In this article, Judge Richard A. Posner presents the first comprehensive economic analysis of the law of evidence. The article is presented in three parts. First, Judge Posner proposes and describes two possible economic models, both a search and a cost-minimization approach, to describe how evidence is obtained presented, and evaluated In both, he incorporates Bayes' theorem to examine rational decisionmaking. Second he examines the evidence-gathering process, comparing and contrasting, in economic terms, the "inquisitorial" and "adversarial" systems of justice. The inquisitorial system, at first glance, appears to be more economically efficient. This, though, may be illusory, a result of the adversarial system's greater public visibility and widespread acceptance of plea bargaining. Finally, the article addresses burden of proof issues, plus specific provisions of the Federal Rules of Evidence: harmless error, limiting instructions, relevance, character evidence, hearsay, expert witnesses, and various privileges and exclusionary rules, lie concludes that American evidence law, rather than simply sacrificing efficiency in order to protect noneconomic values, is actually quite efficient and possibly superior to its Continental, inquisitorial counterparts; but a number of reforms are suggested.
The law of evidence is the body of roles that determines what, and how, information may be provided to a legal tribunal that must resolve a factual dispute. The importance of the accurate resolution of such disputes to an economically efficient system of law has been discussed at length,(1) but the economic literature dealing with the rules themselves is scanty in relation to the scope and importance of evidence law.(2) This article is the first comprehensive (though it is neither exhaustive nor definitive) economic analysis of that law. It is in three parts. The first part proposes and elaborates an economic model (actually two models, a search model and a cost-minimization model) of evidence. The second part examines the basic structure and structural rules of the evidence-gathering process; it includes an economic comparison between the "inquisitorial" and "adversarial" systems of justice and an analysis of issues relating to burden of proof. The third part is an economic appraisal of salient provisions of the Federal Rules of Evidence, the most influential American codification of such rules; it also takes up some issues of evidentiary privilege and exclusion that the rules do not deal with explicitly.
Like many other economic studies of the legal system, mine concludes that the institutional and doctrinal structure of the American law of evidence has a subtle, though intuitive, implicit, and incomplete economic logic. This conclusion will startle. Most evidence professors, and even a few judges,(3) would, if asked, say that of course the American system of finding facts at trial is inefficient, ludicrously so, and redeemed if at all by the noneconomic values that the system protects. But that assessment is founded on incomplete analysis and on misleading anecdotage that is itself an artifact of a worthwhile feature of the American system--the high degree of public scrutiny that it invites and enables.(4) Neither cheap nor highly accurate, our adversarial system is radically imperfect from the Utopian standpoint so often, though mistakenly, used to evaluate social institutions. Yet even from a perspective concerned only with economic efficiency in the sense of wealth maximization or cost minimization, it may not be inferior to the feasible alternatives, including the Continental inquisitorial system much touted in some quarters of the American legal academy.
I both emphasize the close connection between roles of evidence and the use of the jury and make repeated reference to Bayes' theorem. So let me make clear at the outset that I do not propose that juries or, for that matter, judges be instructed in the elements of Bayesian theory or mathematical probability more generally--or any other theory of probability or evidence. Such efforts at formalizing the tacit, intuitive inferential procedures now employed to resolve factual disputes in trials would cause endless confusion. Not only do few members of the general population have even a rudimentary education in mathematical probability theory or in the theory of rational choice that underlies both mathematical probability theory and economic theory; but legal education itself (alas) "produces no improvement in the ability to apply the statistical and methodological rules of the probabilistic sciences to either scientific studies or everyday-life events."(5) The value of Bayes' theorem in the law of evidence is heuristic. The most influential model of rational decisionmaking under conditions of ineradicable uncertainty (conditions which require that decision be based on subjective probabilities), it can be of great help, as we shall see, in evaluating the rationality of rules of evidence.
Although economic theory provides the framework of analysis in this article, I draw heavily on an empirical literature on trials and evidence that is largely psychological in orientation, as well as on Bayes' theorem, other aspects of decision theory, and statistical inference. The approach is therefore eclectic rather than narrowly economic, although it slights epistemological and other philosophical perspectives on the trial process, which seem to me to have only a very limited utility. The article's explanatory findings and reformist suggestions are summarized in the Conclusion--a glance at which will show that, despite the still-widespread belief that economic analysis of law has an inherent tendency toward politically conservative reforms, this is clearly not true with regard to the law of evidence.
THE ECONOMIC APPROACH TO EVIDENCE
How to Proceed
There are many possible ways of coming to grips with the economic issues that the law of evidence presents. The simplest would be to take up the various rules piecemeal and examine their economizing properties. Another way would be to deduce the optimal system of dispute resolution from economic theory and compare it with the actual systems in use in this and other countries. A third would be to begin with the epistemological and psychological literatures dealing with rational inquiry.(6) A fourth would be to build from what is now an extensive empirical literature on the actual operation of the various methods (especially the jury, the focus of this literature) for determining facts at trial.(7) A fifth would be to see how the private sector resolves disputes and to use its methods as a model, since private dispute resolvers have stronger incentives than public ones to maximize the net benefits of the dispute-resolution process. A sixth approach would be to examine all the possible goals of the law of evidence and try to establish the weight that economic goals should be given.
None of these approaches can be disregarded in a responsible economic analysis of evidence, and I shall touch on all of them at some point in this article. It seems best to begin, however, at an abstract level (the second approach sketched above), by asking: If we were writing on a clean slate and trying to design a system for the resolution of factual disputes in litigation that would be economically efficient in the broadest sense, how would we frame our inquiry? I propose two ways, which are equivalent.
The Search Model
The first way is to model factfinding as a problem in search, analogous to searching for a consumer durable,(8) with the correct answer to the question of (say) "Did X shoot Y?" corresponding to the utility-maximizing choice between two brands of dishwasher.(9) It is only an analogy. Not only are the incentives of the participants not the same in the two search processes, but external effects are also more likely in the case of the evidence search. I shall try to account for the differences later.
The search process, which in the litigation setting is the process of obtaining, -sifting, marshaling, presenting, and (for the trier of fact) weighing evidence, confers benefits and incurs costs.(10) Benefits are a positive function of the probability (p) that if the evidence is considered by the trier of fact the case will be decided correctly, and of the stakes (S) in the case. To keep things simple, I shall assume that the benefits are simply the product of the two terms, hence pS, where p is a positive function of the amount of evidence (x), so that the full expression for the benefits of the search is p(x)S. With enough evidence, p might equal 1, meaning that a trial would be certain to produce the correct outcome. The costs of the trial (c) are also a positive function of the amount of evidence (x).
Some of these assumptions will have to be relaxed later, but for now, the net benefits (B(x)) of what I am calling the "evidence search" in a case are given by
(1) B(x) = p(x)S - c(x)
and thus the optimum amount of search--the amount that maximizes net benefits--satisfies
(2) [p.sub.x]S = [c.sub.x]
where the subscripts denote derivatives. In words, the search should be carried to the point at which marginal cost and marginal benefit are equated. The amount of evidence at the optimum point will be greater the higher the stakes in the case, the lower the cost of obtaining evidence, and the greater the effect of evidence in increasing the likelihood of an accurate outcome.
For this optimum to exist, it is enough if p(x) is increasing at a decreasing rate ([P.sub.xx] [is less than] 0) and that [c.sub.x] is nondecreasing ([c.sub.xx] [is greater than or equal to] 0).(11) These conditions are plausible. Begin with the benefits side: As more and more evidence is obtained, the effect of additional evidence on the outcome of the case will tend to decrease, especially if the searcher begins the search with the most probative evidence...