If John Edwards Were To Represent Himself, Would He Have a Fool for a Client?

Questioning the accuracy of an old legal saw.

John Edwards

John Edwards’ attorney Gregory Craig almost struck a plea bargain with prosecutors over charges of campaign-finance improprieties, but the government insisted on some jail time for the former senator. It now looks like the case could go to trial. Is there any reason Edwards, a successful trial attorney, shouldn’t represent himself in court?

There are reasons, but they might not be good ones. As with so much in the world of lawyering, there are arguments for and against attorneys representing themselves, but little data. Edwards knows his case better than anyone, so he might be particularly good at constructing his defense and responding quickly to allegations made in the courtroom. It’s also possible that an impassioned Edwards could sway the jury. Plus, he’d save a lot of money. But Edwards was a plaintiff’s lawyer in civil trials and might not be sufficiently acquainted with criminal procedure. There are also relationship considerations. Part of what you get when you hire a top-flight criminal attorney is familiarity with the prosecutors. Craig has probably dealt with the government’s team in the past and may have a sense of how far they’ll go in settlement negotiations. Finally, and perhaps most importantly, Craig can provide a reality check—defendants can become so convinced of their justifications that they can’t imagine how a jury could possibly disagree with them.

Whether the defendant is a trained lawyer or not, most attorneys have long accepted the conventional wisdom that representing oneself in court, known as pro se representation, is a bad idea. There’s an old saying that a person who represents himself in court has a fool for a client. The Supreme Court has even gotten into the act, quoting a law professor’s statement that “a pro se defense is usually a bad defense.”

A 2007 study, the first of its kind, seriously challenged these aphorisms. Professor Erica Hashimoto of the University of Georgia Law School found that, on the whole, pro se defendants actually achieve better results than their professionally represented peers. About 50 percent of do-it-yourselfers in state courts escape conviction, compared with 25 percent of represented defendants. (Hashimoto got her data from court docketing databases, which don’t list the defendants’ professions, so it’s not known what number of them, if any, were trained as attorneys.)

Of course, these numbers come with a variety of caveats. There were only 234 pro se defendants in the study, which is a small sample. There’s also a self-selection bias. Defendants who reject their attorneys are likely more convinced of their innocence than the average represented client. The difference in plea-bargain rates—44 percent for pro se versus 71 percent for represented defendants—suggests as much. In addition, these statistics mainly compare self-representation with a public defender or court-appointed counsel, not a prominent Washington litigator.

Self-representations can be a major headache for judges, especially when a pro se defendant decides to take the stand. Most judges dispense with the traditional Q&A format and require narrative testimony, but this robs opposing counsel of the opportunity to object before information is disclosed to the jury. Alternatively, some judges make the defendant bring in a proxy to ask the questions. There have even been cases in which the gallery was treated to the absurd spectacle of the pro se defendant both asking and answering the questions.

It’s also confusing for the jury to distinguish between when the defendant is presenting facts under oath and when he is making arguments as a lawyer. Judges typically keep the pro se defendant on a short leash during opening and closing arguments to prevent him from making factual statements without swearing to tell the truth first.

Got a question about today’s news? Ask the Explainer.

Explainer thanks Erica J. Hashimoto of the University of Georgia Law School.