Mass Incarceration and Its Discontents

The contours of mass incarceration are, by now, broadly familiar. The U.S. incarceration rate began an unprecedented ascent in the 1970s. This trend continued through 2007, when 760 of every 100,000 U.S. residents—nearly 1 in 100 adults—lived behind bars, five million others were on probation or parole, more than ten million were booked into jail, and nearly one in three U.S. residents had a criminal record (Kaeble and Glaze 2016, Table 4; PEW Center on the States 2008; Sabol 2014; Subramanian et al. 2016). The scale of confinement now sharply differentiates the United States from comparable countries, where incarceration rates range from a low of 45 per 100,000 residents in Japan to 145 in England and Wales (Walmsley 2015). By 2015, the U.S. incarceration rate had fallen to 670 per 100,000 residents, a drop of nearly 12 percent (Kaeble and Glaze 2016). Still, the United States remains the world’s leading jailer (Wagner and Walsh 2016).

The emergence of mass incarceration in the United States has spawned a tremendous amount of social scientific research. ¹ A number of studies analyze its proximate causes and show that shifts in policy and practice (rather than rising crime rates) were the primary driver of penal expansion. Other studies analyze the consequences of mass incarceration, documenting, for example, its disparate and adverse impact on people, families, and communities of color. Some assess how penal expansion affects not only the incarcerated, but also those who are stopped, frisked, arrested, fined, and surveilled—even in the absence of incarceration or conviction. And a substantial body of research shows that penal expansion has had far-reaching sociological effects thattend to enhance—and mask—racial and socio-economic inequalities.

Although the decline in incarceration since 2007 has been modest, it has nonetheless triggered much discussion regarding the need for, and prospects of, reform. Yet researchers are debating more than the likelihood that meaningful change will occur; they also offer competing understandings of the problems that require attention and the solutions that should be enacted. The books reviewed here—Hard Bargains: The Coercive Power of Drug Laws in Federal Court, by Mona Lynch; Locked In: The True Causes of Mass Incarceration and How to Achieve Real Reform, by John F. Pfaff; and Sentencing Fragments: Penal Reform in America, 1975–2025, by Michael Tonry—speak to these pressing questions and offer surprisingly different ideas about what needs to be done to reverse mass incarceration and improve the quality of justice produced in American courts. In particular, and in contrast to the arguments of Lynch and Tonry, Pfaff makes the case that time served has not increased and therefore that efforts to enact comprehensive sentencing reform are misguided and would have little impact. In my view, this provocative claim is inconsistent with the best available evidence, much of which is brought to life in Mona Lynch’s Hard Bargains.

***********************************

Hard Bargains focuses on the prosecution of drug cases in the federal courts and thus appears to have a narrower focus than the other books reviewed here. Yet Lynch’s concentration on the comparatively small and somewhat unique federal system enables her to provide crucial insights about two key issues: the impact of tough sentencing laws on courtroom dynamics, adjudication processes, and sentencing outcomes; and the (varied) exercise of prosecutorial discretion in courts across the United States. Her findings help clarify what will need to be done not only to reduce incarceration, but also to protect defendants’ constitutional rights.

Lynch’s book is based on extensive ethnographic fieldwork conducted from December 2012 through July 2014 in three regionally distinct federal court districts. The case studies include observations of courtroom proceedings, interviews, and review of case files and other archival materials. Her findings show that although the Sentencing Reform Act (SRA) in 1984 and other tough drug-sentencing laws were intended to prevent the arbitrary exercise of judicial discretion, these laws mainly shifted discretionary power to prosecutors, whose charging decisions acquired extraordinarily important implications for sentencing outcomes and whose leverage in plea negotiations expanded dramatically.

This redistribution of the power to punish had, in turn, a number of important consequences. First, the empowerment of prosecutors triggered a dramatic uptick in federal drug filings, which more than tripled between 1980 and 2000. ² Second, rising caseloads led to institutional growth and increased resource allocation, which, in turn, enhanced the political power of what Lynch calls the “federal law enforcement machine.” Third, prosecutors came to use the threat of very long sentences to pressure defendants to waive their right to trial and accept comparatively unfavorable plea bargains. Indeed, many of Lynch’s case descriptions reveal in painful detail just how risky it has become for federal drug defendants to assert their right to a trial—and how the imposition of very long sentences is often used to punish those who dare to exercise this constitutional right.

And finally, both the likelihood of conviction and the average sentence length grew notably as a result of the adoption of tough drug laws in the 1980s. The share of federal drug defendants who were convicted increased from 82 percent in 1985 to 93 percent in 2009 (Raphael and Stoll 2013:65). Over this same time period, the proportion of federal drug defendants who were sentenced to prison (as opposed to probation) rose from 75 to 90 percent (ibid), and the average federal drug prison sentence increased from 58 to 84 months (Raphael and Stoll 2013:66). ³ As Lynch shows, the adverse impact of these shifts fell disproportionately on African Americans for two reasons: prosecutorial charging practices ensured that white crack users and dealers were rarely prosecuted in federal courts, and the crack-powder sentencing disparity embedded in the Anti-Drug Abuse Act of 1986 ratcheted up penalties for federal crack defendants, the overwhelming majority of whom have been black.

In short, the federal government’s enactment of tough drug sentencing laws notably enhanced prosecutors’ power to punish, which in turn had important consequences for prosecutorial filing practices, the power of the federal law enforcement apparatus, the capacity of federal defendants to assert their rights, the conviction rate, and sentencing outcomes. At the same time, Lynch shows that there is significant variation in the degree to which these shifts have occurred and in the mechanisms that produce them. In the Southeast District, for example, prosecutors place tremendous value on securing offers of substantial assistance (i.e., cooperation) from defendants and reward those who acquiesce—while strenuously punishing those who decline to cooperate. By contrast, in the Southwest, early prosecutorial efforts to secure defendants’ cooperation generally failed due to the even greater threat of (lethal) danger from the drug cartels to anyone who “snitched.” Prosecutors in the Southwest District therefore rely on the “exploding offer,” in which plea deals expire quickly, to pressure defendants to accept relatively unfavorable guilty pleas.

Lynch’s conclusion offers a thoughtful discussion of how justice might be improved. Ideally, Lynch argues, the prosecutorial power to file charges would be regulated. However, because this power in nearly “inviolate” in the U.S. legal system, Lynch concludes that the prospects for directly restraining prosecutors from filing charges in low-level drug cases are dim. Similarly, although the Supreme Court’s ruling in United States v. Booker ⁴ authorized judges to depart from federal sentencing guidelines, this ruling has had little impact in courts where judges remain loyal to the sentencing guidelines. For these and other reasons, Lynch emphasizes the need for sentencing reform aimed at reducing maximum sentences in order to diminish prosecutorial power and reduce sentences. Lynch also recommends revising the statutory “hammers,” including criminal history calculations, career offender guidelines, and protected zone enhancements, that dramatically increase potential sentences and are used by prosecutors to compel guilty pleas and punish federal drug defendants, particularly those who reject plea deals, harshly.

In short, Hard Bargains offers an empirically grounded and insightful analysis of how tough federal sentencing laws shape the balance of power in the courtroom, the nature of the justice process, prosecutors’ informal practices, and sentencing outcomes. Lynch’s analysis also has important conceptual implications, suggesting that policy and practice should be conceived as intertwined and mutually constitutive phenomena. For example, changes in sentencing law triggered an uptick in case filings even though the relevant statutes did not require this. Conversely, the “law enforcement machine” that resulted from rising drug caseloads is now an influential opponent of sentencing reform in Congressional debates. In Lynch’s account, then, sentencing law matters directly but also has a host of consequential indirect effects, including the enhancement of prosecutorial power. Lynch’s discussions of individual cases, involving actual human beings in the throes of an often Kafkaesque federal machinery, bring vivid texture to her broader arguments and underscore the human costs of the system. Although Lynch acknowledges differences in the laws, cases, and caseloads that differentiate state and federal systems and highlights the varied exercise of prosecutorial discretion, her analysis nonetheless sheds important light on the indirect and radiating effects of tough sentencing laws and the prosecutorial power they enhance.

***********************************

In Sentencing Fragments: Penal Reform in America, 1975–2025, Michael Tonry makes a similarly strong case for sentencing reform, albeit one that rests on a very different kind of evidence and reasoning. Tonry, a legal scholar with expertise in international sentencing policy, presents his argument in a characteristically direct manner: “No one admires American sentencing systems. They are arbitrary and unjust, they are much too severe, they ruin countless lives, and they have produced a shameful system of mass incarceration” (p. vii). The chapters provide a historical overview of the evolution of U.S. sentencing policy, the creation of (flawed) federal sentencing guidelines, the normative underpinnings of U.S. sentencing policies, and alternative sentencing principles that inform criminal justice practices in other countries. Throughout, he draws on secondary sources to support and illustrate his key points.

For Tonry, the need for comprehensive sentencing reform is indisputable: “No meaningful progress will be made in reducing mass incarceration, rampant racial and ethnic disparities, or injustice in individual cases until sentencing laws and practices are overhauled” (p. 2). Tonry acknowledges that many states have enacted legislation aimed at reducing incarceration in recent years but emphasizes that only Michigan has repealed a major mandatory minimum, three-strikes, life without parole, or truth-in-sentencing law. The majority of the enacted reforms, he argues, “only nibble around the edges” of mass incarceration; what is needed is repeal of draconian sentencing laws, the enactment of presumptive sentencing guidelines, and the restoration and revitalization of parole.

Tonry begins by providing a historical overview of the evolution of twentieth-century sentencing policy. For most of the twentieth century, indeterminate sentencing existed in all states. Within this framework, judges imposed extraordinarily open-ended sentences under the theory that rehabilitation is possible and that parole boards are best able to determine when an incarcerated person is safe to release. From Tonry’s perspective, this system was imperfect but generally produced acceptable results. Prior to the 1980s, for example, the U.S. incarceration rate was typical of those found in other industrialized democracies. Recidivism rates were similar to those found today. And whileparole boards sometimes made arbitrary and/or racist decisions, racial disparities in incarceration, at least by some measures, were less pronounced than is currently the case.

The situation changed dramatically after 1975, when states and the federal government radically overhauled their sentencing policies. Until the mid-1980s, most such reforms targeted indeterminate sentencing policies’ perceived flaws, including racial disproportionality and the lack of consistency in outcomes. After the mid-1980s, however, sentencing reforms increasingly focused on enhancing penal severity rather than redressing inequities. Indeed, in recent decades, all states enacted mandatory minimums, nearly all adopted statutes authorizing life without parole (LWOP) sentences, and more than half adopted three-strikes and truth-in-sentencing (TIS) laws.

As noted previously, Tonry argues that such changes in U.S. sentencing policy have largely created mass incarceration (p. 12). Tonry supports this argument by providing evidence that people convicted of felonies in the United States are far more likely to besentenced to confinement than is the case in other countries and that U.S. prison sentences are extraordinarily long compared to those imposed in other democracies. Tonry’s deep familiarity with criminal law and policy outside the United States is helpful here, and the comparisons he provides usefully illustrate his key points. For example, we learn that the share of convicted felons sentenced to confinement in select European counties ranges from a low of 3.1 percent in Finland to a high of 23 percent in the Netherlands (p. 26). By contrast, in the United States, 73 percent of people convicted of felonies in 2009 were sent to prison or jail. We also learn that fewer than 10 percent of all felony cases adjudicated in other western democracies result in a prison sentence longer than one year, and only 1 to 3 percent are sent to prison for more than five years. Although the data Tonry provides do not allow for direct comparison, the average prison sentence imposed by U.S. state courts is undoubtedly far longer. In 2009, for example, the average prison sentence was 4.3 years; for violent crimes, it was 7.5 years (p. 29).

It is clear, then, that confinement is imposed in a much larger share of cases and that the sentences enacted in the United States are markedly longer than those levied in comparable countries. Although Tonry further contends that U.S. sentences have increased notably as a result of the enactment of tough sentencing laws, he cites just one study on this point, which shows that the amount of time prisoners stayed behind bars tripled from 1977 and 1988 (p. 25). Given the adoption of harsh sentencing laws by most states since this time, Tonry reasons that sentences (and time served) have likely increased far more since that time. ⁵ In addition, Tonry cites several studies that show that the enactment of tough sentencing laws, especially mandatory minimums andtime-served requirements for violent offenders, are significant predictors of comparatively high incarceration rates across the states (see especially Spelman 2009; Stemen, Rengifo, and Wilson 2006). Although a host of non-policy-related factors, including the size of a state’s black population and its fiscal capacity, also influence incarceration rates, these studies indicate that tough sentencing laws, and especially truth-in-sentencing requirements, have contributed to mass incarceration.

For Tonry, the fact that more people are being sent to prison for longer periods of time is a travesty, for a number of reasons. First, people sentenced to prison are more likely to reoffend than people who are punished in other ways. Second, long prison sentences do not deter more than short ones, but are far costlier in both fiscal and human terms. And third, although long prison stints do incapacitate, this often isn’t necessary because nearly all of the justice-involved are safe to release after reaching middle age.

Tonry provides a number of useful suggestions regarding alternative practices and policies that could (and should) be adopted in order to improve justice and reduce mass incarceration. Most crucially, he contends, sentencing laws should be modified and statutory maximums should be reduced to ensure that fewer people are sent to prison and for shorter periods of time. Release mechanisms (such as parole) must be returned and/or revamped to ensure that people currently serving long or life sentences have a meaningful chance to be considered for release. Tonry also identifies a number of police and prosecutorial practices that could, under existing law, be modified to reduce incarceration and provides examples from abroad to illustrate how these alternative practices might work. For example, police departments could reduce the flow of people into prison (and jail) by minimizing arrests for low-level crimes and enhancing access to diversion and restorative justice alternatives, as they do in some European countries. ⁶ Similarly, prosecutors’ role should be reimagined to promote these and other harm reduction and restorative justice alternatives.

Although Tonry provides many inspiring examples of how such policies work in other countries, he does not discuss how these changes in sentencing policy and criminal justice practice can be achieved given the decentralized and politicized nature of the U.S. criminal justice system. Instead, he provides a compelling vision of what an alternative system might look like, how sentencing law and policy ought to be reformed, and how a wide range of criminal justice actors might reconceive their roles in order to reduce reliance on prison and jails and enhance justice.

***********************************

In Locked In, John Pfaff reaches very different conclusions about the creation and potential dismantling of mass incarceration. To begin, Pfaff contends that there is a dominant but incorrect narrative regarding the causes of mass incarceration, which he calls “the Standard Story.” According to Pfaff, this Standard Story is leading advocates to press for largely toothless reforms. As he puts it, “Reformers still don’t understand the root causes of mass incarceration, so many reforms will be ineffective, if not outright failures” (p. viii).

In Part I, Pfaff makes the case that reformers are focusing on three developments that are not, in fact, the main drivers of mass incarceration: the War on Drugs, tough sentencing laws, and the emergence of private prisons. In Part II, Pfaff identifies what he believes is the true cause of mass incarceration—the increased propensity of prosecutors to file felony charges given arrest—and advances a reform agenda that, he argues, would lead to more meaningful change by reducing felony filings. This agenda includes a long list of reforms, such as improving data collection regarding prosecutorial decision-making, altering the financial incentive structure that currently encourages (county) prosecutors to send people convicted of felonies to (state) prison, and rezoning in order to reduce the power of suburban voters in prosecutorial and judicial elections. He also recommends a series of reforms aimed at addressing the “political defects” that, he argues, lie at the root of prosecutorial excess. Perhaps the most important of these is that prosecutors be appointed rather than elected.

Some of his reform proposals are intriguing, and Pfaff is undoubtedly correct to emphasize the role of prosecutors in the prison build-up, as a number of other analysts also have, including Lynch (see also Barkow 2013; Davis 2007; Gottshalk 2015). I also agree with his argument that successful reform efforts will need to address the problem of violence, as other researchers have also emphasized (e.g., Beckett, Reosti, and Knaphus 2016; Clear and Frost 2013; Forman 2012; Gottshalk 2015; Travis, Western, and Redburn 2014). However, Pfaff’s analysis is undermined by methodological flaws, logical errors, and conceptual limitations. These weaknesses significantly erode confidence in some of his other conclusions.

From my perspective, the most fundamental problem is Pfaff’s claim that efforts to enact sentencing reform are misguided. This argument rests on several inaccurate claims, the first and most important of which is that “there’s not a lot of evidence that the amount of time spent in prison has changed that much—not just over the 1990s, 2000s, and 2010s, but quite possibly over almost the entire prison boom” (p. 6). All other studies of which I am aware reach the opposite conclusion (see, for example, Blumstein and Beck 1999, 2005; Cohen and Canela-Cacho 1994; Courtney et al. 2017; PEW Center on the States 2012; Raphael and Stoll 2009, 2013; Travis, Western, and Redburn 2014; Western 2006).

Pfaff’s divergent findings regarding time served appear to reflect his use of a flawed measure and his reliance on an unusually small and unrepresentative state sample. Time served is most directly captured by observing the amount of time between the admission and release of actual cohorts of prisoners, and Pfaff adopts this approach. But inmates with long or life sentences are rarely released and are therefore undercounted in this observational measure (PEW 2012). Moreover, the lifer population has increased dramatically in recent decades, and one in seven prisoners is now serving a life sentence (Nellis 2017). While the majority of these prisoners are eligible for parole, an increasing number and share are not. Moreover, time served has increased dramatically for lifers who do have the opportunity to go before a parole board (Ghandnoosh 2017). As a result of these trends, direct observation yields underestimates of time served, particularly in more recent years. Indeed, Patterson and Preston (2008) show that the observational measure of time served employed by Pfaff is the least accurate of the available options.

Other researchers have used a variety of alternative methods to estimate the amount of time that cohorts of prisoners will spend behind bars in ways that capture information about the growing number of people serving long or life sentences and take fluctuations in prison admissions and releases into account. These studies also include data from significantly more states than Pfaff’s analysis does. For example, Courtney et al. include 44 states; the PEW (2012) study of trends in time served draws on data from 35 states; and Raphael and Stoll’s (2013) analysis utilizes a 34-state sample. By contrast, in Locked In, Pfaff presents the results of an analysis of a 10-state sample in which, as he acknowledges, southern states are notably underrepresented (p. 253, fn. 16). ⁷

Moreover, the analysis of trends in time served presented in Locked In is limited to the period from 2000 to 2010. This is an odd choice, as it appears that most of the increase in time served occurred in earlier decades (Raphael and Stoll 2013). Most studies also analyze time periods that are far longer than a single decade because, as Beck and Blumstein note, “The contribution of long sentences to rising incarceration rates can be fully observed only over a very long period” (1999:52).

Studies that adopt these and other methodological precautions consistently find that time served did increase notably in recent decades, especially in the 1990s. ⁸ For example, Raphael and Stoll (2013:51) find that average time served for murder and negligent manslaughter increased by 55 percent, for robbery by 44 percent, and for burglary by 21 percent between 1984 and 2004. Western (2006) reports that time served for violent, property, and drug offenses increased by 61, 75, and 71 percent, respectively (p. 45, Table 2.3). Blumstein and Beck (1999, 2005, and in Travis, Western, and Redburn 2014) report similar findings for the period from 1980 to 2010. And PEW (2012) found that time served between 1990 and 2009 increased by more than one-third for people convicted of all offense types. ⁹ Surveys of state prisoners similarly reveal notable increases in expected time served in the 1990s. ¹⁰

Pfaff largely ignores these studies, but when he does acknowledge them, he argues that they are methodologically flawed. His arguments in this regard are unconvincing and, in some cases, misleading. For example, Pfaff dismisses PEW’s (2012) finding that time served increased by more than one-third between 1990 and 2009 by suggesting that 1990 was an aberrational year (p. 254, fn. 22). This critique ignores the fact other studies employ different start dates and that PEW used average time served in 1989, 1990, and 1991 as the benchmark against which later years were compared (PEW 2012:50). Pfaff also argues that although the reported increase in sentence length of “more than one-third” sounds like a lot, it isn’t, and that the impact of such increases on macro-level trends is slight (pp. 58–59). He provides no evidence to support this claim and ignores the many other studies that have found that increases in time served have contributed importantly to the growth of the prison population (Blumstein and Beck 1999, 2005; and in Travis, Western, and Redburn 2014; Raphael and Stoll 2009, 2013; Sabol 2010; Western 2006).

In addition to dismissing abundant evidence that time served has increased, Pfaff also supports his argument that sentencing reform efforts are unnecessary by claiming that they would be less effective than efforts to modify the exercise of prosecutorial discretion such that fewer people were sent to prison in the first place (p. 52). This claim implies that prison admissions are unaffected by sentencing policies, when in fact tough sentencing laws often mandate a prison (as opposed to a jail or probation) sentence. Even setting this issue aside and assuming that prison admissions result solely from (informal) prosecutorial decision-making, Pfaff’s claim that reforms aimed at reducing length of stay will be less effective than those aimed at reducing prison admissions is inconsistent with the available evidence. For example, according to an online prison population forecasting tool created by the Urban Institute, which Pfaff cites favorably in other contexts, reducing prison admissions for violent crimes by 25 percent would yield a 7 percent decline in the state prison population—as would a 25 percent decline in time served for violent offenses. Similarly, reducing prison admissions and length of stay for all offenses by 25 percent would yield declines in the prison population of 19 and 18 percent, respectively. ¹¹ The evidence thus does not support Pfaff’s claims that sentencing reforms aimed at shortening length of stay would be notably less effective than efforts to reduce prison admissions. Clearly, both are warranted—and the repeal of draconian sentencing laws would likely work both directly and indirectly to reduce prison admissions for lower-level offenses that might, in the absence of such laws, yield a probation or jail sentence.

Pfaff’s argument that sentencing reform is unnecessary also rests on his problematic assertion that prosecutors do not actually sentence people to the longest possible sentences because they understand that to do so would increase state correctional costs to unacceptable levels (p. 55). In fact, it is clear that many prosecutors do invoke tough sentencing laws when they are available, although there is significant geographic variation in this practice, as Lynch shows (see also Chen 2014). In California, for example, state courts sent more than 80,000 second strikers and 7,500 third strikers to prison between 1994 and 2004 (Brown and Jolivette 2005). By the end of 2004, strikers made up more than one quarter—26 percent—of the California prison population, and the fact that they were sentenced under the three-strikes law often added many years, and often decades, to their sentences (ibid).

In the federal system, 46 percent of 100,000 federal drug prisoners were sentenced under mandatory minimum sentencing laws (Horwitz 2015). More generally, by 1998, more than two-thirds of people admitted to U.S. state prisons for a violent offense had been sentenced under a TIS law that required that they serve at least 85 percent of their sentence (Ditton and Wilson 1999), and there is evidence that these statutes increased bothtime served and prison populations (see Tonry 2016:78–83). Similarly, the number of prisoners serving life-without-parole (LWOP) sentences skyrocketed as states expanded their LWOP statutes (Nellis 2017). Thus, although some prosecutors do seek to avoid triggering tough sentences, many others eagerly use these laws where they are available.

Moreover, as Lynch shows, harsh sentencing laws have important indirect effects, such as enhancing prosecutorial power and leverage, even if the majority of defendants do not receive the maximum allowable sentence. The research on TIS laws, which require that people serve a certain proportion of their sentence (usually 85 percent) behind bars,also illustrates this point: the mere existence of these laws increases sentences even when defendants plead guilty to lesser crimes that are notsubject to TIS laws (Owens 2011). Similarly, Sutton (2013) finds that California’s Three Strikes law generally enhanced felony sentences, especially in conservative counties and for black defendants. Lynch’s analysis of courtroom dynamics illuminates why this is the case: even if defendants are not charged under harsh sentencing statutes, the existence of those statutes enhances prosecutorial power in plea negotiations, which yields longer average sentences.

Finally, Pfaff’s argument that sentencing reform is unnecessary because prosecutors do not seek long sentences for fear of driving state correctional costs up is in tension with his own claim that prosecutors created mass incarceration through a shift in their filing decisions. If prosecutors are willing to crowd state prisons by filing dramatically more felony charges, why would they refrain from imposing harsh sentences out of concern for state budgets?

In short, Pfaff’s claim that sentencing reform is unnecessary is unconvincing at best. This argument rests mainly on the erroneous claim that time served in prison has not increased in recent decades, a result he apparently obtains by using a measure that is known to yield underestimates of time served, particularly as the lifer population has grown, and by relying on a small andunrepresentative state sample. His discounting of the well-documented growth in life sentences, including life without parole, ¹² is especially disheartening, as these sentences not only contribute to mass incarceration in a quantitative manner but also raise crucial ethical, humanitarian, and human-rights concerns. And because he ignores laws’ radiating and indirect effects, he also fails to appreciate the many other ways that tough sentencing laws have enhanced prosecutorial power, altered courtroom dynamics, undermined rights, and weakened justice.

The flip side of Pfaff’s argument that tough sentencing laws (as well as the War on Drugs and the growth of private prisons) are relatively inconsequential is his insistence that he alone has identified the true engine of the prison build-up: the increased proclivity of prosecutors to file felony charges. As he puts it, “The only thing that really grew over time was the rate at which prosecutors filed felony charges against arrestees” (p. 72). Some background will help explain how Pfaff arrives at this conclusion. Many analysts, including Pfaff, have decomposed the criminal justice process into discrete stages in order to trace shifts in the criminal justice system response to reported crimes. Most such studies examine changes in crime rates as well as the arrest-to-crime ratio, the prison-admission-to-arrest ratio, and time served (see, for example, Blumstein and Beck 1999, 2005; and in Travis, Western, and Redburn 2014; Raphael and Stoll 2009, 2013; Sabol 2010; Western 2006). As noted earlier, these studies indicate that time served has increased, particularly in the 1990s. These studies—as well Pfaff’s research—have also found that the prison-admission-to-arrest ratio increased notably after the early 1980s (ibid).

Most researchers attribute the increased likelihood that an arrest for a felony offense would trigger a prison sentence to changes in sentencing policy, which may, for example, mandate that a second-time drug possessor serve prison time. By contrast, Pfaff attempts to unpack this box and disentangle the impact of shifts in felony filing decisions, the conviction rate, and the likelihood of a prison admission given conviction. This is a laudable goal. Toward this end, he uses (and advocates the use of) National Center on State Court filing data: “This dataset, sitting in plain view on an NCSC server but apparently overlooked by all studies before my own, provides a rare window onto how prosecutorial behavior has changed” (p. 71). Using these data, Pfaff calculates the ratio of felony filings to felony arrests and determines that this ratio increased notably between 1994 and 2008 (p. 71).

Neither in his book nor in the article on which it is based does Pfaff discuss these filing data and their limitations in any detail. ¹³ I obtained the same NCSC dataset provided to Pfaff from NCSC staff (it was not readily available on the NCSC server) and learned that the data for 16 of the 34 states for which data are available are identified by NCSC aseither over-inclusive, under-inclusive, orboth, and the NCSC is unable to provideinformation about the magnitude of the many errors contained in the data. ¹⁴ Although many, perhaps all, large data sets possess limitations, Pfaff neither acknowledges these limitations nor provides any information about them, and he includes the filing data for all 34 states in his analyses (p. 257, fn. 51; see also Pfaff 2011b:12). As a result, it is difficult, perhaps impossible, to assess the reliability of Pfaff’s findings regarding the apparent increase in the filing-to-arrest ratio. Nor can we assess his subsequent claim that the likelihood of prison admission given a felony filing did not increase and hence that prosecutorial filing decisions are the only important driver of mass incarceration.

Pfaff also supports his claim that prosecutors’ increased propensity to file charges was the only criminal justice outcome that changed in recent decades by asserting that the conviction rate has been “fairly stable over time” (p. 73 and p. 258, fn. 58). But the figures he cites to support this claim include people who were convicted of both felonies and misdemeanors, ¹⁵ while his analysis focuses on the drivers of prison growth, which requires felony conviction. If we focus specifically on the felony conviction rate, Bureau of Justice Statistics data show that the share of (urban) defendants who were charged with, and subsequently convicted of, a felony increased from 50 to 59 percent between 1990 and 2004. Similarly, the share that were convicted of the original felony charge rose from 43 to 51 percent. ¹⁶ The available evidence thus suggests that the share of felony filings that resulted in a felony conviction did increase and therefore thatprosecutors’ filing decisions were not the only cause of the increase in the arrest-to-admission ratio. It is also worth notingthat this increase in the felony conviction rate may well reflect the impact of tough sentencing laws, which enhance prosecutorial power in plea bargaining negotiations.

In short, Pfaff’s claim that time served in prison has not increased and is not an important contributor to mass incarceration rests on deeply flawed methods and logic. Conversely, his conclusion that the proclivity of prosecutors to file felony charges given arrest is the only thing that really changed (and hence, by implication, is overwhelmingly responsible for mass incarceration) rests on an analysis of filing data with significant reliability issues and overlooks evidence that the felony conviction rate did, in fact, increase during the prison build-up. Although prosecutors are undoubtedly important and powerful actors, the argument that mass incarceration has a single cause—the increased propensity of prosecutors to file charges given arrest—is simply inconsistent with the best available evidence.

***********************************

The U.S. penal system is composed of myriad complex, decentralized, and highly politicized institutions. Its dramatic expansion has not one cause, but many; shrinking and transforming it will require multi-faceted strategies that address its varied drivers, including the increasingly tough policy response to violence as well as the political power of many criminal justice actors. Although sentencing policies are not the sole cause of mass incarceration, there is significant evidence that they matter both directly and indirectly for incarceration rates and for the quality of justice meted out in our courts. Indeed, it is crucial that we recognize the complex interactions between sentencing law, prosecutorial practices, and correctional decision-making, about which Lynch’s analysis lends significant insight. Either/or thinking is unhelpful here: both prosecutorial power and the sentencing laws that enhance it, as well as numerous other forces and dynamics (such as the abolition of parole in some states and the increasing risk-aversion of existing parole boards), will need to be addressed in the effort to create a less incarcerative and more just society. The analyses presented by Lynch and Tonry shed significant light on the challenges before us and provide useful ideas about how various reform strategies might usefully reinforce each other. They also make clear that comprehensive sentencing reform that reduces maximum sentences, reverses truth-in-sentencing requirements, and revitalizes parole is essential if we are to make progress in redressing one of the most pressing human-rights challenges of our time.