In the eyes of many politicians, activists and most of the voting public, tough punishment of nonviolent drug offenders is the main driver of mass incarceration in general and disproportionate imprisonment of people of color in particular. But in his new book “Locked In,” criminologist John Pfaff challenges that assumption, attributing mass incarceration primarily to violent crime and the public policy response to it.

Which side in this high-stakes debate has more facts on its side?

Within the federal prison system, the “war on drugs” seems to explain the bulk of imprisonment and racial disparities as well: 49.5 percent of all federal inmates are serving time for drug crimes, and the percentages are even higher for black (51 percent) and Hispanic (57.7 percent) inmates. But the federal system is a small and unrepresentative slice of the U.S. prison system. State prisons house almost 90 percent of U.S. inmates, and data from that far larger and more representative system tell a different story.

Consistent with Pfaff’s analysis, drug crime convictions account for the incarceration of fewer than 1 in 6 state prison inmates. Also, drug offenses don’t contribute to the appalling racial disparities in imprisonment. The white rate of being sentenced for drug crimes (15 percent) is actually slightly higher than that for blacks (14.9 percent) and Hispanics (14.6 percent). Reducing sentences for nonviolent drug criminals would thus not only have a small impact on mass incarceration, it could also worsen racial disparities in imprisonment rates.

In contrast, and again consistent with Pfaff’s thesis, violent offenses explain the majority of mass imprisonment. They also drive racial disparities in imprisonment because the rate for whites (46.6 percent) is significantly lower than that for blacks (57.8 percent) and Hispanics (58.7 percent).

Given the outsize role of violent crime in mass imprisonment, what should be done about it?  Pfaff favors “cutting long sentences for people convicted of violence, even for those with extensive criminal histories, since almost everyone starts aging out of crime by their 30s.” He also advocates for relying less on prison altogether and expanding community-based anti-violence programs that have strong evidence of preventing violence in the first place.

Yale University Law Professor James Forman Jr., author of the forthcoming book “Locking Up Our Own,” agrees with Pfaff’s prescriptions but emphasizes the challenges of implementing them. Forman notes that there are “political barriers to reducing sentences for violent crimes. Even more importantly, we have underinvested in violence prevention programs since the 1970s. Programs that work within the nooks and crannies of one particular community aren’t taken to scale because they lack the required funding and human capital.”

The significant political and practical challenges of transforming society’s response to violence may help explain why voters and politicians cling to the myth that prisons will empty out if we simply reduce penalties for nonviolent drug crimes. The road of reform laid out by Pfaff and Forman is a relatively harder one to walk. But if the desired destination is an end to mass incarceration, it’s the only way to get there.

Keith Humphreys is a professor of psychiatry at Stanford University.