Common sense and the opinions of most people actually working on the front lines of criminal justice reveal to us the simple fact that more criminals in prison equals less crime on the streets; but academics have been trying to find a way around this for years using the Marxist-themed term “mass incarceration” but their best results so far are inconclusive and not even very suggestive, as this article from the New York Review of Books indicates.

An excerpt.

More than 2.2 million people are currently incarcerated in US jails and prisons, a 500 percent increase over the past forty years. Although the United States accounts for about 5 percent of the world’s population, it houses nearly 25 percent of the world’s prison population. The per capita incarceration rate in the US is about one and a half times that of second-place Rwanda and third-place Russia, and more than six times the rate of neighboring Canada. Another 4.75 million Americans are subject to the state supervision imposed by probation or parole.

Most of the increase in imprisonment has been for nonviolent offenses, such as drug possession. And even though crime rates in the United States have declined consistently for twenty-four years, the number of incarcerated persons has continued to rise over most of that period, both because more people are being sent to prison for offenses that once were punished with other measures and because the sentences are longer. For example, even though the number of violent crimes has steadily decreased over the past two decades, the number of prisoners serving life sentences has steadily increased, so that one in nine persons in prison is now serving a life sentence.

And whom are we locking up? Mostly young men of color. Over 840,000, or nearly 40 percent, of the 2.2 million US prisoners are African-American males. Put another way, about one in nine African-American males between the ages of twenty and thirty-four is now in prison, and if current rates hold, one third of all black men will be imprisoned at some point in their lifetimes. Approximately 440,000, or 20 percent, of the 2.2 million US prisoners are Hispanic males.

This mass incarceration—which also includes about 800,000 white and Asian males, as well as over 100,000 women (most of whom committed nonviolent offenses)—is the product of statutes that were enacted, beginning in the 1970s, with the twin purposes of lowering crime rates in general and deterring the drug trade in particular. These laws imposed mandatory minimum terms of imprisonment on many first offenders. They propounded sentencing guidelines that initially mandated, and still recommend, substantial prison terms for many other offenders. And they required lifetime imprisonment for many recidivists. These laws also substantially deprived judges of sentencing discretion and effectively guaranteed imprisonment for many offenders who would have previously received probation or deferred prosecution, or who would have been sent to drug treatment or mental health programs rather than prison.

The unavoidable question is whether these laws have succeeded in reducing crime. Certainly crime rates have come down substantially from the very high levels of the 1970s and 1980s that gave rise to them. Overall, crime rates have been cut nearly in half since they reached their peak in 1991, and they are now at levels not seen in many decades. A simple but powerful argument can be made that, by locking up for extended periods the people who are most likely to commit crimes, we have both incapacitated those who would otherwise be recidivists and deterred still others from committing crimes in the first place.

But is this true? The honest answer is that we don’t know. And it is this uncertainty that makes changing the status quo so difficult: for, the argument goes, why tamper with what seems to be working unless we know that it isn’t working?

There are some who claim that they do know whether our increased rate of incarceration is the primary cause of the decline in crime. These are the sociologists, the economists, the statisticians, and others who assert that they have “scientifically” determined the answer. But their answers are all over the place. Thus, for example, a 2002 study by the sociologist Thomas Arvanites and the economist Robert DeFina claimed that, while increased incarceration accounted for 21 percent of the large decline in property crime during the 1990s, it had no effect on the similarly large decline in violent crime. But two years later, in 2004, the economist Steven Levitt—of Freakonomics fame—claimed that incarceration accounted for no less than 32 percent of the decline in crime during that period.1

Levitt’s conclusions, in turn, were questioned in 2006, when the sociologist Bruce Western reexamined the data and claimed that only about 10 percent of the crime drop in the 1990s could be attributed to increased incarceration. But two years after that, in 2008, the criminologist Eric Baumer took still another look at the same data and found that it could support claims that increased incarceration accounted for anywhere between 10 percent and 35 percent of the decrease in crime in the 1990s.

As these examples illustrate, there is nothing close to an academic consensus on the proportion of the decrease in crime attributable to increased incarceration. Last year, a distinguished committee of the National Research Council, after reviewing the studies I have mentioned as well as a great many more, was able to conclude only that while most of the studies “support the conclusion that the growth in incarceration rates reduced crime…the magnitude of the crime reduction remains highly uncertain.”2

Most recently, in February 2015, the Brennan Center for Justice at NYU Law School published a study entitled “What Caused the Crime Decline?” that purports to show that increased incarceration has been responsible for only a negligible decrease in crime. One cannot help but be impressed by the sheer scope of the study. The authors identify the fourteen most popular theories for the decline in crime in the last few decades and attempt to test each of them against the available data.

Five of the theories involve criminal justice policies: increased incarceration, increased police numbers, increased use of statistics in devising police strategies to combat crime, threat of the death penalty, and enactment of right-to-carry gun laws (which theoretically deter violent criminals from attacking victims who they now have to fear might be armed). Another four of the theories are economic in nature, involving changes in unemployment, income, inflation, and consumer confidence. The final five theories involve environmental and social factors: aging population, decreased alcohol consumption, decreased crack use, legalized abortion, and decreased lead in gasoline (which theoretically reduces the supposed tendency of lead fumes to cause overaggressive behavior).

The primary findings of the Brennan study are that “increased incarceration has had little effect on the drop in violent crime in the past 24 years” and has “accounted for less than 1 percent of the decline in property crime this century.” To reach these striking results, the authors rely (as did most of the earlier studies cited above) on the social scientist’s favorite method, a multivariable regression analysis that “controls for the effects of each variable on crime, and each variable on other variables.” But as anyone familiar with regression analysis knows, it rarely speaks to causality, as opposed to correlation; and even to show correlation, the analysis involves a lot of educated guesswork. The authors admit as much, but seek to downplay the level of uncertainty, stating: “There is always some uncertainty and statistical error involved in any empirical analysis.” But when you are dealing with matters as difficult to measure as how much of the decrease in crime can be attributed to everything from decreased alcohol consumption to increased consumer confidence, your so-called “estimates” may be little more than speculations.

In an attempt to adjust to this difficulty, the authors state the percentage of crime decrease attributable to each given factor as a range, e.g., increased police numbers accounted, according to the study, for between 0 percent and 5 percent of the decline in crime between 1990 and 2013. But if you take the low end of each of the ranges, the fourteen factors analyzed in the Brennan study collectively accounted for as little as 10 percent of the decline in crime over that period; and even if you take the high end of each of the ranges, the various factors still accounted for only 40 percent of the decline in crime. Under any analysis, therefore, either the decline in crime in the last twenty years or so was chiefly the product of forces that none of the leading theorists has identified, or (as seems more likely) the regression analysis used by the authors of the Brennan study is too imperfect a tool to be of much use in this kind of situation.

My point is not to criticize the Brennan study. It is in many respects the most ambitious and comprehensive study of its kind undertaken to date. But as the National Research Council report points out in discussing the many similar studies that, as noted, led to a wide range of results, there are simply too many variables, uncertainties, estimates, and challenges involved in the question to rely on a regression analysis that is little more than speculation dressed up as statistics. The result is that one cannot fairly claim to know with any degree of confidence or precision the relative role of increased incarceration in decreasing crime.

Retrieved May 14, 2015 from