Ta-Netisi Coats quoted Bill Stuntz yesterday on the scale of US imprisonment:
In 2008, the imprisonment rate was more than three-and-a-half times that older historical record....The wave of incarceration that produced those numbers extended nationwide: in every region, imprisonment at least tripled in the twentieth century's second half. In some jurisdictions, the increase was much larger. In 1950, 77 of every 100,000 Texans were housed in state penitentiaries; by 2000, the figure was 730. Massachusetts' imprisonment rate rose from a mere 32 in 1972 to 278 in 1997. Between 1973 and 2003, Mississippi's imprisonment rate rose from 76 to 763....If the general imprisonment rat is high, the rate of black incarceration can fairly be called astronomical. The black imprisonment rate for 2000 [1,830 per 100,000] exceeds by one-fourth the imprisonment rate in the Soviet Union in 1950---near the end of Stalin's reign, the time when the population of Soviet prison camps peaked. If jail inmates are included, per capital black incarceration is 80 percent higher than the rate at which Stalin's regime banished its subjects to the Gulag's many camps.
These are large increases in the recent past: 1972 is a year I can remember, but Mississippi has managed to increase it's imprisonment rate ten fold since then.
I went looking for a accounting breakdown of how this large increase came to be, something along the lines of changes in crime rates, changes in arrest rates per crime, change in charging rate per arrest, change in conviction/plea bargaining rate per charge, sentence per conviction/plea bargain, duration of imprisonment per sentence, all suitably accounted for by underlying known causes in these rates (say demographics for crime rates, arrest rate by type of crime, charging rate for type of crime etc.). I didn't come up with much, except for one paper that sort of take stab at this question. It's
John F. Pfaff, The Causes of Growth in Prison Admissions and Populations, 2011
The explosive growth in the US prison population is well documented, but its causes are poorly understood. In this paper I exploit previously-unused data to define precisely where the growth is occurring. In short, the growth in prison populations has been driven almost entirely by increases in felony filings per arrest. All other possible sites of growth - arrests, admissions per filing, convictions per filings and admissions per conviction, and even (perhaps most surprisingly) time served per admission - have barely changed over the past four decades. But the growth in filings tracks that of admissions almost perfectly. This paper demonstrates the importance of felony filings and considers some of the possible explanations for their growth.
Here is Pfaff's graph on felony charging rate per arrest and admission to prison rate per charged felony (if I'm reading the paper correctly).
It would be nice to know more...this paper isn't quite up to good econ paper standards. You'd think the sort of work I outlined above would standard, but it doesn't seem to exist.
So why more felony charges per arrest? Stuntz clearly sees this as an out of control criminal justice system where prosecutors have the wrong incentives, for historically grounded reasons, but I don't know yet if he tells a convincing story -- I suspect not, but that's not a failing, just a call for more people to try to explain what's going on.
There is really a 80% felony charging rate per arrest now?!
Pfaff speculates himself:
The War on Drugs is not increasing admissions directly via the incarceration of drug offenders. But it may be increasing admissions indirectly, by lengthening the records defendants have, and thus the likelihood that prosecutors opt to file charges against them for non-drug crimes. These results may also point to an ongoing collateral cost of the crime boom of the 1960s to the 1990s: by producing cohorts of offenders with longer records, it generated a pool of offenders that may face tougher sanctioning outcome even when all else is constant, thus helping prison populations trend upwards even as crime rates fall.
Somebody should write a paper keeping track of the capital stock of past arrests/convictions and how it plays out for crime rates, arrest rates, felony charges and convictions and imprisonment.
I agree that this is really interesting. The numbers seem quite extraordinary. I am not sure I know what "filings" and "admissions" are in any technical sense, but I am particularly surprised that time served per admission has not changed.
The comparisons to the more distant past are also interesting. Since conservatives like to blame the Warren Court and "1960s liberalism" for weakening the criminal justice system at both federal and state levels, it is striking to see incarceration rates that are so much higher than those in 1950. It would be interesting to see more of those numbers.
Posted by: Doug | September 20, 2011 at 02:19 AM