Fewer Cages

Human nature is hard to change, so hard that even burglary sentences don’t have much effect upon  defendants’ future conduct.

Georgia uses a system of discretionary sentencing where most crimes carry a broad range of potential punishments.  For instance, a first time burglary can carry anywhere from two years of probation to 25 years in a state prison.  Trial judges have almost unbridled authority to impose any sentence in this range.

How to treat first time burglars is arguably the most contested criminal justice issue in the Henry County Superior Court.  In the last decade of his 24-year tenure as District Attorney, Tommy Floyd adopted a restrained approach towards first time burglars.  The “standard sentence” became five years probation with 60 to 120 days in a probation detention center.  Some first time burglars were able to avoid any jail time beyond their initial arrest.  Floyd’s approach was unusual and was more lenient than that of Democratic district attorneys in neighboring counties.  While Floyd’s prosecutors recommended the detention center for first time burglaries, prosecutors in DeKalb and Clayton County typically asked for prison time.  (Any custodial sentence of one year or more in a felony case is typically served in prison, shorter sentences are served in jails or detention centers).  The Superior Court of Henry County was a strangely enlightened place.

In the last decade, there has been push back against Floyd’s gentle ways.  Judge Brian Amero, a former Clayton County prosecutor, took the bench in 2006.  His “standard sentence” for a first time burglary is one year in prison, followed by four years probation.  Early on, Amero largely deferred to Floyd’s sentencing recommendations.  In 2007, the majority of Judge Amero’s first offender sentences for burglary did not involve prison time.  However, by 2011, Amero had become more confident in his sentencing philosophy, almost half of his first offender burglary sentences involved prison time.  Floyd’s successor as district attorney, Jim Wright, shared Judge Amero’s philosophy.  Under Wright, the standard plea offer for a first offender burglary involved a year in prison.  The current District Attorney, Darius Patillo, appears to be continuing Wright’s practice.

The competing approaches to first offender burglary sentences have created an intriguing natural experiment.  Henry County judges imposed at least 212 first offender sentences for burglary in cases filed between January 1, 2007 and December 31, 2011.  Virtually all of these sentences were imposed at least five years ago.  The structure of first offender sentences makes it possible to rigorously compare how defendants given different types of sentences have fared.

Under Georgia law, any person who is accused of burglary is eligible for first offender treatment if they have never been convicted of a felony and have not previously received first offender treatment.  This makes its possible to ballpark the criminal history of defendants given first offender sentences.  A typical felony sentence does not specify the offender’s criminal history.  A long sentence might indicate either an unusually severe offense or a substantial criminal history.  However, any time a first offender sentence is imposed, it is very likely that the defendant has not been convicted of a felony and has not received first offender treatment in the past.  In other words, we know that defendants who received first offender sentences had relatively limited criminal histories.

Every one of the first offender sentences for burglary entailed at least five years of probation.  This makes it easy to measure recidivism.  If a probationer is caught committing a crime or violating the terms of his probation, his probation officer initiates a probation revocation hearing.  When the violations are sufficiently serious, the probation officer also seeks an “adjudication of guilt,” which involves a felony conviction on the original felony charge.  If the judge finds that a first offender has violated his probation, he can adjudicate the defendant guilty and/or sentence the defendant up to the statutory maximum for burglary, which is generally 25 years.  If the judge imposes incarceration, he is required to give the defendant credit for any time served on probation.  Thus, a first offender who violated his probation after two years could receive up to 23 years in prison: the 25 year maximum minus the two years already served on probation.

I used probation revocation proceedings to gauge the success of the judicial system in rehabilitating first offenders.  A case was considered a failure if any of three things occurred: (1) the defendant was adjudicated guilty of burglary and became a convicted felon; (2) the defendant received a custodial sentence of 12 months or more because of a new offense or violation of probation; or (3) the defendant absconded from probation and has not yet been apprehended.  A case was considered a “success” if none of these three things happened.  Many “successful” probationers did violate their probation, but their violations were minor enough that the judge let them keep their first offender status and did not sentence them to 12 months or more.

I sorted the first offenders into three categories.  First offenders who were not sentenced to incarceration, those who received incarceration short of prison, and those who received a year or more in a state prison.  The most common sentence, “incarceration short of prison” typically involved assignment to a detention center for 60-120 days, assignment to a boot camp program for 90 days, or between 60 and 120 days in the county jail.  Most offenders in this category spent about 60 days incarcerated before beginning their probation.  The “no incarceration” category included suspended custodial sentences.  The outcomes for each category were as follows

No Incarceration Incarceration short of prison One plus year of prison Totals
Successes 40 68 15 117
Failures 30 46 13 86
Total 70 114 28 212

Here are the success rates of each category, as well as the standard deviation attributable to sampling error:

No incarceration                               57.1% +/- 5.9%

Incarceration short of prison         59.6% +/- 4.6%

Prison                                                   53.6% +/9.4%

Overall                                                  55.1% +/-4.2%

Superficially, these results suggest that the most effective sentence is incarceration short of prison.  These defendants were 4 to 5% more likely to successfully complete a five year (or longer) period of supervision than defendants who went to prison.  However, these results are nowhere near statistical significance.  The standard deviation for each of the three subsets was at least 4.6%.  This means the plausible ranges for the three figures overlap significantly.  When using sampling, the true value is within one standard deviation of the sample mean 66% of the time, and within two standard deviations 95% of the time.  Thus, there is a 95% chance that the true success rate for “incarceration short of prison” is between 50.2 and 69.4%.  Any value within that range is plausible.  Nor can we reduce the sampling error by adding more observations.  Judge Amero took the bench in 2007, and we need a five year observation period to gauge whether a defendant succeeded on probation.  We might scrounge a few more cases charged in late 2006 and early 2012, but this would not materially reduce the sampling error.

The takeaway from the data is that sentences probably don’t have a big effect on rehabilitation.  This is not a trivial finding.  A year in prison costs taxpayers $23,000.  Two months in a detention center cost about $4,000.  The detention center is probably a more effective punishment, and there is no evidence at all that prison lowers recidivism rates beyond what the detention center achieves.  Though the data are murky, they buttress my respect for an old-fashioned type of conservatism, one I have often seen Judge Crumbley display on the bench.  Human nature is imperfect and stubborn.  Once a defendant gets out of jail, the best predictor of his actions are his temperament, family, and associates.  If prison isn’t demonstrably effective, why not try something else first?  Burglary is a poor predictor of murder, and the hardened criminals will be back in court soon enough.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s