Grits has spent a fair amount of the last year thinking about Fordham law prof John Pfaff's theories on mass incarceration - first in a couple of law review articles he wrote, then on his
Twitter feed, and finally in his recently published book:
Locked In: The True Causes of Mass Incarceration and How to Achieve Real Reform. I've
written before that his data observations hold true for Texas, almost
puzzlingly so, but his recommendations to me never seemed quite right. After reading the book-length exposition of his ideas, I now can say why.
Unfortunately, one can follow Prof. Pfaff's logic all the way to the end and, while much of it is thought provoking, still find it to be of little use to reformers seeking change on the front lines in the states, which is where he rightly insists the real action of criminal-justice reform takes place. I particularly regret coming to that conclusion, both because I've learned a lot thinking about the issues he raises and also because he was kind enough to thank your correspondent in the book's acknowledgements, as well as send me a free review copy. But at the end of the day, he's just wrong about where the American criminal-justice reform movement should focus and why.
Pfaff frames his critique as an antidote to what he calls the "Standard Story" - a straw man to which he attributes every wrong-headed statement he can find made by a reformer while ignoring or downplaying all areas in which their analysis makes sense.
His big critique is to claim that the Drug War is an insignificant contributor to mass incarceration, with Michelle Alexander's
New Jim Crow book coming in for particular criticism. Pfaff's position is that most offenders in prison are there for violent offenses, not nonviolent ones. So, he claims, reformers should focus mainly on sentencing policies for violent offenders. Pfaff considers reforms focused on nonviolent offenses basically small-time and mostly irrelevant, ignoring the Big Kahuna of violent offenders.
On its face, this rings true for Texas, where 56 percent of Texas prisoners were incarcerated for a violent offense in 2015, compared to only 16 percent for drug offenses. (
Source for all TDCJ stats in this post.) That's about the same as the ratios in the 30+ states (not including Texas) for which Pfaff analyzed data. But that's a misleading analysis because it focuses on a snapshot of the prison system instead of treating its population as dynamic and ever changing.
TDCJ released 70,311 inmates in 2015 and took in 69,066 new "receives," so there is a lot of churn. While the majority of offenders housed in TDCJ in 2015 were convicted of violent offenses, people convicted of violent crimes
made up just 23 percent of new admissions that year.
This is particularly ironic because Pfaff's other big criticism of reformers and the Standard Story is that we shouldn't be focused on population totals but reducing the total number of
admissions. But admissions growth has been in the nonviolent property/drug crime area which Pfaff insists reformers over-emphasize!
To be clear,
we cannot embrace Pfaff's two big recommendations simultaneously. If we focus on reducing admissions, we must focus on nonviolent offenses. If we focus on violent offenses, as he insists we must, then they're just a small fraction of annual admissions. These are contradictory suggestions. And they're his two main points.
Indeed, the quickest way to reduce admissions is to reduce penalties along the margin between the lowest-level felonies (in Texas, state jail felonies), and higher-level misdemeanors (in Texas, Class As). These are almost entirely nonviolent drug and property offenses. Pfaff supports these changes, because it would be ridiculous not to, but that stance only highlights the disconnect between the rest of his analysis and what clearly are the most effective near-term decarceration policies.
Pfaff discovered through analyzing state court filings that the average length-of-stay for people released from incarceration is going down, taking this to mean that long sentences aren't a major cause of mass incarceration. But those numbers are averaged and mask a more nuanced dynamic. Really, there are two categories of prisoners: Short-timers who account for the churn, and long-timers who have mostly been convicted of violent offenses and have quite long sentences.
In FY2015, for example, the 70,311 people released from TDCJ served an average of 4.5 years on an average sentence of 8 years. But those are averages. Many sentences are shorter, and the numbers are driven up by long sentences for a smaller number of violent crimes - a fact which complicates his "ignore sentence length" suggestion quite a bit. A whopping 63 percent of people convicted of violent offenses in TDCJ were in 2015 serving sentences longer than 10 years. By contrast, just 12 percent of new admissions had received sentences for longer than a decade. So if Pfaff wants reformers to focus mainly on violent offenses, they'd have to consider sentence length much more than he suggests.
Not only is Pfaff wrong to belittle efforts to reduce penalties for nonviolent offenses, he's also wrong that such efforts preclude focus on reducing incarceration of people convicted of violent crimes in state prisons. In Texas, the state
increased release rates for violent offenders at the same time the Legislature was focused on reforms aimed at nonviolent crimes. It's simply not been my experience that the rhetorical problems he hypothesizes from the ivory tower play out that way in the real world (a sentiment I've expressed to him in private correspondence a time or two).
Pfaff also underestimates, to the point of downplaying, the role the drug war played in reducing civil liberties protections for defendants across the board and incrementally increasing government power across an array of institutions, not just prosecutors (although they were a principal beneficiary). These policies don't directly affect the size of the carceral state in the same fashion as sentencing statutes, at least not in ways which can be measured, but indirectly they empower the prosecutors who Pfaff says are driving the train. The drug war and sex offenders are the two bugaboos which have been used to justify all manner of state empowerment at the expense of the individual, and those losses are as worthy of criticism and reactive reform as are the over-incarceration issues with which Pfaff is more concerned. IMO they're also more related in practice than he appears to realize.
None of which is to say his analysis is entirely wrong, just the political and policy recommendations he draws from them. One of Prof. Pfaff's best observations - and this is another area where Texas' data match his conclusions - was to show that much of the continued incarceration growth in recent years stems from prosecutors securing convictions more often for ever-more petty crimes, with the number of prosecutions per arrest doubling over the period he studied (from one in three to two in three).
That pattern holds for Texas as well.
It's how prosecutors and cops justified ever-growing budgets and staff in an era of declining crime: They just prosecuted more nonviolent offenses, particularly ones with no victims like drug crimes or weapons charges. It's the main reason, for example, why
indigent defense costs continued to rise even as crime declined.
Some of Pfaff's critiques of fuzzy-headed reformers are valid, but he more frequently pretends debates with substantial gray areas are black-or-white questions, even as he himself struggles mightily to suggest valid solutions which would have better outcomes than the ones he criticizes.
The solutions section of the book, in fact, is where Pfaff's critique comes completely apart, to my mind. Bottom line: It shows that the prosecutor-related issues he raises aren't really ones reformers are unwilling to address so much as an area where there aren't very many good reform options. He wants sentencing/plea bargaining guidelines, for example, but the federal sentencing guidelines caused incarceration to increase and there are no real-world examples where they've had the effect he hypothesizes.
IRL, many reforms aimed at prosecutors aren't focused on sentencing outcomes. The innocence movement has been successful in Texas at achieving reforms aimed at prosecutors, from legislatively mandating open-file discovery to requiring corroboration for informants to providing
habeas recourse for convictions based on junk science. Those focus on the accuracy of convictions, not sentencing. But improving accuracy assigns additional burdens to the state which may affect volume at the margins, both by avoiding inaccurate false positives and requiring more resources-per-case to prove. Innocence reforms can also have the benefit of affecting the culture of prosecutors' offices, reinforcing the "seek justice" part of their mandate over seeking convictions. Texas' experience shows they also can change the dynamics surrounding what voters demand from elected District Attorneys, which indirectly supports less punitive carceral policies.
Similarly, bail-reform legislation currently in play at the Texas Legislature could reduce prosecutor power in plea negotiations because more defendants won't be held hostage in jail until they agree. But those sorts of changes only indirectly (if still, probably, significantly) affect sentencing outcomes. It's not that reformers ignore prosecutors, but more that prosecutors are at best targets of opportunity. They are shielded from reform in many ways, and change can only happen in areas where one can secure a toehold. For that reason, one area where I strongly agree with him is the question of
improving metrics to judge prosecutor performance.
Pfaff correctly emphasizes that federal reforms won't do much and the real reform action must happen at the state and local level. Readers might be forgiven, though, if at times reading
Locked In, one mistakenly believed Pfaff were the first person to understand that dynamic. (The notion first occurred to your correspondent sometime in the '90s.)
Two suggestions by Prof. Pfaff have been under-emphasized in popular accounts of his work but would make a big difference. The first is federal funding of indigent defense. He's pointed out that for $4 billion, the feds could double indigent defense spending in America. Not sure that's possible under the current administration, but it's a worthy idea for down the line.
His other (to me) most other notable suggestion was offered as an aside in the recommendations chapter, but I think it has legs. He cited another academic paper (
see here) which suggested a Cap-and-Trade system for incarceration similar to the system used for pollution controls. Basically each county would receive an allocation of sentencing time, and if they wanted to punish people more they'd have to purchase it from other counties in a market setting. That would place financial responsibility for over-incarceration on the local actors most responsible for it. That's an awesome idea and deserves to be developed into a legislative proposal (not this year, though!). Grits will without question return to it later as it is a truly new solution to an old problem. But it's offered as a throwaway line at the very end of the book, it's not a central recommendation.
Bottom line: On the drug war, Pfaff is more right than Michelle Alexander, but less right than he thinks. When it comes to guiding reformers focused on decarceration, however, he is less right than those basing their proposals on what he disdainfully dubs the "Standard Story."
Grits appreciates Pfaff's work and analysis: We need more thoughtful quants looking at big-picture data patterns and testing old assumptions. But implementing reform in a real-life political landscape is a different matter entirely from crunching numbers with your grad students at Fordham. Prof. Pfaff suffers from more than a bit of hubris to imagine that his skill at the latter qualifies him to make strategic judgments regarding the former. Reformers may run across suggestions in the book (like the cap and trade idea) that make sense to pursue. But overall, they would be poorly served to adopt his policy priorities.