In Apprendi v. New Jersey (2000), the Court held that a defendant has a right to a jury trial regarding the facts that may increase the maximum sentence to which he is exposed.  Breyer dissented in Apprendi and has steadfastly maintained ever since that Apprendi was wrongly decided.

Two years later, in Harris, the Court decided not to extend Apprendi to the facts that raise a defendant’s minimum sentence.  Breyer was part of the 5-4 majority in Harris, but stated in a concurring opinion that he could see no reason to distinguish increasing the maximum from increasing the minimum.  Thus, Breyer’s vote in Harris was simply another vote against Apprendi.  This immediately raised the expectation that some day, when Breyer was ready to give up the fight against Apprendi, he would be willing to overturn Harris.

Some day has come.  

In Alleyne, Breyer provided the crucial fifth vote to overturn Harris.  In a short concurring opinion, Breyer reiterated his view that Apprendi was in error, but stated, “Apprendi has now defined the relevant legal regime for an additional decade [since Harris].  And, in my view, the law should no longer tolerate the anomaly that the Apprendi/Harris distinction creates.”

I don’t see much of particular interest in Justice Thomas’s majority opinion; Thomas authored the dissent in Harris, which he has now been able to rewrite as an opinion for the winning side.

Justice Sotomayor’s concurring opinion was more notable for its effort to reconcile overturning Harris with the doctrine of stare decisis.  (Thomas has never had much use for stare decisis, so he predictably gave short shrift to the doctrine in his opinion.)  Sotomayor, joined by Justices Ginsburg and Kagan, affirmed that overturning precedent requires a “special justification.”  Here, though, such a special justification was present:

As an initial matter, when procedural rules are at issue that do not govern primary conduct and do not implicate the reliance interests of private parties, the force of stare decisis is reduced. See United States v. Gaudin, 515 U. S. 506, 521 (1995); Payne, 501 U. S., at 828. And any reliance interest that the Federal Government and state governments might have is particularly minimal here because prosecutors are perfectly able to “charge facts upon which a mandatory minimum sentence is based in the indictment and prove them to a jury.” Harris, 536 U. S., at 581 (THOMAS, J., dissenting). Indeed, even with Harris in place, prosecutors already sometimes charge such facts and seek to prove them to a jury. . . .

In this context, stare decisis does not compel adherence to a decision whose “underpinnings” have been “eroded” by subsequent developments of constitutional law. Gaudin, 515 U. S., at 521. In rejecting a constitutional challenge to a state statute that increased a defendant’s minimum sentence based on judicial factfinding, McMillan relied on a distinction between “elements” and “sentencing factors.”477 U. S., at 86. That distinction was undermined by Apprendi, where we held that a legislature may not “remove from the jury the assessment of facts that increase the prescribed range of penalties to which a criminal defendant is exposed.” 530 U. S., at 490 (internal quotation marks omitted).

In Harris, we squarely confronted the question whether “McMillan stands after Apprendi.” 536 U. S., at 550. Five Members of the Court recognized that the cases were in fact incompatible. See id., at 569 (BREYER, J., concurring in part and concurring in judgment); id., at 572, 583 (THOMAS, J., dissenting) (“[O]nly a minority of the Court embrac[es] the distinction between McMillan and Apprendi that forms the basis of today’s holding”). . . .

We have said that a decision may be “of questionable precedential value” when “a majority of the Court expressly disagreed with the rationale of [a] plurality.” Seminole Tribe of Fla. v. Florida, 517 U. S. 44, 66 (1996). And Harris has stood on especially weak ground because its vitality depended upon the possibility that the Court might retreat from Apprendi. See Harris, 536 U. S., at 569–570 (opinion of BREYER, J.). That has not happened. Instead, while individual Members of this Court have continued to question Apprendi, see post, at 1–2 (opinion of BREYER, J.); post, at 1–2 (ALITO, J., dissenting), its rule has become even more firmly rooted in the Court’s Sixth Amendment jurisprudence in the decade since Harris. . . .

As a result of these decisions, Harris has become even more of an outlier. For that reason, I agree that it is appropriate for the Court to “overrule Harris and to apply Apprendi’s basic jury-determination rule to mandatory minimum sentences” in order to “erase th[is] anomaly” in our case law. Post, at 2–3 (opinion of BREYER, J.). I do not suggest that every single factor that supports the overruling of precedent is present here. Post, at 3, n. * (ALITO, J., dissenting). But particularly in a case where the reliance interests are so minimal, and the reliance interests of private parties are nonexistent, stare decisis cannot excuse a refusal to bring “coherence and consistency,” Patterson, 491 U. S., at 174, to our Sixth Amendment law.

Sotomayor thus suggests a lesser deference to precedent in cases presenting procedural questions, particularly when the precedent lacked majority agreement on the reasoning.  Expect this opinion to be invoked in any number of crim pro cases in the next few terms, especially since Sotomayor’s reasoning was partially, if summarily, echoed in footnote 5 in the majority opinion.

Whatever the significance of its discussion of stare decisis, Alleyene may prove the most practically important decision in the Apprendi line of cases since Blakely v. Washington (2004), which resulted in fundamental changes to many sentencing guidelines systems.  Mandatory minimums are a routine aspect of criminal-law practice in many jurisdictions, including the federal system, often providing prosecutors with a powerful source of leverage in plea negotiations.  Alleyne now reduces that leverage by making it procedurally more onerous for prosecutors to have mandatory minimums applied.  How much more onerous will vary a great deal from case to case; in many, the loss of leverage will be insignificant, but in some it may be decisive.  Although Alleyne is not likely to result in a large increase in the number of jury trials, it may perceptibly alter the prosecutor-defendant balance of power in plea negotiations in certain kinds of cases in some jurisdictions.

Alleyne’s impact, however, will be substantially constrained by another “anomaly” (to use Breyer’s term) in the Apprendi line of cases: the Apprendi jury-trial right does not apply to sentence enhancements based on prior criminal convictions.  The leading case is Almendarez-Torres v. United States (1998).  Since many of the most important mandatory minimums — think, for instance, of “three strikes” laws — are triggered by criminal history, the Almendarez-Torres exception actually constitutes a huge gap in Alleyne‘s coverage.

Might Almendarez-Torres be in jeopardy now?  Alleyne should inspire defense counsel to redouble their efforts to have A-T overturned.  After all, the decision is of about the same vintage as Harris and seems no less anomalous relative to the Apprendi line of cases.  Bearing in mind Sotomayor’s point about the diminished force of stare decisis in procedural cases, the Court should be open to revisiting A-T now.  Notably, Breyer’s vote in A-T was necessary for the 5-4 majority, just as it had been in Harris.  If Breyer really has made his peace with Apprendi and is now committed to eliminating doctrinal anomalies in this area, A-T may not long stand.

So what drives perceptions of disorder?  Some research suggests that an increasing minority presence in a neighborhood tends to increase perceptions of disorder, even holding objective signs of disorder constant.  Other research focuses on perceptions of minority presence, which are associated with perceptions of disorder and fear of crime.

Now, an interesting new article by Rebecca Wickes and coauthors explores how these dynamics vary by neighborhood and city.

Wickes et al. analyzed survey data from individuals in 297 different neighborhoods in the Australian cities of Melbourne and Brisbane.  Four findings stand out.

First, not all minorities have the same impact on neighborhood perceptions.  For instance, a “one percentage point increase in immigrants from Asia increases the perceived non-Anglo Saxon [proportion in the neighborhood] by .5 percent, whereas a similar increase in Middle Eastern immigrants increases the perceived non-Anglo Saxon by 1.6 percent, and a similar increase by Southeastern Europeans increases it by 1.9 percent.” (20)

Second, “when residents overestimate the presence of minorities in their neighborhood, they report significantly higher perceptions of disorder. . . . In both [cities], overestimating the minority presence was one of the strongest predictors of perceived disorder . . . even when controlling for a comprehensive range of individual- and community-level predictors . . . .”  (22)  (Other variables associated with higher levels of perceived disorder were lower levels of income, higher total crime rates, and more residential instability.)

Third, not all neighborhoods are affected equally by higher perceptions of minority presence.  Wickes et al. focused on the effect of social cohesion, which “is commonly understood as a prosocial good that represents a sense of belonging and attachment and brings about positive outcomes for the collective.  In essence, it symbolizes a working trust and a general willingness of residents to work together.”  (8)  The researchers found that “higher levels of social cohesion diminish the association between overestimating the minority presence in a neighborhood and perceiving disorder.”  (26)  This suggests that highly cohesive neighborhoods can absorb a higher minority presence without a corresponding increase in perceived disorder.

Finally, the dynamics were somewhat different not only neighborhood to neighborhood, but also city to city:

We find that city context matters a lot: Overestimates of minority presence vary between the cities of Melbourne and Brisbane.  We suggest this is the result of their different racial, cultural, and ethnic histories.  For example, whereas high concentrations of indigenous residents strongly influenced overestimating minority presence in Brisbane and Melbourne neighborhoods, the strength of these effects differed across cities.  In Brisbane neighborhoods, greater proportions of indigenous people were associated with “seeing” more minorities and perceiving more disorder when compared with Melbourne.  (29)

The article, “‘Seeing’ Minorities and Perceptions of Disorder: Explicating the Mediating and Moderating Mechanisms of Social Cohesion,” is forthcoming in Criminology.

Earlier today, in United States v. Davila (No. 12-167), the U.S. Supreme Court unamimously resolved the circuit split in favor of the general harmless error rule.  As the Court saw it, the legal question was an easy one: “[N]either Rule 11 itself, not the Advisory Committee’s commentary on the Rule singles out any instructions [in Rule 11] as more basic than others.  And Rule 11(h), specifically designed to stop automatic vacaturs, calls for across-the-board application of the harmless-error prescription . . . .”

The Court declined to adopt any bright-line rules regarding the application of the harmless-error rule: “Our essential point is that particular facts and circumstances matter.”  Having determined that the lower court should have applied the harmless-error rule, the Court chose to remand for further consideration of the “particular facts and circumstances.”  At the same time, the Court did say, “Had Davila’s guilty plea followed soon after the Magistrate Judge told Davila that pleading guilty might be the ‘best advice’ a lawyer could give him, this case may not have warranted our attention.”  The suggestion seems to be that a guilty plea entered “soon after” the judge recommended such a plea would pretty clearly not fall into the category of harmless error.  What made Davila’s case more difficult was the three-month delay between the Rule 11 violation and the guilty plea.

The Court converted the federal sentencing guidelines from mandatory to advisory in 2005, but left unanswered many important questions about what exactly it means for the guidelines to be “advisory.”  Several of these questions were answered in a trilogy of 2007 decisions, which effectively established a new and unique sentencing system for the federal courts.  Although sentencing judges are not required to follow the guidelines, the Supreme Court did put a thumb on the scales in favor of guidelines sentences.  Dissenting justices objected that this kinda-sorta advisory system violated the Sixth Amendment, but to no avail.

The new system also raised Ex Post Facto Clause issues, which divided the lower courts.  Peugh nicely illustrates the problem.

Peugh committed bank fraud in 1999 and 2000, but was not sentenced until many years later.  In the intervening years, the applicable guidelines were made substantially harsher.  If sentenced under the guidelines in place at the time of his crime, Peugh faced a recommended guidelines range of 30-37 months.  The new guidelines, however, called for a range of 70-87 months, more than doubling Peugh’s time in prison.  The sentencing judge applied the new guidelines and chose to impose a sentence within the higher range.  On appeal, the Seventh Circuit rejected Peugh’s argument that such a retroactive application of guidelines amendments violates the Ex Post Facto Clause.

In overturning the Seventh Circuit’s decision, the Supreme Court, per Justice Sotomayor, framed the test this way:

The touchstone of this Court’s inquiry is whether a given change in law presents a “‘sufficient risk of increasing the measure of punishment attached to the covered crimes.’” Garner, 529 U. S., at 250 (quoting Morales, 514 U. S., at 509). The question when a change in law creates such a risk is “a matter of degree”; the test cannot be reduced to a “single formula.” Id., at 509 (internal quotation marks omitted).

While the government argued that the risk of a higher sentence was not “sufficient” because the sentencing judge was theoretically free to impose a sentence below the new, higher recommended range, Peugh relied on the thumb on the scales established by the Court’s 2007 trilogy.

By a narrow 5-4 margin, the Court chose Peugh’s “common sense” over the government’s formalism:

The post-Booker federal sentencing scheme aims to achieve uniformity by ensuring that sentencing decisions are anchored by the Guidelines and that they remain a meaningful benchmark through the process of appellate review.  As we have described, “district courts must begin their analysis with the Guidelines and remain cognizant of them throughout the sentencing process.” Gall, 552 U. S., at 50, n. 6 (emphasis added). Failing to calculate the correct Guidelines range constitutes procedural error. Id., at 51. A district court contemplating a non-Guidelines sentence “must consider the extent of the deviation and ensure that the justification is sufficiently compelling to support the degree of the variance.” Id., at 50.

These requirements mean that “[i]n the usual sentencing, . . . the judge will use the Guidelines range as the starting point in the analysis and impose a sentence within the range.” Freeman v. United States, 564 U. S. ___, ___ (2011) (plurality opinion) (slip op., at 5). Even if the sentencing judge sees a reason to vary from the Guidelines, “if the judge uses the sentencing range as the beginning point to explain the decision to deviate from it, then the Guidelines are in a real sense the basis for the sentence.” Ibid. (emphasis added).  That a district court may ultimately sentence a given defendant outside the Guidelines range does not deprive the Guidelines of force as the framework for sentencing. Indeed, the rule that an incorrect Guidelines calculation is procedural error ensures that they remain the starting point for every sentencing calculation in the federal system.

Similarly, appellate review for reasonableness using the Guidelines as a benchmark helps promote uniformity by “tend[ing] to iron out sentencing differences.” Booker, 543 U. S., at 263. Courts of appeals may presume a within-Guidelines sentence is reasonable, see Rita, 551 U. S., at 347, and they may further “consider the extent of the deviation” from the Guidelines as part of their reasonableness review, Gall, 552 U. S., at 51.  As in Miller, then, the post-Booker sentencing regime puts in place procedural “hurdle[s]” that, in practice, make the imposition of a non-Guidelines sentence less likely. See 482 U. S., at 435.

. . .

Peugh points to considerable empirical evidence indicating that the Sentencing Guidelines have the intended effect of influencing the sentences imposed by judges. Even after Booker rendered the Sentencing Guidelines advisory, district courts have in the vast majority of cases imposed either within-Guidelines sentences or sentences that depart downward from the Guidelines on the Government’s motion. See United States Sentencing Commission, 2011 Sourcebook of Federal Sentencing Statistics, p. 63 (Figure G) (16th ed.) (USSC). In less than one-fifth of cases since 2007 have district courts imposed above- or below-Guidelines sentences absent a Government motion. See ibid. Moreover, the Sentencing Commission’s data indicate that when a Guidelines range moves up or down, offenders’ sentences move with it. See USSC, Final Quarterly Data Report, FY 2012, p. 32 (Figure C); USSC, Report on the Continuing Impact of United States v. Booker on Federal Sentencing, Pt. A, pp. 60–68 (2012).

The federal system adopts procedural measures intended to make the Guidelines the lodestone of sentencing. A retrospective increase in the Guidelines range applicable to a defendant creates a sufficient risk of a higher sentence to constitute an ex post facto violation.

The Court’s approach here creates a tension between its ex post facto treatment of the guidelines (in which the guidelines are recognized to be sufficiently weighty so as to preclude retroactive application of amendments) and its Sixth Amendment analysis (in which the guidelines are not seen as weighty enough to trigger a jury-trial right in determining the recommended range).  Does this point to some instability in the great 2007 accommodation?  The realism of Peugh about the continuing practical importance of the guidelines, including the Court’s noteworthy reliance on statistics regarding actual sentencing practices, makes it seem all the more questionable that greater procedural protections are not provided to federal defendants at sentencing.

The Williams controversy showcases the difficulties — medical, legal, political — that can sometimes arise in determining a cause of death, and the fallibility of the experts on whom we rely for these purposes.  Jentzen’s book helps to put all of this into a big-picture historical context.

Jentzen’s central narrative is the contest between coroners and medical examiners.  

The coroner model, ubiquitous in the U.S. at the time of its founding and for a long time thereafter, relies on elected laypeople to determine cause of death, often in conjunction with an inquest jury.  The medical examiner model relies on physicians, preferably with specialized training in both medical and legal aspects of death investigations.

Beginning in the mid-nineteenth century, physicians began to criticize the coroner system and demand more scientific approaches to death investigation.   In response to such agitation, Massachusetts adopted the nation’s first medical-examiner law in 1877.  Under this law, the new office of medical examiner took over the responsibility for determining the cause and manner of suspicious deaths.  The medical examiners, who were appointed by the Governor, were required to be “able and discreet men learned in the science of medicine.”

The medical examiner model spread from Massachesetts, but only quite slowly.  It met fierce resistance from politically entrenched coroners and their allies, including undertakers, many of whom profited from cozy relationships with the local coroner.  Opponents argued that medical examiners represented a wasteful expense that would have to be borne by taxpayers — doctors, of course, don’t come cheap — and that the public benefitted from having an independent and democratically accountable death-investigator.  States only tended to flip to the medical-examiner model in response to highly publicized cases of corruption or incompetence by local coroners.

Jentzen laments that, even today, many states retain the coroner system.  In 2000, he estimates that about forty percent of the U.S. population was still beyond the coverage of medical examiners.

Jentzen’s perspective is that of a classic progressive reformer, reflecting considerable faith in the power of training, specialized expertise, professionalization, and the scientific method to address important social problems.

This is certainly not the only corner of the criminal-justice field in which progressive reformers have engaged in long-running, back-and-forth battles with such competing political forces as populism, fiscal conservatism, and simple institutional inertia.  For instance, looking back in history, the prison, originally developed as a progressive alternative to corporal punishment, also spread slowly and unevenly across the U.S. in the nineteenth century, as did the concept of a professional police force.  More recently, I think of the debates that have emerged in the past generation regarding sentencing commissions and sentencing guidelines.

My own instincts tend to be in line with Jentzen’s.  If our criminal-justice system made more use of science and specialized expertise, it would probably function better along just about every important dimension.

Still, there are limits to what science can do.  Law demands clear, either-or answers to its questions — answers that may be beyond science’s capabilities.  Moreover, law’s questions often blend fact and moral judgment together in uncertain ways.  Thus, for instance, the notion of causation in law combines the factual question of “but-for” causation with the implicit moral judgment of proximate causation.  Such blending can be a source of confusion and frustration when science tries to answer legal questions.  (I should note, by the way, that Jentzen himself is well aware of these challenges and discusses them effectively.)

This brings us back to Derek Williams, of course.  The difficulties here — the inconsistent expert opinions, the seeming disconnect between the ME’s homicide determination and the various decisions not to prosecute — do serve to highlight limitations of the medical examiner system.  This is not to say that a coroner system would be better, but rather that we should be realistic about our progressive ambitions; making criminal justice more scientific does not eliminate its inherent messiness.

Critics often lament the scope of prosecutorial charging and plea-bargaining discretion, which can be difficult to reconcile with rule-of-law ideals.  However, Levine and Wright’s research makes clear that prosecutorial discretion is not a monolithic phenomenon, but may function quite differently in different offices.  They focus particularly on the significance of two variables relating to office structure.  They summarize as follows:  

Two particular features of office structure drew our attention: the flat or pyramidal shape of the organization’s workforce and the preference for hiring experienced attorneys or recent graduates into entry-level positions. These features of the office’s social architecture correlate with distinctive professional identities of the prosecutors who work there. For example, attorneys who work in pyramidal offices and who are hired without experience (as in Metro County) tend to accept bureaucratic and group values. A strong team spirit marks their professional identities. On the other hand, attorneys in an office such as Midway County, characterized by a flatter structure and more experienced hires, display professional identities that are decidedly more independent; they feel no particular obligation to match their own outputs to the decisions of their peers or to the policies of their superiors. These are autonomous, rule-defying prosecutors. Surprisingly, an attorney’s team-member-versus-autonomous-actor identity correlates more strongly with his office’s social architecture than with his gender or race.  (1123)

Levine and Wright do not take a position on which type of office structure is superior, but they do highlight what may be the most important trade-off:

The public expects the professionals who work in American criminal justice to pursue two contradictory aims: to treat in the same way every person who commits the same crime and to treat each suspect and each defendant as an individual. These paradoxical expectations—treat every person the same, except for those who should be treated differently—also shape our views about prosecutorial discretion. Not surprisingly, then, one theme that arose during our interviews was the need for consistency among different prosecutors who work in the same office.

While interviewees in all of our research locations raised this point, some attorneys gave a higher priority to consistency than others, and this variation seemed to correlate with their office environments. Line prosecutors in Metro, an office whose structures promote a strong team spirit, explained that they encourage each other to produce consistent outcomes in the cases they each handle. Conversely, prosecutors in our flat office stocked with veterans (Midway) placed value on individualized outcomes rather than on consistency. (1170-71)

The trade-off implicates questions of accountability, which lie at the heart of concerns about prosecutorial discretion.

I think that what makes people most uncomfortable with prosecutorial discretion is the sense that great power is being exercised by individuals who face no real accountability for their decisions.  Line prosecutors do not normally have to face the voters on a regular basis (although in smaller jurisdictions, it is true that the elected head of the office may actually have to handle cases on a day-to-day basis).  Nor, if democratic accountability is lacking, do judges seem to have much desire or ability to fill the gap.  Occasionally, judges will throw out a plea deal or otherwise attempt to hold prosecutors to account, but this seems more the exception than the rule.

But democratic and judicial possibilities do not exhaust the potential mechanisms of accountability.  Levine and Wright’s research suggests that bureaucratic accountability may function in a robust way in some offices.  Layers of supervision and peer monitoring may help to ensure that some defendants do not suffer (or benefit) unfairly from a single prosecutor’s whims or idiosyncratic priorities.  Moreover, if we imagine an elected prosecutor at the top of the bureaucratic pyramid, we may receive some further reassurance that a measure of democratic accountability is in play.

Still, the picture is troubling in a number of respects.  For one thing, as government bureaucracies go, prosecutor’s offices tend to be unusually opaque, and it is questionable whether voters really have access to the sort of information that would permit a good assessment of how well any given office was performing.  (The fact that incumbent DA’s rarely face serious electoral challenges may, in part, reflect the lack of rigorous public scrutiny.)

Moreover, the way that the accountability mechanisms work may produce a general tendency to over-harshness.  If systematic performance measures are not available, then the public’s views of an office’s performance will be shaped by sensationalistic media coverage of atypical cases.  If the office is lenient to a defendant who then goes out and commits a horrible crime, the DA will take a political hit.  On the other hand, there may be little or political cost to prosecuting cases to the hilt.  The people who find themselves in the prosecutorial crosshairs tend to be politically marginalized to begin with, and if there are ultimately wrongful convictions or overly harsh sentences, blame can easily be shifted to judges and juries.

This may bring us back around to the ideal of an experienced, professional, politically insulated prosecutor who feels empowered  to do the right thing in each case, even at the cost of a certain amount of disparity between cases based on who the prosecutor is.

As Levine and Wright continue their research — and the present article purports to be merely a first installment — it would be fascinating to learn whether different office structures do, in fact, tend to produce different levels of severity.

But how about accountability  in a world of “autonomous, rule-defying prosecutors,” as Levine and Wright put it?

Their research does suggest yet another model of accountability, one based on professional reputation.  They observe that “most prosecutors admit that they want defense attorneys to see them as fair and prepared, rather than as aggressive or game-oriented.”  (1166)  Additionally, they found that some prosecutors develop and value professional relationships with other prosecutors from outside their own offices.  (This seemed to be less true, though, in the more hierarchical office.)  The desire to maintain a good professional reputation may thus establish an effective alternative form of accountability, even for prosecutors who operate with only very weak bureaucratic controls.

Admittedly, though, this type of accountability may be too soft to mollify some of the critics of prosecutorial discretion.

One that will be welcomed by habeas fans is McQuiggin v. Perkins (No. 12-126).  Perkins was convicted of murder in state court, with the judgment becoming final in 1997.  More than eleven years later, Perkins filed a federal habeas corpus petition, alleging that he received unreasonably poor representation by his trial counsel.  The petition plainly violated the one-year statute of limitations for habeas petitions, but Perkins sought to get around the statute by presenting evidence that he was actually innocent of the crime of which he was convicted.  The Supreme Court has long recognized that actual innocence is an exception to the procedural default rule, which normally bars federal courts from considering habeas claims that were not timely raised in state court.  Perkins argued that there should also be an actual-innocence exception to the statute of limitations, and the Supreme Court agreed in a 5-4 decision.

Does this new exception threaten to eviscerate the statute of limitations? 

Hardly.  The Court emphasized the very high standard that a petitioner must satisfy in order to show actual innocence: “The . . . exception , we underscore, applies to a severely confined category: cases in which new evidence shows it is more likely than not that no reasonable juror would have convicted the petitioner.”  Indeed, while remanding to the Sixth Circuit for further proceedings, the Court went out of its way to indicate its view that Perkins had not satisfied the actual-innocence standard.

In truth, as Justice Scalia argued in dissent, it is hard to see in the statute any basis for the new exception.  McQuiggin is perhaps best understood against the backdrop of the Court’s long struggle to deal with claims of innocence in the habeas context.  Habeas is supposed to be a forum for vindicating constitutional rights, and it is not immediately clear that punishing an innocent defendant violates any provision of the Constitution.  Indeed, the Court has steadfastly avoided squarely accepting or rejecting the existence of a “stand-alone” innocence claim in habeas.  However, the strained legal reasoning in McQuiggin seems to reflect the Court’s unwillingness to let procedural obstacles prevent federal courts from reaching the merits in cases presenting serious claims of innocence.  It seems likely that this desire to ensure that habeas courts have the power to correct wrongful convictions will, sooner or later, lead the Court finally to endorse the long-awaited stand-alone actual innocence claim.

A second case offering hope to habeas petitioners is Trevino v. Thaler (No. 11-10189).  The long-simmering background issue in this case is whether there is a constitutional right to counsel (and hence to effective assistance of counsel) in collateral proceedings challenging state-court convictions.  Trevino was convicted of murder in state court in Texas and sentenced to death.  His direct appeals and a collateral challenge were unsuccessful.  Then, in a second collateral attack, Trevino argued for the first time that his trial counsel had provided unreasonably poor representation by failing to investigate and present certain mitigating evidence that might have resulted in a lesser sentence.  The state courts held that this claim was procedurally defaulted for not being raised earlier.  As noted above, such a procedural default in state court normally precludes relief in federal court, too.  However, there is another exception to the procedural default rule besides actual innocence; default may be avoided if a habeas petitioner can show “cause” and “prejudice” as to the default.  In order to show cause, Trevino pointed to the incompetence of his first collateral counsel, who failed to present what has become a common claim in capital cases (that is, failure by trial counsel to develop an effective mitigation case).

The Court was long resistant to the suggestion that poor representation in a collateral proceeding might count as “cause” to excuse a procedural default.  Since the Constitution does not guarantee a lawyer at all in collateral proceedings (in contrast to direct appeals) it has struck the Court as odd to recognize, even obliquely, any sort of a right to effective assistance at this stage.  On the other hand, in light of the well-documented deficiencies of indigent representation in this country, critics have long argued that it is unfair to stick a defendant with an incompetent lawyer at trial and on appeal, and then to deny the defendant a fair opportunity to challenge that incompetent representation in collateral proceedings.

There was finally a modest breakthrough just a year ago in Martinez v. Ryan, in which the Court held that incompetent representation in a collateral proceeding could count as “cause” in excusing a failure to challenge the competence of trial counsel in a state in which the defendant is required to challenge trial counsel in a collateral proceeding.

Trevino was arguably distinguishable, though, because Texas did not formally limit Trevino to a collateral proceeding in order to challenge his trial counsel.  However, the Court, again by a 5-4 margin, chose to expand the Martinez exception so that it also applies in states, like Texas, in which it is “virtually impossible” for an ineffective assistance claim to be presented on direct review.  Important to the analysis was the tight time frames under Texas law for the direct review process and the reality that, in order to comply with the deadlines, ineffective assistance claims might have to be made by post-conviction counsel without even having the benefit of a trial transcript.  It is not clear whether any other states with different post-conviction time frames might also be covered by Trevino’s extension of Martinez.

Despite the fact-specific nature of the holding, habeas fans might nonetheless see in Trevino (and Martinez before it) hopeful signs that the Court now better appreciates the extraordinary difficulties facing defendants who receive incompetent representation at trial and on direct appeal.  Although neither Trevino nor Martinez expressly deals with a constitutional right to counsel in collateral proceedings – just an exception to the procedural default rule — these decisions may possibly indicate a willingness on the part of the Court to reconsider its earlier hard-line stance against a right to counsel after the first direct appeal.

But all is not looking up for habeas petitioners.  The Court ruled against the petitioner in its third recent habeas case, Metrish v. Lancaster (No. 12-547).  Burt Lancaster(!) was charged with murder in Michigan state court and presented a diminished-capacity defense at trial, but was convicted anyway.  The conviction was later overturned and Lancaster retried.  In between the two trials, the Michigan Supreme Court eliminated the diminished-capacity defense, which had previously been recognized by the state’s intermediate court of appeals but never by the high court.  Lancaster was thus prohibited from presenting his defense at the retrial and was again convicted.  On appeal, he argued unsuccessfully that retroactive application of the new Michigan Supreme Court decision violated his rights under the Due Process Clause.

Renewing the claim in federal court on habeas, Lancaster prevailed in the Sixth Circuit, but the U.S. Supreme Court reversed.  The Court did not draw a bright-line rule on the retroactivity of state supreme court decisions, but emphasized the specific facts of the case, including that the state supreme court did not overturn any of its own precedents, but only the decisions of the intermediate court of appeals.  The U.S. Supreme Court also relied on the deference that must be shown to state-court decisions under 28 U.S.C. §2254(d)(1), which limits habeas relief to circumstances in which a challenged state-court decision “was contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court.”

Although the very case-specific analysis of Metrish has direct significance for very few, if any, other habeas petitioners, it does perpetuate the Court’s recent trend of emphasizing §2254(d)(1), which (as I discuss in this article) may be a double-edged sword for defendants: on the one hand, habeas petitioners face yet another challenging hurdle under this line of cases (on top of the procedural default rule, the statute of limitations, and so forth); on the other, a highly restrictive understanding of §2254(d)(1) may leave the Court feeling freer to develop constitutional protections in direct-review cases without the fear that doing so will spawn unmanageable waves of new habeas claims.

The graph depicts year-by-year rates of imprisonment, homicide, and violent crime (the latter based on results from the National Crime Victimization Survey), indexed to 1992 rates. The mirroring effect is most pronounced if you compare imprisonment (green line) with homicide (red): between 1993 and 1999, imprisonment goes up at almost precisely the same rate that homicide goes down; in 2000, there is an abrupt leveling off in both areas; and neither has seen a lot of change since.  The violent victimization line (blue) mostly tracks the homicide line, save for an additional three years of rapid decline (1999-2002) and a notable uptick between 2009 and 2011.

The mirror-image paths might seem counterintuitive.  Shouldn’t less crime translate into less imprisonment?  Let me suggest three theories to account for what has happened.  

First, there is the “rational equilibrium adjustment” theory.  On this theory, increases in imprisonment between 1993 and 1999 drove crime down.  When we stopped increasing imprisonment, we also stopped making progress in fighting crime, thereby establishing a new crime-prison equilibrium that has remained more-or-less constant for more than a decade.

This theory could, in fact, be fit within a larger narrative about crime and imprisonment.  Violent crime exploded in the U.S. between the early 1960′s and the mid-1970′s, but the U.S. imprisonment rate remained relatively constant.  Beginning in the mid-1970′s and continuing through the late 1990′s, the U.S. criminal-justice system belatedly responded to increased violence with increasing severity in charging, sentencing, and corrections administration.  Imprisonment increases continued after crime increases stopped because there was still a lot of “catching up” to do, recalling that violence spiked for more than a decade before imprisonment started to move.  We stopped ratcheting up penal severity as soon as we got violence levels safely below the 1970′s-era peaks that originally triggered the imprisonment boom.

I think there is some truth to this theory, but it also has some significant problems.

For one thing, the relationship between crime and imprisonment is much more complicated and uncertain than I’ve made it out to be, and it is highly doubtful that we could, in any straightforward way, select our crime rate simply by setting our imprisonment level.  Indeed, if we took a longer view than I’ve depicted in the graph, we would see that the apparently neat connection between crime and imprisonment since 1993 may be an anomaly.  While imprisonment rates grew much more rapidly from 1975 to 1993 than they did afterwards, crime rates followed no consistent pattern during that time period.  Indeed, the explosive growth of the American prison population in the 1970′s and 1980′s did not prevent a major escalation of violent crime in the late 1980′s and early 1990′s (the peak of which is reflected at the far left of the graph.) Bottom line: imprisonment rates affect crime rates, but the reverse is also true; moreover, both imprisonment and crime are also affected by any number of other social forces.  It is unlikely that the “great American crime decline” of the 1990′s (to quote Frank Zimring) was solely the result of increased toughness in the criminal-justice system.  (Indeed, as Zimring has observed, the crime decline of the 1990′s was actually a broad, international phenomenon that also touched nations that did not follow the get-tough path of the United States.)

Another difficulty or gap with the “rational equilibrium adjustment” theory is that it does not explain why we stopped where we did.  If crime could be decreased in a linear fashion by imprisoning more people, and if the imprisonment increases of the 1990′s reflected a rational social policy choice to do so, why were we apparently satisfied with returning to the violence levels of the early 1970′s?  Why not go all the way back to the early 1960′s?  Or even lower?

And who is this “we” anyway?  In truth, U.S. imprisonment levels result from a myriad of different decisions made by many diverse actors with competing interests and minimal communication across institutional lines, let alone centralized coordination–cops, prosecutors, judges, parole boards, probation officers, and legislators spread over 50 states and thousands of local jurisdictions.  In some theoretical sense, all are accountable to “the people,” but it seems quite implausible that the American public deliberately “chose” the crime-imprisonment equilibrium that has prevailed since about 2000.  I’ve recently spent some time withthis fascinating report produced by the National Center for State Courts in 2006, which analyzed national polling data regarding sentencing.  The researchers found that a large majority of Americans (about 60 percent) incorrectly believed that violent crime had increased over the past five years.  The researchers also found a sizable minority (nearly 45 percent) who did not know that U.S. incarceration rates are generally higher than those found elsewhere in the world.  Finally, only about 18 percent of respondents expressed the view that sentencing policies were “working pretty well.”  Given such high levels of misinformation and dissatisfaction, it is hard to believe that the current state of affairs is simply a straightforward matter of rational public choice.

So, let’s consider a second possibility, the “crime-wave catchup” theory.  Here’s the story: as violence spiked in the late 1980′s and early 1990′s, criminal-justice resources remained constant, which means that demand for incarceration far outpaced the system’s capacity to deliver incarceration.  Crimes went unsolved; prosecutors declined serious cases or offered generous plea bargains in order to stay on top of exploding dockets; parole boards became more generous in order to save prisons from dangerous overcrowding; and so forth.  By the time the wave of violence began to recede in 1993, there were a lot of people loose in the community who had done some pretty bad things — people who would have been in prison if the system had not been overwhelmed by a crime wave.  The next six years of steadily declining violent crime provided the system with an opportunity to catchup on this backlog, resulting in a burgeoning prison population.

It is not clear, though, why imprisonment rates have remained high even now, many years after the backlog was taken care of.  Prison sentences for violent crime average about eight years, and the median sentence is only about four years, so it is seems unlikely that a large percentage of today’s prisoners are still serving sentences that were imposed in the 1990′s.  To be sure, recidivism rates among recently released prisoners are high, and many return to prison, either as a result of new sentences or revocations of parole.  The revolving door phenomenon means that high imprisonment rates may have a tendency to perpetuate themselves for many years.  But, even at that, the great crime-wave cohort of the late 1980′s and early 1990′s is in its forties now — an age bracket in which criminal tendencies decline markedly.  As that cohort gradually disappears from the system, it seems that the reduced criminality of the younger generation should eventually be manifested in falling rates of imprisonment (assuming that system severity remains constant).

Before proposing my third theory, let me present another graph, which depicts imprisonment rates (blue line) and U.S. gross domestic product (red), indexed to 1992.

This graph may illuminate the “supply-side,” a/k/a “if you build it, they will come,” theory.  Here’s the story: The crime wave of the late 1980′s and early 1990′s created the political impetus for building U.S. imprisonment capacity.  (Such capacity is only in part, I suspect, a function of the number of prison beds, but is also likely a product of such disparate additional factors as the severity of sentencing laws; the number and efficiency of police officers, prosecutors, and judges; and perhaps general social beliefs regarding crime and criminals.)  Capacity grew swiftly across the 1990′s, and imprisonment grew along with it, reflecting to some extent both rational equilibrium adjustment and crime-wave catchup.  Although crime was actually falling over this period of time, the public did not understand this to be the case, and those working in the system, who benefitted from a massive infusion of new resources, had little incentive to correct public misunderstandings.

However, the imprisonment growth of the 1990′s was economically unsustainable; note that imprisonment growth outpaced GDP growth in that decade.  Although short-lived, the 2001 recession was an important reality check for policymakers, and brought the capacity-building era to an end.  It probably helped that the crime-wave catchup needs had been met been then.  Moreover, while the public may not have been consciously aware of declining crime rates, there was likely some inchoate sense that things weren’t quite as bad as they had been.  For instance, as detailed in the NCSC report, the frequency that crime was mentioned as the nation’s top problem in public-opinion polls fell steadily for a decade after its all-time peak in 1994.  In short, the political will to embark on major new spending in the criminal-justice area dried up.

It is one thing to say no to new initiatives, but quite another to start dismantling existing programs.  The public officials, private contractors, and others with an economic stake in existing programs will fight to keep them.  But to continue to justify the programs to policymakers and budget-writers, it is necessary to show that the programs are being fully utilized.

California offers the best example of how this played out with imprisonment in the early 2000′s.  Long after it was obvious to nearly every thoughtful observer that the state’s draconian “three strikes” law — the nation’s harshest — was a train wreck of a criminal-justice policy, the California Correctional Peace Officers Association — the nation’s most powerful prison guards’ union — continued to fight tenaciously and successfully to keep the law in place, thereby ensuring full prisons and plenty of jobs for prison guards.  (This story is recounted in Josh Page’s great book, “The Toughest Beat.“)

Although I would not suggest that this is the only cause for the post-2000 equilibrium — imprisonment is high while crime is low — it does seem likely that the political and economic interests of those who benefitted from the expansion of imprisonment capacity in the 1990′s plays a significant role.

The dynamics I’ve described here seem consistent with the research of William Spelman, which I discussed in more detail in this article.  Spelman has performed a multivariate regression analysis of state prison population trends from 1977 to 2005.  Holding other variables constant, he found a strong relationship between a state’s level of capital spending on prisons and its subsequent level of imprisonment.  As he put it, “Prison populations are largely driven by available capacity; when the money was available to increase capacity, policy makers spent it; when the beds were available, criminal justice agencies filled them.”

In other words, if you build it, they will come.

The United States Supreme Court has recently recognized new constitutional limitations on the use of life-without-parole (LWOP) sentences for juvenile offenders, but has not clearly indicated whether analogous limitations apply to the sentencing of adults. However, the Court’s treatment of LWOP as a qualitatively different and intrinsically more troubling punishment than any other sentence of incarceration does provide a plausible basis for adults to challenge their LWOP sentences, particularly when they have been imposed for nonviolent offenses or on a mandatory basis. At the same time, the Court’s Eighth Amendment reasoning suggests some reluctance to overturn sentencing practices that are in widespread use or otherwise seem to reflect deliberate, majoritarian decisionmaking. This Essay thus suggests a balancing test of sorts that may help to account for the Court’s varied Eighth Amendment decisions in noncapital cases since 1991. The Essay concludes by considering how this balancing approach might apply to the mandatory LWOP sentence established by 21 U.S.C. §841(b)(1)(A) for repeat drug offenders.

The article will appear in print in a forthcoming symposium issue of the Missouri Law Review devoted to the Supreme Court’s year-old decision in Miller v. Alabama.

Measuring public punitiveness has proven difficult.  Simply asking people whether they are punitive seems unlikely to produce helpful results, given the uncertainty and abstraction of the term.  On the other hand, asking about support for any specific criminal-justice policy might or might nor produce answers that are reflective of more general attitudes.  Intuitively, for instance, support for the death penalty would seem a good indicator that a person would also support a range of other policies that are typically characterized as punitive, such as three-strikes laws, but it is hard to rule out the possibility that the death penalty is a unique issue in the minds of many Americans; support may be due, say, to religious beliefs or particular feelings regarding the crime of murder, rather than more general attitudes toward crime and criminals.

Ramirez attempted to overcome this difficulty by aggregating survey responses to several different criminal-justice policy questions.  He identified 24 different survey questions that were asked by national pollsters at least twice between 1951 and 2006.  Many of the questions related to the death penalty, but others touched on three-strikes laws, drug enforcement, law-enforcement spending, imprisonment, and sentencing more generally.  Although the levels of support for different punitive policies varied, they tended to move in unison over time, suggesting that there really is some shifting, underlying attitude that drives support for all of the different policies.

Based on the survey data, Ramirez compiled a year-by-year punitiveness index.

The general directions of change over time will not be especially surprising to readers who have studied trends in criminal law, sentencing, and incarceration rates.  (See, for instance, my article on incarceration trends in three midwestern states.)  Punitiveness fell from the early 1950′s through the mid-1960′s, then rose steadily through the early 1980′s.  Punitiveness was up and down from then until the late 1990′s, but remained consistently higher than it had been in the 1950′s and 1960′s.  Since the late 1990′s, punitiveness has generally been in decline and seemed by 2006 finally to have fallen back to about the level of the early 1950′s.

It is interesting and helpful to see this evidence of movement in public punitiveness.  Ramirez further considers the extent to which a range of other variables correlate with punitiveness over time.

To my mind, the three big questions about punitiveness are these: (1) are public attitudes in this area more an input to, or an output from, the political process; (2) to what extent is punitiveness, either consciously or unconsciously, an expression or manifestation of racial animus; and (3) is punitiveness more instrumental or symbolic?

Although Ramirez’s research does not provide definitive answers to any of these questions, at least as I have framed them, he does offer some suggestive observations at least as to the first two.  (I have blogged a number of times about the third question, e.g., here.)

On the relationship between public attitudes and the political process, Ramirez finds a striking correlation between public punitiveness and the punitiveness of presidential statements on crime.  (Interestingly, as Ramirez measured it, presidential punitiveness reached its post-1950 high not in a Republican administration, but in the first term of the Clinton administration.)  Here is what Ramirez concludes about the impact of the president:

A single [punitive] statement by the president estimated to increase punitive sentiment by an immediate .06 points.  Punitive sentiment will increase another .06 points during the next year.  Moreover, it will continue to shift . . . until it changes a total of .16 points.  (347-48)

As Ramirez points out, it is troubling that the “ups and downs of punitive sentiment are driven by . . . political factors such as the construction of crime by political leaders.  The framing of crime as a problem of a permissive system . . . increase[s] public demand for punitive policies.  [This is] normatively unappealing . . . because it suggests the public can be led into supporting unnecessary and costly policies [even when they are not justified by] objective crime conditions.”  (357)

On the other hand, Ramirez did not find punitiveness to be wholly disconnected from “objective crime conditions.”  For instance, “a point increase in the homicide rate is estimated to increase punitive sentiment directly by .39 points.”  (348)  Additionally, increases in drug usage also show “an immediate and long-term relationship with punitive sentiment.”  (349)

What about race and punitiveness?  Much research has indicated that blacks tend to be less punitive than whites, and Ramirez also found that to be true at least back to the early 1970′s.  (352)  (Women were less punitive than men over the same time period, although the sex gap is less pronounced than the race gap.)

The “racial threat” theory supplies one potential explanation for the punitiveness gap.  This theory posits increasing white demand for social control, including through the criminal-justice system, as the racial balance in a community shifts (or is perceived to shift) in a black direction.  Previous research has found some empirical support for this theory, although the results have not been unequivocal.  (See my posts here and here.)  In any event, Ramirez’s data do provide support for racial-threat effects.  Indeed, a “shift in public perceptions that racial integration is increasing results in an immediate .29 point increase in punitive sentiment.”  (349)  This, too, is a troubling finding, for it suggests another way (along with political grandstanding) in which punitive attitudes may result from factors that do not seem properly part of the policy-making process.

Ramirez’s article is “Punitive Sentiment,” 51 Criminology 329 (2013).