One in five indigent murder defendants in Philadelphia are randomly assigned representation by the Defender Association of Philadelphia while the remainder received court-appointed private attorneys. The authors exploit this random assignment to measure how defense counsel affects murder case outcomes. The primary dataset consisted of a sample of 3,412 defendants charged with murder between 1994 and 2005. Compared to appointed counsel, public defenders reduce their client's murder conviction rate by 19%, lowered the probability that their clients would receive a life sentence by 62%, and reduced overall expected time served in prison by 24%. When the authors applied methods used in past studies of public defenders that did not have the benefit of random assignment, they obtained far more modest estimated impacts, which suggests defendant sorting is an important confounder affecting past research. To explore possible explanations for this large disparity in outcomes, the authors interviewed judges, Defender Association attorneys, and attorneys who took appointments. Interviewees identified a variety of institutional factors that decreased the likelihood that appointed counsel would prepare cases as well as the Defender Association attorneys. Although the authors research is limited by the fact that it is focused on a single jurisdiction, the vast difference in outcomes for defendants assigned different counsel types raises important questions about the adequacy and fairness of the criminal justice system and additional research on the effect of counsel from other jurisdictions is required.