White Collar Sentencing in Federal Court

The sentence almost never corresponds to what the defendant imagined. In federal white collar cases, the distance between expectation and outcome is not a gap; it is a chasm, and it opens widest at the moment a loss figure enters the record. The United States Sentencing Guidelines treat economic crime as an arithmetic problem. The answer, expressed in months, follows from a table. What the table cannot express is the quiet terror of watching a judge accept a number that will dictate the next decade of a life.

Loss Is the Engine of Federal Sentencing

Under Section 2B1.1 of the Guidelines, wire fraud, mail fraud, and bank fraud each begin at a base offense level of 7. That number means almost nothing on its own. What determines the sentence is loss: actual loss, intended loss, or the greater of the two. A fraud involving $150,000 adds 10 levels. At $1.5 million, the addition is 16 levels. At $9.5 million, 20 levels. The table contains 15 graduated categories, and by the time loss exceeds $550 million, the enhancement alone can produce a Guidelines range recommending a century of incarceration.

Consider the arithmetic as it applied to Samuel Bankman-Fried in March of 2024. The government calculated intended loss at no less than $10 billion. His attorneys argued that no customer had suffered actual loss, that the bankruptcy estate would make everyone whole. Judge Kaplan rejected that argument as misleading and logically flawed. He found losses exceeding $550 million to FTX customers, $1.7 billion to investors, $1.3 billion to Alameda lenders. The Guidelines recommended up to 110 years. The sentence imposed was 25 years, plus $11 billion in forfeiture. A quarter century in prison was, in context, a mercy.

That is what loss calculation does. It converts a financial event into a carceral one, and it does so with a precision that can feel arbitrary. A defendant whose scheme caused $400,000 in loss occupies a different moral universe, in the eyes of the court, than one whose scheme caused $4 million. Whether the defendant with the larger number is proportionally more culpable is a question the Guidelines do not ask.

The Numbers Tell a Strange Story

Between fiscal year 1986 and fiscal year 2024, the median federal prison sentence for all white collar offenses was 6 months. The average was 19 months. Those figures are deceptive in their modesty. They include the full range of cases: the accountant who inflated deductions, the wire fraud defendant who accepted a plea to a lesser count, the first offender who cooperated before indictment. They do not capture what happens at the top of the loss table.

In the first half of fiscal year 2025, the median climbed to 14 months and the average to 27 months. That shift is not incidental. It reflects a prosecutorial apparatus that, while filing fewer cases, is selecting larger ones. U.S. Attorney offices filed 4,332 white collar prosecutions in fiscal year 2024. Three decades earlier, in 1994, the number was 10,269. Projections for fiscal year 2025 suggest a further decline to 3,862. The government is prosecuting less than half as many white collar cases as it once did. The cases it does bring tend to involve greater sums, greater complexity, greater exposure.

The Southern District of New York sentenced 1,048 individuals in fiscal year 2024. The average sentence there was 59 months. The median was 36 months. Those numbers belong to a different country than the national figures. They belong to a jurisdiction where the cases are larger, the judges are more experienced with financial crime, and the prosecutors have the resources of Main Justice behind them.

What Happened in 2024 Was Not Routine

The cases that closed in 2024 and early 2025 represent something other than ordinary enforcement. Robert Kowalski received 25 years for embezzlement tied to the collapse of Washington Federal Bank. Abraham Yusuff received over 14 years for orchestrating a $110 million identity theft refund fraud against the IRS. Jack Fisher and James Sinnott, a CPA and an attorney respectively, were convicted for selling more than $1.3 billion in fraudulent conservation easement deductions that generated $450 million in tax losses; both received sentences exceeding two decades. Former Los Angeles City Councilman Jose Huizar received 13 years for bribery and racketeering.

Changpeng Zhao, the founder of Binance, received four months. Four months for the head of the world’s largest cryptocurrency exchange, whose platform processed billions in transactions with inadequate compliance controls. The disparity between Zhao’s sentence and Bankman-Fried’s is not explained by the Guidelines alone. It is explained by cooperation, by the nature of the charges, by the presence or absence of identifiable victims. It is explained, in other words, by the thousand discretionary judgments that occur between indictment and sentencing.

The fraud section of the Department of Justice publicly charged 265 individuals and convicted 235 in 2025, through a combination of guilty pleas and trials. Those numbers represent a fraction of the conduct that occurs. They represent the conduct the government chose to see.

The Commission Is Reconsidering the Table

Something is changing. In August of 2025, the United States Sentencing Commission issued a notice of proposed priorities that included reassessing the role of actual loss, intended loss, and gain in the Guidelines. The Commission signaled interest in simplifying the loss table, adjusting it for inflation, and, more significantly, decreasing the emphasis on loss amount in favor of culpability factors: the defendant’s role as leader or minor participant, the level of planning and sophistication, abuse of trust, the type and number of victims.

In December of 2025, proposed amendments to Section 2B1.1 were published. Part A would restructure and simplify the loss table itself. Part B would amend the specific offense characteristics and add new ones designed to capture individual culpability and victim harm with greater precision. If adopted, these amendments would take effect on November 1, 2026. They represent the most significant potential revision to white collar sentencing in a generation.

The Commission had already moved in 2024. On April 17 of that year, it voted unanimously to adopt two amendments of consequence. The first moved the definitions of actual loss and intended loss from the commentary into the Guideline text itself, giving those definitions the force of law rather than interpretive guidance. The second, and more profound, prohibited courts from using acquitted conduct to calculate a defendant’s Guidelines range. A defendant found not guilty of certain counts could no longer see those counts increase the sentence on the counts of conviction. The amendment, effective November 1, 2024, was the culmination of years of criticism from jurists across the ideological spectrum, including sitting Supreme Court Justices who had urged the Commission to act.

One should be precise about the scope of that reform. The acquitted conduct amendment applies only to federal acquittals. It does not address uncharged conduct, dismissed counts, or state acquittals. And judges retain authority under 18 U.S.C. Section 3661 to consider acquitted conduct outside the Guidelines framework. The amendment limits the arithmetic. It does not eliminate the discretion.

Variance Is Where the Defense Lives

The Guidelines are advisory. That has been true since United States v. Booker in 2005, and the consequences of that advisory status are most visible in white collar sentencing. Among defendants who proceed to trial in white collar cases, 53.4 percent receive downward variances. Among those who plead guilty, 28.6 percent receive non-government-sponsored downward variances. The average sentence for white collar defendants who go to trial and receive a variance is 30.6 months.

Those numbers mean that the Guidelines range is a starting point, not an ending point, and that persuasion at sentencing is not a formality. It is the proceeding that matters most. The factors that produce variances are not mysterious: personal history, community ties, rehabilitation efforts before sentencing, the defendant’s role relative to other participants, the degree to which restitution has been made or can be made. But presenting those factors requires a kind of preparation that begins months before the sentencing hearing, and it requires counsel who understands that the Guidelines manual is a document to be argued against as much as it is a document to be applied.

We have represented individuals facing the full range of white collar charges in federal court. The common error is to treat sentencing as a postscript to the trial or the plea. It is not a postscript. It is, for most defendants, the only proceeding that will determine whether they serve 18 months or 18 years. The distance between those outcomes is not determined by the facts alone. It is determined by how the facts are framed, which facts are emphasized, which mitigating circumstances are documented, and whether the court is given a reason to exercise its discretion.

The Loss Table Is a Blunt Instrument Applied to a Precise Problem

There is something disquieting about a system that can recommend 110 years for a 30 year old defendant on the basis of a spreadsheet. The Guidelines were designed to reduce disparity. In white collar cases, they have produced a different kind of disparity: between defendants whose loss amounts are high and whose culpability is low, and defendants whose loss amounts are modest and whose conduct was predatory. A first time offender involved in a scheme that generated $10 million in intended loss but caused no actual harm faces a higher Guidelines range than a repeat offender who defrauded elderly victims of their savings. The table does not distinguish between the two. The judge must.

DEFENSE TEAM SPOTLIGHT

Todd Spodek

Lead Attorney & Founder

Featured on Netflix's "Inventing Anna," Todd Spodek brings decades of high-stakes criminal defense experience. His aggressive approach has secured dismissals and acquittals in cases others deemed unwinnable.

NY Bar Admitted Multi-State Licensed Federal Courts

Meet the Full Team

The Sentencing Commission’s proposed reforms acknowledge this. By shifting weight from loss amount toward culpability and victim impact, the Commission is attempting to correct an imbalance that practitioners have identified for years. Whether those reforms will survive the amendment process is uncertain. Whether Congress will act to disapprove them is possible but unlikely. What is certain is that the current framework produces outcomes that even its architects recognize as imperfect.

I have watched clients receive sentences that bore no rational relationship to the harm they caused, and I have watched other clients receive sentences that were, given the circumstances, as fair as the system permits. The difference was not the judge. The difference was the preparation.

What a Defense Requires

A white collar sentencing defense begins at the moment of retention, not at the moment of conviction. It requires a complete financial analysis of the loss calculation the government will propose. It requires identification of every applicable reduction: acceptance of responsibility, minor role, aberrant behavior, the new zero point offender provision for defendants with no criminal history. It requires a sentencing memorandum that presents the defendant as a person and not as a cell in a spreadsheet.

The Sentencing Commission reported that Amendment 821, focused on defendants with little or no criminal history, could affect over 7,000 defendants with an average reduction of 15 months, and over 11,000 defendants with status points with an average reduction of 14 months. Those are not abstractions. Those are months returned to lives. Every reduction that applies must be identified, documented, and argued.

The government will present its version of the loss. It will present its version of the defendant’s role. It will present its version of the victims’ suffering. The defense must present a counter-narrative that is not less rigorous but more rigorous, grounded in the same data but organized around a different principle. The principle is this: the sentence must be sufficient but not greater than necessary to comply with the purposes of sentencing. That language comes from 18 U.S.C. Section 3553(a). It is the most important sentence in federal criminal law, and it is the sentence that governs every argument we make.

The first call is not a commitment. It is a diagnosis.