Why the Ninth Circuit’s Relevant Conduct Decision Is a Dangerous Departure From Federal Sentencing Law

The Ninth Circuit's Inference Doctrine: A Procedural End-Run Around the Sentencing Reform Act

In my 25 years as a federal prosecutor and now as a criminal defense attorney, I have never seen a circuit court decision that so fundamentally rewrites the ground rules of federal sentencing as the Ninth Circuit's recent opinion in United States v. Martinez. The court held that a district judge may infer that a defendant engaged in relevant conduct—even where no evidence directly connects the defendant to that conduct—so long as the inference is "logical and reasonable" based on the totality of the circumstances. This is not a subtle shift in sentencing law; it is a seismic departure from the statutory framework Congress enacted in the Sentencing Reform Act of 1984. The Sentencing Guidelines, codified at 28 U.S.C. § 994, require that relevant conduct be proven by a preponderance of the evidence, not by mere logical leaps. The Ninth Circuit's new standard effectively lowers the government's burden to something akin to a "reasonable suspicion" standard at sentencing, which is precisely what Congress sought to avoid when it mandated that sentencing decisions be based on reliable factual findings. This decision invites district judges to speculate about a defendant's conduct without any evidentiary foundation, and it places defense counsel in the impossible position of disproving phantom allegations that the government never needed to prove at trial.

The practical consequences of this decision are devastating for defendants who exercise their constitutional right to trial. In the case before the Ninth Circuit, the defendant was convicted of a single count of possession with intent to distribute methamphetamine, but the district court attributed to him over 500 grams of drug weight from a co-conspirator's activities that occurred months before the defendant's arrest. The only evidence connecting the defendant to that drug weight was the testimony of a cooperating witness who admitted he had never seen the defendant handle those drugs. The district court simply inferred that because the defendant associated with the co-conspirator, he must have been involved in the broader conspiracy. Under the pre-Martinez framework, this inference would have been rejected as speculative under United States v. Ameline (9th Cir. 2005), which required that relevant conduct findings be supported by "sufficient indicia of reliability." The Ninth Circuit's decision effectively overrules Ameline sub silentio, because if a logical inference is sufficient, there is no longer any requirement that the evidence itself carry indicia of reliability. This creates a perverse incentive for prosecutors to decline to charge relevant conduct in the indictment, knowing they can introduce it at sentencing through inference without ever having to prove it beyond a reasonable doubt.

The statutory foundation of federal sentencing law rests on 18 U.S.C. § 3553(a), which directs courts to impose sentences that are "sufficient, but not greater than necessary" to achieve the purposes of sentencing. This parsimony principle is incompatible with a sentencing system that allows judges to pile on unproven conduct based on nothing more than a prosecutor's argument that an inference is "reasonable." The Sentencing Commission's own policy statement at U.S.S.G. § 1B1.3 explicitly defines relevant conduct as "acts and omissions committed, aided, abetted, counseled, commanded, induced, procured, or willfully caused by the defendant." Notice that the Guideline does not say "acts that a judge can reasonably infer the defendant committed." It says acts the defendant actually committed. The Ninth Circuit's decision reads this requirement out of the Guidelines entirely, because if a judge can infer conduct from mere association, then the distinction between actual conduct and inferred conduct collapses. I have litigated hundreds of sentencing hearings, and I can tell you that this decision will produce wildly inconsistent sentences across the circuit, with some judges drawing aggressive inferences and others requiring actual evidence—all under the guise of applying the same legal standard.

The Collision With Federal Rule of Criminal Procedure 32: How the Ninth Circuit Undermines the District Court's Fact-Finding Obligations

Federal Rule of Criminal Procedure 32(i)(3)(B) imposes a mandatory duty on district courts to "rule on any unresolved dispute about a fact in the presentence report" or to "determine that a ruling is unnecessary because the matter will not affect sentencing." This is not a discretionary suggestion; it is a procedural command that ensures defendants have a meaningful opportunity to contest the factual basis for their sentences. The Ninth Circuit's inference doctrine directly undermines this rule by allowing district courts to sidestep factual disputes altogether. Under Martinez, a judge can simply announce that a reasonable inference supports a finding of relevant conduct, and then treat any defense objection as irrelevant because the inference itself constitutes the factual finding. This is procedurally indefensible. The Advisory Committee Notes to Rule 32 make clear that the district court must "make a finding as to each controverted matter" and that "the court cannot simply adopt the presentence report's factual findings without resolving the defendant's objections." Yet that is precisely what the inference doctrine permits: a judge can adopt an inference that the government suggests, declare it reasonable, and then treat the defendant's objection as moot because the inference has already been drawn.

I have seen this play out in practice in a post-Martinez sentencing hearing I observed in the Central District of California. The defendant objected to the presentence report's attribution of 400 grams of cocaine base, arguing that the only evidence was a confidential informant's uncorroborated statement that the defendant was "involved" with a drug trafficking organization. The government conceded it had no direct evidence but argued that the court could "reasonably infer" involvement because the defendant lived in the same apartment complex as a known dealer. The district court accepted this argument, stating on the record that "it's a logical inference" without making any finding about the reliability of the informant's statement. Under pre-Martinez law, this would have been reversible error because the court failed to assess whether the informant's statement carried "sufficient indicia of reliability" under United States v. Huckins (9th Cir. 1995). But under the new inference standard, the court's failure to evaluate reliability was deemed harmless because the inference itself was deemed reasonable. This is a profound procedural abdication that strips defendants of their right to have the government prove relevant conduct by competent evidence.

The implications for appellate review are equally troubling. Under the pre-Martinez framework, the Ninth Circuit reviewed relevant conduct findings for clear error, meaning the appellate court would examine the record to ensure the district court's findings were supported by evidence. But if a district court can now base a finding on a "reasonable inference" without any underlying evidentiary support, what exactly is the appellate court reviewing? The Ninth Circuit's opinion suggests that the clear-error standard still applies, but in practice, it is nearly impossible to overturn a finding based on an inference because the appellate court must defer to the district court's assessment of reasonableness. This creates a virtually unreviewable sentencing discretion that Congress never intended. The Supreme Court in Gall v. United States (2007) emphasized that appellate courts must ensure that district courts do not abuse their discretion, but the Ninth Circuit's inference doctrine effectively insulates relevant conduct findings from meaningful appellate review. I advise every defense attorney practicing in the Ninth Circuit to object to any inference-based finding and to specifically request that the district court identify the specific evidence supporting each inference, as required by United States v. Romero-Rendon (9th Cir. 2000), which held that "a sentencing judge may not rely on speculation or unsubstantiated allegations."

The Dangerous Precedent for Acquitted Conduct and the Erosion of the Beyond-a-Reasonable-Doubt Standard

The most troubling aspect of the Ninth Circuit's decision is its implicit endorsement of sentencing enhancements based on acquitted conduct. While the Supreme Court held in United States v. Watts (1997) that a sentencing court may consider acquitted conduct if proven by a preponderance of the evidence, the Court also cautioned that "a jury's verdict of acquittal does not prevent the sentencing court from considering conduct underlying the acquitted charge if that conduct has been proven by a preponderance of the evidence." The key phrase is "proven by a preponderance." The Ninth Circuit's inference doctrine effectively eliminates the requirement of proof, because an inference is not proof—it is a conclusion drawn from assumptions. In Martinez, the defendant had been acquitted of conspiracy, yet the district court used the same conduct that formed the basis of the conspiracy charge to enhance his sentence on the possession count. The Ninth Circuit held that the acquittal was irrelevant because the court could infer the defendant's involvement based on his association with co-conspirators. This directly contradicts the Supreme Court's holding in United States v. Booker (2005) that the Sixth Amendment requires any fact increasing the statutory maximum to be found by a jury beyond a reasonable doubt, and while relevant conduct does not increase the statutory maximum, it does increase the Guidelines range, which remains a critical factor in sentencing.

The practical effect of this decision is that prosecutors can now obtain sentencing enhancements for conduct that a jury explicitly rejected, so long as the district court can articulate a "logical inference" that the defendant engaged in that conduct. This creates a perverse incentive for prosecutors to overcharge in the indictment, knowing that even if the jury acquits on the overcharged counts, the conduct can still be used at sentencing. I have seen this tactic used repeatedly in my practice, and it fundamentally undermines the finality and integrity of jury verdicts. The Sentencing Commission's data shows that defendants who go to trial receive sentences that are, on average, 300% longer than defendants who plead guilty, and the Ninth Circuit's decision will only widen this disparity. When a district court can infer relevant conduct from acquitted charges, the trial penalty becomes even more severe, punishing defendants for exercising their constitutional right to put the government to its proof. This is precisely the kind of sentencing disparity that the Sentencing Reform Act was designed to eliminate, and it is deeply troubling to see a circuit court embrace a standard that rewards prosecutorial overreach.

The constitutional implications extend beyond the Sixth Amendment to the Due Process Clause of the Fifth Amendment. The Supreme Court held in McMillan v. Pennsylvania (1986) that sentencing factors must be proven by at least a preponderance of the evidence, and in Apprendi v. New Jersey (2000), the Court emphasized that "the relevant inquiry is one not of form, but of effect." If the Ninth Circuit's inference standard allows a judge to impose a sentence that is effectively based on acquitted conduct without any evidentiary foundation, then the effect is to deprive the defendant of due process. The Ninth Circuit's opinion attempts to distinguish Apprendi by noting that relevant conduct does not increase the statutory maximum, but this misses the point. The Guidelines range remains the starting point for every federal sentence, and a dramatic increase in the Guidelines range based on inferred conduct has a profound effect on the ultimate sentence. In Martinez, the defendant's Guidelines range increased from 51-63 months to 168-210 months based on the inferred drug weight—a tripling of the sentence based on nothing more than a logical inference. This is not the system Congress created, and it is not the system that the Constitution permits.

Strategic Imperatives for Defense Counsel in the Post-Martinez Landscape

Given the Ninth Circuit's dangerous departure from settled sentencing law, defense counsel must immediately adapt their advocacy to protect their clients from inference-based sentencing enhancements. The first and most critical step is to file a pre-sentencing memorandum that specifically challenges any relevant conduct attribution that relies on inference rather than direct evidence. I recommend citing United States v. Ameline (9th Cir. 2005) and United States v. Huckins (9th Cir. 1995) for the proposition that relevant conduct must be supported by "sufficient indicia of reliability," and arguing that Martinez did not overrule these precedents because they remain binding on district courts. The government will inevitably cite Martinez for the inference standard, so counsel must be prepared to distinguish the case by arguing that the inference in Martinez was supported by some evidence—specifically, the co-conspirator's testimony that the defendant was present during drug transactions—while in your client's case, there is no evidence whatsoever. This distinction is critical because even the Ninth Circuit acknowledged that "reasonable inferences must be grounded in the record, not in speculation."

Second, counsel must demand specific factual findings under Rule 32(i)(3)(B) and insist that the district court identify the precise evidence supporting each inference. I recommend submitting proposed findings of fact that require the court to specify: (1) what evidence supports the inference, (2) why that evidence is reliable, and (3) how the inference logically follows from the evidence. If the district court refuses to make these findings, counsel should object on the record and preserve the issue for appeal. The Ninth Circuit held in United States v. Romero-Rendon (9th Cir. 2000) that a district court's failure to resolve factual disputes is reversible error, and this holding remains good law even after Martinez. Counsel should also request a continuance of the sentencing hearing if the government introduces new evidence of relevant conduct at the hearing, because Rule 32(i)(1)(C) requires that the parties be given "an opportunity to comment on the probation officer's determinations and on other matters relating to an appropriate sentence." If the government springs inference arguments at the hearing without prior notice, counsel should object that this violates due process and the rule's notice requirements.

Third, counsel should consider filing a motion to strike any relevant conduct allegations that are based solely on speculation or association, citing the Sentencing Commission's policy statement at U.S.S.G. § 1B1.3 and the commentary that relevant conduct must be "within the scope of the jointly undertaken criminal activity." The commentary specifically states that "the defendant's accountability for the conduct of others is limited by the scope of the agreement and the foreseeability of the conduct." If the government cannot show that your client agreed to the conduct or that it was foreseeable, then the inference standard cannot save the government's case. I have successfully used this argument in several post-Martinez hearings by focusing on the lack of evidence that my client knew about or agreed to the conduct attributed to him. The Ninth Circuit has long held in United States v. Banuelos (9th Cir. 2003) that "a defendant may not be held accountable for the conduct of others unless that conduct was within the scope of the defendant's agreement and was reasonably foreseeable," and this holding directly limits the reach of the inference doctrine.

Finally, counsel must preserve the record for appeal by objecting to any inference-based finding and by requesting that the district court state on the record whether it would impose the same sentence if the relevant conduct finding were reversed on appeal. This is critical because if the district court states that it would impose the same sentence regardless of the relevant conduct finding, the error may be deemed harmless under United States v. Munoz-Camarena (9th Cir. 2011). However, if the district court refuses to make such a statement, counsel should argue that the error is structural and requires resentencing. The Ninth Circuit has not yet addressed whether a violation of the inference standard is structural error, but I believe a strong argument can be made that it is, because the error goes to the very heart of the sentencing process—the factual basis for the sentence. The Supreme Court in Neder v. United States (1999) held that harmless error analysis applies to most constitutional errors, but the Court also recognized that structural errors "defy analysis by harmless error standards." If a district court imposes a sentence based on an inference that has no evidentiary support, the entire sentencing proceeding is fundamentally unfair, and no amount of harmless error analysis can cure that defect.