« The Poor are Still Losing: Gideon's Empty Promise | Main | Expenditures, Tuition, and U.S. News »

Tuesday, July 09, 2013

Prosecutors and Prison Admissions: The Main Story

In my last post, I briefly stated that prosecutorial filings have driven prison growth, at least since 1994. Here, I’ll demonstrate more carefully the impact of filings on admissions—and do so almost entirely in pictures, so no real reading required! My next few posts will then explain why longer sentences have not mattered that much. Taken together, these results imply that prosecutorial filings drive overall prison population growth.

In a criminological/graphical version of “Name That Tune,” I can show that prosecutorial filings are at the heart of prison admissions growth nationwide1 since crime began falling in five pictures.

First, crime has dropped steadily since 1991: violent crime by 48%, property crime by 43%. So for the period starting in 1994, crime only declines.

Screen Shot 2013-07-09 at 10.07.48 AM

Note though—and I’ll come back to this important-but-often-overlooked point in a future post—that crime is still high by historical standards. Between 1960 and 1991, violent crime soared by 371% and property crime by 198%. Crime is at life-time lows only for those under about 42 years old. For Boomers, the country remains substantially more violent than when they were young, and much of the crime drop they have experienced has come about from self-protective measures—not going out at night, car alarms, etc.—that don't make us feel safer. There's a reason crime has been such a powerfully salient issue for so long. 

But for our purposes here, the pool of violent and property offenders has declined since 1991.

Second, clearance rates for index crimes have been flat. Huge changes in police strategy and technology, no real change in clearance rates. Except for declines for murder, aggravated assault, and rape. This is puzzle beyond the scope of this post, but it’s a fascinating one.

Screen Shot 2013-07-09 at 10.11.22 AM

Ah, you say, what about drug arrests? Those are more discretionary. Fair question. But even if we add in non-marijuana drug arrests,2 we see that total arrests decline by about 8.5% in my sample of 34 states (and by slightly more in all 50 states). There is an increase in total arrests after 2000 due to increased drug arrests, but nowhere near enough to offset the general decline in all arrests.

Screen Shot 2013-07-09 at 10.13.04 AM

So total inputs into the criminal justice system have declined, whether you start with “crime” or “arrests.” Yet between 1994 and 2010, prison admissions in my sample rose by 40%. How?

The next graph plots total felony filings and total admissions. The former grows by 37%, the latter by 40%. This certainly suggests that filings are playing a major role.

Screen Shot 2013-07-09 at 10.17.26 AM

We can make the point even more clearly. The next graph plots filings per arrest and admissions per filing. (Recall that convictions are hard to see in the data.) The results are striking: filings per arrest soar, while admissions per filing remain flat.

Screen Shot 2013-07-09 at 10.21.04 AM
The basic lesson: once you commit a crime, you are no more likely to get arrested than before (except for some increased risk for drug offenders). Once arrested, though, you are much, much more likely to face a felony charge: the probability an arrest results in a felony filing rose from 0.375 to 0.573 over the period. Once charged, though, a defendant is no more likely to go to prison—the probability hovers at around 0.26 the whole time.

So these results, combined with those I’ll post soon showing the stability in time served, add up to a surprisingly simple story: the post-1991 prison growth is due to one actor—sort of 3—and more specifically to one choice by that actor. The policy implications of this are significant, and I will discuss them in more detail after my posts on sentence length.


1A note before we begin. These results, which come from this paper, and which are summarized here, draw on a sample of thirty-four states. As the map below indicates, these states are a fair cross-section of US states in terms of geography, industrialization, politics, etc. (Ignore the colors: if a state is any shade of blue, I use it here.) Moreover, when I’ve compared some of the numbers from my sample to nation-wide numbers, they all track fairly closely. I’m not counting this as one of my graphs, though—it’s background info.

Screen Shot 2013-07-09 at 10.02.08 AM

2It's very hard to go to prison on a marijuana charge. A short jail term, perhaps. But prison is unlikely.

3While the ultimate decision to file charges rests with the prosecutor, he may get substantial assistance from other actors. Police can bring better or worse cases, making the prosecutor’s job easier or harder, and legislators can provide prosecutors with more or fewer crimes to use as leverage to extract more and better pleas.

Posted by John Pfaff on July 9, 2013 at 10:29 AM | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference Prosecutors and Prison Admissions: The Main Story:


"Once charged, though, a defendant is no more likely to go to prison—the probability hovers at around 0.26 the whole time."

Fascinating. Also noteworthy, though, is that the defendant is no LESS likely to go to prison. It seems prosecutors are prosecuting more cases now--- which presumably means weaker cases and less important cases. Yet they are getting prison sentences just as often, from these "smaller" cases. This implies one of two things:

1. The arrest pool is deep enough that prosecutors formerly let go many cases that were just as strong and important as the cases they did bring. So underfunding must have been a huge problem.

2. Prosecutors are now prosecuting cases that would have gotten a lower sentence or no sentence before, but judges and juries have gotten tougher.

This will be important to your Sentencing post, because "no change" actually could mean "got tougher".

Posted by: Eric Rasmusen | Jul 12, 2013 12:22:34 PM

The comments to this entry are closed.