Principles of Prosecutor Lenience

Article - by Volume 102 - Issue 7: Mercy Symposium

Once “the Darth Vader of academic writing,”[1] American prosecutors are making a comeback. In recent years, “progressive prosecutors” have leveraged prosecutors’ one true superpower—lenience—to “reform the criminal justice system from the inside.”[2] There is so much scholarly enthusiasm for this project that the existing commentary can be summarized as offering a one-word principle to govern considerations of prosecutorial lenience: yes.[3]

But there is surely more to say.[4] American criminal law covers a broad array of offenses with vast differences in punitiveness across jurisdictions and courts. And even harsh critics of the system’s severity tend to pivot when it comes to certain offenses, like crimes committed by police.[5] Plus, some scenarios call for lenience but not necessarily prosecutor lenience. Prosecutors may be poorly positioned relative to police, judges, legislators, and governors for some exercises of lenience, and poorly suited by demeanor or experience to recognize the need for “equitable discretion.”[6]

Lenience can certainly be problematic. For example, in 1204, England’s King John freed “all prisoners, whatever the cause for which they may have been detained, whether for murder, felony, or larceny, or breaking the forest laws, or for any other accusation brought against them whatsoever.”[7] The blanket pardon was intended to engender “the love of God . . . [for the] salvation of the soul of our dearest mother.”[8] Jews were ineligible.[9]

While there is broad support for lenience, few commentators would support a proclamation like King John’s, particularly if it came from the local district attorney. I can say this with confidence, having recently suggested that President Biden commute the sentences of “over 10,000 federal prisoners.”[10] The “law and order” crowd watched from the sidelines as progressives blasted the proposal.[11] I suggested that Biden commute sentences not just of thousands of people convicted of federal drug, immigration, and gun crimes, but also reach across the aisle to include a handful of January 6 rioters “who are not accused of assaulting or interfering with police.”[12] By doing so, “Biden could signal the need to return to the way we used to think about mercy—as applying even to those who deserve punishment.”[13] The New York Times Pitchbot, Pod Save America, and everyone on the Internet objected . . . strenuously.[14]

All of which is to say that there are profound questions about the when and why of lenience, and particularly prosecutor lenience. The answers speak to one of the great mysteries of American criminal law: the role of the prosecutor.[15] I have taken on this mystery in recent years and continue the effort here by offering a skeletal framework for prosecutor leniency.[16] The framework proposes three principles of prosecutor lenience. Prosecutor lenience should be (1) non-arbitrary, (2) equal, and (3) abundant. Each principle is summarized below and then explained more thoroughly in later discussion.

The first principle (non-arbitrary) requires the most unpacking. Given the immense factual variation and countless considerations, there is no precise formula for evaluating arbitrariness. But it can be evaluated by assessing whether the justification for lenience is reversible: i.e., would the reason for lenience support severity, if reversed? For example, just as a criminal record or use of violence supports increased severity, these factors’ absence supports lenience. By contrast, lenience based on a defendant’s physical attractiveness or campaign contributions would fail this test because the reverse—more harshly punishing unattractive defendants who contributed to the opposition—would be unpalatable.

Lenience should also be applied equally. Assuming a prosecutor can justify a specific application of lenience, that prosecutor should be able to explain why lenience was not offered to another defendant in a seemingly analogous case. Few will object to this principle, which is in essence, a subset of non-arbitrariness. The challenge of equality is application. There are thousands of individual prosecutors facing an endless variety of cases and considerations.[17] Guidelines and transparency are the natural response.[18] But as foreshadowed by the federal sentencing guidelines, efforts to increase equality may decrease the overall quantity of lenience (principle three).

Finally, prosecutorial lenience in the United States should be abundant. This consideration is contingent and requires some justification. In part it speaks to the important question: why prosecutors? Prosecutors are part of a larger system that includes other actors who might be better positioned to exercise lenience. Prosecutor lenience may not even be needed if the overall system is just: for example, a system with reasonable, evenly enforced criminal laws, impartial factfinding, moderate penalties assessed by independent judges, a well-functioning parole system, and a robust clemency apparatus. But that is not an accurate description of much of criminal law enforcement in the United States.[19] American prosecutors work against a backdrop of severe criminal laws, uneven enforcement, sometimes mandatory and frequently overly punitive sentences, coercive plea bargains, limited parole release, and rare commutations.[20] If one accepts this characterization, it follows that prosecutors should embrace lenience. That’s the best answer to the question: why prosecutors? Lenience may theoretically be best offered by other actors, such as a legislature who can best reform draconian laws, a judge who can impose reasonable alternatives to incarceration, or a governor who could remedy individualized injustices through commutations. But if those actors cannot or will not take on that role, it becomes hard to object when prosecutors fill the void. Moreover, if American legislators are creating an overly punitive default system with the expectation that prosecutors will blunt its sharp edges, resisting prosecutor lenience undermines legislative intent.[21]

This Symposium Essay explores prosecutorial lenience through the lens set out above. Part I defines prosecutorial lenience and proposes three principles to guide its exercise. Part II applies the principles to common prosecutorial lenience scenarios like insufficient evidence, justice-based lenience, transactional lenience, triage, nullification, and mercy. The analysis is necessarily incomplete and tentative. But it reveals that some forms of prosecutorial lenience are more easily justified than others and offers a rough outline for exercising lenience within each category. Part III highlights the dilemma that, in some circumstances, the principles conflict. Specifically, insisting on the first two principles may jeopardize the third. This means that prosecutors, and their critics, will have to consider not just the overall desirability of lenience, but tradeoffs between the quality of prosecutorial lenience and its quantity.

I. Lenience Principles

This Part defines prosecutorial lenience and then posits three principles that can aid in its exercise. These principles can separate desirable from undesirable forms of lenience. In addition, lenience principles can reduce bias and conceptual errors that will otherwise be inevitable in highly pressurized, ad hoc decision-making.

A. Defining Lenience

For purposes of this Essay, I define prosecutorial lenience to include any scenario where a prosecutor can select from alternatives and chooses a less severe option.[22] This definition does not differentiate between any of the potential motives for choosing lenience except in one respect. It only reaches scenarios where the prosecutor can lawfully pursue a more severe option and chooses not to. If the law dictates lenience, for example when there is not probable cause to pursue a charge, then there may be lenience. But it is not prosecutorial lenience. [23]

The clearest examples of prosecutorial lenience arise when the prosecutor declines to formally charge a case. Other examples include decisions to charge a misdemeanor instead of a felony, or to charge fewer counts, offenses, or enhancements than the facts and law permit. Beyond the charging decision, prosecutorial lenience arises when the prosecutor dismisses previously filed charges. This lenience can occur unilaterally or in the context of a non-prosecution agreement, or a plea bargain. Finally, prosecutorial lenience can surface in bail or sentencing recommendations as well as efforts to void convictions.

B. Avoiding Arbitrariness through Reversibility

The first principle of prosecutor lenience is that it should not be arbitrary.[24] A standard definition of arbitrary action is an action that is irrational, “without adequate determining principle,” and “not done or acting according to reason or judgment.”[25] Given the broad array of considerations that come into play in any criminal case, a prosecutor will usually be able to point to some reason for an action. The challenge is evaluating that reason in a manner that is not completely subjective.

One way to test rationality is to assess whether a reason cited for lenience is reversible. Lenience and severity are not distinct concepts but opposite directions on a single line. Consequently, prosecutors who seek severity based on a particular factor should recognize that lenience is called for when that factor is reversed. By the same token, if a prosecutor cites a consideration to support lenience, the absence or reversal of that consideration should support severity.

The reversibility inquiry will not resolve the question of arbitrariness, but it offers structure to an otherwise amorphous and subjective exercise. There is, of course, the subsidiary question of the degree of lenience. In theory, this exercise can also determine degree. If a consideration supports lenience to a certain degree, its absence will support severity to that same degree. But it is foolish to pretend that this is math. There will always be disagreement about whether and to what degree a particular consideration warrants lenience. As Josh Bowers has pointed out, “No one can know such things.”[26] For many questions, particularly those governed by “justice” (discussed below), all commentators can offer is general guidance.

C. Equality

Lenience should be applied equally.[27] Stated more precisely, if a prosecutor’s office offers lenience to one defendant, that same leniency should be available to similarly situated defendants. While the concept seems uncontroversial, ensuring equal lenience presents a logistical dilemma.

One complication in ensuring equal lenience is identifying the institution across which we should expect consistency. Most obviously, we can expect a single prosecutor to act consistently across defendants. This is also true for a single prosecution office. Beyond that, the expectation of consistency seems unreasonable since inconsistency is a necessary consequence of empowering elected district attorneys to direct their office’s policies.

Ensuring equal treatment across a single prosecution office is itself a daunting challenge, particularly for larger offices with hundreds of prosecutors and frequent turnover. District Attorneys must delegate prosecution tasks, resulting in a principal-agent dilemma.[28] One solution is to enact guidelines for line prosecutors. The more precise the guidelines, the more likely they will ensure equal treatment of similarly situated defendants. Guidelines may also serve to promote lenience in offices where line attorneys are less lenient than the chief prosecutor.

For progressive district attorneys seeking to differentiate themselves from their predecessors, guidelines have become widely promoted lenience blueprints.[29] These guidelines serve multiple purposes. First, they ensure consistency. By issuing internal guidelines, chief prosecutors can standardize and control the actions taken by subordinates. Second, by publicizing the guidelines, district attorneys deputize interested outside parties (e.g., defense attorneys) who can rein in incompetent or rebellious prosecutors by bringing deviations to the attention of prosecutorial supervisors, increasing abundance.

The problem with publicized guidelines is that they can be criticized as an enticement to commit certain (non-prosecuted) crimes and serve as a formal admission of lenience that political opponents can attack. In fact, prosecutorial guidelines fall into the political trap that legislators have set for decades: passing broad sweeping criminal laws and then blaming police, prosecutors, and judges for non-enforcement when crimes continue or the laws are applied in an unpopular manner.

D. Abundance

The third principle of prosecutorial lenience—abundance—is contingent on time and place. In modern America, there is a strong argument to encourage prosecutorial lenience because the system has, over time, steadily reduced other, more natural exercises of lenience, such as informal policing, jury verdicts, parole release, and commutations.[30] But the determination depends on jurisdiction. In jurisdictions with vibrant applications of other forms of lenience, there is less need for prosecutorial lenience.

Prosecutorial lenience makes the most sense in areas where prosecutors have the greatest expertise, like measuring the strength of the evidence, assessing caseload and court capacity, and comparing one case in a certain category to others.[31] Prosecutors often claim expertise in other areas, like public safety and justice, but there is an argument police are better suited to the first and judges the second. When it comes to overall policymaking, legislators may have a stronger claim to institutional expertise and legitimacy.

Nevertheless, there are good reasons for prosecutors to exercise lenience beyond their narrow expertise. The multitude of actors in the American system creates a kind of collective action problem. If no single official feels responsible for a defendant’s fate, no one takes responsibility for necessary exercises of lenience. When lenience is warranted, and not forthcoming from any other actor, it seems strangely formalist to object to prosecutor lenience. For example, while judges may be best suited to determining an appropriate sentence, the judge’s hands can be tied by a mandatory sentence. Similarly, when police arrest authority is mandated, prosecutorial lenience becomes the sole avenue for an early exit from the system. Prosecutor discretion has also increased relative to judges and juries as the courts became heavily dependent on plea bargains and not trials. The abolition of parole and rarity of commutations in many jurisdictions further restricts the availability of lenience. In sum, through numerous mechanisms over time and across jurisdictions, prosecutors have increasingly become the right actors to exercise lenience because no one else can or will.

II. Evaluating Common Forms of Prosecutorial Lenience

This Part catalogues the variety of motivations that spur prosecutors to lenience. It then evaluates these motivations through the lens of the principles of lenience described in the preceding Part.

A. Insufficient Evidence

Perhaps the least controversial justification for prosecutor lenience is a lack of evidence of guilt. In fact, when the available proof drops below the legally mandated “probable cause” threshold, dismissal is required and so not an example of prosecutorial lenience at all.[32] But prosecutors also dismiss cases that are supported by “probable cause.” Any dismissal of a case supported by probable cause falls into the prosecutorial lenience category. While there is no consensus on the exact threshold, I have suggested the following: “[A] prosecutor should only charge a case when the prosecutor expects that the evidence introduced at trial will prove the defendant’s guilt beyond a reasonable doubt.”[33] To foster abundant lenience, district attorneys can adopt that (or a similar) threshold to make it clear that prosecutors should only proceed with cases that reach a significantly higher proof threshold than probable cause.

Once a threshold is established, dismissal of cases or charges based on lack of evidence will not be arbitrary. The consideration is reversible; prosecutors regularly proceed when the evidence is strong. Indeed, American law consistently authorizes increasing levels of government intrusions and punishments depending on the strength of the evidence of culpability.[34]

Equality is also more easily enforced for this lenience category because the considerations are concrete. Prosecutors across the office can follow a uniform evidence-of-guilt threshold—even if individual applications necessarily vary. And since evidentiary assessments fall squarely within prosecutors’ expertise, these decisions will be less vulnerable to critique.

B. Doing Justice

For better or worse, American prosecutors are tasked with “doing justice.”[35] As I have written elsewhere, the term offers little guidance for prosecutors operating “in the grey landscapes of the criminal law, where justice has little uncontested content.”[36] But it’s a starting point. For this Essay, I define a prosecutor’s pursuit of “justice” as ensuring that defendants are punished as much as is necessary and no more.[37] While this definition is contested and does not provide precise answers, it is intended to capture the rough consensus across the political spectrum of what American prosecutors should be doing. Most significantly, this standard leaves ample space for prosecutorial lenience. When American criminal law is too severe, all should agree that a prosecutor seeking justice can exercise lenience.[38] (While the focus in this Essay is lenience, the “do justice” talisman provides broad rhetorical license for prosecutors to seek harsh penalties as well.[39])

As an example, a prosecutor might forego the death penalty in a capital case because the defendant, while eligible for execution, is not among “the worst of the worst.”[40] Other examples include:

  • Declining to trigger a firearm-use enhancement when a conviction on the underlying offense alone provides sufficient sentencing authority, and the enhancement would lead to disproportionate punishment.
  • Declining to charge a minor offense as a third strike.
  • Selecting a charge with no immigration consequences to ensure that a minor transgression does not result in life-altering consequences.
  • Declining to prosecute a defendant who broke the law but deserves no punishment.

Depending on the precise facts, the application of lenience in each of these scenarios will likely be contested. But the general principle—that prosecutors tasked with doing justice can appropriately apply lenience—should be uncontroversial.[41]

In a system of severe laws and harsh enforcement, justice can lead to abundant prosecutorial lenience. The problem is indeterminacy. To reduce arbitrary actions and cabin justice-based lenience somewhat, prosecutors (and their critics) can examine rationales for reversibility. Some of the most common examples easily pass this test, such as the presence of violence, harm to victims, or a record of serious crimes. When present, these are all factors that support severity; their absence, then, logically supports lenience. Other examples of justice-based lenience include a person who steals to meet basic needs. Prosecutors could offer lenience assuming they would view less noble motives as aggravating factors. Moving up the severity spectrum, a person who commits a mercy killing could receive more lenience than a killer seeking an early inheritance. An intriguing type of justice-based lenience could involve supporting shorter prison terms in light of harsh prison conditions in the jurisdiction. The worse the prison conditions, the shorter the period of incarceration needed to reach the deserved punishment.[42]

Other rationales for justice-based lenience fare poorly. For example, a prosecutor would have difficulty justifying lenience on the grounds that the defendant has a promising future, a loving family, or was enrolled in college. Reversing the argument, it would be problematic to advocate for more severity for defendants with unloving families, few prospects, or who dropped out of college.

Due to its broad scope, justice-based lenience creates tensions with the second principle of prosecutorial lenience, equality. Justice-based concerns can be subjective and amorphous, making it likely that different prosecutors in a single office will treat similar defendants unequally. This will be especially true if prosecutors suffer from conscious or unconscious biases, or if the background laws or societal realities make defendants appear different when their relevant culpability is equivalent.

Clear guidelines are a common way to address equality concerns. Publicizing the guidelines promotes feedback that may correct oversights and allows defendants and their attorneys to bring deviations to the attention of prosecutorial supervisors. Precise guidelines, however, may restrict prosecutors’ freedom to offer lenience. In addition, publicized guidelines invite blowback from members of the public and other officials who disagree with a district attorney’s view of what justice entails. These criticisms, essentially a variation of the why prosecutors? critique, may resonate with the public because prosecutors are not necessarily seen as having specialized expertise about what justice requires.

C. Transactional Lenience

Perhaps the most substantial and variable form of prosecutorial lenience arises when prosecutors bargain. Defendants have three bargaining chips that they can exchange for lenience. The first is cooperation. Since prosecutors cannot lawfully compel defendants to divulge information about their offense or assist in its prosecution, prosecutors offer lenience in exchange for testimony or other forms of cooperation. A second form of transactional lenience lies at the heart of plea bargaining. Here, prosecutors offer lenience in exchange for the defendant’s waiver of trial rights. A third, related form of transactional lenience arises when prosecutors act as problem solvers, conditioning the dismissal of charges on non-conviction alternatives, like the defendant’s agreement to community service, drug treatment, restitution, or counseling.[43]

At least at the level of the individual, transactional lenience is critical to fostering abundance.[44] All things being equal, less negotiation would mean less lenience. Indeed, many of the most powerful instances of progressive prosecution involve lenient plea deals in serious cases.[45]

Transactional lenience certainly raises arbitrariness concerns, although some forms of cooperation-based leniency do not seem arbitrary. A prosecutor who offers lenience to a low-level accomplice in return for testimony against a crime boss would likely seek increased severity for an offender who refuses to cooperate. Perhaps the most justifiable distinction is between an offender who expresses remorse and pledges to discontinue the charged crime versus the unapologetic offender who proclaims a desire to continue committing the offense. It is important to recognize that cooperation-based lenience will create perceived injustice as defendants’ treatment turns not on their culpability but on the value of their cooperation. But most such critiques target prosecutors’ evaluation of the public value of cooperation in particular instances (and the need for justice), rather than the transaction itself. For example, it would be hard to fault a prosecutor who offered to take the death penalty off the table in return for a defendant’s cooperation in rescuing a kidnapped victim.

The arbitrariness calculus becomes cloudy as the withheld cooperation maps onto defendants’ constitutional rights, and this cloudiness becomes an impenetrable fog when the prosecutor offers lenience for a trial waiver. A prosecutor’s offer of lenience for a trial waiver is not reversible. It is, in fact, unconstitutional to punish defendants for refusing to waive their constitutional rights.[46] This familiar argument extends to diversion agreements. While it may seem unobjectionable to propose a lenient disposition in exchange for an agreement to enter drug treatment or do community service, the argument becomes problematic when flipped. Should someone be penalized for declining to accept the prosecutor’s suggested trial alternatives?

Part of the problem in this context is that the prosecutor collapses two distinct roles: (1) adjudicating guilt and (2) setting punishment. A defendant who is remorseful and taking steps toward a crime-free future may deserve lenience, but that lenience applies at sentencing, not in the process of adjudicating the offense. Prosecutorial lenience in this context pressures defendants to concede guilt to obtain sentencing concessions, conflating two questions that should be distinct.

Transactional lenience also raises equality concerns. These concerns can, again, be addressed through guidelines and transparency but only with the attendant risks of restricting prosecutorial lenience and inviting public and political disagreement.

D. Resource Constraint Lenience

Perhaps the most common motivation for prosecutorial lenience is triage. American prosecutors and courts are not capable of processing, much less trying, all the cases that are brought to them by police.[47] This leads to pressure to decline and dismiss prosecutions and plea bargain cases. By shedding weaker, more trivial cases, prosecutors can focus their attention on those that are more important. This type of lenience can be a necessity: “Prosecutors who cannot satisfy the legal obligations attendant to every case on their docket must dismiss cases until they can.”[48]

Resource constraint lenience is not arbitrary, as indicated by an intuitive kind of reversibility. A prosecutor who dismisses a case for lack of resources could prosecute a similar case when those resources later materialize. And since such dismissals rely on relatively concrete variables (i.e., resource availability), they can be applied equally. Resource constraints have the potential to foster abundant lenience. This is why some commentators suggest that one path to lenience (or at least chaos) is to overwhelm prosecutor offices.[49]

E. Disagreement with the Law

While there is general agreement that prosecutors can use lenience to address individualized justice concerns, the consensus frays when lenience is applied more broadly. For example, prosecutors have begun to embrace blanket lenience (or quasi-blanket lenience), such as refusing to trigger a particular law seen as unjust.[50] Such an action is exemplified in the Special Directives issued by Los Angeles District Attorney George Gascón.[51] As explained in a later court challenge,

[T]he Special Directives prohibited deputy district attorneys in most cases from alleging prior serious or violent felony convictions (commonly referred to as “strikes”) under the three strikes law or sentence enhancements and required deputy district attorneys in pending cases to move to dismiss or seek leave to remove from the charging document allegations of strikes and sentence enhancements. The Special Directives’ stated objectives, through these policies, were to promote the “interests of justice and public safety” by reducing “long sentences” that “do little” to deter crime.[52]

Similar policies exist in district attorney offices across the country, where prosecutors discourage the filing of certain, typically low-level charges except in exceptional cases or with supervisor approval.[53] Prosecutors could also seek to exercise leniency by using their charging discretion to avoid certain collateral consequences, such as deportation or the loss of a professional license.

Lenience based on disagreement with the law increases the abundance of leniency and follows a clear principle, at least in the abstract. A prosecutor who disagrees with a law declines to prosecute violations of that law while preserving resources (and severity) for enforcing other laws that the prosecutor deems more important. If the prosecutor is consistent, these decisions will not be arbitrary. The challenge for prosecutors will be explaining how they determine which laws to enforce and why prosecutors (as opposed to, say, legislators or police) should make these determinations. Presumably the legislature has decided to enact (or keep) the nullified law and so prosecutors arguably undermine democratic will by overriding that determination.[54] One answer is that district attorneys are (usually) elected officials who can also be voted out of office, perhaps preserving the democratic legitimacy of blanket non-prosecution.[55] That said, blanket lenience policies are likely to provoke responses from other official actors, like legislators, governors, and judges, since prosecutors are not necessarily viewed as experts in evaluating the relative importance of laws.

With respect to the second principle of prosecutorial lenience, blanket approaches are well positioned to meet equality concerns, particularly if the policies are announced in advance. The potential application of the (unenforced) law to the defendant, not any individualized determination of the equities, controls. Of course, equality concerns resurface when there are exceptions to the blanket policies, and in the broader decisions about which laws to enforce.

F. Mercy

Most scholars agree that “the exercise of mercy within the criminal justice system” is necessary to temper “the draconian harshness of our current penological regime.”[56] But a few prominent scholars contend that “mercy” is “better eliminated from the realm of criminal justice.”[57] Digging into this debate, it appears that at least some of the disagreement turns on what one means by “mercy.”[58] Those who define mercy as lenience are strong proponents.[59] Those who squeeze every virtuous concept out of their definition of “mercy” oppose its use. One suspects that, if locked in a room, both sides would agree that justice is good, lenience is necessary to achieve justice in the American system, and to the extent justice does not account for all the lenience that is required, mercy could be helpful depending on precisely what we mean by the term.[60]

Given that I have broadly defined both lenience and justice already, I only have room for a narrow definition of mercy. Thus, I throw my lot in with the anti-mercy crowd in terms of definition. When someone does not deserve to be punished, justice (not mercy) supplies all necessary justification for lenience. Mercy comes into play only after justice concerns have been exhausted. This fits with the philosophical understanding that mercy is a gift to which the recipient is neither legally nor morally entitled.[61]

Understood in this manner, mercy-based lenience seems a poor fit for prosecutors. Mercy fits best at the end of the process when all the facts are gathered, the law has been applied, and one can assess whether mercy, as opposed to justice, is necessary to mitigate the punishment. It is hard to do this at the charging or plea-bargaining stages.[62] For this reason, it seems more fitting to think of a judge, parole board, or governor dispensing mercy.

In addition, mercy seems better situated to officials who broadly represent the populace. Other officials have the support of many more engaged voters and are evaluated across a broader range of values. Stated another way, mercy seems best exercised by those we would identify (say, to space aliens) as “our leaders.” We typically don’t view district attorneys as our leaders. Consequently, a prosecutor who determines that someone morally and legally deserves punishment but nevertheless exercises a kind of personal prerogative to grant mercy seems to be acting beyond their role.[63] Blackstone saw mercy as “one of the great advantages of monarchy” and it has that feel.[64] Perhaps no one fills the role of monarch in the United States, but if someone does, it is a mayor, governor, or President, not the local district attorney.

Returning to our framework, mercy-based lenience will often seem arbitrary. Devoid of a justice basis, prosecutors will be hard pressed to offer satisfactory examples of reversibility. It also may be difficult to enforce any kind of equality across mercy-motivated-lenience (“Mercy Mondays!”). Perhaps that is the nature of mercy, as exemplified by King John’s pardon described in the introduction, Bill Clinton’s pardon of a childhood friend who saved him from drowning,[65] or the Governor of Missouri’s commutation of a death sentence to avoid the awkward optics of a state execution during a visit from the Pope.[66]

All that said, mercy certainly helps promote abundant lenience. And given the broad grey areas in the concepts sketched so far, it may be important to hold onto mercy as a catchall for ensuring abundant lenience in scenarios where the application of other rationales is unclear.

G. Hybrid Rationales

Scholars can isolate lenience rationales, but real-world prosecutors will have mixed motives. In fact, one can imagine scenarios involving every lenience motivation referenced so far: a prosecutor offers a lenient plea deal because the defendant agrees to cooperate (transactional lenience), the evidence is less strong (insufficient evidence), the deal reduces the prosecutor’s caseload (resource constraints), there are equitable reasons that favor the defendant (doing justice), the prosecutor does not support the governing criminal law (disagreement with the law), and the prosecutor feels forgiving (mercy). This illustrates the complexity of prosecutorial decisions. The presence of multiple motivations complicates but does not alter the analysis. Prosecutors (and their critics) will need to interrogate lenience decisions to determine which motives truly matter and to what degree.

III. Conflicting Lenience Principles

Prosecutors seeking to exercise lenience sail in smoothest waters when the three principles of prosecutorial lenience align, for example by setting a robust threshold (well above “probable cause”) for insufficient evidence-based lenience. But sometimes the principles conflict. Maximizing justice- and mercy-based lenience enhances abundance while increasing risks of arbitrariness and inequality. The same is true for transactional lenience. If policymakers restrict prosecutors’ freedom to do things like offer steep plea discounts and creative problem solving, there will be less overall lenience. Relatedly, defendants could maximize resource-constraint lenience by overwhelming prosecutor offices with trial requests and starving them of resources, but the resulting quality of prosecutorial lenience would suffer.[67]

The tension between the principles of prosecutorial lenience raises an important dilemma. Is it fair to insist that prosecutors follow all three principles? Perhaps all we can ask is that prosecutors consider all three principles, while acknowledging that they will be forced to prioritize depending on jurisdiction—or even offense-specific factors. These factors would include the background criminal law, the district attorney’s political support both in the jurisdiction and in the state as a whole, and the degree to which line attorneys support the district attorney.

As noted already, against a background of moderate laws, non-punitive judges, frequent parole releases and commutations, the district attorney could be more attentive to the first two principles, seeking to perfect prosecutorial lenience without worrying as much about limiting its abundance.

Lenience will be most abundant in offices where subordinate prosecutors share the preferences of their lenient boss. This likely explains why newly elected prosecutors seek to transform not just the office’s policies but its personnel. If line attorneys mirror the district attorney’s lenience priorities, the district attorney can quietly encourage (or turn a blind eye to) lenience, hoping that external critics do not notice individualized lenience to which they might object. The quality of lenience might suffer but it will be most abundant. This is basically the approach that prosecutors took to lenience in the era before progressive prosecution. District attorneys could maximize discretionary lenience by declining to memorialize it, using the lack of transparency to dodge criticism and external restraint.[68]

When the background law is punitive and line attorneys either intentionally or inadvertently do not reflect the district attorney’s lenience preferences, tradeoffs are inevitable. District attorneys will need to promulgate and publicize precise guidelines to generate lenience. But the guidelines may provoke a response from voters, legislators, police and mayors, judges, and governors. Depending on the jurisdiction and the district attorney’s political strength, this may not matter. In fact, a district attorney with thoughtful, transparent guidelines will be able to respond to critics with a clear justification for any instance of prosecutorial lenience.[69] But for district attorneys in more precarious political positions, the guidelines may embolden political critics, leading to external constraints and, ultimately, less lenience.

Conclusion

The conceptual difficulties inherent in prosecutorial lenience should be no surprise. These difficulties are a natural outgrowth of the American prosecutor’s uncertain role. Without greater consensus on that role, prosecutors and their critics will struggle to answer questions like when and how frequently prosecutors should be lenient. Still, prosecutors must continue to make decisions amidst this uncertainty. Consequently, it is important to offer some guidance based on whatever limited consensus exists. This Essay suggests three governing principles. Prosecutorial lenience should be: (1) non-arbitrary, (2) equal, and (3) abundant. But in applying these principles in common scenarios, the analysis reveals that the principles sometimes conflict. Consequently, prosecutors must prioritize. In some jurisdictions, it may turn out that the best principle of prosecutorial lenience remains the one we started with: “yes.”

  1. . Jeffrey Bellin, Reassessing Prosecutorial Power Through the Lens of Mass Incarceration, 116 Mich. L. Rev. 835, 837 (2018) (“Prosecutors are the Darth Vader of academic writing: mysterious, powerful and, for the most part, bad.”). ↑
  2. . Jeffrey Bellin, Defending Progressive Prosecution, 39 Yale L. & Pol’y Rev. 218, 218 (2020); see also Stephanie Holmes Didwania, Redundant Leniency and Redundant Punishment in Prosecutorial Reforms, 75 Okla. L. Rev. 25, 32–33 (2022) (discussing and documenting the trend); W. Kerrel Murray, Populist Prosecutorial Nullification, 96 N.Y.U. L. Rev. 173, 176 (2021) (“[P]rosecutors are beginning to stretch their power beyond mine-run resource-driven nonenforcement and one-off ex post declinations in ‘anomalous cases’ of factual guilt.”). ↑
  3. . See Jessica A. Roth, Prosecutorial Declination Statements, 110 J. Crim. L. & Criminology 477, 480 (2020) (“[T]he question of how and to what extent prosecutors should be held accountable for their exercise of negative discretion has not received sufficient attention.”); Rachel E. Barkow, The Ascent of the Administrative State and the Demise of Mercy, 121 Harv. L. Rev. 1332, 1352 (2008) (“[M]ost commentators worry more about the coercive power of prosecutors rather than their power to be lenient.”). ↑
  4. . See Roger A. Fairfax, Jr., Prosecutorial Nullification, 52 B.C. L. Rev. 1243, 1250–51 (2011) (“[C]omplete and unfettered discretion might too easily lead to the sort of arbitrary or improperly discriminatory criminal law enforcement our legal culture has shunned as a normative and constitutional matter.”). ↑
  5. . See Jeffrey Bellin, Mass Incarceration Nation 169 (2023) (“Everybody wants the government to punish someone. . . . In isolation, each argument for severity can sound compelling. Add up all these arguments and you get Mass Incarceration.”); Benjamin Levin, Prosecuting the Crisis, 50 Fordham Urb. L.J. 989, 1007 (2023) (“[L]iberal, progressive, and left commentators criticize prosecutors for failing to be sufficiently punitive when addressing police violence, financial crimes, race- and gender-based violence, and so forth.”). ↑
  6. . See Josh Bowers, Legal Guilt, Normative Innocence, and the Equitable Decision Not to Prosecute, 110 Colum. L. Rev. 1655, 1660 (2010) (arguing that “prosecutors may be ill-suited to adequately consider relevant equitable factors”). ↑
  7. . Thomas J. McSweeney, Salvation by Statute: Magna Carta, Legislation, and the King’s Soul, 25 Wm. & Mary Bill Rts. J. 455, 462 (2016). ↑
  8. . Id. ↑
  9. . Id. at 463. ↑
  10. . Jeffrey Bellin, The Road to End Mass Incarceration Could Begin with Mercy for Some Jan. 6 Rioters, NBC News (Jan. 6, 2023, 7:40 AM), https://www.nbcnews.com
    /think/opinion/biden-show-nonviolent-jan-6-rioters-mercy-end-mass-incarceration-rcna64552 [https://perma.cc/P8G9-89P2]. ↑
  11. . See, e.g., Pod Save America, Biden’s Lawless Raid on Biden, Crooked Media, at 1:12 (Jan. 24, 2023), https://crooked.com/podcast/bidens-lawless-raid-on-biden/ (describing the idea of commuting the sentences of January 6 rioters as a bad “take”); Pod Save America, Wikipedia, https://en.wikipedia.org/wiki/Pod_Save_America [https://perma.cc/AXW3-CSY6] (Apr. 3, 2023) (describing the show as “an American progressive political podcast” that “averages more than 1.5 million listeners an episode”). ↑
  12. . Bellin, supra note 10.  ↑
  13. . Id. ↑
  14. . See supra note 11 and accompanying text. ↑
  15. . See Bruce A. Green & Rebecca Roiphe, A Fiduciary Theory of Prosecution, 69 Am. U. L. Rev. 805, 806 (2020) (“Generations of scholars have failed to arrive at a unifying theory of prosecution, one that explains the complex role that prosecutors play in our democratic system.”). ↑
  16. . See Jeffrey Bellin, Theories of Prosecution, 108 Cal. L. Rev. 1203, 1211 (2020) (attempting “to construct a coherent normative theory of prosecution”). ↑
  17. . See Adam M. Gershowitz, The Prosecutor Vacancy Crisis, 50 BYU L. Rev. (forthcoming 2024) (manuscript at 2) (arguing that prosecutors have “astronomical caseloads” consisting of crimes that cannot be ignored, such as “[m]urders, robberies and other violent crime[s]”), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4666047 [https://perma.cc/X68Q-79CX]. ↑
  18. . See W. Kerrel Murray, Populist Prosecutorial Nullification, 96 N.Y.U. L. Rev. 173, 226, 232–33 (2021) (advocating “a transparent, reasoned, pre-election promise” to justify blanket non-enforcement policies). ↑
  19. . See generally Jeffrey Bellin, Mass Incarceration Nation (2023) (discussing inequities in the criminal justice system). ↑
  20. . Id.; Lee Kovarsky, Prosecutor Mercy, 24 New Crim. L. Rev. 326, 337 (2021) (“Modern clemency is, for the most part, an empty husk . . . .”). ↑
  21. . Bowers, supra note 6, at 1664 (explaining that legislators “seek to leave determinations of optimal enforcement to the executive”); Fairfax, supra note 4, at 1270 (flagging “implicit . . . delegation of authority to prosecutors to determine which, if any, of these laws are to be enforced”); see also William J. Stuntz, The Pathological Politics of Criminal Law, 100 Mich. L. Rev. 505, 509 (2001) (arguing that “lawmaking and adjudication pass into the hands of police and prosecutors . . . [and they] determine who goes to prison and for how long”). ↑
  22. . See Dan Markel, Against Mercy, 88 Minn. L. Rev. 1421, 1435 (2004) (conceptualizing “‘leniency’ towards an offender as a value-free umbrella term under which an offender receives less punishment than is possible”). ↑
  23. . Bordenkircher v. Hayes, 434 U.S. 357, 364 (1978) (“[S]o long as the prosecutor has probable cause to believe that the accused committed an offense defined by statute, the decision whether or not to prosecute, and what charge to file or bring before a grand jury, generally rests entirely in his discretion.”). ↑
  24. . See James Vorenberg, Decent Restraint of Prosecutorial Discretion, 94 Harv. L. Rev. 1521, 1555 (1981) (“Giving prosecutors the power to invoke or deny punishment at their discretion raises the prospect that society’s most fundamental sanctions will be imposed arbitrarily and capriciously and that the least favored members of the community . . . will be treated most harshly.”). ↑
  25. . Arbitrary, Black’s Law Dictionary (4th ed. 1951). ↑
  26. . Bowers, supra note 6, at 1724. ↑
  27. . Fairfax, supra note 4, at 1272 (describing prosecutors’ “obligation of consistency”); Kay L. Levine, Should Consistency Be Part of the Reform Prosecutor’s Playbook?, 1 Hastings J. Crime & Punishment 169, 174 (2020) (arguing that reform prosecutors should focus on “consistency of process” as the “principal value” for reform). But see Bowers, supra note 6, at 1675 (“genuine uniformity is illusory”). ↑
  28. . Note, The Paradox of “Progressive Prosecution,” 132 Harv. L. Rev. 748, 760 (2018) (highlighting “standard principal-agent problem” that creates a “significant potential for noncompliance from those on the lower rungs of the hierarchy due to a lack of buy-in to the goals of the head prosecutor”). ↑
  29. . See Angela J. Davis, The Perils of Private Prosecutions, 13 Cal. L. Rev. Online 7, 16–17 (2022) (offering examples from Larry Krasner and Rachel Rollins). ↑
  30. . See generally Bellin, supra note 5 (exploring the decline in these traditional forms of lenience since the 1970s). ↑
  31. . Green & Roiphe, supra note 15, at 844 (“Prosecutorial independence . . . assumes that experience and expertise are the best guarantee that prosecutors will seek the public interest.”); Barkow, supra note 3, at 1354 (contending that prosecutors avoid criticism when they “are seen as making an ‘expert’ determination about priority-setting when they choose not to bring charges”); Kovarsky, supra note 20, at 341 (“[T]he prosecutor’s office also has the best information and expertise to make the pertinent decisions.”); Bowers, supra note 6, at 1686 (“Prosecutors are best positioned to gauge evidentiary merit and to set and negotiate administrative priorities.”). ↑
  32. . See Bordenkircher v. Hayes, 434 U.S. 357, 364 (1978) (“[S]o long as the prosecutor has probable cause to believe that the accused committed an offense defined by statute, the decision whether or not to prosecute, and what charge to file or bring before a grand jury, generally rests entirely in his discretion.”); Am. Bar Ass’n, ABA Standards for Criminal Justice Prosecution Function and Defense Function § 3-3.9(a) (3d ed. 1993) (“A prosecutor . . . should not institute criminal charges when the prosecutor knows that the charges are not supported by probable cause.”); U.S. Dep’t of Justice, Just. Manual § 9-27.200 cmt. (2018) (“[F]ailure to meet the minimal requirement of probable cause is an absolute bar to initiating a federal prosecution . . . .”). ↑
  33. . Bellin, supra note 16, at 1223. ↑
  34. . See Irene Oritseweyinmi Joe, Regulating Mass Prosecution, 53 U.C. Davis L. Rev. 1175, 1187 n.40 (2020) (enumerating the various standards required to prevail on a variety of different court matters). ↑
  35. . See Berger v. United States, 295 U.S. 78, 88 (1935) (exhorting that the prosecution’s goal is “not that it shall win a case, but that justice shall be done”). ↑
  36. . Bellin, supra note 16, at 1208. ↑
  37. . Kathleen Dean Moore, Pardons: Justice, Mercy, and the Public Interest 12 (1989) (“Granting a pardon is a duty of justice that follows from the principle that punishment should not exceed what is deserved.”). ↑
  38. . Bowers, supra note 6, at 1664 (“It is necessary and desirable for prosecutors to exercise a measure of discretion because codes are too expansive to do otherwise.”). ↑
  39. . Bellin, supra note 16, at 1216–17. ↑
  40. . Kansas v. Marsh, 548 U.S. 163, 206 (2006) (Souter, J., dissenting). ↑
  41. . See Mortimer R. Kadish & Sanford H. Kadish, Discretion to Disobey: A Study of Lawful Departures from Legal Rules 82 (1973) (“It is widely accepted that . . . the prosecutor must . . . ‘balance the admonitory value of the invariable and inflexible punishment against the greater impulse of the quality of mercy.’”). ↑
  42. . See, e.g., United States v. Chavez, No. 22-CR-303 (JMF), 2024 WL 50233, at *1 (S.D.N.Y. Jan. 4, 2024) (“It has gotten to the point that it is routine for judges . . . to give reduced sentences to defendants based on the conditions of confinement in the MDC.”). ↑
  43. . See Bruce A. Green & Rebecca Roiphe, When Prosecutors Politick: Progressive Law Enforcers Then and Now, 110 J. Crim. L. & Criminology 719, 761 (2020) (contrasting conventional prosecutors being “primarily case processors” with progressive prosecutors, who believe in sometimes “providing mental health and social services to offenders rather than prosecuting and imprisoning them”); Ronald F. Wright & Kay L. Levine, Career Motivations of State Prosecutors, 86 Geo. Wash. L. Rev. 1667, 1710 (2018) (“Problem-solving skills, and the broad set of motivations on which they necessarily rest, will define which prosecutors claim a place at the frontier of criminal justice reform.”). ↑
  44. . Cf. Jeffrey Bellin, Plea Bargaining’s Uncertainty Problem, 101 Texas L. Rev. 539, 571, 584 (2023) (noting the “common critique” that plea bargaining “increases the overall efficiency of the State’s ability to impose criminal punishment” but suggesting that the primary individual effect of plea bargains is reducing sentencing exposure). ↑
  45. . See id. at 541 (identifying instances in which plea deals were negotiated for defendants facing life sentences). ↑
  46. . United States v. Goodwin, 457 U.S. 368, 372 (1982). ↑
  47. . See Shu-Yi Oei & Diane M. Ring, “Slack” in the Data Age, 73 Ala. L. Rev. 47, 58 (2021) (“A key context in which slack arises is when enforcers have scarce resources, requiring prioritization.”); Fairfax, supra note 4, at 1257 (“Prosecutors often must decide not to pursue one matter (or category of matters) in order to have the investigative or prosecutorial capacity to prosecute other matters . . . .”). ↑
  48. . Bellin, supra note 16, at 1214; Joe, supra note 34, at 1246 (highlighting “the prosecutor’s responsibility and ethical duty to refrain from engaging in practices that overwhelm the public defender” to “comply with national and state ethical guidelines while maintaining the executive function in the criminal justice system”). ↑
  49. . See generally Andrew Manuel Crespo, No Justice, No Pleas: Subverting Mass Incarceration Through Defendant Collective Action, 90 Fordham L. Rev. 1999 (2022) (discussing a proposal that suggests a collective refusal to engage in plea bargaining as a method of undermining mass incarceration). But see Gershowitz, supra note 17 (manuscript at 2, 23) (suggesting that a lack of prosecutors leads offices to simply “drop and refile charges” that aren’t immediately ready for trial). ↑
  50. 50. See, e.g., Ass’n of Deputy Dist. Att’ys for Los Angeles Cnty. v. Gascón, 79 Cal. App. 5th 503, 514, 555 (2022) (affirming preliminary injunctions enjoining the district attorney from directing deputy district attorneys to not plead prior strikes and to move to dismiss or withdraw allegations of strikes and sentence enhancements).

    51. Id. ↑

  51. ↑
  52. . Id. at 514. ↑
  53. . See Davis, supra note 29, at 16 (providing an example of a Philadelphia prosecutor who has discouraged the prosecution of low-level offenses). ↑
  54. 54. See Milan Markovic, Charging Abortion, 92 Fordham L. Rev. 1519, 1524 (2024). ↑
  55. . Murray, supra note 2, at 208–09 (defending prosecutors who “nullify not unilaterally, but consistent with a reasonably ascertainable popular will”); Bellin, supra note 2, at 226 (arguing that “prosecutorial lenience is itself subject to restraint through political accountability”); Fairfax, supra note 4, at 1268 (“[T]hat the prosecutor represents the interests of, and is accountable to, the citizenry in the enforcement . . . is a powerful argument for the authority of prosecutors to engage in ‘nullification.’”). ↑
  56. . Carol S. Steiker, Tempering or Tampering? Mercy and the Administration of Criminal Justice, in Forgiveness, Mercy, and Clemency 16, 19 (Austin Sarat & Nasser Hussain eds., 2007) (“[T]he creation of institutional possibilities for the exercise of mercy within the criminal justice system is extremely attractive as a way of mitigating the draconian harshness of our current penological regime.”). ↑
  57. . Markel, supra note 22, at 1433; see also Jeffrie G. Murphy & Jean Hampton, Forgiveness and Mercy 174 (1988) (arguing that prosecutors and other officials should “keep their sentimentality to themselves for use in their private lives with their families and pets”). ↑
  58. . Steiker, supra note 56, at 16, 22 (“[U]nder [the] skeptical view of mercy, justice embraces a piece, perhaps a very large one, of what in common parlance goes by the name of mercy.”). ↑
  59. . Compare Kovarsky, supra note 20, at 327 (“When I use the term ‘mercy,’ I refer to an official act that reduces a lawfully imposed sentence.”), and Eric L. Muller, The Virtue of Mercy in Criminal Sentencing, 24 Seton Hall L. Rev. 288, 329, 341 (1993) (defining “mercy [as] a frame of mind induced by the imaginative effort to see both the impact of the possible sentences and the nature of the criminal conduct from the defendant’s perspective” and arguing that “mercy, as defined in this Article, is an independent moral virtue of the criminal sentencing process”), and R. A. Duff, The Intrusion of Mercy, 4 Ohio St. J. Crim. L. 361, 362–63 (2007) (explicitly declining to define the term but asserting that “mercy is sometimes relevant to the criminal law, but that it is relevant as a justified intrusion,” especially, but not exclusively, in the sentencing context), with Markel, supra note 22, at 1435 (“I will define mercy in a more narrow way so that there can be a distinct difference between reasons that serve to lessen punishment for purposes of justice and reasons that lessen punishment for purposes of mercy.”), and Murphy & Hampton, supra note 57, at 166 (explaining that mercy “is never owed to anyone . . . as a matter of desert or justice”), and J. Budziszewski, Categorical Pardon: On the Argument for Abolishing Capital Punishment, 16 Notre Dame J.L. Ethics & Pub. Pol’y 43, 43–44 (2002) (arguing that “mercy is punishing the criminal less than he deserves”). ↑
  60. . Bowers, supra note at 6, at 1682 (suggesting examples of mercy that may not be connected to justice, such as the parent who negligently kills her own child or a terminally ill defendant). Arguably, even these examples can be subsumed into the justice analysis under the heading of whether these defendants deserve less punishment. ↑
  61. . Aziz Z. Huq, A Right to a Human Decision, 106 Va. L. Rev. 611, 661 (2020) (“Mercy is generally understood as a discretionary act to which one has no entitlement.”); Stephanos Bibas, Forgiveness in Criminal Procedure, 4 Ohio St. J. Crim. L. 329, 333 (2007) (“The offender has no right to mercy, but the punisher grants it as an act of grace out of love or compassion.”); Murphy & Hampton, supra note 57, at 166 (explaining that mercy is “best viewed as a free gift”). ↑
  62. . See Kovarsky, supra note 20, at 327–28 (advocating for institutional mercy “after a conviction and sentence is final” and declining to “address discretion exercised in favor of a defendant at earlier phases of the criminal process, such as a prosecutor’s decision to undercharge”). ↑
  63. . Cf. Austin Sarat & Conor Clarke, Beyond Discretion: Prosecution, the Logic of Sovereignty, and the Limits of Law, 33 Law & Soc. Inquiry 387, 412 (2008) (discussing the prosecutors’ power to decline prosecution as a “[f]ragment[] of sovereignty”). ↑
  64. 64. 4 William Blackstone, Commentaries *233. ↑
  65. . Bellin, supra note 5, at 45. ↑
  66. . Mary Sigler, Mercy, Clemency, and the Case of Karla Faye Tucker, 4 Ohio St. J. Crim. L. 455, 481 (2007). ↑
  67. . See Jennifer E. Laurin, Progressive Prosecutorial Accountability, 50 Fordham Urb. L.J. 1067, 1090 (2023) (highlighting related “accountability trade-offs” inherent in progressive prosecution movement). ↑
  68. . See Ronald F. Wright, Prosecutorial Guidelines and the New Terrain in New Jersey, 109 Penn St. L. Rev. 1087, 1104 (2005) (“Perhaps the only way to remove some of the severity is to allow prosecutors to operate quietly, dispensing mercy in a few cases, even if it is done inconsistently.”); Barkow, supra note 3, at 1353 (noting that “prosecutors have not received the same scrutiny” as others who dispense lenience in part because “fewer noteworthy examples of improper exercises of discretion come to the public’s attention”). ↑
  69. . Cf. Laurin, supra note 67, at 1090–91 (recommending “well-mediated disclosures” rather than “direct-to-market data”). ↑

Download PDF