The Public Administration Case Against Participation Income

The Public Administration Case Against Participation Income 1 Jurgen De Wispelaere Department of Philosophy, Trinity College Dublin Email: jurgen.dewispelaere@tcd.ie Lindsay Stirton The Norwich Law School, University of East Anglia Email: l.stirton@uea.ac.uk [Under review, please do not quote or cite without permission] Abstract – Anthony Atkinson’s proposal for a participation income (PI) has been acclaimed by social philosophers and policy experts as a workable compromise given the problems besetting an unconditional basic income (UBI). While some see PI as the first step towards a fully unconditional scheme, others regard PI as superior to UBI on ethical grounds as well as in terms of political feasibility. Both of these views disregard the administrative complications associated with introducing a broad participation requirement into welfare entitlements. In this paper we identify three essential administrative tasks that any welfare scheme must perform — establishing criteria of entitlement, determining compliance with those criteria, and allocating benefits to qualified beneficiaries — and show that PI performs poorly in terms of all three tasks. This suggests what we call the trilemma of participation income: only at significant costs to administrators and welfare clients does the scheme retain its apparent ability to satisfy the requirements of both activation and universal approaches to welfare. Consequently, the main apparent strength of PI — its capacity to unite different factions within the basic income debate — is shown to be illusory precisely because competing factions strongly prefer different resolutions of the trilemma. This, we argue, has far-reaching implications for the political strategy of basic income advocates as well as the wider debate on universal welfare reform. 1. INTRODUCTION This article challenges the idea that the modern welfare state would be improved by the introduction of a participation income (PI) to accommodate the twin objectives of universal social protection and broad social participation. Many scholars of the modern This draft was written partly while Lindsay Stirton was a Visiting Fellow in the Regulatory Institutions Network, Research School of Social Sciences, Australian National University. Earlier drafts were presented at the Social and Political Theory Research Seminar, Australian National University, 1 November 2005 and the 5th US Basic Income Guarantee Network Conference, Philadelphia, 24-26 February 2006. We are grateful to audiences at both events for comments, and to Geoffrey Brennan, Richard Caputo, David Casassas, John Ferejohn, Bob Goodin, Michael Harker, Michael Lewis, José Antonio Noguera, Claus Offe, Arvind Thattai, Yannick Vanderborght and Almaz Zelleke for helpful suggestions. The usual disclaimer applies. 1 1 welfare state agree with Goodin (2000) that the crumbling of its traditional pillars limits the ability of welfare policies to achieve a range of often contradictory objectives, but disagree as to which policies might be introduced in their place. In particular, a fault line separates those favoring increased emphasis on universal mechanisms, such as unconditional basic income (UBI) or its variants, and those who favor increased targeting of benefits through a host of conditional and activating measures commonly captured under the label workfare.2 At first sight PI appears extraordinarily well-suited to bridge the divide between “universalists” and “selectivists” in welfare reform. Like UBI, PI grants each adult citizen an individual right to a secure basic income irrespective of living arrangement or alternative sources of income or wealth (Atkinson 1993, 1996, 1998). However, unlike UBI, PI insists that recipients satisfy a broad participation requirement, consisting of a range of socially useful activities such as caring for an elderly relative, volunteering in a neighborhood project or engaging in a human capital enhancing activity. PI is also conceptually distinct from workfare because it expands the notion of social participation well beyond the narrow notion of labor market participation. Nevertheless, PI retains a strong element of conditionality attached to the receipt of welfare benefits. It is the latter feature that seemingly allows PI to bridge the positions of those who believe universal social protection and inclusion are the key objectives of the modern welfare state and those who think its primary task ought to be the increased activation of adult citizens.3 In this article we take issue with this view. Unlike those who regard PI as a policy solution that merges the best features of two worlds — a welfare scheme that appears genuinely inclusive and activating — we argue that PI has very little to offer either side of this debate. By critically assessing PI from the standpoint of administrative analysis and drawing evidence from experience of existing programs, we identify several concerns at the level of implementation which have largely been ignored in existing debates. Once these concerns are confronted, it can be seen that the apparent attractions of PI as a political compromise do not translate into a stable administrative solution to the competing priorities of those who support PI on different grounds. Inevitably, because our purpose is to assess an abstract policy proposal, rather than to evaluate an existing program, empirical evaluation of our argument is problematic. Our analysis is therefore less an exercise in hypothesis testing as one of hypothesis formulation.4 2 While workfare advocates occupy the mainstream in contemporary welfare policy, a substantial literature on basic income and related policy proposals has developed over the past two decades. See in particular Van Parijs (1992, 1995), Atkinson (1995), Fitzpatrick (1999), Ackerman and Alstott (1999), Standing (1999, 2002, 2005), Van der Veen and Groot (2000), Van Parijs, Cohen and Rogers (2001), Dowding, De Wispelaere and White (2003), White (2003b), Cunliffe and Erreygers (2004), Groot (2004), Widerquist, Lewis and Pressman (2005), Ackerman, Alstott and Van Parijs (2006) and Murray (2006). In June 2006 Basic Income Studies, the first academic journal on basic income published its inaugural issue (http://www.bepress.com/bis/). 3 Van Parijs (1996, 2004) proposes an unconditional basic income as a solution for simultaneously combating “exploitation” and “exclusion”, but most policy analysts insist a more activating approach is required to secure inclusion through labor market participation. 4 We claim support here from F. A. Hayek who argues (contra Popper) that applying theories to new questions is a central part of scientific inquiry, and one which is separate from that of testing theories. Hayek (1967: 6) argues that although it is desirable to check one’s conclusions against observations as far as possible, “[t]he conclusions which we 2 The first aim of the present paper is to demonstrate that, in its current underdeveloped state, PI does not constitute a workable proposal.5 An examination of the administrability of PI — that is, its capacity to be administered in a practical and efficient manner in accordance with its primary objectives and within existing constraints — reveals significant weaknesses which makes its endorsement by UBI supporters in particular premature, if not altogether gratuitous. While such a negative assessment primarily affects what we refer to as the first-best justification of PI, it also impacts importantly on the second-best case for PI. This brings us to the second goal of this paper. Comparing PI with both UBI and workfare from the point of view of administrative analysis casts doubt on PI’s ability to serve as a stable compromise between the competing camps of universalists and selectivists. The administrative instability of PI, we maintain, confronts would-be supporters of the proposals with the following trilemma of participation income: only by imposing substantial economic and moral costs on both welfare bureaucrats and clients in order to enforce a broadlydefined participation requirement effectively can PI satisfy the requirements of both a UBI and a workfare program. Avoiding the horn of bureaucratic cost might require the scheme to either compromise on weak enforcement of the participation requirement, thus rendering PI barely distinguishable from UBI, or designing the participation condition around selective eligibility criteria which effectively would mutate PI into workfare. And since advocates of PI hold both explicit and opposing views on how to resolve this trilemma, it is difficult to see how PI can attract sufficient support for an enduring coalition. In short, because administrative instability produces political instability, far from being a perfect compromise between proponents and opponents of universal welfare reform PI may well constitute the worst of all worlds. The implications of this conclusion for the broader basic income debate are considered in the final section of this article. 2. PARTICIPATION INCOME: AN ANALYTICAL OUTLINE PI is the brainchild of Anthony Atkinson (1993, 1996, 1998), an Oxford economist with a distinguished track record of research into poverty and social exclusion. Increasingly dissatisfied with the reliance upon means-tested benefits in the UK since the 1980s, Atkinson came to look sympathetically upon arguments in favor of universal grants allocated without a means-test. At the same time, he recognized that “it will be difficult to secure political support for a citizen’s income while it remains unconditional on labor market or other activity” (Atkinson, 1996: 67). The solution for Atkinson is a social assistance program that retains the politically salient notion of social participation while giving up on means-testing — thus participation income was born. The broad outline of PI is best described by quoting Atkinson at some length: can draw from a combination of well-established hypotheses will therefore be valuable though we may not be in a position to test them.” 5 Fairness requires us to mention that Atkinson is well aware of problems with PI that might not arise with UBI, such as some people failing to secure a PI where they would have received a UBI, or behavioral adjustments to the program (Atkinson, 1996: 69). 3 “In my proposal, the basic income would be paid conditional on participation. I should stress at once that this is not limited to labor market participation. While the qualifying conditions would include people working as an employee or self-employed, absent from work on grounds on sickness and injury, unable to work on grounds of disability and unemployed but available for work, it would also include people engaging in approved forms of education or training, caring for young, elderly or disabled dependents or undertaking approved forms of voluntary work, etc. The condition involves neither payment nor work; it is a wider definition of social contribution.” (Atkinson, 1996: 68-9) The precise nature of a PI depends crucially on the sort of activities included in the notion of social participation. It is perfectly possible for a PI to barely expand on existing labor market-oriented notions of economic participation6; alternatively, a PI might encompass such a wide range of social activities effectively rendering it virtually indistinguishable from a genuine UBI. While this flexibility offers valuable maneuvering room for policy entrepreneurs to construct an enacting coalition – along the lines of Gibson and Goodin’s (1999) “veil of vagueness” – it raises concerns about how the precise scope of participation is to be determined and by whom. A second key concern relates to the question of how compliance with such a requirement is to be enforced. As Loek Groot and Robert van der Veen (2000) suggest, a PI with a broad notion of social participation and lax enforcement will evidently bear a close resemblance to a UBI, while the same policy with a narrowly defined participation requirement and strict enforcement will in practice bear a close resemblance to a workfare program. This apparent malleability of PI might lead one to regard PI as either a mutated UBI proposal or an expanded workfare scheme. However, the apparent symmetry masks an important distinction between PI and UBI when compared to workfare schemes: where UBI claims to eradicate a host of social and economic ails solely by providing cash support, workfare and closely related activating welfare policies in addition employ two separate but complementary tools. In addition to using sanctioning devices to “push” workers into the labor market, workfare also purports to offer additional forms of support in negotiating barriers to employment, such as training, placement, assistance with transport or care problems, and so on. Granted, in practice the latter function often becomes crowded-out by a variety of other priorities (Handler 2004; Handler and Babcock 2006) but the analytical distinction nevertheless stands, suggesting a close alignment between PI and UBI. Atkinson’s proposal has gained almost universal, if at times implicit, support among basic income supporters as well as from a number of UBI’s more sympathetic critics, despite the fact that little work has been undertaken to flesh out Atkinson’s original proposal.7 On the other hand, perhaps it is precisely because both the concept and the policy details of PI remain vague that it has managed to retain the support of scholars and advocates with widely divergent perspectives. In order to advance the debate, however, it is now necessary to pierce the veil of vagueness and to consider those features of PI which make the scheme attractive to different positions within For instance, Socio-Economic Participation Income (SEPI) discussed by Jordan et al. (1999). Zelleke (2005: 640-641) is one of the few exceptions, explicitly denouncing PI for its selective treatment of various tax and benefit measures. 6 7 4 contemporary welfare debates. Analytically, we can discern three reasons for endorsing PI, each of which may take priority for some subset of those currently in favor of the proposal. First, one might endorse PI believing it to be a first-best welfare policy under contemporary conditions, for example because it allows us to combine generous income support with the freedom of recipients to engage in a wide range of socially useful activities. Thus, Stuart White’s relentless and sophisticated arguments against unconditional welfare, rooted in the view that unconditionality fails to satisfy the basic requirement of “fair reciprocity”, has led him to endorse PI on several occasions.8 In addition to such justice-based arguments, first-best proponents of PI may also advance an incentive argument in which PI encourages an optimal production of socially valuable activities without relying on overly coercive measures. Here, we could imagine PI fitting nicely with both feminist approaches to welfare reform (Robeyns 2000, McKay and Van Every 2000, McKay 2001, Pateman 2003, 2004) and with post-productivism in which narrow employment-driven policies are supplanted by schemes that induce the broad valuation of social activity, including care work (Offe 1992, Fitzpatrick 1999, Standing 1999, 2002, Van der Veen and Groot, forthcoming). In practice, the justice-based and the incentive-based perspectives often coincide, but the arguments are sufficiently distinct to warrant separate mention.9 This approach differs markedly from the views of second-best adherents of PI, who (like Atkinson himself) favor radically unconditional policies but acknowledge that UBI may not be a feasible option, for example because it does not command sufficient political support. Many basic income supporters, including Philippe Van Parijs, Claus Offe, Robert Goodin, Brian Barry, and many others have subsequently adopted this position. However, second-best justifications for PI themselves come in two subtle but distinct variants. First, there are those who regard political feasibility in static terms and who — perhaps reluctantly — accept PI as a permanent settlement. This seems to be Atkinson’s position, for example. This position can be contrasted by those who see PI as a first step towards a more radically unconditional scheme. This more dynamic view assumes that once a PI is in place, an important benchmark in favor of universalizing welfare is established; subsequent strategy would then be directed towards adjusting the scheme along various parameters — such as increasing the level of the grant to render it more generous, widening the conditions attached to receipt to render it more universal and encompassing, etc.10 — in order to move from a PI to a mature UBI. Recently, a number of basic income scholars have drawn attention to the many ways in which existing welfare and social support schemes can be made to mimic basic income in one of its versions. The three most common strategies are: (1) start with a 8 See especially White (2003b); an early view of this requirement can be found in Amy Gutmann & Dennis Thomson (1996). While White is arguably one of the most eloquent defenders of PI on first-best or justice-based grounds, he is certainly not the only one. 9 In the language of political philosophy, the justice-based argument is substantially (though not solely) backwardlooking, primarily concerned with ensuring that people are held responsible for past actions, while the incentive-based argument is genuinely forward-looking, aimed at ensuring that individuals bring about a sizeable social product. 10 For an analytical outline of the various dimensions across which broadly universalist policies can be varied, see De Wispelaere and Stirton 2004a. 5 partial income and then move towards a full UBI at the level of the poverty line; (2) start with a UBI in one sector or for part of the population, such as a basic pension or child benefit, and extend that gradually to the whole target group; or (3) introduce a combination of schemes that jointly operate as a UBI (Vanderborght 2000, 2005; Van Parijs 2001, 2004; Van der Veen and Groot 2000 and Standing 2005). Unfortunately, thus far we have not encountered a genuinely convincing strategy that indicates how to move from PI to a UBI; in the absence of such a strategy it remains equally plausible that opponents of universalism might use PI to benchmark a fairly narrow criterion of social participation and then progressively restrict access to eligibility — by imposing time constraints, for instance, as in the 1996 Clinton welfare reforms,11 — or effectively decreasing the level of support. In summary, the first-best justification sees PI as an ethical compromise between competing normative claims, while the second-best perspective regards PI essentially as a political compromise in a constrained policy environment. It can be argued that PI consequently emerges as a policy that has the capacity to unite a range of different groups and as such stands out as the welfare policy par excellence — capable not only of sustaining a coalition based around shared goals, but also allowing different groups to avoid having to chose for a more extreme policy, be it UBI or workfare. It is this apparent capacity to unite that in large part explains PI’s success amongst basic income advocates, and the apparent willingness of most advocates to lend their support without fully assessing how a PI would operate in practice. Once such an assessment is carried out, we believe, this apparent capacity to unite may be illusory. In our view, this illusion is largely maintained by avoiding discussing PI beyond its most basic contours. To illustrate, even sophisticated thinkers like Philippe Van Parijs (2004) only ever offer an impressionistic analysis of PI. Once such an exercise is carried out, as we propose to do in the remainder of this paper, the illusion rapidly disintegrates, and PI reveals itself as a risky and costly strategy. 3. THE ESSENTIAL FEATURES OF AN INCOME SUPPORT MECHANISM From the point of view of administrative analysis, there are three essential administrative tasks which any welfare scheme must perform. First, a scheme must establish operational criteria of eligibility which define the intended beneficiaries; second, it must identify those within the population who meet these criteria of eligibility, and distinguish them from those not eligible; third, it must transfer payments to eligible beneficiaries at the appropriate level, frequency, etc. Accordingly, the design of an income support mechanism involves, at a minimum, taking a position on how each of these tasks is to be performed. The following three sub-sections deal with each task in turn. We argue that PI performs poorly in terms of administering each of these three tasks. This is an important conclusion because the inability to perform one or more of these tasks as required by a particular scheme may undermine or negate the reasons for preferring that scheme in the first place. Even where such differences in administerability are not decisive, we may still find that how we resolve 11 The idea of a time-limited unconditional grant is proposed by White (2003a). 6 questions of administrative design has important consequences for the way the scheme may be expected to operate in practice, and thus for the conclusions we draw about the normative desirability of different schemes.12 A. STANDARDS CONFERRING ENTITLEMENT The first step in the design of any welfare benefit is to determine what criteria, if any, should govern eligibility for a grant, and how these criteria are best encapsulated in a set of rules or standards. Debates on welfare entitlement have, for the most part, focused on the first, normative part of this task, neglecting the second, technical-legal aspect.13 However, these two aspects cannot be completely divorced from one another, and the relative difficulty of reducing alternative criteria of eligibility into a set of suitably precise rules is relevant to the balance of considerations for and against different income support schemes. Indeed, Colin Diver (1983) has attributed the failure of many public policies to achieve their intended purposes to inappropriate degrees of rule-precision. Diver (1983: 67-68) identifies three separate dimensions of rule-precision. First, transparency is defined as the extent to which the language in which rules are formulated is clearly defined and has universally accepted meaning. Second, accessibility is the extent to which rules are readily applicable by those that are required to interpret them in concrete or “street-level” situations. The third dimension is the extent to which verbal formulations are congruent with the underlying policy objectives, so that compliance with the terms of the rules should lead to the outcomes consistent with the policy. A key part of Diver’s argument is that these three dimensions are not co-linear, and so policy-makers need to make compromises and trade-offs between the three dimensions, in order to produce effective rules. Applying Diver’s analysis to the problems of designing an effective PI scheme reveals the complex administrative judgment involved in implementing a PI scheme. As discussed in the previous section, the key feature that distinguishes PI from UBI on the one hand, and from workfare on the other, is the incorporation of a broad participation requirement. Applying Diver’s analysis to the problem of legislating Atkinson’s participation criterion into a set of rules, consider the following three alternative model standards (inspired by Diver 1983: 69): Model I: Any person working for at least ten hours per week in employment, selfemployment, for a charitable organization, or in study at university or college, or caring for dependent parents, children or spouses shall be entitled to receive a PI. Model II: Any person substantially engaged in an appropriate socially valued activity shall be entitled to receive a PI. Polities with limited administrative capacity may find such choices even more severely constrained. The importance of this point should not be underestimated, given that interest in basic income and related policies of late has been greatest in some less developed countries (Nattrass and Seekings 2005, Suplicy 2005). 13 Part of the problem may be that the two aspects of this question fall on different sides of a disciplinary divide. Whereas discussion of abstract criteria comes naturally to the political philosopher or student of public policy, the problem of reducing abstract criteria into concrete rules has traditionally been part of the subject matter of administrative lawyers and public administration scholars (Breyer 1982, Chapter 5; Black 1995; Baldwin 1990, 1995; also Hood 1986). 12 7 Model III: A person shall be eligible to receive a PI, provided he or she undertakes activities as prescribed in the following table. This is followed by a detailed table displaying various types of employment, positions within voluntary organizations, approved college and university courses, disabilities, care work activities, etc. Comparing these three alternatives, we observe that model I is clearly both transparent and accessible — it is relatively easy to know what it means, and whether it applies to any particular case, provided one is familiar with the relevant facts — but it performs poorly in terms of congruence with the underlying policy of promoting an inclusive conception of active citizenship as demanded by reciprocity theorists (Anderson 2001, Galston 2001, White 2003b). Specifically, the formulation is in certain respects “under-inclusive”: one can envisage many socially valuable activities that are not captured by the formulation, such as non-university-based education and other non-credited forms of training, or caring for dependents outside of the prescribed relationship. It may also be “over-inclusive” in other respects: it is debatable whether all charitable organizations contribute unambiguously to the public good.14 Moreover, in the absence of clear standards we can expect a proliferation of charitable organizations, many of which may be set up with the explicit objective of exploiting the scheme. This is an example of the phenomenon of “creative compliance”, discussed below. The fact that a PI might end up inadvertently supporting negatively or neutrally valued activities, while at the same time failing to support activities that are generally considered beneficial to society at large, is a topic of some concern amongst advocates as well as adversaries of the scheme.15 It is also a concern well-documented in the literature on welfare reform, where it is referred to as goal congruence. Goal congruence requires that policy systems allow political agents who decide on formal policy goals to adequately transmit these goals throughout the system, and provide street-level case workers with both resources and incentives to implement policy accordingly. Recent research argues that congruence between formal policy goals and street-level, operational goals is a prerequisite for effective welfare administration (Meyers et al. 2001). However, goal congruence is immensely difficult to achieve when either policy goals or the operational policy environment (delivery systems) are complex; complex policy goals invite conflict and ambiguity regarding the proper interpretation and ranking of goals, while complex administrative environments introduce conflicting incentives into the equation. Unsurprisingly, Marcia Meyers and her colleagues show that where both policy goals and the administrative environment are complex, a decoupling of formal and substantive policy goals often occurs. One strategy to avoid such decoupling would be to start with clearly defined policy goals (Riccucci et al. 2004). This is the approach taken in Model II, which directly elucidates the policy intention behind the rule. However, Model II manages to avoid 14 As an example, for a period one of us played in an amateur orchestra, established as a charitable organization which has never, in its forty-year existence, given a single public performance. 15 The problem is a political one in the sense that the content of the participation requirement ideally should form part of a political consensus. In practice, however, it is likely that policy-makers and bureaucrats decide where to draw boundaries. 8 this problem only by offering a vague, all-encompassing criterion of participation, leaving it open to conflicting interpretations by clients, welfare officers, policy-makers and even the general public, and thus scoring poorly in terms of accessibility. Of course, for various political reasons, often “elected officials are pre-disposed to providing vague or ambiguous policy directives” (Meyers et al. 2001: 166). This is particularly the case when policy goals are contested, providing incentives to decision-makers to shift responsibility to administrative agencies and case workers (Brodkin 1990, 1997). While some, as noted, regard a political predisposition towards vagueness an opportunity, it may also encourage strategies of blame-avoidance (Weaver 1986). Model III, in turn, may get round this problem, providing both high accessibility and congruence, provided policy-makers can foresee all the socially valuable activities of which individuals can conceive, but is likely to be extremely complex; so these gains come at the cost of poor transparency.16 Of course, we could attempt to get round this particular problem by combining different versions of the models. For example, we could combine Model II with a set of guidelines similar to Model III, leaving residual discretion of some official to “include” other cases. However, this may combine the weaknesses of different approaches as well as their strengths, and may furthermore come at the cost of considerable complexity. While Models II and III may avoid problems of goal congruence, Model II confronts another central problem in welfare administration, namely that it confers excessively broad discretion on welfare administrations to determine unilaterally the eligibility of clients (Diller 2000). Poor accessibility exacerbates this problem because unless there is a firmly established consensus among the “interpretative community” on which activities are socially valued, highly unlikely in modern plural societies (Black 1995), key decisions are typically passed through the system to street-level case workers. Similarly, Model III may be regarded as highly congruent with the underlying policy goal of instituting a broad participation criterion, but avoids these problems only by creating another: poor transparency increases the de facto discretionary power of street-level administrators because of imperfect and asymmetrical knowledge of the regulations amongst and across welfare workers and clients (see Baldwin, 1995: pp. 1533). The literature on the implementation of welfare policies demonstrates that case workers are frequently overwhelmed by forms, rules and regulations and are often forced to take on responsibilities for which they are insufficiently trained, leaving little or no time for actual engagement with clients (Sandfort 2000; Handler and Hasenfeld, forthcoming).17 Consequently, non-transparent rules have been shown in some cases to lead administrators to base decisions on cues and heuristics that may themselves be incompatible with underlying policy concerns, or be otherwise arbitrary. This has been shown to lead to the creaming of easy cases and the dumping of problem clients, the selective provision of key information to clients and the arbitrary use of compliance and sanctioning (Handler 2004). Paradoxically, attempts to curtail welfare workers’ The manifold presumptions that drive current social policy is often regarded as a key advantage of universal measures which, in the words of Goodin (1992), are “minimally presumptuous”. 17 To illustrate, Jody Sandfort notes how the mere list of standardized forms to be used in the processing and follow-up of welfare claimants adds up to an astounding 33 pages (Sandfort 2000: 735)! 16 9 discretion by expanding rules, regulations and other attempts at routinizing welfare implementation have proven counter-productive because they incite case workers to ignore cumbersome procedures and to use their discretion to achieve what they perceive as the primary goal, typically case-load reduction (Meyers et al. 1998; Handler 2004; Hander and Hasenfeld, forthcoming). “Because staff recognize the limited ability of the programs to move all clients into work, they develop alternative definitions of organizational success” (Sandfort 2000: 739). How does PI compare to rival schemes in this respect? It is often assumed that UBI side-steps the problem of rule formulation, since eligibility is nominally unconditional. This point, while not entirely without force, should not be overstated: even in its most inclusive formulation such a grant must restrict entitlement to some subset of the inhabitants of one country, or perhaps of some supra-national region such as the EU (Van Parijs 1995; Vanderborght and Van Parijs 2001; also Howard 2006). In addition there will be rule specifications pertaining to the level of the grant, its uniformity, the time scale at which it operates and so on, all of which involve considerably more regulation than is appreciated by many UBI supporters (De Wispelaere and Stirton 2004a). Nevertheless, reducing a proposal for a radical UBI scheme into a set of transparent, accessible and congruent rules is likely to be a relatively straightforward matter, based on comparably well-defined criteria compared with other forms of welfare grants, including PI. At the other extreme, in our view workfare schemes may also be superior to PI in terms of ease-of-administration even though they incorporate strict conditions. There are a number of reasons why workfare might outperform PI in terms of defining a standard conferring entitlement. First, the sheer volume of activities captured by a broad participation requirement compared to the eligibility requirements under a workfare program is likely to impact on the performance of each of these two respective policies. Second, precisely because of its formality a formal work condition is likely to be more easily expressed as a rule. Third, because work conditions are more easily embedded in existing practices and labor market institutions, they are more likely to engender accessible and transparent regulations. This again contrasts favorably with PI which demands the institution of newly valued activities; or at least activities that are “new” from a rule perspective. Of course, research has shown that workfare in practice faces many complicated problems and in fact manages to gather support despite its record of obtaining goals and objectives (Meyers et al. 1998; Sandfort 2000; Handler 2004; Handler and Hasenfeld, forthcoming). Nevertheless, from both a political and an administrative perspective, it makes a considerable difference whether the task of conferring standards of entitlement is initiated from a clear and coherent perspective or one that is by its very nature covers a diverse and heterogeneous range of activities. Irrespective of actual outcomes, workfare and PI do not appear to start on a similar footing in this respect. In short, all welfare schemes face the sort of hard choices and trade-offs inherent in Diver’s optimal precision calculus, but different schemes score differently in how well they negotiate these trade-offs. The basic formulation of PI as a non-means-tested grant subject to a broad participation requirement ignores the distinctive ways in which the latter can be operationalized. Bringing the operational dimension to the foreground, as 10 we have done in this section, demonstrates PI faces considerably greater problems in implementation compared with both UBI and workfare. B. IDENTIFICATION AND MONITORING OF BENEFICIARIES The task of determining whether individuals satisfy criteria of eligibility, and whether beneficiaries have complied (or continue to comply) with conditions attached to the receipt of a grant is a familiar problem in welfare administration. From the perspective of administrative analysis, this question focuses attention on the informational demands associated with different schemes, as well as the potential for strategic behavior on the part of potential beneficiaries, who may evade conditions of entitlement or engage in creative compliance (McBarnett & Wheelan 1991). As we shall see, these issues are closely intertwined with the issues of rule design addressed in the previous section: rules that perform poorly along any of the three dimensions of rule precision discussed in the previous section typically give rise to further monitoring and enforcement problems.18 This section addresses these issues by examining the features of benefit schemes that lend themselves to the effective identification of beneficiaries, by assessing how far the alternative formulations of the participation requirement discussed in the previous section can readily embody these features, and how the challenges associated with monitoring and enforcement of a participation requirement compares with equivalent implementation issues posed by UBI and workfare, respectively. In several writings Christopher Hood (1985; 1986: 74-81; 1994: Ch. 7) has discussed in detail the features that make a tax more or less administerable for a given sociotechnical environment. The features identified by Hood can also be applied to the crediting of income support grants. First, Hood (1994: 118) coins the term standard clarity to denote “the property of being relatable to values that can be ascertained relatively economically and ‘objectively’.” While this is arguably just an enforcement perspective on Diver’s criterion of accessibility discussed in the previous section, it reinforces the point that where standards cannot readily be applied to particular circumstances, determining compliance is likely to be costly and difficult — potentially leading, as Diver suggests, to policy failure. Next, Hood (1994: 118; 1986: 76-77) argues that ease of administration requires that the scheme’s targets must be cadasterable; that is, they must possess “the property of being applicable to a readily identifiable population of [‘creditable’] units, if possible from sources that are available in advance and cannot be easily contaminated by evasive action.” This also implies that individual circumstances that are likely to change frequently — for example, the amount of hours worked out by casual employees — constitute a poor basis for entitlement because they lead to a rapidly changing and therefore unstable target population. This is not merely a problem for welfare administrators, but also for clients themselves, since they often fail to understand when they are in violation of an eligibility constraint (Handler and Hasenfeld, forthcoming; Handler and Babcock, 2006)19 It is worth noting that one of the key reasons identified in the regulation literature for regulatory failure is a lack of sensitivity on the part of rule-makers to the challenges of enforcement and compliance (Baldwin 1990). 19 Conversely, as often pointed out by basic income supporters, rule complexity prevents many eligible individuals from understanding when they qualify for a particular benefit. 18 11 Building on the argument of the previous section, our hypothetical implementation of a participation requirement again helps to illustrate how alternative rule descriptions can lead to divergent outcomes by virtue of their effects on determining compliance. 20 In Model I and Model III, decisions about what constitutes active citizenship are made in advance; determining compliance requires only knowledge of the facts of individual circumstances. While information about employment can be gleaned from payroll information and enrolment in education can be assessed from matriculation records, there are no such comprehensive or reliable sources of information about the distribution and extent of care work in the home. Effective monitoring of care work would indeed require overly intrusive forms of monitoring; alternatively, some form of self-certification might be used, but this would in turn be susceptible to manipulation by those not meeting the participation requirement, in particular when not all eligible clients might be aware of their entitlement. Model II, faces all of these problems and more. In particular, this model performs poorly in terms of standard clarity. In the absence of a strong interpretative consensus, basing entitlement on an abstract definition of a “socially valued activity” places a broad discretion on administrators to decide whether different activities qualify. As well as raising concerns about due process, such as what (if any) appeals procedures should be made available to enable those who are denied a grant to challenge such a decision (Danz, 2000; Lens and Vorsanger 2005), this may also be a recipe for inconsistent application—an acute concern, especially where different agencies are involved (Stoker and Wilson 1998; Meyers et al. 2000). There is, however, at least one respect in which Model II outperforms its competitors: it is less susceptible to the form of manipulation termed creative compliance in the legal literature; that is, compliance with the technical requirements of a rule or standard in such a way as to undermine the purpose of the rule (McBarnett & Whelan, 1991). Model I, for instance, might encourage enrolment at a university, leading to a proliferation of what former UK Higher Education Minister Margaret Hodge has called “Mickey Mouse courses” — a B.Sc. in Surfing, anyone? Tackling such opportunistic exploitation through further detailed rules along the lines of Model III, which might specify “approved” university courses, in turn requires further detailed monitoring to allow administrators to distinguish between those courses which are deemed valuable and those which are not. By avoiding formalism, Model II at least potentially permits administrators to respond flexibly to such attempted exploitation—albeit at the price of placing substantial discretionary authority in the hands of administrators. Consideration of the task of identification of beneficiaries and monitoring compliance also gives rise to questions as to what (if any) sanctioning mechanisms should complement information-gathering. When clients fail to abide by conditions attached to the receipt of a grant they normally face some form of sanction — typically in the form of partial or complete withdrawal of the benefit itself (Handler 2004). The 20 This is a separate concern from the distinction made by Stoker and Wilson (1998) between client-based and organization-based forms of claims verification. While this distinction draws attention to the important point that clients too need to be able to verify their eligibility to particular benefits, reliance on clients’ assessments to determine compliance with conditions of entitlement could be expected to result in a systematic bias in favor of overpayment. For an analytical argument see Prendergast (forthcoming). 12 link between monitoring and sanctioning appears straightforward, and this has arguably had the effect that strategies of moral persuasion to induce behavioral compliance and change have been relatively undervalued (Mead 1996; Hasenfeld and Weaver 1996; Riccio and Hasenfeld 1996). However, recent research casts doubt on various assumptions implicit in the design and operation of common approaches to sanctioning welfare recipients (Wilson et al. 1999; Hasenfeld et al. 2004; Handler and Hasenfeld, forthcoming). First, it remains unclear whether sanctioning is mandated by moral or utilitarian arguments (Hasenfeld et al. 2004; Handler and Hasenfeld, forthcoming). In addition, it is questionable whether sanctioning really produces desired outcomes; that is, do non-compliant welfare recipients really learn a moral lesson or are sanctions merely pre-selecting (“creaming”) less demanding clients where welfare systems face too many pressures (Wilson et al. 1999), without taking account of the manifold “barriers to compliance” that may intersect with clients’ behavioral disposition (Hasenfeld et al. 2004)? Again, comparing the analysis of PI with that of UBI and workfare, respectively, is instructive. The unconditional nature of UBI again implies that monitoring and enforcement is strictly speaking a non-issue. If we grant each and every individual citizen a grant unconditionally, there is neither need for, nor a point to, monitoring recipients of the grant. There might nonetheless be other reasons for retaining some form of monitoring, if only to ensure that all citizens end up receiving the UBI they are entitled to: in the absence of existing mechanisms to monitor citizens generally it is impossible to assess their take-up (a point we elaborate in the next section). Although monitoring might constitute an important built-in safety mechanism, the need for a monitoring apparatus is nonetheless significantly reduced, if not rendered superfluous, in a basic income welfare state. Similarly, UBI would reverse the sanctioning philosophy implicit in current welfare programs. What about workfare? Workfare programs stipulate a narrow conception of active citizenship based on labor market categories and institutions. Most of these categories already feature prominent in various cadasters and this would suggest they are more easily to monitor, at least in theory. In practice, evidence shows that monitoring remains one of the most challenging issues in the administration of workfare (Handler and Babcock, 2006; Handler and Hasenfeld, forthcoming). The evidence suggests that monitoring is both costly and error-prone, with case workers often making decisions on the basis of outdated, inaccurate or incomplete information. This in turn has potentially disastrous implications concerning the subsequent assessment of eligibility, sanctioning, and so forth. The situation becomes more complicated if we move from pure “people-processing” technology to the “people-changing” requirements inherent in workfare’s ambitious activation goals (Hasenfeld and Weaver 1996). As has been reported throughout the relevant literature, agencies embracing an eligibilitycompliance culture find it hard simultaneously to endorse notions of self-sufficiency (Bane and Elwood 1994). We do not wish to downplay the problems associated with current welfare-towork administration, but the point remains that where workfare faces difficulties, monitoring and enforcement of a broad participation requirements becomes practically unworkable because of its inherent ambiguities. One important variable to consider is 13 the default-position when there is implicit uncertainty about the outcome of a verification procedure: do we assume that a client is eligible unless proof to the contrary is offered, or do we shift the burden of proof entirely on the client (Stoker and Wilson 1998)? Another variable is a measure of proportionality between “infraction” and “sanction”. There is ample evidence that the bulk of sanctioned TANF clients in the US, some of who lose their entire entitlement, are effectively being sanctioned for missing an appointment or not properly filling out a form (Cherlin et al. 2002). Clearly both concerns bear a close relation to the distinction between narrow and broad participation requirements and their strict or lax enforcement, discussed in a previous section. In addition, PI also faces a major problem of inter-agency coordination since the broadening of approved activities also implies an implosion of agencies that need to cooperate to produce the required information (cf. Stoker and Wilson 1998). Finally, the inherent ambiguities of PI will also interact with the doctrine of sanctioning: the uncertainty about when clients deserve to be sanctioned under PI may introduce further arbitrary and unequal treatment.21 The task of determining the target population eligible for a particular entitlement and its subsequent monitoring and, if required, sanctioning, proves to be one of the major challenges for any welfare system. In this section we argued that, for reasons closely related to its failures in the specification of rules or standards of eligibility, PI fares badly compared to UBI and workfare. When properly considered, the difficulties in identifying and monitoring the relevant target population appear rather insurmountable. Again, we must conclude that looked at from a public administration perspective, we have little reason to embrace PI. C. CREDITING PAYMENTS TO BENEFICIARIES The third essential feature of a welfare scheme is the means to credit payments to those who have been identified as the proper beneficiaries of the grant. Discussion of alternative welfare proposals often takes this aspect of a welfare scheme for granted, but its importance should not be underestimated. A set of standards may confer eligibility on a set of beneficiaries with an appropriate degree of precision, and those who meet eligibility criteria may have been identified, yet unless the design of welfare scheme includes the effective means of crediting payments to beneficiaries it must be judged a failure. Further, vulnerability to fraud or opportunism may undermine the legitimacy as well as the program efficiency of a scheme. Again, Christopher Hood’s work on tax administration offers theoretical inspiration for our attempt to identify the place of the administrative factor in debates on welfare reform. Hood’s concept of the conduitability of a tax, defined as “the property of being assessable and collectable through a relatively small number of surveillable channels or ‘bottlenecks’ at which oversight can be applied” (Hood, 1994: 118) has its counterpart in the administration of welfare grants. In the present context, we can think of conduitability in terms of the degree of fit or complementarity between At least one study shows that what clients learn from sanctioning interventions is not any deeper moral lesson on what constitutes appropriate behavior, but simple lessons on how to properly cooperate with welfare workers who determine eligibility. “In the end, paternalist reform seems to be a lesson about power, not responsibility” (Wilson et al. 1999; see also Hasenfeld and Weaver 1996). 21 14 the way in which beneficiaries are defined and identified, and the means by which payments are credited. Relatedly, the arrangements for paying a benefit should ideally lend themselves to “cross-checking” entitlement to a benefit (see Hood, 1986: 78-79). There are two ways in which conduitability, defined in this way, may be considered important to the administerability of a welfare scheme. First, the monitoring of payments is essential to ensuring that benefits are received by all those who are entitled, combating “false negatives”. Second, by incorporating as far as possible robust oversight mechanisms into the arrangements for payment of a benefit, the design of a scheme plays an important part in ensuring that payments are only made to those who meet — and continue to fulfill — conditions of entitlement, eradicating “false positives”. To the extent that these two desiderata have conflicting implications for the design of a grant, addressing the question of the conduitability calls for skilful administrative judgment. Broadly speaking, there are two strategies for ensuring the effective crediting payments to claimants. The first is to rely on as few alternate payment channels as are sufficient to reach the target population, in the extreme case, making use of a single but universally accessible payment mechanism. It is often suggested that a scheme with a smaller number of surveillable channels is to be preferred on administrative grounds because it lacks the complications arising from connecting up multiple devices. For example, in the UK, the National Audit Office (NAO) has argued with reference to the frequently maligned Child Support Agency, that “the parallel running of separate systems makes the customer interface more complex” (NAO 2005: ¶2.29). However, we need to be careful with such cursory assessments since the specific design of payment schemes matters a great deal. Consider, for instance, the institution of a single payment channel with heavy monitoring, such as where a case-officer presents a cheque after a face-to-face encounter. This design may perform well in terms of further ensuring compliance with the general eligibility requirements, but is likely to prove costly, especially if a scheme is intended to benefit a large population of beneficiaries. A single payment scheme that eschews monitoring might perform better in terms of reaching eligible claimants, but at the same time the absence of monitoring may make a scheme vulnerable to false positives.22 The alternative strategy is to positively embrace redundancy and to credit payments through multiple, overlapping mechanisms. While this foregoes the ease-ofoversight provided by a single channel, overlapping devices may make it easier to reach the full spectrum of beneficiaries. Overlapping systems have a decisive advantage in cases where the very circumstances that make a particular beneficiary group difficult to reach make such a group a priority in terms of welfare policy. Homelessness, for example, or exclusion from the formal labor market, and hence from firms’ payroll systems, are typical examples. Even so, the advantages of the multiplechannels approach may not always be so decisive in practice. Complementary payments may suffer from common mode failure: for example, both electronic payment of funds and account payee cheques require that the intended recipient has access to a In addition, the absence of any monitoring implies we often do not know precisely how comprehensive a delivery system is. 22 15 bank account. And as the NAO report (2005) mentioned above suggests, any advantages of redundancy comes at the cost of complexity, potentially leading to mistakes on the part of claimants and benefit staff and to increased fraud and barriers to accessing particular benefits. In summary, effective design of payment systems achieves an appropriate balance between targeting the full spectrum of beneficiaries and delivering benefits through a small number of surveillable channels. Administrative failure can lead to unequal access to benefits and to irregular take-up rates in diverse ways. Two general observations follow if we apply this formulation of the concept of conduitability to the design of welfare schemes. First, crediting payments cannot be considered merely as an afterthought to other aspects of the design of welfare schemes. There is a central administrative dimension to securing effective channeling of payments to beneficiaries which goes beyond simply instituting the proper technological environment, as suggested by some scholars and practitioners. Specifically, it is a mistake to assume that in societies with well-established payment systems, existing payment systems will render this aspect of welfare design unproblematic. 23 Second, as with our discussion of identifying and monitoring beneficiaries above, the requirement that a scheme must make — and monitor — transfers to a sub-set of beneficiaries who fall outside institutionalized payment and oversight mechanisms challenges the extent to which any scheme will fully achieve its intended results. In the remainder of this section we consider the performance of PI in comparison to its rivals in relation to the criterion of conduitability. Given UBI’s radically inclusive philosophy, ensuring payments reach all those found eligible can be considered to be the more important aspect of conduitability for this scheme. UBI takes this approach to its extreme by eschewing conditionality and altogether dispensing with the need for payment systems to act as a check on entitlement. In practical terms, discussion of UBI usually propose one of the following payment schemes: UBI can be implemented through integration into existing tax-credit schemes, turning it de facto into a refundable tax credit; alternatively, UBI proponents occasionally advance the idea of a “basic income debit card” from which recipients draw their grant as and when it is convenient to them, just like a regular cash withdrawal using a bank card. Both of these approaches reflect the implicit – and sometimes explicit – assumption among the basic income community that removing the “gatekeeper” role of welfare case-workers is desirable both because it would increase claimants’ autonomy and improve take-up rates. One potential shortcoming of both tax-benefit integration and the basic income debit card may be that these approaches offer little or no means of checking whether people have actually received their grants, a point that so far seems to have escaped the radar of UBI advocates.24 23 Consider, for example, the finding of the UK Office of Fair Trading: “In Birmingham in the early 1990s, five of the 39 electoral wards had no bank or building society branches, while a further six wards had only one. But 28% of the city’s population lived in the relevant areas—including some of its poorest inhabitants.” (Office of Fair Trading, 1999: 22-23). One can imagine how these issues multiply in rural areas or in countries with less developed electronic payment systems. 24 It is worth noting a crucial difference between the two types of delivery mechanism. While a refundable tax-credit scheme retains a measure of monitoring – typically in terms of means testing – both the extent and the approach to monitoring payments differ considerably from standard welfare programs. Nevertheless a basic capacity to monitor 16 Workfare approaches, on the other hand, eschew payment systems which dispense with oversight altogether. Quite the contrary, workfare relies on extensive intervention from case-workers. This approach provides for face-to-face oversight but faces considerable variable costs which increase as the number of claimants increases. While these might be considered to be tolerable as long as benefits are targeted more narrowly at those in formal employment and training, this approach may well be regarded as infeasible as the target population expands. An obvious solution to this problem is restricting access to benefits and pro-actively reducing the number of beneficiaries; however this solution significantly shifts the goals of workfare (Handler 2004; Handler and Babcock 2006). A second mechanism in favor of workfare is that the aforementioned focus of eligibility on training and formal employment provides a host of institutionally embedded oversight mechanisms, including payroll systems and university matriculation records, which in theory allow for effective cross-checking – assuming the necessary administrative resources are kept in place. Both mechanisms make it possible for workfare to properly monitor the delivery of payment systems, even if conduitability under workfare is achieved by restricting eligibility largely those who are part of the formal employment sector. UBI and workfare clearly take opposite approaches to the issue of conduitability: in the case of workfare by restricting the class of beneficiaries to those who can be reached by existing well-defined channels; in the case of UBI by radically minimizing the oversight requirements associated with ensuring that conditions of eligibility are met. Each of these has problems but both may be regarded as broadly viable strategies. In contrast, the problem of PI is that it possesses the strengths of neither of these two schemes. Since eligibility is conditional on the satisfaction of a participation requirement, PI does not share UBI’s ability to dispense with oversight mechanisms required to prevent payment without the fulfillment of conditions. Combining either tax-benefit integration or the basic income debit card idea with monitoring and oversight of conditions associated with PI seems to invite the risks associated with complexity, identified most recently by the NAO.25 Similarly, because PI is intended to be more broadly-targeted than workfare, the variable costs of face-to-face monitoring by case-workers may become prohibitive. Furthermore, the intention that the participation requirement should be satisfied by informal activities (care work, etc.) limits the extent to which control can be exercised through surveillable payment channels, such as employer payroll systems or payment to students at the point of college registration. Returning to a point made earlier, PI may be regarded as having one modest advantage over UBI. Because the scheme is targeted at a narrower set of beneficiaries it may retain some of the capacity for ex post monitoring of benefit receipt. However, depending on how broad the participation criterion is interpreted, this capacity may be very modest indeed, effectively leaving a considerable subclass of the intended group remains in place. In contrast, while the basic income debit card mechanism in theory also allows for the monitoring of payment into the account, this scheme retains little capacity to monitor whether the benefit actually reaches the intended beneficiary. 25 Specifically, complexity poses risks for the implementation of bespoke IT systems for welfare payments that may defeat attempts at automation and require substantial case-worker intervention (NAO, 2005 paras. 3.24-3.31). 17 unprotected – in particular those who find themselves near the informal end of the participation spectrum. This violates the goal of providing equal security across the population, one of the core concerns of a wide range of policy analysts (Standing 1999, 2002). But this modest advantage of PI is in turn trumped by workfare simply because the latter set of beneficiaries will always be more restricted than that associated with PI, which implies both a smaller and more homogeneous set. It seems then that, as with the previous two administrative tasks, PI performs worse than either UBI or workfare in terms of the crediting of payments to beneficiaries. Admittedly, compared to the previous two tasks, these arguments may not seem decisive to the same extent. What is important to carry the overall argument of this paper is that administrative analysis of the task of channeling payments does not offer countervailing arguments in favor of PI that might outweigh the disadvantages of PI discussed in the previous sections. 4. THE TRILEMMA OF PARTICIPATION INCOME From the analysis of the previous two sections, it follows that PI proponents face a trilemma of participation income. The first horn of the trilemma consists of the requirement that PI must remain substantively inclusive if it is to satisfy the concerns of basic income advocates. The second horn is constituted by the equally important requirement, insisted upon this time by reciprocity theorists and “soft” workfare proponents, that recipients must satisfy a genuine participation requirement, which in turn demands that any PI scheme must be capable of differentiating between those who fulfill their social obligations, however defined, and those who do not. Finally, the third horn is made up of the concerns pertaining to the economic and human costs related to administrability, discussed in detail in the previous sections. The trilemma arises because PI can only avoid two of three horns simultaneously, inevitably getting pinned on a third. This leaves us with three possible routes to move the policy forward. This remainder of this section discusses each of these three routes in turn, and makes the argument that each of these routes will appeal, if at all, only to a small subset of those attracted at a general level to PI, with obvious negative consequences for the political feasibility of PI. A. THE IRONCLAD ADMINISTRATION STRATEGY A first possible resolution of the trilemma would be to reject any compromise on easeof-administration considerations, taking the view that where these conflict with a participation requirement the former should always give way. In this way, PI might hope to retain the support of both UBI and workfare supporters, as well as those who support PI on first-best grounds. While this constitutes a coherent position which some PI proponents implicitly seem willing to adopt, there are nevertheless important consequences which adherents to this position may be reluctant to accept. In particular, the ironclad administration strategy sacrifices some of the chief advantages that are claimed for non-means-tested benefits, in particular the ability to use bureaucracy sparingly. Christopher Hood (1983: pp 142) has identified two senses in which one may seek to use bureaucracy sparingly: first, by economizing on the 18 administrative resources or on the administrative capacity of the state required in the execution of welfare policy, and second, by minimizing what Adam Smith calls the “trouble, vexation and oppression” visited upon the citizenry by the welfare administrators. Concerns with using bureaucracy sparingly in the former sense are captured by the notion of program efficiency in welfare analysis (Goodin, Headey et al. 1999). The latter sense is captured in concerns about welfare administrations constantly interfering with individuals’ lives — an all too familiar issue in welfare policy, often cited as one of the key advantages of universal and unconditional measures (Van Parijs 1992, 2004, Offe 2005, Standing 1999, 2002, Fitzpatrick 1999; also Wolff 1999).26 To the extent that one or both of these two senses of using bureaucracy sparingly may be considered material to the feasibility and desirability of any income support scheme, considerations of administerability are obviously material to the choice between rival income support mechanisms. A PI scheme which fails to use bureaucracy sparingly in the first sense may achieve its objectives only at considerable administrative cost, satisfying demands for improved target efficiency — if at all — only at the cost of poor program efficiency (Goodin, Headey et al. 1999). Poor program efficiency in turn is likely to impact negatively on the capacity of PI to attract a necessary enacting coalition; it would also impact negatively on the wider public opinion because one of the core arguments of UBI, namely, low-cost administration, is no longer attainable. When a proposed scheme also fails to use bureaucracy sparingly in Hood’s second sense, the administrative demands of effectively enforcing a broad participation requirement are borne not only by administrators, but also directly by welfare recipients, and even by the general population. Intrusive inspections, threats and punitive sanctions, and the like may all prove necessary to discourage cheating where compliance cannot be easily and objectively determined. This is an important consideration with clear impact on the freedom and respect considerations that often inform universal approaches to income support (Wolff 1999, Eyal 2006). Even where UBI supporters do not endorse a robust libertarian or left-libertarian position (Van Parijs 1995), lack of freedom and respect still constitute a personal cost that counts against the scheme. In short, negotiating the “trilemma” by embracing ironclad administration comes at considerable cost, with workfare proponents typically lamenting the increased resource cost of welfare administration and UBI proponents primarily denouncing the interference cost associated with increased monitoring and disciplining. B. THE SOFT WORKFARE STRATEGY A second approach would be to deliberately formulate the participation requirement in an under-inclusive fashion; effectively drafting a participation requirement in such a way as to include only those activities that are easy to specify precisely and to monitor Incidentally, while UBI is often justified because of its optimal efficiency in regard to both senses of using bureaucracy sparingly, we urge some caution in this assessment: UBI has the inherent flaw of not only making part of the administrative apparatus redundant, but as such possibly also breaking down administrative redundancies and even administrative capacity in general. The former we believe to be a good thing, but the latter may cause problems that UBI proponents have thus far failed to appreciate. See our analysis in De Wispelaere and Stirton 2004b. 26 19 efficiently, and including — as far as possible — those beneficiaries to whom it is administratively simple to make payments. As discussed previously, the latter strategy of creaming “suitable” welfare applicants is familiar in the workfare literature (Handler 2004; Wilson et al. 1999). While this approach is less attractive in principle than the broad participation requirement envisaged by Atkinson and others, it might be regarded as a viable compromise between the theoretical advantages of PI and the requirements of ease-ofadministration outlined above. At the same time it clearly narrows the gap between PI and selective workfare schemes, for in practice the scheme will necessarily focus on those social categories or activities that already feature prominently in current administrative systems. In other words, here PI exhibits the conservative tendency of reinforcing existing institutionalized practices and activities at the cost of an expanding valuation of alternative social activities, such as voluntary work or care work. And here lies the problem: such a solution would be unacceptable to those UBI proponents – notably feminists and post-productivists – who have made the recognition and valuation of non-traditional forms of work a key objective of welfare reform.27 Moving away from the broad acceptance of non-traditional forms of work is likely to lose the support of political factions representing these points of view, or possibly even incur active resistance. In the extreme, there is even a danger that PI would prove less inclusive than workfare because, as mentioned before, it lacks the myriad of support programs pertaining to training, in-work assistance, or placement that constitute an integral part of workfare (at least in theory).28 Of course PI advocates may well acknowledge the need to supplement basic security with additional activation support. But such “PIPlus” programs face two important constraints. First, faced with permanent austerity and the sort of political constraints that we witness in relation to workfare programs in the US and UK (Pierson 2001), one must retain a healthy skepticism about the likelihood of such programs being instituted at a large scale, with appropriate levels of funding, if they are to be designed and implemented on top of a PI scheme. But this is not all. Even if we could resolve such concerns as they pertain to labor market activation, part of the attraction of PI is that it expands the notion of social participation to include a wider range of activities. It remains unclear whether a “PI-Plus” proposal would offer support for individuals to engage in such activities over and beyond the grant they receive; and if so, what form such support will take.29 Failure to offer assistance across the whole range of social participation appears to reinforce the current employment-focused paradigm. Taken these difficulties together it becomes clear that moving ahead with PI on this first approach raises serious questions in terms of whether PI will still command the sort of support amongst UBI supporters it enjoys 27 More generally, Guy Standing (2006: 180) argues that “a subversive politics should be about wrestling control over time for the ‘dispossessed’, and it should recognize that such control is the essence of real security.” 28 PI remains crucially underspecified in this respect: it is compatible with but does not explicitly require the installment of programs offering support to individuals facing difficulties entering the labor market. 29 Perhaps we are too pessimistic here. Both Claus Offe and Steve Shafarman have independently pressed the point upon us that civil society might take over at this point. Even if we were to accept this argument – and there are reasons to be cautious – the asymmetric valuation of formal employment compared to alternative economic or social activities implied in such a scheme is still worth noting. 20 at present time – or, indeed, that it should. C. THE LAX ENFORCEMENT STRATEGY A third resolution of the trilemma suggests an alternative compromise between the incorporation of a participation requirement and ease-of-administration by adhering to a broad specification of the participation requirement, accepting all the time that effective monitoring of compliance with this requirement is going to be problematic – the welfare state equivalent of de Toqueville’s “rigid rules and lax enforcement”. If the previous approach narrows the gap between PI and workfare, the present strategy all but obliterates any distinction between PI and UBI. Instituting an overly broad PI implies that a participation requirement de facto only exists as a symbolic commitment to eliminating free-riders. This would more than likely be acceptable to those who would ideally adopt a UBI and only supported PI as a politically feasible, second-best option. In addition, it is likely to satisfy political entrepreneurs who oppose UBI on purely strategic grounds, and who regard this second route as a unique opportunity to save on administrative resources, opposing big government and the like, while retaining a firm and public stand on the need for social participation in return of state support.30 However, this route out of the trilemma would hardly be acceptable to those who favor PI as a first-best policy. Symbolism aside, lax enforcement of a broadly specified notion of social participation does not engender commitment to promoting reciprocity or a stand against free-riding. Of course, once PI no longer serves the need of one of the core partners of a fragile compromise they will be looking elsewhere for other, more suitable policies with robust activation components. In addition to more “genuine” workfare schemes, research of the past 10 years has opened up some interesting possibilities. One option being investigated is the use of time-limited policies. Stuart White (2003b), for instance, suggests combining unconditional, time-limited policies with conditional unlimited schemes. Another possibility is the recent surge of interest in (conditional) asset-based welfare schemes in both US and UK (Ackerman and Alstott 1999; Nissan and LeGrand 2003; Goodin 2003; Paxton et al. 2006). 5. IMPLICATIONS FOR THE WIDER BASIC INCOME AGENDA Our analysis of the difficulties in administering a PI scheme, and of the resulting trilemma, points to several potentially far-reaching implications for the basic income debate. First, it calls into question the assumption that a PI can be easily slotted into existing social security and welfare assistance arrangements. Our discussion has emphasized the challenges posed in the formulation, application and fulfillment of standards instituting a broad participation requirement, the defining feature of PI, as part of welfare policy. Because both UBI and workfare negotiate these difficulties, albeit in very different ways, these alternative visions of welfare may be regarded as better able to meet these challenges, which are faced to a greater or lesser extent by all Charles Murray’s recent proposal for introducing a “Plan”, his version of what essentially amounts to a neo-liberal UBI scheme for the US, would fit this scenario (Murray 2006). 30 21 income support mechanisms. Our analysis of the trilemma points to various resolutions, each of these has important consequences for first- and second-best supporters of PI alike. Supporters of PI on first-best grounds must squarely confront the issue of whether the advantages they claim for PI are sufficient to justify the high costs to administrators and welfare clients of ironclad administration; if not, are these same advantages sufficiently realized either by a scheme which is close to workfare or one which is barely distinguishable from UBI? For supporters of PI on second-best grounds, the trilemma is still more uncomfortable: because support for PI on second-best grounds represents a political compromise between different positions, the risk arises that such a compromise would evaporate once the tough choices in the design and implementation of a scheme are confronted, specifically about how each of the three tasks discussed in the previous section are to be accomplished. Except under the somewhat implausible assumption that supporters of universal and active welfare could compromise on ironclad administration, these two positions opt for incompatible routes out of the trilemma. In short, at some point in the implementation process the political compromise is expected to break down. These considerations lead to a second significant implication for basic income debates. Implicit in many discussions of universal welfare reform is the assumption that in key respects PI closely approximates a UBI. The trilemma suggests, however, that depending on the fine-grained details of any proposed scheme, PI may just as likely come to resemble workfare schemes (though without the benefit of the activation mechanisms that in theory are part of the workfare program). This undermines the dynamic as well as the static version of the second-best case for PI. In a recent essay Brian Barry writes: “I guess something like participation income might be necessary politically to get a basic income introduced, but that the expense and intrusiveness of administering it (as well as lending itself so easily to fraud) would lead either to abandoning the whole experiment or moving to an unconditional basic income” (Barry 2001: 66). There is, however, a third plausible alternative, namely that given the problems identified by Barry, criticism of such a scheme would lead to a tightening of both the formulation and enforcement of the participation criterion. The resulting outcome is essentially a workfare scheme. It is worth pointing to two more general implications of our analysis for the basic income agenda. First, UBI advocates need to consider seriously whether they ought to lend their support to PI. Our analysis shows that for UBI supporters, endorsement of PI comes at considerable risk. We caution against those such as Barry, Offe, Vanderborght or Van Parijs who apparently see the political and administrative instability of PI as an opportunity for a strategy of introducing basic income by stealth. In addition to the risk that these features might lead to the abandonment of experiments with universalism, there is a further and equally plausible possibility that the fine-tuning of such schemes might open the door for workfarist schemes to which they are implacably opposed. Furthermore, our analysis shows how this “backdoor” strategy might plausibly require universalists to sacrifice one of the greatest advantages of their preferred policies, namely ease-of-administration and non-intrusiveness. 22 The final lesson of our analysis is that supporters of basic income and related policies cannot ignore the administrative realities of welfare reform. Abstract discussion of the pros and cons of a broad participation requirement or other features of a universal grant proposals can be extremely limited, potentially obliterating crucial conditions on which the desirability or otherwise of particular proposals depends. While falling short of proper empirical study and comparison, we have shown how administrative analysis generates important insights beyond abstract theorizing, which UBI supporters ignore at their peril. Much progress can be made from applying the insights gleaned from studies of existing welfare policies in order to anticipate the likely effects of proposed schemes. In this particular case, such analysis clearly suggests important weaknesses of PI as a welfare policy. While none of our conclusions are necessarily decisive, together they amount to a strong public administration case against participation income. REFERENCES Ackerman, Bruce and Anne Alstott 1999. The Stakeholder Society. New Haven: Yale University Press. Ackerman, Bruce, Anne Alstott and Philippe Van Parijs eds. 2006. Redesigning Distribution: Basic Income and Stakeholder Grants as Cornerstones for an Egalitarian Capitalism. London/New York: Verso. Anderson, Elisabeth 2001. “Optional Freedoms” in What’s Wrong With a Free Lunch?, edited by Philippe Van Parijs, Joshua Cohen and Joel Rogers. Boston: Beacon Press. Atkinson, Anthony 1993. Beveridge, the National Minimum, and Its Future in a European Context. Working Paper WSP/85. London: London School of Economics, STICERD. Atkinson, Anthony 1995. Public Economics in Action: The Basic Income/Flat Tax Proposal. Oxford: Oxford University Press. Atkinson, Anthony 1996. “The Case for a Participation Income”, The Political Quarterly 67 (1): 6770. Atkinson, Anthony 1998. Poverty in Europe. Oxford: Blackwell. Baldwin, Robert 1995. Rules and Government, Oxford, Clarendon Press. Baldwin, Robert 1990. “Why Rules Don’t Work”, Modern Law Review 53 (3): 321-337. Bane, Mary Jo and David Elwood 1994. Welfare Realities. Cambridge, Mass.: Harvard University Press. Barry, Brian 2001. “UBI and the Work Ethics” in What’s Wrong With a Free Lunch?, edited Philippe Van Parijs, Joshua Cohen and Joel Rogers. Boston: Beacon Press. Black, Julia 1995. “Which Arrow? Rule Type and Regulatory Policy”, Public Law: 94-117 Breyer, Stephen 1982. Regulation and its Reform, Cambridge, Mass.: Harvard University Press. Brodkin Evelyn 1990. “Implementation as Policy Politics” in Implementation and the Policy Process, edited by D. Palumbo and D. Calista. New York: Greenwood Press. Brodkin Evelyn 1997. “Inside the Welfare Contract: Discretion and Accountability in State Welfare Administration”, Social Service Review 71 (1): 1-33. Cherlin, Andrew, Karen Bogen, James Quane and Linda Burton 2002. “Operating within the Rules: Welfare Recipients’ Experiences with Sanctions and Case Closings”, Social Service Review 76: 387-405. Cunliffe, John and Guido Erreygers eds. 2004. The Origins of Universal Grants. An Anthology of Historical Writings on Basic Capital and Basic Income. Basingstoke: Palgrave. Danz, Sheri 2000. “A Non-Public Forum or a Brutal Bureaucracy? Advocates’ Claims of Access 23 to Welfare Center Waiting Rooms”, New York University Law Review 75 (4): 1004-1044. De Wispelaere, Jurgen and Lindsay Stirton (2004a) “The Many Faces of Universal Basic Income”, The Political Quarterly 75 (3), pp. 266-274. De Wispelaere, Jurgen and Lindsay Stirton (2004b) “The Administration of Universal Welfare” paper presented at the 10th Basic Income European Network Congress, Barcelona. Diller, Matthew 2000. “The Revolution in Welfare Administration: Rules, Discretion, and Entrepreneurial Government”, New York University Law Review 75 (5): 1121-1220. Diver, Colin 1983. “The Optimal Precision of Administrative Rules”, Yale Law Journal 93: 65-109. Dowding, Keith, Jurgen De Wispelaere and Stuart White eds. 2003. The Ethics of Stakeholding. Basingstoke: Palgrave. Ewalt, Jo Ann and Edward Tenning 2004. “Administration, Governance and Policy Tools in Welfare Policy Implementation”, Public Administration Review 64 (4): 449-462. Eyal, Nir 2006. “Basic Income Guarantee and the Social Basis of Self-Respect” paper presented at the 5th USBIG conference, Philadelphia, 24-26 February 2006. Fitzpatrick, Tony 1999. Freedom and Security: An Introduction to the Basic Income Debate. Basingstoke: Palgrave. Galston, William 2001. “What About Reciprocity?” in What’s Wrong With a Free Lunch?, edited by Philippe Van Parijs, Joshua Cohen and Joel Rogers. Boston: Beacon Press. Gibson, Diane and Robert Goodin 1999. “The Veil of Vagueness: A Model of Institutional Design” in Organizational and Institutional Factors in Political Life: Essays in Honour of Johan P. Olsen, edited by Morton Egeberg and Per Laegreid. Oslo: Scandinavian University Press. Goodin, Robert 1992. “Towards a Minimally Presumptuous Social Policy” in Arguing for Basic Income, edited by Philippe Van Parijs. London/New York: Verso. Goodin, Robert 2000. “Crumbling Pillars: Social Security Futures”, The Political Quarterly 71 (2): 144-150. Goodin, Robert 2003. “Sneaking Up on Stakeholding” in The Ethics of Stakeholding, edited by Keith Dowding, Jurgen De Wispelaere and Stuart White. Basingstoke: Palgrave. Goodin, Robert, Bruce Headey, Ruud Muffles and Henk-Jan Dirven 1999. The Real Worlds of Welfare Capitalism. Cambridge University Press. Groot, Loek 2004. Basic Income, Compensatory Justice and Unemployment. Dordrecht: Kluwer Academic Publishers. Groot, Loek and Robert van der Veen 2000. “How Attractive is a Basic Income for European Welfare States?” in Basic Income on the Agenda, edited by Robert van der Veen and Loek Groot eds. Amsterdam University Press. Groot, Loek and Robert van der Veen (forthcoming) “Post-Productivism and Welfare States: A Comparative Analysis” British Journal of Political Science. Gutmann, Amy & Dennis Thomson 1996. Democracy and Disagreement. Cambridge, Mass.: Belknap Press of Harvard University Press. Handler, Joel 2003. “Social Citizenship and Workfare in the US and Western Europe: From Status to Contract”, Journal of European Social Policy 13 (3): 229-243. Handler, Joel 2004. Social Citizenship and Workfare in the United States and Western Europe: The Paradox of Inclusion. Cambridge University Press. Handler, Joel and Amanda Sheela Babcock 2006. “The Failure of Workfare: Another Reason for a Basic Income Guarantee” Basic Income Studies 1 (1): 1-22. Handler, Joel and Yeshekel Hasenfeld 1997. We the Poor People: Work, Poverty and Welfare. New Haven: Yale University Press. Handler, Joel and Yeshekel Hasenfeld (forthcoming) Blame Welfare, Ignore Poverty and Inequality. Cambrige: Cambridge University Press. Hasenfeld, Yeshekel and Dale Weaver 1996. “Enforcement, Compliance and Disputes in 24 Welfare-to-Work Programs”, Social Service Review 70: 235-256. Hasenfeld, Yeshekel, Toorjo Ghose and Kandyce Larson 2004. “The Logic of Sanctioning Welfare Recipients: An Empirical Assessment”, Social Service Review 78: 304-319. Hayek, F.A. 1967. “Degrees of Explanation” in Studies in Philosophy, Politics and Economics, London: Routledge and Kegan Paul Hood, Christopher 1983. The Tools of Government, Basingstoke: Macmillan. Hood, Christopher 1986. Administrative Analysis. Brighton: Wheatsheaf. Hood, Christopher 1994. Explaining Economic Policy Reversals, Buckingham, Open University Press Howard, Michael 2006. “Basic Income and Migration – A Moral Dilemma?”, Basic Income Studies 1 (1): 1-22. Jordan, Bill, Phil Agulnik, Duncan Burbridge and Stuart Duffin. 1999. Stumbling Towards Basic Income: The Prospect for Tax-Benefit Integration. London, Citizens’ Income Trust. King, Desmond 1995. Actively Seeking Work? University of Chicago Press. Lens, Vicky and Susan Vorsanger 2005. “Complaining After Claiming: Fair Hearings After Welfare Reform”, Social Service Review 29: 430-453. McBarnett, Doreen & Wheelan, Christopher. 1991. “The Elusive Spirit of the Law: Formalism and the Struggle for Legal Control”, Modern Law Review 54 (6): 848-873. McKay, Ailsa 2001. “Rethinking Work and Income Maintenance Policy: Promoting Gender Equality through a Citizens Basic Income”, Feminist Economics 7 (1): 97-118. McKay, Ailsa and Jo Van Every 2000. “Gender, Family and Income Maintenance; A Feminist Case for Citizens' Basic Income”, Social Politics 7 (2): 266-288. Mead, Lawrence 1986. Beyond Entitlement. New York: Free Press. Meyers, Marcia K. 2002. “How Welfare Offices Undermine Welfare Reform” in Making Work Pay: America After Welfare, edited by Robert Kuttner. New York: The New Press. Meyers, Marcia, Bonnie Glaser and Karin MacDonald 1998. “On the Frontlines of Welfare Delivery: Are Workers Implementing Policy Reform?”, Journal of Policy Analysis and Management 17 (1): 1-22. Meyers, Marcia, Norma Riccucci and Irene Lurie 2001. “Achieving Goal Congruence in Complex Environments: The Case of Welfare Reform”, Journal of Public Administration Research and Theory 11 (2): 165-201. Murray, Charles 2006. In Our Hands. Washington D.C.: IAE Press. National Audit Office 2005. Department of Work and Pensions: Dealing With the Complexity of the Benefits System. London: The Stationery Office. Nattrass, Nicoli and Jeremy Seekings 2005. “The Political Economy of the Basic Income Grant in South Africa” in The Ethics and Economics of the Basic Income Guarantee , edited by Karl Widerquist, Michael Lewis and Steven Pressman eds.. Aldershot: Ashgate. Nissan, David and Julian LeGrand 2003. “A Capital Idea: Helping the Young to Help Themselves” in The Ethics of Stakeholding, edited by Keith Dowding, Jurgen De Wispelaere and Stuart White eds. Basingstoke: Palgrave. Offe, Claus 1992. “A Non-Productivist Design for Social Policies” in Arguing for Basic Income. Edited by Van Parijs ed. London/New York: Verso. Offe, Claus 1996. “The Acceptance and Legitimacy of Strategic Options in Social Policy” in Modernity and the State: East, West. Cambridge: Polity Press Offe, Claus 2001. “Pathways from Here” in What’s Wrong With a Free Lunch?, edited by Philippe Van Parijs, Joshua Cohen and Joel Rogers. Boston: Beacon Press. Offe, Claus 2005. “Wasteful Welfare Transfers” in Promoting Income Security as a Right: Europe and North-America, edited by Guy Standing. London: Anthem Press. 25 Office of Fair Trading 1999. Vulnerable Consumers of Financial Services: The Report of the Director General's Inquiry. London: The Office of Fair Trading. Orbach, Barak 2006. “Unwelcome Benefits: Why Welfare Beneficiaries Reject Government Aid”, Law and Inequality 24 (107): 107-150. Pateman, Carole 2003. “Freedom and Democratization: Why Basic Income Is to Be Preferred to Basic Capital” in The Ethics of Stakeholding, edited by Keith Dowding, Jurgen De Wispelaere and Stuart White eds. Basingstoke: Palgrave. Pateman, Carole 2004. “Democratizing Citizenship: Some Advantages of a Basic Income”, Politics and Society 32 (1): 89-106. Paxton, Will, Stuart White and Dominic Maxwell 2006. The Citizen’s Stake. Bristol: The Policy Press. Pierson, Paul ed. 2001. The New Politics of Welfare. Oxford: Oxford University Press. Prendergast, Canice (forthcoming) “The Motivation and Bias of Bureaucrats”, American Economic Review. Riccio, James and Yeshekel Hasenfeld 1996. “Enforcing a Participation Mandate in a Welfare-toWork Program”, Social Service Review 70: 516-542 Riccucci, Norma, Marcia Meyers, Irene Lurie and Jun Seop Han 2004. “The Implementation of Welfare Reform: The Role of Public Managers in Front-Line Practices”, Public Administration Review 64 (4): 438-448. Robeyns, Ingrid 2000. “Hush Money or Emancipation Fee? A Gender Analysis of Basic Income” in Basic Income on the Agenda, edited by Robert van der Veen and Loek Groot. Amsterdam University Press. Sandfort, Jodi 2000. “Moving Beyond Discretion and Outcomes: Examining Public Management from the Front Lines of the Welfare System”, Journal of Public Administration Research and Theory 10 (4): 729-756. Standing, Guy 1999. Global Labor Flexibility. Seeking Distributive Justice. Basingstoke: Palgrave. Standing, Guy 2002. Beyond the New Paternalism. London/New York: Verso. Standing, Guy 2006. “CI, COAG and COG: A Comment on a Debate” in Redesigning Distrbution: Basic Income and Stakeholder Grants as Cornerstones for an Egalitarian Capitalism, edited by Bruce Ackerman, Anne Alstott and Philippe Van Parijs. London/New York: Verso. Standing, Guy ed. 2005. Promoting Income Security as a Right: Europe and North-America. London: Anthem Press. Stoker, Robert and Laurie Wilson 1998. “Verifying Compliance: Social Regulation and Welfare Reform”, Public Administration Review 58 (5): 395-405. Suplicy, Eduardo Matarazzo 2005. “The Approval of the Basic Income Guarantee in Brazil” in The Ethics and Economics of the Basic Income Guarantee, edited by Karl Widerquist, Michael Lewis and Steven Pressman. Aldershot: Ashgate. Tennings, Edward and Dale Krane 1994. “Coordination and Welfare Reform: The Quest for the Philosopher’s Stone” Public Administration Review 54 (4): 341-348. Van der Veen, Robert and Loek Groot eds. 2000. Basic Income on the Agenda. Amsterdam University Press. Van Parijs, Philippe 1992. “The Marriage of Justice and Efficiency” in Arguing for Basic Income, edited by Philippe Van Parijs. London/New York: Verso. Van Parijs, Philippe 1995. Real Freedom for All. Oxford: Oxford University Press. Van Parijs, Philippe 1996. “Basic Income and the Two Dilemmas of the Welfare State”, The Political Quarterly 67 (1): 63-66. Van Parijs, Philippe 2001. “Reply” in What’s Wrong With a Free Lunch?, edited by Philippe Van Parijs, Joshua Cohen and Joel Rogers. Boston: Beacon Press. Van Parijs, Philippe 2004. “Basic Income: A Simple and Powerful Idea for the Twenty-first 26 Century”, Politics and Society 32 (1): 7-39. Van Parijs, Philippe and Yannick Vanderborght 2001. “From Euro-stipendium to Euro-dividend. A Comment on Schmitter and Bauer”, Journal of European Social Policy 11: 342-46. Van Parijs, Philippe, Joshua Cohen and Joel Rogers eds. 2001. What’s Wrong With a Free Lunch? Boston: Beacon Press. Van Parijs, Philippe ed. 1992. Arguing for Basic Income: Ethical Foundations for a Radical Reform. London/New York: Verso. Vanderborght, Yannick 2000. “'The VIVANT Experiment in Belgium” in Basic Income on the Agenda, edited by Robert van der Veen and Loek Groot. Amsterdam University Press. Vanderborght, Yannick 2005. “The Basic Income Guarantee in Europe: The Belgian and Dutch Back Door Strategies” in The Ethics and Economics of the Basic Income Guarantee, edited by Karl Widerquist, Michael Lewis and Steven Pressman. Aldershot: Ashgate. Weaver, Kent 1986. “The Politics of Blame Avoidance”, Journal of Public Policy 6: 371-398. White, Stuart 2003a. The Civic Minimum. Oxford: Clarendon Press. White, Stuart 2003b. “Freedom, Reciprocity and the Citizen’s Stake” in The Ethics of Stakeholding, edited by Keith Dowding, Jurgen De Wispelaere and Stuart White. Basingstoke: Palgrave. Widerquist, Karl, Michael Anthony Lewis and Steven Pressman eds. 2005. The Ethics and Economics of the Basic Income Guarantee. Aldershot: Ashgate. Wilson, Laura, Robert Stoker and Dennis McGrath 1999. “Welfare Bureaus as Moral Tutors: What Do Clients Learn from Paternalistic Welfare Reforms”, Social Science Quarterly 80 (3): 473-486. Wolff, Jonathan 1999. “Fairness, Respect, and the Egalitarian Ethos”, Philosophy and Public Affairs 27: 97-122. Zelleke, Almaz 2005. “Basic Income in the United States: Redefining Citizenship in the Liberal State”, Review of Social Economy 63 (4): 633-648. 27