The official SQ 780 savings calculation rests on flawed assumptions

Two years ago, Oklahoma voters passed State Questions 780 and 781, together known as the Smart Justice Reform Act. SQ 780 reclassified simple drug possession and many low-level property crimes as misdemeanors rather than felonies. SQ 781 directs the Office of Management and Enterprise Services (OMES) to calculate the savings to the state as a result of the changes made by SQ 780 and to distribute that amount to counties to provide mental health, substance abuse, and other rehabilitative services.

As required, OMES released the savings calculation for Fiscal Year 2018 on July 31. To the surprise of many, they estimated that the changes made by SQ 780 would save the state $63.5 million in FY 2018 and a total of $137.8 million from FY 2018 to FY 2026. Corrections Director Joe Allbaugh strongly criticized the report, saying that the Department of Corrections had not saved any money over the last year. Why is there such a divergence between the two agencies? Our analysis shows that the assumptions that OMES made are not supported by data, and they lead to an unrealistic picture of what SQ 780 accomplished in its first year.

The main disagreement: How many people avoided prison?

The disagreement between OMES and DOC hinges on the estimate of how many people would have gone to prison in FY 2018 if SQ 780 hadn’t gone into effect. OMES put that number at 9,182, while DOC said that there had been, at most, only 80.

OMES used criminal case filings from District Courts 

In their report, OMES starts with the total number of criminal cases filed in District Courts in FY 2018 that involved crimes affected by SQ 780. That number, provided by prosecutors’ offices, comes to 27,492. They then assumed that all 27,492 cases would have resulted in a conviction and a prison sentence if SQ 780 hadn’t gone into effect, and that one third of them – 9,182 people – avoided the prison sentence because their cases included only SQ 780 offenses.

Where did that one-third proportion come from? OMES looked at all DOC receptions that included offenses affected by SQ 780 in recent years, estimating that number at about 4,500 per year. Of that 4,500, OMES found that one third of them were serving sentences only for SQ 780 offenses, coming to about 1,500 per year in recent years. But in their calculation, OMES applied that one third proportion to all cases filed (over 27,000) rather than the much smaller number of cases that were sentenced to prison (only 4,500 in recent years).

Next, OMES had to estimate how long people sentenced to prison would serve. Based on recent averages, OMES assumed that those 9,182 people would have spent an average of 145 days in county jail waiting to be transferred to a state facility, 528 days in a state facility, and 1,140 days after release on community supervision.

Finally, OMES had to determine how to divide the savings over time in order to find the amount that was saved just in FY 2018. They decided to count the sentences of all 9,182 people charged in FY 2018 as starting their prison sentence on July 1, 2017 – the first day of FY 2018.

DOC used historical and current prison reception data

As OMES prepared their calculation, DOC Director Allbaugh sent a letter providing prison admissions data. The data showed that 1,483 people were admitted in FY 2017 for only SQ 780 offenses. In FY 2018 – the same year that OMES estimated 9,182 people were diverted – DOC received 1,403 people for only SQ 780 offenses. The number actually diverted in FY 2018, the data suggests, was a mere 80 people – less than one percent of OMES’ estimate.These hard numbers should have been taken into account, but OMES chose to use court cases instead of this historical prison reception data.

How could DOC receive 1,403 people for only SQ 780 offenses after those offenses were made ineligible for prison sentences? Those people were arrested before SQ 780 went into effect but weren’t sentenced to prison until after SQ 780 went into effect, so they were sentenced under pre-SQ 780 law. The numbers of receptions are basically the same the year before and after SQ 780 went into effect, demonstrating just how slow the justice system moves even with cases involving minor crimes. Thus, while the number of felonies filed across the state has dropped sharply, the drop in prison admissions will lag far behind.

OMES made assumptions not supported by historical data

In forming their calculation, OMES made several assumptions that are difficult to reconcile with historical data:

• Assuming a 100 percent conviction rate for all 27,492 criminal cases is clearly too high. Though Oklahoma doesn’t track conviction rates, we know that a significant portion of criminal cases are dismissed for a wide range of reasons. Rates of dismissal range from a low of 14 percent in Pennsylvania to 46 percent in North Carolina, among six states with comprehensive data available. Beyond dismissal, many people charged with low-level crimes receive deferred prosecution, suspended sentences, and many other specialty courts and programs. Though there are major differences in state justice systems that create wide disparities in conviction rates, that rate is certainly much lower than 100 percent in Oklahoma.
• Assuming that 100 percent of convictions would have resulted in a prison sentence is also unrealistic. We know that Oklahoma’s prisons have admitted between 8,000 and 10,000 people annually in recent years. We also know that only one in four of those convicted of felony drug possession were sentenced to prison in a 2010 study of Tulsa County. The OMES calculation estimated that all 27,492 people charged with SQ 780 offenses would have gone to prison in FY 2018 if the reform hadn’t gone into effect, but with the reform, only 18,310 did. That’s almost twice as many as the 10,778 people that DOC actually received in FY 2018.
• Assuming that all people charged in FY 2018 would enter DOC in FY 2018 fails to account for the slow pace of our justice system. Many who are admitted to prison for low-level offenses like drug possession first go to probation, drug court, or other forms of supervision, but violate terms of their supervision – by failing a drug test or missing meetings with their probation officer, for instance – and are admitted to prison after their supervision is revoked. DOC’s data shows just how slow the effects of reform kick in: despite implementing SQ 780 at the beginning of the year, the number of people admitted to prison for only SQ 780 offenses dropped by only 80 people, from 1,483 to 1,403. The savings will grow gradually as more people are diverted, but in early years, the savings are modest at best.

In short, OMES ignored two critical stages in the justice process, and assumed that everyone who enters the courts ends up in prison. But as the data show, the number of people who go to prison is a fraction of the people who are charged in court, and an even tinier fraction go to prison in the same year they were charged. 

Justice data is scarce, but a realistic calculation must take into account what we do know

OMES had a very difficult job. There’s little or no data available on most of the major pieces of Oklahoma’s justice system: information on controlling offenses, conviction rates, case dispositions, and sentencing patterns is nearly impossible to come by. As the footnotes to the OMES report make clear, even those within our most important justice agencies – the Administrative Office of the Courts, the District Attorneys Council, and the Department of Corrections – have limited ability to provide reliable data on what is going on in our justice system.

Nevertheless, the OMES calculation failed to take into consideration the best data that is available. From the OMES report itself, we know that about 1,500 people were received by DOC solely due to SQ 780 offenses in recent years. That’s similar to the number of people admitted whose most serious offense of Possession of a Controlled Dangerous Substance in recent years. It’s nowhere near the neighborhood of the 9,182 people OMES used, and it creates an unrealistic picture of what SQ 780 accomplished in its first year.

A more realistic calculation would estimate the number of people who would be convicted and sentenced to prison based on what we know. Based on recent caseloads, conviction patterns in other states, and research on prison sentences, we know that the number of people going to prison is far, far lower than OMES estimated.

The road ahead hasn’t gotten any shorter

SQ 781 directs OMES to “use actual data or best available estimates where actual data is not available” in calculating savings, and it directs that the savings be appropriated by the Legislature to a fund that is distributed to counties. It seems very unlikely, however, that the Legislature will find the $63.5 million OMES estimated that we saved to appropriate in the budget in the coming year, even with growing revenue.

There is still an enormous amount of work to be done to reduce our highest-in-the-world incarceration rate, and unfortunately, the OMES calculation sends a signal that Oklahoma has already come a long way. The Justice Reform Task Force bills passed last session are a great start, but they only reduce expected growth and won’t change our number one ranking. The worst outcome now would be complacency in the face of a towering problem that is just as immediate as ever.