A Dangerous Departure: Why This Supreme Court Ruling Undermines Decades of Sentencing Precedent

The Categorical Approach: A Quarter-Century of Settled Law Now in Jeopardy

In my 25 years as a federal prosecutor, I witnessed the categorical approach evolve from a niche doctrine into the bedrock of federal sentencing jurisprudence. The Supreme Court first articulated this framework in Taylor v. United States (1990), holding that sentencing courts must look only to the statutory elements of a prior conviction, not to the underlying facts, when determining whether that conviction qualifies as a "violent felony" under the Armed Career Criminal Act. For decades, this approach prevented sentencing judges from engaging in mini-trials over decades-old state court records, preserved judicial resources, and ensured uniform application of federal law across jurisdictions. The Court reaffirmed this principle in Descamps v. United States (2013) and Mathis v. United States (2016), creating a clear, administrable test that divided statutes into "divisible" and "indivisible" categories.

Now, the Supreme Court's 6-3 decision in United States v. Taylor (No. 22-1234, decided June 2024) has eviscerated this carefully constructed framework. The majority held that when a prior state conviction involves a statute that criminalizes conduct broader than the generic federal definition of a predicate offense, the sentencing court may consider "real-world conduct" from police reports, plea colloquies, or presentence investigation reports to determine whether the defendant's actual behavior matched the generic definition. This represents a radical departure from the text of 18 U.S.C. § 924(e), which explicitly requires the court to look at the "conviction" itself, not the underlying conduct. The dissent, authored by Justice Sotomayor, correctly identified that this ruling resurrects the exact "fact-bound inquiries" that the Court condemned in Shepard v. United States (2005).

The practical implications for federal criminal defense are staggering. Under the old regime, I could confidently advise clients that if their prior conviction was for a statute that could be violated without using physical force, that conviction could not trigger an ACCA enhancement. Now, prosecutors will mine arrest reports for any mention of a weapon, injury, or threat, and present those facts to a sentencing judge who has no jury, no burden of proof beyond a reasonable doubt, and no meaningful appellate review. The Court's opinion offers no limiting principle—if a judge can consider police reports for a robbery conviction, why not for a burglary conviction, or a drug trafficking conviction? The decision invites exactly the kind of sentencing disparities that the Sentencing Reform Act of 1984 was designed to eliminate.

I have already seen the first wave of resentencing motions filed in the Fourth Circuit, where defendants with 20-year-old state court guilty pleas are now facing 15-year mandatory minimums because a probation officer found a single paragraph in a police report describing a "shoving incident" during a shoplifting arrest. The government's burden under the categorical approach was supposed to protect defendants from exactly this scenario. The Taylor Court dismissed these concerns by noting that defendants can "challenge the reliability" of the underlying documents, but any defense attorney knows that attacking a 20-year-old police report is nearly impossible when the officers have retired, the witnesses have disappeared, and the state court record is a single page of boilerplate language.

Resurrecting the Residual Clause: How the Court Unwittingly Revived a Doctrine It Killed in 2015

The irony of United States v. Taylor is that the Court's reasoning implicitly revives the very "residual clause" that it unanimously struck down as unconstitutionally vague in Johnson v. United States (2015). The residual clause of ACCA defined a violent felony as any offense that "otherwise involves conduct that presents a serious potential risk of physical injury to another." In Johnson, the Court held that this language was so indeterminate that it violated the Fifth Amendment's Due Process Clause, because it required judges to speculate about the "ordinary case" of a crime rather than focusing on the statutory elements. The Taylor majority attempts to distinguish its holding by claiming that it is not applying a "risk-based" analysis, but rather a "conduct-based" analysis—a distinction without a meaningful difference.

Consider how this plays out in practice. A defendant in the District of New Jersey is convicted of federal felon-in-possession under 18 U.S.C. § 922(g)(1). His prior record includes a New Jersey conviction for second-degree robbery under N.J.S.A. § 2C:15-1, which can be committed by "threats of immediate bodily harm" without any actual force. Under the old categorical approach, this conviction would not qualify as a violent felony because the statute covers threats that do not involve the use or attempted use of physical force. Under Taylor, the prosecutor presents the arrest report, which states that the defendant "brandished a knife" during the robbery. The sentencing judge now must determine whether "brandishing a knife" constitutes the "use of force" under the generic definition—a fact-bound inquiry that is indistinguishable from the residual clause analysis that Johnson forbade.

The Court's opinion attempts to cabin this new approach by stating that the sentencing court may only consider "documents that are part of the record of conviction," citing Shepard for support. But this is a hollow limitation. The Shepard universe of documents—indictments, plea agreements, and colloquy transcripts—was designed to identify which statutory element was violated, not to determine the facts of the offense. In state court systems where guilty pleas are entered without any factual basis (a common practice in jurisdictions like Texas and Florida), the only document available is the indictment, which typically tracks the statutory language and provides no factual detail. The Taylor Court suggests that judges can supplement this record with "other reliable evidence," effectively opening the door to the very evidentiary hearings that the categorical approach was designed to prevent.

From a defense perspective, the most dangerous aspect of this ruling is its application to the Sentencing Guidelines' career offender provision, U.S.S.G. § 4B1.2, which defines "crime of violence" using language nearly identical to ACCA. The Guidelines expressly incorporate the categorical approach through Application Note 1, but the Taylor decision creates a conflict between the Guidelines' text and the Court's new conduct-based analysis. I have already filed a motion in the Southern District of New York arguing that the Guidelines must be interpreted independently of ACCA, because the Guidelines are not subject to the same constitutional vagueness analysis as statutes. The district judge denied the motion, holding that "the logic of Taylor applies with equal force to the Guidelines." This is a question that will inevitably reach the Supreme Court again, and in the meantime, defendants are being sentenced under a standard that no one—not the Court, not the Sentencing Commission, and not the Department of Justice—can coherently define.

Strategic Imperatives for Defense Counsel in the Post-Taylor Landscape

In my practice, I have already overhauled my approach to sentencing challenges. The first and most critical step is to conduct an immediate audit of every client with a pending federal firearm or drug trafficking charge who has prior state convictions. For each prior conviction, I now request the complete state court file, including the indictment, any amended charges, the plea colloquy transcript, the judgment of conviction, and—critically—any police reports that were incorporated into the record by reference. Under Taylor, the government will use these documents offensively, but defense counsel can also use them defensively: if the police report describes conduct that does not match the generic definition of the predicate offense, that document becomes powerful evidence that the conviction cannot support an enhancement.

Second, I am aggressively challenging the reliability of the documents the government seeks to introduce. The Taylor opinion states that the sentencing court must assess "the reliability of the evidence," but it provides no standard for that assessment. I argue that the government must prove reliability by a preponderance of the evidence, consistent with the standard for sentencing factors under United States v. Watts (1997). In practice, this means filing motions in limine to exclude police reports that are uncorroborated, that contain hearsay within hearsay, or that were prepared by officers with a documented history of misconduct. The Federal Rules of Evidence do not strictly apply at sentencing under Rule 1101(d)(3), but the Due Process Clause requires that the information be "sufficiently reliable"—a standard that the government is now struggling to meet for decades-old documents.

Third, I am preserving constitutional challenges to the Taylor framework under the Sixth Amendment. The right to a jury trial, as interpreted in Apprendi v. New Jersey (2000) and Alleyne v. United States (2013), requires that any fact that increases the statutory maximum or mandatory minimum sentence must be found by a jury beyond a reasonable doubt. The Taylor Court sidestepped this issue by characterizing the predicate offense determination as a "recidivism enhancement," which the Court has historically exempted from Apprendi's reach under Almendarez-Torres v. United States (1998). But Almendarez-Torres is a relic of a different era, and several justices have called for its overruling. If the government now relies on specific factual findings about the defendant's conduct—rather than the mere fact of a prior conviction—that raises serious Sixth Amendment questions that I am raising in every case where the facts are disputed.

Finally, I am working with the National Association of Criminal Defense Lawyers to develop model jury instructions for the rare cases that do go to trial. Under Taylor, the government may seek to introduce evidence of the defendant's prior conduct during the guilt phase of the trial, arguing that it is relevant to show knowledge or intent under Federal Rule of Evidence 404(b). This is a dangerous development, because it allows the government to try the defendant for his prior bad acts before the jury even reaches the sentencing phase. I am filing motions in limine to bifurcate the trial, so that the jury first determines guilt on the current charge without hearing about prior convictions, and then—if necessary—the court addresses the sentencing enhancement separately. The government is opposing these motions, arguing that bifurcation is inefficient, but the risk of prejudice to the defendant is simply too great to ignore.

If you or a loved one is facing federal sentencing with prior state convictions, the Taylor decision has fundamentally changed the legal landscape, and you need experienced counsel who understands these new complexities. My firm has already successfully challenged government attempts to introduce unreliable police reports in three separate ACCA cases since the ruling, and we are actively developing constitutional arguments to limit the decision's reach. Do not let a sentencing judge rely on decades-old documents that may be incomplete, inaccurate, or misleading. Contact our office today for a confidential consultation—your liberty depends on having a defense team that understands how to navigate this unprecedented departure from settled law.