If I told you that Hillary Clinton wasn’t a real progressive because she supported the Violence Against Women Act (VAWA) and the federal assault weapons ban, you’d think I was either crazy or joking. But in fact, the 1994 Crime Bill, which included both provisions, has become one of the standard anti-Hillary-Clinton soundbites used by supporters of Sen. Bernie Sanders. (Yes, the president at the time was Bill Clinton, not Hillary Clinton, and a Vermont congressman named Bernie Sanders voted for the bill, but never mind. Facts have a well-known a counter-revolutionary bias.)

That bill was a mix of provisions, including, as usual, the good, the bad, and the ugly. Some of what seemed good at the time to progressives actually hasn’t turned out to be of much use; some of what seemed bad at the time to progressives (especially “100,000 cops”) has turned out to be hugely beneficial to poor minorities living in big cities; and much of the ugly stuff just hasn’t mattered, though the really ugly prison-building provisions mattered a lot. But you can’t make much sense of any of it without putting your mind back to the circumstances of 1994.

1994 was the fifth consecutive year in which the FBI counted more than 23,000 murders: an all-time high. Then, as now, about half of the victims were African-American. As it turned out, that was the peak. Today’s murder rate (and total violent-crime rate) is just about half of its 1994 level. From our perspective now, making crime a primary political issue looks a little bit nuts. That’s in part, of course, because the suffering of poor people and minorities continues not to count very much in the political calculus. If the homicide mortality in the white population were at the current African-American level – let alone at the 1994 African-American level – that would be considered a national crisis.

But history is made looking forward, not backward. No one knew then that we’d seen the worst. All we knew is that the number of murders had more than doubled, that the total number of violent crimes had increased six-fold in the previous thirty years, that no reversal of trend seemed to be in sight, and that the street-level arms race financed by the crack trade had expanded the age range of killers and their victims down into adolescence. If you weren’t seriously worried about crime in 1994, you just weren’t paying attention.

The question was what to do about it, and no one had much of an answer. The Great Society era had also been the era of the first great crime explosion, and the possibly wrong but inevitable conclusions had been drawn that we didn’t know how to control crime with social welfare programs. In fact, aside from the Job Corps, none of the Great Society programs had demonstrated much in the way of crime-control benefits. The scale of the social service effort simply wasn’t in proportion to the scale of economic devastation created by de-industrialization, and “crime prevention” was not a field either bubbling with convincing new ideas or seriously devoted to the rigorous evaluation of its efforts.

Putting more cops on the street, Bill Clinton’s favorite approach, had little in the way of evidence behind it. The accepted wisdom, among police chiefs and criminologists alike, was that policing responded to crime rather than controlling it. Most of what police departments did was random patrol in cars plus rapid response to reports of crime. But George Kelling’s Kansas City experiment seemed to show that “preventive patrol” didn’t actually prevent anything, and “rapid response” almost never made a difference in either stopping a crime in progress or catching the criminal.

“Community” or “problem-solving” policing, going back to the work of Herman Goldstein in the 1960s, had been developed and brought to prominence over the previous several years at the Harvard Kennedy School’s Executive Session on Policing, whose intellectual sparkplugs included Kelling, Mark Moore, Edwin Meese III, Francis Hartmann, Robert Wasserman, Malcolm Sparrow, David M. Kennedy, James Q. Wilson, and several current and former police chiefs, including a rising star named William Bratton. The central idea was to make residents and police departments partners in crime-fighting, with the tantalizing promise of not merely lower crime rates but more lawful, less intrusive policing leading to improved police-community relationships in high-crime neighborhoods. A related idea was to use information technology to reallocate officers from low-crime areas and low-crime times of the day to high-crime times and places: “Putting the cops on the dots.” There was also some hope that “order maintenance” or “broken-windows” policing – an idea developed by Kelling and James Q. Wilson a decade earlier – might be integrated into the problem-solving/community-policing framework, with the likely effect of making people feel safer while potentially actually reducing serious crime.

At the time the 1994 crime bill passed, Bratton had just moved from Boston to take over the New York Police Department, and had made his famous commitment to getting the homicide rate down by 10%. But that commitment reflected a gamble on a new strategy, not the application of well-documented practice. There was reason to hope that adding more cops might do some good, especially if paired with a shift in management. But “reason to hope” isn’t the same as “reason to believe.”

Among the worst ideas out there was building more prisons. I say that as someone who had thought, a decade earlier, that prison capacity was inadequate. But by 1994, there were 1 million felons in state prison – a doubling in a decade, a quadrupling since the trough a decade before that – in addition to half a million misdemeanor convicts and pre-trial detainees in local jail (also representing a doubling over a decade). In addition to whatever deterrent effect the threat of incarceration might have, putting people in prison has a direct, mechanical effect on crime through incapacitation; people inside can’t commit crimes against people on the outside. (Prisoner-on-prisoner crime other than homicide mostly isn’t counted officially.)

But both deterrence and incapacitation are subject to the law of diminishing returns. When the prison population is small, it consists mostly of serious, high-rate offenders, because prosecutors and judges try to single them out. Therefore, when the population grows, it grows mostly by adding less and less dangerous people. So there was really no reason to think that adding yet more prisons would do much to stem the crime wave.

William Spelman has since estimated that the total crime-reduction gain from the entire prison-building binge might have accounted for something like 20% of the overall crime drop. Since more than half the new building was pre-crime bill, and since the crime control benefit per prisoner-year falls as the number of prisoners rises, the additional contribution of the 800,000 prisoners added post-1994 cannot have been very large, even if we ignore – as we should not – all the ways in which increasing incarceration can eventually increase crime by weakening communities, changing family structure, and reducing the deterrent value of the threat of incarceration by making it commonplace. Bert Useem and Ann Morrison Piehl estimate that, under today’s conditions, adding an additional prisoner is more likely to cause a crime increase than a crime decrease.

With nothing really compelling in the way of strategic ideas available, and faced with a range of ideological and interest-group demands, the (Bill) Clinton administration decided to offer a little bit of everything. The thing the President and his advisers really seemed to believe in was carrying out his campaign pledge to put “100,000 new cops on the beat” by providing federal money to hire local police officers. But given the apparent uselessness of just adding more resources to a system that didn’t know what it was doing, the Administration added a twist: to get the money, departments had to commit – at least on paper – to implement “community policing” to move departments away from the demonstrated futility of the random-preventive-patrol-plus-rapid-response-to-calls-for-service model that had dominated the “reform” style of policing since the 1920s.

Since then, of course, crime has fallen dramatically; the extent to which the crime bill’s contribution of more cops and encouragement of new policies contributed to that fall is subject to dispute. Massive social trends tend to be over-determined, and no one at the time the bill passed predicted the spectacular crime decline that ensued. But it’s worth noticing that most of the explosion in incarceration for which the bill (and President Clinton as its sponsor, and Hillary Clinton because her last name is “Clinton”) get the blame occurred before the bill became law, while virtually all of the decline in crime decline for which neither the bill nor either Clinton gets any credit came after the bill’s passage. And there is now both theory and evidence supporting the idea that adding cops tends to drive down crime rates, especially if the increase in workforce is paired with reforms in management and strategy. All we know for sure is that about 10,000 fewer Americans will be murdered this year than were murdered in 1994, and about half of those avoided deaths were African-Americans.

In addition to “100,000 cops,” the bill was a grab-bag of programs popular with liberals or conservatives: VAWA (a good idea, later partly gutted by the Supreme Court), the assault weapons ban (mostly an irrelevancy because few murders involve “assault weapons,” however defined, and because the definition of “assault weapon” in the law allowed functionally equivalent firearms to continue to be sold), federal “three-strikes” legislation (again mostly an irrelevancy because the federal system holds only about a tenth of the nation’s prisoners), 41 new capital punishment crimes (none of them ever used; the federal government has executed only three people since the law passed: Oklahoma City bomber Timothy McVeigh and two other murderers); a really nasty provision denying educational grants to prisoners; and a bunch of “crime prevention” social service programs, few of them with any demonstrated crime prevention value.

The one disastrously bad provision of the law was grants to states to build new prisons. Not only was that a bad idea on its own, but it was made much worse by the addition of another conservative shibboleth, “truth in sentencing.” Truth in sentencing is an excellent example of giving a vicious twist to an otherwise sensible idea. Under the old system of “indeterminate sentencing,” there was only a fuzzy relationship between the sentence pronounced from the bench and the time actually served by the offender. It was modified by a number of statutory rules including “good time,” but also by the parole system, which allowed an administrative board to release a prisoner after a fraction – typically one-third – of his or her nominal sentence. Everyone in the courtroom understood that “twelve years” actually meant something like four years: except the victim and the victim’s family, who were often horrified to discover that the offender they had thought safely stashed away was in fact back on the street. So there was a reasonable case for bringing the actual sentence in line with the nominal sentence, even aside from well-justified concerns that the parole system, which operated largely without published standards, might introduce another source of racial bias.

However, there were two ways to achieve truth in sentencing: shorting the sentence pronounced from the bench (but making the offender serve all of it), thus eliminating the falsehood without changing the time to be served, or eliminating discretionary release while leaving the nominal sentences unchanged, thus in effect tripling effective sentence lengths. The Republicans, who had made prison-building and truth in sentencing the price of their acquiescence with the rest of the bill, insisted that in order to get federal prison-building money the states introduce this concept for violent crimes without shortening nominal sentences. (“State’s rights,” you say? “Federalism,” you object? But in Republican minds that’s only for nice policies; cruelty should be uniform.)

As Kevin Drum has pointed out, it’s simply not the case that the 1994 crime bill created the mass-incarceration problem; most of the damage had already been done. (In 1994, the prison-plus-jail headcount had reached 1.5 million, about triple its 1975 level, on its way to the current peak of 2.3 million.) But the crime bill certainly made the problem worse, and it contributed – directly through the money, indirectly through the rules, and even more indirectly by spreading truth in sentencing as a policy meme – to the policy of continuing to increase the size of the prison population as the crime rate collapsed. The prison-building money, and especially the truth in sentencing provision, were predictably bad policies, and the people who fought for those provisions should be held accountable for what was at best a bad mistake and at worst an act of pointless, politically-motivated cruelty.

But the people who fought for those provisions almost all had (R) behind their names. In the politics of the crime bill, the death penalty, the anti-prisoner-education provision, and especially prison-building and truth in sentencing were understood to be the conservative interests, while the VAWA, the assault weapons ban, and the “prevention” agenda were designed to appeal to liberals. (The cops provision was designed to appeal to the broader public, but the political muscle behind it came mostly from the White House and, obviously, local law enforcement.)

Again, it’s crucial to remember the political mood of 1994. The alternative to the 1994 crime bill was not a bill written by the American Civil Liberties Union – though with better presidential leadership and more adroit maneuvering from congressional progressives we might have passed something better than what actually emerged by getting something through in 1993, a year farther out from the midterm elections – but a truly appalling crime bill the Gingririchized Congress would have passed in 1995.

Hillary Clinton did make one serious blunder: she allowed herself to be taken in by a then-slightly-fashionable criminological theory – since then not merely refuted by the facts but retracted by its most prominent academic sponsor – that the upsurge in homicides by teenagers reflected not merely the transient effects generated by the spread of the street markets for crack cocaine, but the emergence of a new class of violent juvenile “super-predators” whose psychological makeup was allegedly different from the makeup of the typical ordinary criminal. Based on that (as it turned out) fallacious bit of science, dozens of states passed laws allowing juveniles to be tried and sentenced as adults, with predictably disastrous results. On at least one occasion – after the passage of the bill, and the course of arguing for community policing – Clinton used the phrase “super-predator.”

Sanders, so far as the record shows, never talked about “super-predators.” That’s to his credit. But it’s less to his credit that he has rarely talked about, or tried to legislate on, crime at all.

Sanders’ theory seems to be that crime would mostly go away if the country offered better economic opportunity to poor people. That it’s hard to offer economic opportunity without first offering security of person and property doesn’t seem to have occurred to him; but then, it’s not much of a problem in Burlington. If Burlington had the crime rate of Brooklyn, or if Sen. Sanders had stayed in Brooklyn rather than moving to the peace, quiet, and safety of Burlington, he would have had to pay more attention to crime, because his constituents (of all races and economic classes, but especially minorities and the poor) would have been deeply worried by it.

Indeed, Sanders, more than most politicians or academics, might have noticed that crime control – protection of people and their property from victimization, and protection of the social and economic life of the community from the costs created by the fear of crime and the avoidance of victimization – is a first-order requirement of class justice and of racial justice. As Randall Kennedy argued years ago in Race, Crime, and the Law, and as his Harvard Law School colleague William Stuntz argued even more eloquently in The Collapse of American Criminal Justice, our actual social practice does not meet the Fourteenth Amendment standard of providing “equal protection of the law.” You are more likely to be killed or robbed, and your home is more likely to be burglarized in the poorer, darker parts of Brooklyn than it is in lighter-skinned and more prosperous areas of Manhattan or Burlington. And the person who victimizes you is less likely to be caught and punished, and more likely – holding constant the severity of the offense – to get off lightly even if he (or, less likely, she) is caught. Impunity is one of the factors that maintains high rates of victimization.

That’s what made it important to pair community policing with “putting cops on the dots,” thus giving more police protection to those more in need of it, who were also generally poorer than those who lived in relatively over-policed low-crime neighborhoods.

To be sure, that reallocation of resources, if not handled delicately and with due deference to neighborhood leadership, risks subjecting young men in high-violence neighborhoods to intolerable levels of police intrusion; New York City, for example, has radically cut back on street stops in response to community outcry. Police management remains one of the toughest of all public-management tasks, and no one does it perfectly. David Kennedy and the National Network for Safe Communities assembles academic talent, police management, and community advocates to carry “community policing” to its next level, in which it becomes a tool of community empowerment for the communities that need it most.

And that is why someone passionately concerned about crime and about racial and social justice could reasonably prefer Hillary Clinton’s record on the issue – warts and all – to Senator Sanders’s policy of benign neglect.

Mark Kleiman

Mark Kleiman is a professor of public policy at the New York University Marron Institute.