The luck of the draw: Using random case assignment to investigate attorney ability

One of the most challenging problems in legal scholarship is the measurement of attorney ability. Measuring attorney ability presents inherent challenges because the nonrandom pairing of attorney and client in most cases makes it difficult, if not impossible, to distinguish between attorney ability and case selection. Las Vegas felony case data, provided by the Clark County Office of the Public Defender in Nevada, offer a unique opportunity to compare attorney performance. The office assigns its incoming felony cases randomly among its pool of attorneys, thereby creating a natural experiment free from selection bias. We find substantial heterogeneity in attorney performance that cannot be explained simply by differences in case characteristics, and this heterogeneity correlates with attorneys' individual observable characteristics. Attorneys with longer tenure in the office achieve better outcomes for the client. We find that a veteran public defender with ten years of experience reduces the average length of incarceration by 17 percent relative to a public defender in her first year. While we find no statistical difference based on law school attended or gender, we find evidence that the public defender's race correlates with sentence length, with Hispanic attorneys obtaining sentences that were up to 26 percent shorter on average than those obtained by black or white attorneys. We also find evidence suggesting that differences in sentencing may be driven partly by different plea bargaining behavior on the part of the public defenders.