Substantive criminal law defines the conduct that the state punishes. Or does it? If the answer is yes, it should be possible, by reading criminal codes (perhaps with a few case annotations thrown in), to tell what conduct will land you in prison. Most discussions of criminal law, whether in law reviews, law school classrooms, or the popular press, proceed on the premise that the answer is yes. (1) Law reform movements regularly seek to broaden or narrow the scope of some set of criminal liability rules, always on the assumption that by doing so they will broaden or narrow the range of behavior that is punished. Opponents of these movements operate on the same assumption -- that the law determines who goes to prison and who doesn't, that the distribution of criminal punishment tracks criminal law as it is defined by code books and case reports. Of course, participants in these debates understand that the law does not by itself determine who is and isn't punished. Some criminals evade detection, police and prosecutors frequently decline to arrest or charge, and juries sometimes refuse to convict. Still, if the literature on criminal law is an accurate gauge, all that is just a gloss on the basic picture, a modification but not a negation of the claim that criminal law drives criminal punishment.
But criminal law does not drive criminal punishment. It would be closer to the truth to say that criminal punishment drives criminal law. The definition of crimes and defenses plays a different and much smaller role in the allocation of criminal punishment than we usually suppose. In general, the role it plays is to empower prosecutors, who are the criminal justice system's real lawmakers. Anyone who reads criminal codes in search of a picture of what conduct leads to a prison term, or who reads sentencing rules in order to discover how severely different sorts of crimes are punished, will be seriously misled.
The reason is that American criminal law, federal and state, is very broad; it covers far more conduct than any jurisdiction could possibly punish. The federal code alone has thousands of criminal prohibitions covering an enormous range of behavior, from the heinous to the trivial. State codes are a little narrower, but not much. And federal and state codes alike are filled with overlapping crimes, such that a single criminal incident typically violates a half dozen or more prohibitions. Lax double jeopardy doctrine generally permits the government to charge all these violations rather than selecting among them. Since all change in criminal law seems to push in the same direction -- toward more liability -- this state of affairs is growing worse: legislatures regularly add to criminal codes, but rarely subtract from them. In a world like that, lists of crimes in statute books must bear only a slight relation to the conduct that leads to a stay in the local house of corrections.
Of course, criminal law's breadth is old news. It has long been a source of academic complaint; (2) indeed, it has long been the starting point for virtually all the scholarship in this field, which (with the important exception of sexual assault) consistently argues that existing criminal liability rules are too broad and ought to be narrowed. Yet the implications of this piece of old news are not well understood.
Consider two defining features of criminal law's large literature. First, it is relentlessly normative. Almost all writing about American criminal law argues that some set of criminal liability rules is morally wrong or socially destructive, and that a different (usually narrower) set of rules would be better. (3) Second, these normative arguments almost always presuppose that changing the liability rules would change the behavior the system punishes -- again, the assumption is that criminal law drives criminal punishment, not the other way around. (4)
Both features are at odds with the way American criminal law actually works. Normative legal argument makes sense on the assumption that lawmakers care about the merits, that the side with the better policy position has a better chance of getting its preferred rule adopted. But the legislators who vote on criminal statutes appear to be uninterested in normative arguments. (5) To take an obvious example: for the past generation, virtually everyone who has written about federal criminal law has bemoaned its expansion. (6) But the expansion has continued apace, under very different sorts of Congresses and Presidents. Normative argument does not seem to have mattered. One can put the point more generally: American criminal law's historical development has borne no relation to any plausible normative theory -- unless "more" counts as a normative theory. Criminal law scholars may be talking to each other (and to a few judges), (7) but they do not appear to be talking to anyone else.
And changes in criminal liability rules do not necessarily mean changes in the scope or nature of behavior the system punishes. In a system structured as ours is, the law on the street may remain unchanged even as the law on the books changes dramatically. Rather, broader substantive criminal law chiefly affects the process, the way law-on-the-street is made and the way guilt or innocence is determined. As criminal law expands, both lawmaking and adjudication pass into the hands of police and prosecutors; law enforcers, not the law, determine who goes to prison and for how long. The end point of this progression is clear: criminal codes that cover everything and decide nothing, that serve only to delegate power to district attorneys' offices and police departments. We have not reached that point yet; substantive criminal law has not wholly ceased to operate. But we are closer than we used to be -- the movement is very much in that direction. In a criminal justice system that incarcerates two million people, criminal law is becoming a sideshow. It seems like, and is, an unhealthy state of affairs.
Which raises an obvious question: How did all this happen? How did criminal law come to be a one-way ratchet that makes an ever larger slice of the population felons, and that turns real felons into felons several times over? The conventional answer is politics. (8) Voters demand harsh treatment of criminals; politicians respond with tougher sentences (overlapping crimes are one way to make sentences harsher) and more criminal prohibitions. This dynamic has been particularly powerful the past two decades, as both major parties have participated in a kind of bidding war to see who can appropriate the label "tough on crime." Congress's enactment of the famous hundred-to-one crack sentencing provision in 1986 (9) is the best-known example -- that ratio rose steadily as the relevant legislation wound its way through Congress, with members vying with one another to see who could propose the toughest crack penalties. (10)
This explanation has a good deal of power, but it is incomplete. Criminal defendants have not always been the political bogey they are today, nor has crime always been such a salient national issue. Criminal law's expansion, though, is a constant, going back (at least) to the mid-1800s. And while it is easy to see how public opinion would push toward harsher sentences (as with the 1986 drug legislation), it is hard to see how it would produce broad criminal codes that cover a range of ordinary, fairly innocuous behavior. The more natural assumption is that the public would want to criminalize only the kinds of things criminals, understood in the ordinary sense of that word, do. Yet contemporary criminal codes cover a good deal of marginal middle-class misbehavior -- a very odd state of affairs, politically speaking. The question remains: Why are criminal codes so broad, and why are they always getting broader?
A large part of the answer involves not the politics of ideology and public opinion, but the politics of institutional design and incentives. Begin with the basic allocation of power over criminal law: legislators make it, prosecutors enforce it, and judges interpret it. In this system of separated powers, each branch is supposed to check the others. That does not happen. Instead, the story of American criminal law is a story of tacit cooperation between prosecutors and legislators, each of whom benefits from more and broader crimes, and growing marginalization of judges, who alone are likely to opt for narrower liability rules rather than broader ones. This dynamic does not arise out of any particular ideological stance, and does not depend on the partisan tilt of the relevant actors. Criminal law seems to expand as much, and as fast, under Democrats as under Republicans. Rather, it arises out of the incentives of the various actors in the system. Prosecutors are better off when criminal law is broad than when it is narrow. Legislators are better off when prosecutors are better off. The potential for alliance is strong, and obvious. And given legislative supremacy -- meaning legislatures control crime definition -- and prosecutorial discretion -- meaning prosecutors decide whom to charge, and for what -- judges cannot separate these natural allies.
So two kinds of politics drive criminal law. Surface politics, the sphere in which public opinion and partisan argument operate, ebb and flow, just as crime rates ebb and flow. Usually these conventional political forces push toward broader liability, but not always, and not always to the same degree. A deeper politics, a politics of institutional competition and cooperation, always pushes toward broader liability rules, and toward harsher sentences as well. The current tough-on-crime phase of our national politics will someday end; indeed it seems to be ending already, as the current controversies over the death penalty and racial profiling suggest. (11) (The war on terrorism may reverse this trend.) But the deeper politics...