timejob of 30 or more hours per week paying at least the minimum wage. In addition, therecipient cannot simultaneously receive welfare and the SSP earnings supplement.

The SSP financial incentive operates by paying people who meet the full-time workrequirement a supplement equal to half the difference between a target earnings leveland actual earnings (see note 4 for the exact formula used). For purposes of thispaper, program effects are examined using two target earnings levels: $20,000 and$30,000.17 For the target earnings level of $20,000, if the person works 40 hours perweek for 50 weeks per year and earns $7.00 per hour (about the average wage in these

16 The simulation samples in Minnesota and Florida come from environments where the experimentalenhanced disregards offered by the study existed. The simulations, therefore, cannot provide estimates ofoffering the enhanced earnings disregards used in these studies. Instead, they provide estimates of theeffects of taking away those financial incentives. The effects of taking away a financial incentive will differfrom the effects of adding a financial incentive to the extent that the samples the simulations started withwould have differed depending on whether the financial incentive had actually been offered. After takingaway the enhanced disregards from the experimental study samples, we then simulated adding back in theTANF enhanced disregards for the same samples. In Florida, where the experimental study enhanceddisregard is the same as the statewide TANF enhanced disregard, the simulated effects of taking away theenhanced disregard are the same (but opposite in sign) as the effects of adding back in the enhanceddisregard. In Minnesota, where the enhanced disregard used in the MFIP experimental study differedfrom the enhanced disregard eventually adopted under the statewide TANF program, the magnitude of thesimulated effect (in absolute value) of taking away the MFIP enhanced disregard is different from theeffect of adding back in the TANF disregard.17 In the actual SSP programs being tested in Canada, the target earnings levels (in Canadian dollars) are$37,000 in British Columbia and $30,000 in New Brunswick. At an exchange rate of $0.75 U.S. dollars perCanadian dollar, these target earnings levels in U.S. dollars are $27,750 and $22,500, respectively.


samples), the annual SSP subsidy would be $3000, or roughly one-fifth of annualearnings of $14,000. For the target earnings level of $30,000, the SSP subsidy wouldbe $8000, or roughly three-fifths of annual earnings.

While these subsidy amounts seem substantial, it should be kept in mind that therecipient is required to give up AFDC (TANF) benefits in order to receive the subsidy.In Portland, the average annual AFDC benefit was nearly $6000 per year, which issubstantially more than the SSP subsidy under the program with the $20,000 targetearnings level, and about $2000 less than the SSP subsidy under the program withthe $30,000 target earnings level. Furthermore, people who work might also payfederal and state income taxes, which further reduces government costs of theprogram. However, people who work qualify for the EITC and some people whowould have left AFDC (TANF) without being offered the SSP financial incentivewill receive subsidies.

Figure 1a-c shows for the three sites the net weekly income by weekly work effortfor a single mother of two earning $8.00 per hour under SSP programs with a $20,000and $30,000 target earnings level. For reference, net weekly income is also shownunder the traditional AFDC earnings disregard ($120 per month disregarded, taxedat the rate of 100 percent thereafter) and the post-TANF earnings disregards adoptedby states (a 50 percent variable disregard in Portland, the same $200 flat and 50percent variable disregards as used in the experimental FTP program in Florida, anda 36 percent variable disregard in Minnesota that differed from the disregard used inthe experimental MFIP program).18

The figures illustrate how the traditional AFDC earnings disregards provide littleincentive for welfare recipients to work. As shown in Figure 1a for Portland, forexample, net weekly income is relatively constant between four and 18 hours of workper week for a mother of two who earns $8.00 per hour. This is the range over whichearnings have exceeded the AFDC flat disregard ($120 per month) and are deducteddollar for dollar from the AFDC benefit.

The post-TANF enhanced earnings disregards adopted by the states improve thefinancial incentives to work part-time, but leave the financial incentives to work full-time pretty much unchanged. In contrast, SSP does not provide an incentive to workfewer than 30 hours per week, but substantially increases the incentive to work 30 ormore hours. This is especially true of the SSP program simulated with a $30,000target level.

Outcomes Examined

Reported for each of the SSP programs simulated are: effects on labor force outcomes(full-time and part-time employment, hours of work, and earnings), welfare outcomes(receipt of AFDC, food stamps, and SSP), various components of net income (AFDC,food stamps, EITC, SSP, and income taxes),19 and estimated effects on these outcomesof the TANF earnings disregards. It is important to note that these effects are for thechosen sample. Namely, they are effects for people who were receiving AFDC foralmost the entire second year of the study follow-up period—that is, sample memberswho would be eligible for the SSP supplement at the end of the second year of follow-up according to how the program operates in Canada. In general, the effects of the

18 Minnesota’s TANF variable earnings disregard was 36 percent until October 1999, when it increased to38 percent. The simulations presented in this paper were performed using the 36 percent disregard.19 The details of how each of these outcomes is calculated are available from the authors, and in Robins,Michalopoulos, and Pan (2000). Certain components of net income are ignored in the calculation of netgovernment assistance and net income, such as social security benefits and sales taxes.


Figure 1. Budget constraints under alternative financial incentive schemes singlemother with two children and wage rate of $8/hour.

a.Portland

b.Florida

c.Minnesota

Hours per Week

Net

Wee

kly

Inco

me

Net

Wee

kly

Inco

me

Net

Wee

kly

Inco

me

Hours per Week

Hours per Week

AFDCTANFSSP 20KSSP 30K




SSP financial incentives and the TANF earnings disregards for the full welfarepopulation will differ from their effects for the simulation samples because the fullwelfare population includes people who received welfare for shorter lengths of time.

Characteristics of the Samples

Characteristics of the three welfare-to-work samples are presented in Table 2. Thefirst three columns show the average characteristics of all people who were randomlyassigned to the program group in each site. The last three columns show thecharacteristics of the samples used in the simulations. The simulation samples havefewer people than the program group samples because they were limited to peoplewho were receiving AFDC at the end of the second year of follow-up and had beenreceiving AFDC for 11 of the 12 months before the end of the second year. Thesimulation samples have real-world relevance because an SSP-type financial incentiveoffered in the United States would most likely be offered only to people still receivingTANF benefits at the time the SSP program would be introduced.

Never-married women head almost all families in the program groups. In PJOBS,about two-fifths were never married at the beginning of the follow-up period, in FTPthree-fifths were never married, and in MFIP about two-thirds were never married.The average mother was about 30 years old and more than two-thirds had childrenyounger than 6 years of age. In PJOBS, about one-quarter were black, in FTP morethan half were black, and in MFIP about two-fifths were black. Roughly one-third ofthe sample lacked a high school diploma or a GED and roughly one-third of thesample had received AFDC as a child. At the start of the follow-up period, just over 10percent of the samples worked, but only a small fraction worked full-time.

In the year prior to the start of the welfare-to-work program, about 30-40 percentof the full program group samples of mothers were employed. Thus, many of thesemothers had some recent work experience. In PJOBS and FTP, about half the programgroup received AFDC or food stamps the full year, while in MFIP more than three-quarters received AFDC or food stamps the full year. The MFIP sample is somewhatmore disadvantaged than the PJOBS and FTP samples because it includes only long-term recipients from urban counties. These are sample members who had been onAFDC for at least 24 of the 36 months before random assignment, and is the groupthat was most affected by the MFIP program (Miller et. al., 1997).

Each of the experimental welfare-to-work programs significantly increasedemployment (Bloom et al., 1998; Miller et al., 1997; and Scrivener et al., 1998).20

Table 3 shows selected effects of each of these programs. Because the welfare-to-work programs had already successfully increased employment, adding the SSPfinancial incentive may generate smaller increases in employment than if the SSPprogram had been superimposed on the old AFDC system. A similar argumentcan be made for adding the SSP incentive to a transfer system already having theEITC. In Canada, SSP was introduced into an environment without an EITC-typeprogram. Hence, some of SSP’s effects in Canada may have already occurred forthe types of welfare recipients who responded to the EITC in the United States. Ininterpreting the effects of the SSP programs in the United States, therefore, itshould be kept in mind that many of the people who would have responded to

20 Changes in employment over time for members of the program group do not necessarily representeffects of the welfare-to-work program because other changes are also occurring that affect employment.To validly measure effects, behavior of a randomly selected control group must also be tracked and com-pared with the behavior of the program group. Such an experimental approach to measuring effects isbeing used in each of the welfare-to-work programs examined in this paper.


SSP by finding full-time employment may have already responded to either the EITCor the welfare-to-work program.

The last three columns of Table 2 indicate that the simulation samples (those whowere on AFDC at the end of the second year of follow-up and were on AFDC for atleast 11 of the 12 previous months) are somewhat more disadvantaged than the fullexperimental program groups. They are less likely to have a high school diploma(although not by very much in MFIP); they are less likely to have worked in the yearprior to random assignment; they were much less likely to have worked during thefollow-up period; and their average wages were lower.

Simulation Results

Tables 4 shows the estimated effects from the simulations for each of the three studysamples. The first column for each sample (columns 1, 4, and 7) shows the estimatedeffect of the earnings disregard used in the state’s TANF program. Each effect ismeasured relative to the pre-TANF AFDC flat earnings disregard of $120. The nexttwo columns for each sample (columns 2-3, 5-6, and 8-9) show the estimated effectsof the two SSP-type financial incentive programs.

In Portland’s PJOBS program, column (1) illustrates strikingly how the variableearnings disregards many states currently use may be quite successful in movingwelfare recipients to work, yet at the same time might not succeed in movingrecipients to economic self-sufficiency. The TANF earnings disregard Oregoncurrently uses is estimated to have increased employment considerably among long-term recipients. (Recall that the simulation selects program group members in thestudy who were still on AFDC in the last month of the second year of follow-up andin at least 11 of 12 previous months.) The 50 percent earnings disregard is estimatedto have increased employment by just over 23 percentage points. All of thisemployment, however, is part-time.

Comparing columns 2 and 3 with column 1 provides an estimate of what wouldhave happened if PJOBS had instituted an SSP-type financial incentive instead of a50 percent earnings disregard. Unlike the 50 percent earnings disregard, the SSP-type financial incentive programs increase full-time employment. Furthermore, forboth SSP-type programs, they do so without any net increase in cash transfers fromthe government. An SSP program with target earnings of $20,000 would increasethe full-time employment rate by 5.2 percentage points, annual earnings by $406,and net annual income by $261 (column 3). All this occurs while decreasing netcash transfer payments from the government to recipients by $146. An SSP programwith target earnings of $30,000 would increase the full-time employment rate by10.4 percentage points, annual earnings by $884, and net annual income by $946,without increasing the cash transfers to recipients by a statistically significantamount (column 3).21

Despite their effects on full-time employment, the SSP-type programs do not generatenearly as much overall employment in Portland as the TANF earnings disregard. The

21 As noted by Citro and Hanushek (1991), error in simulation models can stem from the sample chosen,uncertainty in parameter estimates, model misspecification, measurement error, miscoding of data, andprogramming errors in constructing the simulation model. In determining the statistical significance ofthe simulated effects, we took into account only the first source of variation. Specifically, significance testswere based on t-tests in which the standard error of the simulated effects is based on the variance acrosssample members in their actual survey responses and their simulated responses to the subsidies as if thelatter represented actual behavior. As a result, the effects in Table 4 overstate somewhat the likelihood thatsimulated impacts are statistically significant.


Tab

le 2

. Ch

arac

teri

stic

s of

pro

gram

gro

up

mem

ber

s in

th

ree

wel

fare

-to-

wor

k p

rogr

ams.

All

Pro

gram

Gro

up

Mem

ber

sP

rogr

am G

rou

p M

emb

ers

inth

e S

imu

lati

on

Sam

ple

s


Sou

rces

: M

DR

C c

alcu

lati

ons

usi

ng

Bas

elin

e In

form

atio

n F

orm

s, T

wo-

year

Cli

ent

Su

rvey

dat

a fr

om t

he

Por

tlan

d, O

rego

n W

elfa

re-t

o-W

ork

Eva

luat

ion

, Tw

o-ye

ar C

lien

t S

urv

ey d

ata

from

th

e F

lori

da

Fam

ily

Tran

-si

tion

Pro

gram

Eva

luat

ion

, 36-

mon

th C

lien

t S

urv

ey d

ata

from

th

e M

inn

esot

a F

amil

y In

vest

men

t P

rogr

amE

valu

atio

n,

and

Un

emp

loym

ent

Insu

ran

ce e

arn

ings

rec

ord

s an

d p

ub

lic

assi

stan

ce b

enef

it r

ecor

ds

dat

afr

om O

rego

n, F

lori

da

and

Min

nes

ota.

Not

e:

Th

e si

mu

lati

on s

amp

les

incl

ud

e p

rogr

am g

rou

p m

emb

ers

wh

o w

ere

not

liv

ing

wit

h a

sp

ouse

or

par

tner

at

the

foll

ow-u

p in

terv

iew

an

d w

ho

rece

ived

AF

DC

in t

he

twen

ty-f

ourt

h f

ollo

w-u

p m

onth

an

d in

at

leas

t 11

of

the

12 m

onth

s p

rior

to

that

.

Tab

le 2

. co

nti

nu

ed.

All

Pro

gram

Gro

up

Mem

ber

sP

rogr

am G

rou

p M

emb

ers

inth

e S

imu

lati

on

Sam

ple

s


SSP program with target earnings of $30,000 would increase employment by justover 6 percentage points and the SSP program with target earnings of $20,000 wouldincrease employment by only 1 percentage point. Nonetheless, the SSP-type incentivesare estimated to substantially reduce AFDC and food stamp payments. Such reductionsare possible because many of the people who are estimated to go from part-time tofull-time work when offered an SSP supplement are people who work very few hoursand receive practically their full welfare grant amount.

These effects imply that an SSP program as generous as the ones tested in Canadawould modestly increase employment and income among this group of persons inthe Portland PJOBS program at no additional cost to the government. Full-timeemployment would have increased by a somewhat greater amount. For the moregenerous SSP-type program, net family income would increase by almost three timesthe estimated increase generated by the TANF earnings disregard. It is important toemphasize that these effects are those in addition to the effects already generated bythe expanded EITC. Without the EITC, the effects of SSP and the TANF disregardmight have been substantially larger.

Thus, in the present welfare environment, if the policy objective is to increasefull-time employment and income without any additional cost to the government,these effects suggest that a moderately generous SSP program could be somewhatmore effective than the enhanced disregard actually adopted by Oregon under itsTANF program.

Table 3. Effects on employment and earnings in the Portland, Florida, and Minnesotastudies.

Sources: The Portland numbers are taken from Scrivener et al., 1998. The FTP numbers are taken fromBloom et al., 1998. The MFIP numbers are taken from Miller et al., 1997.Notes: The MFIP numbers are for long-term recipients in urban counties. MFIP defined long-time receiptas two years or more in the prior three years. Random assignment dates were: Portland, February 1993-December 1994; FTP, May 1994-October 1996; MFIP, April-December 1994.Following the Portland and MFIP studies, Quarter 1 is defined as the calendar quarter in which randomassignment falls. In the FTP study Quarter 1 was defined as the quarter after the calendar quarter in whichrandom assignment fell. Quarter 4 in the table therefore corresponds to Quarter 3 in Bloom et al., 1998.A two-tailed t-test was used to assess the statistical signifiance of differences between outcomes for pro-gram and control group members. Statistical significance levels are indicated as *** = 1 percent, ** = 5percent, and * = 10 percent.Rounding may cause slight discrepancies in sums and differences.a Sample sizes differ from Table 2. Sample sizes include program and control group members who re-sponded or did not respond to the follow-up survey.


Simulation effects for Florida’s FTP program are presented in columns 4-6 of Table4. Unlike Portland’s experimental welfare-to-work program, FTP included an enhanceddisregard ($200 flat and 50 percent variable) as part of its program. The enhanceddisregard in Florida’s TANF program is identical to the enhanced disregard tested inthe FTP program. The estimated effects of this enhanced disregard for the simulationsample are shown in column 4 of Table 4.22 Columns 5 and 6 show estimated effectsof the two SSP-type programs.

Column 4 shows that the simulation predicts no changes in employment from theFTP (TANF) earnings disregard. Hours of work are estimated to increase slightly, butthe incentive does not induce anyone to start working, nor does it induce anyone toswitch from part-time to full-time work. Furthermore, the FTP financial incentive isestimated to increase welfare caseloads. The FTP financial incentive does increaseannual net income, but much of this increase (more than two-thirds) comes from anincrease in government spending.

Both of the SSP programs would have substantially increased full-time employment.Under the less generous SSP program (target earnings of $20,000), full-timeemployment would have risen by just over 15 percentage points. Under the moregenerous SSP program (target earnings of $30,000), full-time employment wouldhave risen by just over 24 percentage points. For the less generous SSP program, allof the increase in full-time employment is estimated to be the result of people switchingfrom part-time to full-time work, with no net increase in overall employment. For themore generous SSP program, nearly all of the increase in full-time employment isestimated to be the result of people switching from part-time to full-time work, withonly a 1.5 percentage point increase in overall employment. The SSP programs wouldhave increased annual hours of work by between 126 and 240 hours and annualearnings by between $595 and $1159. The less generous SSP program would haveincreased net family income by $881, at a net cost to the government of $286. Themore generous SSP program would have increased net family income by more than$2000 (which represents more than a 20 percent increase), at a net cost to thegovernment of about $1000.

The less generous SSP program costs about the same as the enhanced disregard ofFTP, although the sources of costs are different for the two strategies (SSP inducesgreater decreases in welfare payments and more taxes, but adds EITC costs plus SSPsupplement payment costs). However, the SSP program more than doubles familyincome, primarily because SSP induces a substantial amount of full-time employment.Overall, then, the main difference between SSP and the enhanced disregard in FTP isthe greater full-time employment associated with the SSP program. Thus, thesimulation model predicts that many of the people still on FTP and working part-time because of the EITC and the non-financial components of FTP would have beeninduced to work full-time under an SSP program that conditioned benefits on full-time employment.

Simulation effects for Minnesota’s MFIP program are given in columns 7-9. Column7 presents the estimated effects of Minnesota’s earnings disregard under TANF.

22 These effects do not represent experimental effects of the financial incentive component of the FTPprogram. Such a program was never tested experimentally. Instead, the numbers represent the simulatedeffects of FTP’s enhanced earnings disregard for the simulation sample of long-term welfare recipients.Furthermore, the numbers were derived from taking away the financial incentive from the simulationsample. To the extent that the sample of long-term recipients would have been different if there had notbeen a financial incentive as part of the FTP program, the effects of offering the financial incentive willdiffer from the effects of taking away the financial incentive from people who were long-term recipientswhen the financial incentive was offered.


Tab

le 4

. Sim

ula

ted

eff

ects

of

SS

P f

inan

cial

in

cen

tive

s on

par

tici

pan

ts i

n t

hre

e w

elfa

re-t

o-w

ork

pro

gram

s in

cen

tive

s in

trod

uce

d a

fter

tw

o ye

ars

for

fam

ilie

s st

ill o

n A

FD

Ca .

ee

e


Minnesota implemented an enhanced disregard that was slightly different fromthe one used in the experimental MFIP programs. The earnings disregard used inthe Minnesota TANF program is a 36 percent variable disregard with no flatdisregard (but see note 18).

According to column 7, the simulation model predicts that the TANF earningsdisregard increased part-time employment for the sample, but had virtually no effecton full-time employment.23 With the TANF financial incentive, overall employmentis estimated to be 11.4 percentage points higher, annual hours of work 93 hours higher,and annual earnings $617 higher than if there had been no enhanced earningsdisregard. The simulation model predicts that the TANF enhanced earnings disregardhas virtually no effect on welfare receipt for this sample. Net government costs arenearly $500 higher and net family income is about $1000 higher as a result of theenhanced earnings disregard.

Similar to the FTP sample, the simulations predict that the SSP programs wouldnot increase employment, but would cause a substantial number of people to switchfrom part-time to full-time work (columns 8 and 9). Under the less generous SSPprogram ($20,000 target), full-time employment would increase by 13 percentagepoints. Under the more generous SSP program ($30,000 target), full-time employmentwould increase by almost 24 percentage points. For the less generous SSP program,net government assistance would rise by $193 and net family income would risemore than twice that amount. For the more generous SSP program, net governmentcost would increase by $591 per person, while annual net family income would againrise by more than twice that amount.

SUMMARY AND CONCLUSIONS

Since the enactment of PRWORA in 1996, most states have included financialincentives in their overall welfare-to-work programs. These financial incentives havetaken the form of enhanced disregards. Before PRWORA, under the AFDC program,no earnings were disregarded after one year of earnings. This meant that recipientsfaced a 100 percent “tax rate” on their earnings, which can impose a significantdisincentive to work. Given PRWORA’s time limits, and the importance of movingwelfare recipients into the work force, most states have reduced this 100 percent taxrate through enhanced disregards. The intent of the enhanced disregards is to providea work incentive and to ease the transition from welfare to work.

Sources: Data from Portland, Florida, and Minnesota Welfare-to-Work Programs, MDRC simulation model.aEffects are measured relative to outcomes under the welfare to work program with a flat earnings disre-gard of $120 per month and a child support disregard of $50 per month (the pre-TANF disregard).bOregon’s TANF program has a 50 percent variable earnings disregard.cFlorida’s TANF program has a $200 flat disregard and a 50 percent variable disregard.dMinnesota’s TANF program has a 36 percent variable disregard.eMinnesota’s TANF benefit includes the cash value of food stamps.fNet Government Assistance is TANF + Food Stamps + EITC + SSP - Taxes.gNet Income is Net Government Assistance + Earnings.

Table 4. continued.

23 In effects not shown here but reported in Robins, Michalopoulos, and Pan (2000), we found that the MFIPexperimental earnings disregard reduced full-time employment and increased part-time employment.


One of the problems with enhanced disregards is that they often provide an incentiveto work only part-time and possibly a disincentive to work full-time. Faced with asudden loss in welfare benefits when the time limit is reached, many recipients mayfind that part-time earnings are not enough to allow them to be economically self-sufficient. Although these families will still be eligible for the EITC, the EITC alsoprovides only limited incentives to work full-time.

This paper has presented effects from a simulation model to examine whether analternative form of financial incentive could increase full-time employment amonglong-term welfare recipients. The alternative financial incentive scheme consideredin this paper is based on the Self-Sufficiency Project (SSP), a program being conductedexperimentally in two provinces in Canada. The SSP provides a direct incentive towork full-time because it conditions benefits on full-time work. The SSP is beingevaluated using a randomized experiment and the effects thus far indicate that theprogram is having sizable effects on full-time employment, at little additional transfercost to the government. Because the welfare-to-work programs under TANF in theUnited States are different from the welfare-to-work programs in Canada, and becauseCanada does not have an EITC, it is not clear what the additional effects an SSP-typeprogram might be in the United States. This paper has used a microsimulation modelto predict what the effects might be of an SSP-type program in the United States.

The simulation effects indicate that an SSP-type program in the United States, inplace of the enhanced disregards currently in use, could have significantly greatereffects on full-time employment for long-term welfare recipients at modest additionalcost to the government. Perhaps the most attractive feature of SSP is that it canachieve gains in family income that are as much as three times the increase ingovernment cost. Such a high “efficiency” ratio is rarely seen in financial incentiveprograms. Thus, an SSP-type program in the United States may be an attractive wayof partially easing the transition from welfare to work under a system of time-limitedwelfare benefits.

The research for this paper was supported by a grant to MDRC from The Annie Casey Foundation.We are grateful to Mike Laracy for his encouragement and support. An earlier version of thepaper was presented in a seminar at MDRC. The authors wish to acknowledge the helpfulcomments of Gordon Berlin, Barbara Goldman, Judy Gueron, and two anonymous referees.

PHILIP K. ROBINS is Professor of Economics at the University of Miami, Coral Gables.

CHARLES MICHALOPOULOS is Senior Research Associate at MDRC, New York.

ELSIE PAN is Research Associate at MDRC, New York.

REFERENCES

Bane, M.J. & Ellwood, D. (1994). Welfare realities: From rhetoric to reform. Cambridge, MA:Harvard University Press.

Berlin, G. (2000). Encouraging work, reducing poverty: The impact of work incentive pro-grams. New York: Manpower Demonstration Research Corporation.

Blank, R., Card, D., & Robins, P.K. (2000). Financial incentives for increasing work and in-come among low-income families. Pp. 373–419 in R.M. Blank & D. Card (Eds.) FindingWork: Jobs and Welfare Reform. New York: Russell Sage.


Bloom, D., Farrell, M., Kemple, J.J., & Verma, N. (1998). The family transition program: Imple-mentation and interim impacts of Florida’s initial time-limited welfare program. New York:Manpower Demonstration Research Corporation.

Card, D. & Robins, P.K. (1998). Do financial incentives encourage welfare recipients to work?Evidence from a randomized evaluation of the Self-Sufficiency Project. Vol. 17, pp. 1–56 inS.W. Polachek (Ed.), Research in Labor Economics. Stamford, CT: JAI Press.

Card, D., Robins, P.K., & Lin, W. (1998). Would financial incentives for leaving welfare leadsome people to stay on welfare longer? An experimental evaluation of ‘entry effects’ in theself-sufficiency project. Working Paper 6449. Cambridge, MA: National Bureau of EconomicResearch.

Citro, C.F. & Hanushek, E.A. (Eds.). (1991). Improving information for social policy decisions:The uses of microsimulation modeling, Vol. 1. Washington, DC: National Academy Press.

Gallagher, L.J., Gallagher, M., Perese, K., Schreiber, S., & Watson, K. (1998). One year afterfederal welfare reform: A description of state Temporary Assistance for Needy Families (TANF)decisions as of October 1997. Occasional Paper Number 6. Assessing the New FederalismProject. Washington, DC: The Urban Institute.

Greenberg, D.H., Long, D., Meyer, D., Michalopoulos, C., & Robins, P.K. (1995a). Usingmicrosimulation to help design pilot demonstrations: An illustration from the CanadianSelf-Sufficiency Project. Evaluation Review 19(6):687–706.

Greenberg, D.H., Michalopoulos, C., Robins, P.K., & Wood, R. (1995b). Making work pay forwelfare recipients. Contemporary Economic Policy 13(3):39–52.

Lin, W., Robins, P.K., Card, D., Harknett, K., & Lui-Gurr, S. (1998). When financial incentivesencourage work: Complete 18-month findings from the self-sufficiency project. Ottawa: So-cial Research and Demonstration Corporation.

Michalopoulos, C. (1999). Comparison of simulated impacts of SSP and MFIP with effectsfrom evaluation (memorandum, May 14), Manpower Demonstration Research Corporation.

Michalopoulos, C., Robins, P.K., & Card, D. (1999). When financial incentives pay for them-selves: Early findings from the self-sufficiency project’s applicant study. Ottawa: Social Re-search and Demonstration Corporation.

Michalopoulos, C., Card, D., Gennetian, L.A., Robins, P.K., & Harknett, K. (2000). The self-sufficiency project at 36 Months: Effects of a financial work incentive on employment andincome. Ottawa: Social Research and Demonstration Corporation.

Miller, C., Knox, V., Auspos, P., Hunter-Manns, J.A., & Orenstein, A. (1997). Making welfarework and work pay: Implementation and 18-month impacts of the Minnesota Family Invest-ment Program. New York: Manpower Demonstration Research Corporation.

Moffitt, R. (1983). An economic model of welfare stigma. American Economic Review73(5):1023–1035.

Pavetti, L.A. (1995). Who is affected by time limits? In I. Sawhill (Ed.) Welfare reform: Ananalysis of the issues. Washington, DC: The Urban Institute Press.

Robins, P.K., Michalopoulos, C., & Pan, E. (2000). Financial incentives and welfare reform inthe United States. New York: Manpower Demonstration Research Corporation.

Scrivener, S., Hamilton, G., Farrell, M., Freedman, S., Friedlander, D., Mitchell, M., Nudelman,J., & Schwartz, C. (1998). Implementation, participation patterns, costs, and two-year im-pacts of the Portland, Oregon Welfare-to-Work Program. New York: Manpower Demonstra-tion Research Corporation.

United States General Accounting Office (1998). Welfare reform: States are restructuring pro-grams to reduce welfare dependence, GAO/HEHS-98-109. Washington, DC: U.S. Govern-ment Printing Office.

financial incentives and welfare reform in the united states

Documents