The William B. Lockhart Lecture The following lecture was delivered in April of 1986 at the University of Minnesota Law School. The Lockhart Lecture Series is in honor of Dean William B. Lockhart Judicial, Legislative, and Administrative Lawmaking: A Proposed Research Service for the Supreme Court* Kenneth Culp Davis** The American legal system is distinctive in the world in the extent of its reliance on judicial lawmaking. We have far more judicial lawmaking than any other country in today's world and far more than any other country in world history. Yet no planning group has ever deliberated and decided that judges should make much of our law, including fundamental policies. Even though the Constitution explicitly puts the legislative power in Congress, judicial legislation is so deeply established that the legal profession takes it for granted, as though nature provided it. Although criticism of judicial lawmaking is plentiful, its focus is largely on activism versus restraint, and my concern in this lecture is not with substantive law but is with lawmaking, that is, with the process by which law is made. I believe that both legislative lawmaking and administrative lawmaking are superior to judicial lawmaking in three main ways: (1) The product is better in clarity, reliability, and freedom from conflict; (2) the legislative process and the administrative process are more democratic than the judicial process; and (3) the fac* ©1986, Kenneth Culp Davis. Professor of Law, University of San Diego; John P. Wilson Professor Emeritus, The University of Chicago. For valuable suggestions, I am indebted to Professor Carl Auerbach, Judge Stephen Breyer, Judge Henry Friendly, Professor Walter Gellhorn, Dean Sheldon Krantz, Professor William Lockhart, and Judge Carl McGowan. ** MINNESOTA LAW REVIEW [Vol. 71:1 tual base for legislation and for administrative rules is normally much stronger than the factual base for judge-made law. The third item-the factual base-is what matters most, for it is often what makes the difference between beneficial law and detrimental law. The factual base for lawmaking is the subject of a new idea I shall advance near the end of this lecture. Judge-made law, in comparison with statutes and rules, is generally inferior in clarity and reliability; indeed, the degree of inferiority may be greater than is commonly recognized. Supreme Court law may now be more conflicting than ever before; during each recent decade the conflicts seem to be greater than in the preceding decade. Although Supreme Court law changes significantly in every term of Court, express overrulings are very few-only 146 in 154 years, an average of about one a year. The changes come almost entirely through departures from precedents without overruling, and the consequence is conflicting decisions, which may now be more numerous than ever before. Stare decisis now seems to be the weakest it has ever been, and it is steadily growing weaker. Huge bodies of case law have about zero authoritative effect because other case law cancels them out. An extreme example is a 1985 en banc opinion of the Third Circuit: "There are literally hundreds of cases in the Courts of Appeals in which various aspects of this issue have arisen and no clearly discernible thread has evolved." Davidson v. O'Lone, 752 F.2d 817, 823 (3d Cir. 1984), aff'd sub nom. Davidson v. Cannon, 106 S. Ct. 668 (1986). The court's careful remark has deep meaning. The hundreds of decisions created confusion, not clarification; the product of the hundreds of decisions was not law but absence of law. The example is extreme; far more significant are the numerous examples of the same phenomenon in lesser degree. In contrast is any law that is governed by detailed statutes or detailed regulations. For instance, the main part of federal tax law is embodied in 2,000 pages of statutes and 9,000 pages of regulations. Despite its great magnitude and extreme complexity, tax law was litigated in only 600 appellate cases last year. The contrast with the relatively small subject of damages liability of employees of federal, state, and local governments for constitutional violations-what the Supreme Court calls constitutional torts-is striking; the law of the subject is in twenty or thirty thousand appellate opinions that are extraordinarily con- 1986] 196VAVAKING flicting. Appellate decisions last year were 7,300-more than twelve times the number in the huge federal tax field. In the large perspective of all federal law, the federal tax statutes and regulations are a magnificent lawmaking accomplishment. Probably no large body of law created by courts comes up to the same quality. A good way to measure the degree of uncertainty of judgemade law is by the proportion of trial court decisions that are appealed on questions of law. Such appeals are few when the law is clear and are many when it is unclear. During the period from 1960 to 1984, the number of district court decisions multiplied by two and two-thirds, while the number in courts of appeals multiplied by eight. During the 24-year period, the increase in appellate courts was thus three times the increase in district courts. The reasons for the faster increase in appellate cases are complex, but the main one was probably the increasing uncertainty of judge-made law. The basic movement seems strongly away from, not toward, clarification. Examples might be chosen from almost any area. The preface to volume 4 of my Administrative Law Treatise (2d ed. 1983) asserts: "The potential for improving the Supreme Court law discussed in this volume lies much less in ideas about administrative law than in the Court's practice concerning stare decisis." The 140-page discussion of the law of standing has many horrible examples. The worst might be the Supreme Court's opinion in Warth v. Seldin, 422 U.S. 490 (1975), which laid down ten propositions of law, every one of which is contrary to earlier or later Supreme Court law. In each of ten numbered paragraphs of § 24:34 I quoted one proposition from the Warth opinion and then set forth the contrary law. The first example quoted one sentence from the 1975 opinion and then set forth Supreme Court decisions to the opposite effect in 1973, 1976, 1978, and 1980. The total number of Supreme Court opinions on the law of standing, including concurring and dissenting opinions, exceeds 500. A good system of judicial lawmaking would carefully build a body of precedents that could be the basis for predicting results on any new problem. The Court's decisions are not at all of that character. They are often more harmful to prediction than helpful, for some are extreme in one direction and others are extreme in the opposite direction. Almost every decision may be matched by an opposite decision. The lawmaking standards the Supreme Court applies to it- MINNESOTA LAW REVIEW [Vol. 71:1 self are far more lenient than the lawmaking standards it applies to a rulemaking agency. The amusing fact is that the Court in one case imposed a lawmaking requirement on an agency and then in the same case violated that same lawmaking requirement. The Court held that "'an agency changing its course must supply a reasoned analysis . . . .'" Motor Vehicle Mfrs. Ass'n v. State Farm Mutual, 463 U.S. 29, 57 (1983) (quoting Greater Boston Television Corp. v. FCC, 444 F.2d 841, 852 (D.C. Cir. 1970), cert. denied, 403 U.S. 923 (1971)). That holding seems to me thoroughly sound, except that, in so holding, the Court changed its own course without supplying a reasoned analysis, for it had previously declared emphatically that the courts may not add to the procedural requirements of the Administrative Procedure Act. Vermont Yankee N.P. Co. v. NRDC, 435 U.S. 519, 546 (1978). The APA did not require a reasoned analysis, but the Court required it. The Court thus clearly departed from its Vermont Yankee requirement without a reasoned analysis, doing precisely what the Court in the same case held the agency must not do. An additional weakness in recent Supreme Court lawmaking is the sharply growing proportion of concurring and dissenting opinions, which vastly complicate the task of one who is trying to find out what the law is. Fifty years ago, about sixty percent of the opinions in United States Reports were opinions of the Court. Recently, only about twenty percent are opinions of the Court and about eighty percent are concurring or dissenting opinions. The unfortunate splintering has two harmful effects: It adds to the uncertainty of the law, and it transfers mental energy of the Justices from the central job of lawmaking to the unimportant rivalry among the Justices. Altogether, the law made by judges seems to me clearly inferior to statutes and administrative rules in clarity, reliability, and freedom from conflict. The two subjects of the democratic base for lawmaking and the factual base for lawmaking may best be discussed together because they overlap. I shall approach these two subjects by first describing four main patterns of lawmaking procedureideal procedure, congressional procedure, administrative procedure, and judicial procedure. Ideal procedure is hard to find, but easy to imagine. The lawmaking is solidly based on whatever scientific or professional understanding is relevant, and that is the main basis unless public opinion pulls toward something else. Ideal 1986] LAWMAKING procedure may start with studies by qualified specialists of the appropriate kinds, and the one task may be to determine what parts of those studies must be modified in deference to the democratic element. Legislative procedure in the American government has evolved over two centuries and might now be about at its alltime best. The First Congress probably legislated with hardly more facts than were already in the minds of Members of Congress, and with little more democratic arrangement than the representative system. Some of the legislation of the NinetyNinth Congress is comparable, but most is not; most is based on investigation or studies, some extensive, nearly all openly available to the public in tentative form, along with ample opportunity to make comments and apply pressures. Lobbyists play a vital role. The staffs who help Congress are of higher and higher quality, including many who are among the leaders in their fields. The legislative committee system at its best is a superb procedure for the development of understanding and for the reflection of democratic desires. Administrative rulemaking procedure has developed over the last half century, but the main development has been during the last twenty years. In 1970 I first called it one of the greatest inventions of modern government, and it has developed a great deal since then. Like legislative committee procedure, it is open and democratic, but it is also quicker and less expensive. With the help of its specialized staff, the agency prepares a proposed rule, which it publishes along with an invitation for written comments. Anyone at all may send in a written response, from a simple letter to a comprehensive study. The staff sifts the comments, and the agency heads may then have a sound foundation for preparing the final rule. Both the factual base and the democratic base usually are adequately developed. The Supreme Court is a major lawmaker, but it has no procedure designed for lawmaking. Its only procedure is designed for adjudication. The Justices look at the record from the lower court, read the briefs, listen to oral arguments, confer with each other, sometimes look up law in a library or have their clerks do so, and prepare an opinion. The only significant change since the pattern became fixed in 18th century England is the use of law clerks, who are always young lawyers, never specialists in nonlegal fields. The Justices never openly consult specialists, either about law or nonlaw, and neither the Justices MINNESOTA LAW REVIEW [Vol. 7/1:1 nor the law clerks are qualified to make studies or investigations of the kind that specialists in sciences or social sciences customarily make. The Court has no facilities for notice and comment procedure, no matter how much it would sometimes improve the lawmaking. The procedure is essentially what it would be if it had been deliberately planned to make it as undemocratic as possible, for the Justices avoid lobbyists, take no opinion polls, make no inquiry about public opinion, and use no procedure of notice and comments. The weak spots are the insufficient factual or scientific base and the lack of a democratic base when one is needed. Procedures used for lawmaking depend more on who is the lawmaker than on the needs of the particular lawmaking task. For the same lawmaking task, a legislator may consult lobbyists, sample public opinion, and call for a staff study of the relevant legislative facts; an administrator may have such a staff study and a notice and comment proceeding; and a judge listens to arguments but neither consults lobbyists nor samples public opinion nor calls for a staff study of legislative facts. If choices among lawmakers were planned on the basis of relative advantages and disadvantages of each, as they never have been, courts would make no law except when neither a factual base nor a democratic base is needed for lawmaking. The contrast between the way judges make policy and the way legislators make policy is a strong contrast. Judges do not invite influence and pressures; legislators do. Judges seldom make inquiries about public opinion and bend to it; legislators usually do. Judges never prepare proposed positions and invite public criticism; legislators do. Legislators do not protect themselves from informal influence by those who have interests at stake; judges do. Legislators do not insulate themselves from ex parte conferences with advocates; judges do. Legislators do not emphasize precedents created by their predecessors; judges do. Legislators have no problem about considering the way a policy may affect a nonparty; judges usually focus mainly on parties, even when a decision may vitally affect nonparties. The contrast between the way judges make policy and the way administrators make policy may be even stronger. Judges do not have specialized staffs; administrators usually do. Judges do not assign policy questions for special studies or investigations; administrators often do. Judges never announce 19861 LAWMAKING tentative or proposed policies, publish them, and invite written comments; administrators often do. No one has planned the present system under which the procedure of appellate courts, designed for adjudication of questions of law, is used for a large portion of all the lawmaking that is done in the whole society. No one would plan such a system. What has happened is that appellate procedure was originally planned for finding and applying law; and then appellate courts, without changing their procedure, have gradually increased their lawmaking activity. Major policy, no matter who makes it, usually needs a democratic base and often needs a factual base. Unlike legislators and unlike administrative rulemakers, courts are often inadequately informed about democratic desires; they protect themselves from lobbyists, they take no polls, and they never use procedure of notice and comments. They always have the needed adjudicative facts, that is, the facts about the immediate parties-who did what, where, when, how, and with what motive or intent. But courts often have inadequate legislative facts, that is, the facts that bear on the court's choices about law and policy. No judge can think about law, policy, or discretion without using extrarecord facts. Even the meaning of the word "the" has to come from outside the record. What we call wisdom is largely factual. In its higher reaches, understanding of legislative facts is the essence of making law or policy, whether the actor is a judge, a legislator, or an administrator. When Lord Mansfield was creating English commercial law during the 18th century, he did not limit himself to what the parties presented. He freely went outside the record for his legislative facts. He needed to know the practices, customs, and attitudes of merchants, and to do so he associated with some of them in a local tavern; then he would go back to his chambers to make law on the basis of what he had learned. He acknowledged in a formal opinion: "I... endeavored to get what assistance I could by conversing with some gentlemen of experience in adjustments." Lewis v. Rucker, 97 Eng. Rep. 769, 772 (K.B. 1761). As early as 1908, Justice Holmes said: "As the judge is bound to declare the law he must know or discover the facts that establish the law." Prentis v. Atlantic CL. Co., 211 U.S. 210, 227 (1908). Justice Brandeis, while he was a Boston lawyer, submitted a factual brief to the Supreme Court, and such briefs are now called Brandeis briefs. Such briefs, devoted to legisla- MINNESOTA LAW REVIEW [Vol. 71:1 tive facts, are filed in many cases. Unfortunately, they are needed in many cases in which they are not filed. Those who are affected should have opportunity to try to influence any lawmaking that affects them, no matter who is the lawmaker. Congressional committees open their doors to facts, ideas, and pressures. So do rulemaking agencies. A common pattern of both the committees and the agencies is to work out a proposed policy, to publish it, and to receive comments. That fundamental procedure is tending to become the essence of democratic government on complex issues, except when courts are lawmakers. But the courts are now making much of our law, without a democratic base and, often, without a factual base. I believe that the procedure of appellate courts is exceedingly good procedure when it is used for the purpose for which it was designed, that is, finding and applying established law, but it is exceedingly poor procedure when used for a purpose for which it was not designed, that is, legislating new law. I believe that, for lawmaking, procedure designed for lawmaking is better than procedure not designed for lawmaking. Congressional procedure and administrative procedure have been designed for lawmaking. Procedure of appellate courts, including procedure of the Supreme Court, is not designed for lawmaking. Of course, much judicial lawmaking does not require either a factual base or a democratic base. The Supreme Court is often at its best on complex thinking problems, on philosophical or ethical or moral issues, on analysis or reasoning, and on issues of interpretation. But the Court may often be at its worst on policy issues that are dependent upon understanding or instincts about legislative facts. Indeed, my impression is that, typically, the Court is basically baffled in trying to deal with legislative facts. The Court's treatment of legislative facts may be an especially weak area in its whole performance, and it seems to be fully aware of that weakness. We may sympathize with the predicament of the Justices, but whatever our sympathy, it won't cure the inadequacy of the factual base and of the democratic base for lawmaking. When the Supreme Court in deliberating about a case realizes that it needs legislative facts it does not have, what should it do? Either a legislative committee or a rulemaking agency would assign the problem to a qualified staff for an appropriate study or investigation; but the Supreme Court has no such staff 1986] LAWMAKING available to it. The committee's inquiry would be open, and any interested person would have opportunity to influence it. An agency might have a staff study and also a notice and comment proceeding. The Supreme Court has no facilities either for a staff study or for a notice and comment procedure. What should the Supreme Court do? And what does it do? What it does not do is easier to state. It never assigns the problem to a qualified staff for a study or investigation; it never gets the parties' response to such a study or investigation; and it never uses notice and comment procedure. What the Supreme Court affirmatively does is quite various. It has no system. It tries one method after another, almost always with unsatisfactory results. I shall now list, with citations, seven solutions the Supreme Court has tried: 1) The Supreme Court sends the case back to the trial court for taking evidence on questions of legislative fact. Borden's Farm Products Co. v. Baldwin, 293 U.S. 194, 210 (1934). That idea has some merit, but it has failed to take hold, perhaps because the procedure is especially awkward. 2) The Supreme Court simply asserts an emphatic view of the legislative facts, with nothing to support its view. That is what the Court did when it called the agricultural dislocation of the 1930s "a widespread similarity of local conditions," contrary to the view of all economists of the time. United States v. Butler, 297 U.S. 1, 75 (1936). 3) Even worse is judicial notice of what the Supreme Court calls "common experience," without mentioning any facts, and directly in the face of a dissenter's abundant and convincing evidence to the contrary. Jay Burns Baking Co. v. Bryan, 264 U.S. 504, 517 (1924). 4) Another unfortunate way is to examine a published source and to find what is not there. The Supreme Court looked at a report of a 19-member investigating commission, which had divided 17-2. The majority of the commission said quite clearly that "empirical research . .. has found no evidence.., that exposure to explicit sexual materials plays a significant role in the causation of ... criminal behavior .... " Paris Adult Theatre I v. Slaton, 413 U.S. 49, 108 n.26 (1973). The Court completely ignored what the majority said, and relied only on what the minority said. On top of that, the minority did not at all say what the Supreme Court said it said. 5) On the question whether government employees should be made liable for damages in some circumstances, even though MINNESOTA LAW REVIEW [Vol. 71:1 they had previously enjoyed absolute immunity under Supreme Court decisions of 1845 and 1896, a main question was whether the change would cause "disruption of government." Butz v. Economou, 438 U.S. 478 (1978). Four dissenters plausibly asserted it would. But the Supreme Court imposed the new liability, while maintaining complete silence on the disruption question. The Court may have had no means of answering such questions as these: To what extent are offers of government positions rejected because of tort liability of officers? To what extent do officers resign on account of such liability? To what extent does Congress by private laws indemnify officers? How many judgments do officers pay, and in what amounts? Is insurance available? To what extent are officers insured? For all that appears, the Court had no information on such questions. Neither did the four dissenters. 6) One way the Court resolves questions about legislative facts without getting the facts is by imposing the burden of proof on one of the parties. In a challenge of state legislation, should the burden be on the challenger or on the defender? In 1973, the Court put the burden on the challenger; in 1977, it put the burden on the defender. In the 1977 case, it rejected the view of the 1973 decision without mentioning it. ParisAdult Theatre I v. Slaton, 413 U.S. 49 (1973); Carey v. Population Services Intl, 431 U.S. 678 (1977). 7) In two major cases, the Supreme Court diligently did extrarecord research and made decisions based on its findings of legislative facts, but the Court in both cases failed to give the parties a predecision chance to challenge its extrarecord facts. Ballew v. Georgia, 435 U.S. 223 (1978); Roe v. Wade, 410 U.S. 113 (1973). Judge Friendly pointed out in an article that the Supreme Court decision in Roe v. Wade rested heavily on extrarecord abortion facts that the parties had had no chance to challenge, and he accurately asserted: "If an administrative agency, even in a rulemaking proceeding, had used similar materials without having given the parties a fair opportunity to criticize or controvert them at the hearing stage, reversal would have come swiftly and inexorably." Friendly, The Courts and Social Policy, 33 U. Miami L. Rev. 21, 37 (1978). The seven ways, as a bundle, show that the Supreme Court has been earnestly striving to find a solution, but that nothing it can do is as satisfactory as what either a legislative committee or an agency would normally do, for the Supreme Court has no properly qualified staff to make an appropriate study or investi- 1986] LAWMAKING gation of legislative facts, and it has no facilities for a notice and comment proceeding. The conclusion is overwhelming that the Supreme Court lacks the essential institutional arrangement for developing the legislative facts on which some of its lawmaking should rest. When legislative facts are needed for a sound decision, a trial court can do better than an appellate court, because it is free to take evidence on questions of legislative facts. Some trial courts do so, with desirable results. Even so, I have to point out that the normal evidence-taking process may be a total misfit for legislative facts. An example is a case in which a court needed to answer the question whether drivers under forty are safer than drivers over forty. The trial court applied Rule 803(8) of the Federal Rules of Evidence and excluded as hearsay "assorted publications by eminent authorities on effects of aging upon functional abilities." The court also excluded what it called "many scientific studies." Usery v. Tamiami Trail Tours, Inc., 531 F.2d 224, 243, n.50 (5th Cir. 1976). As the law was interpreted, the court was barred from considering precisely the kind of information that was most helpful and most needed. The Supreme Court's deficiency in developing needed legislative facts on which to base its lawmaking is so important and so pronounced that I shall present some additional examples. The Supreme Court invalidated a New York adoption statute on the basis of facts it simply assumed, even though the facts were susceptible of study; the result was invalidation of similar legislation throughout the country. Caban v. Mohammed, 441 U.S. 380 (1979). The statute required consent of mothers, but not of fathers, for adoption of an illegitimate child. In finding unconstitutional discrimination against fathers, the Supreme Court assumed that illegitimate children's mothers and fathers do not significantly differ from each other in their relationships with their children. The Court did not allow the losing party a pre-decision chance to challenge the assumed facts, even though three of the four dissenters asserted that the assumed facts were erroneous. I think the Court's system is procedurally unfair but, more importantly, widespread legislation should not be invalidated on the basis offactual assumptions when the subject is susceptible of factual study. In Barefoot v. Estelle, 463 U.S. 880 (1983), the main question was whether a death sentence could be based on testimony MINNESOTA LAW REVIEW [Vol. 71:1 of two psychiatrists about the criminal defendant's "future dangerousness," or whether the Court should rely on the statement of the American Psychiatric Association, in an amicus brief, that "[t]he unreliability of psychiatric predictions of longterm future dangerousness is by now an established fact within the profession." Id. at 920 (dissenting opinion, quoting Association's brief). The Association said that its best estimate was that two out of three predictions of long-term future violence made by psychiatrists are wrong. The three dissenting Justices asserted: "Neither the Court nor the State of Texas has cited a single reputable scientific source contradicting the unanimous conclusion of professionals in this field that psychiatric predictions of long-term future violence are wrong more often than they are right." Id. at 921. (Note the word "unanimous.") Even in the face of what the Association said, the Court upheld the jury finding based on testimony of the two psychiatrists; the result was execution of the individual. The Court came close to contradicting the professional view the Association presented to the Court: "We are not persuaded that such testimony is almost entirely unreliable and that the fact-finder and the adversary system will not be competent to uncover, recognize, and take due account of its shortcomings." Id. at 899. Either a legislator or an administrator who wants to locate the limits of psychiatrists' understanding would call some psychiatrists for questioning, including representatives of the Association. The Court could not do that because it had no procedural mechanism to do it. Even the important decision in Miranda v. Arizona, 384 U.S. 436 (1966), was based on loose impressions about police practices, not on reliable facts. The Supreme Court had no evidence of police practices; it had only police "manuals and texts." Justice Harlan for three dissenters specifically rejected the Court's factual assumptions. Justice White said for three dissenters that the Court "has not examined a single transcript of any police interrogation," and then he said, accurately, I think, that: "Judged by any of the standards for empirical investigation utilized in the social sciences the factual basis for the Court's premise is patently inadequate." Id. at 533. The Supreme Court continues its exclusionary rule without knowing even the main facts about the effects of the rule; in fact, the Court has specifically acknowledged that it has no facts as to how many felons are released by the police on the basis of illegality of searches or seizures. United States v. Leon, 1986] LAWMAKING 468 U.S. 897, 907-08 n.6 (1984). In footnote 6, the Supreme Court relied on law review articles which made estimates. Justice Blackmun in a concurring opinion complained: "Like all courts, we face institutional limitations on our ability to gather information about 'legislative facts.'" Id. at 927. Two of the three dissenters also acknowledged need for facts by asserting the extreme difficulty of ascertaining "whether the incidence of unlawful conduct by police is now lower than it was prior to Mapp." Id. at 942. The meaning of that statement is that, despite the importance of the law the Supreme Court is making, the Court not only lacks a factual base for its decisions but it has no information about the effect of its decisions on police practices. Because of the Court's lack of facts, Justice Blackmun said that "the scope of the exclusionary rule is subject to change in light of changing judicial understanding about the effects of the rule outside the confines of the courtroom." Id. at 928. The "judicial understanding" was thus avowedly based on inadequate facts. Yet any good research staff could probably produce the facts needed for sound lawmaking. Two dissenters protested against what they called "[a] doctrine that is explained as if it were an empirical proposition but for which there is only limited empirical support.... ." Id. at 943. One statement I have just quoted from Justice Blackmun is of major importance. It is precisely correct, and it states a true fundamental about the Supreme Court's lawmaking. It is worth repeating. He said: "Like all courts, we face institutional limitations on our ability to gather information about 'legislative facts.'" The meaning is that the institutional limitations prevent the Court from getting the legislative facts that are indispensable for making a sound decision. When the parties fail to supply needed legislative facts, the Court often has no procedural mechanism for developing those facts, and the Court's lawmaking may be seriously deficient for that reason. In the Leon case, the lawmaking was deficient, for the Court had no means of finding out whether the estimates in the law reviews were reliable. In holding that the exclusionary rule does not apply to deportation, the Supreme Court divided on questions that were essentially factual. INS v. Lopez-Mendoza, 468 U.S. 1032 (1984). One of the four dissenters asserted that "the majority exaggerates the costs associated with applying the exclusionary rule in this context." Id. at 1055. An appropriate research staff could probably have replaced the guesses with facts. MINNESOTA LAW REVIEW [Vol. 71:1 The Justices are often perturbed by their lack of legislative facts needed for lawmaking. An example that invites close examination is Furman v. Georgia, 408 U.S. 238 (1972), in which the Chief Justice affirmatively asserted that lack: "The case against capital punishment... rests primarily on factual claims, the truth of which cannot be tested by conventional judicial processes." Id. at 405. Note that the Chief Justice specifically says that factual claims cannot be tested by conventional judicial processes! The Court lacks capacity to get the legislative facts that are essential for a proper decision! The Chief Justice goes further. He asserts that five opinions of his colleagues "share a willingness to make sweeping factual assertions, unsupported by empirical data .... " That seems to me to be entirely true. But just consider how serious the statement is that five Justices make sweeping factual assertions that are unsupported, even though lives of those about to be executed are at stake. Then the Chief Justice acknowledges in broad terms the insufficiency of the Court's tools for doing its job: "Legislatures will have the opportunity to make a more penetrating study of these claims with the familiar and effective tools available to them as they are not to us." The Chief Justice thus makes one of the main points I am emphasizing: The Supreme Court lacks capacity to find legislative facts that a legislative body has capacity to find; when that is so, lawmaking by the Court is inferior to the law that Congress or a state legislature could make. A legislative body has what the Chief Justice calls "effective tools" for getting the facts, and the Supreme Court has no such tools. The Chief Justice was surely right in saying that legislatures have tools for making factual studies and courts do not. And Justice Blackmun was surely right in emphasizing the Supreme Court's "institutional limitations" on gathering legislative facts. Yet in both cases the Court made important law without the legislative facts it knew it needed. The Court yielded to the necessity of deciding the cases, for its established system allowed no other course. The obvious fact is that the Court often makes new law without legislative facts it knows it needs. Supreme Court cases in which legislative facts are inadequate are far more numerous than the sixteen specific cases I have discussed. Policymaking always reflects the legislative facts the lawmakers know or believe, and the quality of the decision reflects the accuracy or inaccuracy of the knowledge or beliefs. LAWMAKING In the first decade of the 20th century, a brief filed by the Attorney General of Kentucky with the United States Supreme Court asserted that brains of blacks were smaller than those of whites, that brains of mulattoes were smaller than those of blacks, and that intelligence varied with the size of the brains. The Court may have believed the brief, for it held constitutional a criminal statute forbidding a college to teach both white and black students in the same classes. Berea College v. Kentucky, 211 U.S. 45 (1908). The Supreme Court's lawmaking is sometimes based on understanding of the relevant legislative facts and sometimes not. When the Court lacks the needed information, it usually makes guesses. Much of our law is based on wrong assumptions about legislative facts. What the Court would have held in the Berea College case if it had had the true facts about brain sizes may never be known. Just as factual assumptions of eight decades ago now seem absurd, factual assumptions of today may seem absurd eight decades from now. Today's Constitution is better than yesterday's because the underpinning with legislative facts is better, and tomorrow's Constitution will be still better for the same reason. Furthermore, the Court, perhaps in the near future, will develop a system for its ready access to the best legislative facts anyone in the society can provide. What sort of system might be established to supply the Court with needed studies of legislative facts? Brandeis briefs should be encouraged; the Court now welcomes them, but not many are filed. Perhaps law offices should more often engage appropriate scientific or professional specialists to contribute to the preparation of briefs. That would help but at best would be only a partial solution. The Supreme Court might employ specialists in sciences and social sciences to assist the Justices in somewhat the way law clerks do. But the Court is not geared to employing and managing such a staff, and a small staff would probably lack the needed diversification. Although having such a staff might be preferable to not having it, a much better idea is creation of a research organization outside the Court to make studies at the Court's request. Such a research organization already exists, but it serves only Congress. Since 1856, congressional committees have employed professional staffs. Such staffs are in addition to those who serve each Senator and each Representative. In 1946, Congress created the Legislative Reference Service to make special studies upon request by committees. That system was so suc- MINNESOTA LAW REVIEW [Vol. 71:1 cessful that Congress strengthened and expanded it in 1970, changing the name to Congressional Research Service and providing that any single Member of Congress may request studies. The CRS is now a large and vital institution. It has 587 research and information specialists, supported by 273 clerical and administrative staff. Most congresional requests are merely for information, but many of them require in-depth policy analysis and research. The CRS budget for the current year is about thirty-nine million. The CRS Review reports some of the studies. It is published ten times a year and finds its way into many libraries. Anyone who examines the Review is likely to conclude quickly that the published studies are of exceedingly high quality. Of considerable importance is the determination by Congress that research for lawmaking should not be limited to what can be done by the staff of the Congressional Research Service. Congress provided for research by the CRS but it did not stop there. It realized that the outstanding specialists on particular problems often are in the employ of universities, private corporations, or nongovernmental research organizations. Accordingly, it authorized CRS to "procure the temporary or intermittent assistance of individual experts or consultants... and of persons learned in particular or specialized fields of knowledge" and also of "organizations of experts and consultants ... and of educational, research, and other organizations of persons learned in particular or specialized fields of knowledge." 2 U.S.C. § 166(h). The intent behind that provision is that, as far as feasible, lawmaking by Congress should be guided by whatever specialized understanding may be relevant and useful, no matter what the source. All the doors to legislative facts of all kinds-all the doors to specialized understanding-are opened. Now, in our present stage of development, we have three principal makers of federal law-Congress, the Supreme Court, and rulemaking agencies. All three need access to scientific and professional understanding that is necessary for proper lawmaking. Congress has such access in many ways-through association with lobbyists, through legislative committee hearings, through a special research organization with a budget of thirty-nine million, and through access to specialists and private research organizations. The agencies have such access through their specialized staffs and through their use of notice and comment procedure. The Supreme Court has no such ac- 1986] LAWMAK!7G cess, except in the few cases in which Brandeis briefs are volunteered. Its only professional helpers are legally trained clerks. A single Member of Congress may set the machinery in motion for special studies by the Congressional Research Service or by private specialists or private research organizations. But neither a single Justice nor the entire Supreme Court may obtain such research assistance. I propose that the Supreme Court after due deliberation should formally ask Congress to explore the potential for creating a research service to assist the Court. The sole purpose should be to increase the Court's freedom to obtain whatever research assistance it decides it needs. The Court should have the privilege of asking for research either on a problem about a pending case or about a narrow or broad area of law. All rules, practices, and customs of the Supreme Court and its staff concerning confidentiality would apply to the research staff. The Court's use of the research service would rarely delay decisions. The service might be used in only about a half dozen decisions in each term, although it might be quite important in some of those cases. Many questions would be quickly answered. Even substantial research could usually be fitted into the Court's normal time schedule, but not always. Some decisions might be made earlier with the research service than without it. Justice Blackmun's research in the library of the Mayo Clinic on medical facts about abortion might have been both more efficient and more reliable if he had had the help of research assistants with medical training. Now, in concluding, I shall requote three significant statements made by Justices in formal opinions: Justice White in Miranda,"the factual basis for the Court's premise is patently inadequate." Chief Justice Burger in Furman, "factual claims, the truth of which cannot be tested by conventional judicial processes." Justice Blackmun in Leon, "we face institutional limitations on our ability to gather information about 'legislative facts.'" ADDENDUM What I have learned after delivering the foregoing Lockhart Lecture in April of 1986 impels me to accentuate my belief that the Supreme Court should formally ask Congress to ad- MINNESOTA LAW REVIEW [Vol. 71:1 dress the whole problem of a research service for the Supreme Court. During a conference I had with the director and three top officers of the CRS, all of them expressed their view that the Court will benefit by using their service. The director is currently inquiring into the extent of the Court's need for such service, by systematically sending selected Supreme Court opinions to appropriate staff specialists for review. A Supreme Court law clerk sought CRS assistance in Bowen v. American Hospital Ass'n, 106 S. Ct. 2101, decided June 9, 1986, and the Court cited a CRS study in footnote 30, using the names of authors but not mentioning the CRS. The record shows no predecision chance for the losing party to respond to the study. A gradual drift toward informal consultation by law clerks with CRS personnel might naturally develop, but a statutory authorization is preferable because (1) what the Court does should be open, not clandestine, (2) each party should have a chance to respond to significant extrarecord information and ideas, (3) the Court's informal use of CRS may be deemed a departure from the statutory design of 2 U.S.C. § 166, and (4) an explicit statutory authorization will give a needed boost to extrarecord research and thereby strengthen the factual foundations of Supreme Court lawmaking.