« Second Circuit panel rules that district court lacks ancillary jurisdiction to expunge a valid conviction | Main | Could "conservative Latinos religious groups" become a significant voice and force in death penalty debates (at least in California)? »

August 11, 2016

"Public Defenders vs. Private Court Appointed Attorneys: An Investigation of Indigent Defense Systems"

The title of this post is the title of this intriguing empirical paper recently posted on SSRN authored by Yotam Shem-Tov. Here is the abstract (with one line emphasized by me):

Individuals facing criminal charges in the U.S. have a constitutional right to attorney representation.  If they cannot afford one, the court is required to appoint and finance a legal counsel. Indigent defense systems are usually composed of private court appointed attorneys and/or a public defenders’ organization.  I investigate the causal effect of being assigned a public defender as oppose to a private court appointed attorney on defendants’ trial outcomes using a new “twins design” identification strategy.

I argue and show empirically that in co-defendant cases, the decision of who is assigned to the public defender organization can be treated as close to a randomized experiment, which can be exploited to measure the effectiveness of court appointed private attorneys relative to public defenders.  Using data from all multiple defendant cases in federal courts between 2001-2014, I find that defendants assigned a public defender in co-defendant cases had slightly worse outcomes: a higher probability of being convicted, an average of three months longer expected prison sentence, longer court proceedings and a higher probability of reaching a plea bargain.  However, there is large heterogeneity in public defender effectiveness across federal districts ranging from a 13.8 month longer prison term to a 16.1 month shorter prison term.

I lack the empirical skills to question or assess the new “twins design” identification strategy used in this paper, and the full paper is full of challenging empirical jargon like "The main regression specications is Yit = B . PDi + aj(I) + X'it􀀀 + Eit".  That said, my first reaction is to be quite suspicious of these findings/conclusions due to (1) my own experiences and high regard for the work of nearly all federal public defenders, and (2) my own normative instinct that, at least for a significant number of federal defendants, experiencing "longer court proceedings and a higher probability of reaching a plea bargain" is not at all a marker of a "worse outcome."

UPDATE months later: As noted in this post, the author revised this paper based on feedback and a new data analysis and reached a significantly different finding: "I find that public defenders out-perform court-appointed attorneys in a range of sentencing outcomes."

August 11, 2016 at 05:15 PM | Permalink


Your write, " my own experiences and high regard for the work of nearly all federal public defenders,"

That misunderstands his conclusion. He doesn't make any remarks about the quality of the PD. His claim is prosecutors CHARGE defendants differently when they are represented by a PD vs a CA attorney.

"Table 16 columns (1) and (2) show that gaps in charge characteristics are translated to large and significant differences in the expected sentencing outcomes under the control (CA) regime. Almost the entire observed difference in predicted sentenced prison length of −10.83 between defendants represented by a PD or a CA can be explained by this selection index of −9.72." (page 7).

So if I understand him correctly his claim is perceptual--prosecutors see public defenders as weaker than CA attorneys and so they charge more aggressively. This weakness may have nothing to do with the PD qualifications but could equally reflect the fact that PDs are overburdened, overworked, etc.

As to the quality of the research design it is fair enough, it's a quasi-random sample. My biggest problem is that while he finds some difference in sentence length he doesn't demonstrate that this difference is significant. To my mind a 1% difference in sentence length is so minor as to not merit comment.

Posted by: Daniel | Aug 11, 2016 7:21:55 PM

An additional issue to be considered is that when more than one defendant is indigent, the public defender's office usually gets the first listed defendant in the indictment. In most districts, the first-listed defendant is often the one considered by the prosecutor to be the most culpable, therefore, the premise that "decision of who is assigned to the public defender organization can be treated as close to a randomized experiment," is flawed, since it is more likely the public defender is getting the person considered more/most culpable.

Posted by: APD | Aug 11, 2016 7:29:51 PM

APD: Yes!!

Actually, I would say the high figure of 13.8 months longer sentence is evidence of the superior work of the PDs, who are almost always starting with a defendant positioned by the Government and the charges as deserving of a significantly higher term than many co-defendants.

Nobody familiar with how the system works would think this study design was smart.

Posted by: USPO | Aug 11, 2016 8:07:38 PM

Daniel -- I think APD's theory makes more sense. Prosecutors typically don't know who a defendant's lawyer is going to be before they charge them, so (for that and other reasons) the idea that charging decisions are based on whether a defendant has a FD or CJA attorney seems pretty far fetched to me.

Posted by: Jay | Aug 11, 2016 8:15:39 PM


I agree that APD explanation makes more sense than mine. I didn't realize that, "the public defender's office usually gets the first listed defendant in the indictment. In most districts, the first-listed defendant is often the one considered by the prosecutor to be the most culpable," as APD wrote.

If that is an accurate depiction of reality it would explain why defendants assigned to PDs receive graver charges that those assigned to CAs and since we know that it is the charges that determine the difference in sentence length we have the solution to the puzzle.

Note that this is a well-known drawback to quasi-random sampling which is that the sample may be biased in some way that the research does not anticipate. Usually this problem is avoided by some type of literature review or researcher triangulation.

Posted by: Daniel | Aug 11, 2016 8:56:26 PM

It could be that the court regularly appoints the FPD to represent the first defendant on a long list of defendants in a multi-defendant indictment. Usually, the defendants are listed in order of culpability. The ring leader is listed first. For that reason, he got the results that he did.

Posted by: T_J | Aug 12, 2016 12:42:32 AM

I think there are many reasons why the attorney of choice in a federal prosecution, especially for the indigent, would be the federal defender. First, it's true that the federal defender will almost always get the first name on a multi-defendant indictment. The case goes to the grand jury, and if an Indictment is issued (I laugh at the word "if" but...), the list of names is set before any defense attorneys know it exists. It is also true that the names are generally in order of culpability. Second, federal defenders only do federal criminal work. Many private attorneys seldom see federal criminal cases. Therefore, federal defenders are far more well versed in federal criminal law. Third, the federal defenders work day in and day out with the same prosecutors and the same judges. They have a very good inkling of how to work with them, who's the toughest, who is reasonable (or not), how to make deals, etc. So in most cases, you're going to get the best result possible with a federal defender.

Yes, I work for a federal defender. 22+ years. By and large, the attorneys are very good at their job.

Posted by: Paralegal | Aug 12, 2016 9:20:19 AM

in the district where I am a federal defender, the practice of assigning the FD the top (i.e., most culpable) defendant in the multi-defendant indictment is indeed common.

Posted by: FPD | Aug 12, 2016 2:40:10 PM

Just to add another data point - in our district, in multi defendant cases, our office frequently asks the Government which of the defendants is facing the most serious time and take that one, regardless of where he or she is listed in the indictment.

Posted by: Another AFPD | Aug 12, 2016 3:24:35 PM

In our county the appointed court attorney's work for the system and usually friends with the DA prosecutor and lawyer/judge. They are usually older and semi-retired needing the work. Their concern is not of justice or truth. They do not investigate to find the truth because investigators are not funded for the defendant by the people like the DA prosecutor is. The appointed public defender makes money for the court or they lose work and are not appointed in that court again. The problem of so many innocent prisoners is that the prosecutor has very deep pockets funded by the people and are never accountable.

Posted by: LC in Texas | Aug 14, 2016 10:50:53 AM

In my state, all appointed state court cases go to the local public defender who then farms out the conflicts, the first conflict typically going to a neighboring office followed by a contracted private attorney if there are multiple conflicts. The local public defender is supposed to keep the most serious case. While the study took its sample from the federal system, I would expect to see a similar result comparing attorneys who work for the public defender system full-time and those private attorneys contracted to handle a specific conflict case, just due to the relative culpability of the clients.

Posted by: tmm | Aug 15, 2016 10:55:14 AM

Post a comment

In the body of your email, please indicate if you are a professor, student, prosecutor, defense attorney, etc. so I can gain a sense of who is reading my blog. Thank you, DAB