An Explicit Test of Plea Bargaining in the “Shadow of the Trial”

According to the Shadow of the Trial Theory, a defendant will plead guilty if the sentence they are offered is less than or equal to the expected value of their sentence at trial. This theory explains the variation in plea discounts provided to those who plead guilty. While risk-neutral defendants are unaffected by uncertainty and thus indifferent to their choice between plea and trial, risk-averse defendants prefer the certainty of bargained pleas. This study developed a simple econometric regression model of the theory, where plea value is a function of the probability of conviction and the sentence at trial. The study then demonstrated the model’s strength through an online survey of 1,585 participating prosecutors, judges, and defense attorneys. The participants reviewed a set of thirty-one fact files from a hypothetical robbery case and provided their estimates for the probability of conviction at trial, the outcome and average sentence at trial, and the plea deal for which they would advocate. Based on the participants’ responses, the researchers adjusted the regression model to include dummy variables for state, county size, caseload, and participant demographic information. The researchers found that, while judges tend to offer fixed plea discounts and thus do not act in a manner consistent with the Shadow of the Trial Theory, defense attorneys and prosecutors behave in a manner largely consistent with the model.

“The shadow model, in contrast, is a formal theoretical (i.e., mathematical) model that starts from the perspective of the defendant, but ultimately can be used to predict the behavior of other actors interacting with the defendant. The outcome of the model is a clear prediction about the maximum sentence the defendant would willingly accept as part of a plea bargain.” p. 727