Much research has been done on institutional allocation of a scarce resource, starting with Guido Calabresi and Philip Bobbitt’s classic 1978 study of situations of “tragic choice”. Most studies have focused on the medical field (Elster and Herpin, 1992; Dalgalarrondo and Urfalino, 2000 [RFS English Selection 2002]) and distribution of goods that prolong or improve life (organs for transplant purposes, treatments for fatal diseases such as AIDS). Distribution of financial aid may seem to raise less dramatic questions in that the very lives of potential recipients are not at stake; it is more prosaically concerned with the ordinary exercise of social work (Messu, 1991). Reading applicants’ files, however–their budget calculations, duly substantiated by the requisite supporting documents, and their letters to the commission– we are justified in believing that poverty, too, puts an individual’s survival on the line, though in a different way. The following text, written after a year of unemployment by a 35-year-old Frenchman, divorced, father of five, recipient of the Revenu Minimum d’Insertion [1], which came to 2,138 francs [326 € ] while his monthly expenses for rent and alimony amounted to 2,345 francs, gives some idea of the pleas addressed to the prefectures :

This text may lead us to wonder how exceptional the writer’s situation is, and about the emotional effect he meant to have on his readers. For a more objective idea of applicants’ socio-economic reality, we should keep in mind the following set of figures : a third of unemployed persons receiving aid from the ASSEDICs, a fifth of RMI recipients, and a tenth of wage earners have fixed expenses above income. To use the expressive terminology of social action, they have a “negative remainder to live on” (Fassin and Defossez, 2000). The concept of tragic choice is thus clearly relevant, since this was a matter of allocating a limited-quantity good exclusively to persons whose material and even biological existence depend on it; as we shall see, the themes of hunger, cold, and hygiene were recurrent in applicants’ letters. Moreover, administrative use of the term “vital needs” for precisely those needs the state was accepting to meet through this fund, makes it clear that this particular type of government-distributed financial assistance involved the zoe, to use Giorgio Agamben’s word (1997), borrowed from Aristotle to designate “bare existence” as opposed to “political life” or bios. As the above plea, among many others, attests, not just personal dignity but bodily integrity were under threat here. Living under tight economic constraints has physical consequences, frequently expressed in applicants’ “motivation statements” with the word “survival”. By identifying the aid fund’s aim as handling “the most difficult situations”, ones that could not be resolved by “welfare minimums” (Join-Lambert, 1998), the government clearly expressed the idea that this was intervention at an infrapolitical level, in Hannah Arendt’s sense of the term (1958); ie, aid aimed at “maintaining subsistence conditions”. Indeed, several members of the FUS allocation commission for the department of Seine-Saint-Denis, who said it was a “discovery” to them that they would be dealing with people nearly unable to feed themselves, declared that the aid should enable applicants to “hold their head above water, at least for a time”. The decision to provide aid or not, and the sum chosen, thus took on, for them, the attributes of a tragic choice.

There are, however, two fundamental differences between the classic medical situations usually designated by this expression and the present cases of financial aid. Both concern how aid is distributed. First, financial aid is, in theory, infinitely divisible, which is obviously not the case with transplant organs or even drug treatments. At one extreme, we could imagine a strictly egalitarian mode that would involve allocating the same sum to all applicants, on the model of the “Christmas bonus” for jobless persons in the BouchesduRhône. In such a system, called for by certain social actors, all possibility of choice is eliminated. But this would have gone against the spirit of the measure as announced in the ministerial circular and fully internalized by commission members. The money was not to be distributed as a kind of complementary income to insufficient welfare minimums–that would seem to accredit the demands of advocacy groups for the unemployed. The purpose of this aid was instead to ensure that “the amount allocated is sufficient to provide aid proportionate to each identified situation of distress”. Clearly then, while financial aid is a rare resource that cannot be allotted to every applicant, it is also a resource of a very specific kind in that level can vary. The judgment procedure used concerned both eligibility (who receives) and amount (how much to give). As with a court jury summoned to determine accused’s guilt (or innocence) and fix the sentence, the commission’s decision was twofold. It thus constituted a kind of modulable choice.

Second, the persons who allocate aid were not merely informed of applicants’ objective situations, as when one has to decide who receives a transplant organ or a new drug treatment. Appeals for help were made to them in the form of subjective accounts. It is true that in medical situations, doctors may be informed of candidates’ personal characteristics, just as external pressure can be exerted in favor of a given case, but these are exceptional situations or at least presumed to be, whereas in distributing emergency financial aid, the applicant’s case in his or her own behalf was an essential part of the decision-making process. For the FUS, “the presentation of financial difficulties and motivation behind the request” were conceived as a means of circumventing the usual external “assessment by the social services”, for both the practical reason that it would save time and the ideological justification of avoiding the normalizing effect of the “social assessment” procedure. According to a person in charge of department-level social services, the fundamental principle of the procedure was “not to determine attribution on the basis of social service assessment, but rather to recognize individuals’ autonomy and their ability to clearly express their needs”. This approach put applicants in the position of having to justify their request in writing, giving them the feeling that the final decision would depend on that justification. This explains the fact that applicants overinvested the exercise of presenting their situation, misfortunes and suffering, leading one head of the departmental office of Health and Social Affairs (DDASS) to say, with regard to administrative agents doing applicant-file follow-up, that “the poor found themselves dealing with social workers who cried along with them”. Divulging situations of private distress in the public space of fund attribution commission meetings gave commission deliberations an emotional side, and the decision became a matter of what could be called a pathetic choice.

The singular twofold nature of such choice–the fact that it involves modulation and mobilizes the power of pathos– makes FUS attribution particularly relevant for understanding how “local justice” is administered (Elster, 1992). First, we can identify subtle normative and affective distinctions, made possible because the commission had the power to vary amount allocated. As mentioned, such variation is virtually impossible with the all-or-nothing law operative in medical situations. This opens up the field to near-ethnographic analysis of the practical functioning of a particular “sphere of justice” (Walzer, 1983). Moreover, identifying the modes by which scarce resources are divided up is all the more relevant in that we are dealing with a kind of casuistry that nonetheless does not preclude establishing certain rules for allocation. Deliberation on a case-by-case basis, as recommended in the ministerial circular, and the implementation of that recommendation in terms of the twofold principle of determining aid that was both adapted and proportionate to the case at hand, bring to light explicit and implicit criteria operative in commission deliberations. On the other hand, the commission’s way of proceding reflects a rejection of the opposition generally made between judgment and feeling, implicitly asserting that judgment can be fueled or oriented by feeling –as we know, feeling is often appealed to in court pleas for the defense. It thus suggests that “politics of pity” are not in fact exclusive of “politics of justice”, even though conceptually the two can be separated (Boltanski, 1993). The compulsory exercise of addressing one’s request to the state and having it read by representatives of the state in fact makes such decision-making both normative and affect-based, with compassionate attention to others’ misfortunes and troubles becoming mixed in with rational analysis of their needs. Applicant’s plea transmits his or her feelings to decision-maker, thereby significantly affecting the judgment procedure.

Studying how local justice is practiced in matters of social action or aid is often done in situ, by observing the functioning of commissions charged with either proposing activities to integrate people (Astier, 1991) or distributing public monies (Ogien, 1999). Such observation can provide a wealth of material about the deliberations that lead to decisions, enabling us to reconstitute conflicting arguments and even commentaries made during meetings. In the present case, however, methodology was determined only after the administrative work had been done. This study, conducted in the department of Seine-Saint-Denis, is thus retrospective. It is based on two types of material. 1) Interviews were conducted with actors directly or indirectly involved in implementing the FUS; these included allotment commission members but also administrative agents doing application follow-up investigation, local social workers, and representatives of associations for the unemployed. 2) I examined a sample of three hundred applicant files, randomly selected from among the twenty thousand received by the prefecture services. One hundred were from job-seeking recipients of unemployment insurance, one hundred from RMI recipients, and one hundred from wage earners (categories to be identified respectively as J-S, R, and S).

While interviews with actors provided useful clarifications on the principles applied in fund allocation, and interesting justifications of decisionmaking criteria, the analysis that follows is based primarily on qualitative and some quantitative study of the three hundred application files. These include basic socio-demographic information; a household budget including income, expenses, debts; a written argument justifying the request; and the ultimate decision of the commission; sum of aid attributed (if any); commission’s written response to the applicant, and, in many cases, handwritten annotations explaining or detailing the decision. In the spirit of local justice studies, which on this point may be contrasted to “global justice” theories (Schneider-Bunner, 1998), the purpose here is to grasp not merely the formal rules that decision-makers say they are following, but the actual procedures used in resource allocation. Accepting Harold Garfinkel’s hypothesis (1984, p. 114) that a trial juror “defines retrospectively the decisions that have been made”, which amounts to saying that “the outcome comes before the decision”, we can in this case set rules announced a priori against solutions determined a posteriori. I shall therefore first analyze the principles of justice identified, explained, and justified by allocation commission members, then compare these principles with the judgment practices that were actually used, as discerned through examination of applicant files.

First-order determinations for the FUS involved both national and local, ie departmental, levels. The sum of a billion francs was determined by the national government on the basis of the amount distributed by the ASSEDICs; that is, an average of 1.4 billion francs a year for the preceding three years. The figure was highly symbolic, clearly displaying government good will and aimed at containing the discontent of unemployed job-seekers, though modest compared to the 2 trillion annually allotted to “the nation’s social effort”. The Ministry of Employment and Solidarity was charged with distributing the sum among the departments on a prorata basis for number of long-term job seekers registered with the ANPE (Agence Nationale Pour l’Emploi) [national job placement service] and RMI recipients, these statuses being considered two clear indicators of poverty. The department of Seine-Saint-Denis received 30 million francs, a figure which may be compared to the seven billion distributed by the Family Allocations Bureau (Caisses d’Allocations Familiales : CAF) the same year.

Second-order determinations involved the departmental level only, though some general instructions were given in the January 12 ministerial circular and a follow-up note dated January 19. The prerogative of decision-making was officially granted to an allocation commission set up by the Mission d’Urgence Sociale under prefectoral authority, and delegated to the department director of Health and Social Affairs (Direction Départementale des Affaires Sanitaires et Sociales : DDASS). The commission was to be presided by the department director or another senior civil servant from the DDASS, either the deputy director or general inspector, and it consisted of representatives of the Department Council, the CAF and the ASSEDIC. In slightly under six months, the commission was to decide which of the twenty thousand applications would receive a favorable response. It met four or five times a week, with sessions often lasting half a day and mobilizing senior administrators from the participating institutions. It met a total of 115 times. In March and April it regularly examined 400 applications per session. The arrangement was quite remarkable in that the management of aid distribution, a task that usually falls to social workers or administrative agents, was here to be performed by a company of department-level institutional heads. For casebycase decision-making this was clearly not the ideal arrangement, since even with so many sessions programmed, the average time given to each application file amounted to one minute, even less in the heaviest work periods. At the end of the procedure, 72% of applications in Seine-Saint-Denis were accepted, somewhat less than the national average of 80%. I will be focusing here on second-order decision-making, especially the fugitive, immaterial intersecting of second- and third-type actors, ie, allotment commissioners and applicants. Though applicants were not physically present, their fate was to be decided on the basis of administrative applications they themselves had assembled, applications so precise and expressive as to give their authors a kind of virtual presence at commission deliberations.

The Ministry of Employment and Solidarity’s instruction to the commission to establish decision-making criteria seems both precise in terms of policy aims and technically vague. Two policy demands were clearly formulated. First, the commission’s response must be personalized; it must correspond to “individualized treatment of social situations”, as recommended by Pierre Rosanvallon (1995, p. 197, p. 210, and p. 217); the state should “take charge of individuals”, with the awareness that each is in a particular situation, instead of developing and implementing universal measures to be applied to entire population segments. In the same spirit, the case-by-case approach, which according to the circular should result in “the amount allocated [being] adapted to the situation and proportionate to each case of distress”, required that the commission examine each application in an unbiased manner. It is stressed that applicants need not be job-seekers; this meant that the fund extended to the “unstably employed and economically insecure” of the protest movement. Second, the problems must be urgent and the money granted must not be conceived as a mere complement to insufficient resources; this in accordance with an aid approach which, as Serge Paugam demonstrated (1993, p. 85 and p. 87), has been applied in France since the Revolution and explicitly affirms “the collectivity’s obligation to guarantee adequate means of existence” to the poor. However, the assistance should in no way appear as something owed to the applicant, a point expressed quite crudely by one commission president : “The idea behind the aid” was to “avoid [giving] the three thousand francs demanded by the jobless movement” and thereby to “break their activism”. By proposing emergency aid under these conditions, the idea was to distinguish national solidarity from trade union and advocacy group demands. This was to be neither an automatic “Christmas bonus” or an improvement on the ordinary “welfare minimums”. Instead, the FUS would give to each poor person in accordance with his or her situation.

Despite the intended break from habitual welfare procedures, the two governmental instructions show it to be perfectly consistent with “modern assistance”, which for Georg Simmel was “a unique public institution” (1998, p. 47 and p. 66). First, “it is completely personal; its entire purpose is to help meet particular needs, palliate particular situations of distress”. Second, “emphasis is on [giver’s] moral duty rather than [receiver’s] moral right, although the two are correlated”. This is what distinguishes solidarity in the form of welfare allocation from solidarity in the form of insurance, which pertains to law and is addressed to entire populations.

When it comes to applying this political philosophy of assisting the poor, tension develops between 1) the theoretical purpose of ensuring stateguaranteed fairness and 2) practical reasoning, which leads to instituting local implementation and thereby necessarily to unequal treatment. In this case, the choice of allocation criteria was left entirely up to the commissions; this amounted simply to applying the proximity principle for management of aid allocation and was also the consequence of the rule that aid must be individualized. A variety of procedures for drawing up applications and, most importantly, for attributing aid were put in place over the national territory, with each prefecture developing its own criteria.

One commission president’s comment clearly points to interinstitutional borrowing : “We got ideas from looking at aid applications submitted to the FAJ [Fonds d’Aide aux Jeunes; youth welfare program], the FSL [Fonds Solidarité Logement; housing assistance], Électricité de France-Gaz de France [state-owned energy companies], and for the Departmental Council’s monthly aid program for children, [since] nothing more closely resembles a financial aid application than another financial aid application”. Other institutions seem to have provided “inspiration” not only for what to require in applicants’ files but also for decision-making criteria. The expression “remainder to live on” is certainly more remarkable as a linguistic than technical invention, and perfectly represents the minimalism of administrative language. It should be noted that for nearly one in four persons in the population group targeted for state assistance, the “remainder to live on” is negative.

If existing social welfare instruments were used in calculating sum available per person per month after expenses, how was the 1,000-franc eligibility threshold decided on ? Two reasons were generally given for the choice. One was common usage : this amount is often used in assessment by the social services, though the reasoning behind it has never been made explicit. The other concerns the attraction of “round numbers” in establishing limits. In recalling their first working sessions, however, one commission president proposed a third interpretation : “There were many applications from immigrant workers living in hostels for single men. We decided early on to turn them all down. Most of these men received the RMI and, after rent, they still had about 1,000 to 1,500 francs to live on. We reasoned : they’ve got a roof over their heads, they can eat, they don’t fit the strict framework set out by the circular, whose purpose is not to combat poverty, but to respond to situations of distress. And often they were asking for aid for their families in the home country, whereas the fund was not established for that. So since we wanted to apply the same attitude toward everybody, that’s where the thousand francs came in. Yes, that’s where that rule came from.” Though it is impossible to determine which of these three explanations was the operative one, it can be pointed out that they do not all involve the same type of justification. The first two involve relatively bland generalities of the sort with which people can effectively, plausibly rationalize their decisions after-the-fact. The third, on the contrary, consists in singular information that delivers not the ultimate truth on how the eligibility threshold was determined but an unexpected truth about the practices of a state administration.

In fact, it is because the third explanation is singular that it has general scope. First, foreign inhabitants living in workers’ housing are understood to be exceptional, as their situation can in no way be considered part of the common lot of “grave distress”. Here the commission president was only corroborating a common representation of this population group, who are confined by immigration and housing policy to a kind of extra-territorial status that applies to both physical and mental space. Second, this justification maintains the appearance of equality for all when it comes to “national solidarity” by establishing a rule excluding immigrants which meanwhile masks that exclusion in the name of a principle that seems universal. In this paradoxical way, the logic of French republicanism could be reaffirmed. Of course, the only way that a contradiction can be avoided between attributing exceptional status to immigrant workers and affirming the principle of equality is by not taking into account the fact that immigrant workers living in housing for single men regularly send money to their families overseas, thereby lowering their own real income to below the fixed eligibility threshold. In an application from a father of six originally from Mali who did not specify that his children were not living with him–the commission inferred they were in Mali from the fact that the address given was a hostel– we read : “I don’t receive more than the RMI and I have a family to support. Since [19]95 I have not found work because of my age. I’m 57. Nonetheless, I have work 33 years in France. That’s why Im requesting exceptional financial aid, in order to balance my financial situation.” (R 71). Actually, the commission’s automatic rejection only applied to foreigners living in housing for single men. Many immigrant families living in the department did receive FUS aid, in some cases substantial amounts.

Whatever actually explains how and why this way of proceding was arrived at, it seems inconsistent with the ministerial instructions. First, in place of the casuistry-based charity recommended in the circular, the commission substituted an elementary arithmetic norm. Second, they twisted the objective of bringing relief to people in emergency situations by in fact having the aid function as a mere palliative for insufficient income. Two things, however, need to be said about this departure from officially announced policy. First, the commission did abide by the case-by-case examination instruction, despite institutionalized use of a norm in the form of a round figure. All application files were expertly evaluated by the commission and all applicant statements read. Since each situation was evaluated individually, it was possible to reverse the automatic decision to accept or reject an application that would have been made if the only criterion taken into account was whether applicant’s “remainder to live on” fell below or above the fixed threshold. The commission president who affirmed that “examination of applications was highly individualized” was therefore correct, even though the indicator chosen played a more important role in deliberations than had been planned. Denis Bouget’s nationwide synthetic study (1999, p. 43), based on departmental studies, confirms this observation. He also notes that “the allocation procedure was characterized by vagueness and imprecision”, which “led to valorizing the principle of strong personal conviction in final decisions”, and that there was therefore a tendency “to apply classic social-action and assistance logic based on evaluation of individual needs”. Far from indicating methodological defects, these comments show how closely commission procedure was in line with the government project. On the other hand, calculating ordinary budget by subtracting expenses from in-flows shows how empty the notion of social emergency is. That simple calculation shows something quite basic and obvious : the few hundred francs remaining to applicant households once rent, water, electricity, and phone bills had been paid could not suffice to ensure their subsistence : the financial resources of four out of five ASSEDIC recipients, three out of four RMI recipients, and nearly two out of three employed applicants fell below the fixed threshold. The administration’s discovery of this “chronic poverty” invalidates its presupposition that applicants’ difficulties amounted to financial accidents that justified, and could be solved by, one-shot financial assistance. As one DDASS director put it, “True social emergencies don’t exist, except when you’re dealing with persons with no roof over their heads and nothing to eat” –a fairly rare situation. Moreover, according to the national evaluation done by the Ministry for Employment and Solidarity (1998, p. 8), “83% of applicants alleged long-term insufficient resources (necessities, survival)” while only “10% referred to an unanticipated accident (sudden job loss, family break-up, illness)”, observations leading the authors of the report to conclude that “there is a profound feeling that poverty is becoming chronic”. As certain commentators were quick to point out, the kind of “emergency” the FUS was designed to handle was in fact more political than social.

The “remainder to live on” calculation and its use as a discriminating criterion in commission responses can ultimately be considered a method aimed at constructing an objective standard for a system of local justice. It was a practical means of conducting a first sorting procedure on an empirical basis, a procedure validated by other fund-distributing welfare institutions. And the sorting proved effective overall (see Table I): eight out of nine applicants whose “remainder to live on” fell below the limit did obtain assistance, while two out of three above the threshold were turned down. This also indicates that, proportionally, differences from decisions that would have resulted from automatic application of the criterion involved more generosity rather than less.

In reading the applications we see that, while the threshold was first considered indicative, it rapidly became psychological. Several applications where remainder per household member was a few dozen francs above the fixed level show traces of having been “corrected” downward, and the initial affirmative answer being replaced with a definitive rejection, which was then communicated to the applicant in the following standard terms : “Your situation does not correspond to the conditions specified by the Ministry of Employment and Solidarity circular of January 12,1998; ie, ‘persons or families in a situation of grave distress and in danger of not being able to maintain themselves in subsistence conditions’.” This justification of rejection often contradicted applicant statements. It was sent, for example, to an RMI recipient whose statement of financial difficulties and supporting documents indicated fairly clearly that he fell within the framework of the government circular :

TABLE I.
Social distribution and the discriminating power of the “remainder to live on per household member” criterion

The exceptional nature of this applicant’s need and his demonstration of good intentions overcame the threshold effect, but only for a moment. The sum of 3,000 francs initially granted him was then scratched out, for the stated reason that he was over the limit : his “remainder to live on” had been calculated at 1,115 francs. Clearly for this person, the norm weighed in as the decisive element. It should be noted that “remainder to live on” was sometimes checked in near-obsessive fashion. In this case, applicant’s resources came to 2,115 francs a month and fixed expenses to 1,743 francs, leaving 372 francs–an amount which should certainly have qualified his application. In fact, the commission had gone back and corrected its original calculation, now excluding transportation costs and the debt repayment mentioned in the letter, thus rounding off to 1,000 francs the amount to be subtracted from his income. This illustrates how painstakingly “remainder to live on” was evaluated. Not only were supporting documents required, but applicant’s own figures were checked, in some cases to calculate monthly amount of an expense paid out over several months or a year, in others to subtract an amount seen not to qualify as a fixed expense, in yet another to exclude a person not considered a household member from calculation of remainder per person. Painstaking calculation of this sort, attested to by the presence of administrators’ handwritten figures next to those given by applicants, was nonetheless limited.

It should be noted that applicants’ presentations of their situations often contained inconsistencies that made it practically impossible to decide in a co-herent manner. Consider the example of a 30-year-old Frenchman whose only income was the RMI and whose only declared expense was 800 francs a month to his mother, in the form of rent. He also mentioned 8,000 francs in debt–fines for public transport fare-dodging, explaining : “I would also like to be able to pay to travel legally in public transport to look for a job.” (R 60). Clearly he did not include the cost of a monthly transportation pass in his own calculation of fixed expenses. And yet not having a pass is precisely what explains his debt. Also, since he knew it was imperative for him not to get any more fines, it became more difficult for him to look for a job. If, instead of expressing his wish, he had indeed been able to buy transportation passes, his “remainder to live on” would have been calculated in such a way as to qualify his application. This was also the case for homeless persons who, because they cannot pay rent or hotel fees, have no regular expenses and thus fall above the established limit. In general, the commission’s evaluation of economic situation did not take into account expenses that poverty makes it impossible for people to pay out but that nonetheless would have enabled applicants to acquire integration-facilitating goods. The choice of accounting method here redefines the state’s intervention procedure in such a way that it is no longer based on reasoning in terms of individual needs but on reasoning in terms of household budget.

In fixing a single eligibility criterion, the commission was as much expressing a concern for convenience in aid distribution as it was a concern for fairness. While keeping open the possibility of deciding on the basis of other criteria once it had become familiar with the individual situation, namely once it had read applicant’s statement, the commission members endowed themselves with what appeared to be a relatively simple and dependable decisionmaking instrument. This explains the success of the criterion, testified to by both the aforecited effort to check calculations systematically and the general use of the criterion, which was quickly approved by the commission as the best way to proceed. In practice, however, the criterion proved more complex and less dependable. Since it is always a delicate business to assess an economic situation, since the instrument for doing so was sometimes applied in haste and without any clear rules, but also since situations of poverty are particularly difficult to analyze and minimal interpretive variations can have major effects given the small amounts involved, the “remainder to live on” criterion had limitations which the actors were not fully conscious of, especially when an amount of a few dozen francs above or below the threshold pushed them toward accepting or rejecting an application. Statistical analysis of the results and qualitative examination of the application files shows that this criterion was the main one used in determining who was eligible to receive financial aid from the fund and who was not. The “remainder to live on” was generally accepted as the best instrument for objectifying what the ministry circular specified as ability to “maintain subsistence conditions”.

Two competing principles were involved in fixing financial aid amounts : 1) a very rough evaluation of household make-up and 2) the application of specific criteria to each case. The first principle led to instituting an economic scale where “single applicants” stood in contrast to “families”. The second involved sending a “sign” to individual applicants, a kind of social semaphore. According to one commission president, the second principle prevailed : “In determining amount to be allocated, we sought to include a part of debt or fixed expenses. We could either grant an amount that would work as overall aid, or an amount that made particular sense for the individual applicant, enabling him or her to pay off a particular debt, an electricity bill for example, or handle an unplanned expense, such as a water heater breaking down, or lastly to realize a project, such as enabling a child to go on vacation. We tried as much as possible to grant amounts that people could identify. And we wanted at all costs to avoid giving complementary income.” The remarks of another president, however, suggest the priorities were reversed : “There were not really rates, but norms. In principle, it was 1,000 or 1,500 francs for single individuals, 3,000 or 3,500 for couples or people with children, unless a specific amount was justified by a particular feature of the file, such as a debt, a major accident, a particular project. In such cases we chose that amount, to send a message to the person. But in the great majority of cases, it was impossible to identify a particular amount.” This apparent contradiction between the two decision-makers’ accounts does not reflect a difference in the analysis on which decisions were based so much as a difference in commission practices. Presidents stamped commission deliberations with their particular approach, and this could be either individualizing-moralizing, with a tendency to personalize amount granted so that it would have a particular “meaning” for the applicant, or more universalist-objective, with a tendency to proceed on the basis of systematic “norms” applicable to all.

Still, these stated preferences are not necessarily reflected in the decisions, as is shown in the following two cases. The first concerns a French woman job-seeker, mother of one, and recipient of a specific “solidarity” allocation, with a “remainder to live on” evaluated at 558 francs per person. In her statement she indicates : “Live-in partner left the household leaving me with debts in the form of unpaid rent and car payments. A dependent child of three, often ill. Difficulty taking care of my health and clothing my son.” (J-S 20). The commission president, explaining that he preferred to grant an amount the recipient could identify, allocated 3,000 francs, which corresponds to the amount he most often granted households with children. To make his intention clear he wrote “insurance, hospital + water” in the margin of the application, though the recipient would never know anything of this because all she received was a form letter without any mention of what the sum was supposed to represent–a sum that was, moreover, all the less “readable” for being perfectly round. The second case concerns the application of an Algerian man, married with five children, living off the RMI and family allocations, with a “remainder to live on” estimated at 655 francs per person. “I’m at the end of my rope”, he writes. “The children are quitting school and there’s no work and that means they will stay at home so as not to be homeless and I also have to fed and cloth them.” (R 21). The president of the commission handling this case claimed in general to use a scale, but did not apply it here, instead allotting the amount of 2,608 francs, corresponding, according to a handwritten annotation, to “one month of expenses”. Later, however, he changed his mind, crossing out this highly specific sum and granting 2,600 francs.

In both these cases, as in many others, notes in the margins, supplementary calculations, and cross-outs reflect with what care aid amount was fixed. But while the “message” was aimed at the applicant, the sum allotted had meaning above all for the decision-maker, the paradox being that applicant would never see or know of the “sign” that decision-maker wanted to send, except perhaps in the 4% of cases where the amount, specified to the tens of francs, might give him a hint. Receiving 1,075 or 3,773 francs might lead her to suppose that this corresponded to a precise reality, though that remained to be discovered behind the (inaccessible) calculation that had produced the particular sum. In more than two out of three decisions where marginal annotations make it clear that the commission wanted to enable the recipient to repay a debt, overcome a specific difficulty, or realize a particular project, the sum was ultimately rounded-off, making the subtlety of the commission’s choice imperceptible. Thus in the two examples above, the intention–clearly indicated by marginalia– was invisible to the recipient. The administration was merely talking to itself.

Furthermore, fixing a sum that applicants may identify as pertaining specifically to them did not go against the notion of a scale. The two situations just cited were fairly similar in socio-economic terms and the commission accorded similar amounts, whatever the reasoning behind its calculations. Often the final amount seems to have been fixed near the established norm yet with intent to communicate a message. A Moroccan woman working half-time at minimum wage and mother of a young child, for whom per person household income after expenses came to 500 francs, was at first allotted 3,000 francs, in accordance with the scale; this amount was then corrected to 2,800 francs in accordance with the “legibility” principle, with the specification “rent” noted in two different places in her file (S 13). In this case applicant loss came to only 200 francs. In other cases, however, a sharp dissociation between the two principles may be noted. Take the case of a Senegalese couple on the RMI with expenses exceeding income. Their statement specifies that their financial situation has made them two months’ late with the rent, for a total of 4,600 francs. Instead of what the norm provides for, the commission granted them 1,340 for “three months of remaining rent”, their calculation based not on the real amount due but on the difference between rent and housing aid already received. For this couple, the loss incurred by the commission applying not the scale but the “legibility” principle was over 1,600 francs (J-S 11). Sending applicants a “message” could penalize them.

In general, the tension between the two competing principles was resolved by preferring “the message” on condition of not exceeding the “norm”. The “message” delivered must not be too costly, either in financial or moral terms (these two components are in fact indissociable, since granting excessive sums to certain applicants necessarily means depriving others). Ideally, the two principles converged. This was the case for a Moroccan woman in a four-person household (husband and two children) with 827 francs per person; she received 2,845 francs, exactly the sum of their stated debt (R 49). When an identifiable amount fell below scale amount, the former was most often chosen. A French couple with a young child living on 939 francs per person after expenses had been deducted from their single salary, was allotted 1,075 francs, corresponding exactly to its stated debt (S 3). The commission decided to allot one month’s rent, 1,200 francs, to a Tunisian woman earning half minimum wage, single head of household with a daughter, per person remainder of 602 francs (S 47). In these two cases applicants received much less than if the distribution scale principle had been applied.

When “message sending” sum would be grossly above scale amount, however, the commission tended to allocate the latter. The tension between the two principles in such cases seems to have been resolved either by abandoning the “message” principle or modulating it. An example of the first solution is a Peruvian couple with one child whose remaining income came to 702 francs per person : after initially deciding to grant them the equivalent of “half their rent debt”, 4,800 francs, the commission then backed away from such a high sum and applied the scale principle, granting them 3,000 francs (S 34). In an example of modulation, the commission allotted 4,000 francs to a French woman, RMI recipient, single head of household with one child and expenses higher than income, so she could pay off debts accumulated at a previous residence. The debts were too high to be paid in full, and a note inserted into the file (R 40) reads : “Nantes arrears (electricity, tel, hospital, rent), not including housing tax.” It is through such complex adaptations that actors can maintain a sense of consistency between two justice principles, one involving a universal rule, the other a subtle kind of casuistry.

In effect, the commission had to juggle a principle of equality that instituted the same amount for everyone, an amount fixed on the basis of objective, if coarsely defined, criteria (only two possible types of household : one person or more than one person), and a principle of fairness, granting to each in accordance with his or her need (on the basis of three situations : debt, accident, and project). In the general approach here described, fairness prevailed as long as it did not threaten equality “from above” (Figure I). If we consult the numbers, scale seems to have been more widely applied than semaphore : 32% of applicants were granted 1,500 or 2,000 francs; 36%, 2,500 or 3,000 francs. As we have seen, however, real intentions were more complex, because a scale amount could be allocated with a particular intent. There remain two oft-allotted amounts that cannot be accounted for by this twofold approach : 1,000 francs, granted in 11% of cases, often in cases where applicant seems not to have fully convinced the commission, such as single persons living with their parents, and, in 4% of cases, 4,000 francs, often granted to large families.

Given how meticulously “remainder to live on” was calculated to determine eligibility, it is striking that that criterion was not used at all in determining amount to be allotted. Monthly remainder per person was simply not used in deciding how to distribute fund resources. Not only did commission members never mention using this criterion, the application files show no trace of it. A person with less was not allotted more. Once it had been clearly determined that the “remainder” was under the 1,000 francs that made a favorable response highly likely, the applicant was evaluated according to another system. Remainder per person could therefore no more predict amount granted than it could predict aid eligibility. Just as a household with a negative “remainder” had no greater chance of receiving aid, statistically speaking, than one with a remainder of between 500 and 1,000 francs per person, so the former household could not expect to receive more than the latter. In the case of the FUS, the Christian parable of the vineyard worker hired in the eleventh hour was applied in the area of financial aid. Once a person has entered the moral space of aid-receiving, at the moment of determining how much he will be compensated for his pain, he is no longer judged in accordance with his degree of poverty, but only by means of a rudimentary scale and a semantic system, techniques with which the public authorities–first among them, the state– manifest their mastery of the rules of the game.

But local justice cannot be described merely in terms of the rules that actors say they follow in their decision-making. Above and beyond the justice principles underlying allocation of financial aid–principles that actors do not formulate beforehand to guide them but rather proclaim afterwards to justify their decisions– there are ordinary types of reasoning, affect-based reactions, and moral perspectives that determine acts of judgment or judgment practices.

Applicants were assisted in assembling their files either by an advocacy group for the unemployed (the Association pour l’Emploi, l’Information et la Solidarité and the CGT-Unemployed are the only two such associations represented in Seine-Saint-Denis [2] ) or a local administration (the city center for social action or other local social service), or directly by the DDASS, where agents had been mobilized to carry out this work for the entire period in question. At this stage there were major disparities in what may be called access to the agency window, since there were major variations from one town to another in when and how an application could be submitted (Fassin, Defossez and Thomas, 2001). All applications, regardless of where assembled and sub-mitted, were closely checked, with particular attention to the validity of documents substantiating applicant’s statement on income, expenses, and debt. The main information–marital status, household make-up, employment situation, budget and debt summary, amount requested– was reproduced on what was called an “analysis” sheet. The full file, with applicant’s letter, supporting documents, and information summary sheet, was then examined by the allocation commission. Once the commission had gone through the initial stage of elaborating its “doctrine”, a phase in which the highest-level department administration representatives participated, the commission itself was set up, composed in most cases of a senior civil servant from the DDASS (either director, deputy director, or a general inspector) –he or she systematically presided– mid-level administrators, and in some cases social workers representing the other welfare institutions (Department Council, CAF, ASSEDIC). In what some called the “second generation” of the commission, which handled far more applications, the status imbalance of members led to near-hierarchical authority relations : decisions were systematically finalized by the president, deliberations were sometimes very short and, for most applications, did not allow for any real discussion. Differences in treatment from commission to commission were thus quite sharp; everything depended on who was presiding (one of three persons could be in charge of this task). It is therefore by considering this procedure in terms of the two-fold constraint of individualized treatment (presentation of each case) and strictly limited deliberation time (one minute per application on average) that we will be able to see just what effective power the rules had and how readily liberties were taken with them. Three elements attest to the maneuvering room the commision gave itself and that justified in members’ eyes the way that this system, unusually heavy for a decentralized state administration, functioned.

First, interviewed commission members expressed judgments and feelings about singular cases and the decisions taken on them; this shows the importance of subjective elements in their casuistry. “We were really happy when we could help welfare recipients who had had a nasty blow–funds to replace a washing machine that had broken down, for example”, or “There was the case of a Mali man who wanted to send his daughter on a trip to Spain to study Spanish. Raising his children all alone. He would’ve done anything in the world for his daughter.” Administrators’ interest in applicants’ texts was not only functional; they were not merely concerned to determine eligibility and, possibly, amount. For some there was also an ideological dimension, involving an implicit critique of the principle of external assessment by the social services : “People are perfectly capable of articulating their requests themselves. What struck me was the quality of their writing and even the sensitivity of their letters. They were always very clear. This confirmed our initial idea about the autonomy of such persons”, explained a local civil servant in charge of social services.

Second, in the absence of dependable data for the entire set of applications, statistical study of the sample clearly indicates that the commission left itself considerable maneuvering room in decision-making. For 53 eligibility decisions, it did not respect the prescribed threshold : 24 were harsher than the norm, rejecting applicants whose “remainder to live on” was below the 1,000 francs; 29 were more generous, granting aid to households whose “remainder” exceeded the fixed limit. However, we must be cautious in using quantitative methods for amount allotted as this was not always explicitly justified. With this reservation, only 96 of 224 favorable decisions were based exclusively on the scale; for the 128 other applications, personal information on applicant influenced the decision, with debt reimbursement noted in 31 cases. Overall, the degree of liberty taken by the commission with regard to the norms was greater in determining allotted amount than in recognizing eligibility, if only because the declared principles explicitly included “message sending” as well as the scale norm, thus generating more than one amount to choose from.

Third, qualitative analysis of applications reveals wide variation in decisions taken, particularly in determination of amount. Applicant texts often played a decisive role in these judgment fluctuations, as we shall see. A single person declaring a salary of 7,000 francs and 4,500 francs of fixed expenses, a “remainder” equal to 150% of eligibility threshold, was nonetheless allotted 1,500 francs –“given debts and overall situation”, reads the note on the application (S 70). The debts in question came to 41,657 francs and the statement presented the context thus :

I live alone in a three-room apartment, until last month I recieved the APL [Public aid for housing]. My income does not allow me to live decently; too many loans to repay. The ASSEDICS don’t want to pay me, and my work contract ends 4/17/98, and what then ? I want to make a new start, without loans, without owing anything. Today I’m paying for the errors of my youth; I’m 27 and have no future. No one wants to or can help me, perhaps you can do something for me.

When we look at decisions “after the fact”, it is clear that applicants did well to write their statements with particular care. Argumentative ability is a great help in this self-justification exercise and truth test. In the absence of any external social services assessment, the written statement was supposed to relate a true story. Commission members’ often highly favorable assessments of the quality of these texts are in fact less directly concerned with their literary value or grammatical correctness than their clarity and veracity. It was not enough for the text to touch the reader; it must do so legitimately, in a context where commission suspicion could immediately disqualify an application. Still, the doubts manifested by the commission nearly always concerned supporting documents, whether present or missing, and in some of these cases applicants’ statements were perceived by the commission as powerful proof of veracity.

Two types of commission response may be observed. In some cases, applicant’s account convinced the commission, even though the objective application file was not very credible. In this case, reservations were more likely to be expressed about amount to allot–which often came out lower than that indicated by the scale. A young French woman raising two young children by herself stated that she received RMI welfare of 2,160 francs and gave vague figures for fixed expenses, with no supporting documents (R 83). It was therefore impossible for the commission either to check her expense figures or calculate remainder per person. Her statement seems to have convinced the commission while not entirely counterbalancing the insufficiencies in her file, because she did receive aid, but only 2,000 francs :

Often, however, the text was not enough to reverse the poor impression made by the application as a whole. As soon as suspicions were raised by the discovery of a disturbing detail, no demonstration of good faith on the part of the applicant made any difference. A young, single Peruvian woman, mother of two, with a one-year residence permit, explained that she received half-time minimum wage, with expenses exceeding income (S 15). The bills attached to her application are in another name, a point which the examiner notes at two places in the file, first next to her debt statement, where he writes : “No ! bills not in applicant’s name”; then next to the final decision : “Electricity bills, tel, and rent receipt in the name of Mr N (same address), Mr N’s name… not on the ‘family members list’.” A careful reading of the application might have removed all doubt or at least shown that the application was honest on the following two points : as is the custom in Latin America, the Peruvian woman’s children bore both parents’ last names, making it possible to identify Mr N as their father; the bills in his name attest to the fact that the young woman did not make the proper changes after she separated from Mr. N, probably because her resident permit was so tenuous. Paradoxically, the very fact that she did not explain this in her statement attests to her honesty, suggesting that she thought the objective data she did give, complete with supporting documents, did not require further explanation. Indeed, she goes no further than to articulate her needs :

In questionable cases such as these, one feature was decisive : whether or not the application had gone through a social worker and bore his or her approval in the form of a comment, signature, or stamp (in many such cases, the social worker helped the applicant prepare the file). The value attributed to this kind of support, especially when the social worker expressed a clear position on the application, reflects commission recognition of the abilities this profession is credited with, but also of precisely the kind of social assessment decried by some commission members. Social worker’s approval was none theless a source of real legitimacy for the commission. A sum of 4,000 francs more than what the scale indicated was allotted to a French family with two children whose application bore the following remark from a social worker (remark judiciously adapted to the kind of reasoning often involved in attributing aid): “Mrs Mr X wish to expunge their debts. They don’t dare indicate sum owed.” (J-S 30). Social worker’s remark for a Mali family with nine children was more directive (they obtained sum requested, well above established norm): “At the end of December Mr Y received two huge bills : second notice for a year’s worth of cold water consumption and paying off supplementary housing fees for 1997. Could he be given 5,000 francs of financial aid to help him pay these bills ?” (J-S 8). In some cases social workers’ requests were even more pressing, particularly for second-time applications. A 25-year-old French woman, RMI recipient, whom the commission had turned down twice for not furnishing sufficient “elements (motivation and supporting documents) to enable us to make even a cursory evaluation of your situation”, submitted a third application on which the social worker for the Departmental association for the promotion of Roms had written as follows :

The social worker’s perseverance won out over the commission’s reservations and the applicant was granted 2,000 francs. The maneuvering room the commission gave itself and actually used thus made all sorts of accomodations possible. There were two main bases for such accomodations. One, the commission may have been affected by certain features of an application. Particular justice values were then mobilized, for a greater show of either generosity or severity. In this case, decisions belonged to a specific sphere of justice. Two, deliberations respected neither the general principles the commission had set itself or particular criteria. In this situation we are presented with exceptions that cannot be accounted for in terms of either rules or reason. In such cases, there was no longer any concern to justify decisions. We shall now examine the substance of these two types of decision-making foundations.

In composing their statements, candidates for FUS funds were compelled to try to convince their readers. To this end, and to overcome both the physical and social distance separating them from the decision-makers, their main means was to appeal to their emotions; that is, to restore a relation of proximity through their story of suffering that would enable writer and reader to share in a common humanity despite the two-fold distance between them. In their attempt to create a tie of sympathy that would give them a virtual if only ephemeral presence in the deliberation space, they were obviously not aware of the principles and criteria used by the commission (understandably, since the commission only discovered those principles and criteria in the process of examining applications). Still, applicants were hardly without ideas of how they would be judged (if only because most of them had already been in prolonged contact with social action and welfare services).

Four types of self-justification may be identified in the rhetoric of applicant letters : appeals to necessity, compassion, merit, and justice. Each goes together with a particular style (expressive tone), vein (type of evidence), and argument (nature of facts invoked) (for a full analysis of applicants’ rhetoric see Fassin, 2000). The impact of applicant texts on commission decisions can be assessed in two ways. One is statistical; we can calculate rate of favorable responses by rhetorical figure identified (Table II). A strong point of this approach is that it involves exhaustive reading of applications and deliberations. A weak point is that we cannot consider the effects of each argumentation feature “other things being equal”, as this would entail multivariate analysis, which presupposes a high number of variables and, therefore, in order for each configuration to be represented, a high number of applications. The second approach is case-by-case, and involves analyzing decisions individually; this approach enables us to take into account all components of a given application, and in some cases to compare similar situations, closely following the logic behind the procedure adopted by the commission. The limitation of this approach is of course the singularity of each case and each deliberation. Both of these approaches have been used here.

TABLE II.
Effect of various rhetorical figures on commission decisions (%)

Appeals to necessity were made in a factual style and an accounting vein; texts presented factual data on income, expenses, and debt, and reached the conclusion of insufficient income and unmet needs. Applicant’s letter thus seems largely redundant given information in administrative file, adding only a few explanatory or self-justifying comments and having little effect on the decision. This was both the most likely configuration and the only one that produced a below-average rate of favorable response. If we then distinguish arguments that describe income insufficiency or financial difficulties only from those that evoke the risk of not being able to take care of vital needs such as food, housing, protection against the cold, we see that only 67% of first-type cases received aid, whereas 80% of second-type cases did. Argumentation appealing to necessity thus seems not to have been very effective overall, except when it suggested that applicant’s very survival was at stake.

Appeals for justice were written in vindictive style and used a protestation vein, either collective, on behalf of all poor people, or individual, with references to being personally wronged by an employer or administration. This was a marginal approach; depending on rhetorical feature studied, it is to be found in one out of 10 to one out of 25 applicant letters. In accordance with the role they think of themselves as playing when making a formal request to the public powers, applicants rarely made demands or suggested they were entitled to do so, so this type of argument had little overall effect on commission decisions. Though we should be cautious in analyzing such cases since there were so few of them, we can observe that statements denouncing general social inequalities or the unfairness of a particular situation were effective in 83% of cases. Contrary to what applicants may have believed, the commission considered such denunciations legitimate. The only feature that worked against a favorable response in this type of letter was using a threatening tone.

Demonstrations of merit, made in dignified style and a vein that appealed to a sense of applicant’s virtue, seem to have been particularly effective in convincing the commission. Though less used than necessity or compassion, deservedness was generally more decisive. As we have known for at least two centuries (Katz, 1990), a good poor person is a deserving one. Merit as used by applicants is to be understood in terms of the economic order, implying efforts to find a job; the social order, implying rejection of deviant practices; and the moral order, with reference to each person’s family responsibilities. First, there is the expressed will to become socio-economically integrated through work–a kind of leitmotiv that applicants had effectively internalized. A single jobless man wrote :

This applicant obtained 1,500 francs, nearly twice the amount requested. Moreover, the promise to manage one’s meager income carefully seems to have been considered strong proof of good intent, whereas getting into debt was seen by the administration as a deviant practice. A single male wage-earner wrote : “The financial problems is that Im having trouble getting by, specially eating. Given that I have a lot of expenses to pay, as well as payments that I make for my two children. I’m always overdran. So Im making this request to find out if you can help me pay the backup dets and eat.” To justify granting him 1,000 francs when his net income was 1,140 francs, the president of the commission noted : “Is making a real effort to balance his budget.” (S 76). Lastly, showing concern for one’s children or elderly parents elicited the commission’s benevolence. A divorced man, father of two children he was not supporting, with income after expenses assessed at 1,315 francs, wrote : “The RMI is my only income, I have my children every other weekend and half the school vacation. I have to feed them, and feed myself. My motivations are better health, and to find a job corresponding to my occupation. Am presently in a training program at L, to get remotivated. When I have a little more money, I would be able to ensure my children’s upbringing and more frequent visits, so that they will be in more comfortable conditions.” (R 17). This way of arguing, which links occupational and family projects, got the highly favorable response of 1,500 francs, even though applicant’s remainder was above the scale norm.

Judgment on the basis of merit can also be read in the negative in rejections, though it was rarely explained as such, and even in calculated reductions of initially allotted amounts. The response to application S 7 illustrates rejection for moral reasons. A young French man on the RMI requested aid for his family, consisting of his widowed mother and his two brothers, one employed, the other in high school. The file contained all necessary supporting documents for calculating “remainder to live on”, which came to 739 francs, making him theoretically eligible for aid. But the commission was reluctant to give it : the 1,000 francs initially set aside for him was scratched out. The problem was his stated debts, and not because of the amount, but the purchase : “a full set of living room furniture from the Leather center”. It is likely that this cause of debt seemed unacceptable to the commission. Three sets of arguments are given to justify rejection (two more than usal). First, the commission recalculated his expenses, though without managing to move him above the eligible “remainder”; second, it used an “imprecision about the situation” to cast doubt on income figures, even though the situation in question was clearly explained; third, it fell back on recommending that the application be sent to the CAF for possible family allocations, though it is hard to see how that could be justified for a family where the youngest child was 24. With application R 41 we have an illustration of reducing amount on moral grounds. A Mali mother receiving the RMI, net income estimated at 368 francs per person, described her situation thus (she obviously received help writing her text from a local administrative agent): “We are requesting financial aid to pay off a fine and an electricity bill. We ensure the upkeep and upbringing of 3 children, all in school, and my sister, for which facts we have obtained [the confirmation of] a reliable third party. Together we cope with the expenses incurred by our household, but to do this we have only the RMI and family allocations.” The applicant requested 2,030 francs but was only granted 1,580 with the note “electricity bill”. Despite her modest request, far below the scale amount, the commission had subtracted the 450-franc tax fine from allotted sum. For the administrators, this was a bad debt.

Lastly, applicants used an imploring style to elicit compassion; the approach was to win sympathy, often by a detailed description of misfortune. Compassion is part of how people in a welfare framework relate to poor people (Geremek, 1987). Because this feeling is produced when the suffering of one party meets the gaze of the other, a sense of shared membership in humanity is experienced, even though the distance between a long-term jobless person and the civil servant in charge of assessing his situation could hardly be greater. Emotion is a major wellspring in the exercise of charity, operating on two levels : empathy with the unfortunate person and satisfaction in assisting her. For the agents who assembled the applications, whether they were working out of administrations or advocacy groups, as well as for commission members deciding whether or not to allot aid, this combination of feelings was what cemented the collective experience. Everyone interviewed spoke about it. To hear them tell of the feelings elicited in them by poor people’s distress (“We cried along with them”) and the symbolic benefit of relieving that distress in whatever measure (“a nudge in the right direction”), it might seem that compassion had invaded the space of moral judgment. This was not the case, however, and neither the imploring style, with its numerous versions of desperate plea, nor the sympathy approach, which aims to create a kind of closeness through the use of concrete, painful details, affected application acceptance rates, which stand respectively at 77 and 76%, both close to the average. Only the mention of singular ordeals–sudden illness, accident, death of spouse or parent, even divorce, but not loss of a job, too common in this context– seemed capable of increasing proportion of favorable responses.

This insensitivity on the commission’s part, unexpected in a procedure where applicants present their own plight and thus reach for empathy, is readily accounted for by the mere recurrence of themes of misfortune, unemployment, illness, frustration, failure, and inability to cope. Sympathy for those in distress becomes a kind of routine affect that permeates judgment, no longer functioning very well to discriminate among cases. This is attested to by the commission’s increasing severity over time, perceptible in the decrease in proportion of favorable responses and sums allotted, whereas available socio-demographic indicators reveal no changes in applicant profile and suggest no particular financial constraints intervening to make the commission more parsimonious. For the 19,726 applications submitted, the rate of favorable responses decreased from 75.1% in January-February to 70.5% in March-April to 62.5% in May-June. Likewise, and once again for all applications, amount allocated in each case decreased from an average of 2,668 francs in the course of the first fortnight the commission was in session to 1,880 francs in its last two weeks. Emotion is blunted by prolonged contact with ordinary applicant distress.

Contrary to what may be thought and what the applicants themselves surely assumed, displaying suffering only had a slight effect on financial aid decision-making. A woman with a husband and two children and net income per person of 320 francs wrote : “My husband has been unemployed and without compensation since September 96. We collect a bit of RMI and my maternity leave + family allocations. There is not a single month that we manage to get by. For us each month is very difficult once the expenses and debts are paid we have nothing left. Our children have not had Chrismas like all the other children because no ressources.” (R 100). This pathos-charged account had no effect. Financial aid was granted–the “remainder to live on” threshold was applied– but no more than 2,500 francs, the lowest amount for a family, according to the norm. The commission was not moved by the plea.

Yet even this iron law was not immutable. A single man, whose “remainder” was calculated at 416 francs, presented his situation thus : “Havin been seriously ill and hospitalized for three and a half months in November 1995, and job-seeker since 2/2/96 and separated from my spouse since 11/29/96, I receive only the RMI and even with all the good will in the world and the privations that I have made it is impossible to make ends meet in the hope of finding a job to get on the up and up I hope to pay off a maximum of my dets so as to have delayed payment and show my good faith in taking care of my dets.” (R 72). This time the commission, presided by the same civil servant as for R 100, granted the applicant 2,500 francs, nearly twice what the scale indicated for a single person. The applicant’s accumulated misfortunes (illness, unemployment, separation) reasonably explain this generosity, especially since there is also a demonstration of deservedness (applicant is looking for work, wants to pay off his debts, manifests good intent). But if we consider this application in light of all the others handled by this commission, another interpretation, not exclusive of the first, is suggested; namely that this commission president was more sensitive to the distress of single men than families (a point which, as we shall see, differentiates him from other commission presidents, who often showed distrust of single men and sympathy for families).

The differentiated effects of applicant attempts to mobilize compassion underline the unstable nature of the criteria on which decisions were based, both from one commission to another and, as in the above example, within the same commission. We should therefore look into the limitations of this exercise of local justice with regard to explicitly cited norms and implicit values.

Such apparently incomprehensible decisions can be categorized with the help of André Lalande’s distinction between “arbitrary” and “contingent” (1993, p. 75 and p. 182). The noun arbitrariness is defined thus : “Concretely, the individual will of an authority; capricious decision-making in a matter where one should proceed by reason or the application of a rule.” The adjective arbitrary is defined as “arising exclusively from individual decisionmaking, rather than an established order or reasoning valid for all”. The following clarification is given : “Arbitrary differs from contingent in that contingent does not include the idea of will.” Contingent is defined thus : “General meaning : that which is contingent is conceived as having the possibility of being or not being; relative meaning : a fact is contingent in relation to a given general law or specific type when it consists not in the application of that law or type but in some circumstance particular to this or that individual object to which that law and type apply.” For decisions where it is clear that no explicit rule or strong reason has been applied, we can speak of exercising authority without reference to principles or values of justice. Such decisions will here be qualified as arbitrary when made through will alone. They will be designated contingent when chance of some sort prevails over intention. In the first case, power acts excessively. In the second it is manifest by default. In both situations, authority is unfairly exercised.

One type of unfair decision-making stemmed from the fact that the commission lacked adequate knowledge to do what it was set up to do. Commission members were not fully aware of how welfare scales work, and this may have led to their making decisions contrary to those they would have made had they been correctly informed. A Yugoslavian man, father of two, RMI recipient with expenses exceeding income, was suspected by the commission of cheating. The president noted on his application file, “Some income not included (applicant states 1,739 francs of RMI !)”, and his request was turned down. What the commission did not take into account was the fact that the RMI, like other minimum welfare benefits, is a differential allocation determined on the basis of applicant’s existing resources so as to reach a maximum total amount. In December 1997, for example, the nationwide average total paid to a couple with two children was 4,736 francs, representing housing aid plus family aid plus RMI. In fact, the Yugoslavian applicant’s statement –“I ask you to help me with the sum of 3,000 francs because I’m not managing on my income of 4,478 F/mo.”– together with the supporting documents he sub-mitted, clearly indicate as income what he received in housing aid and family benefits, and we see that that income is perfectly comparable with what is normally allotted to a household such as his. With this information, the commission should not have suspected the applicant of duplicity and, given his negative “remainder”, it should surely have allotted aid. The unjust decision here resulted from misapplying the rule. Rather than arbitrariness, this was a matter of incomplete knowledge.

Above and beyond this particular case–which, it should be noted, involves a structural matter : incomplete knowledge led to injustice here precisely because public generosity toward the poor is always dispensed against a backdrop of suspicion about how real their distress is and how true their declarations– unfair decisions, ie, decisions that do not stem logically from universal principles, even debatable ones, or from particular but regularly applied reasons, were made in three areas.

First, there is how the commission calculated “remainder to live on”, which as we have seen could be decisive for an applicant’s obtaining aid. Though it was done with special care and attention–documents were examined and calculations checked and redone in almost Byzantine fashion– it was also done with varying degrees of rigor, and this in three areas : 1) expenses, namely transportation, loan payments, local taxes, and sums paid to parents; 2) how family situation was understood (an adult son was sometimes not counted as a household member; conversely, an absent father raised doubts); 3) the very concern to check the data (absence of supporting documents could but did not necessarily disqualify an application; amounts might or might not be recalculated). In this way, different calculation practices, in some cases a result of omitting to use the practice that had implicitly taken over in the absence of precise codification, could lead to inconsistent, indeed, contradictory decisionmaking, including for two consecutively examined applications. We have seen how excluding the price of transport passes or debt reimbursements from expense calculations could put an RMI recipient’s “remainder to live on” above the limit, depriving her of aid that her situation and argument amply justified (R 51). Later on, the same commission examined the case of a young French woman who was also an RMI recipient. She included as expense a 1,000-franc pension paid to her parents, for which she presented no supporting document, as well as a monthly transport pass. This time the two expenses indicated were subtracted from income, so her income came out to be exactly the same as the preceding applicant’s and her budget was very similar. In her case the commission calculated a remainder below the limit and she was allotted 2,000 francs (R 54). If we remember that here the sum allotted to one and refused to the other corresponds to almost a month of income, it is obvious that microdecisions of this sort–including or not including a given amount in the numerator or person in the denominator– can have non-negligeable consequences.

Second, injustice affected decisions about applicants’ eligibility. This is manifest in exceptions to the 1,000-franc rule. In the case of homeless applicants, whose net income often put them over the limit given that they had either non-existent or very low regular expenses, some received aid whereas for others the limit was strictly applied (R 57 and R 60). For young persons living with parents, the response was generally no, since they were not considered a household in themselves. Some were granted aid, however, for reasons that are neither indicated or understandable in terms of their application (R 24 and R 20). High debts were sometimes a motive for rejection; the application was then sent to the debt commission. In other cases being in heavy debt did not work against the applicant. Consider the following two applications. A couple with two children, the husband a technical assistant in the building industry and the wife unemployed. Remainder per person came to 382 francs. The couple was 64,688 francs in debt, a quarter of which was for medical bills. The statement explained their efforts to “reduce expenses, repay our loans, and cope with daily first-necessity expenses” and even mentions the “promise of a half-time job to start March 21,1998” (S 35). Nonetheless, the application was rejected for being above the “remainder to live on” limit, and rerouted to the debt commission. The rejection letter includes the standard formula indicating that applicant’s situation is not that of a “person or family whose subsistence is seriously threatened”. We may well wonder how that letter was received in a four-member household heavily burdened with loans and living on 1,500 francs a month. Another couple, also with two children but wife employed, had income below its expenses because husband had lost his job. They were heavily in debt, and a repayment schedule had been drawn up and was included in the file. This the commission considered a serious commitment and it granted 3,000 francs–here, then, debt was not an obstacle (S 97). In general, there was no common reference that could systematically explain decisions to grant aid to an applicant whose “remainder” was above the limit and refuse it to an applicant whose remainder was below.

Third, the process of determining amount to be allocated was deeply affected by unfairness. Wide variations in amount could have very subtle causes : decision-makers claimed in some but only some cases to be sending applicant a message; they were affected by one feature in applicant’s statement more than another, and all these details varied by composition of the commission and mood of the moment, leading in some cases to incomprehensible severity, in others to unexpected generosity. Unfair decisions could as easily be made in applicant’s favor as to applicant’s detriment. A single woman raising two children, RMI recipient with net income per person of 940 francs, explained her situation thus :

To the applicant’s request of 5,000 francs, the commission initially wrote 2,500 francs on the file, then changed the amount, ultimately granting her 5,395 francs for “movable goods debts” –the highest amount allotted of all 300 applications studied. A few weeks later, the commission, now presided by another DDASS director, examined the file of another woman, also a single mother raising two children, also an RMI recipient. Her household’s net income per person was 610 francs and her debts were comparable to the aforecited applicant. “Alone with two dependent children”, she writes simply, ’”it is hard to meet my needs. At present I have ten months of unpaid rent. Thus even a tiny amount of financial aid would be very welcome.” (R 48). The aid allotted was indeed small : 1,500 francs, approximately half the sum indicated by her situation and three and half times less than what the preceding applicant received for a similar situation. Charity is always at the giver’s discretion.

Given these decision-making realities, there is something troubling in the suggestion to applicants to fill out a “sum requested” box. In fact, 226 out of the 300 applicants did so, but in no case did the commission grant the sum, even though a departmental social services director expressed amazement at how “modest” the amounts were. The median request was 4,500 francs; a third of applicants requested 3,000 francs or less; only 14% asked for more than 10,000 francs, most in cases of heavy debt. Comparing wishes with amounts granted shows that the commission simply did not grant sum indicated, even if the difference was only a few dozen francs. A couple living on a Contrat-Emploi-Solidarité with 83 francs of disposable income per person per month asked for 2,000 francs, rounding up a recently contracted debt for electricity to the closest hundred. They were granted 1,900, to let them know the money was to go exclusively to paying the bill (S 4). A household with two children where remainder came to 895 francs per person requested 3,000 francs, corresponding exactly to the amount of their debts. To them the commission granted only 1,000 francs, a little less even than the 1,200 francs mentioned in the letter as a crucial expense (glasses for the woman, who had already lost the use of one eye) –no explanation given (S 67). By refusing to comply with applicant requests, the commission seems to have wanted to show that it had exclusive control of decision-making. In the first of the above cases, the “message” principle was operative (“pay off your debt”), although the applicants probably never got the message since in the month between submitting their application and receiving the FUS check, the household debt had practically doubled (we know this because four weeks after submitting his first request, the applicant sent in a second, probably because he was concerned not to have heard anything in response to the first). The second case demonstrates even more clearly that even a perfect fit between amount requested by applicant and the two-fold logic of “norm” and “sign” (amount requested was identical to amount indicated by the scale) might not produce the favorable response that could logically be expected. There is no clearer demonstration that “we, the commission, know best what is good for the poor”.

The arbitrariness and contingency operative in these decisions, which I am here calling unjust, should not, however, be seen as manifesting anomic justice. The problem was not that norms were absent, but that there was more than one norm to choose from, and that one was not systematically given priority over the other. In this, aid allotment differs less than we may think from organ allocation (Fox and Swazey, 1973). The fact that amount in one case was determined by the general scale and in another by a specific need; that from among the possible signs the commission could send, it chose a sign concerning debt, or some component of debt, for one applicant and a sign concerning applicant’s proposed project for another; and lastly that certain commissions or commission presidents were more generous toward applicants who had suffered misfortune or single women raising children while others were more generous toward homeless men or applicants who convincingly demonstrated merit, clearly indicates how arbitrariness and contingency can be generated when there are too many norms rather than too few. It should therefore come as no surprise that in this analysis of “unfair” judgments all principles and values thus far described are present. Arbritrariness is no more opposed to norm than contingency contradicts morality; instead, it is the unpredictable and sometimes unjustifiable application of norms and “morality” that leads to unfair decisions. Arbitrariness and contingency both correspond to irregular use of more or less explicit rules (Figure II). The difference is that the first is intentional and the second fortuitous. But the boundary between the two is not as easy to draw as it may seem. When the same commission, a few days apart, grants 2,500 francs to a male RMI recipient with a negative remainder who did not submit a written statement, and 1,000 francs to another man in exactly the same situation whose statement was written in the “dignified” style (R 34 and R 36), it is reasonable to assume that the two decisions taken together are merely an instance of the fluctuation characteristic of ordinary bureaucratic functioning. Still, we cannot entirely reject the idea that they also express a conscious will to establish a difference. Excepting cases where such will was explicitly indicated with reference to a moral or pedagogical value, and cases where, on the contrary, the desire seems to have been to break radically from expected decision or average amount, it is difficult, when confronted with unfair decisions, to determine whether gratification given or penalization inflicted reflect commission’s good or bad will or if they were the result of favorable or unfavorable circumstances arising from commission members’ fatigue or mood.

The common foundation of these two figures of injustice is composed of two elements. First, financial aid is a matter of obligation. The effect of public generosity is not to create citizens who lay claim to what they are owed according to the principle of national solidarity; rather, it creates poor persons who solicit the benevolence of the state. This has two consequences. On the one hand, the receivers, those who are “much obliged” for the aid, cannot allude to a social entitlement or demand that specific rules be applied; they can only submit to whatever methods are used to assess their requests, methods of which they know nothing. On the other hand, the givers, the “obligers”, feel they have a duty toward the society whose resources they are spending and the state whose charitable mission they are executing. The fact that senior civil servants from the DDASS were implicated in the entire process shows how seriously it was taken. Second, the decisions were to be based on individualized treatment of applications. By seeking to take into account the singularities of each situation, the empirical case-by-case approach thus practiced had a built-in tendency to pull away from common and transparent justice principles. This in turn had two effects. On the one hand, the “obliged” needed to show a certain skill in justifying their requests; in comparable situations, this generates unequal treatment. On the other hand, the “obligers” had to have recourse to moral values as the basis for their decision-making, as is shown by the “added value” accorded merit and misfortune, or more exactly how well merit and misfortune were put into words.