A newly released report from RAND, summarized in a NYT editorial, reports radically better outcomes for public defender represented murder defendants that those assigned to a private lawyer. The Philadelphia study was randomized, removing most of the likely methodological objections.  (However there are complications from “crossover” effects in which defendants moved from one pool to the other.  The researchers made serious attempts to correct for these possibilities.)  Here is the abstract, which summarizes the results:

One in five indigent murder defendants in Philadelphia are randomly assigned representation by public defenders while the remainder receive court-appointed private attorneys. We exploit this random assignment to measure how defense counsel affect murder case outcomes. Compared to appointed counsel, public defenders in Philadelphia reduce their clients’ murder conviction rate by 19% and lower the probability that their clients receive a life sentence by 62%. Public defenders reduce overall expected time served in prison by 24%. We find no difference in the overall number of charges of which defendants are found guilty. When we apply methods used in past studies of the effect of counsel that did not use random assignment, we obtain far more modest estimated impacts, which suggests defendant sorting is an important confounder affecting past research. To understand possible explanations for the disparity in outcomes, we interviewed judges, public defenders, and attorneys who took appointments. Interviewees identified a variety of institutional factors in Philadelphia that decreased the likelihood that appointed counsel would prepare cases as well as the public defenders. The vast difference in outcomes for defendants assigned different counsel types raises important questions about the adequacy and fairness of the criminal justice system.

There are, of course, numerous possible contributing explanations for these differences.

• Caps on preparation time.  According to the study, counsel appointed in murder cases in Philadelphia “receive flat fees for pre-trial preparation—$1333 if the case is resolved prior to trial and $2000 if the case goes to trial. The $2000 also includes the first half-day of trial.“
• Hourly rates are also relatively low.  “While on trial, lawyers receive $200 for three hours of court time or less, and $400/day for more than three hours.”
• As the study puts it: “For reasons of local institutional history, the appointed counsel system seems very likely to result in comparatively poor defense counsel function.”
• Economic conflicts of interest in the appointment process.  “Thus judges have incentives to appoint counsel who file fewer pre-trial motions, ask fewer questions during voir dire, raise fewer objections, and present fewer witnesses.75 Quite apart from reducing the expenditures paid to counsel, this also allows judges to process more cases in less time.”
• Political conflicts of interest:  “Historically, judges have also purportedly assigned cases to lawyers with whom they had political connections. . . . Today, opinion is mixed with respect to whether political considerations continue to play a role in the appointments with most respondents indicating that this played much less of a role than in the past.”
• Relative difficulty of hiring expert witnesses, investigators, and mitigation specialists. (Court approval is required for assigned counsel to do so.
• Relative isolation versus integration into a defense culture.

The conclusion to the paper includes the following:

Consider the following thought experiment: suppose the 2,459 defendants in our sample represented by appointed counsel had been represented instead by Defender Association counsel? Based on the results in Table Two, we would expect 270 defendants who were convicted of murder to have been entirely acquitted of this charge with Defender Association representation. Three hundred ninety-six individuals who received life sentences would have been spared a life sentence. In aggregate, we would expect the time served by the 2,459 defendants for the crimes observed in our data to decrease by 6,400 years.

Prison costs for these crimes would also be affected. Recent estimates place the cost of incarcerating a prisoner for one year in Pennsylvania at roughly $32,000 so a decrease of 6,400 years would reduce prison costs for these crimes by over $200 million. (footnotes omitted.)

There are, of course, a number of different ways of explaining the underlying phenomenon.  Which one you choose (or prove by doing additional research) will have major implications for how you think rights to access to justice should be made possible in both the civil and criminal areas.  Some possibilities:

• The inherent quality argument (one that, when in the past I was a public defender, I was perhaps guilty of feeling some sympathy for.)  This theory is that those who are salaried are inherently better lawyers.  I have now met too many dedicated private lawyers for that theory to hold water for me .
• The defender culture argument.  This is the idea that public defenders are (usually) in a subculture that rewards intensely zealous representation, withe the result that the same person, if employed in a public defender office, will do a better job.
• The overall institutional environment argument.  This is the idea (well illustrated by the study), that the overall environment of payment rates, judicial involvement in assignment and payments, availability of support services, makes it very hard for assigned lawyers to do as good a job as salaried defenders.
• The impossibility of equality of instutional environemnt.  This is a version of the above argument, but goes further and says that, as a political matter, it will always be the case that an institutional defender organization will be better able to protect its lawyers from outside pressures and financial incentives tending to lower quality representation.  Such a theory would best be tested in an environment such as Massachusetts in which one Board and administrative structure manage and (hopefully) protect both salaried and assigned lawyers.

I have long advocated a mixed model, noting that in European legal aid systems, the majority of services are delivered through private lawyers, and the salaried part of the system is in the minority — and is subject to the same kinds of criticisms to which private attorney participation is subjected in the US.  A better understanding of the dynamics suggested in the study would help in system design as we hopefully expand the US civil system in some way.