The negotiated guilty plea: Vacancies as an alternative to the caseload pressure explanation

Plea bargaining is often explained as a product of caseload pressure in the trial courts. Adherents of this position argue that plea bargaining expedites case flow and moves cases more speedily to sentencing. Yet there is substantial empirical evidence that the proportion of its population that a society incarcerates at any given moment has tended historically to remain relatively constant—suggesting that, even with plea bargaining, crowded courts do not necessarily produce more convicted and sentenced offenders. More importantly, there is also substantial empirical evidence that changes in caseload have little effect on the prevalence of plea bargaining. Thus, the claim that crowded courts induce plea bargaining as part of an effort by prosecutors, judges and attorneys to move cases more rapidly has been called increasingly into question.

The purpose of this paper is to advance an alternate explanation that contemporary plea bargaining, far from responding to case backlog, constitutes a flexible systemic response to fluctuations in the vacancies available in the prisons. The argument is formalized for subsequent empirical testing. The paper begins by developing the argument that prison vacancy, instead of caseload pressure, induces plea bargaining. The paper then formalizes the argument by means of an innovative application of a discrete time Markov process model. Drawing on the distinctive properties of the Markov chain, it is demonstrated how the model may be used to test the competing vacancy and case flow hypotheses, and to explore the incidence and dynamics of plea bargaining, speed of case flow, conviction and sentencing effects of bargaining, and the systemic potential for racial discrimination that bargains generate. Implications of this vacancy model for differential sentencing, racial discrimination, and possible reform of the criminal justice system are also delineated.

In developing the ideas presented in this paper, the author has benefitted from valuable comments and advice on the part of Duncan Luce and David Karen. Ronald Breiger, Harrison White and John Padgett also read and commented on an earlier version of this manuscript. Their contributions are appreciatively acknowledged.