Abstract

There is a deepening crisis in the funding of legal services in the USA with cut backs in Legal Services Corporation and Interest on Lawyers Trust Account funding, rendering more visible the fact that there is and always will be persistent scarcity in the availability of both criminal and civil legal assistance. This article examines how existing Legal Service Providers (LSPs), both civil and criminal, should ration their services when they cannot help everyone. I draw on the bioethics literature on the allocation of medical goods (organs, ICU beds, vaccine doses, etc.) to illuminate the problems facing LSPs and the potential rationing principles they might adopt.

1. INTRODUCTION

There is a deepening crisis in the funding of legal services in the USA. The House of Representatives has proposed cutting the budget of the Legal Services Corporation (LSC), one of the main funders of legal assistance to America’s poor, to an all time low in inflation-adjusted terms (Ruger 2012). If it becomes law, the House Appropriation bill will have cut funding to the LSC by 21 percent from its 2010 levels—6 percent this time and 17 percent in the prior budget cycle. Other sources of funding, such as Interest on Lawyers Trust Account (IOLTA), are also way down due to low interest rates (id.). For example, Connecticut has seen a drop in IOLTA revenue from $ 20.7 million in 2007 to $1.7 million in 2013, a 92 percent decline (Eppler-Epstein Interview 2012). More than 135 state and local organizations providing LSC assistance are now in a precarious position (Ruger 2012). The community was already decimated by the last round of cuts in January 2011, which led to the laying off of 1,226 lawyers and support staff at LSC-funded organizations, and 81,000 fewer low-income Americans receiving aid (id.). This is all occurring at a time of extremely high unemployment and state budget cuts in services supporting low-income people, meaning demand for many of these services is going up (id.).

The deepening crisis in funding of legal services only makes more pressing and manifest a sad reality: there is and always will be persistent scarcity in the availability of both criminal and civil legal assistance (Bellow & Kettleson 1978; Legal Services Corporation 2009).2 Given this persistent scarcity, this article will focus on how existing Legal Service Providers (LSPs), both civil and criminal, should ration their services when they cannot help everyone.

To illustrate the difficulty these issues involve, consider two types of LSPs, the Public Defender Service and Connecticut Legal Services (CLS), that I discuss in greater depth below. Should the Public Defender Service favor offenders under the age of twenty-five years instead of those older than fifty-five years? Should other public defenders offices with death eligible offenses favor those facing the death penalty over those facing life sentences? How should CLS prioritize its civil cases and clients? Should it favor clients with cases better suited for impact litigation over those that fall in the direct service category? Should either institution prioritize those with the most need? Or, should they allocate by lottery?

These are but a small number of the difficult questions faced by those who have to ration legal services. Very little has been said as to what principles should govern the rationing of legal services.3 This is surprising given that civil and criminal LSPs are often funded through a mixture of government funding and charitable support in such a way that they should be answerable on questions of justice, and because their decisions whether or not to support a client is likely to have significant effects on that person’s life prospects.4 Thus, it seems as though the rationing decisions of LSPs deserve significant ethical scrutiny.

In this article, I seek to remedy this deficit in the existing literature by engaging in a comprehensive analysis of how LSPs should allocate their resources given the reality of persistent scarcity.5 Luckily, this work does not have to begin at square one. There is a developed literature in bioethics on the allocation of persistently scarce medical goods (such as organs, ICU beds, and vaccine doses) that I use to illuminate the problems facing LSPs and the potential rationing principles they might adopt.

Section 2 of this article begins by explaining why bioethical thinking about rationing medical goods is a useful starting place, while highlighting some key potential differences between the two contexts that are taken up later in the article. I also situate my focal question, relating to rationing from a fixed budget, among other questions regarding the funding of legal services such as the case for funding them at all. I then examine the way three existing institutions, the Public Defender Service for the District of Columbia (PDS), Connecticut Legal Services (CLS), and the Harvard Legal Aid Bureau (HLAB), currently ration services. Additionally, I discuss the ways in which the constitutional right to counsel jurisprudence and the rules of professional responsibility do and do not constrain LSPs’ room to ration.

Section 3, the heart of the article, develops potential rationing principles for LSPs, drawing some inspiration from bioethics. For this Section I adopt a series of simplifying assumptions, most notably that the mission of a LSP is solely to produce the outcomes desired by clients, and ask what rationing principles would be most ethical to adopt.6 I discuss six possible families of ‘simple’ rationing principles: first-come-first-serve, lottery, priority to the worst-off, age-weighting, best outcomes, and instrumental forms of allocation. Each contains multiple possible principles, and exploring the ethical defensibility of each as applied to LSPs introduces many complex normative questions. While I ultimately tip my hand on my views of each of these sub-principles, my primary aim is to enrich the discourse on rationing legal services by showing LSPs and legal scholars that they must make a decision as to each of these issues, even if it is not the decision I would reach.

Section 4 presses on the analogy to medical services by focusing on some distinguishing characteristics of legal services that might alter the appropriate rationing principles. First, I examine how bringing in dignitary or participatory values complicates the allocation decision, drawing in particular on Jerry Mashaw’s work on Due Process values (Mashaw 1979, 1981). Second, I ask whether it makes a difference that, in some cases, individuals who receive legal assistance will end up succeeding in cases where they do not “deserve” to win. I also examine whether the nature of legal services as “adversarial goods”, the allocation of which increases costs for those on the other side of the “v.”, should make a difference. Third, I relax the assumption that funding streams and lawyer satisfaction are independent of the rationing principles selected, and examine how that changes the picture. Finally, I respond to a potential objection that I have not left sufficient room for LSP institutional self-definition to allow departures from these rationing principles.

Section 5, a concluding section titled “Some Realism about Rationing”, takes a step back to look for the sweet spot where theory meets practice. I use the foregoing analysis to recommend eight very tangible steps LSPs might take, within their administrability constraints, to implement more ethical rationing.

While I have no illusion that even this complex model perfectly captures the reality of allocating legal services, part of the power of this analysis is its illustration that even a simpler model demonstrates many flaws in current systems for rationing legal services and why this neglected area deserves much more scrutiny than it has heretofore received. My goal is to begin this conversation.

2. STARTING POINTS: WHY FUND LEGAL SERVICES, RATIONING IN THE REAL WORLD, THE ANALOGY TO MEDICAL GOODS, PROFESSIONAL RESPONSIBILITY, AND THE RIGHT TO COUNSEL

2.1. Rationing in the Real World

Before we can understand how rationing legal services can be made better, we must understand what kinds of rationing systems currently exist. In this article, I cannot offer an exhaustive survey of rationing practices of all LSPs. But I try to ground the theoretical analysis in the real world by examining the current rationing approaches of three LSPs: PDS, CLS, and the HLAB.

2.1.1. Public Defender Service

PDS was created by federal statute in 1970, and funded by Congress to satisfy the constitutional mandate to provide defense counsel to poor people in criminal, juvenile, and mental health proceedings (Pub. L. No. 91–358, Title III, § 301 (1970); PDS 2012). It is widely considered to be the gold standard for public defenders across the nation.

PDS handles a majority of the most difficult, complex, time-consuming, and resource-intensive criminal cases, while private attorneys (Criminal Justice Act (CJA) attorneys) handle the majority of the less serious felony, misdemeanor, and regulatory offenses (PDS 2012, 4). PDS attorneys also handle criminal appeals, almost all parole revocation hearings, and most Superior Court Drug Intervention Program (Drug Court) sanction hearings and represent people facing involuntary commitment in the mental health system, children with special education needs facing delinquency charges, and clients in civil proceedings whose issues were triggered by their criminal charges or their incarceration (id.). PDS routinely represents clients in the D.C. Superior Court, the D.C. Court of Appeals, and at times in the United States District Court, Court of Appeals for the D.C. Circuit, and the United States Supreme Court (id.).

PDS is organized into several divisions. I will focus on its Trial Division, which provides assistance in criminal cases in the superior court and child delinquency matters, and the Appellate Division, which handles PDS’s own appeals and some non-PDS appeals at the request of the courts of appeals and sometimes CJA attorneys (Leighton Interview 2011).7 PDS also contains a Special Litigation Division, Parole Division, Mental Health Division, Civil Legal Division, and Community Defender Division.

Even before it begins allocating attorneys and their time to potential clients, PDS has made several structural rationing decisions. The first is the allocation of resources and personnel to the divisions other than the trial and appellate ones. PDS could give more resources to criminal defendants if it eliminated divisions such as those handling mental health and child delinquency. It could also change the balance between appellate- and trial-level resources.

Moreover, PDS has made structural rationing decisions as to what kinds of cases it has recommended to the courts that PDS take as opposed to the cases it has recommended be given to CJA attorneys, a division the courts have largely followed resulting (in combination with some statutory rules) with CJA attorneys getting primarily the “less serious felony, misdemeanor, and regulatory offenses” (Leighton Interview 2011; PDS 2012). Though PDS does handle some misdemeanors involving sex offenses where the collateral consequences can include sex offender registration, eviction, and deportation its docket consists primarily of more serious offenses.

Beyond these built-in threshold rationing decisions, PDS makes some more fine-tuned distinctions. First, because of conflicts (between co-defendants and between clients who develop adverse interests as their cases proceed), PDS must on occasion choose who it can continue to represent. In deciding which defendant to represent, PDS tries, to the extent permitted by the Rules of Professional Conduct and by the appointing authority, to choose those individuals “in the most trouble and facing the more serious consequences if found guilty” (id.). Attention is also paid to financial need in these circumstances: PDS will favor the defendant whose case will be the most expensive, or who needs the most specialized resources (e.g. mitigation experts, forensic experts, translators, appellate expertise, and mental health expertise) (id.).

Second, attorneys themselves are rationed, with the more experienced attorneys promoted to handling exclusively the more serious offenses (id.). Most attorneys begin by litigating juvenile delinquency cases (where no matter the offense charged, the accused can only be incarcerated until age 21) (id.). Typically, after a year, they are promoted to dealing with adult cases, initially felony drug and gun charges (id.). After that point, all promotions to the next offense grade are through attorney application (id.). The next promotion is to cases with the possibility of a life sentence (including armed robbery, armed car jacking, and others), and the final level includes homicide (id.). There is no death penalty in the District of Columbia, so (unlike other criminal defense LSPs) PDS does not have to decide which attorneys to allocate to death eligible offenses. With each promotion, attorneys take their cases with them, but from that point on receive new cases only at the higher level (id.). PDS does not have rules about how many lawyers ought to be representing clients at each level at any time, and instead promotes based on ability, and in the past “PDS has maintained its ability to handle the majority of the life offenses and an even higher percentage of the homicides” (id.).

Within each of these categories, neither attorneys nor the office administrators have a policy of allocating resources to the more serious cases under any definition (id.). Instead, attorneys are taught to “fight every case like it is a serious case because it is a serious case to the client. This culture, that every case is an important case even when the charges are less serious, serves both the expressed interest of the client and facilitates training attorneys to practice at a level that will eventually allow them to handle the most serious cases” (id.). Nor are attorneys instructed to try and sort cases based on chance of success or the likelihood of a plea offer. Instead, “PDS lawyers are trained to be client centered and to pursue the client’s goals and objectives. While focusing on a client’s objectives will drive some decisions (for example focusing on negotiations or investigation or the filing of certain motions) the ethos is to do that which is needed to secure the best plea offer and simultaneously do that which is needed to be in the best posture to try the case should the client choose to go to trial” (id.).

The Appellate Division also does not do much self-conscious rationing. It does not pick and choose which appeals from the trial division to handle or even focus on (id.). The quality of the briefing on appeal is uniformly high, with little if any premium being given to cases with issues that are more likely to make important precedent (id.). In other words, even on appeal the focus is on direct services much more than impact litigation with respect to case selection. The Division is, however, aware of the issues being addressed in any given appeal and may assign appellate lawyers or supervisors with specific expertise to be involved in the case (id.). In addition, the court of appeals typically requests an amicus brief from PDS on matters of importance or involving novel legal issues (id.). The only office at PDS with an impact litigation focus is the Special Litigation Division, which occasionally files class actions and often writes specialized pleadings on issues facing many accused, such as Brady or eyewitness litigation (id.).

PDS does not engage in limited representation agreements in the criminal or mental health arenas (id.). For parole cases, PDS does not generally handle an appeal absent an explicit agreement to do so (id.). In its small civil practice it typically puts limitations in place as to whether it will handle an appeal or whether it will take a case to federal court (id.).

2.1.2. Connecticut Legal Services

CLS is a “private, non-profit, civil law firm dedicated to helping low-income families and individuals obtain justice”, that serves all areas of Connecticut except Hartford and New Haven, and operates six full-service offices and five satellite offices (CLS Website 2012).8

In fiscal year 2010–2011 it handled 8,124 newly opened cases (and carried over cases) with the following distribution: housing and homelessness 24 percent, consumer (mostly for elderly) 19 percent, domestic violence, divorce, child support, and other family matters 18 percent, public benefits and employment 10 percent, education and juvenile law 9 percent, Social Security 8 percent, health law (including Medicaid, Medicare, and nursing home matters) 6 percent, intellectual disabilities 1 percent, other cases 5 percent (CLS Website 2012). It also provides a limited number of status adjustments for domestic violence victims, self-petition, U-Visas, and, as an experiment, is deploying a new program to help clients with the “deferred action” policy recently started by President Obama by which individuals who came to the US at young ages, under certain circumstances, can petition to avoid deportation and obtain work permits for two years (Eppler-Epstein Interview 2012). Due to overwhelming demands for its other services, CLS does not currently take many cases in family law (outside of domestic violence), its intellectual disability work is limited to certain geographic areas, and its consumer practice focuses almost entirely on the elderly, largely because it believes that the elderly are more stressed by repeated debt collection calls then other client populations (id.) As CLS explains:

Although we serve thousands of individual clients and their families each year, the need within our service population for legal aid far outstrips our current resources.

CLS case selection priorities focus our resources on helping indigent clients meet basic life needs, for example:

Protecting employment rights, or obtaining a means of support when they are incapable of working or cannot find a job; Avoiding or escaping homelessness, and obtaining decent, safe, and affordable housing safety from domestic violence and other forms of abuse; A stable, integrated family; Medical and behavioral health care; A good education, especially for children with disabilities; Autonomy and dignity, especially for persons who are elderly or coping with disabilities; Protection against consumer scams, especially those that target the elderly and disabled, and; Avoiding or overcoming harmful discrimination based, e.g., on race, ethnicity, disability, or source of income.

We seek to maximize the impact of our efforts by undertaking cases and projects that cost-effectively benefit large numbers of needy people.

We provide community education programs to clients and social services agencies to help our clients know when they need a lawyer, how to avoid legal pitfalls, and how to solve or deal with some kinds of legal problems on their own (CLS Website 2012).

In 2010–2011, CLS was funded by government-based grants ($2,136,397), private grants (including bar foundation distribution of IOLTA, court filing fees, and a state budget contract) ($7,389,111), and “donations and other income” ($929,425), for a total of $10,45,933 in annual income (CLS Annual Report 2011). It employs fifty lawyers as part of its total staff of seventy-six lawyers. Most of the lawyers work full time, although it also receives minor level of support from private attorneys working pro bono and law school interns (Eppler-Epstein Interview 2012).

CLS gets cases through three main channels: Statewide Legal Services, other nonprofit agencies/community partners, and “repeat business” from individuals they have already aided who reach out to their lawyer for help with a new matter. Let me say more about each.

In 1995, when Congress introduced rules that limited LSPs funded through LSC funding in several significant ways—most notably prohibiting funding of LSPs that use class actions—CLS, among other Connecticut LSPs, decided to opt out of LSC funding altogether (LSC v. Velasquez, 531 U.S. 533, 538–539 (2001); Eppler-Epstein Interview 2012). They ultimately established “Statewide Legal Services” (“Statewide”) as a clearing house that would handle initial intake and distribution of cases for Connecticut LSPs. Statewide receives over a thousand calls a month from individuals seeking help with legal problems. It does an initial phone intake with individuals and then, where the cases meet pre-set criteria, sends them to CLS, among other LSPs. As will become particularly relevant when we discuss first-come-first-serve rationing below, Statewide’s phone is often literally ringing off the hook, and at certain times callers will get a voice message saying that Statewide’s staff are currently unable to take the call and that the caller should try and call back later (id.).9

Like the other participating LSPs, CLS gives Statewide parameters for the cases to send to it, which are essentially a first stage rationing system. It only takes cases in the subject matter areas described above. It accepts only cases for individuals whose gross income is at or below 125 percent of the federal poverty level (FPL), or whose gross income is at or below 187 percent of the FPL and have enough work or medical-related expenses such that their net income is at or below 125 percent of the FPL. Beyond these screens, CLS also provides more specific parameters to Statewide. For example, CLS indicates to Statewide it wants individuals with housing subsidies facing landlord–tenant disputes, not individuals without such subsidies. The reason is that although CLS representation helps clients without such housing subsidies—ensuring orderly eviction and retention of moveable property, or avoiding homelessness—those without the housing subsidy are unlikely to be able to afford the rent even when CLS is successful in its representation. Even more specifically, CLS will instruct Statewide not to send them cases where the basis for the adverse housing action was an allegation that the tenant was dealing drugs unless the Statewide interviewer sees evidence that the allegation is false. The reason given for this rule is that individuals who are dealing drugs are considered bad for the rest of the low-income housing population, and thus the LSP has decided not to try to serve them when it cannot serve everyone (id.).

Once Statewide has determined that a case fits CLS’ parameters it does a check for conflicts of interest. Assuming none, Statewide’s case management system ordinarily transfers the case to CLS’ case management system, and in turn to a secretary who will transfer it to one of CLS’ lawyers. The decision as to which attorney to send it to is made based on subject matter expertise and geographic coverage of various CLS attorneys. One deviation from this system occurs when a CLS attorney is oversubscribed or on vacation. Much like hospitals, the CLS system allows notice to SLS that a CLS attorney is currently “closed for new business”. In such cases, CLS or Statewide will attempt to try and find a private attorney interested in pro bono work to take the case or, if the matter is less urgent, instruct the potential client to try again several weeks later. The decision that a particular CLS lawyer is “oversubscribed” is always approved by the office supervisor, but may be initiated by an individual attorney or her supervisor. For cases referred to CLS by its community partners, CLS treats this decision as a softer boundary and has a commitment to “figuring something out”, in part to maintain good working relations with these institutions and in part in deference to referring agencies’ expertise in identifying priority emergencies (id.).

Assuming the case gets transferred to CLS and the CLS attorney has capacity to take on the case, a second round of intake evaluation is done in person by the CLS attorney assigned the case. At this point, the CLS attorney’s assessment of the case may reveal that it is no longer an appropriate case for CLS to handle; they may learn new information, see new documentation, determine facts from an in-person interview that were not ascertainable from a telephone screen, etc. When a case that was transferred is now deemed unsuitable for representation, the CLS attorney still endeavors to provide additional advice to the client on how to represent himself in court, or other more limited forms of representation I discuss below. Let me emphasize that the decision as to whether to take a transferred case after this in-person intake interview, as well as the decision of how much resources to invest in a particular case, are usually left in the hands of the individual CLS attorney who will handle the case in question, though they may sometimes talk to others in the practice if they want more input (id.).

A second source of cases that CLS handles comes from local nonprofit agencies and community partners. One prominent example of this is representing victims of domestic violence, which makes up a large share of CLS’ family law docket. Community partners involved with helping victims of domestic violence will, like Statewide, engage in a first round of intake screening before sending the case to CLS; if an attorney is available, that attorney will then do a second intake interview and screening. The domestic violence program is also a good example of the way in which CLS’ docket mix is in part a function of its funding sources, growing as grant funding opportunities in this area became available. Another similar example has to do with representing minors in education law issues where CLS has increased its representation of children referred from juvenile probation offices because of a funding stream from the Connecticut court system, even though CLS believe it is more difficult to achieve good results in these cases than for minors referred by teachers or other community referral sources (id.).

The third source of cases is “repeat business” from prior clients. For example, a client may call the lawyer that represented him in an eviction proceeding and say “you helped me with housing, and I recently applied for disability and got turned down, can you help me again?” Sometimes these are clients that are currently being represented by CLS in other subject matter areas; in other cases they are clients with closed cases. While CLS does not give these clients priority over other potential clients, it often indirectly gives them an advantage in that they get evaluated for intake by the relevant CLS attorney. As a result, these clients do not face the de facto rationing of calling Statewide to speak to an intake evaluator, nor do they run the risk of being screened out by Statewide at that stage, or referred to another Connecticut LSP (id.).

CLS’ ability to engage in limited representation is itself constrained by Connecticut law, making it one of only six US states with significant restrictions on “limited scope representation”. CLS cannot say to a client “this is the most vital hearing in your divorce case, the judge is going to make temporary orders on visitation and custody that will be hard to change once they are in place, so I will help you at this hearing but after that you are on your own”. That said, CLS can and does engage in other limitations that may be considered a kind of limited representation. For example, for clients it turns down after an in-person intake interview, it never says “now that I understand your problem, it is not a good fit, so good luck”. Instead the lawyer will give a brief session on self-representation, provide self-help publications and materials that CLS has on hand, and give general advice such as “when in court, this is the kind of thing you should say”. On other occasions the CLS attorney will say “I cannot take your cases but I think you are well-poised to do this yourself. I will coach you before each proceeding for a half an hour, but you are representing yourself”. The CLS Executive Director I spoke to was hopeful that Connecticut’s judicial rules would allow for limited court representation agreements in the foreseeable future (id.).

When choosing cases, CLS’ internal procedures do, to some extent, consider impact litigation potential. This might take the form of pursuing a class action or a case with an individual plaintiff that nonetheless will have major relevance for others in the client class (such as attacking the constitutionality of a statute relevant to the class). While the general docket is focused on direct service work, there is a process—especially for class actions—whereby a CLS attorney can get authorization to pursue these more resource-intensive impact cases. The attorney who initiates the request will fill out a form for his or her supervisor indicating what the case is about, what they expect to get if they win, how resource intensive they expect the case to be, etc. That form is reviewed by the head of the practice group, the head of litigation, and ultimately CLS’ executive director. If approved, the attorney handling that case will have his or her work plan adjusted such that he or she is assigned fewer individual cases. CLS gets about ten requests a year for such impact litigation authorization, of which they approve seven or eight, of which usually only one is a class action. CLS shies away from class actions in part because they are both resource intensive and often take a long time to achieve results. CLS is more likely to ration toward impact litigation involving a smaller number of plaintiffs seeking to establish an important precedent (id.).

CLS engages in a full-scale review of its rationing system every fifteen years, and a more focused review for strategic planning every five years. This is an extremely time-intensive project, with all staff members participating. The main reason why they do not do it more often is because doing so would divert staff from helping clients. This echoes a theme we will see a few places in this article, that deciding on and implementing rationing principles must itself be traded-off against helping clients, or put more flippantly: how to ration is itself a rationing decision. The CLS board has authorized the management to “experiment” by temporarily introducing new practice areas in between these formal reviews. CLS has most recently experimented with offering a clinic in Stamford to help workers (especially undocumented ones) to get advice and/or representation on collecting unpaid and underpaid wages, as well as advising/representing undocumented individuals in President Obama’s new Deferred Action program mentioned above (id.).

2.1.3. Harvard Legal Aid Bureau

HLAB, established in 1913, is the nation’s oldest student legal service organization (HLAB 2012).10 It is an entirely student-run nonprofit law firm composed of forty-five to fifty second- and third-year Harvard Law Students who are competitively selected (id.; Grossman Interview 2011). HLAB employs seven or eight practicing attorneys with extensive public and private litigation experience, who are responsible for training students on how to become effective advocates. These attorneys supervise the students’ actual casework, teaching the various skills necessary to practice law and understand the ethical and moral components of the profession. Staff attorneys also accompany students to court, provide strategic advice, and assist in case management (Grossman Interview 2011; HLAB 2012). HLAB also has a managing attorney/faculty director, and administrative director (id.; Lumley Interview 2011). Most of the case assignments at HLAB are handled by a subset of its student members (id.). HLAB is highly respected in the legal services community (Greiner & Pattayanak 2012).

HLAB specializes in four major practice areas. First, housing law, which includes public and private housing eviction and occasionally public subsidy termination (Lumley Interview 2011; HLAB 2012). More recently, it has shifted resources to the foreclosure crisis and its effects on low-income families and communities in Boston and surrounding towns, which has involved a greater focus on mobilization, impact litigation, and legislation (Grossman Interview 2011). Second, family law, which includes divorce, child custody, child support, alimony, and paternity (HLAB 2012). Third, government benefits, which involves assisting clients who have been denied government benefits, especially SSI (Social Security), unemployment benefits, but also on occasion SSDI (Disability), TAFDC (Welfare), EAEDC (Emergency Assistance), Medicaid, Veteran’s benefits, and Food Stamps (id.; Grossman Interview 2011; Lumley Interview 2011). Finally, HLAB has a wage and hour practice that helps individuals who seek to recover unpaid or underpaid wages or benefits from employers (Lumley Interview 2011).

Like PDS, HLAB has a series of structural rationing decisions built in to its organization. HLAB limits its representation to potential clients living in Massachusetts’ Suffolk County and parts of Middlesex County (id.). Its formal policy is to only represent clients who fall below 187.5 percent of the poverty level within the four practice areas listed above, but it refers clients that do not meet those criteria to other area providers (Greiner & Pattanayak 2012).11 The screens are applied during an intake telephone conversation when potential clients first call HLAB (Lumley Interview 2011). Cases that pass the telephone screen are further screened through an in-person meeting, and those that survive are considered by the intake committee (chaired by students), where all HLAB members (including staff and supervisors) get a vote (id.).

HLAB roughly divides its intake period from mid-August to November and January to early April, tracking the law school’s academic calendar, although it sometimes also takes cases in the summer (Grossman Interview 2011; Lumley Interview 2011). HLAB has limits on capacity for each intake period based on student and supervisor time, with a target of three to seven cases per student, and accepts clients during these two intake periods in a way that smoothes intake over the period (Lumley Interview 2011).12 Intake meetings generally occur twice a week (id.). A certain number of cases are accepted by the committee at each meeting, with an eye kept on needs in the later segments of the intake process (id.). HLAB parcels out the clients it accepts by practice area, although it takes more cases in housing and family sectors than employment or benefits (id.; Grossman Interview 2011). While HLAB sets targets as to how many clients to accept in any given meeting, its intake committee sometimes takes more or fewer cases in a particular meeting, affecting the number of cases that it can accept later (Lumley Interview 2011). While HLAB aims to smooth acceptances over each of the intake periods, at some point it reaches capacity and tells potential clients that they will have to wait until the next intake period to be considered, again largely dictated by the academic calendar (id.). Thus, there is some element of first-come-first-serve rationing that occurs due to the break in intake periods and capacity.

Once a case has been accepted, it is allocated by the executive director to a student based on availability, interest, and training (id.). In the main there is no attempt to allocate more student hours or higher skilled students based on the importance of the case (on any set of criteria) (id.). However, there are exceptions. When “it is clear that a case is going to be complicated (involving an upcoming jury trial, say) or is otherwise important (an appellate case that has the potential to set key precedent), [HLAB] will assign multiple students to the same case – and will ensure that a student with the requisite skills is on the team” (Grossman Interview 2011). Moreover, HLAB tends to assign less complex cases to 2Ls in the first two weeks of the semester as “training cases” (id.).

HLAB does sometimes handle appeals from its own cases. For housing and benefits appeals, more so than trial-level work, it has indicated more openness to prioritization of impact litigation (id.). Impact litigation is prioritized less on family law appeals largely because HLAB sees fewer opportunities for doing so (id.). For housing, and more rarely for benefits cases, HLAB does occasionally take appeals where it did not handle the case at trial, but usually only selects cases in this manner if they involve impact litigation (id.). HLAB’s retainer specifically indicates that its representation does not cover appeals, and the decision to handle an appeal in any case comes before the intake committee for a separate approval (id.). Moreover, in its new foreclosure initiative, HLAB does a significant amount of limited representation work where it counsels, advises, negotiates for, and makes court appearances for the hundreds of homeowners and tenants it has mobilized, all without full-representation retainers and all without going through HLAB’s formal intake process (id.).

HLAB uses limited representation agreements in two ways. First, it participates in an attorney-for-the-day program at Boston Housing Court, and limits its representation of clients in that program for that day of hearings (id.). Sometimes a client who seeks assistance as part of the regular intake process is told that their case is more appropriate for the attorney-for-the-day-program and advised to seek HLAB assistance through that program. Second, if there is a close call at intake as to whether the client is a good candidate for representation, HLAB will agree to handle a sub-set of the issues at first and then evaluate whether to take the clients’ other issues (id.). Similarly, clients occasionally present complex, involved, and/or very resource-intensive cases where the intake committee feels it more appropriate to work on only one discreet legal issue presented (id.).

That is the extent of HLAB’s formal guidelines for rationing. Nevertheless, members of its intake committee often implicitly bring up reasons to favor certain cases (some of which mirror those I discuss in Section 3) when deciding whether to take a particular client or not (id.). Rationing is also discussed at length in the law school’s Introduction to Advocacy and Advanced Clinical Practice courses that are required of HLAB members, in HLAB’s Public Policy Working Groups and members’ retreats, and informally in the hallways (id.).

2.2. Situating the Rationing Question

Before getting in to the substance of the analysis in this article, let me situate my focal question. In discussions of bioethics and health care policy, we can usefully separate out four questions: (i) Should the government fund or provide health care for poor individuals at all? Norman Daniels’ book length treatment of this issue in Just Health is one of the most prominent philosophical examination of the question, but there are other places one might look, such as Martha Nussbaum’s Capabilities approach argument (Nusssbaum 2000, 2006; Daniels 2008). (ii) How much should the government spend on assistance in health care for the poor, as opposed to other funding priorities, such as education? (iii) Imagining we had a set budget for health care services for the poor (such as in Medicaid, although in actuality that budget really is not fixed) or supply of a health care good (such as in the allocation or organs), how should health care resources be distributed to rival claimants when one cannot fully help all of them? This question, the rationing question, is the focus of most of the bioethics literature I draw on. (iv) Who should implement the rationing system we arrive at: individual doctors, hospital administrators, insurers, funders, or some combination of the above?

We could imagine four parallel sets of questions on the law side: (i) Should the state fund or provide legal services to poor individuals at all? 13 (ii) How much should the state spend on assistance in legal services as opposed to other funding priorities?14 (iii) Given a set budget for legal services for the poor, coming both from the state and charitable organizations, how should those resources be distributed to rival claimants when one cannot fully help all of them? (iv) Who should implement the rationing system we arrive at, individual LSP lawyers, office managers, funders, or someone else?

My paper is very self-consciously aimed at answering the third question and not the others. I largely (with some exceptions in Section 5 of the article) take the budget set for legal services as given and ask how it should be allocated, not how big it should be. I also largely put to one side the question of whether we should be funding legal services for the indigent at all, which could occupy an article in its own right, and has, as a philosophical and legal matter, been well covered by others (Luban 1984; Wertheimer 1988; Rubenstein 2002, 1873–1874; Solum 2004, 258; Rosen-Zvi 2010, 719). On the criminal side this right is well established in the Sixth Amendment of the Constitution, where the Court held that “any person haled into court, who is too poor to hire a lawyer, cannot be assured a fair trial unless counsel is provided for him” and the fact that “government hires lawyers to prosecute and defendants who have the money hire lawyers to defend are the strongest indications of the widespread belief that lawyers in criminal courts are necessities, not luxuries”, Gideon v. Wainwright, 372 U.S. 335, 344 (1963). There is no equivalent federal constitutional right to counsel on the civil side, although some state courts and legislatures have recognized such a right (Rulli 2011). One argument offered for such assistance stems from Mark Galanter’s famous observation, procedural rules, institutional facility, quality of lawyering, and the nature of the parties in litigation all give a decisive advantage to “Repeat Players” (RP) such as large corporations (Galanter 1974). Galanter among others has argued that free legal services for the “Have Nots” is essential, even if not sufficient, to change the way the system works to their disadvantage and level the playing field the playing field (id. at 139).

What about the fourth of the possible questions, who should do the rationing? Again, there is a parallel discourse in medicine, which in its polarities feature those who think it is monstrous and corrosive of the profession to have physicians be the rationers, and those who think it is a positive development to have physicians consider their job as promoting the health of the population more generally even if it means denying a particular patient beneficial but not cost-effective care (Strech et al. 2009; Callahan 2012; Gruenewald 2012). One could imagine a similar normative conversation about legal services. There are also pragmatic institutional design questions for which we currently have little available data: to what extent is the information gained through more detailed intake transferrable to the lawyers who actually handle the case if the case is selected, or is most of it “wasted” if we too heavily segment intake from litigation in terms of the relevant personnel? Are litigators in LSPs more susceptible to cognitive biases (such as confirmation and overconfidence biases) than would be a separate set of intake coordinators? Would LSPs performance be hampered by having those who do intake feel less ownership of the cases they select? And so on. All I want to say about these issues here, though I hope to pursue it in further work, is that what the rationing principles should be and who should apply them are separate questions, and the answer I give to the former is compatible with multiple possible answers to the latter, from individual LSPs’ intake managers to LSC or other funders doing the rationing. Currently on the legal side we have nothing like the complex architecture of hospital administrators and insurance claims managers that exist on the medical side. That may be for the better or for the worse. If, however, my project succeeds in getting lawyers, academics, and policy makers to think seriously about the rationing of legal services I envision a separate institutional design project aimed at how such rationing systems might be implemented at LSPs and whether we ought to be creating equivalent positions on the legal side to implement our rationing.

2.3. Why Bioethics?

Why examine the allocation of medical goods to illuminate how LSPs ration legal services? As a scholar who works at the intersection of law and bioethics, is this merely an instance of “looking over a crowd and picking out your friends” or is this the old joke about the drunk who looks for his keys under the lamppost? At the risk of self-deprecation, it is partially the latter in that there has been more than forty years of bioethics literature on (and experience in) implementing rationing systems for medical goods, making this an illuminating place to start.

There are also, however, important analogies between LSPs and hospitals, physicians, and public health systems that should draw us to the comparison. Both doctors and lawyers enjoy a governmentally enforced monopoly on providing their services that is achieved by strict licensing. This carries with it duties of professional responsibility as to their conduct. This also in some ways exacerbates the scarcity problem by preventing those outside the profession from offering the service in question, and in so doing increases the responsibility of those inside the profession to ensure that their allocation systems comport with the dictates of justice. This responsibility is heightened by the fact that LSPs, like hospitals and other health care institutions, are publicly and charitably funded such that it is appropriate to impose considerations of justice on them. For those who rely on LSPs, access to legal services makes a crucial difference to their life prospects, just as does access to ICU beds, organs, and health care in general. Finally, unlike other high-level rationing decisions—such as whether to build a particular road or power plant with attendant costs and benefits—LSPs and medical allocators are allocating assistance to rival claimants, rather than making more general action and omission investment decisions.

All of this is a good reason why the medical analogy makes sense as a starting point. That said, there are also important differences between the two contexts. Here are five: part of the benefit of legal services to a claimant is the recognition of her dignity that participating in the legal system affords her; while the medical system has achieved its ends when patients receive the care that helps them, we have more misgivings (especially on the criminal side) when a claimant receives legal assistance that gets them an outcome they did not “deserve”; legal services are an “adversarial good” that increases costs to opposing parties; there is more variation in the goals and institutional self-definitions of LSPs then their medical equivalents; and finally, the rules on professional responsibility and the right to counsel that are particular to the legal profession do not leave room for as robust rationing as on the medical side. There is much to be said here, and I examine the first four of these points in Section 4 of this article. I discuss the last point in this section because I view it as more of a threshold objection. While I provide full responses to each of these questions, I should make clear that I am not claiming a perfect analogy. All I claim is that the analogy is close enough to make bioethics literature on rationing a useful jumping-off point for LSPs that will enrich the discourse.15

2.4. Room to Ration?

At the threshold, one might worry that unlike medical service providers LSPs face greater difficulties in forecasting allocation needs ex ante, and face professional and right to counsel limitations on initial allocations and particularly re-allocations. LSPs do not necessarily know who will step in the door demanding services next. Health care systems also face this problem when rationing scarce goods like ICU beds, but over time they have been able to make predictions (based on their catchment areas and patient population), which LSPs could also do. This is what HLAB does, holding off filling slots at some of its intake meetings based on anticipated needs down the road. LSPs will also find it hard to estimate how much time an individual case will take, a problem also faced by their medical brethren and sistren, but over time they can make fairly good judgments about how much time a group of cases will take (Interview with Eppler-Epstein 2012).

Moreover, the allocation of legal services can in theory be adjusted over time in a way most rationed medical goods cannot. If on Day 1 of the LSP’s existence a client shows up with a claim to resources, and on Day 50 a client with a superior claim to such resources appears, in theory the LSP could re-allocate more time from the first to the new client on Day 50.

That sounds good in theory, but would such rationing violate rules of professional responsibility? To be clear, here I am not discussing the merits of these rules but instead examining them as a real world current potential limit on what LSPs might do, whether well grounded or not. There is no obligation for a civil LSP to take on a particular client vel non, and attorneys have considerable discretion in deciding which clients to take. Once they have started to represent a client, however, there is a responsibility to give competent representation. Model Rule of Professional Conduct 1.16 demarcates some limited situations where an attorney who has begun representation can withdraw, including where “withdrawal can be accomplished without material adverse effect on the interests of the client”. These restrictions are not a problem for the re-allocation I am contemplating, however, because the lawyer is not withdrawing or abandoning the client.

Can re-allocation nonetheless constitute de facto abandonment? There is no violation of these rules as long as re-allocation from one client to another never puts the first client below the floor of “competent representation” that is required by the rules. This would suggest that, at most, the provider has an obligation to meaningfully inform the client of the possibility of re-allocation in the future so he or she can instead choose an alternative provider.16 As long as a baseline of competent representation will be given to every client, the rules do not prohibit redistributing resources above that line.

In the private firm context, several courts have found a professional responsibility violation when a law firm drops a client like a “hot potato” or substantially changes its representation of a client to meet the needs of a new more remunerative client, although these cases center on the conflicts-of-interest rules. E.g., Picker Intl, Inc. v. Varian Assocs., Inc., 869 F.2d 578, 584 (Fed. Cir. 1989); Santacroce v. Neff, 134 F. Supp. 2d 366, 367 (D.N.J. 2001) (citing Stratagem Dev. Corp. v. Heron Intl N.V., 756 F. Supp. 789, 794 (S.D.N.Y.1991); Harte Biltmore Ltd. v. First Pa. Bank, 655 F. Supp. 419, 421 (S.D. Fla.1987)). The equivalent Restatement (Third) of the Law Governing Lawyers (2000) rule makes clear that the problem is limited to cases where the second representation is on the “same or a substantially related matter in which the interests of the former client are materially adverse”, meaning that matter “involves the work the lawyer performed for the former client” or a risk that it “will involve the use of information acquired in the course of representing the former client, unless that information has become generally known”. That is inapposite to our context.

For these reasons, there is no clear rule prohibiting rationing by re-allocation of the type I am contemplating. Suppose, however, I am wrong about that. After all, the matter is far from clear. Even if there is a professional responsibility issue here, it could be cured by a limited representation device. According to Model Rule of Professional Responsibility 1.2(c): “A lawyer may limit the scope of the representation if the limitation is reasonable under the circumstances and the client gives informed consent.” The comment to the rule further elucidates that “limited representation may be appropriate because the client has limited objectives for the representation”, and that “the terms upon which representation is undertaken may exclude specific means that might otherwise be used to accomplish the client's objectives”, including “actions that the client thinks are too costly or that the lawyer regards as repugnant or imprudent”. Model Rule of Professional Responsibility 1.2(c) cmt. 6 & 7. While recognizing the “substantial latitude to limit the representation, the limitation must be reasonable under the circumstances”. The comment gives the example of restricting representation to a brief telephone call for “a common and typically uncomplicated legal problem”, but cautions that such a limitation would not be reasonable “if the time allotted was not sufficient to yield advice upon which the client could rely” (id.). Finally, the comment makes clear that although “an agreement for a limited representation does not exempt a lawyer from the duty to provide competent representation, the limitation is a factor to be considered” (id.).

All this suggests that, at least in the forty-six states where they are allowed, LSPs could use limited representation clauses to reserve the right to re-allocate services to other clients with more pressing needs, so long as it does not result in incompetent representation to a client. In terms of informed consent, as one authority puts it the lawyer “must clearly explain the limitations of the representation, including the types of services which are not being provided and the probable effect of limited representation on the client’s rights and interests”, and where it is “foreseeable that more extensive services probably will be required” the lawyer may not accept the engagement unless “the situation is adequately explained to the client” (Althoff 2004, 71 (citing Comment, Colo. RPC 1.5; Ethics Opinion 101, January 17, 1998)). For pro se litigants this ought to “include a warning that the litigant may be confronted with matters that he or she will not understand” (id.).

Currently, limited representation agreements usually take the form of limiting particular lawyer activities, for example representing the client in negotiations but not in litigation or in pre-trial proceedings should the case go to trial (id. at 72). Our earlier discussion showed that HLAB engages in this kind of limited representation practice already. LSPs could make even more use of such practices, specifying specific issues or court proceedings in which they will represent the client. It may also be possible to push the envelope a little here—much will depend on the state in question and its specific rules—and be a bit less concrete in that they involve the resources devoted to representation rather than the stages of representation. Indeed, a limitation on the number of lawyers involved or hours devoted may be easier to understand than distinctions between pre-trial and trial representation. In any event, even more traditional limited representation frameworks can give significant room to ration. That said, even if not formally prohibited by the rules, I realize that many will feel queasy about the notion of re-allocating resources away from existing clients just as many doctors are resistant to the notion of not doing everything they can for a patient at whatever cost, even when not doing everything would better serve the health of the population17. Those particularly bothered by re-allocation could restrict the rationing principles I propose to initial decisions of whether to take on a client. A middle course would be to impose a higher “burden of persuasion” in the re-allocation context, which would permit re-allocative rationing only in cases where the gains would be particularly compelling based on the principles I argue for below. In any event, most of my analysis will focus on the initial allocation decision rather than the re-allocation decision, with the hope of giving more specific focus to re-allocation dilemmas in future work.

That answers the professional responsibility side of the issue, but what about the constitutional right to counsel? On the criminal side, as mentioned above, Gideon v. Wainwright and its progeny have held that all defendants have a right to counsel that attaches for “trial-like” critical stages of adjudication if actual imprisonment could be the penalty. See Gideon v. Wainwright, 372 U.S. 335 (1963); Argersinger v. Hamlin, 407 U.S. 25 (1973); Scott v. Illinois, 440 U.S. 367 (1979); Rothgery v. Gillespie County, 554 U.S. 191 (2008), (Metzger 2003, 1645). This excludes most probable cause hearings, most misdemeanors, and prison disciplinary proceedings, See Gerstein v. Pugh, 420 U.S. 103 (1975); Minnesota v. Murphy, 465 U.S. 420 (1984); Nichols v. U.S., 511 U.S. 738, 743 n. 9 (1994), and possibly more, see Rothgery, 554 U.S. at 217 (Alito, J., concurring). Where Gideon and its progeny do not apply, they obviously do not limit LSPs’ ability to ration. However, even where the Gideon and its progeny apply, it will not substantially limit rationing of the kind I discuss. While in a case involving paying counsel the Supreme Court has held that an erroneous deprivation of a defendant’s choice of counsel can constitute reversible error, there is no equivalent right of an indigent defendant to choose his counsel. United States v. Gonzales-Lopez, 548 U.S. 140 (2006). There is no right to be represented by a particular high-quality LSP (such as PDS) or to receive a certain amount of resources from the LSP that represents you; so long as the representation is “effective”, Gideon and its progeny are satisfied, and indeed the Supreme Court has rejected the suggestion that this even requires a “meaningful attorney-client relationship” (Morris v. Slappy, 461 U.S. 1, 14 (1983)).18 Therefore, the criminal right to counsel jurisprudence constitutes at most a de minimus constraint in LSP rationing of the kind I will discuss below.

On the civil side, as we saw, the right to counsel is even less of a constraint on rationing (Lassiter v. Department of Social Services, 452 U.S. 18 (1981); Gardner 2007, 64). Even in the few states that have by legislation or court decision recognized a civil right to counsel (Rulli 2011), there is no legal right to a particular counsel or anything more than competent representation, such that there remains much room for rationing.

In sum, both civil and criminal LSPs have significant freedom to decide whether to allocate resources to claimants in the first place, and even whether to re-allocate resources between existing clients. I now turn to how they should make such decisions.

3. PRINCIPLES FOR JUST RATIONING OF LEGAL SERVICES

Civil and criminal LSPs inevitably face rationing questions. As the discussions of PDS, CLS, and HLAB in Section 2 suggest, these organizations do—at least implicitly—engage in some rationing. However, as this section will show, their rationing is fairly rudimentary compared with what they might do. I draw on bioethics principles to develop a more robust approach to just rationing of legal services.

I begin this section with an obviously skeletal and simplified view of what a LSP does and what its goals should be. I assume that both the funding stream of these institutions and the satisfaction of their lawyers are entirely unaffected by their choice of rationing system. I assume that the goal of LSPs is to produce good outcomes for their client population (in the specific sense of the outcomes their clients desire), just as the doctor’s mission is to improve the health of her patients. I then ask, under these assumptions how should LSPs ration? All of these limitations and assumptions are relaxed in the next section, where I ask among other things: should it matter that that student legal aid bureaus like HLAB also have educational goals in addition to their goal of producing good outcomes? Should it inflect the analysis that providing PDS assistance based on the rationing approach of this section will sometimes end up helping factually guilty clients avoid criminal liability? And so on.

In this section, I describe six types rationing of principles and how they might be applied to legal services. The principles I cover are: first-come-first-serve, lottery, priority to the worst-off, age-weighting, best outcomes, and instrumental forms of allocation. I examine them one-by-one in a series of hypotheticals and analyses. I call these “simple” rationing principles to contrast them with more complex combinations of these principles that I will take up below, but, as my discussion reveals, they raise normatively complex issues and are better thought of as families of possible rationing principles.

At the end of this section, the reader will understand how these “simple” principles might apply to the rationing of legal services, and the normative arguments for and against each principle and its variants. There is a lot of content here, so I summarize my conclusions by way of Table 1 at the end of this section. I also discuss how these simple allocation systems might be combined into more complex systems, again looking to medical allocation systems for illumination.

Table 1.

Summary of “simple” rationing principles for legal services

Principle type/principle Rejected or endorsed? 
First-come-first-serve  
    First-come-first-serve Rejected 
Lottery  
    Weighted lottery/lottery Endorsed only to allocate between rival claimants when all other endorsed rationing principles have already been applied 
Priority to the worst-off  
    Priority to the worst-off in the sphere of legal needs Endorsed for criminal but not most civil legal service allocation, and even as to criminal it should form only part of the definition of worst-off 
    Priority to the worst-off generally Endorsed for civil legal service allocation in full, and as part of the definition of worst-off for criminal legal service allocation 
Urgency/priority for identified lives Rejected 
    Statistical lives approach Endorsed 
Age-weighting  
    Age-weighting for quantity of life Endorsed. Most direct impact is to favor the young for assistance with death-eligible criminal offenses, but also potentially other criminal and civil legal needs with clear ties to lifespan. 
    Age-weighting for quality of life Rejected. However, best outcomes and contribution/indirect benefit rationing principles should de facto lead to significant preferences for younger claimants 
Best outcomes  
    Save the most lives/rule of rescue Rejected to the extent it gives avoiding imminent death additional importance over other possible outcomes 
    Save the most quality of life years Endorsed, subject to the qualification that small differences in outcomes should not have huge consequences in terms of allocation, a principle that can be accommodated via proportional priority models 
    Take into account chance of success Endorsed only to the extent we have confidence that LSPs can do it well, and there are not countervailing negative effects on attorney morale, harm to rejected clients, or attorney–client relations 
Aggregation  
    Aggregation I: save more rather than fewer people facing comparatively severe consequences if not helped Endorsed 
    Aggregation II: add together smaller benefits to many individuals such that it may outweigh a very large benefit to one or fewer individuals Endorsed 
    Aggregation III: add together very small benefits (“irrelevant utilities”) to many individuals such that it may outweigh a very large benefit to one or fewer individuals Endorsed, although more tentatively 
Instrumental  
    Responsibility-sensitive rationing Rejected 
    Contribution/indirect benefit/indirect Endorsed in part: endorsed for initial allocation more than re-allocation, and for contribution measures keyed to the impact/direct services distinction and number of dependents of claimants and deterrence of harm to client population, but not more robustly 
Principle type/principle Rejected or endorsed? 
First-come-first-serve  
    First-come-first-serve Rejected 
Lottery  
    Weighted lottery/lottery Endorsed only to allocate between rival claimants when all other endorsed rationing principles have already been applied 
Priority to the worst-off  
    Priority to the worst-off in the sphere of legal needs Endorsed for criminal but not most civil legal service allocation, and even as to criminal it should form only part of the definition of worst-off 
    Priority to the worst-off generally Endorsed for civil legal service allocation in full, and as part of the definition of worst-off for criminal legal service allocation 
Urgency/priority for identified lives Rejected 
    Statistical lives approach Endorsed 
Age-weighting  
    Age-weighting for quantity of life Endorsed. Most direct impact is to favor the young for assistance with death-eligible criminal offenses, but also potentially other criminal and civil legal needs with clear ties to lifespan. 
    Age-weighting for quality of life Rejected. However, best outcomes and contribution/indirect benefit rationing principles should de facto lead to significant preferences for younger claimants 
Best outcomes  
    Save the most lives/rule of rescue Rejected to the extent it gives avoiding imminent death additional importance over other possible outcomes 
    Save the most quality of life years Endorsed, subject to the qualification that small differences in outcomes should not have huge consequences in terms of allocation, a principle that can be accommodated via proportional priority models 
    Take into account chance of success Endorsed only to the extent we have confidence that LSPs can do it well, and there are not countervailing negative effects on attorney morale, harm to rejected clients, or attorney–client relations 
Aggregation  
    Aggregation I: save more rather than fewer people facing comparatively severe consequences if not helped Endorsed 
    Aggregation II: add together smaller benefits to many individuals such that it may outweigh a very large benefit to one or fewer individuals Endorsed 
    Aggregation III: add together very small benefits (“irrelevant utilities”) to many individuals such that it may outweigh a very large benefit to one or fewer individuals Endorsed, although more tentatively 
Instrumental  
    Responsibility-sensitive rationing Rejected 
    Contribution/indirect benefit/indirect Endorsed in part: endorsed for initial allocation more than re-allocation, and for contribution measures keyed to the impact/direct services distinction and number of dependents of claimants and deterrence of harm to client population, but not more robustly 

3.1. First-Come-First-Serve

Many LSPs use some version of this form of rationing already, turning away clients when they are oversubscribed or when intake periods are closed. For example, HLAB does the latter, although within each intake period it attempts to smooth subscription. CLS also occasionally has to turn away clients when the lawyer working on a particular subject matter area in a particular geographic area is oversubscribed or away on leave. Moreover, for cases that arrive at CLS via the Statewide hotline, there has already been a de facto first-come-first-serve rationing approach, since many callers will get the “call back later” message and never call back.

First-come-first-serve privileges those with more expertise who know how to access and find services, as well as the person who shows up at a particular moment in time over those who are more needy or more deserving but show up later (Daniels 2005, 169–171; Persad 2009, 424). The fact that many LSPs, such as HLAB and CLS, condition access to their programs on financial need is one bulwark against gaming by the rich, a concern in organ allocation, for example. Still, among the population that meets that filter, first-come-first-serve will favor the more sophisticated or more persistent consumer of legal services, and there is no reason that doing so is just (Luban 1988, 308).19 Indeed, with something like CLS/Statewide’s hotline, we may worry that the individuals most in need of legal services are the ones who have the least resources to enable them to continually call the hotline with the hope of getting through.

There is no reason to privilege someone who shows up at a particular moment in time over someone later in the time stream, if the person who shows up later has a superior claim to assistance based on the factors to be discussed below. Even when they escape these problems, first-come-first-serve rationing ends up being, at best, an imperfect substitute for a lottery, a mode of rationing that I also find problematic for reasons discussed below.

Thus, it appears that one of the most common forms of rationing by LSPs seems unethical. That point alone is an important contribution to the literature. That said, there may be instances where LSPs cannot avoid some use of first-come-first-serve. For example, PDS often sends lawyers to courthouses to “pick up” new clients in a way that both PDS and the courts find essential. Where LSPs cannot avoid using first-come-first-serve altogether, they ought to minimize its importance. This might be accomplished by supplementing their existing first-come-first-serve intake procedures with additional client recruitment that follows the rationing principles set out below. Additionally, even when they cannot avoid using first-come-first-serve as to initial formation of an attorney–client relationship, they ought not to give those first in time priority in the intensity of resources devoted to their cases, and may also re-allocate resources to blunt the impact of first-come-first-serve, at least to the extent permitted by the rules of professional responsibility.

3.2. Lottery

Lotteries give everyone who is eligible and demands a service a chance of receiving the good, with the choice between claimants made through some random system such as a roll of the dice (Persad et al. 2009, 429). In simple lotteries, every eligible claimant is given an equal probability of getting the good. Such lotteries are often defended on the basis of a Kantian principle of respect for persons which instructs us never to treat others as “mere means” but always as “ends in themselves” (Kant 1785). In bioethics it is sometimes said that lotteries reflect the idea that “each person’s desire to stay alive should be regarded as of the same importance and deserving the same respect as that of anyone else” (Harris 1985).

A more sophisticated version of this approach uses a “weighted lottery”, which could incorporate some of the principles discussed below but also retain some element of randomness. Metaphorically, weighted lotteries use dice that are weighted toward certain numbers, but still have a chance of rolling up on the unweighted ones. Some bioethicists argue in favor of weighted lotteries on the theory that they give everyone “some chance”, even if a low one (Brock 1988, 88; Weir 1995, 100).

Lotteries are very easy to implement in legal service provision. Indeed, in a recent experiment run by my colleague Jim Greiner, HLAB actually implemented a lottery system to determine whether it took clients in one of its practices in order to create the equivalent of a randomized clinical trial of its legal services (Greiner & Pattanayak 2012).

The very thing that makes some favor lotteries—that everyone has a chance of getting the desired good—is an ethical problem because they sometimes result in someone who is less deserving getting the good over another, more deserving, person (Elhauge 1994, 1501; Persad 2009, 423). As Luban (1988, 309) notes, “by presuming not to judge the worth of cases, a lottery mechanism in effect indulges in the fiction that such judgments of worth cannot be made by one person about another’s projects and problems”, and it problematically “treats every potential client with equal concern and respect only by treating every potential client with no concern and respect”.

For these reasons, LSPs should not rely on lottery allocations when they can use other criteria to ration more justly. However, allocation by lottery may be useful when other more just rationing principles are impracticable, or after these principles have been used and LSPs still have more equally situated claimants than their resources can accommodate.

3.3. Priority to the Worst-off

This is essentially Prioritarian distribution principle. Prioritarians do “not give equal weight to equal benefits, whoever receives them”, but instead give more weight to “benefits to the worse off” (Parfit 1997, 213). In allocating medical goods, bioethicists think of this principle as “Sickest First”: give priority to those with the worst prospects if left untreated (Persad et al. 2009, 424).

Why give priority to the worst-off? The argument is typically defended on Prioritarian grounds that the “moral value of achieving a benefit for an individual (or avoiding a loss) is greater, the greater the size of the benefit as measured by a well-being scale, and greater, the lower the person's level of well-being over the course of her life apart from receipt of this benefit” (Stanford Encyclopedia of Philosophy 2008). On this view, the person who deserves priority for a resource (such as an LSP’s assistance or an organ) is the person whose life will be worst-off without it. For Prioritarians, this is true even if the unit generates the same improvement of welfare for either person. For instance, if an organ will give whoever receives it five years of life, Prioritarians would give it to the person whose life would go the worst without it. Indeed, Prioritarians sometimes favor allocation to the worst-off even if the resource generates less welfare for that person than a better-off claimant.

This dispute is currently being played out among LSPs in the housing context, to give one example. Some advocates argue that LSPs should prioritize representation of very low-income tenants faced with eviction from low-income housing. Others argue that LSPs should divert resources to foreclosure defense because, even though foreclosure clients are “better-off” more generally, the intervention is likely to have a more positive effect in terms of outcomes, both individually and systemically (Grossman Interview 2011).

Priority to the worst-off can sometimes also be grounded on Utilitarian theories committed to maximizing good states of the world. Due to diminishing marginal utility from resources, access to the good is likely to “raise the welfare of the poor more than it is that of comparably richer individuals” in cases where being well-off is measured by wealth (Fisher & Syed 2007, 602–605; Cohen 2011, 18). If anything, this approach seems more plausible as to legal services than health resources, because there is likely to be less diminishing marginal utility for health resources.

While, in theory, priority to the worst-off might lead to moral hazard problems where individuals try to become worse off to gain priority, in practice this seems unlikely for legal services: it requires unrealistic amounts of knowledge about the condition of other claimants, legal services are unlikely to fully compensate for the requisite self-injury, and one would bear a huge risk of worsening one’s condition and still not receiving the service.20

To put “priority to the worst-off” into practice requires taking a position on several applied ethics debates I now discuss.

3.3.1. Worst-Off Globally? The Question of Separate Spheres

How global is our assessment of who is “worst-off?” One could allocate in favor of claimants with the greatest legal need, that is those whose life will go worse in a legal sense without the aid. Or, one could adopt a more general conception of need, and consider how badly the person’s life will go over all domains without the legal assistance. To illustrate, imagine that J.R. is an oil tycoon who has lived a life rich in all goods and pleasures has recently lost it all. Bobby, in contrast, has received much more of the short end of the stick in life, including being born blind. Both are now indigent and have been accused of crimes. J.R. has been accused of embezzlement, and faces twenty-five years in prison, while Bobby has been accused of criminal fraud and faces twenty years in prison. Under the “legal need” criteria, J.R. should be favored because his life will go worse in terms of the legal domain than Bobby’s if he is denied aid. In contrast, if the question is whose life will go worse in all domains, there may be reasons to favor Bobby.21 While devising a comparative measure of how needy two individuals are on all domains is practically quite difficult, we could use something like expected lifetime earnings as a very crude approximation.

In allocating medical goods, bioethicists face a similar choice between allocating based on greatest medical need or “priority to the generally worst-off” (Brock 2002, 368). Choosing between the two variants depends on the position one takes as to what is known as the “separate spheres” debate (Brock 2003; more generally Walzer 1983). Priority to the overall worst-off threatens to lead to some initially counter-intuitive results in allocating health goods in that “we would have to give lower priority to treating the rich than the poor, even when the rich are much sicker than the poor, if the overall well-being of the rich is higher despite their much worse health” (Brock 2002, 368). Because access to legal services are usually means-tested to begin with, and those who can afford better care can buy it outside of the public distribution system (unlike organ donors), one might think we have dodged this problem. In fact the problem persists, albeit in more subtle form. Among those who fall beneath 185 percent of the FPL, which is CLS/HLAB’s cut-off, there will be some whose overall well-being is very low (e.g., at 20 percent of FPL) but whose legal needs are not as great, and those with the opposite configuration, reproducing the dilemma.

I believe that the way to resolve this question is to reflect on why we give priority to the worst-off in the first place. Whether grounded in Prioritarian or Utilitarian justifications, priority to the worst-off is focused on the fact that these individuals’ well-being is overall worse off. Therefore, it seems natural to say that claimants deserve priority as to a particular element of well-being in so far as it will increase their overall well-being, which is our goal. A disadvantage “a disadvantage in one aspect of well-being can be compensated for by an advantage in a different area”, for example, that “loss of income from taking a less pressured job may be compensated for by increased time to spend with one’s family” (id.).22

There are, however, arguments, for the sphere-specific interpretation. The philosopher Frances Kamm has argued that when it comes to the “sphere of lifesaving enterprises”, such as health care allocation, not all resource expenditures are fungible such that “the goal of improving health or lifesaving is sufficient unto itself, and there is corruption in the achievement of this aim if achieving some other good is combined with it in the selection of persons, even if as much life or health is attended to” (Kamm 1993, 259). when it comes to allocating health care resources, we violate the Kantian prohibition on treating individuals as mere means rather than ends in themselves if we allocate based on criteria of need other than those pertaining to saving the person’s life or health (id. at 369).

This version of the argument leaves me relatively unpersuaded, in that it seems just as compatible with the Kantian prohibition—indeed possibly more compatible—to treat individuals as beings whose holistic well-being matters, such that our goal ought to be to give overall better lives to those with overall worse lives and view health care allocations as merely one tool to do so.

A different re-construction of the argument is that health care is special as an “all-purpose means necessary for the pursuit of nearly all people’s aims and ends” such that “its loss may not be able to be compensated for by other goods”, and therefore it should have its own separate priority division rather than allocating it to those who are generally worse off (Brock 2002, 368). Even if that argument holds true for health care—and it seems only partially true even there—the claim is somewhat more strained as to legal services, which are more substitutable for other elements of well-being. This is especially true on the civil side where legal services are primarily valued to maintain access to other goods such as housing, wages, governmental benefits, etc. Nonsubstitutability is more plausible on the criminal side because what will be lost if one is sentenced to jail or death—one’s life, years of one’s freedom—seem closer to being uniquely valuable and not as commensurable with other goods. That said, there may be some civil matters—such as loss of custody of one’s child, orders to enhance physical safety in domestic violence cases, special education rights—for which this also true. I think the right approach is to group these kinds of cases with the criminal ones for the purposes of this analysis, rather than to draw a sharp criminal versus civil line.

Putting this category of cases to one side, for the remainder of civil cases, the overall well-being criteria generally seems most justifiable for allocating legal services, while for criminal legal services there is a stronger argument for allocation based on the sphere-specific harm of the criminal sentences faced by the accused. That said, even on the criminal side I think a hybrid measure that gives extra weight to the legal sphere but still takes into account overall well-being is warranted. In practice, because the effect on one’s overall well-being of the death penalty or a long prison sentence is enormous, these distinctions in metric are unlikely to make much of a difference. Moreover, even where overall well-being seems normatively more desirable, as the metric for being worst-off it is unlikely that LSPs can, in a cost-effective or accurate way, adequately measure it. When administrability constraints are factored in, it is much more likely they will only be able to use a crude proxy, perhaps income level in relation to the FPL, to sort claimants.

3.3.2. Worst-Off When? (Statistical versus Identified Lives)

We also face a temporal question of timeframe: worst-off when? One approach focuses on who is in the most imminent need at the time of allocation, an issue of “urgency”. The other approach attempts to determine who will be worst-off without the good over a lifetime. This can be viewed as an instantiation of a more general problem in rationing known as the “statistical versus identified life” problem (Adler & Williams Sanchirico 2006, 355; Daniels 2011). As Charles Fried (1969, 1416) classically put it:

We are prepared to expend far greater resources in saving the lives of known persons in present peril, than we are prepared to devote to measures that will avert future dangers to persons perhaps unknown or not yet even in existence. The anomaly arises insofar as a consistent policy favoring the expenditure of resources for persons in immediate peril can be shown in the long run to lead to a smaller number of lives saved, with the same or perhaps even a larger long-run expenditure of resources.

In bioethics, Norman Daniels (2011, 7) has recently illustrated the issue with a hypothetical that tries to remove the related aggregation problem:

Suppose we have only five tablets of a medicine that can be used either as an effective treatment for a terminal disease, provided that all five tablets are given to the person who has contracted the disease, or as an effective vaccination, given in one-tablet dose to five people exposed to the disease. Without vaccination, there is 20 percent chance of contracting the disease once exposed. So the two cases look like this:

Treatment: Alice has the disease. We can give her the whole dose – all five tablets.

Vaccination: Betty, Cathy, Dolly, Ellie, and Fannie have been exposed to Alice. We can vaccinate all of them with one-fifth of the dose we can give Alice.

In both cases, we can suppose that one expected life saved (thus avoiding issues about aggregation). In addition, the people at risk in the vaccination are identified with regard to who will receive preventative treatment, though we do not know which of them will get the disease if they are not vaccinated.

Since one statistical life will be saved either way, if we truly believed one of the five vaccination candidates would die in the future, favoring Alice can only be justified by favoring identified lives and the imminence of the threat over statistical lives. On the other hand, if we dismiss the preference for identifiable lives as merely a bias the system should avoid, then our views about the treatment/vaccination case should be the same if we changed the numbers such that there were 100 pills needed for treatment, one for each vaccination, and each of the potential victims had only a 1 percent chance of infection, and so on (id.).

In making rationing decisions, LSPs face a similar dilemma. If PDS takes on the defense of someone facing a five-year sentence in February, it may not have a lawyer available to take on a client facing a twenty-year sentence later in the calendar year.23 HLAB does its best to avoid favoring those facing imminent risk of eviction today over those that face it later on in their intake period, but it turns away people (or at least asks them to wait) when its intake periods are closed. I suspect that some LSPs are even less careful on this score, but I have not undertaken an exhaustive survey of their practices. More abstractly, resolving these issues is important in determining whether LSPs can permissibly engage in impact litigation—focused on the interests of potential statistical clients who are “out there” but have not made claims—or mobilization lawyering—aiming for system reform that eliminates individuals’ need for litigation in the first place (Tremblay 1999, 2502–2504).

Most bioethicists have taken the view that the preference for identified lives is a mistake, that it “myopically bases allocation on how sick someone is at the current time – a morally arbitrary factor in genuine scarcity” (Fried 1969, 1420–1428; Brock 2004; Persad et al. 2009, 425). Allocations based on identified lives’ imminent need can make sense in situations of temporary scarcity, when more of the good is on its way and if we do not give priority to the person about to die they will not be someone we can save when the “ship comes in”; in contrast, this is not true for the persistent scarcity that is the case for most organs and for LSPs’ services (Brock 2004, 42). Giving a resource to the client or patient who needs it most urgently will not increase the number saved, and when the less urgent claimant needs it later there will still not be enough to go around.

The minority view in favor of individual lives tends to focus on notions of the “symbolic value”, that we “demonstrate the value we place on human life” by showing preference for “individuals in immediate peril” (Fried 1969, 1425). I tend to agree with Fried that such arguments for favoring identified lives in bioethics are “confused, wrong, or morally repugnant” (id.). It is unclear why showing respect for identified lives better captures this symbolic value of life; indeed, one might think it is statistical lives that captures the notion that all lives are equal and of the same value. After all, what is spent on those in immediate need must come from allocations to others not present (id.; Persad et al. 2009, 425). The argument is still weaker for resources that do not save lives but merely make them better, which is true of many legal services.

Another less deontological argument for favoring identified lives suggests that there is additional utility derived by rescuers and by society in general from the notion that heroic attempts have been made to save a life (McKie & Richardson 2003, 2411–2417). There are numerous problems with this argument. First, where is the evidence for this utility bump? Second, the argument depends on the “rescue” having high media visibility. There are some contexts where this may occur. Consider one episode of The Simpsons, where Bart mischievously places a walkie-talkie in a Springfield well and convinces the people of the town he is a little boy stuck in the well—leading to massive spending to try to rescue the boy and even a charity fundraiser led by Sting that produces the song “We Are Sending Our Love Down a Well”, a potshot at “We Are the World” (The Simpsons 1992). In contrast, in the typical case taken on by a LSP the “rescue” is largely invisible (McKie & Richardson 2003, 2416).24 Third, we might derive the same or larger utility bump by educating people on the importance of saving more life years or even more people, by focusing on statistical lives. Fourth, if we decide that the preference for dramatic rescue is ultimately based on an irrational favoring of identifiable over statistical victims, on some theories of preference-laundering we ought to exclude any utility derived therefrom (Sumner 1996, Chang 2000, 183–194). After all, if it turned out antipathy toward the obese (or women, or minorities) was such that their rescue garnered less utility for rescuers or the public, most would resist the idea that these groups deserved less priority for this reason. Why should things be different here?

Daniels has recently offered a series of more subtle arguments for favoring identified lives in order to show that the question is at least an open one, including: the identified lives approach better motivates life-saving in the long run because the statistical approach requires us to ignore our psychological drives and partially dehumanizes us, and that people with current needs have an arguably stronger claim of entitlement than those with a mere chance of future needs (Daniels 2011, 4–7). These arguments seem speculative to me, but we need not resolve the issue once-and-for-all since our task is only to evaluate the argument as to rationing legal services. Because the rationing is done through alterations to the intake process, and because LSPs will always be helping someone even if not this particular person at this particular time, the arguments about dehumanization seem weaker in this context; indeed it may be more humanizing to be able to re-allocate to those now more in need, at least when prior clients are not being “abandoned” but merely given fewer resources. This is particularly true when the attorneys and the clients are informed of the rationing rules up front.

For these reasons, in the context of LSPs, the statistical rather than the identified life conception of what it means to help the worst-off seems more defensible. This does not mean that identified lives do not count, but rather that they deserve no preference over the equivalent number of statistical lives. Among other implications, this means that LSPs have no reason to favor direct services over impact litigation or mobilization lawyering if they were certain that the same number of clients would be helped. That the claimants in direct services work are identified lives is of no moment.

Being committed to priority to the worst-off does not tell you how much priority to give those individuals. Answering that question depends in part on whether you support priority to the worst-off for Prioritarian or Utilitarian reasons. For Utilitarians, priority to the worst-off is only justified because one gets more “bang for your buck” in helping them due to diminishing marginal utility, such that they should get only the amount of priority dictated by the slope of the respective utility curves. For Prioritarians, in contrast, the priority is fairness-based. How much priority Prioritarians should give depends on one’s distributive justice function. The two extremes on the continuum are easy to see: priority for the worst-off could be used to break ties where individuals are matched on all other criteria, versus something like John Rawls’ Difference Principle, which gives maximal priority to the worst-off for primary goods to the extent it benefits them to do so (Rawls 1999, 2001, 172; Daniels 2008). In between are a huge number of possibilities, and LSPs like medical allocators can reasonably disagree on how much priority should be given.

3.4. Age-Weighting (Youngest First)

3.4.1. Age-Weighting and the Quantity of Life

Age-weighting is a rationing principle that gives priority to those who are younger as opposed to those who are older, which in bioethics has been applied to organs, vaccine doses, and other medical goods (Williams 1997, 117–132; Persad et al. 2009, 425). Age-weighting can easily be applied to access to legal services by privileging younger claimants’ need for civil or criminal legal assistance. Because age is objectively verifiable from government-issued identifications, age-weighting would be among the easiest rationing principles to apply. HLAB and CLS currently do not do have an intake policy that uses age-weighting.25 To the extent PDS takes age into account, it is by allocating less experienced lawyers to those facing juvenile detention, although this appears to reflect the potential length of the sentence and that juvenile cases that proceed to trial are generally not as complex and more appropriate to the skill level of newly trained lawyers still receiving close supervision, rather than attempts to age-weight (Leighton Interview 2011).26 PDS like most LSPs gives intensive training before trial lawyers begin to pick up cases, and more periodic training throughout a lawyer’s career (id.)

There are two quite different justifications offered for age-weighting. The first stems from fairness: that is, everybody should as much possible have the opportunity to live a “complete” life (often called “fair innings”) and therefore those who are older and have had more of their share of life years ought to get less priority than the younger who have not (Harris 1985, 87–102; Bognar 2008, 178–179).27 Such an approach, it is argued, does not constitute discrimination because unlike sex- or race-based preference, everyone will have been young at one point and would at that point have received priority (Persad et al. 2009, 425 (citing Daniels 1988); but seeKerstein & Bognar 2010 (arguing that very young children denied life-saving health care goods will never “benefit from an arrangement that gives priority to young adults”)). The second rationale, which underlies the WHO’s age-weighting in allocating resources toward the global burden of disease, is that the young are more productive and therefore giving them priority “grows the pie” (Williams 1997, 127; Bognar 2008, 174–175). This approach is more akin to the instrumentalist indirect benefit principles I will discuss below.

Within the camp of those who advocate for prioritization based on age, there are further debates about when the priority should “start” and “end.” In bioethics, for example, Govind Persad, Alan Wertheimer, and Zeke Emanuel, argue that priority should increase from age zero to age twenty, be relatively flat from twenty to thirty, decline somewhat from thirty to fifty, then decline more sharply from fifty to seventy (Persad et al. 2009, 425–428). Ronald Dworkin (1993) has pushed a similar idea (Persad et al. 2009, 428).28

The fair innings justification for age-weighting in bioethics applies most strongly for LSPs when it comes to allocating resources to accused individuals facing death-eligible sentences. Just as allocating ICU beds and organs have the intended effect of extending life, so does legal assistance of this kind. Thus, age-weighting would lead us to prioritize younger death-eligible defendants older ones in terms of resources allocated.29

But legal aid to death-eligible defendants is but a small part of all criminal and civil legal aid. Another sub-category of legal aid might be thought to indirectly extend life without too much attenuation. Time spent in prison exposes inmates to significant health risks, including “elevated prevalence of communicable disease … [h]igh levels of violence, including sexual violence” and increased HIV transmission due to “consensual sex without condoms and unsterilized tattooing”, leading to reductions in lifespan (Hogg 2008, 4). This would justify age-weighted prioritization for criminal offenses that involve only prison time. Some civil legal needs, most notably housing and medical benefits, also seem plausibly connected to lifespan without too much attenuation. Therefore, for these services there is also a strong argument for age-weighting based on “fair innings”, although only in proportion to the effect those services have on clients’ lifespans.

3.4.2. Age-weighting and the Quality of Life

What about age-weighting for legal services that improve the quality of clients’ lives but not the quantity of life years? Again, this is most visible in the criminal law context. If a defendant is allocated legal assistance she would not otherwise get, she stands a better chance of avoiding or diminishing time spent in prison and improving her quality of life as against the counterfactual of incarceration. An older individual facing criminal prosecution ordinarily already enjoyed more years of high quality nonimprisoned time than one who faces criminal prosecution earlier in life.

To the extent civil legal interventions improve the quality of litigants’ lives compared with a counterfactual world without them, one can similarly argue for age-weighting. A year of life with housing, employment, or with health benefits is better than year of life without them. This just raises the question, though, if a twenty and a fifty-year-old each face two years of unemployment in their lifetime if they do not receive free legal assistance, does the mere fact that it happened later for the fifty-year-old give a reason why he should get less priority in allocation?

One might be tempted to suggest that there is a “domino effect” such that a denial of needed legal services early on will have negative repercussions for quality of life going forward, as an argument for favoring the younger claimant. This seems quite plausible on the criminal side—given the effect of incarceration on one’s future employment and other life prospects—and on the civil side—denial of unemployment or disability benefits, loss of housing, etc., are likely to produce a cascade of negative life consequences. Upon reflection, however, this is not actually an argument for age-weighting as such. Instead, it is an argument about differential sizes of benefit from providing the service for people at different points in the lifespan, a form of the best outcomes rationing metric I discuss below. Expending the same amount of resources “buys” a larger amount of benefit if we happen to give it to the younger, but we are not giving it to the younger for outcome-independent reasons.

It is helpful to illustrate this point using the “Quality Adjusted Life Year” (QALY) measure, which I will discuss in greater depth below. This is a Cost Effectiveness Analysis health care allocation tool developed by the WHO, NICE (the UK health rationing agency), and others (McKie & Richardson 2003, 2407–2408). A year of perfect health is treated as 1.0 QALY, while death is treated as 0 QALY (id. at 2408). A year of life with any illness or disability can be placed somewhere between 1.0 and 0 on this scale; for example a year on dialysis is scored at 0.57 (id.). Typically, QALY measurements are generated by testing large numbers of people using one of three techniques: time-trade-off, standard gamble, or the visual analog scale (id. at 2412).

While QALYs have been used to rate and compare health states, similar methods could be used to generate the equivalent metric for legal states of being, “QALeYs” if you will for “Quality Adjusted Legal Years”. Again, application on the criminal side is easiest to imagine. If a year of ordinary life is rated a 1, we could determine how much a year incarcerated in a maximum (or medium) security prison is worth to the individual. We could also do the same thing for a year without housing, etc.

Applying age-weighting to the quality and not just the quantity of life means that a QALY (or QALeY) allocated to a fifty-year-old is not valued the same as one allocated to a twenty-year-old. Suppose that a year on kidney dialysis is valued at 0.57 QALY. If age-weighting does not apply, all other things being equal, providing a kidney transplant or other intervention which will save a fifty-year-old one year on dialysis is equivalent to the same intervention for a twenty-year-old, it results in a gain of 0.57 QALY per year of life. In contrast, if we age-weight on quality, we ought to add a multiplier to the fact that the 0.57 QALY gain is going to someone who is younger (Williams 1997, 128).

We theoretically could adopt the same principle for LSPs and give priority to younger individuals for services that only improve quality and not quantity of life. We should reject such a principle because the reasons for age-weighting do not apply to quality-of-life-improving legal services. There is no argument from justice for preferring preventing a bad year—a year of unemployment or imprisonment—in someone’s twentieth rather than fiftieth year of life. It may be that the fifty-year-old who must go to prison or who loses his job suffers more, or that the opposite is true because of a domino effect. However, if this is the reason to favor certain ages, then it is no longer age-weighting but instead a form of best outcomes rationing.

In sum, I believe that both civil and criminal LSPs should age-weight and give priority to younger claimants for legal services that directly extend the individual’s life, as well as those services that do so indirectly in ways that are relatively verifiable. Where rationing legal services merely improves the quality and not quantity of life, age-weighting as such is not appropriate. However, other rationing principles, such as best outcomes, may still lead us indirectly to strongly favor younger individuals if it is the case that they will benefit more by being given legal representation when they are younger.

3.5. Best Outcomes

“Best Outcomes” is consequentialist and forward-looking principle that looks at how much an individual will gain (or avoid losing) from the intervention as opposed to other possible claimants. It depends both on the expected value of the outcome and how likely it is to occur (Persad et al. 2009, 425). It comes in three major variants, which I discuss in turn: save the most lives, save the most life years, and save the most QALYs.

3.5.1. Save the Most Lives or Life Years? The Rule of Rescue and the Specialness of Death

“Save the most lives”, as the slogan suggests, pushes for allocating based on the number of lives saved. Standing alone it would suggest that we treat all lives as equal in terms of the calculus and prioritize interventions that save a life over those that improve quality of life or add life years.

Adopting this principle for the rationing of legal services would mean that an LSP like PDS (in a jurisdiction where there is a death penalty) ought to heavily overinvest in the defense of death-eligible accused, because doing so saves lives as compared with merely defending those faced with long prison sentences.30 (Again, I am bracketing the question of likelihood of factual guilt and its effects on allocation until the next section). The same might be true for civil claimants with life-or-death needs, such as an individual suing their insurer for reimbursement for a life-saving treatment. The vast majority of both civil and criminal legal services, however, do not represent this kind of urgent life-or-death need, so adopting this principle would cause us to invest heavily away from these services.

Is such a principle merited? This raises the question of whether death is special in terms of allocation, often represented by what is called the “Rule of Rescue”, the “imperative to [] rescue identifiable individuals facing avoidable death” (McKie & Richardson 2003, 2407–2409). To evaluate this question one has to understand the simplest alternative: Save the Most Life Years. A bioethics example will prove illuminating. Suppose two individuals, twenty-five-year-old Smith and twenty-two-year-old Jones, are ill.31 Smith is in need of an emergency heart transplant, but due to many factors about his poor prognosis the transplant will gain him only an additional two years of life. In contrast, Jones is very ill and in declining health, but not immediately at risk of death, and his current heart will last him another three years. Still, for a number of reasons, if provided the heart today then Jones is expected to live to sixty years. Just to avoid the easy way out of the case, imagine that a heart becomes available for transplant that is a tissue matching for Smith and Jones but that the doctors do not believe that another such heart will become available before Jones will die. A principle of “Save the Most Lives” would favor allocation to Smith since he is in urgent and imminent need of the organ and will die without it; in contrast, maximizing the number of life years saved would favor allocation to Jones since he will gain thirty-five life years compared with Jones’ two life years.

Similar hypotheticals can be constructed for LSPs involving trade-offs between assisting a death-eligible older accused versus one or more younger accused facing long prison sentences, etc. The basic question is whether it is right to invest resources to save someone from imminent death when to do so we must forego giving more years of life to someone else.32 There is some evidence showing that when asked to hypothetically ration, many individuals adopt something akin to the Rule of Rescue (id. at 2408–2409). The issue is intimately related to the statistical versus identified lives question discussed above, but now the identified life will be imminently extinguished. The same arguments can be marshaled for the Rule of Rescue—the importance of affirming the symbolic commitment to the value of life, the utility derived by society of knowing that those in peril will be rescued, etc.—and the same responses I offered above show why I find the argument wanting. Indeed, it seems to me that the argument for the Rule of Rescue is, if anything, weaker than the more general argument for favoring identified lives. All interventions that gain the claimants life years actually “rescue” them from death at some point, so it is a conceptual error to frame the issue as rescuing those who face death and saving lives versus merely achieving more life years. To the extent proponents of the Rule argue for a burden to rescuers and society in having to watch people die when death could be avoided, such arguments seem even more hollow in the context of the delivery of legal services: unlike physicians, lawyers do not view saving lives as paramount to their “role morality” as opposed to achieving other good outcomes for clients. The allocation choices are very often made at the institutional level through intake processes such that individual lawyers will not bear the psychic burdens of watching one of their clients die in a way that could be avoidable (Tremblay 1999, 2518–2519), although this depends somewhat on how we set up the allocation scheme. Finally, these allocation decisions are largely off the radar of the general public, such that there is no “We Are Sending Our Love Down a Well” moment that is being given up.

For these reasons, LSPs like PDS and HLAB should not treat preventing the death of their clients as especially important and re-allocate resources in keeping with a Rule of Rescue. That said, in practice even rejecting the Rule of Rescue will generate significant reasons to allocate resources that will prevent imminent death because in some legal contexts—the death penalty is the clearest example—preventing an individual’s death counts as “saving” all of the life years they have left. This is why I suggested that even a best outcomes approach that did not incorporate age-weighting would give reason to allocate more to young claimants—preventing the execution of a 21-year-old as opposed a 50-year-old, for example, if we estimate a common life expectancy of 70 years, will save forty-nine life years as opposed to twenty.

Public Defender’s Offices currently spend dramatically more on death-eligible defendants. For example, a 2008 Maryland study suggested that the state spent an average of $1.9 million per death-eligible defendant, roughly $1 million more than on nondeath-eligible defendants (Liebman & Clarke 2011, 311 (citing Roman et al. 2008)). A study by the Kansas government found that trial costs for death penalty cases were sixteen times greater for capital than for noncapital murder cases (id., citing Performance Audit Report 2003). Other studies also suggest much higher costs for the court system as a whole in capital trials without breaking out the defense costs (e.g., Report to the ALI Concerning Capital Punishment 2010, 404–405). My analysis shows that such disproportionate spending appears unjust.

3.5.2. Should Quality of Life Matter?

3.5.2.1. The simple case

If life years are what matter, should all life years be treated the same or ought we to consider the quality of those life years in rationing? We have already discussed the QALY measure considers quality of life and how we could develop a similar QALeY measure for legal deficits. If available, should LSPs use these kinds of measures, or to put the question another way should we try to develop such measures? There are some difficult normative questions here, but at the threshold it is impossible that rationing of legal services could totally ignore quality of life. This is easiest to see on the criminal side. Suppose we were deciding whether to allocate a criminal defense claimant any legal assistance at all. Further, suppose I was to assure you that, contrary to the empirical evidence, those who are sentenced to prison do not face diminutions in their lifespan. If only lifespan mattered for allocation purposes, then we ought to be indifferent as to whether the accused spent a year of his life in prison or not, despite its quality effects. The whole reason for rationing the good in the first place negates that assumption, thereby proving that quality of life must matter in a threshold sense.

This alone suggests an important rule of rationing on the criminal side: LSPs ought to ration their services such that those who face longer sentences get priority over those who face shorter sentences. The number of years of incarceration avoided by offering the legal service can be understood as high-quality years gained. Just as giving someone a kidney allows her to avoid several years “chained” to a dialysis machine in metaphorical detention, giving someone legal assistance frees her from several years of literal detention.

PDS currently only incorporates this rationing principle indirectly by rationing more experienced attorneys to more serious crimes that carry longer sentences. It does not ration more attorney time to the more serious offenses per se, but it does take into account the complexity of the case (which has some correlation with the seriousness of the offense) as well as the experience of the lawyer involved and time to trial of the case when it sets caseloads for a practice level and an attorney. Further, its offense bands in terms of attorney experience-level allocations are extremely wide, encompassing offenses with quite heterogeneous sentences. The one place where PDS has more opportunity to select specific cases is in the context of conflicts (between co-defendants and between clients who develop adverse interests as their cases proceed). There, I am told that when it can PDS chooses the defendant “with the more difficult and more serious, for example, cases involving complex legal issues, cases involving experts, or cases with clients who suffer from a mental illness” (Leighton Interview 2011).

To be sure, estimating something even as objective and quantifiable as the number of years of incarceration faced by an accused may create pragmatic complications. While in theory one could use the statutory maximum for the offense charged as the measure, doing so is inexact in that in our current criminal law tends to overcriminalize and provide very large sentences on the books for a broad set of offenses (Stuntz 2004, 2558; Larsen 2011, 1599–1601). Rationing based on the statutory maximum might lead to similar rationing of resources to an accused charged for selling a dime bag of cocaine and manslaughter, but an experienced PDS lawyer will know that the expected sentence for the two offenses will be quite different. There is thus a strong argument for using an expected sentence calculus in rationing rather than the statutory maximum or even the Federal Sentencing Guidelines. This, however, introduces not only the possibility of bias or inaccurate estimation, but in some cases may require more upfront investment by PDS attorneys in developing the background facts of the case to determine what sentence to expect, or what plea is likely to be offered. That investment will itself come from attorney time that would otherwise be available for helping clients. Thus, in practice, LSPs like PDS would have to make trade-offs between investment and accuracy, but these are of a piece with many other trade-offs they currently make. Moreover, there are hard practical limits in how much information can be gathered by PDS in that most defendants are presented to the court within 24 hours of arrest and have counsel appointed for that initial appearance (Leighton Interview 2011). For present purposes I am more interested in the goal than the exact mechanism for realizing it.

Just as on the medical side, in theory this calculus ought also to be inflected by expected value calculations of the chance of success. For example, those with strong chances of avoiding shorter sentences might receive priority over those with poor chances of avoiding longer sentences.33 This is largely at variance with what PDS currently does, since its lawyers are trained “fight every case like it is a serious case because it is a serious case to the client”, not to sort cases that are likely winners versus losers in terms of the investment of their energies (Leighton Interview 2011).

Is such a change desirable? Although far from an exact science, in the medical context there are some objective indicia (age, comorbidities, etc.) that can be used to predict prognosis. On the legal side, there is much more subjectivity in attempting to make such assessments and therefore more chance for implicit biases or mere uninformed hunches (Mosteller 2010a, 959–973). There are also costs to implementing more thorough early evaluations. Thus, we would need to know whether the relevant allocating decision-makers are (or could become) good enough at sorting so that the improvements in outcomes from that sorting outweigh the potential errors and costs. We lack good data on these questions, and it will vary vastly between LSPs. Therefore, I do not think there is a single right answer as to whether LSPs ought to build chance of success into their best outcomes measures in the way it is standard for medical allocation systems to do so.

3.5.2.2. Harder cases

We face harder questions in taking quality of life adjustments further.

Even when length of incarceration remains constant, not all forms of incarceration are the same. Life in a maximum-security facility is quite different from a minimum-security one in terms of the limitations on one’s recreational, social, and health prospects. A robust counting of quality of life would lead PDS to favor those whose sentences will be served in more restrictive facilities, because if those individuals are convicted and sentenced they will face larger diminutions in the quality of each life year spent in the facility as against the baseline of a year of freedom. That is, illustratively, if a year of freedom is measured as 1.0 QALeY, a year in minimum-security prison might be rated 0.60, while a year in a maximum-security prison might be rated at 0.20, should that be part of the Best Outcomes analysis?34

Still harder is whether to take into account differences in the quality of life of those to whom we would allocate the legal assistance that will persist even if they are “treated”. A bioethics example is helpful in seeing this problem: Tom and Jerry are nearly identically situated fifty-year-olds each in need of a heart transplant without which each will die in the next year, but Tom uses dialysis due to kidney problems while Jerry is otherwise in perfect health. Assume that their chance of a successful outcome from the heart transplant is identical, but the heart transplant will, of course, do nothing to cure Tom’s need for dialysis, which will persist after the transplant. Thus, every year of life the transplant gives Jerry will result in 1.0 QALYs, while Tom will get only 0.57 QALYs for each of those years saved. A best outcomes rationing principle focused on maximizing quality of life years ought to favor Jerry significantly, because for each of the life years it saves he gains 0.33 more QALYs than does Tom.

We can construct a similar problem for LSPs. Imagine the prior example except instead of needing a transplant both Tom and Jerry face a death-eligible offense. Or to use a civil example, suppose both Chip and Dale approach HLAB seeking assistance to avoid eviction from their respective apartments in the same building. Both are at 130 percent of the FPL, and both are otherwise identically situated, except that Chip is also facing deportation to Paraguay—where deportation is a service HLAB does not assist with. Suppose this means that even if his housing problem is resolved successfully by HLAB, for the foreseeable future his quality of life will be considerably worse than that of Dale, and that the intervention will “buy” him less of an improvement for that reason.35 The principle we have been discussing would seem to favor allocation to Dale over Chip.

Should one endorse such a principle? Some bioethicists argue that any use of quality of life measures is discriminatory because, for example, those with disabilities will get lower priority since each year of life they gain is valued at less than 1.0 QALY. Advocates of the principle counter that it treats all QALYs as equal, to which their opponents retort that what matters is treating all people equally not their QALYs equally (Persad et al. 2009, 427–428).

Ignoring quality of life altogether seems to lead to some counterintuitive conclusions. Suppose we had to decide to allocate a life-saving treatment between two individuals. Each will gain three life years, but for one they will be healthy years with “active engagement with his loved ones and his projects” and for the other it will be three years with dementia such that he would often be unaware of his surroundings and unresponsive (adapted from Kerstein & Bognar 2010, 39). Some have argued that it would be wrong to hold a lottery in this instance, and we should disfavor the patient with dementia (id.). If that seems right there is reason to favor a best outcomes principle that robustly takes into account quality of life. More abstractly, people pursue medical treatment as well as legal services to improve their well-being. If our goal is to promote well-being, it would be strange to ignore quality of life since it is an important determination of it.

While the matter is close and I do not purport to completely resolve it here, there seems to be a strong argument for building in quality of life judgments in allocating legal services based on outcomes, as with medical services. Doing so, however, adds significant complications both in terms of measurement and ethics. One central problem involves allocation to the disabled and determinations about whether to use healthy individuals’ estimates of how bad a year with a disability would be, or the estimates of those who have the disability and have “adapted” to it (Menzel et al. 2002). I do not think there is a clear “right answer” on this issue. Fortunately, we can take solace in the fact that this problem is less likely to come up in legal allocation compared with medical allocation contexts.36

While I generally support factoring in quality of life judgments into best outcomes analysis, I share with others some concerns about cases where small differences in quality of life are used to make allocation decisions that will have huge impact on those who do not receive the medical or legal service. Suppose two almost identical individuals, Jacob and Esau, are in need of a heart transplant, and we have only one heart available to save them (this example is adapted from Kamm 1993, 144–163, 1994). Jacob has TMJ, a persistent jaw condition that sometimes causes him headaches. As a result, each year that Jacob gets from the heart transplant will be worth only 0.97 QALYs while Esau’s will be valued at 1.0 QALY. Should TMJ dictate who gets a heart transplant? What if Jacob and Esau were both facing the death penalty, or both facing eviction?37 Kamm argues that the answer should be no, supported by the intuition that it seems unfair that small differences in quality of life should be the basis for decisions with huge effects on the claimant’s life, such as life and death.38

Of course, any principle using terms like “small” and “huge” is quite vague and presents hard line-drawing. However, in practice I think this is less of a concern in the allocation of legal services than medical services for two reasons. First, while theoretically we would want to holistically measure all expected quality of life posttreatment in determining allocation, in practice legal services providers have neither the time, resources, expertise, nor inclination to do so, such that they are likely only to pick up large variations between claimants. Second, unlike organs or vaccine doses, the assistance of a lawyer is to some extent a divisible resource in the form of attorney hours and other expenditures. LSPs can therefore ration attorney resources more proportionally to the difference in quality of life, rather than the all-or-nothing allocation that doctors often face.

That said, in most contexts LSP assistance is unlikely to bear a linear relationship to chance of success. Investments are likely to clump, such that there may be no meaningful difference for chance of success between investing one hour or ten in preparing a criminal defense, eleven or thirty, etc. Research that rigorously measures the effectiveness of different approaches to providing legal services is in its infancy (Greiner & Pattanayak 2012), so one should not expect answers to these questions any time soon. While for simplicity I will often describe the relationship between lawyer investment and output as though it is a continuous variable, the reality is more complex and the roughness of the approximation will differ by practice setting.

In sum, I argue that LSPs should build quality of life posttreatment measurements into their rationing systems, but do so in such a way that small differences on the margins do not have huge consequences for competing claimants by having the amount of resources allocated be proportionate to the difference in expected outcome measures as compared with other claimants.

3.5.3. Aggregation

Thus far I have repeatedly used pairwise comparisons between two individuals to illustrate the “simple” rationing principles, largely for expository reasons. In the real world, LSPs rarely face cases that are so neat and instead must decide whether to provide some benefit to many individuals or concentrate that benefit in a single person or a smaller number. This is known as the aggregation problem and comes in easier and more difficult variants.

Let me begin with an easy variant, which involves saving fewer or more individuals facing the same fate. Suppose (to use a stylized example that) a criminal defense LSP like PDS has to decide which of three similarly situated possible clients to help, Alvin, Simon, or Theodore, who all face forty years in prison. Because of resource constraints, it cannot take on all three as clients. Because of forensic evidence collected at the crime scene, winning Alvin’s case will be significantly more resource-intensive than the other two—PDS will have to hire and prepare witnesses to discredit the chain of custody over the evidence and explain the limits of DNA evidence. As a result, if PDS chooses to take on Alvin’s case it cannot take on either Simon or Theodore’s. If it takes on Simon and Theodore’s cases it cannot take on Alvin’s. Suppose hypothetically (I do not mean to say anything about the quality of or availability of CJA attorneys in the District of Columbia) PDS also has no reason to believe that if it fails to help one of them another LSP will represent them nearly him as well as PDS could. What should it do, save one (Alvin) or two (Simon and Theodore)?

We can construct a similar example on the civil side. Suppose Sookie, Tara, and Erik have each been fired and seek an LSP’s assistance to sue for employment discrimination under Title VII. Due to the differing nature of their employment—Sookie and Tara work at a bar, Erik at a “Gentleman’s Club” (the genteel name for a strip club)—only Erik’s employer will raise a Bona Fide Occupational Qualification (BFOQ) defense to the claim that the termination was unlawful (42 U.S.C. § 2000e-2(e); see Dothard v. Rawlinson, 433 U.S. 321, 334 (1977)). Litigating the BFOQ defense will require significantly more resources to represent Erik, such that their choice is either to represent both Sookie and Tara or only Erik. Or suppose HLAB can help a married couple with three children or one without children to avoid eviction, should it favor the first?39

Many people’s pretheoretical intuition will lead them to favor the many over the few in such cases, and we can push them further in this direction by increasing the numeric disparity (e.g., aid 100,000 versus 2).

There are also cases that are harder at least in terms of our intuitions. Suppose PDS is presented with 100 individuals that seek help, each of whom faces two years of prison time. Because the offenses charged tend to be simpler for PDS to handle—the prosecutors who are given such offenses are less experienced and prepared, more willing to plead, or other reasons—PDS can prepare defenses for these individuals at relatively low cost in terms of attorney time and resources. In contrast, defending individuals on murder charges requires many more hours of PDS’ best attorneys’ time, such that with the same resource expenditure it could save two individuals from sentences that it estimates cap out at forty years. Assume for the sake of simplicity (unrealistically) that each population would serve their sentence in the same facility and that all are otherwise identically situated. To which of the two potential sets of claimants should PDS allocate its resources? If we did pairwise comparison of each of the two facing murder to each of those facing the low-level offense, those facing murder would dominate in terms of their claim to assistance. However, if we aggregated the quality of life years saved by each of the low-level claimants then their claim seems to dominate (200 years versus 80). Should PDS aggregate the life years saved or favor those facing major prison time, or engage in a lottery? We can construct a similar case on the civil side. Suppose CLS or HLAB can represent either 100 claimants in a housing eviction matter each of whom will be able to avoid eviction and stay in their house for six months, or four individuals facing eviction that it can help become owners of the homes establish long-term tenancies?40

The aggregation problem has been a persistent preoccupation of normative and applied ethics, especially since John Taurek’s 1977 paper “Should the Numbers Count” (Taurek 1977) and has generated a vast body of philosophical literature.

On one end of the continuum, consequentialists (of which Utilitarians are the most familiar variant) strongly favor aggregation in rationing. What matters is welfare (sometimes called “well-being”) and (to use a phrase beloved by the critics of this moral philosophy) individuals themselves are thought of largely as containers for this well-being. Improving the lives of a large number of people a little bit can be better than improving one person’s life a lot. Accepting that principle does not inexorably commit one to the proposition that random individuals can be murdered to distribute their organs to six other claimants, to use one classical example used to criticize Utilitarianism. One can accept aggregation for the purpose of rationing scarce goods while endorsing side-constraints on doing injustice or unwarranted harm to others for deontological reasons, or because those side-constraints themselves are necessary to maximize well-being over the long run.

On the other extreme are more deontologically oriented theories suggesting that aggregation problematically treats individuals as mere means in contravention of the Kantian Categorical Imperative (Kant 1785, 434). By asking a claimant to sacrifice for the sake of the aggregated benefit to others, it is argued that we treat him as a mere means to the improvement of their welfare and that aggregation thus fails to take seriously the separateness of persons (Rawls 1971, § 5, 23–24; Nagel 1979, 115; Luban 1988, 308 (discussing this objection as to legal services)).

A third approach to aggregation that falls in between the consequentialist and deontological approaches has been defended by Tim Scanlon from the contractualist school of normative ethics. For Scanlon, in determining whether a particular principle should govern ethical decision-making, we ought to ask whether the principle could be reasonably rejected by a party affected by it (Scanlon 1998, 238). Scanlon argues that “contractualism supports a principle according to which, in situations in which aid is required and in which one must choose between aiding a larger number of people all of whom face harms of comparable moral importance, one must aid the larger number” (id.).

Scanlon reaches this conclusion by imagining the situation from the perspective of the many (let us say two) who would be saved and the few (let us say 1one who will not be saved. If the presence “of the additional person, however, makes no difference to what the agent is required to do or to how she is required to go about deciding what to do” that should be unacceptable to a person who is part of the larger group “since his life should be given the same significance as anyone else’s in this situation” (id.). That is, the interests of one of the people on each side and the valuation of their lives are treated equally, and the presence of the second person merely serves a tie-breaking role (id.). Indeed, Scanlon goes so far as to reject use of a weighted lottery that would give the smaller group some chance of winning, because “[if] there is a strong reason, other things being equal, to save this additional person, then deciding on this ground to save the two-person group is not unfair to the person who is not saved, since the importance of saving him or her has been fully taken into account … [t]here is no reason, at this point, to reshuffle the moral deck by holding a weighted lottery, or an unweighted one” (id. at 234).

While Scanlon endorses aggregation in our easy cases, he departs from the consequentialists in arguing that “contractualism does not require or even permit one to save a larger number of people from minor harms rather than a smaller number of people who face much more serious injuries” (id. at 238). He reaches this conclusion by way of a hypothetical wherein during a World Cup match Jones suffers an accident in the transmitter room of a television station (id. at 235). There is one hour left in the World Cup game, watched by millions, but we cannot save Jones without shutting down the power and thus interrupting the broadcast for fifteen minutes. Jones’ condition will get no worse if we wait, but he is being subject to very painful electric shocks in the meantime. Should we wait to rescue him? No, says Scanlon, “[i]t seems to me we should not wait, no matter how many viewers there are” (id.). He does not think the principle discussed above, where the additional member acts as a tie breaker, works in this case because “when the harms in question are unequal, we cannot create such a tie simply by imagining some of the people in the larger group to be absent” (id.). All that said, even in cases like this Scanlon wants to leave some leeway based on how far apart the categories of harm suffered are in terms of their seriousness, suggesting that preventing paralysis or blindness for many may outweigh preventing a single death (id. at 239).

Frances Kamm offers a similar argument: Suppose two almost-identical individuals Huey and Duey are mortally ill and we have only enough serum to save one, but because of tiny differences in how much serum they need if we save Huey there will be enough serum left over to also cure a third person, Louis’, sore throat, but if we save Duey there will not be (adapted from Kamm 1993, 144–163). Kamm argues that most of us would believe it would be unjust in this circumstance to allocate the serum to Huey rather than Duey on this basis as opposed to holding a straight lottery between the two. If the sore throat is not enough to justify giving Huey preference over Duey when everything is equal, says Kamm, it is an “irrelevant utility” such that even if by saving Huey we could save not only Louis’ sore throat but a million such sore throats, it would not matter; the utility bonus is irrelevant and therefore even aggregated in large quantities cannot count (id.). Quite different, she claims, would be a case where he serum enables us to save Louis’ leg, which would be a relevant utility (id.).

Scanlon and Kamm’s examples have some initial plausibility, but upon reflection also have counterintuitive implications. To use an example suggested by John Broome, the National Health Service (the UK’s universal health care system) gives out millions of analgesics for headaches; at some level, due to health care rationing and fixed budgets, that means that someone’s life will not be saved (Broome 2002, 727–728). If Kamm and Scanlon are right, it is immoral for the health system to try to cure headaches. And yet, upon reflection, that does not seem right. I think Broome has the better of the argument when he suggests that Kamm’s thought experiment proves that curing a sore throat does not matter more than fairness in this context, not that curing a sore throat is irrelevant or would never count for anything (id.).41

That said, I recognize that this question may hinge on resolving large moral theory debates about the morality of consequentialism, something I can obviously not do in this article (or likely even a lifetime of articles). Luckily, very few of the cases for which individuals seek LSP assistance are likely to involve “irrelevant utilities”, so we are unlikely to have to resolve these hardest cases for aggregation in rationing legal services.42

Where does this leave us? For the easy cases, involving saving more or fewer people from the same or comparatively grave fates through allocating legal services, there is a very strong argument for aggregation in the overlapping consensus between consequentialist and contractualist theories. For cases where the losses are significant (eviction, loss of employments benefits, jail time) there is a also a (somewhat weaker) overlapping consensus. I would be inclined to allow aggregation even for “irrelevant utilities”, the hardest cases, but this is an area that comes up infrequently, where disagreements seem more reasonable and LSPs could permissibly diverge.

To make explicit what I have been noting throughout, Best Outcomes rationing will often conflict with some of the other rationing principles, as is indeed true of all the rationing principles. The clash between Best Outcomes and Priority to the Worst Off is particularly common in the day-to-day decisions made by LSPs. For example, as we discussed above with CLS, while its clients without housing subsidies are often worse off when threatened with evictions given the consequences to them if the eviction takes place, it nonetheless largely restricts its representation to those with subsidies, because it is more confident that if its representation succeeds those individuals will be able to stay in their homes; the unsubsidized are less likely to be able to pay the rent on a recurring basis, even if CLS can win them some time to remain in the residence, or an opportunity for a more orderly move.

3.6. Instrumental

A final family of principles ties allocation preference to responsibility and benefits to others. I will discuss two variants, each of which has analogies on the medical side: the first is to allocate based on the client’s responsibility for having the legal need, the second is based on the client’s contribution to others. I will break out for particular emphasis a variant specific to legal services relating to direct service versus impact litigation strategies.

3.6.1. Responsibility

On the medical side, responsibility-sensitive rationing would suggest that we give priority to those who have been responsible in their health choices, so that perhaps an alcoholic should get less priority for the liver transplant not because of future expectations of his sobriety but because he “drank away” his first liver (Wikler 2002, 47–48). This principle is instrumental in the sense that it seeks to incentivize desired behaviors and deter irresponsible behavior, although the principle also appeals to fairness.

Responsibility-sensitive rationing on the legal side would add significant additional complications.43 Let us begin with the criminal context. If we had beliefs about the guilt or innocence of the accused, one might be tempted to allocate more to the innocent on the view that the guilty are more responsible for their need. I explain in the next section in greater depth why I think this is undesirable.

A more modest proposal would be to base allocation on how much use individuals had already made of PDS’ services, giving him more priority the first time an accused sought these services than the fiftieth. PDS’ services would then become almost like a voucher system. This would introduce record-keeping and measurement burdens. It would also require coordination between criminal LSPs (if more than one might take the case), lest the accused simply switch providers. Moreover, it would rely on the accused to prudentially manage when to call on PDS services during his lifetime, raising questions of forecasting and the foreseeability of needs. Further, we may be worried that some “frequent fliers” in the system do not bear responsibility for the fact that they are arrested more often—for example if African-Americans are more likely to be arrested for “driving while black” (Harris 1999, 277–288).

Responsibility-sensitive rationing could also be applied on the civil side. If there were whole categories of disputes where the individuals seeking services are thought to be more responsible, civil LSPs could refuse to provide those services or reduce their resource investments as to these disputes. In theory, LSPs could also attempt to measure responsibility claimant-by-claimant. For example, in a mortgage foreclosure case, HLAB could attempt to determine if a particular claimant was irresponsible with their nonmortgage spending or by investing in multiple houses. It could also disfavor tenants whose eviction is based on criminal activity or other misbehavior, or favor tenants whose eviction is based on “no-fault” grounds, such as a landlord who wants to convert a property to condominiums (Grossman Interview 2011).

I have serious concerns about such an approach. There is a huge epistemological challenge in determining which areas of law or subclasses of cases involve individuals more responsible for their legal needs. Such determinations are likely to activate the biases of rationers. Further, if in order to construct an administrable system we must rely on categorical sorting there will be significant over- and underinclusivity in that some individuals within those categories will be less responsible for their needs, even if the average responsibility for individuals in the category is high.

Beyond these pragmatic concerns, there are also more theoretical reasons to be suspicious of responsibility-sensitive rationing. Very roughly speaking, “luck egalitarianism” is the idea that individuals should not be held responsible for “brute luck” things they could not help (such as genetic traits), but should be held responsible for “option luck” choices they do make (Anderson 1999; Eyal 2007; Markovits 2008, 272–273). For example, your genetic propensity to develop lung cancer is “brute luck”, while your increased chance of developing lung cancer due to smoking would be “option luck”. Shlomi Segall, one of the main defenders of luck egalitarianism as to health, defines brute luck more precisely as “the outcome of actions (including omissions) that it would have been unreasonable to expect the agent to avoid (or not to avoid, in the case of omissions)” (Segall 2010). The reference to reasonableness is meant to cabin the principle away from responsibility for avoidable risks that we think it unfair to demand of the risk-taker, for example the risks of living in California due to its location on a fault line or the risks of pregnancy for women (id. at 21). In theory, LSPs could also try to allocate legal resources in a way that was sensitive to whether the claimant’s need was a function of that person’s brute or option luck.

Even cabined in the way Segall has in mind, however, the luck egalitarian approach runs into a large number of difficulties. First, very few decisions that create the need for legal services are informed, voluntary, uncoerced, and deliberated upon in such a way that we feel as though a judgment of responsibility is fair, especially because lifestyles and habits giving rise to these needs can become ingrained at a very young age (Wikler 2002, 50). The relationship between socioeconomic status and the need for legal services (as with health services) is particularly fraught in this regard, and it is likely unavailing to search for an “original sin” for which the individual is responsible (id. at 51). Moreover, even if in theory we could make these distinctions, it seems hard to imagine that LSPs would have the resources, expertise, or freedom from bias to adequately perform this role.

Second, requiring LSPs to screen for responsibility might poison their relationship with potential clients and/or lead clients to hide crucial details of the circumstances in order to avoid being screened out.

Third, in responsibility-sensitive rationing “in order to lay claim to some important benefit, people are forced to obey other people’s judgments of what uses they should have made of their opportunities” requiring “intrusive moralizing judgments of individual’s choices” that “interferes with citizens’ privacy and liberty” (Anderson 1999, 310).

Fourth, there is the “harshness” objection: even if responsibility differentials should count somewhat, it is unfair for access to a crucial service to turn on small “sins”. One common example involves whether a reckless driver deserves hospital assistance after crashing into a tree (Anderson 1999, 300–302; Wikler 2002, 52).44 One could construct legal equivalents where a small amount of moral responsibility results in denial of legal services and thus the death penalty, life imprisonment, or the loss of a house.

Segall has argued that using a weighted lottery might protect luck egalitarians against the harshness objection, in that even if we believe that “the reckless driver is more at fault compared to her innocent passenger”, this does not imply that she deserves “an automatic death sentence for what could be a rather small amount of imprudence (not stopping at a stop sign, say)” so we should instead “toss an imaginary coin between the two patients, one that is slightly weighted in favor of the innocent passenger” that would result in a “responsibility-sensitive account that is not unduly harsh” (Segall 2010, 72). Nir Eyal (unpublished) has countered that this solution cannot be justified for the luck egalitarian, because “ethicists use weighted lotteries where there exists reason to prioritize some over others”. In this hypothetical, “not only are there reasons to treat [the] passenger better than we treat the driver as an incentive against reckless driving”, but according to Segall’s own theory “the driver lacks a just claim to be rescued, whereas the passenger has one” (id.).

In fact, because legal services are at least partially divisible, the luck egalitarians have a better response to the harshness objection here: rather than a weighted lottery for the full good, assistance vel non, we can to some extent have a responsibility-based reduction in the amount of resources allocated to the particular claimant. Still, all in all, the pragmatic and theoretical arguments against rationing in responsibility-sensitive ways seem strong enough to reject this as an element of a just rationing system for legal services.

3.6.2. Contribution, Indirect Benefits, and Indirect Harms, with a Focus on Direct Services versus Impact Litigation

The second variant is contribution based. One way of conceptualizing this variant is that it dictates giving priority to those who will grow the pie. In the health care context, for example, the Department of Health and Human Services Pandemic Flu Plan recommends that priority for vaccination be given to “[v]accine and antiviral manufacturers and others essential to manufacturing and critical support” because of the “[n]eed to assure maximum production of vaccine and antiviral drugs”, as well as “[m]edical workers and public health workers who are involved in direct patient contact, and other support services essential for direct patient care, and vaccinators” (HHS 2005; Persad et al. 2009, 426). The reason given is that “[h]ealthcare workers are required for quality medical care” (HHS 2005, D-13). The plan also gives priority (albeit less) to “[p]ublic health emergency response workers critical to pandemic response” and “[k]ey government leaders” (id.).

LSPs could adopt similar rationing techniques. In this discussion I will broaden my skeletal conception of an LSP’s goal slightly, in that I am examining whether the LSP ought to consider benefits and harms to third parties that are not its clients in its rationing decisions. At the crudest level, LSPs could record the occupation of the applicant and favor teachers and doctors, for example. More plausibly they could look to dependencies. Suppose that a widowed mother of four children and a single woman come to CLS seeking assistance with wrongful termination, or in defending a drunk driving charge that might result in a multi-year prison term. If the mother goes unaided, are the potentially significant negative effects on her dependents a good reason for prioritizing her claim?

This characterization also captures a different argument for age-weighting offered by the WHO mentioned above: that the younger are more productive and therefore giving them priority ultimately benefits others (Williams 1997, 127; Bognar 2008, 174–175). For legal services, one might analogously claim that those who avoid potential incarceration early in life are more likely to contribute more, especially given evidence that incarceration has very detrimental effects on postrelease life prospects, including employment. A similar argument also seems plausible for some civil legal services such as housing, and disability accommodation.

A final example of indirect-benefit rationing without a bioethics analog pertains to the distinction (or more accurately continuum) between direct and impact litigation work. Suppose two clients approach HLAB or CLS seeking assistance with an eviction proceeding. They are identically situated, except that one’s case has a legal posture such that a win in this case will allow HLAB or CLS to successfully prevail in 100 other cases. Or suppose two identically situated potential clients ask PDS to represent them on a drug possession case, except that one’s case raises the important legal issue of whether GPS tracking constitutes a Fourth Amendment search, an issue for which a win that will generate benefits for other PDS clients, while the other case represents merely garden-variety legal and factual issues? (Cf. United States v. Jones, 2012 BL 14420 (U.S. January 23, 2012) (deciding the issue)). Recall from our earlier discussion that PDS does very little rationing based on whether impact litigation benefits are available, whereas HLAB puts it front and center at the appeal selection stage (and to some extent at every intake decision), and CLS builds in this consideration very explicitly with a separate mechanism for authorizing class action type litigation (Grossman Interview 2011; Leighton Interview 2011; Eppler-Epstein Interview 2012).45

Deterrence might be thought of as a kind of indirect benefit as well, in which case, to the (fairly poor) extent it could be predicted, LSPs would want to consider not only the way in which their case selection would help remedy other existing actions harming their client population, but also the harms from clients that never come about because of the success of their litigation behaviors or the threat that they will represent a party aggrieved in the future. This is connected to the identified versus statistical lives issue discussed above—those who favor identified lives will care less about deterrence—and has an analogy in the medical literature in HIV resource allocation to the decision of treatment versus prevention (E.g., Faust and Menzel 2012)

To try to understand whether/when LSPs should count indirect benefits, I find the bioethics literature illuminating. Here are adaptations of some hypotheticals from that discourse that can help us.

The Surgeon Case: We must choose whether to allocate an avian flu vaccine to a retired science teacher, Leslie Arzt, or a practicing surgeon, Dr Jack Shepard, with the knowledge that if we save the surgeon he will go on to save five additional lives, while the teacher will not (example adapted from Brock 2003, 2). In the pandemic flu vaccine policy discussed above, providing vaccination priority to health care or disaster-relief workers, mirrors this case.

The Flo-Jo Case: In an isolated island we have a vaccine against a virulent avian flu strain and two individuals who need it to save their lives: Florence Griffith Joyner, called “Flo-Jo” (the deceased Olympic gold medalist runner) and Newman. The vaccine comes in a package that can be subdivided to vaccinate six people, but the vaccine must be customized to a blood type that will be used for all six doses. Newman and Flo-Jo are of different blood types. The vaccine must be taken before contact, and within thirty minutes all individuals on the island will have contact with the virus. If we give the vaccine to Flo-Jo, she can run to a nearby village and distribute the remaining five doses to individuals who match her blood type. If we give it to Newman, who runs slowly, it will be too late by the time he reaches the village.46 In the real world, the priority for those administering vaccines in the pandemic flu case and to those who manufacture the anti-viral arguably mirrors this case.

The Industrialist Case: Charles Widmore is the industrialist genius behind Widmore Industries. His company employs three million people worldwide. Much of the company’s success depends on Widmore’s personal genius and reputation, such that his death will plunge the company into turmoil leading to lay-offs of more than two million employees, many of whom have few good job prospects. Ben Linus is a loner, with no family, friends, or business. Both need a heart transplant. Should Widmore get priority, because of his death’s negative impacts on the employees? In the pandemic flu vaccine policy, giving priority to key government leaders somewhat parallels this case.

The three cases differ in the combination of indirect benefits and the separate sphere issue we discussed above. In the Industrialist case, allocating a health good (the organ) to Widmore would lead to indirect benefit to third parties in a nonhealth sphere (employment). In the Flo-Jo case, giving the vaccine to Flo-Jo would lead to indirect benefit to third parties that is very related to the purpose (or to use Aristotle’s term, the telos (Aristotle 1984, 1094a1–1094a5)) of the good, since the five villagers who benefit do so by being given the vaccine, such that this is a within-sphere indirect benefit. The Surgeon case falls somewhere in between depending on characterization: one can argue that the sphere should be understood as “health benefit” and that giving the vaccine to the surgeon would lead to indirect benefits in that sphere. But if we understood the purpose of the vaccine as “preventing pandemic flu infection”, then the Surgeon case appears to be an across-sphere indirect benefit since while saving the surgeon will improve the health of those needing surgeries, it does not do so by preventing the flu.

We can construct analogs to each of these cases on the legal side. If a public interest lawyer working in housing himself needs housing support to avoid eviction, that is more like the Flo-Jo case. If he needs assistance with a criminal conviction that seems more like the Surgeon case. If someone accused of a serious crime claims that his priority stems from the fact that he has many dependents, or that his business will collapse if he goes to prison, then the case resembles the Industrialist case. I think that the most important version of the problem on the legal side, the choice to engage in impact litigation, most closely parallels the Flo-Jo rather than the Surgeon case,47 in that the victory can be seen as carried over to other pending cases the same way the runner brings over the vaccine, but the point is somewhat arguable.48

In which cases is rationing based on the indirect benefit justifiable? Frances Kamm argues that preference is justifiable in her version of the Flo-Jo case but not in the Surgeon case, because although in the Flo-Jo case priority depends on a personal characteristic that she has, it is a personal characteristic (speed) that is relevant to help “better distribute” the health care good in a way that shows respect for persons (Kamm 1993, 107–114). In other words, she says, there is a “general background limit on our goal: we do not do with our resources whatever will result in the as much good as possible”, but instead “we try to achieve the best outcome for which our resource was specifically designed”, that is “we limit the sphere in which an item can maximize the good” (id.). In contrast, Kamm finds the Surgeon case problematic because it distributes the good (vaccine) for something it is not meant for (promoting health through surgery) (id.). For this reason she would likely argue there is, if anything, still more unfairness in the Industrialist case.

If impact litigation is like the Flo-Jo case, then even on Kamm’s restrictive view counting indirect benefits is justifiable here. LSPs should favor clients whose cases are more likely to lead to impact litigation results, where the expected results are that there will be larger benefits from doing so when aggregated. On the criminal side more than the civil one, though, this enthusiasm for impact litigation should be somewhat mitigated by the role that institutions like PDS view themselves as having in keeping the “prosecutors in check” and “test the government’s case”. To the extent that too much of a focus on impact litigation would disserve this function, or more cynically cause prosecutors to dynamically alter their own behavior in a way that exacerbates the problem, that would provide a reason to do more direct services work than these rationing principles might otherwise recommend.49 In any event, PDS has certainly not given up entirely on impact litigation work and carries it out through its Special Litigation Division as well as the filing of amicus briefs inside and outside its jurisdiction (Leighton Interview 2011).

Should we go further and count indirect benefits more robustly? Without purporting to fully prove the matter, let me tip my hand as to why I think we should go somewhat broader but perhaps not all the way. Kamm’s argument seems to suffer from a level of generality problem. It is unclear why we do not define the good being distributed as “medical resources”, such that there is no reason to focus on the fact that lives will be saved by the drug rather than by surgery (Brock 2003, 8). Just as I was skeptical above about sphere-limitations in defining who is worst-off, I am also skeptical in giving them much weight as to indirect benefits. Arguments about sphere differentiation are either conventionalist (limited to a particular time and place and concerned with the way in which those people value the good) or essentialist (a claim about the essence of a good in a more metaphysical sense) (Cohen 2003, 693–702). While desirable, because more persuasive, I think essentialist accounts of particular sphere differentiation are quite difficult if not hopeless to sustain (id.). In this way, I take a quite different approach than my colleague Michael Sandel, who has argued, for example (in reflecting on a case before the Supreme Court regarding whether Casey Martin should have been allowed to use a golf cart in the PGA) that there is an “essential purpose” or “telos” to the game of golf that could be used to decide the case (Sandel 2009, 204–210). I find it hard to believe that appeals to the essence of the goods could persuasively distinguish the Surgeon and Flo-Jo case; when it comes to LSPs and legal services, we will find deep disagreements on the finer points of the essence or telos of the institution with no path of breadcrumbs to lead us to a clear resolution. Moreover, even Kamm (1993, 109) concedes, in a value pluralist moment, that the unfairness of giving the drug to the surgeon “could be overridden by significant utility”, which suggests that as the numbers grow we ought to be less squeamish about how indirect the benefit is.

I think the better approach to the issue is, as Brock (2003, 11) suggests in the medical context, that as a “rough generalization and all other things being equal, the higher the level a macro health care resource allocation or prioritization decision, the more defensible it is to give weight to the indirect nonhealth benefits and costs of alternative resource uses in health care”. For example, deciding the budget of one’s department of health versus deciding which of two claimants should get an organ. The reason is that these macro-decision-makers are charged explicitly with making trade-offs between health and nonhealth goods like education (id.). It seems to me that the allocation policies for LSP intake are more like the macro-allocation decisions, while decisions to re-allocate resources away from existing clients are more like micro-allocation and ought to admit less room for consideration of indirect benefits. In both instances, though, in theory LSPs should count indirect benefits and give proportionally more priority to those with indirect benefits.

All that said, for both practical and theoretical reasons, I think LSPs should not try and count indirect benefits to the fullest extent. First, any full-scale analysis of indirect benefits will be costly to undertake. Second, too much probing may lead clients to falsify information about themselves. Third, LSPs will face problems of comparability in trying to determine whether, for example, a priest contributes more to society than does a single mother. Fourth, judgments about the worth of one’s clients are likely to harm the attorney–client relationship and make it more difficult to achieve the zealous advocacy for which our legal system aims, which is especially damaging in the re-allocation context.

This is a place, though, where the separate questions of what are the rationing principles and who applies them intersect. If we had a professionalized group of intake coordinators equivalent to positions in hospital or insurance administration that managed intake—indeed perhaps they would not even be lawyers—separate from the litigating attorneys, the effects on the lawyer–client relationship might be muted, although that would depend to some extent on whether the clients were equally able to compartmentalize their reactions to the two sets of LSP personnel rather than developing reactions to the LSP as such.

There also may be more normative reasons to avoid a full consideration of indirect benefits that has to do with a conception of the institutional role of LSPs. Suppose a Miami-based LSP focused on immigration believed with complete confidence that if it challenged a particular immigrant detention policy and won the results would alter the feeling of the Latino community and make it less likely that the Republicans would win in the state, denying them control of the Senate, which would be beneficial (in the LSP’s view). In deciding whether it should take on this particular case, would it be permissible to consider the positive effects that such a change in the Senate would have on the Latino community, poor individuals, or the world at large?

While I think Kamm’s attempt to distinguish between the surgeon and runner case is too finicky and I am skeptical about going too far down the road toward essentialism, I also think we should also resist turning LSPs into a “roving commission to inquire into evils” in the world that perfectly adjust their rationing to combat them (Panama Refining Co. v. Ryan, 293 U.S. 388, 435 (1935) (Cardozo, J. Dissenting)). I think that conclusion lies at the intersection of a few different views to which I subscribe: one is legal process or institutionalist—the idea that the legal system assigns a particular role for LSPs to play and “global welfare maximizer” is not it; one is a second-order consequentialist view that sometimes, due to epistemological failings or the reactions of other parties in a system, it maximizes good consequences not to try to take all consequences into account when making a judgment within a consequentialist frame (Cass Sunstein offers a similar argument as to judging in Sunstein 2007, 175–178); and finally despite my skepticism of essentialist arguments as guides to the finer distinctions, it does seem to me that we might reach consensus based on the telos of an LSP on rougher questions such as whether to count very indirect effects. To these points I will also add a more in-depth discussion of theories of adversarialism in the next section.

For these reasons, in practice I would recommend limiting consideration of indirect benefits to the number of dependents and whether a case has impact potential, but not to go further in making assessments about indirect benefits related to individual client characteristics. More so than other parts of my argument, I think there is not a single “right” normative answer on indirect benefits and a lot will depend on one’s general moral theory preferences for more or less consequentialist systems. Consequently, this may be a place where more LSP experimentation and discretion is more warranted. Like many middle positions—here between a polar consequentialist and Aristotelian positions on rationing theory—the result is admittedly not entirely satisfying and will raise line drawing questions, but on the whole seems largely right in articulating a useful standard for rationing.

While most of the bioethics literature is focused on indirect benefits, with the allocation of legal services the possibility of indirect harm from decisions as to allocation is also very salient. To be sure, even medical goods may generate negative externalities for some other parties. For example, saving someone from dying from lung cancer with a lung transplant may increase the costs to the health care system when they later die of something else that is more expensive to care for; expanding what is covered by an insurance plan may increase premiums paid by other insured patients. A useful example is antibiotic resistance, which pits doctors as agents of the welfare of individual patients against obligations to the health of the population. The Institute of Medicine has estimated “that antibiotic-resistant infections generate costs as high as $4 to $5 billion per year in the United States alone”, and a recent study suggests that methicillin-resistant staphylococcus aureus (MRSA), a drug-resistant bacteria associated with hospital-acquired infection, kills 18,000 patients a year in the USA, more than HIV-AIDS, Parkinson’s, emphysema, or homicide (Saver 2008, 432). This case is particularly useful in our setting because it is an instance where fulfilling the patient’s health care needs to the utmost (i.e., giving the antibiotic) negatively impacts the larger population.

LSPs can also set back the welfare of others based on the client they represent in a way that obtains even if the legal merit of her claim is held constant. The fact that a particular test case is likely to generate bad precedent is a potential indirect harm. Although the additional harm is not to the client one is representing, it is still quite closely tied to the interests of the client population that LSP serves qua clients. The harms involved could be still more indirect. For example, imagine that Walter and Pinkman are both residents of a housing development facing eviction on the same grounds, which are grounds that the LSP views as unlawful. Walter, is a kind high school teacher who teaches the local children about wonderful real world applications of chemistry after school. Pinkman sells meth to neighborhood kids and intimidates his neighbors. The two are otherwise identically situated. The LSP could consider the indirect harm of Pinkman’s drug dealing and threats in deciding whether to allocate to him, and indeed CLS does exactly this by instructing Statewide not to send its housing cases where there is a credible allegation that the reason for eviction was that the tenant was selling drugs. Or suppose a client approaches a housing-oriented LSP about a rent increase in violation of the community’s rent control ordinance, but the LSP is of the view that the rent control ordinance is itself a bad idea for the local community it serves in that it reduces housing opportunities. Would it be appropriate to consider that in the rationing calculus? Note how, in both cases, successfully representing the client in question will produce results that are directly adverse not to other potential clients per se, but to the larger community from which one’s client population comes.

One could also imagine still more indirect harms: suppose an LSP is considering representing two clients challenging a drug possession charge stemming out of a drunk driving traffic stop. One case would involve challenging the constitutionality of the drunk driving traffic stop program more generally as unjustified under the Fourth Amendment’s special needs exception (cf., Illinois v. Lidster, 540 U.S. 419 (2004)), while the other involves a more narrow fact-specific objection. While both clients will benefit from the challenge to their conviction, if the challenge succeeds the LSP believes the police will face much greater difficulties in prosecuting and deterring drunk driving leading to a predictable increase in road deaths of innocent pedestrians and drivers. Should that be relevant in determining which of the two individuals to whom to allocate representation?

Although, as a theoretical matter, there need not be symmetry in the analysis of indirect benefits and harms, I do not see a reason for asymmetrical treatment in the analysis in this instance. The same epistemological, practical, and normative concerns that lead me to recommend only a narrow consideration of indirect benefits also lead me to think that LSPs should not attempt to robustly consider indirect harms when making rationing decisions.

To step back from fine-grained argumentation to a more general sense of why I think this is a bad idea for LSPs, a bioethics parallel may again prove illuminating. In the 1960s, the Seattle Artificial Kidney Center chartered a committee, comprised of “a doctor, a lawyer, a housewife, a businessman, a labor leader, a state government official, and minister” (all white) to develop rationing criteria for potential dialysis patients. The committee, nicknamed “the God Squad”, sought to measure patients’ social value to decide who should get life-saving treatments, looking at “patients’ mental acuity, family involvement, criminal history, occupation, educational background, employment record, transportation ability, and future potential”, frowning on “[d]ivorce, homosexuality, and unemployment,” while favoring wealthier “patients who held good jobs” (Goodwin 2004, 339–340; Sanders & Dukeminier Jr 1968, 377–378). At the time, Sander and Dukeminier attacked the social worth premises behind the committee deliberations suggesting it provided “numbing accounts of how close to the surface lie the prejudices and mindless clichés”, of a “middle-class suburban value system”, and wryly observed that “[t]he Pacific Northwest is no place for a Henry David Thoreau with bad kidneys” (id.). While I have argued for some room for consideration of indirect benefits relating to the benefits of impact litigation for an LSP’s client base and perhaps consideration of the number of dependents who can be thought of as clients, I think it would be wrong to go further and turn LSP personnel into a God Squad of their own.

In this section I have focused on the way in which rationing decisions faced by LSPs parallel those made in bioethics. In the next section I consider more fully considerations unique to (or at least more salient in) the legal aid context. In so doing I will return to two issues related to the indirect benefit/harm question: factual guilt of criminal claimants and “adversarial goods”.

3.7. ‘Simple’ Rationing Principles for Legal Service Providers: A Summary

I have examined six families of rationing principles, discussed the normative issues they raise, and made recommendations as to which ones should be considered by LSPs on the criminal and civil side. My conclusions from that discussion are summarized in Table 1.

While I have taken a stand on all of the principles, for many the issues are admittedly close. My goal is not to definitively resolve each, but to offer LSPs a checklist of considerations they must face in evaluating their current policies. What matters most to me is to make clear that all of these are questions LSPs cannot avoid answering. As quasi-public institutions responsible to the taxpayers and to those who charitably fund them, LSPs must be open about their decisions on these issues and provide reasons why they have chosen the rationing principles they have.50 The checklist could also be used by LSC or other funders in deciding which LSPs to fund and to what extent.

3.8. Complex Rationing Systems

I have thus far presented basic rationing principles one by one, but as on the medical side it is not plausible to adopt only one of them. Combining multiple rationing principles requires making two kinds of decisions. The first is about selectionwhich principles to incorporate into a complex system. Table 1 summarizes my own views on this matter. The second decision is about proportionhow much to afford each principle in the system? Once again, the medical context can give us some ideas about how to generate complex rationing systems. Here are sketches of three complex medical rationing systems.

3.8.1. United Network for Organ Sharing

United Network for Organ Sharing (UNOS) allocates organs in the USA via a point system. It combines a first-come-first-serve principle (which I have argued against) with a priority to the worst-off system that focuses on the health sphere, and a best outcomes measure focused on antigen, antibody, and blood type matches of the donor and recipient (Persad et al. 2009, 426). Historically, these principles are weighted differently by organ type, with more focus on first-come-first-serve for kidneys and pancreases, more focus on sickest first in hearts, and a balancing of all three for lungs and livers (id.).

3.8.2. Disability-adjusted Life-years

The WHO uses Disability-adjusted Life-years (DALYs) to make decisions on resource allocation for disease prevention internationally. To simplify significantly, the DALY system begins with QALY measures and then modifies the QALY number with an age weight giving increasing priority for those in the twenty to fifty range (id. at 428).

3.8.3. The “Complete Lives System”

In the Lancet, Persad, Wertheimer, and Emanuel have argued for a “Complete Lives System” that combines modified age-weighting (prioritizing adolescents and younger adults over young children and those over age fifty), best outcomes (restricted to health spheres) and a principle of saving the most lives—not on the quality of life years (id.). Somewhat more vaguely they also say that “[i]n a public health emergency, instrumental value could also be included to enable more people to live complete lives” and that “[l]otteries could be used when making choices between roughly equal recipients, and also potentially to ensure that no individual – irrespective of prognosis – is seen as beyond saving” (id.). They do not specify what proportion each of the principles gets in the weighting.

The UNOS and DALY systems combine their “simple” principles in a way that is mathematically precise and generates fixed priority scores, while the advocates of “Complete Lives System” have answered only the selection and not the proportion question. That said, if we looked “under the hood” of the UNOS and DALY systems and tried to determine why, for example, age-weighting was given this rather than that modifier number, I do not think we would find determinate, philosophically well-reasoned arguments.

All this suggests to me that in generating complex rationing systems for LSPs we will also be unable to use normative reasoning to perfectly reach such answers and instead will at most be able to achieve rough parameters. Therefore, unlike the selection question, I do not think the proportion question has definitive “right answers”, and I would be inclined to give LSPs more leeway in fixing their own priority weights as to each of the “simple” rationing principles I have endorsed, subject to the kind of transparent deliberation discussed above.51 Would it be better if normative analysis could take us further? Undoubtedly, but as Aristotle reminds us, “the educated person seeks exactness in each area [only] to the extent that the nature of the subject allows” (Aristotle 1984, 1094b24–1094b25).

4. DIGNITY, DESERT, AND ADDITIONAL COMPLEXITIES ABOUT THE LEGAL CONTEXT

Section 3 developed principles for rationing legal services under a skeletal account of the goal of LSPs and some simplifying assumptions about the endogeneity of the rationing system to the available funding stream and lawyer satisfaction. This section examines how incorporating other values and relaxing those assumptions might alter the picture. I begin by considering dignitary participatory process values. I then consider whether LSPs should factor in desert, in the sense of whether the “objectively right outcome” was achieved in a case (especially criminal matters), or its effect on parties on the other side of the “v.”. I then discuss what happens when we bring funding streams and lawyer satisfaction back into the picture. At the very end, I address whether LSPs should have discretion for institutional self-definition, allowing them to depart from the rationing principles I have identified.

4.1. Dignity and Participation as Due Process Values

In a justifiably famous article and book, Jerry Mashaw argued that one of the unsatisfying aspects of the Supreme Court’s constitutional procedural Due Process jurisprudence was its failure to consider values other than efficiency that are part of Due Process. I will focus, in particular, on his discussion of “dignity,” which he connects to the distinction “between losing and being treated unfairly” and the way in which we might “describe process affronts as somehow related to disrespect for our individuality, to our not being taken seriously as persons” (Mashaw 1985, 162–163). Would recognizing dignitary process values substantially alter the rationing principles derived in Section 3? I will argue largely no, but concede one area where it should make a difference.

One argument against the claim is that it confuses the ends of the justice system with the ends of an LSP. While dignitary process values might merit consideration when developing a set of procedural requirements for the courts, they are not factors that should move the LSP in determining its own rationing systems. This kind of argument has an Aristotelian flavor; that the LSP’s role or telos is to meet the needs of its clients in outcome not dignity terms.

The argument seems a bit too pat to me. The telos of an institution is always being contested, subject to change, etc. While I think it is fair to say that LSPs’ primary role is helping their clients meet their tangible needs through litigation, allowing their clients to feel heard by the system should also be considered part of the goal. In the extreme case where the individual has no chance of winning at all—when their claim is frivolous or entirely blocked by settled law—I think a LSP would satisfy this duty by merely briefly explaining to a client why his case was not selected.

What of other cases where the potential client has a chance of winning, but less than other claimants? Here one might argue that even if the dignitary values count for something, they play a “cancelling” role as between the person to whom legal services are and are not rationed. To whichever of the two it fails to allocate the service there will be a failure to vindicate his dignitary interest, so the LSP might as well give the resource to the person favored by the rationing principles.

This answer seems, on the whole, right, but it should cause us to slightly alter our understanding of aggregation. To return to an earlier example, suppose an LSP can represent either 100 claimants in a housing matter, each of whom will win two months of rent rebate, or four individuals facing eviction who will lose housing altogether if not helped. Now let us imagine, for present purposes, that when aggregated the benefits to clients are exactly the same, such that if we were allocating based only on best outcomes we would hold a lottery. By allocating it to the 100 we vindicate 100 people’s dignitary participatory interest, as opposed to four people’s, giving a reason to favor the larger numbers. Notice how this is different from impact litigation where the benefits may diffuse to large number but not the dignitary value of being represented.

To operationalize this one could add a fixed “dignitary participatory bonus” per client that gets aggregated along with the outcomes. How big should that bonus be? It depends on what weight one assigns that value in relation to outcome measures. My own view would accord it a small weight since I do not think this value primary in the telos of LSPs, and deep skeptics of dignitary values might reduce it further. There are also more subtle ways to operationalize this idea. One could modify the “bonus” by our Prioritarian measure of how badly off the individual is, if one thought that the dignity of participation mattered more to the down-trodden. One might also modify the “bonus” by the size of the outcome if success is achieved, if one thought that one’s dignity is harmed more by being denied representation in a case where more was on the line. Perhaps this would lead one to give a bigger dignitary “bonus” for criminal rather than civil cases as a rough approximation, or more precisely grow the bonus with the size of the negative outcome the individual faces if they lose. One reaches a point where, both in theory and in practice, complexity threatens to overwhelm. At this point, I think we are there. I would be inclined merely to add a small fixed dignitary bonus rather than implement the more complex proposals, even if it will get things wrong some of the time. But perhaps heartier souls will want to further refine the more complex approaches I have laid out.

4.2. Of Adversarial Goods, Just Outcomes, and LSPs’ Commitment to Justice

Should LSPs count under “best outcomes” cases where the claimant wins a case she “deserved” to lose from an “objective standpoint” (scare quotes intentional)? This is a major difference from the medical context. Gains in life years or quality of life gained, while not an unalloyed good, are usually thought of as a job well done by the medical system. In contrast, one might argue that when PDS gets someone acquitted for a crime they actually committed, or when CLS or HLAB prevents eviction of someone who actually “deserves” to be evicted, that is not an outcome that should be given positive weight in a rationing system. One can call this “just outcomes” as a gloss on “best outcomes”.52 This is an approach that PDS explicitly rejects in the way it trains its lawyers, instructing them to “fight every case like it is a serious case because it is a serious case to the client.” In contrast, Darryl Brown (2004, 818) has urged criminal LSPs to “distribute limited defense resources [] toward strategies more likely to vindicate factual innocence”.

A related objection is that while organs and vaccine doses are goods that primarily generate value for those who receive them without negatively impacting the lives of those around them, legal assistance is different in that it has externalized effects for those on the other side of the “v.” and beyond. We saw above that even as to medical goods, the matter is actually more complicated—antibiotic resistance, for example—but it is clear the issue is more pressing for LSPs: when HLAB assists a tenant to resist eviction, it also sets back the interest of the landlord who has another tenant to whom she prefers to rent. Legal services are ordinarily what I call “adversarial goods”: allocation of services to a client may set back the interests of that client’s opponent as against a counterfactual world where the LSP did not assist.

Indeed, as Galanter classically recognized, on some occasions litigation involves “One-Shotters” (OS) versus OS (e.g., parental custody fights, divorce proceedings, etc.) rather than configurations of RP against each other or against OS (Galanter 1974, 107). Thus, an LSP that sees itself as furthering the interests of OS will sometimes make things worse for other OS. Further, even if it only represents OS against RP, some of the RP may be within the communities they care about. Some landlords will be mom-and-pop types trying merely to make ends meet with relatively low incomes. Should LSPs at least consider the effects of its assisting tenants on such landlords?53

There are both pragmatic and theoretical responses to these objections, both of which augur against significantly incorporating these concerns into a rationing system. That said, I suspect that this may be one of the most contentious set of claims I make in the article, and while I will defend my ground, I also want to offer a concession to those who disagree with me and place the objection in perspective: if you think I am wrong about this, you need not reject anything I have said thus far in this article. One could accommodate either form of the objection within the rationing principles I have set out. To operationalize it within a best outcomes rationing principle, one could treat as zeros any gains made by criminal or civil clients who did not “deserve” to win, apply a discount factor, or treat such gains as having negative values. An LSP could also do something similar with the more general adversarial good idea, for example, by sifting out cases where OS are suing other OS or other parties whose interest the LSP believes deserves special attention, or apply a discount to priority in such cases. It could do any or all of those things as an “add-on” to the rationing system I have thus far developed, either as a method of screening out claims and applying the rationing system to what remains, or through discounting.

Now that we understand the possible “fixes” one might implement, should we make them? I would argue against LSPs robustly incorporating either of these “fixes” for the following reasons.

First, the objections (especially the “just outcomes” version) assume there are metaphysically “right answers” to the question of what should happen in an individual case. There are deep traditions of legal scholarship that dispute the claim and press for indeterminacy in at least some cases. Compare, for example, Frug (1984, 1290) and Singer (1984, 5) with Dworkin (1978). While I am not particularly drawn to the stronger rejections of this claim, for those who are, any talk of just outcomes is a nonstarter.

There is a more modest epistemological version of this response suggesting that even if there are right answers we may not be able to know them. Even more contextually, one might argue that LSPs are particularly ill-equipped to make these judgments at the intake stage, and that inviting them to do so is more likely to lead to bias (implicit and explicit). Further, the resources required to do even the most rudimentary screening of this sort would divert a significant portion of the LSP’s budget and dramatically reduce the number of clients they can help.54 Indeed, there is an argument that the adversary system and zealous representation is itself the procedure by which we find the right answer. As Lon Fuller beautifully put it: “the role of the lawyer as a partisan advocate appears not as a regrettable necessity … but an expression of human insight in the design of a social framework within which men’s capacity for impartial judgment can attain its fullest realization” (Fuller 1978, 383). Moreover, the kind of probing required to pre-judge cases in this way may sour the developing lawyer–client relationship and further exacerbate the dignitary affront (discussed above) to those who are denied services, although again it may be possible to blunt some of this negative effect by certain intake system designs and answers to the who rations question.

Second, as Galanter classically noted, those pressing for legal change often face off against well-heeled individuals who hold the relevant entitlements under the status quo (Galanter 1974, 125). There is likely to be significant overlap between those individuals and the ones (especially on the governmental funding side) making decisions regarding the funding and allocation rules relating to LSPs. By introducing a constraint on the rationing system that discouraged LSPs from bringing cases that would impose significant costs on their adversaries or other third parties, funders and their allied constituencies could use their control of LSP rationing to protect their status quo entitlement against legal change through the court system. One might be particularly worried about this in the case of litigation challenging government programs. These kinds of concerns prompted the Supreme Court to declare unconstitutional on First Amendment grounds a Congressional statute that would have prevented LSC from funding any LSP that sought to amend or challenge existing welfare laws (Legal Services Corp v. Velasquez, 531 U.S. 533 (2001)). If the court system is meant to be a meaningful reviewer of the status quo distribution of legal entitlements, and LSPs are meant to broaden access to the courts, such a constraint threatens to short circuit the entire process.

Third, may of these forms of rationing would impose a burden on LSPs that do not apply to their brethren and sistren in private practice. While they cannot make frivolous arguments (see, e.g., Fed. R. Civ. P. 11.) and face other professional responsibility limits on representation, they currently face no obligation to select only clients whose victories would be just or whose success would further advance the social welfare apart from the legal merit of their claims. Although some thinkers on legal ethics have urged that private attorneys be held to more onerous standards than the floor of what the law and the codes of professional responsibility (e.g., Simon 1988), at the moment attorneys providing services to private clients do not face hard constraints in that way. To the extent legal service provision is about evening the playing field in the legal system (Galanter 1974), there is an equality argument that those who cannot afford private counsel ought not to face an additional restriction not faced by their opponents. This has particular bite in the claim that LSPs should consider the costs litigation will place on the adversary. It seems unfair to suggest that the worst-off claimants in the legal system should have their access to legal services cabined because of possible ill effects on their better off privately represented opponents, when those opponents face no obligation to perform the same analysis or restrict themselves in this way.

Matters seem closer to me when it comes to litigation of the OS versus OS variety and/or when the opponent is also a member of the community of interest that the LSP represents. It seems to me that as part of its rationing system, an LSP might simply decide not to represent clients who are going after other potential clients that would qualify for its services when it feels the representation will make things worse for its client base, or at least apply a discount weight to that particular client in determining its priority rank. For example, perhaps a housing-oriented LSP might decide to represent poor tenants against landlords but not against other poor tenants in nuisance claims if it worries that the precedent it will set will in turn be used by landlords in future attempts to evict poor tenants. One complexity here is that the indirect benefits of deterrence from the availability of LSP representation might push somewhat in the opposite direction. Adversaries that are themselves represented by LSPs or otherwise fit the OS model may be easier to deter than true RP, in that the additional costs of litigating against an LSP-represented opponent versus an opponent without representation may matter less to a RP than it would to a OS opponent.55 This is one of the reasons why I find it hard to take a firm position on this “fix”, and perhaps some experimentation in this regard would be desirable.

Fourth, especially in the criminal context, such rationing might end up leaving those who are rejected worse off than if there were no representation available, since it might signal that the LSP views their case as unmeritorious or their client as guilty. Thus, a rationing system that incorporated these values might actually harm potential claimants rather than merely fail to help them. Of course, to the extent the LSPs’ own preliminary review of the cases would make it easier for other actors (such as prosecutors or judges) to gain this kind of information, that might be thought of as a benefit, especially to more efficiency-oriented analysts. The dynamics of how this would actually work out are complex and hard to predict. Would some individuals with “good” cases nonetheless worry about the chance of a negative signal and thus refuse to seek review by the LSP if they can, such that the risk of negative signaling would fall most on the worst off? Could those with “bad” cases successfully adopt a mimicking strategy? How would rules regarding confidentiality permit or forbid LSPs from disambiguating the reasons why they did not take a case? The complexities make matters hard to predict, but there does seem to be some risk of transforming LSPs into signaling adjuncts to other parties in this way, a kind of lawyer as double agent problem.

Fifth, and at a more theoretical level, are responses connected to the ideals of adversarialism and role division in the US legal system. There is a significant literature pressing the notion of adversarialism and critiquing it within more typical professional responsibility discourse, which focuses on how lawyers should modify their representation of clients when faced with dilemmas where the seemingly less ethical course is within bounds of the formal rules of professional responsibility: for example, whether to tell an adversary that a statute on contributory negligence that benefits one’s own client has been abolished when one has reason to believe the other side is unaware of it (Simon 1988). This literature can be adapted to our context, where the question is not how to represent a particular client in an ethically fraught strategic decision, but whether to represent him and with what intensity of resources.

As Bill Simon classically characterized it, there are two conventional polar views of lawyers’ role in our adversarial system. On the “libertarian approach”, the “lawyer is obliged—or at least authorized—to pursue any goal of the client through any arguably legal course of action and to assert any nonfrivolous legal claim”, an approach Simon characterizes as privileging “form over purpose by authorizing appeals to interpretations of rules that frustrate the purposes of the rules”, and “narrow ways of framing ethical issues over broad ones”. This includes “permitting nonlegal advantages and disadvantages to exercise a relatively broad influence over the resolution of legal controversies”, such that “relatively wealthy clients may invoke procedures in order to impose prohibitive expense on relatively poor ones, and publicly subsidized lawyers for poor clients may engage in tactics that impose expenses on opposing parties required to pay for their counsel” (Simon 1988, 1085). On the “regulatory approach”, in contrast, the “lawyer should facilitate informed resolution of the substantive issues by the responsible officials”, with a basic duty “to clarify the issues in ways that contribute to a decision on the merits, not to manipulate information to serve the client's goals”, recognizing that the “job still involves advising the client on ways to advance her interests and presenting the client’s case, but it also involves a duty to develop and disclose adverse information that would be important to the responsible official” (id. at 1085–1086). Against these two poles, Simon offered a third alternative: “ethical discretion”, whose maxim was that “[t]he lawyer should take those actions that, considering the relevant circumstances of the particular case, seem most likely to promote justice”, namely “a kind of noncategorical judgment that might be called pragmatist, and hoc, or dialectical” (id. at 1090). Robert Gordon also captures a similar idea with his beautiful phrase of lawyers as “the social curators of the values of legalism—promoting understanding and respect for the values of human rights, due process, equal legal treatment, universal access to justice, and coming to the aid of persons injured by the deprivation of these rights” (Gordon 1988, 24).

One might suggest that we can recognize equivalent poles (again merely ideal types, few seriously advocate the polar views) in conceptions of the duty of LSP lawyers in selecting which clients to represent and the intensity of resources to allocate. On the libertarian pole, the LSP should take potential clients as it finds them and nonjudgmentally ration services based on the kinds of outcome and needs-based criteria I set out above, seeking to facilitate (within the bounds of proper professional conduct) the interests of the client rather than to try to judge whether the client “deserves” to win at the outset. On the regulatory pole, the LSP acts more like a pre-litigation court that judges the cases before deciding whether to take them or not.

It would be tempting to think of the objection as suggesting Section 3’s rationing system errs too much on the side of the libertarian pole and away from the regulatory pole, and that it seeks merely to re-right the balance. I think it would be more accurate, though, to say that both my hypothetical interlocutor and I are actually fighting over what it would mean to adopt the equivalent of “ethical discretion” in the rationing context, and the contours of what it means for an LSP to “promote justice” in its rationing approach. I have argued LSPs would promote justice in rationing decisions by using mechanisms like age-weighting, priority to the worst off, best outcomes, etc., and the question posed by the objection is whether it would further promote justice to allow them to consider things like factual guilt or effects on the needs of the adversary as well.

On the criminal defense side, the argument that “just outcomes” would not promote justice departs not only from the constitutional command of Gideon discussed earlier, but also from “[t]he guiding premise of the entire system is that maintaining the integrity of rights-guarding procedures is more important than obtaining convictions or enforcing the substantive law against its violators” (Gordon 1988, 11–12). That is, the role of the LSP is that of a protector of interests “flowing from the command of the Sixth Amendment under our adversarial system for equal justice and the contention that in protecting those least loved, defenders support the foundation of the liberties of all” (Mosteller 2010b).

Matters are closer on the civil side, where, as others have recognized (e.g., Gordon 1988), the line between zealous advocacy and forsaking one’s role as curator of legalism is often harder to draw. There are reasonable, and I think largely intractable, disputes on this issue. In facing particular ethical dilemmas, I think Simon is quite correct that noncategorical, more situation-sensitive approaches are essential. Even Simon, though, recognizes as a justified convention in the profession an “advocate’s duty to make any effective non-frivolous argument to a court for a client’s position regardless of whether she is personally convinced of its merit”, which he views as representing only “a minimal departure from ordinary morality” and sees “grounded in the belief that, by deferring private judgment, the advocate facilitates a more reliable and accountable resolution by the judge” (Simon 2010, 994–995). Nothing I advocate for LSPs to consider doing runs afoul of that understanding. In any event, even if one would urge a much more robust and judgmental posture to possible litigation strategies of clients one represents, it does not follow that the same should be true in the decision whether, as an LSP, one should ration resources to or away from a client who will in all likelihood be unable to secure representation elsewhere. Indeed, the existence of a more regulatory approach to litigation conduct might in fact somewhat reduce the need for robustly pre-judging claims for their merits, in that a responsible advocate, in his or her litigation behavior, will both shape the client’s interests and the means by which it may be pursued in a way that enables the court to do its job. Refusing to represent the client at all would be more like a death sentence, that forsakes the client entirely rather than trying to lead him or her on the ethical path.

In combination, these arguments provide good reasons to reject “Just Outcomes” rationing and also too much of a foray into the “adversarial goods” approach, though I have suggested more openness to the idea of taking into account whether the presumed opponents will be OS or RP in Galanter’s sense. As in my discussion of indirect benefit counting in Section 5, both rule consequentialist views suggesting there may be value in keeping LPSs “pure”, and Aristotelian conceptions of the telos of LSPs are doing some of the work in my analysis. Those with quite different normative theory commitments about what will make the system run best may come out differently. In any event, if one concludes I am wrong and that a just rationing system would review considerations of this kind, I have indicated how the principles articulated in Section 3 might be modified to do so.

4.3. Funding Streams and Lawyer Satisfaction

The analysis in Section 3 explicitly makes the choice of rationing systems independent of an LSP’s funding streams and satisfaction (and thus performance) of its lawyers. Let me now relax those assumptions. First, we can imagine that the rationing principles chosen will determine the willingness of private or even public funders to support LSPs. It is difficult to predict the effects of adopting any one of the “simple” rationing principles, let alone a combination.We can tell “just-so” stories where funders are turned-off by LSPs “playing G-d” through indirect benefit counting, or care deeply about particular kinds of cases that the principles in Section 3 de-prioritize. We can also tell opposite plausible stories. We currently do not know which stories are correct, and the answer will likely vary by LSP.

Let us therefore play with both possibilities. If adopting these rationing principle increases provider funding, that is just another reason (apart from those discussed above) to adopt them. What if adopting these principles led to a decrease in overall funding levels, such that potentially fewer people were helped, but the distribution of who was helped was more just? We should divide the rationing principles between those focused on welfare-maximization as such, and those focused on distributional justice issues. For the former, the question is whether the reduction of funding leads to a diminution in welfare that is greater than the welfare benefits achieved by adopting rationing principles better keyed to welfare-maximization. If so, that is a strong argument not to adopt these particular rationing principles. Things are less clear for rationing principles defended on justice/equity bases. It may depend to some extent on who suffers from the diminutions in funding and the justice function being used. If adopting the rationing principles improved the welfare of the worst of group but diminished total welfare significantly, that would not necessarily count against adopting these principles on the Rawlsian Difference Principle approach.

Second, adopting some of the rationing principles might lead to decreased lawyer satisfaction.56 Recall that, as discussed above, in PDS the best lawyers work their way up to handling homicides. Imagine that PDS changed its policy and kept these lawyers working on less serious crimes because they were very efficient at doing so, and the aggregated benefits of assisting with a large number of such cases outweighed the benefit they could produce on a few homicides. If under this policy PDS lawyers became less satisfied with their jobs, PDS might face more difficult recruiting or higher lawyer turnover. On the other hand, some of my rationing principles might lead to increased lawyer satisfaction. For example, the ability to invest one’s time in impact litigation cases might excite some lawyers much more than direct services work. Part of the problem is that in all of these cases both things may be true. Changing PDS’ focus would have an impact on which lawyers end up staffing the institution.

Imagine we knew—as we most certainly do not—which version was likely to generate better outcomes for clients. Once again, if the net effect on lawyer satisfaction is positive, then there is an even stronger argument for adopting these rationing principles. If, however, the net effect was negative, this might lead us in at least two directions. We might focus on the output of LSPs. If the quality/quantity of the work decreases as a result of decreased lawyer satisfaction due to adopting these principles, this result is similar to the reduction in funding stream discussed above and the same kind of analysis applies.

Or we might focus on the lawyer’s perspective. Even if the output remains the same, is it not dehumanizing to view them as cogs in a machine producing good outcomes for clients? Does their happiness not count for something? My views are tentative here, but I think that from the point of view of a just rationing system the answer is that their happiness may count, but not for very much. Those who have chosen to work for a particular LSP, knowing the rules of its rationing system,57 presumably have decided that all-things-considered they feel as actualized in this life as any open alternative. I do not think prospective lawyers can make strong normative claim for a rationing re-design that was less just but more personally satisfying. If transplant surgeons would get more satisfaction from the challenge of transplanting organs into patients posing more difficult medical cases and thus reduced chances of success, that would not justify changing our organ rationing policy to reflect their desires. If anything, LSP lawyers can make a less compelling case in that they could theoretically move to the private legal sector, which is not something transplant surgeons can do since the supply of organs is centrally controlled.

4.4. The Allure of the Big Tent: How Much Room Should There be for Legal Service Providers’ Institutional Self-Definitions?

Finally, some LSPs might object to my championing of impact litigation work, suggesting that they are just not that kind of LSP and there ought to be room for all types in a big tent. Tremblay, for example, helpfully develops a four-point continuum of LSPs encompassing those focused on “Individual Case Representation”, “Focused Case Representation”, “Law Reform”, and “Mobilization Lawyering”(Tremblay 1999, 2500–2004). Similarly, one might imagine PDS or another LSP protesting “we leave impact litigation to other institutions; that is not what we are about”.

The strongest response would be that this objection mistakenly treats the question of the nature of the LSP and “what it is about” as normatively prior to the rationing question. In situations of persistent scarcity, when there is not and will not be sufficient governmental, charitable, and lawyer resources for LSPs to help everyone, allocation of resources to specific LSPs and issues regarding LSP missions are themselves rationing questions on which institutions must answer to the demands of justice. The point is not that a particular LSP with a particular conception of its mission has no right to exist. It most certainly does. What it has no right to do is to take resources from a common pool set aside for helping ensure access to justice and use those resources in such a way that do not conform to the principles of just rationing I have set out. Again, the medical analogy can be helpful. Suppose the WHO was faced with the choice of allocating its fixed budget of health care funding to fund health care workers specializing in direct testing of HIV in sub-Saharan Africa or “impact” oriented interventions encouraging circumcision in the same region. Suppose it was crystal clear that its allocation to the circumcision group would prevent many more HIV cases and the benefits would be distributed at least as fairly and equitably as the alternative. There would be a very strong argument for allocating to the circumcision group. The fact that the defunded group might say “our vision is direct testing, not wide scale prevention efforts”, would not justify WHO making a different allocation. Why should a different conclusion obtain as to LSPs?

There is much to be said for this strong response. However, although I will not be able to fully discuss these issues in this space, I want to tentatively suggest why on my view there ought to be some room for institutional self-definition. HLAB provides an excellent example (in contrast to, say, CLS): While doing the best for its clients is part of its institutional mission it also serves an educational purpose in teaching law students to litigate and a value inculcation purpose in helping them internalize the ethical norms of our profession. Sometimes these values will be put in conflict at the intake stage, such that there a case would arise where priority is warranted on the rationing principles I have set out, and yet the case is poorly suited for educating its students. What should an institution like HLAB do in such cases?

If we made it slavishly follow the rationing rules the result would be a homogenizing of LSPs, where all would operate on a single institutional definition. Something would be lost: experimentation in institutional form and methods, the ability to accommodate divergent political or religious views,58 the insulation from funding cuts by the government made possible through a broader base of claimants, the benefits of community formation and self-actualization for its lawyers, and other benefits. As the same time, these considerations seem insufficient to justify a full departure from just rationing, especially to the extent that LSPs are taking public and not simply charitable funding. Here are two very tentative suggestions for accommodating this clash. First, all the rationing principles I endorsed could be applied after a threshold “mission screen”. For example, HLAB could screen out any cases that did not meet its educational mission, but then allocate resources among remaining cases based on the rationing principles discussed. Second, we could add “mission suitability” as an additional variable in a complex rationing system that modified one or all of the other variables discussed above. Choosing between the methods would be a very context-dependent determination and depend on a more robust normative analysis of the nature of LSPs and their relationship to society. I do not purport to do that work here, but merely suggest that making room for institutional self-definition is not antithetical to the rationing project that has been my focus.

5. CONCLUSION: SOME REALISM ABOUT RATIONING

In this article I have tried to comprehensively analyze how criminal and civil LSPs should allocate legal services given the reality of persistent scarcity, drawing illumination from bioethics. Admittedly, the normative analysis and the principles they have engendered are in some instances complex. Given this complexity, is developing a just rationing system for LSPs realistic? After all, LSP attorneys are overworked and underpaid, with funders wanting to see investment in results, not intake procedures. Let me offer two responses to this concern.

First, health care allocators like UNOS, WHO, and the NHS have faced similar complexities and challenges as well as pressure from their funders, and I believe lawyers are, if anything, more attuned to thinking about justice and its many forms than our medical brethren and sistren. It may be that at the end of the day these principles are better applied to institutions like LSC making funding decisions, rather than individual LSPs. Second, while I have sketched the outlines of an ideal approach to rationing legal services, we could make significantly headway toward a more just allocation system without adopting every recommendation. In particular, here are eight tangible recommendations that could be adopted without too much administrative burden or cost:

  • Where possible, do not use first-come-first-serve rationing.

  • Do not favor identified lives, or those with urgent needs.

  • For criminal cases, do not favor those facing the death penalty apart from the number of years of life saved.

  • For criminal cases, give priority to clients based on the number of years of incarceration they face if convicted.

  • For cases where allocating legal services will have a significant effect on the client’s expected lifespan, favor younger clients over older clients. Even in cases where representation promises only to improve the client’s quality of life, favoring younger clients may prove a good rule of thumb since it captures outcome and indirect benefit differentials.

  • Aggregation is not only ethically permitted but ethically desirable in determining rationing most cases.

  • Give priority to cases that raise the likelihood of positive and successful impact litigation results. Treat every person who will benefit from that impact litigation like a “client served”, and do the opposite for cases to generate negative precedent that is harmful to one’s client base.

  • Favor clients with larger number of dependents.

Again, I think the contribution of this article is much richer than this list, but at the end of the day I want to make clear the immediate practical impact of adopting just a few of these suggestions. To be sure, I have left many questions unanswered, among them: who exactly should make the rationing decisions in question? For areas where I have suggested that normative theory cannot provide clear answers, what are the most defensible processes for adopting policies? Should individual LSPs adopt these rules even if they know other LSPs will not? These are important questions, and ones I hope that others will join me in trying to answer going forward. My more modest goal in this article has been to show that the rationing of legal services deserves just as much scrutiny as its medical equivalents, and begin the conversation as to what fair principles might look like.

2 This scarcity will persist for many reasons, not the least of which is that there is no way the government can afford to supply all the free legal assistance individuals would want.
3 The small existing legal literature, which had its heyday in the 1960s, 1970s, and early 1980s, tended to focus on lawyers’ duties under the code of professional responsibility rather than designing just rationing principles (e.g., Matthews & Weiss 1967; Bellow & Kettleson 1978, 343–362) to focus on describing the day-to-day decisions made by legal service providers (e.g., Silver 1968–1969, 221–241; Menkel-Meadow & Meadow 1983), or to focus specifically on the direct services versus impact litigation question (e.g., Breger 1981; Failinger & May 1984, 18–32), which forms only a small part of the issues I address. Further, almost all of this writing studies only civil LSPs, but, as I show, there are equally important rationing issues in criminal cases.
Two works come closer to the methods and aims of this article. The first is ten pages devoted to rationing legal services in David Luban’s book Lawyers and Justice (1988, 306–316), and the second is Paul R. Tremblay’s article Acting “A Very Moral Type of God”: Triage Among Poor Clients (1999). Luban’s work is excellent but quite short, again dealing only with civil LSPs, and is framed as a response to political attacks on LSPs and focused on discrediting lotteries and first-come-first-serve distribution and briefly touching on the aggregation issue. I take up much more in this article and in much greater depth, and engage more deeply the philosophers opposed to aggregation. Tremblay’s outstanding work is focused only on poverty law and departs from his Aristotelian conception of what poverty law is for. I take a broader perspective driven by the bioethics literature and analyze both civil and criminal legal service provision and engage more with consequentialist, deontological, and contractualist perspective, as well as considering many rationing principles that Tremblay does not.
On the criminal side, the only rationing question to get in-depth attention has been whether to prioritize clients who are actually innocent (Brown 2004; Mosteller 2010a). I discuss this issue, but my analysis goes far beyond it.
4 Much of what I say here may also have implications for how law firms allocate their pro bono services. Because they do not take public or charitable funding, though, they may operate under different conceptions of justice more connected to theories of corporate social responsibility.
5 This is a particularly exciting time to examine this question because new scholarly methods are becoming available to allow us to evaluate the success of different legal assistance programs (Greiner & Pattanayak 2012). However, to determine what research questions to study through these methods and to decide what to do with the results requires a normative theory of how LSPs should ration services. My goal in this article is to present exactly that.
6 Throughout the article I assume, for the sake of simplicity, that all service providers are offering litigation services only. Considering trade-offs between offering litigative and transactional services would add an additional layer of complexity that for present purposes I will bracket.
7 In addition to reviewing publicly available information, I also interviewed Julia Leighton, General Counsel of the Public Defender Service for the District of Columbia, on November 22, 2011, and she had an opportunity to review the relevant text. I refer to this source as “Leighton Interview”.
8 In addition to reviewing publicly available information, I also interviewed Steven Eppler-Epstein, Executive Director, Connecticut Legal Services, on September 10, 2012, and he had an opportunity to review the relevant text. I refer to this source as “Eppler-Epstein Interview.”
9 As one CLS lawyer emphasized to me, there is a level of “hidden rationing” going on here in several ways. Many individuals will not know they have a legal problem, i.e., that the problems they are experiencing are ones for which they may have a claim in the legal system. Denial of services to Medicaid patients was, in particular, mentioned as an example. Second, some who know they have a problem for which legal redress may be possible will not know they may be able to get assistance from an LSP. Third, some who may know they are eligible to receive LSP assistance may not call. Fourth, of those who call, some will be discouraged by not being able to talk to an intake staff member immediately and will never call back (Eppler-Epstein Interview 2012). One set of interesting questions I only gesture at here, but hope to discuss in more depth in future work, is how LSPs should think about these populations of individuals who never make it to intake for rationing purposes. In particular, under what circumstances should they divert resources used to represent the existing pool of clients toward attempting to reach these populations?
10 I obtained more information on HLAB’s rationing through a series of structured interviews with Michael Lumley, a student, and HLAB’s intake director from 2010 to 2011. He also added details that modified some of the statements on HLAB’s public documents. Lumley and others at HLAB were given the opportunity to verify all of the information ascribed to him about HLAB. I refer to this source as “Lumley Interview”. These materials were also reviewed and approved by HLAB student leadership. I also had these materials reviewed by David Grossman, Clinical Professor of Law and Director of the Harvard Legal Aid Bureau. When Prof. Grossman suggested additional information or changes to what was publicly available about HLAB, I cite it as “Grossman Interview”.
11 However, when it undertakes community lawyering and focused case representation activities in its foreclosure practice, HLAB does relax these guidelines somewhat (Grossman Interview 2011).
12 When they are unable to take on a case, HLAB makes efforts to refer clients to other legal service providers in the area or where appropriate other clinical programs at Harvard Law School. They will also encourage clients to use limited representation opportunities, such as attorney for the day programs (Grossman Interview 2011; Lumley Interview 2011).
13 A peer reviewer raised the question of whether it would be better to merely provide indigent individuals the equivalent in cash rather than fund these LSPs. Again, there is a parallel question sometimes raised in the health care context. Without going too far afield, there are several reasons to doubt that this would be a good idea. First, it would assume that individuals are fully capable of self-insuring against their need for future legal services using these funds, but given that many of these individuals are living near the poverty level it is far from clear such prudential behavior is likely, especially given what we know about general bounded rationality problems. Second, the benefits of LSPs are not merely at the individual level but through systemic effects and there may be “strength in numbers”—for example, the police may alter behavior knowing that defendants will be represented, a welfare agency may change its review procedure or a bank its foreclosure procedure when threatened with a deluge of coordinated suits brought by an LSP rather than a small number of potential individualized suits (Galanter 1974, 139). Third, there is a debate in the philosophical literature about the fungibility of entitlement claims to health care versus welfare that may be applicable here. For example, Norman Daniels has argued that health has a special importance because of its connection to equality of opportunity, and thus that individuals can make claims relating to (and that governments can properly concern themselves with) the health of citizens in a way that they cannot with a more global conception of well-being or welfare: that is, that I may have a good rights claim against the government to provide me health care but not a Magritte painting, even if the Magritte in fact improves my welfare more and costs the same price (Daniels 2008, 33–35, 51–78). The same may be true as to claims for legal assistance. Finally, notice that even if we were to shift to cash payments the rationing question would persist in a different form: who would get priority in receiving cash payments of what size?
14 Here there is also a particularly interesting sub-question about funding civil versus criminal legal service providers and whether the state should view the two as fungible. While some of what I say at various points in the article touches indirectly on this, I leave full examination of the issue for a future project.
15 Kamm, 2008, offers an illuminating discussion of the application of bioethics principles to other rationing problems that focuses on education.
16 Of course, if what I say about ethical obligations to re-allocate in some circumstances is true, and everyone believed and adopted it, there might be no alternative providers without re-allocation policies.
17 For a discussion of a parallel issue of re-allocation in the medical context of disaster relief see Eyal 2012.
18 Further, the Supreme Court’s ineffective assistance of counsel jurisprudence, especially since Strickland v. Washington, 466 U.S. 668 (1984), has made it exceedingly difficult to prevail on these types of claims. See, e.g., Berguis v. Thompkins, 130 S. Ct. 2250, 2264 (2010).
19 See Luban 1988, 308 (“What, after all, is the moral significance of the fact that you heard about the Public Interest Law Center before I did, or that your problem arose before mine did, or that you caught your bus downtown to the center’s office and I missed mine?”). One possible defense is that more sophisticated clients are likelier to be more successful plaintiffs or defendants. That assumes that success is an important metric for allocation, a matter I discuss below. Even if that was true, it seems more defensible to ration on that basis rather than use this as a very imperfect proxy, unless the trade-off in terms of cost of rationing system versus quality of information gained weighed in favor of first-come-first-serve. I am skeptical, but at the very least, it does not seem that the LSPs using these techniques have any data suggesting this is the case.
20 That said, there is a separate more general risk of malingering or exaggerating one’s description of one’s condition in order to qualify for services. For example, HLAB prioritizes family law cases involving domestic violence, and “[m]any potential clients know of this prioritization system and some indeed game it by emphasizing – and, in some cases, exaggerating – their D[omestic] V[iolence] histories” (Grossman Interview 2011).
21 While this distinction is meaningful, I do not want to exaggerate it; law has so many roles in life that the distinction between “legal need” and other kinds of needs will involve some blurriness.
22 But see Nussbaum 1996, 166–167 (arguing that on her theory “the capabilities are radically non-fungible: deficiencies in one area cannot be made up simply by giving people a larger amount of another capability”).
23 Of course, when savings will be achieved much later in time we face the question of whether there should be discounting and what the discounting rate should be. For an illuminating examination of these kinds of questions with application to law (see Kaplow 2007). In all the examples I use in the article I will not apply a discount rate and treat future gains as equivalent to current ones, out of a desire for expository simplicity. This is a simplification, but one that I think where simplicity can justifiably be favored over accuracy. Those troubled by this can simply adjust the examples to take into account their preferred discount rate.
24 One might object that for some LSP lawyers there is a reverse-visibility effect—that it is the fact that they alone are the one who cares for this particular client that gives them the utility bump. I deal with changes to LSP lawyer satisfaction more in-depth in the next section. However, it is important to emphasize that a focus on statistical instead of identified lives may not change whether an LSP lawyer gains the utility bump from the “rescue” of the “invisible” client. Instead it alters whether that bump accrues from helping the presently identified client, or one later in the time stream who at the moment is only a statistical client, but at the time the service is delivered will be a real flesh and blood individual in need of the service.
25 While I am told that a number of HLAB students do informally advocate taking age into account, they do so to favor the elderly, the opposite of what the philosophy behind age-weighting would direct us to do (Grossman Interview 2011). It is more likely they are implicitly thinking of this as priority to the worst-off in the identified life conception.
26 They make good training cases in part because, as one former PDS attorney now-academic explained to me, there is no jury trial for juvenile cases in D.C. such that the supervisor can confer with the lawyer she is training at PDS without worrying about appearing inexperienced to the jury.
27 Earlier versions drew a line between those (over the age of seventy in one version) who had their “fair innings”, and everyone else (Harris 1985, 87–102; Bognar 2008, 179–81). The more recent versions I will focus on use age-weighting in a more general way, whereby people gradually gain and lose priority as their ages change (Persad et al. 2009, 425).
28 The lower priority of a two-year old over a twenty-year-old on these views is typically justified based on a separate claim that the two-year-old has not yet “received substantial education and parental care, investments that will be wasted without a complete life”, and/or the idea that the two-year-old lacks a developed personality or complex life goals such that it has “less to lose” (Persad et al. 2009). This particular claim has been the subject of significant critique (see, e.g., Kerstein & Bognar 2010, 39–40). I will not dwell on the priority for very young children as it comes up infrequently for LSPs.
29 In passing, Darryl Brown has suggested that “[w]hile attorneys can be tempted to favor the young first-offender over the older recidivist when each is wrongly accused, it is more defensible”, not to do so because it is “ethically fraught” (Brown 2004, 819)—without explaining the ethical problem.
30 What about the “delayed action” of something like prison’s influence on one’s lifespan, does preventing that count as saving lives? If “save the most lives” and “save the most life years” are to remain distinct ideas, it cannot, and the Rule of Rescue is usually framed as preventing imminent death. One might object to the analogy by suggesting the presence of state action here in carrying out the death distinguishes the case. In fact, though, the mere presence of state action does not to distinguish the medical from LSP context: every time that Medicare refuses to cover a potentially life saving treatment through a state rationing principle someone has died whose death may have been averted. Can one nonetheless distinguish the cases as involving state action versus inaction, hurting versus failing to help? I think this is an incorrect description of the cases. The LSP that does not devote resources to a death eligible accused fails to help him avoid being executed just as the organ allocation system that fails to allocate an organ needed by someone who will otherwise die fails to help him avoid being killed by the disease. It is of no moment that in one case the cause of the death that might be avoided is a state actor (a sort of human agency) and in the other a disease. What matters is that in each case the cause of the death is something other than the entity that is providing assistance, and whose assistance must be rationed.
31 The age difference between the two is deliberate in order to make their life expectancies without transplant the same (nullifying a fair innings argument).
32 Here I am imagining that the life years saved would be “concentrated” in another individual to momentarily bracket the aggregation question.
33 Brown and Tremblay have pushed for a similar principle among criminal defense lawyers and poverty lawyers respectively (see Brown 2004, 818–819; Tremblay 1999, 2490–2491).
34 In theory, there is no reason why a criminal defense LSP like PDS ought to focus exclusively on the length of sentence rather than also taking into account its collateral consequences for the accused. For example, an individual facing charges on her first offense faces bad outcomes for employment, housing, etc., if convicted, that an individual who already has a long list of prior convictions may not. If the rationing system was put in place prospectively only (i.e., for new accused going forward from X date), then there is an argument that this is not discriminatory, paralleling the argument as to age-weighting: when the accused with a long list of prior offense was “young” in the sense of being a first-time accused, she too benefitted from this preference, so it is not unfair to favor other first-timers now. Nor, for reasons discussed below, do I think it is a problem that avoiding these collateral consequences would provide “indirect benefits” to individuals rather than direct benefits. The more serious objection to considering collateral consequences is, once again, the likelihood of error and the cost of investment necessary to adequately take account of this data. It may be that once these concerns are factored in, it is superior to ignore these collateral consequences or to consider them but reduce their weight. This is the kind of administrability question that will depend to some extent on the on-the-ground reality of the way the LSP currently operates, and possible feasible institutional re-alignments.
35 In the present discussion we are taking the “simple” rationing principles one-by-one, but this example illustrates how when combined they may pull in opposite directions. Best outcomes favor Dale, but a variant of priority to the worst-off with a wide sphere definition may favor Chip.
36 Less likely but not impossible. There may be an analogous problem in determining how bad going to a particular type of prison should be rated, and whether individuals “adapt” to prison conditions as to disabilities.
37 If readers are bothered by the mixing of separate spheres, we can make the disability a legal one such as garnishing of Jacob’s wages due to a prior debt.
38 Tim Scanlon (1998) makes a similar contractualist argument.
39 Cf. Tremblay (1999, 2491).
40 To be sure, the comparison is a little less clear here where we have to compare rent rebates to eviction rather than varying years of incarceration.
41 In later work, Kamm has expanded upon these points in such a way that might suggest a slightly different reading of her view. First, she has made clear that "[w]hether a good is irrelvant is context dependant" such that "[c]uring a sore throat is morally irrelevant when others' lives are at stake, but not when other ear aches are." (Kamm 2000, 221). She also frames the inquiry in terms of what a nonconsequentialist duty to"to suffer a relatively minimal loss (e.g., a sore throat) to save another person's life" or to "give someone a chance at life" (id. at 221). Moreover, similar to a point Brock (2003, 11) makes as to indirect benefit counting that I will discuss below, Kamm seems open to the idea that macro decision-making may have different aggregation rules. She writes that for macro "decisions, for example, whether to invest in research to cure a disease that will kill a few people or in research to cure a disease that will only with an arm in many," we might "permit aggregation of significant (not insignificant) losses to many people to outweigh even greater losses to a few, even when no individual in the larger group will lose as much as each individual in the smaller group" (Kamm 2000, 222). She think this last principle distinguishes her views from the most common contractualist ones (id.).
42 Of course, this depends somewhat on the contours of “irrelevant utilities”. HLAB for example, gets occasional phone calls for legal assistance in small claims matters it usually screens out (Grossman Interview 2011).
43 Tremblay (1999, 2495) briefly discusses this possibility under the rubric of rationing by “social worth”, but he mixes together the responsibility and contribution-based versions in such a way that he misses the strongest objections against responsibility-sensitivity allocation while at the same time condemning forms of contribution-based allocation that I find defensible.
44 For more on attempts to save luck egalitarianism from the harshness objection, see Segall (2010, 58–73).
45 This differs from the more common questions about professional responsibility duties in impact and cause lawyering (Cause Lawyering 1997). I am asking a prior question: would it be ethical to choose between claimants to represent based on the impact the cases would have on other cases?
46 This hypothetical is adapted from a case suggested by Kamm (1993, 107–108), but I have added blood-type locking to prevent the possibility of giving it to both of the individuals immediately present and then distributing what is left to those more distantly located.
47 One objection considered by Luban (1988, 315) to aggregation in the context of impact litigation is that not all those who benefit from the impact litigation are “clients” who have come directly to the LSP for aid and thus should not be counted. Luban says this objection is premised on the incorrect assumption that a lawyer does not “fulfill someone’s right to access the legal system by ‘representing’ that person, [w]hen the person has not requested it”, and confuses the client control objection to cause lawyering with the question of distribution of resources (id. at 315–316). I think he is right, and the Surgeon and Flo-Jo cases help us see that the objection is problematic in that in both cases the five people who will be helped are not yet clients, but that does not defeat the claim that just allocation requires favoring the Surgeon or Flo-Jo. The objection seems to again focus improperly on identified lives.
48 Other than in the impact litigation context, I suspect that cases like the Flo-Jo and Surgeon case will be less common on the legal side in part because many people with important jobs generating significant indirect benefits will have the resources to purchase assistance outside the LSP model.
49 One way of operationalizing this would be by adopting a fixed per-client “bonus” for testing the prosecution’s case similar to the one I discuss in Section 4 as a way of accommodating the role of dignity concerns.
50 On the medical side, for what he calls “unsolved rationing problems”, Norman Daniels (2008, 274–296) has made a similar push for a process like this that he calls “Accountability for Reasonableness” In this work I am less concerned with the exact contours of what such a public deliberation about rationing would look like, and open to many possible configurations. I am sensitive to the worry that too much publicity about rationing principles might cause potential clients to fake their histories in order to get priority.
51 A different alternative would be to use something like the experimental survey techniques discussed above that are used to generate QALY time-trade-off, standard gamble, or the visual analog scale (McKie & Richardson 2003, 2412), to instead weight the different proportions given to each consideration. That would tell us how individuals would set it, not, however whether those conclusions are justified or the result of a reflective process.
52 Of course even use of the word “just” here is slippery. LSP lawyers may say that victories which would not be legally “correct” are “just” if they advance social justice goals (Grossman Interview 2011).
53 It may depend on whether one defines RP individually or by type. A particular small-fry landlord may not be able to reap the advantages that Galanter typically associates with RP— advance intelligence in setting up the transaction, access to specialists, relations with institutional incumbents, ability to play the odds, etc.)—even if Galanter so classifies them (1974, 101–104, 107).
54 See Mosteller (2010a, 965–973) (noting these problems on the criminal defense side). But see Brown (2004, 218–220 (arguing that defense attorneys can routinely determine the innocence or guilt of their clients with a “reasonable level of confidence”).
55 Some institutional litigants, like a welfare agency or the prosecutor’s office, may be somewhat insensitive to the cost of litigation against different kinds of parties and instead be driven by different incentive structures.
56 Tremblay (1999, 2498) briefly notes a similar concern,but suggests it should play only a limited role as a “tie-breaking factor among competing important matters”. This is puzzling. If lawyer satisfaction has a significant role in lawyer performance (as Tremblay appears to accept), it seems though as we ought to give it as much weight as we would differences in the likelihood of success in individual cases. While in this section I treat the effect of rationing systems on lawyer satisfaction as analytically distinct from the aggregate effect on the outcomes valued by the rationing principles, I fully acknowledge that in reality the two are likely to be heavily intertwined.
57 If instead the system is changed midway through the LSP’s existence, there may be some fair notice issues, although presumably most LSP lawyers know that client selection policy is always susceptible to change.
58 Compare, for example, an LSP that represents religious individuals facing discrimination and one focused on promoting secularism.

REFERENCES

Adler
Matthew
William Sanchirico
Chris
Inequality and Uncertainty: Theory and Legal Applications
Univ. Penn. Law Rev.
 , 
2006
, vol. 
155
 (pg. 
279
-
377
)
Althoff
Barrie
Ethical Issues Posed by Limited-Scope Representation – The Washington Experience
Prof. Law.
 , 
2004
, vol. 
2004
 (pg. 
67
-
100
)
Anderson
Elizabeth
What is the Point of Equality?
Ethics
 , 
1999
, vol. 
109
 (pg. 
287
-
337
)
 
Aristotle (W.D. Ross & J.O. Urmsom trans.). 1984. Nicomachean Ethics. In Jonathan Barnes (ed.), The Complete Works of Aristotle, Bk 1, ch. 1. Princeton, NJ: Princeton University Press
Bellow
Gary
Kettleson
Jeanne
From Ethics to Politics: Confronting Scarcity and Fairness in Public Interest
Boston Univ. Law Rev.
 , 
1978
, vol. 
58
 (pg. 
337
-
391
)
Bognar
Greg
Age-weighting
Econ. & Phil.
 , 
2008
, vol. 
24
 (pg. 
167
-
189
)
Breger
Marshall J
Legal Aid for the Poor: A Conceptual Analysis
North Carolina Law Rev.
 , 
1981
, vol. 
60
 (pg. 
281
-
364
)
Brock
Dan W
Mathieu
Deborah
Ethical Issues in Recipient Selection for Organ Transplantation
Organ Substitution Technology: Ethical, Legal, and Public Policy Issues
 , 
1988
, vol. 
86
 
Boulder and London
Westview Press
Brock
Dan W
Rhodes
Rosamond
, et al.  . 
Priority to the Worse Off in Health-Care Resource Prioritization
Medicine and Social Justice: Essays on the Distribution of Health Care
 , 
2002
, vol. 
362
 
New York
Oxford University Press
Brock
Dan W
Separate Spheres and Indirect Benefits, Cost Effectiveness & Resource Allocation
 , 
2003
 
Brock
Dan W
Guttman
Thomas
, et al.  . 
The Misplaced Role of Urgency in Allocation of Persistently Scarce Life-Saving Organs
Ethical, Legal, and Social Issues in Organ Transplantation
 , 
2004
, vol. 
41
 
Lengerich, Germany
Pabst Science Publishers
Broome
John
Murray
Christopher J L
, et al.  . 
All Goods are Relevant
WHO, Summary Measures of Population Health: Concepts, Ethics, Measurement and Applications
 , 
2002
, vol. 
727
 
Geneva
World Health Organization
Brown
Darryl K
Essay, Rationing Criminal Defense Entitlements: An Argument from Institutional Design
Columbia L. Rev.
 , 
2004
, vol. 
104
 (pg. 
801
-
835
)
Callahan
Daniel
Must We Ration Health Care for the Elderly?
J. L. Med. Ethics.
 , 
2012
, vol. 
40
 (pg. 
10
-
15
)
Chang
Howard F
A Liberal Theory of Social Welfare: Fairness, Utility, and the Pareto Principle
Yale L. J.
 , 
2000
, vol. 
110
 (pg. 
173
-
235
)
Cohen
I Glenn
Note, The Price of Everything, the Value of Nothing: Reframing the Commodification Debate
Harvard L. Rev.
 , 
2003
, vol. 
117
 (pg. 
689
-
710
)
Cohen
I Glenn
Medical Tourism, Access to Health Care, and Global Justice
Virginia J. Int’l L.
 , 
2011
, vol. 
52
 (pg. 
1
-
56
)
Connecticut Legal Services
About CLS
 , 
2012
 
Connecticut Legal Services
Annual Report
 , 
2011
 
Daniels
Norman
Am I My Parents Keeper? An Essay on Justice Between the Young and the Old
 , 
1988
New York
Oxford University Press
Daniels
Norman
Fair Process in Patient Selection for Antiretroviral Treatment in WHO’s Goal of 3 by 5
Lancet.
 , 
2005
, vol. 
366
 (pg. 
169
-
171
)
Daniels
Norman
Just Health: Meeting Health Needs Fairly
 , 
2008
, vol. 
24
 
Cambridge, UK
Cambridge University Press
Daniels
Norman
Reasonable Disagreement About Identified v. Statistical Victims
Hastings Cent. Rep.
 , 
2011
, vol. 
41
 (pg. 
35
-
45
)
Dworkin
Ronald
No Right Answer?
New York Univ. L. Rev.
 , 
1978
, vol. 
53
 (pg. 
1
-
32
)
Dworkin
Ronald
Life’s Dominion
:
An Argument About Abortion, Euthanasia, and Individual Freedom
 , 
1993
New York
Knopf, Inc
 
Egalitarianism. In Stanford Encyclopedia on Philosophy. Available at http://plato.stanford.edu/archives/fall2008/entries/egalitarianism/#5.2
Elhauge
Einer
Allocating Health Care Morally
Calif. L. Rev.
 , 
1994
, vol. 
82
 (pg. 
1449
-
1544
)
Eyal
Nir
Egalitarian Justice and Innocent Choice
J. Ethics Soc. Phil.
 , 
2007
, vol. 
2
 (pg. 
1
-
18
)
Eyal
Nir
Firth
Paul
Repeat Triage in Disaster Relief: Questions from Haiti
PLoS Curr.
 , 
2012
, vol. 
4
 pg. 
e4fbbdec6279ec
  
Eyal
Nir
Deep Exclusionary Reasons: The Case of Luck Egalitarianism and Personal Responsibility for Health
unpublished
 
Faust
Halley S
Menzel
Paul T
Prevention vs. Treatment: What’s the Right Balance?
 , 
2012
New York
Oxford University Press
Failinger
Marie A
May
Larry
Litigating Against Poverty: Legal Services and Group Representation
Ohio St. L. J.
 , 
1984
, vol. 
45
 (pg. 
1
-
56
)
Fisher
William W
Syed
Talha
Global Justice in Health Care: Developing Drugs for the Developing World
Univ. Calif. Davis L. Rev.
 , 
2007
, vol. 
40
 (pg. 
581
-
678
)
Fried
Charles
The Value of Life
Harvard L. Rev.
 , 
1969
, vol. 
82
 (pg. 
1415
-
1437
)
Frug
Gerald E
The Ideology of Bureaucracy in American Law
Harvard L. Rev.
 , 
1984
, vol. 
97
 (pg. 
1276
-
1388
)
Fuller
Lon L
The Forms and Limits of Adjudication
Harvard L. Rev.
 , 
1978
, vol. 
92
 (pg. 
353
-
409
)
Galanter
Marc
Why the “Haves” Come Out Ahead: Speculations on the Limits of Legal Change
L. Soc’y Rev.
 , 
1974
, vol. 
9
 (pg. 
95
-
160
)
Gardner
Debra
Justice Delayed is, Once Again, Justice Denied: The Overdue Right to Counsel in Civil Cases
Univ. Balt. L. Rev.
 , 
2007
, vol. 
37
 (pg. 
59
-
77
)
Greiner
D James
Pattanayak
Cassandra Wollos
Randomized Evaluation in Legal Assistance: What Difference Does Representation (Offer and Actual Use) Make?
Yale L. J.
 , 
2012
, vol. 
121
 (pg. 
2118
-
2214
)
Goodwin
Michele
Altruism’s Limits: Law, Capacity, and Organ Commodification
Rutgers L. Rev.
 , 
2004
, vol. 
56
 (pg. 
305
-
407
)
Gordon
Robert
The Independence of Lawyers
Boston Univ. L. Rev.
 , 
1988
, vol. 
68
 (pg. 
1
-
83
)
Gruenewald
David A
Can Health Rationing Ever Be Rational?
J. L. Med. Ethics.
 , 
2012
, vol. 
40
 (pg. 
17
-
24
)
Harvard Legal Aid Bureau
About Us
 , 
2012
 
Harris
David A
The Stories, the Statistics, and the Law: Why “Driving While Black” Matters
Minnesota L. Rev.
 , 
1999
, vol. 
84
 (pg. 
265
-
326
)
Harris
John
The Value of Life: An Introduction to Medical Ethics
 , 
1985
, vol. 
80
 
New York
Routledge & Kegan Paul
HHS, Pandemic Influenza Plan
Appendix D, Table D-1
 , 
2005
 
Hogg
Robert S
, et al.  . 
Years of Life Lost in Prison: Racial and Gender Gradients in the United States of America
Harm Reduction J.
 , 
2008
, vol. 
5
 (pg. 
1
-
5
)
 
Interview with Steven Eppler-Epstein, Executive Director, Connecticut Legal Services. 2012
 
Interview with David Grossman, Clinical Professor of Law and Director of the Harvard Legal Aid Bureau, in Cambridge, MA. 2011
 
Interview with Julia Leighton, General Counsel of the Public Defender Service for the District of Columbia. 2011
 
Interview with Michael Lumley, Intake Director (2010–2011) for Harvard Legal Aid Bureau, in Cambridge, MA. 2011
Kamm
Frances M
Morality, Mortality: Death and Whom to Save from It
 , 
1993
New York
Oxford University Press
Kamm
Frances M
To Whom?
Hastings Cent. Rep.
 , 
1994
, vol. 
24
 (pg. 
29
-
32
)
Kamm
Frances M
Hugh
LaFollette
Nonconsequentialism
2000
 
The Blackwell Guide to Ethical Theory, 205. Malden, MA: Blackwell Publishers
Kamm
Frances M
Thomas
Laurence
Moral Issues in Rationing Scarce Resources
Contemporary Debates in Social Philosophy
 , 
2008
 
263. Malden, MA: Blackwell Publishing
Kaplow
Louis
Discounting Dollars, Discounting Lives: Intergenerational Distributive Justice and Efficiency
Univ. Chicago L. Rev.
 , 
2007
, vol. 
74
 (pg. 
79
-
118
)
Kant
Immanuel
1785
 
(James W. Ellington trans., 1981). Grounding for the Metaphysics of Morals, 434. Indianapolis/Cambridge: Hackett Publishing Company, Inc
Kerstein
Samuel
Bognar
Greg
Complete Lives in the Balance
Am. J. Bioethics.
 , 
2010
, vol. 
10
 (pg. 
37
-
45
)
Larsen
Allison Orr
Bargaining Inside the Black Box
Georgetown L. J.
 , 
2011
, vol. 
99
 (pg. 
1567
-
1612
)
Legal Services Corporation
Documenting the Justice Gap in America: The Current Unmet Civil Legal Needs of Low-Income Americans: An Updated Report of the Legal Services Corporation
 , 
2009
 
Liebman
James S
Clarke
Peter
Minority Practice, Majority’s Burden: The Death Penalty Today
Ohio St. J. Crim. L.
 , 
2011
, vol. 
9
 (pg. 
255
-
351
)
Luban
David
Luban
David
The Adversary System Excuse
The Good Lawyer: Lawyers’ Roles and Lawyers’ Ethics
 , 
1983
, vol. 
83
 
Totowa, N.J.
Rowman & Allanheld
Luban
David
Lawyers and Justice: An Ethical Study
 , 
1988
Princeton, NJ
Princeton University Press
Markovits
Daniel
Luck Egalitarianism and Political Solidarity
Theoretical Inq. L.
 , 
2008
, vol. 
9
 (pg. 
271
-
308
)
Mashaw
Jerry L
The Supreme Court’s Due Process Calculus for Administrative Adjudication in Mathews v. Eldridge: Three Factors in Search of a Theory of Value
Univ. Chicago L. Rev.
 , 
1979
, vol. 
44
 (pg. 
28
-
59
)
Mashaw
Jerry L
Administrative Due Process: The Quest for a Dignitary Theory
Boston Univ. L. Rev.
 , 
1981
, vol. 
61
 (pg. 
885
-
931
)
Mashaw
Jerry L
Due Process in the Administrative State
 , 
1985
New Haven, CT
Yale University Press
Matthews
Arther R
Weiss
Jonathan A
What Can Be Done: A Neighborhood Lawyer’s Credo
Boston Univ. L. Rev.
 , 
1967
, vol. 
47
 (pg. 
231
-
243
)
McKie
John
Richardson
Jeff
The Rule of Rescue
Soc. Sci. Med.
 , 
2003
, vol. 
56
 (pg. 
2407
-
2419
)
Menkel-Meadow
Carrie
Meadow
Robert G
Resource Allocation In Legal Services
L. Pol’y Q.
 , 
1983
, vol. 
5
 (pg. 
237
-
256
)
Menzel
Paul
, et al.  . 
The Role of Adaptation to Disability and Disease in Health State Valuation: A Preliminary Normative Analysis
Soc. Sci. Med.
 , 
2002
, vol. 
55
 (pg. 
2149
-
2158
)
Metzger
Pamela R
Beyond the Bright Line: A Contemporary Right-to-Counsel Doctrine
Northwestern Univ. L. Rev.
 , 
2003
, vol. 
97
 (pg. 
1635
-
1699
)
Mosteller
Robert P
Protecting the Innocent: Part of the Solution for Inadequate Funding for Defenders, not a Panacea for Targeting Justice
Missouri L. Rev.
 , 
2010a
, vol. 
75
 (pg. 
931
-
988
)
Mosteller
Robert P
Why Defense Attorneys Cannot, But Do, Care About Innocence
Santa Clara L. Rev.
 , 
2010b
, vol. 
50
 (pg. 
1
-
78
)
Nagel
Thomas
Equality. In
Mortal Questions
 , 
1979
, vol. 
106
 
Cambridge, UK
Cambridge University Press
Nussbaum
Martha C
Women and Human Development: The Capabilities Approach
 , 
2000
Cambridge, UK
Cambridge University Press
Nussbaum
Martha C
Frontiers of Justice: Disability, Nationality, Species Membership
 , 
2006
, vol. 
166–167
 
Cambridge, MA
Harvard University Press
Parfit
Derek
Lindley Lecture at the University of Kansas: Equality or Priority?
Reprinted in Ratio
 , 
1997
, vol. 
10
 (pg. 
202
-
221
)
Persad
Govind
, et al.  . 
Principles for Allocation of Scarce Medical Interventions
Lancet.
 , 
2009
, vol. 
373
 (pg. 
423
-
431
)
Performance Audit Report
Costs Incurred for Death Penalty Cases: A K-GOAL Audit of the Department of Corrections, Kansas Legislative Post Audit Committee
 , 
2003
 
http://www.kansas.gov/postaudit// audits_perform/04pa03a.pdf
The Public Defender Service for the District of Columbia
Who We Are
 , 
2012
 
Rawls
John
A Theory of Justice, Revised Edition
 , 
1971, 1999
Cambridge, MA
Harvard University Press
Roman
John
, et al.  . 
The Cost of the Death Penalty in Maryland
 , 
2008
Washington, D.C.
Urban Institute
Rubenstein
William B
The Concept of Equality in Civil Procedure
Cardozo L. Rev.
 , 
2002
, vol. 
23
 (pg. 
1865
-
1915
)
Ruger
Todd
Once More, LSC Budget at Risk
National L. J.
 , 
2012
 
Rosen-Zvi
Issachar
Just Fee Shifting
Florida St. Univ. L. Rev.
 , 
2010
, vol. 
37
 (pg. 
717
-
767
)
Rulli
Louis S
On the Road to Civil Gideon: Five Lessons from the Enactment of a Right to Counsel for Indigent Homeowners in Federal Civil Forfeiture Proceedings
J. L. Pol’y.
 , 
2011
, vol. 
19
 (pg. 
683
-
750
)
Sandel
Michael J
Justice: What’s the Right Thing to Do?
 , 
2009
New York
Farrar, Straus, & Giroux
Sanders
David
Dukeminier
Jesse
Jr
Medical Advance and Legal Lag: Hemodialysis and Kidney Transplantation
UCLA L. Rev.
 , 
1968
, vol. 
15
 (pg. 
357
-
413
)
Sarat
Austin
Scheingold
Stuart A
Cause Lawyering: Political Commitments and Professional Responsibilities
 , 
1998
Saver
Richard S
In Tepid Defense of Population Health: Physicians and Antibiotic Resistance
Am. J. L. Med.
 , 
2008
, vol. 
34
 (pg. 
431
-
491
)
Segall
Shlomi
Health, Luck, and Justice
 , 
2010
Princeton, NJ
Princeton University Press
Scanlon
Thomas M
What We Owe to Each Other
 , 
1998
Cambridge, MA
Harvard University Press
Silver
Carol R
The Imminent Failure of Legal Services for the Poor: Why and How to Limit Caseload
J. Urban L.
 , 
1968–1969
, vol. 
46
 (pg. 
217
-
248
)
Simon
William H
Ethical Discretion in Lawyering
Harvard Law Rev.
 , 
1988
, vol. 
101
 (pg. 
1083
-
1145
)
Simon
William H
Role Differentiation and Lawyers’ Ethics: A Critique of Some Academic Perspectives
Georgetown J. Legal Ethics.
 , 
2010
, vol. 
23
 (pg. 
987
-
1009
)
The Simpsons
Radio Bart
 , 
1992
Fox
Singer
Joseph William
The Player and the Cards: Nihilism and Legal Theory
Yale L. J.
 , 
1984
, vol. 
94
 (pg. 
1
-
70
)
Solum
Lawrence B
Procedural Justice
Southern Calif. L. Rev.
 , 
2004
, vol. 
78
 (pg. 
181
-
321
)
Steiker
Carol S
Steiker
Jordan M
Report to the ALI Concerning Capital Punishment
Texas L. Rev.
 , 
2010
, vol. 
89
 (pg. 
367
-
421
)
Strech
D
Persad
G
Marckmann
G
Danis
M
Are Physicians Willing to Ration Health Care? Conflicting Findings in a Systematic Review of Survey Research
Health Pol’y
 , 
2009
, vol. 
90
 (pg. 
113
-
124
)
Stuntz
William J
Plea Bargaining and Criminal Law’s Disappearing Shadow
Harvard L. Rev.
 , 
2004
, vol. 
117
 (pg. 
2548
-
2569
)
Sumner
LW
Welfare, Happiness, and Ethics
 , 
1996
New York
Oxford University Press
Sunstein
Cass
If People Would Be Outraged By Their Rulings, Should Judges Care?
Stanford L. Rev.
 , 
2007
, vol. 
60
 (pg. 
155
-
212
)
Taurek
John
Should the Numbers Count?
Phil. Public Affairs.
 , 
1977
, vol. 
6
 (pg. 
293
-
316
)
Tremblay
Paul R
Acting “A Very Moral Type of God”: Triage Among Poor Clients
Fordham L. Rev.
 , 
1999
, vol. 
67
 (pg. 
2475
-
2532
)
Walzer
Michael
Spheres of Justice: A Defense of Pluralism and Equality
 , 
1983
New York
Basic Books
Weir
Robert F
The Issue of Fairness in the Allocation of Organs
J. Corp. L.
 , 
1995
, vol. 
29
 (pg. 
91
-
108
)
Wertheimer
Alan
The Equalization of Legal Resources
Phil. & Public Affairs.
 , 
1988
, vol. 
17
 (pg. 
303
-
322
)
Wikler
Daniel
Personal and Social Responsibility for Health
Ethics Int’l Affairs.
 , 
2002
, vol. 
16
 (pg. 
47
-
55
)
Williams
Alan
Intergenerational Equity, an Exploration of the ‘Fair Innings’ Argument
Health Econ.
 , 
1997
, vol. 
6
 (pg. 
117
-
132
)

Author notes

1 Assistant Professor, Harvard Law School. Co-Director, Petrie-Flom Center for Health Law Policy, Biotechnology, and Bioethics. J.D., Harvard Law School. E-mail: Igcohen@law.harvard.edu. I thank Libby Adler, Rachel Brewster, Dan Brock, Gabby Blum, Cansu Cana, John Coates, Norman Daniels, Einer Elhauge, Steve Eppler-Epstein Nir Eyal, Terry Fisher, Charles Fried, Jim Greiner, David Grossman, Alon Harel, Janet Halley, Howell Jackson, Frances Kamm, Andy Kaufman, Duncan Kennedy, Adrian Laani, Julia Leighton Holly Lynch, Kristin Madison, Dan Meltzer, Michelle Meyer, Wendy Parmet, David Phillips, Mark Ramseyer, Ben Roin, James Rowan Ben Sachs, Michael Sandel, Jed Shugerman, Jeff Skopek, Henry Smith, Carol Steiker, Holger Spamann, Matthew Stephenson, David Wilkins, and Mark Wu for comments on earlier drafts. I also thank participants at the Harvard population-level bioethics reading group, the Northeastern University Law Faculty Workshop, and the Northeast Consequentialist Workshop. Alex Anzalone, David Attanasio, Mollie Bracewell, Joseph Brothers, Sean Driscoll, Samara Fox, Rebecca Livengood, Keely O’Malley, Erik Graham-Smith, Lisa Sullivan, and Desta Tedros provided excellent research assistance.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0/), which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.