Revisited Tomer Blumkin, Yoram Margalioth and Efraim Sadka September - PowerPoint PPT Presentation

The Desirability of Workfare as a Welfare Ordeal – Revisited Tomer Blumkin, Yoram Margalioth and Efraim Sadka September 2012, Munich

Some Background � In a second best setting, an egalitarian government seeking to target benefits to the least well-off members of society is faced with a fundamental screening problem (mental problems and back pains). � The government often resorts to means-testing as a measure to enhance the target efficiency of its welfare system [alongside other measures such as ‘tagging’ (observable disability) and in-kind transfers (wheelchair)].

� In recent years, public outcry about welfare dependency and the allegedly weak incentives provided to claimants to leave welfare have caused many developed countries to introduce some form of work requirements (workfare) for welfare eligibility [such as the 1996 welfare reform in the US (PRWORA)]. � Work requirements serve a dual purpose: (i) enhancing the recipient’s job prospects via acquisition of skills; (ii) serving as a form of welfare ordeal, due to the opportunity cost associated with participation.

� This latter aspect has been emphasized by Besley and Coate (1992 and 1995) claiming that when the government goal is income maintenance, workfare would be a desirable supplement to income testing even when it does not affect the productivity of participants; however, when the government objective is utility maintenance (a Rawlsian welfare function) workfare becomes undesirable.

� The latter result has remained, by and large, unchallenged by the literature [Cuff (2000) is an exception, but is using a paternalistic objective; Brett (1998) and (2005) consider productive workfare]. � The goal of our paper is to revisit the Besley and Coate (1995) negative result, suggesting the undesirability of workfare under utility maintenance.

The Model � Individuals differ in (unobserved) innate earning ability, w . � Ability groups (high- and low-skill) of equal size normalized to unity. � CRS technology with perfect substitutability across skills. � Perfect competition. � Preferences represented by: � � � � � � (1) U ( c , l , d , ) c h ( l ) d

Remarks: (i) c and l denote consumption and labor, where h is increasing and convex. (ii) d is an indicator function which assumes that value of one if the individuals misreports his income in order to be eligible for benefits, and zero otherwise. The parameter � measures the (moral) cost of cheating and is distributed uniformly over the [ � for both types. interval 0 , ]

� The government is seeking to maintain some pre-specified minimal � ˆ � level of utility, u ˆ , satisfying: V u V , where V and V , denote the utility levels derived by a low-ability and a high-ability individuals, respectively, in the absence of government intervention. � The government offers a means-tested benefit schedule, where benefits are based on reported (not necessarily truthfully) levels of income (simplify to set focus on the role of workfare in addressing the issue of misreporting!)

The Government Program � The government is minimizing its costs: � � � � � � � � � � � � ( 1 / ) ( ) ( 1 / ) ( ) (2) E c y c y 0 0 Subject to: � � � � (3) c h ( y / w ) u ˆ , � � (4) c h ( y / w ) V � � � � � ( 5 ) c h ( y / w ) c h ( y / w )

� � � � � ( 6 ) c h ( y / w ) c h ( y / w ) � � � � � � ( 7 ) ( / ) , c y V c h y w 0 � � � � Where V max [ y h ( y / w )] . y Remarks: (i) Two income dependent benefits & voluntary participation w.l.o.g. The parameter � denotes the work requirement (in hours). (ii)

� is the (high-skill) individual indifferent b/w misreporting or (iii) 0 not. (iv) Constrains (3) and (5) will be binding (notice that in general FB may be feasible unlike optimal tax setting!). (v) We only care about high skill ‘cheaters’.

Results � Assuming no workfare in place ( � � 0 ) we solve for the optimal program. � Standard properties are obtained (efficiency at top and implicit positive MTR at bottom). � We then examine whether introducing some workfare requirement is � � � 0 � desirable: L / 0 � �

� W/o misreporting ( � � � ) we replicate the Besley and Coate (1995) negative result: workfare is undesirable as a supplement to means- testing. � With some misreporting we show that when (i) moral costs are small enough ( � small enough) and (ii) skill differences are large enough ( w / w large enough), workfare is desirable as a supplement to means- testing.

� In such a case, means-testing is replaced by workfare as a screening � � � � tool (we show that y / 0 ).

Rationale and Implications � When skill differences are large, a workfare requirement can serve as an efficient screening device, as it is much more costly for the high-ability individuals to take part in these programs relative to low-ability ones. However, workfare programs entail a large deadweight loss (in light of our assumption that they serve purely for screening purposes).

� The previous literature has shown that when income testing is allowed, there is, therefore, no desirable supplementary role played by workfare for utility maintenance purposes. � However, when agents can misreport their income thereby rendering the income testing less effective, workfare proves to be a useful policy tool as it mitigates the high-ability individuals’ incentive to misreport.

� Indeed, when misreporting is sufficiently manifest and the screening efficiency of workfare is significant, workfare becomes a desirable supplement to income testing. � The model has a testable implication (work in process): in countries with weak ‘civic mindedness’ (as measured by recent surveys, see Cahuch and Algan 2009 study of flexicurity) governments should replace means-testing with workfare (cross-country variation can be used for identification).