United States v. Griffin

Opinion

United States Court of Appeals For the First Circuit

Nos. 24-1475 24-1540 24-1541

UNITED STATES OF AMERICA, Appellee,

v.

WILLIAM W. ROBERTSON and DANIEL J. GRIFFIN, Defendants, Appellants.

APPEALS FROM THE UNITED STATES DISTRICT COURT FOR THE DISTRICT OF MASSACHUSETTS

[Hon. Margaret R. Guzman, U.S. District Judge]

Before

Gelpí, Lynch, and Thompson, Circuit Judges.

William W. Fick, with whom Daniel N. Marx, Amy Barsky, and Fick & Marx, LLP, were on brief, for appellant Robertson.

Thomas M. Hoopes, with whom Jason M. Strojny and Libby Hoopes Brooks & Mulvey, P.C., were on brief, for appellant Griffin.

Randall E. Kromm, Assistant United States Attorney, with whom Leah B. Foley, United States Attorney, was on brief, for appellee.

December 19, 2025

THOMPSON, Circuit Judge. Between 2015 and 2017, Lieutenant Daniel Griffin and Sergeant William Robertson of the Massachusetts State Police ("MSP") clocked hundreds of hours of overtime to the tune of practically $100,000. (And that's not counting the troopers they supervised.)

The problem? A lot of that time, they weren't actually working. Instead, they routinely showed up late or cut out early yet still charged for the whole shift. They also began overtime- specific tasks during their normal work hours but still billed like they had worked a separate overtime shift later in the day, an impermissible form of double-dipping. And they regularly encouraged their subordinates to do all that, too.

The Feds were none too pleased once they caught wind of this scheme, because the money that these troopers (and their supervisees) stole came from federal grants funding overtime hours for law enforcement initiatives to promote highway safety. (Think DUI checkpoints and "Click-It-Or-Ticket" programs.) So Robertson and Griffin were charged with a host of fraud-based federal crimes.

After an extended trial featuring more than twenty testifying witnesses, a jury convicted the duo of all the charges: wire fraud, theft of federal funds, and conspiracy to commit the latter. (And on the brink of trial, Griffin separately pled guilty to a set of tax falsification charges and some unrelated wire fraud charges.) The district court then sentenced Robertson and Griffin

3

to some time behind bars, and it ordered each defendant to fork over what they made during the scheme (and some more) via substantial restitution and forfeiture orders.

On this omnibus appeal, the pair protests the propriety of the prior proceedings practically soup-to-nuts. Mostly, we affirm, though with a bit of disquietude about the district court's tone during Robertson's sentencing (more on that later). But we vacate and remand one narrow issue: the forfeiture order against Griffin. Read on as we explain what led to this mega-appeal. We'll start with Griffin and Robertson's time with the MSP, their fraud(s), the investigation, and the proceedings below.

HOW WE GOT HERE

A. The Defendants

We kick things off with a bit about our defendants'

histories, both personal and professional.1 A Massachusetts native

To start our recitation of the facts, we note a wrinkle 1

complicating our presentation of the record. One of Robertson's claims is a challenge to the sufficiency of the evidence, and "we typically recite those facts relevant to sufficiency claims" in "the light most favorable to the verdict." United States v. Burgos-Montes, 786 F.3d 92, 99 (1st Cir. 2015). But that's not the only challenge here, and for others (like the constitutional challenge or the facts underlaying the application of a sentencing enhancement), we're instead supposed to take a more "balanced" treatment where we "objectively view the evidence of record." Id. (cleaned up). Yet we can't "simultaneously recite the facts in both manners," so we'll instead "limit our initial summary of this lengthy record to those details essential to framing the issues on appeal." Id. Further complicating this is (1) that the district court is afforded some discretion in factfinding for sentencing

4

with a blue-collar upbringing and a college education, Griffin joined the MSP in 1986 at age 23. In 2011, he was promoted to Lieutenant, a rank he held until his retirement in 2020. And from 2011 to 2019, he led MSP's Traffic Programs Section ("TPS"), the division whose conduct gave rise to this case. (More on TPS shortly.) While at MSP, Griffin also founded a private security company. (More on that company shortly, too.)

Robertson, also a Massachusetts native, joined the force in 1986 (the same year as Griffin) at age 26. He served for thirty-two years, and throughout that long tenure (all as a Massachusetts State Trooper), he earned several awards for his service and departed at the rank of Sergeant. And while those accolades are surely commendable, they aren't why we write. Instead, relevant to this appeal, Robertson joined TPS as a Sergeant in 2008 and got involved with its doings (both good and not so good) until he was transferred in 2018.

So, what is TPS? It's a small division (usually around five troopers at a time) that seems best understood in large part as a federal/state administrative go-between: it receives hundreds of thousands of dollars in grants from the U.S. Department of Transportation, and then it oversees the distribution of those grants into funding for traffic safety initiatives throughout the

purposes and (2) that some facts relate not to the trial but to Griffin's guilty plea. But we'll get more into each later.

5

Commonwealth.2 The TPS's leader (at the time relevant here, Griffin) is also responsible for applying for grants to fund these programs. Along with all those administrative responsibilities, TPS also directly operates some programs itself.

That last point is what really matters for our purposes.

The trial focused closely on Griffin and Robertson's overtime fraud related to two buckets of federally funded traffic safety programs that TPS directly implemented: sobriety checkpoints and safety initiatives (like "Click-It-Or-Ticket"). So that's where we next turn.

B. The Scheme

We'll start with the sobriety checkpoints and how they were supposed to be operated. On Friday and Saturday nights, troopers could sign up to work eight-hour shifts at BAT (short for "breath alcohol testing") checkpoints on highways. Much of that time would be spent in the "BAT Mobile," a camper parked by a sobriety checkpoint and manned by troopers where identified drunk

2 We'd be wrong to say TPS is the only go-between. So too is the Executive Office of Public Safety and Security, another branch of Massachusetts's government that also works with the federal government to secure money and, for at least some of these grants, was the direct source of the money. But we'll spare the reader from more details of this seemingly endless bureaucratic labyrinth unless they're legally material here (or otherwise absolutely necessary).

6

drivers would be processed and held until they could get escorted to jail.3 The eight-hour shifts were often geared towards evening surveillance, with a start time of 8:00 p.m. and running until 4:00 a.m. It was generally understood that because the overtime was governed by federal rules and interdepartmental service agreements (or "ISAs"), troopers were obligated to work the full eight-hour shift in order to be paid for the full shift.4 (That's apparently unlike other MSP-governed shifts, which aren't as stringent in their billing practices.)

The Griffin-Robertson fraudulent twist on these? For one, the TPS trooper team members often wouldn't report to the barracks (where they'd load up onto the BAT Mobile) until 10:00 p.m., despite the billing clock starting at 8:00 p.m. Griffin, on at least one instance, arrived at 11:30 p.m. And the troopers often would wrap up and go home at 1:00 or 1:30 a.m., but, at

3 We assume that the name of the "BAT Mobile" finds its origins in the famous vehicle of the same name owned by the Caped Crusader and Dark Knight himself: the DC Comics superhero "Batman." We'll note that corrupt coppers also play an important part in Christopher Nolan's The Dark Knight (2008), but thankfully, the inappropriate conduct of the MSP troopers in this case doesn't come close to matching the deep-rooted corruption and greed in the Gotham Police Department on display throughout the legendary film.

4For instance, one ISA required that personnel "paid under this grant must actively participate in the enforcement effort." According to one government official, it was expected that reported overtime hours "would be accurate for actual hours worked."

7

Griffin's direction, would still bill as if they worked until 4:00 a.m. to get their "full time" pay. Troopers testified to knowing it was wrong but still doing it anyway -- relying on Griffin's orders.

Along with sobriety checkpoints were a hodgepodge of other federally funded road safety initiatives, our second bucket. These initiatives paid for shifts where troopers would conduct traffic stops specifically for unbelted drivers (to encourage seatbelt use) and those glued to their phone or otherwise inattentive at the wheel (to discourage distracted driving). Troopers could sign up to do four-hour blocks on these shifts, but importantly, they couldn't "pyramid" their shifts, meaning they couldn't charge time for federally funded initiatives while they were still on the clock for their normal MSP work hours. Ordinarily, a federally funded traffic shift would start at 3:30 p.m., after the regular TPS workday (which ran from 7:00 a.m. to 3:30 p.m.) ended.

The Griffin/Robertson nefarious spin on these is a bit more nuanced. For one, the TPS troopers would indeed "pyramid" shifts, and thus improperly double dip. (And we know that they pyramided because the troopers, on at least once instance, issued a grant-related safety ticket at 3:25 p.m., despite their regular TPS hours running until 3:30 p.m.) For another, after the officers would write eight or nine tickets at the first location that they

8

posted up at, Griffin would give them a signal over the radio: "Let's move to another location." But that other location wasn't really a cue to move to a new spot on the interstate; it was his code to his team to pack it up and go home. Griffin (as the officer-in-charge of TPS) would then approve for a full four hours' worth of labor, despite only an hour or two actually worked. In one trooper's recollection, TPS "never" worked the full shift, and the troopers knew it was wrong the whole time.

This went on for years under the radar and without notice. But there is one noteworthy episode to mention here that, as you will learn, crucially factors into the scheme's unraveling. One day at the barracks in 2017, apparently nervous about the discovery of similar overtime fraud in other MSP sections, Robertson approached Trooper Dennis Kelley, a lower-ranking TPS member.5 Robertson smacked a folder in Kelley's chest and told Kelley to "get rid of these." Robertson then got in Kelley's face and said, "Shred these fucking things. Go shred them." Kelley did, and as the files were going into the shredder, he noticed that they were "637s," an important billing form. Once he realized

5 Details of the fraud investigation that caused Robertson's outburst are thin in Kelley's testimony. And throughout the trial, discussion of fraud in other units was skirted around, presumably so as to avoid prejudicing the defendants too much. But as the reader will later learn in our discussion of sentencing disparities, Robertson and Griffin certainly aren't the only two MSP troopers with overtime fraud convictions these days.

9

what was happening -- that he was destroying TPS documents that by federal rule should have been preserved -- he expressed his dismay to Robertson. But Robertson at that point didn't care: the deed was done.6 C. The Investigation

By now, the thoughtful reader might be wondering how, exactly, someone figured out that Griffin, Robertson, and their small squad were skimming off the top of these programs. If everyone in TPS was in on it and the files were destroyed, who would ever know?

The answer: we can point to and commend a scrupulous federal agent working for the Department of Transportation's Office of the Inspector General. One day, Agent Tayna Chavez noticed oddities in TPS's billing while investigating potential fraud in other MSP units. Thinking it strange that the same troopers were working the same number of hours at the same time,

6 Later on, Kelley learned that higher-ups (higher than Griffin and Robertson) were sniffing around and looking for important TPS billing forms. Kelley then "got scared" and "did the stupidest thing" he had "ever done." He took home the rest of the files and burned them because he was "afraid" that the MSP higher-ups "were going to find out" that he destroyed other files at Robertson's direction. In his words, all of those files "would have showed everything that we had done there with the times and the hours and everything." Although Kelley didn't say that Robertson told him to do it, Kelley called Robertson to tell him that the files were destroyed; to that news, Robertson simply remarked "Good."

10

week after week (after week), Agent Chavez started to piece together some more information about their activities.

Some serious sleuthing ensued. Agent Chavez pulled together the following records:

• "Paystation" data, to see what hours the troopers were reporting and what the officers were approving;

• additional paperwork (dubbed "636," "637," and "638"

forms) to accomplish those same ends;

• fuel records, to figure out how far, when, and in which cars the troopers drove;

• cruiser "radio affiliation" data, to discern when the troopers were turning their cars on and off, and to see when the cars' radios were pinging certain radio towers to triangulate their location; and • citations that the troopers issued during their shifts, to see what patterns, if any, were discernable from their ticketing practices.

From all that, Agent Chavez and her DOT compatriots knew something was way off. So, after postulating from the data that TPS must have been short-shrifting the shifts, they ran the numbers to figure out the estimated loss arising from the inside dealing. And they ultimately attributed hundreds of "loss hours" (time

11

charged but not worked) to the troopers for each year between 2015 and 2017.

D. The Side Hustles

We'll digress here for a minute to talk about what other info happened to surface during the investigation. All of the above exposition related to Griffin and Robertson's shared TPS skullduggery, but as it turned out, Griffin simultaneously had his own separate unholy schemes on the side. And because the reader will want to know about that by the time we start discussing the indictment (coming soon), we'll explain it now.

Recall that we earlier mentioned how Griffin also ran a private security business. It was called Knight Protection Services, or "KNIGHTPRO." Back in the 1990s, MSP permitted him to run KNIGHTPRO (which focused on private security details for college events as well as art transportation), but it came with an important caveat: he couldn't let it conflict with his MSP work hours or responsibilities.

Well, it did. The first conflict for this officer of the law was straight-up tax fraud: from 2012 to 2019, KNIGHTPRO records showed about $1.9 million in revenue deposits, but Griffin only reported $1.2 million to the IRS -- the point being that he hid over $700,000 of income from the IRS during that period. And that's especially alarming given that he held KNIGHTPRO out to the IRS as a money-pit almost every year. But that false reporting

12

wasn't all. Between 2014 and 2019, Griffin also siphoned off tens of thousands of dollars in KNIGHTPRO money to pay for a country club membership, home improvements for his Cape Cod property, and other personal expenses, while, for tax purposes, writing off those same remittances as business expenses.

And that leads to Griffin's second conflict, this one specifically MSP-themed. Several times from 2015 through 2017, Griffin used his KNIGHTPRO credit card to pay for MSP-related training and travel. Afterward, he wrote off those KNIGHTPRO charges as business expenses (again, for tax purposes) but also applied for and collected personal reimbursement from MSP's federal grant money for those same outlays. (If it isn't obvious, that kind of double-dipping isn't cool.)

All told, the IRS figured out that Griffin was about $186,389.00 short on his tax payment obligations through the years.

But the IRS and MSP weren't the only ones that Griffin pulled a fast-one on. Starting in 2013, he began telling his children's private school for boys, the Belmont Hill School ("BHS"), that he was making less money than he did. The reason? So he could benefit from the school's financial aid program.7 Over

7 If the reader will indulge us, we'll share a vignette from the record. At one point, the Griffins were awarded too much financial aid when their application was confused with that of another family. They wrote to the school to let it know about the mix-up. In response, the school sent an official letter correcting the amount. And on the letter, the school's financial aid director

13

seven years, his family obtained $176,700 in financial aid from BHS. (As we'll later discuss, it's an open question how much of that, exactly, stems from Griffin's fraud.)

E. The Proceedings Below

All that fraud naturally spawned a court case. And that case started in 2020, when the government obtained an indictment against Griffin and Robertson charging the following:

• one count of conspiracy to commit theft concerning programs receiving federal funds, in violation of 18 U.S.C. § 371;

• one count of theft concerning programs receiving federal funds, in violation of 18 U.S.C. § 666(a)(1)(A); and • four counts of wire fraud, in violation of 18 U.S.C.

§ 1343.

These charges all related to the overtime fraud scheme. Separately, the indictment charged Griffin with four additional counts of wire fraud relating to the BHS financial aid applications and ten counts of filing false tax returns related to KNIGHTPRO's cooked books.8

added a hand-written note to Griffin in large print: "Again, I can't thank you enough for your honesty!" In hindsight, those thanks may have been a bit premature.

The careful reader might be wondering where the MSP 8

underlings are in all this. Most of the ordinary troopers involved in TPS took immunity deals with the government to avoid prosecution, and they testified against Griffin and Robertson in

14

Unremarkable pretrial proceedings went on for nearly three years, teeing up the case for trial.

Before opening arguments, Griffin opted to fess up to the BHS wire fraud and KNIGHTPRO tax return counts (and thus pled guilty to them). Then a jury trial commenced to thrash out the remaining counts. We'll recount a bit more of what happened at the trial later on, but for now, we'll note that the defendants thrice moved for judgments of acquittal (at appropriate times) and the district court thrice denied them. Following eleven days of trial where it heard from 25 witnesses, the jury returned a guilty verdict against both defendants on all counts on the twelfth day of the saga.

In due course, the district court sentenced each defendant separately. We'll start with Griffin. The court calculated Griffin's United States Sentencing Guidelines ("U.S.S.G" or "Guidelines") offense level at 25, and his criminal history as Level 1. (We'll get to the math later.) That combination resulted in a Guidelines Sentencing Range of 57 to 71 months, and the court sentenced Griffin to 60 months' imprisonment and three years of supervised release. Along with that, he owed a special assessment of $2,100, restitution of $329,163.77 (with

exchange for getting off the prosecutorial hook. (We say "most" because we couldn't track down what happened with Trooper Pete Morawiak, who Trooper John Jakobowski testified was in the unit while the fraud happened.)

15

$142,774.77 going to the DOT and $186,389.00 to the IRS), and forfeiture of $239,796.75 (the bulk of it related to the BHS financial aid).

Next came Robertson. As to the Guidelines, the court calculated Robertson's offense level at 21 and his criminal history as Level 1. (Again, the math will come later.) That pairing resulted in a Guidelines sentencing range of 37 to 46 months, and the court sentenced Robertson to 36 months' imprisonment and three years of supervised release. Along with that, he owed a $600 special assessment, restitution to the DOT totaling $142,774.77 (the same figure as Griffin, because the judge found them jointly and severally liable), and forfeiture of $32,180.50.

Each timely filed a notice of appeal challenging a litany of issues related to these proceedings. And here we are.

OUR TWO CENTS

With all that history squared away, we'll turn to the appellants' arguments. In doing so, we note that because their claims necessitate that we deploy a few different standards of review as we undertake our appellate task, we'll infuse those standards along the way as necessary for our various discussions. We start then with the appellants' joint constitutional challenge to the statutes undergirding their convictions before moving into each trooper's individual protestations. Because there's a lot to cover, we implore the reader to buckle up for a long, winding ride.

16

A. Combined Constitutional Challenge First, Griffin and Robertson jointly attack the constitutionality of their conviction. They argue that neither had fair notice that billing in four- and eight-hour blocks (rather than hour-by-hour) was fraud under the wire fraud and federal program theft fraud statutes; in their view, that lack of notice makes those statutes unconstitutionally vague as applied to them. In their briefing they walk through the gamut of MSP customs and practices that, by their lights, purportedly justified their billing methodology.

But we don't need to tarry long over this argument because, as the government correctly points out, the pair lost their chance to pursue this constitutional claim by raising it too late. Here's why.

When the indictment issued back in 2020, neither defendant moved expeditiously to dismiss it based on these alleged constitutionally inadequate notice defects. Then, as usually transpires with the district court's trial management responsibilities, the parties were ordered to submit any pre-trial filings by November 20, 2023. Well, that day came and went, but again, neither defendant moved pre-trial to dismiss the charges on constitutional grounds. Instead, those challenges didn't surface until mid-trial at the close of the government's case.

17

According to the government, Federal Rule of Criminal Procedure 12(b)(3) tells us how to think about this one. Conveniently titled "Motions That Must Be Made Before Trial," the rule explains (perhaps rather obviously) that arguments falling under this rule "must be raised by pretrial motion if the basis for the motion is then reasonably available and the motion can be determined without a trial on the merits." Fed. R. Crim. P. 12(b)(3). That said, the law isn't so rigid that there are no exceptions. Elsewhere in the Rule is a fail-safe provision: an untimely Rule 12(b)(3) filing can be excused if the party can show "good cause" for the delay. Fed R. Crim. P. 12(c)(3).

So we first ask, does Griffin and Robertson's vagueness argument fall within the purview of Rule 12(b)(3)? There are some reasons to think that it does, though the First Circuit has not had occasion to squarely address the appellants' particular challenge. See United States v. Seuss, 474 F.2d 385, 387 n.2 (1st Cir. 1973) (explaining that an unconstitutional vagueness claim constitutes a "defense of failure of an indictment to charge an offense" under an older version of Rule 12). As we've past noted, two other circuits have found Rule 12(b)(3) to encompass constitutional (but not jurisdictional) attacks. See United States v. Cardona, 88 F.4th 69, 77 n.6 (1st Cir. 2023) (citing United States v. Mullet, 822 F.3d 842, 847-48 (6th Cir. 2016); United States v. Herrera, 51 F.4th 1226, 1282-85 (10th Cir. 2022)).

18

Also, we've considered an as-applied Second Amendment challenge (another constitutional-but-not-jurisdictional assault on a prosecution) as falling within Rule 12(b)(3)'s sweep, though without much resistance. United States v. Turner, 124 F.4th 69, 77 (1st Cir. 2024) ("Turner also does not claim that an as-applied Second Amendment argument for dismissal of a count of an indictment falls outside of Rule 12(b)(3)'s scope."). But we need not resolve the question definitively today given that neither Robertson (in his reply brief) nor Griffin (who chose not to file a reply brief) contested the government's position, expressly raised in its appellee brief, that Rule 12(b)(3) applies to their argument. See Cardona, 88 F.4th at 77-78 ("Cardona does not dispute that Rule 12(b)(3) applies to his vagueness claim, and so we assume that his motion is covered by that provision." (collecting cases)); Turner, 124 F.4th at 77.

So now we turn to the "good cause" analysis and dispose of it with some dispatch: Robertson and Griffin have offered us no cause, let alone a good one, for not raising their constitutional claim earlier.

Because we assume that Rule 12(b)(3) covers the appellants' claim but they failed to show good cause for not raising the argument in a pre-trial motion, "the consequence is that the matter is waived on appeal." Turner, 124 F.4th at 77. And the stakes are unusually high here, because a lack of good

19

cause also means that there's no plain error review, either. See, e.g., Cardona, 88 F.4th at 75 ("Where a defendant does not show good cause to consider an unpreserved Rule 12(b)(3) argument on appeal, he is not entitled to plain error review." (cleaned up)); United States v. Reyes, 24 F.4th 1, 16 n.8 (1st Cir. 2022) (same). Therefore, consistent with our case law, we soldier on. See Turner, 124 F.4th at 78; Cardona, 88 F.4th at 78.

B. Robertson's Remonstrances Seeing no error on the constitutional front, we'll -- to quote Griffin -- "move to another location" in this combined appeal and turn to the defendants' individual remonstrations.

We start with Robertson, whose grievances are best split into four groups. We'll take them in the following order: first the sufficiency of the evidence, then the Guidelines calculation, next the sentencing disparity, and finally the restitution. The bottom line? We don't see errors in any.

i. Proving the Case (Sufficiency of the Evidence)

Now on his own, Robertson contends that there wasn't enough evidence to convict him on any count. Properly preserved sufficiency challenges (like this one) are reviewed de novo. United States v. Reyes-Ballista, 146 F.4th 100, 108 (1st Cir. 2025). So we'll first look at the evidence, "both direct and circumstantial, in the light most favorable to the prosecution," and then "decide whether that evidence, including all plausible

20

inferences drawn therefrom, would allow a rational factfinder to conclude beyond a reasonable doubt that the defendant committed the charged crime." Id. at 109 (quoting United States v. Martínez- Mercado, 132 F.4th 61, 72 (1st Cir. 2025)).

Robertson launches his attack on his conviction from two different beachheads. The first relates to the sufficiency of the evidence as to all counts whereas the second addresses the evidentiary sufficiency of Counts 3 and 4 only.

a. Robertson's Intent to Steal or Defraud On front one, Robertson opens by contending that the government failed to prove his intent to steal or defraud, an essential element of each of the six counts he was convicted for. In doing so, he incorporates the facts undergirding the waived constitutional attack -- facts about the longstanding billing practices and customs at MSP, as well as regulations, guidelines, and mores about time-keeping protocols (and sometimes a lack thereof) -- to show how he acted in good faith. And he points the finger at Griffin, his superior officer, saying that he (meaning Robertson) was just following his (meaning Griffin's) lead.

In response to Robertson's first point, the government says there was evidence aplenty to support the jury's verdict. Witness testimony and documentary evidence were introduced to show that Robertson knew that his submission of hours violated applicable rules and that he acted with fraudulent intent in

21

getting paid for hours he didn't spend working. The government also points out Robertson's presence during an interview (back in 2012) wherein Griffin admitted that he knew over-billing was illegal. And, of course, the government mentions that Robertson ordered a subordinate to shred folders containing damning information implicating the fraudulent billing practices; that evidence, urges the government, powerfully shows that he knew he was doing something wrong.

We start our analysis with a well-worn principle we often repeat -- sufficiency challenges are an "uphill battle" for defendants. See United States v. Pena-Santo, 809 F.3d 686, 696 (1st Cir. 2015); United States v. Lipscomb, 539 F.3d 32, 40 (1st Cir. 2008); United States v. O'Shea, 426 F.3d 475, 479 (1st Cir. 2005). As we explained more fully above, "reversal is warranted only if we find that no levelheaded jury could have found" Robertson guilty. Reyes-Ballista, 146 F.4th at 109 (cleaned up).

No one contests that intent to steal or defraud, in one form or another, is, as Robertson contends, relevant to each of the counts of his conviction: theft of federal funds, conspiracy of the same, and wire fraud. So we ask ourselves: what, exactly, did the government need to produce to prove intent? In the district court's formulation (which we'll use, because Robertson uses it too and he isn't challenging the jury instructions), the conspiracy charge required the government to put forth evidence to

22

prove that Robertson "willfully joined in" the "agreement specified in the indictment, namely to steal, embezzle, [or] obtain by fraud" the money that he wasn't entitled to. The theft count similarly required the government to advance evidence showing that Robertson "embezzled, stole, obtained by fraud, knowingly converted, or intentionally misapplied property" (here, money) under the "care, custody, and control" of the MSP. And the wire fraud counts likewise required the government to factually prove that Robertson "knowingly and willfully participated in the scheme with intent to defraud."

In surveying the record in the light most favorable to the prosecution, we see plenty of evidence of Robertson's intent to do all that. Most obvious is the trial testimony of several TPS troopers who worked alongside Robertson and under his command; they said, over and over again, that they knew what they did was wrong but that they did it anyway. From their admissions, the jury could reasonably infer that Robertson, given his superior perch on the food chain, and his personal participation in filling out misleading billing forms and ordering the destruction of records, had the very same intent to steal.

The jurors also heard about codewords (like "let's move to another location") that Griffin would utilize to dismiss the troopers early. The words' cloaked meaning was understood by all involved in the scheme, including Robertson, and the codewords

23

were necessary because (1) other troopers from different troops were on the same radio frequencies and (2) the TPS squad didn't want them to know that TPS was leaving. The jury could appreciate that such codewords would not have been needed if what the TPS team was doing was copacetic.

And the jury heard testimony from other troopers and MSP leaders (including some who were not involved in the events of the case) stating how it was clear that the TPS's billing practices (i.e., working three or four hours of an eight-hour shift but charging for the whole thing anyway) were inappropriate for federally funded programs. No matter what might have been permitted in unrelated MSP duties, the jury heard from Lieutenant Eric Bernstein (who now has Griffin's job as head of TPS) that it was "written into the ISA, in the grant" that "troopers were not going to get paid" for the full shift "if they went home early." To underscore the point, the jury heard from Trooper Dennis Kelley (reading from an email chain between other troopers) that "when federal money is involved, we must cross all Ts and dot all Is."

All that could lead to the reasonable inference on the jury's part that Robertson, a sergeant, knew that he wasn't acting above-board in his billing practices (our shorthand for intent across these three separate crimes of conviction).

And if that wasn't enough, the jury heard about Robertson's orders to Kelley to axe the paperwork. For us, that

24

incriminatory directive practically seals the deal for our de novo review. If Robertson didn't know he was stealing or defrauding, specifically after investigations into related MSP branches, why else would he order the files (files that contained the information that could lead an investigator to uncover the fraud) shredded? And that this destruction order happened after much of the charged conduct had occurred is, in our view and despite Robertson's argument to the contrary, of no matter, because it's well- established that "later events often may shed light on earlier motivations." United States v. Rodriguez, 858 F.2d 809, 816 (1st Cir. 1988).

Robertson whistles past the graveyard with his alternative arguments. For one, that Robertson was "just following orders" from Griffin -- what we'll refer to as his "Nuremberg defense" -- doesn't negate the illegality of the acts on these facts; the testimony of his subordinates showed that they all (including Robertson) knew that it was wrong and that they were stealing. On this record, it would not, as Robertson suggests, be a reasonable inference to conclude that he would be more ignorant of the fraudulent scheme than the troopers he helped supervise to carry it out.

For another, we don't buy Robertson's defense, presented to but rejected by the jury, that because of longstanding MSP practices and protocols, his conduct was okay on paper. It's not

25

our place to usurp the jury's role in crediting the live testimony of troopers testifying about intent saying that they knew what they did was wrong, over the pile of documents that Robertson threw at the jury to combat that point -- including the collective bargaining agreement, division commander orders, ISAs, and others. See United States v. Vázquez Rijos, 119 F.4th 94, 102 (1st Cir. 2024) (explaining that "we cannot reweigh witness credibility on a sufficiency challenge"). And again, Robertson's order to Kelley to shred the incriminating documents shows further why his I-was- just-following-orders-and-doing-everything-by-the-books defense fails. If he thought he was following regular MSP department orders, why would he have needed to give Kelley such a blatantly questionable directive?

We address one last point before we steer clear of Robertson's sufficiency argument. Robertson contends that because he didn't break any internal rules, the convictions can't stand. There are two reasons why we don't make much of this. First, we don't think his characterization of the evidence is an accurate description of the record. Again, the jury heard testimony from leaders in the MSP discussing how, under both ISAs and the collective bargaining agreements, work on federally funded grants couldn't happen at the same time as regular shift work. That pyramiding was plainly verboten, but Robertson and his compatriots did it anyway. The jury also heard how certifying false hours ran

26

afoul of the Paystation requirement that hours be truthful. And the jury heard it was necessary to keep forms for a certain period; from that, it could readily infer that destroying them (as Robertson ordered Kelley to do) seemed to run afoul of the formal requirement that grant-related documents be retained for a period of six years.

Second, whether Robertson did or did not violate any internal rules is beside the point: that is "not the basis for his criminal liability." United States v. Ransom, 642 F.3d 1285, 1291 (10th Cir. 2011). Robertson is "criminally liable because there was sufficient evidence for a rational jury to convict him on each element of the charged crimes," including his intent to defraud the government and steal something of value. Id. at 1291. Those regulations "helped establish underlying facts necessary to prove the crimes," but they weren't what he was on trial for. Id. at 1292. So we don't buy this one on the facts or the law.

All that's to say: we see no reason to vacate based on a purported dearth of evidence of Robertson's intent. Sufficient evidence supports the jury's verdict.

b. Robertson's Liability for Griffin's Fraud Recall that we mentioned how Robertson attacked his convictions from two separate beachheads. Now we'll relocate to beachhead two, a position from which he attempts to launch a more targeted advance. Robertson offers an alternative attack on why

27

he should not have been convicted for wire fraud under Counts 3 and 4. He asserts that because those counts only relate to money deposited into Griffin's bank account, he has no criminal culpability for this conduct. And while (at least in Robertson's view) the government relied on an "aiding and abetting" theory to hold Robertson accountable for Griffin's particular wire fraud, Robertson protests that he didn't aid or abet Griffin's fraud because he didn't do anything to facilitate fraudulent payments to Griffin.9 The government retorts that Robertson misapprehends what's needed for a wire fraud conviction. Notwithstanding his aiding-and-abetting disclaimers, Robertson's individual contributions to the broader "Griffin-Robertson TPS Overtime Fraud Scheme" were sufficient to convict him even under counts concerning Griffin's fraud. And as for Robertson's aiding-and-abetting avoidance argument, it flops (in the government's view) because the overtime fraud scheme was a bona fide group project -- its success for one member depended on the support of all participating members.

9 Robertson also argues that, because the government leaned on an "aiding and abetting" theory of conviction but its evidence didn't line up with the elements of an aiding and abetting conviction, the claim fails. More specifically, Robertson hangs his hat on the "affirmative act to help or cause the charged crime" prong, arguing that he took no affirmative acts to aid Griffin's fraud. But for reasons we'll discuss, we decline to get into that one today.

28

We think the government's right on this one, too. To prove wire fraud related to Griffin's procurement of payments on false pretenses, the government needed to show Robertson's "knowing and willful participation in a scheme to defraud and the use of interstate wires to further that scheme." United States v. Tum, 707 F.3d 68, 72 (1st Cir. 2013). But the government doesn't need to show that Robertson "personally used the wires" that got Griffin the money -- just "that such use was a reasonably foreseeable part of the scheme in which he participated." Id. (cleaned up); see also United States v. Fermin Castillo, 829 F.2d 1194, 1198–99 (1st Cir. 1987) (explaining that "[w]hether or not the appellant had foreknowledge of the precise series of [electronic communications] . . . is beside the point," and stressing that "[a]s long as some use of [wires] in the course of the endeavor was reasonably to be anticipated, the causation requirement is met").

Based on the record, the jury could reasonably infer that payments to Griffin were a "reasonably foreseeable part of the scheme" that Robertson participated in. Tum, 707 F.3d at 72 (cleaned up). Testimony of troopers showed that Robertson and Griffin (along with their handful of uncharged compatriots) jointly cut corners on their overtime duties and on their timesheets. Part of the scheme required each trooper's individual submission of false statements through Paystation. And because

29

Robertson submitted false hours on Paystation at Griffin's instruction, the jury could find that Robertson knew how Griffin, in order to be paid for his fraudulent hours, also procured the money relevant to wire fraud Counts 3 and 4 -- via Paystation submissions containing false hours. So Robertson and Griffin used the same means to submit the same false hours from the same shifts as part of the same scheme; Griffin's submissions are thus a "reasonably foreseeable" part of the scheme "in which [Robertson] participated." Id. at 72 (cleaned up). That's sufficient to sustain Robertson's convictions under Counts 3 and 4, especially given that Robertson didn't need to handle the wires that moved the money into Griffin's account.

Again, Robertson's arguments to the contrary fail to persuade. For one, that Robertson didn't enter or approve Griffin's false overtime in Paystation is irrelevant, because, as we've explained, Griffin's fraudulent use of the wires was still a reasonably foreseeable part of their joint scheme, and that's all that's needed. Id. at 72.

For another, Robertson's handwringing about the expansion of liability is misplaced. He frets that "every MSP employee who submitted inaccurate Paystation entries would be jointly and severally liable for every payment made to every other MSP employee engaging in the same behavior." But Robertson's liability for Counts 3 and 4 arises not just from Griffin's use of

30

the same means of Paystation but also from their joint participation in the specific TPS scheme. See id. at 72 (noting the necessary connection to "the scheme in which he participated" (cleaned up)). As to the broader liability concern which he contends "would unfairly distend the concept of 'foreseeability' beyond recognition": hypothetically, the handful of TPS troopers from 2015 to 2017 might have all been on the hook for each other's wire fraud, because they were all in on the same scheme, but that wouldn't extend out to every MSP trooper who may have submitted false information more generally in that timeframe just because all of them used Paystation. Only the TPS troopers who worked in concert to further the particular scheme of relevance here would be foreseeably liable. After all, the jury was instructed that one way to find Robertson guilty was by finding that he "personally committed or participated in the particular offense charged," and, for the reasons explained above, the record supports the jury's conclusion that Robertson participated in that offense. See United States v. Pena, 24 F.4th 46, 74 (1st Cir. 2022) ("When a jury returns a general guilty verdict on a sole count . . . insufficiency of the evidence as to the other theory of guilt will not undermine the conviction.").

So, in sum, because (1) Robertson participated in the scheme that led to Griffin's wire fraud convictions and (2) the payments to Griffin were a "reasonably foreseeable" part of their

31

joint scheme, we see no issues with Robertson's conviction on these counts premised upon the arguments he advances before us. And because we find this theory of conviction supportable, we don't need to reach the alternative aiding-and-abetting argument that Robertson proffers. So his convictions will stand and we motor along to the sentencing challenges.

ii. Calculating the Guidelines Next, Robertson asserts that the district court made three mistakes in calculating his Guidelines range.10 First, the district court failed to apply the "zero-point offender" reduction to his total offense level. Second, it erred in applying the "leadership role" enhancement. Third, it inappropriately applied the "abuse of trust" enhancement.

We'll take these in turn, reviewing "the sentencing court's interpretation and application of the Sentencing Guidelines de novo, its factfinding for clear error, and its exercise of judgment for abuse of discretion." United States v.

10 Here's the quick math on the Guidelines calculation.

Robertson's base offense level was 7. The district court added 8 levels because the "loss" (a complex term that we'll have more on later) from his fraud was between $95,000 and $150,000. It added another 2 levels because Robertson abused a position of public trust. Then, it added 2 more levels because he was a leader of the scheme. Finally, it added another 2 levels for willful obstruction of justice. All told, the district court calculated Robertson's total offense level as 21 and his criminal history level as I. That pairing led to a recommended sentencing range of 37 to 46 months.

32

Guía-Sendeme, 134 F.4th 611, 616 (1st Cir. 2025). But to headline our conclusion, we again see no reversible error.

a. Zero-Point Offender Reduction Robertson opens his assault on his sentence by arguing that the district court erred in denying him what's called the "zero-point offender reduction." In essence, this is a two-level Guidelines reduction available if a defendant meets a host of circumstances related largely to their personal history and the gravity of their crime. See generally U.S.S.G. § 4C1.1(a). So the result of the court denying Robertson that reduction was that Robertson's Guidelines offense level was two points higher than if it had applied.

More than usual, the procedural history undergirding this enhancement is important, so we'll start there. Griffin and Robertson's draft presentence reports (or "PSRs") each included the zero-point offender reduction. When probation issued Robertson's final PSR, two weeks before Griffin's sentencing, it reversed course and found Robertson ineligible for the reduction. Yet Griffin was first to be sentenced, and, at that hearing, the government brought to the court's attention an intervening U.S. Supreme Court decision, Pulsifer v. United States, 601 U.S. 124 (2024), which it said changed the reduction's applicability. While the government recognized that it had not protested Griffin receiving the reduction and believed it had thus waived its chance

33

to object to it, it noted that it did plan to press the issue at Robertson's later-in-time sentencing. And that's exactly what it did. Then, when Robertson faced sentencing, the district court -- apparently in implicit agreement with the government's Pulsifer take -- declined to afford him the zero-point reduction. Such is the alleged error that Robertson contends requires reversal.

From the court's declination, Robertson identifies three flubs. First, the district court's change of heart violated the "law of the case" doctrine. Second, because Robertson and Griffin were identically situated, an unwarranted disparity arose from application of the reduction to one but not the other. Third, the district court incorrectly interpreted the text of the Guidelines that were in effect at the time of the sentencing, leading it to the misplaced denial of the reduction. The government contests all three of those propositions.

We'll jump straight to Robertson's third argument, because it's the clearest to us, and then work backwards. We agree with the government that the district court, based on its holding that Robertson earned a leadership enhancement, correctly declined to apply the zero-point offender reduction.

Let us say more. At the time of Robertson's sentencing, § 4C1.1(a) provided for a two-level reduction if a long list of

34

criteria were met.11 Among them was the condition that "the defendant did not receive an adjustment under § 3B1.1 (Aggravating Role)" (what we call colloquially the leadership enhancement) "and was not engaged in a continuing criminal enterprise, as defined in 21 U.S.C. § 848." U.S.S.G. § 4C1.1(a)(10) (2023 version) (emphasis added).

Robertson argues that this language -- its use of "and" -- created "conjunctive" disqualification requirements. That's a fancy way of saying that the reduction applied unless Robertson both received a leadership enhancement and was involved in a continuing criminal enterprise.12 If only one requirement was met, he remained eligible, because he didn't receive an adjustment AND participate in a continuing criminal enterprise.

The government responds that the language created two alternative disqualifiers: if there was either a continuing criminal enterprise or a leadership enhancement, Robertson could not get the zero-point offender reduction.

Between the two competing interpretations of the language, we believe the government offers the better read. One

11The reader will see in a moment why we emphasize the text in effect when Robertson was sentenced. Still, everyone agrees that what we've quoted above is the version of the Guidelines that we should be interpreting.

12Robertson contends also that he should never have received a leadership enhancement, but we'll get to that argument momentarily.

35

important reason is an amendment made to this provision in November 2024, several months after Robertson's sentencing. Then, the Sentencing Commission amended the section to divide the "Aggravating Role" and "Continuing Criminal Enterprise" requirements from Subsection (a)(10) into two separately numbered grounds of disqualification. See U.S.S.G. § 4C1.1(a)(10)-(11) (2024). The Commission said that it did so to "clarify [its] intention that a defendant is ineligible for the adjustment if the defendant meets either of the disqualifying conditions in the provision." See 89 Fed. Reg. 36853, 36866 (May 3, 2024). That's a point in the government's favor.

And so far, every court that's taken a stab at the pre-amendment language (again, the language that the district court used to sentence Robertson) has read it the same way as the government, consistent with what the Commission enacted later in 2024.13 United States v. Milchin, 128 F.4th 199, 201-02 (3d Cir. 2025) (collecting cases). We see every reason to join the trend.

13 Among those cases in this consensus are: United States v.

Milchin, 128 F.4th 199, 201-02 (3d Cir. 2025); United States v. Shaw, 2024 WL 4824237, at *1 (4th Cir. Nov. 14, 2024) (per curiam); United States v. Morales, 122 F.4th 590, 594 (5th Cir. 2024); United States v. Ashrafkhan, 129 F.4th 980, 985 (6th Cir. 2025); United States v. Cervantes, 109 F.4th 944, 946-47 (7th Cir. 2024); United States v. de la Cruz, 135 F.4th 1127, 1129 (8th Cir. 2025) (per curiam); United States v. Gonzalez-Loera, 135 F.4th 856, 859- 60 (9th Cir. 2025); United States v. Alvarez, 2025 WL 1450840, at *2 (11th Cir. May 14, 2025) (per curiam). The government identified even more in the consensus in a post-argument 28(j) letter, too.

36

Like others, we find the Supreme Court's recent decision in Pulsifer v. United States, 601 U.S. 124 (2024), instructive. There, the Court considered how a different Guidelines provision applied to defendants with "(A) more than 4 criminal history points . . . (B) a prior 3-point offense . . . and (C) a prior 2-point violent offense." 601 U.S. at 129 (quoting 18 U.S.C. § 3553(f)(1) (emphasis added)). There, as here, the dispute turned largely on what work the word "and" did to connect the conditions. And as with our issue under § 4C1.1(a)(10), the parties fiercely fought over whether the provision applied only to defendants who satisfied all three criteria or to defendants who satisfied any of the three criteria. See Pulsifer, 601 U.S. at 130-31.

The Court acknowledged that both interpretations were "grammatically possible" and looked to "legal context" to determine the meaning of the provision. Id. at 140-41. And it held that the provision established an "eligibility checklist" with three requirements; to benefit from the provision, the defendant must "not have more than four criminal-history points, not have a prior three-point offense, and not have a prior two- point violent offense." Id. at 142 (emphasis added). That was because reading the section as imposing "conjunctive" requirements would render other parts of the section's calculation scheme superfluous (a big no-no in the world of statutory interpretation). See id. at 143, 153.

37

With that holding in mind, the question posed here is straightforward. Recall the relevant preface: "If the defendant meets all of the following criteria," then they receive the enhancement. U.S.S.G. § 4C1.1(a) (2023). The provision then enumerates ten items, and in the tenth rests the issue: one of the necessary criteria is that "the defendant did not receive an adjustment under § 3B1.1 (Aggravating Role) and was not engaged in a continuing criminal enterprise, as defined in 21 U.S.C. § 848." Id. § 4C1.1(a)(10) (2023) (emphasis added). To put the preface and the criteria together and paraphrase, here's how we read it: "If the defendant did not receive a leadership adjustment and did not participate in a continuing criminal enterprise," then they get the reduction (as long as they satisfy the other conditions, too). Or more simply, here, "and" means "and," and the defendant needs to satisfy both criteria to remain eligible. But if they either have a leadership enhancement or participate in a criminal enterprise, then they're disqualified.

Robertson's position, that the "and" creates two separate conditions, both of which must be satisfied to trigger disqualification under Subsection (a)(10), has it backwards. As other circuits have explained, "§ 4C1.1(a) does not set out things that the government must show or that the district court must find in order to disqualify" Robertson "from receiving the reduction." United States v. Morales, 122 F.4th 590, 595 (5th Cir. 2024)

38

(summarizing United States v. Cervantes, 109 F.4th 944, 947 (7th Cir. 2024)). Instead, the provision "sets out requirements that" Robertson "must meet to qualify for the reduction." Id. So "to receive the zero-point-offender reduction," Robertson needed to show "both that he did not receive a § 3B1.1 adjustment and that he was not engaged in a continuing criminal enterprise," with "the failure to meet either one of those requirements" dispositive on the question. Id. (emphasis in original). Or more simply, it's on Robertson, not the government, to run the table here.

With the law cutting so clearly against Robertson's position on the provision's plain meaning, Robertson's assertion that the rule of lenity should apply here has no footing. See United States v. Place, 693 F.3d 219, 229 (1st Cir. 2012) (explaining that "the rule of lenity counsels that ambiguities in criminal statutes should be resolved in a defendant's favor" (emphasis added)).

Statutory interpretation aside, Robertson offers two alternative reasons as to why the district court erred in not granting him the zero-point reduction: the law of the case doctrine and the ever-present need for federal courts to avoid unwarranted sentencing disparities. Neither moves the needle.

As for the law of the case, we see the parties as talking past each other a bit. Let us explain. The law-of-the-case doctrine establishes that "when a court decides upon a rule of

39

law, that decision should continue to govern the same issues in subsequent stages in the same case." United States v. Moran, 393 F.3d 1, 7 (1st Cir. 2004) (quoting Arizona v. California, 460 U.S. 605, 618 (1983)). There are "two branches" of the doctrine. Id. The first, known as the mandate rule, prevents "relitigation in the trial court of matters that were explicitly or implicitly decided by an earlier appellate decision in the same case." Id. That branch is what the government points to; it argues that because there's no appellate decision, the law of the case doesn't yet matter.

But that ignores the second branch, which is what Robertson's talking about. Under the second branch, the doctrine also establishes that "a legal decision made at one stage of a criminal or civil proceeding should remain the law of that case throughout the litigation, unless and until the decision is modified or overruled by a higher court." Id. Usually, this branch binds "a successor appellate panel in a second appeal in the same case" or "a successor trial judge who steps in to complete a pending case." Id. (collecting cases). But it also more generally has been described as the principle that "a court ordinarily ought to respect and follow its own rulings, made earlier in the same case." Ellis v. United States, 313 F.3d 636, 646 (1st Cir. 2002). And we can understand why Robertson tries to

40

find some refuge in that more general understanding of the doctrine.

But what Robertson critically overlooks is that the law of the case doctrine is not without exception. See United States v. Holloway, 630 F.3d 252, 258 (1st Cir. 2011). Before we get to the exception that seals the deal against Robertson's argument, we'll remind the reader of the quirky circumstances that led to the difference between Robertson's and Griffin's Guidelines calculations on this point.

Recall that the district court rejected Robertson's attempt to secure the zero-point offender reduction only days after the government (during Griffin's earlier sentencing) explained how Pulsifer (the decision by the Supreme Court that we described just a few short pages ago) changed things. To be more specific, at the second day of Griffin's sentencing hearing (again, only days before Robertson's hearing), the government stated as follows:

So with respect to Mr. Griffin, he was given the zero-point offender reduction of two points. At this point in time, the government has waived any ability to object to it. It's not timely, but in the intervening period our appellate section has looked at a Supreme Court case, Pulsifer [v. United States, 601 U.S. 124], which pertains to the safety valve section of the [G]uidelines; and in brief, the Supreme Court says, If a defendant has to meet every factor of a sentencing guideline, they have to meet every factor.

So for the zero-point offender . . . [t]here's one to ten

41

guideposts that a defendant has to reach. The tenth factor is the defendant cannot have an aggravating role. Our appellate section has pointed that out to us. I think Probation acknowledged that in the Robertson PSR and did not give the zero-point offender two-point reduction for Robertson . . . .

. . . We can't lodge an objection now to those two points [for Griffin] because it's not timely, but just previewing for your Honor, we will be arguing that the zero-point offender should not apply to Mr. Robertson.

True, the government has surmised that it made a mistake when it did not timely object to Griffin's zero-point reduction based on Pulsifer. But, as it here noted, Robertson's PSR was timely "amended" (probation officer's words) to remove the two-level decrease because he had a leadership enhancement. While Probation didn't mention Pulsifer in reaching that decision (as far as we can tell), the timing of the change is consistent with Pulsifer's publication: it was decided on March 15, 2024, so it was not available at the time Probation prepared Griffin's PSR (March 6, 2024), but it was available at the time Robertson's was issued (April 16, 2024). And the government connected the dots explicitly at Griffin's sentencing, raising the Pulsifer issue and saying it would press that point at Robertson's sentencing. If that all wasn't enough, Robertson's sentencing memo also engaged with Pulsifer's relevance on this point (albeit only in a footnote).

So by the time the district court sentenced Robertson, it had read an amended PSR using Pulsifer-style reasoning to change

42

the enhancement's applicability, it had heard the government's argument four days earlier (albeit for a different defendant, but in the same case) saying that Pulsifer was highly relevant for Robertson's case, and it had read some more from Robertson about why Pulsifer didn't actually change the law.

That leads back to the exception we hinted at above: the law of the case doctrine doesn't apply in these circumstances, where a change in controlling legal authority is brought to the court's attention. See Holloway, 630 F.3d at 258. And that's true even when the decision isn't dead-to-rights on-point (insofar as Pulsifer concerned a very similar issue but in a different part of the Guidelines). See id. (explaining how an intervening Supreme Court opinion need not be directly on point to render the law of the case doctrine inapplicable); see also United States v. Bad Marriage, 439 F.3d 534, 537-38 (9th Cir. 2006) (explaining how an intervening Supreme Court opinion displaces the law of the case, even under the mandate prong).14 So we don't think the law of the case doctrine has any applicability here.

14That the district court didn't shout-out Pulsifer by name during Robertson's hearing is to us no matter. For one, the district court ruled in a way consistent with the opinion's rationale as it applies here (saying that Robertson did not qualify under "certainly at least one prong" of § 4C1.1(a)(10)). For another, the district court noted all the documents that it read prior to sentencing, including both the PSR and Robertson's memo. See, e.g., United States v. Romero, 906 F.3d 196, 210 (1st Cir. 2018) (imputing knowledge to district court based on the fact that "the judge started the sentencing hearing by saying that he read

43

As for the unwarranted disparities argument Robertson offers, it cannot win the day. The heralded principle of "two- wrongs-don't-make-a-right" has some weight here. Our sibling circuits have recognized as much in a variety of criminal contexts. See, e.g., United States v. McNair, 605 F.3d 1152, 1232 (11th Cir. 2010) ("That a district court may have used the wrong version of the guidelines in a co-defendant's separate sentencing (to the benefit of a defendant) does not make another defendant's sentence under the correct version unreasonable in any way."); Kinnard v. United States, 313 F.3d 933, 936 (6th Cir. 2002) (explaining that "the fact that one defendant received a benefit to which he was not entitled" usually "does not entitle another defendant to the benefit of the same mistake" in the ineffective-assistance-of- counsel context).

And let's remember that sentencing, as caselaw stresses, is an individualized affair, and the sentencing court has an obligation to correctly calculate the Guidelines for each particular defendant. See Gall v. United States, 552 U.S. 38, 49 (2007). If the Guidelines change (or become clearer) over the life of a case, it's not an unwarranted disparity if the judge correctly applies the Guidelines at the time of sentencing, even

the PSR and the parties' sentencing memos"). For a third, everyone else had by this point mentioned Pulsifer -- for example, the government made the point quite clearly at Griffin's sentencing, which the district court presided over only days earlier.

44

if it means a co-defendant was previously sentenced under a different regime or interpretation. See 18 U.S.C. § 3553(a)(4)(A)(ii) (explaining that the court must consider "the kinds of sentence and the sentencing range established" for "the applicable category of offense committed by the applicable category of defendant as set forth in the guidelines" that "are in effect on the date the defendant is sentenced" (emphasis added)).

So, although we might appreciate Robertson's angst, we see no abuse of discretion in the district court's decision not to apply the zero-point reduction to Robertson.

b. Leadership Enhancement Next, Robertson contests the district court's application of a two-point leadership enhancement. The relevant provision here, § 3B1.1(c), allows a two-level enhancement if "the defendant was an organizer, leader, manager, or supervisor" in a criminal activity. To apply, the government must demonstrate by a preponderance of the evidence that the criminal enterprise involved at least two complicit participants, and that the defendant, in committing the offense, exercised control over, organized, or was otherwise responsible for superintending the activities of at least one of those other persons. United States v. Cruz-Ramos, 987 F.3d 27, 44 (1st Cir. 2021). In determining the enhancement's applicability, the judge can look "to a variety of factors, including the nature and degree of the defendant's

45

participation, planning, and control -- and whether he exercised decisionmaking authority, drafted collaborators, or claimed a bigger piece of the spoils." Id. In our examination of Robertson's claim, we'll note that the "determination of an individual's role in committing an offense is necessarily fact- specific," so "appellate review must be conducted with considerable deference." United States v. Soto-Peguero, 978 F.3d 13, 23 (1st Cir. 2020) (cleaned up).

Robertson argues that the district court didn't have enough evidence to characterize him as an "organizer, leader, manager, or supervisor." He accuses the district court of improperly putting too much weight on his sergeant title, his one- time after-the-fact document destruction incident, and what he calls the judge's "vague perceptions" about his role in the scheme itself. In his view, the record only supports a finding that Griffin could be deemed a leader. Robertson also raises an additional challenge that he frames like this: "whether a single act of direction to another in connection with post-offense obstruction satisfies § 3B1.3."15

15We pause to note that in our scrutiny of Robertson's briefing, his arguments get quite confusing. He clearly seems to be arguing a couple more things yet ultimately and puzzlingly disclaims them. First, he says that the document destruction directive cannot legally support a leadership enhancement "because it was already the subject of a separate obstruction of justice enhancement" (a finding that he does not contest). Essentially, his briefing seems to advance a theory that amounts to an assertion

46

The government responds with a few points. First, it emphasizes that Robertson had a position of authority in TPS, and he exercised that authority over those ranking below him in the commission of the charged crime, including ordering Kelley to shred documents. Second, it says that Robertson's continued participation in the scheme, where he exercised authority over the troopers, is further evidence of his leadership. Third, the government submits that the judge's so-called "vague perceptions"

that, because an obstruction enhancement looks to conduct extrinsic to the underlying offense and a leadership enhancement requires conduct that is part of the underlying crime, the shredding directive cannot, as a matter of law, serve as evidence supporting both enhancements. The government understood him to be making this specific argument and responded accordingly in its briefing. However, in his reply brief, Robertson says the government misconstrued his position and that he specifically disclaims making such an argument. Rather, what we've quoted above from his reply was meant to clarify what he set forth in his opening brief and not raise a new argument. So we'll say no more about that.

We also note that Robertson's briefing seems to raise a distinct but related leadership enhancement argument: because the underlying offense was completed before the destruction incident occurred, Robertson's directive to Kelley could not have been a leadership act "in committing the offense" because it was later in time and served only to obscure the already-completed offense. But aside from his evidentiary insufficiency claim, Robertson (like we said above) insists he is only advancing the argument that we've quoted above.

Finally, we note that the quoted portion of the brief indeed says "§ 3B1.3," which leads the reader to the provision about the abuse of trust enhancement, but we assume (based on context clues, like the placement of that quote within the structure of Robertson's brief) that Robertson meant § 3B1.1, the leadership enhancement provision.

47

were actually well-supported facts in the record. Fourth, it offers that Robertson's legal argument -- that the destruction of documents directive is beyond the scope of the leadership enhancement -- is waived and otherwise bunk.

At bottom and as we explain, we think the district court's decision to apply the leadership enhancement was supportable. We first briefly address Robertson's second legal claim. According to the government as it understood Robertson's leadership enhancement argument, he failed to raise it below so plain error should guide our review. Robertson disagrees and contends his preserved legal argument should be reviewed de novo. We think the government probably has the better argument on this front when we zero in on Robertson's two-sentence presentation of the issue to the district court.16 See United States v. Zannino, 895 F.2d 1, 17 (1st Cir. 1990).

But we need not decide this preservation spat because Robertson's broader argument as to the enhancement's applicability

16 In totality, Robertson said this:

Probation argues that the enhancement is appropriate because "defendant directed a co-conspirator to shred incriminating TPS records." However, this single alleged act of after-the-fact obstruction, which is already accounted for in a separate Guideline enhancement, does not remotely make Mr.

Robertson an "organizer, leader, or manager"

of the core underlying criminal activity, which is the ongoing overtime abuse.

48

is built on a fatally faulty factual premise and fails for that reason alone. More specifically, the district court never indicated (despite Robertson's contentions) that his directive to Kelley to destroy records, standing alone, was sufficient to support a leadership role enhancement. Rather, the district court, which presided over a twelve-day trial and observed Griffin, Robertson, and their former subordinates, looked to a mix of evidence pointing to Robertson's leadership. See Soto-Peguero, 978 F.3d at 23 (accounting for the fact that the district court oversaw a trial in affirming its decisions on enhancements). And it concluded that Robertson "had a very significant hand" in the scheme which warranted application of the enhancement. With the district court's ruling properly understood, we agree that such evidence was sufficient.

For one, Robertson was a sergeant, a person who held a position of authority as second-in-command of his TPS unit. And while that title by itself may not have demonstrated his leadership role in the criminal activity, the evidence indicated he did in fact exercise his authority in furtherance of the core, underlying criminal scheme. To be a manager or supervisor for § 3B1.1 purposes, we require evidence of "some degree of control or organizational authority over others," and we see that here. United States v. Voccola, 99 F.3d 37, 44 (1st Cir. 1996) (cleaned up). For instance, as the government pointed out, Robertson

49

personally participated in the overtime scheme from 2015 through 2017 and he never sought to extract himself from the normal chain of command when engaging in the illegal conduct. Instead, the district court explained how Robertson -- along with Griffin -- helped put the team together and made sure certain trustworthy-enough persons were on the team to keep the fraud running smoothly. Along with that, the district court heard trial testimony from one trooper that Robertson taught him how to fill out the forms and from another trooper that he submitted false timesheets to Robertson, presumably for Robertson's approval. "Managerial status attaches if there is evidence that a defendant, in committing the crime, exercised control over, or was otherwise responsible for overseeing the activities of, at least one other person." Id. (cleaned up). We think all this evidence clears that bar.

Yes, Robertson's directive to Kelley to destroy documents was part of the district court's reasoning, as it specifically demonstrated to the court Robertson's control over one scheme participant. Yet we doubt Robertson's argument that because this single act was extrinsic to the underlying offense it therefore could not be viewed as leadership in service of the core crime itself. After all, we've past held that in determining a defendant's role in the offense for purposes of a leadership enhancement, "the sentencing court need not look only to the

50

elements underlying the conviction, but may consider the whole of the defendant's relevant conduct." United States v. Fosher, 124 F.3d 52, 56 (1st Cir. 1997) (cleaned up). But even if Robertson's argument about the directive's relevance were right, other evidence was sufficient to satisfy the government's preponderate burden. So we carry on.

c. Abuse of Trust Enhancement Finally, Robertson urges us to reverse the district court's decision to apply the two-level "abuse of trust" enhancement under U.S.S.G. § 3B1.3. That provision requires a two-level increase if the "defendant abused a position of public or private trust" in "a manner that significantly facilitated the commission or concealment of the offense." U.S.S.G. § 3B1.3.

In our circuit, applying this enhancement is a two-step pavane. First, the sentencing judge "must see if the defendant held a position of trust," either public or private. United States v. Zehrung, 714 F.3d 628, 630 (1st Cir. 2013). Second, "if the answer is yes, then (and only then) they must see if the defendant used that position in some significant way to facilitate or conceal the offense." Id.

Robertson makes a few points cutting against the enhancement's relevance here. For one, he says that the "position of trust" analysis is from the point of view of the victims, here DOT and MSP. And because DOT and MSP didn't view Robertson as a

51

fiduciary, the enhancement is irrelevant. After all (at least in Robertson's view), there wasn't any evidence that he had special authority or responsibility over the funds beyond his own time submissions. And the government's position, Robertson thinks, is so all-encompassing that if Robertson abused a trust, so did every other fraud defendant ever.

The government responds that First Circuit caselaw clearly shows that police officers are in a position of trust, establishing the first part of the enhancement. As for the second prong, the government argues that Robertson's repeated theft of MSP funds depended on his status as a police officer; he couldn't do it if he didn't work there. And that Robertson was both a sergeant and the Executive Officer of TPS makes the case all the clearer.

We can all agree that Step One, the "position of trust"

analysis, is easily satisfied. We've held before that police officers do qualify, and there's no need to reinvent the wheel (or reconsider these holdings). See United States v. Flecha- Maldonado, 373 F.3d 170, 180 (1st Cir. 2004) ("This court has at least twice said that police officers do occupy a position of public trust within the meaning of the Guidelines because they exercise significant discretion." (collecting cases)).

And we discern no abuse of discretion on the district court's factfinding as to Step Two: that Robertson "used that

52

position in some significant way to facilitate or conceal the offense." Zehrung, 714 F.3d at 630. The district court pointed in part to emails of Robertson and Griffin "making sure that certain people were on" the team "so there wasn't anybody on the outside of the team" involved in the scheme. The point there, as we read it, was that Robertson (in tandem with Griffin) exercised his discretion as a leader to keep the scheme chugging along. Together, he and Griffin wanted to load the squads only with those they could see were trustworthy enough to keep the scheme going. And along with some more obvious facts -- that he abused his position as a sergeant to "conceal" the offense by ordering Kelley (his subordinate) to destroy the documents and that he had access to these public funds in the first place only because he was a police officer -- we can see how he used his position to "facilitate" the offense. See id.

Robertson's arguments to the contrary don't persuade us.

For one, we don't think he's characterizing the government's position right: it wasn't just Robertson's position as a police officer that justified the enhancement. We agree -- that only satisfies Step One. But, as we've just explained, Robertson did more.

For another, Robertson's reliance on United States v.

George is, in our view, misplaced. 841 F.3d 55 (1st Cir. 2016). George explicitly acknowledged that it dealt with a unique

53

circumstance: "whether -- or under what circumstances -- a high- ranking employee of a government contractor can be said to occupy a position of trust vis-à-vis a defrauded government entity." Id. at 67. In other words, it reckoned with deciding when a private actor can be held liable for abusing the public trust. Id. (discussing how "a government agency sometimes may rely too heavily on a high-ranking employee of a contractor and thereby place that individual in a position of special trust").

But George's statements about those unique circumstances don't help us here. By contrast, we see this offense -- a public official abusing the public trust -- as closer to a public-sector twist on the Sentencing Commission's Commentary Notes' classic example of "a bank executive's fraudulent loan scheme." See U.S.S.G. § 3B1.3 app. n.1.17 And, to Robertson's point that he's more like the bank teller or hotel clerk whom the Sentencing Commission specifically disclaimed from this enhancement, we can't agree. Robertson's second-in-command position in TPS -- the MSP division responsible for the application and distribution, and in some cases the direct use, of federal highway funds -- puts him much closer to the bank executive than the teller. With that in mind, it's easy to see why MSP and DOT qualified as victims of his

17 The difference, here, is that the bank executive held a private trust while Robertson holds a public trust. But the enhancement applies to either circumstance, and that's why we draw the analogy across the private-public line.

54

fraud -- in the same way the Commentary's hypothetical bank would be a victim of the twisted bank executive.

Finally, as we've said before, Robertson's argument that he wasn't a legal fiduciary of TPS or DOT is no matter. See, e.g., United States v. Sicher, 576 F.3d 64, 70 n.5 (1st Cir. 2009) (rejecting the argument that "there must be a fiduciary or fiduciary-like relationship between the defendant and victim of the defendant's fraud" because that "is not the law of our circuit").

All three challenges considered and rejected, we see no reversible error in the district court's calculation of Robertson's recommended Guidelines range.

iii. Issuing the Sentence Guidelines calculations settled, we next turn to the sentence that Robertson received. Here, he's got one big argument: that the district court, in giving him a 36-month sentence, violated its duty to consider the "need to avoid unwarranted sentence disparities among defendants with similar records who have been found guilty of similar conduct." 18 U.S.C. § 3553(a)(6). He casts this error as both procedural and substantive.

Like many other sentencing issues, we look at his disparity arguments through an "abuse of discretion" lens, where we review "findings of fact for clear error and issues of law de

55

novo" while also giving "due deference to the district court's analysis of the section 3553(a) factors." United States v. Freeman, 147 F.4th 1, 37 (1st Cir. 2025) (internal quotations omitted) (first quote); United States v. Rosario, 143 F.4th 41, 45 (1st Cir. 2025) (second quote).

In his sentencing memo below, Robertson contended that a substantial prison sentence would create an unwarranted disparity from other similarly situated defendants. To illustrate, he included a chart of eleven Massachusetts state trooper overtime fraud cases -- all but two resulting only in probation and the two prison sentences only being two and three months. Eight were federal cases, three were state. Importantly, the chart included just the defendants' names, docket numbers, and the sentence received. Robertson noted that three defendants who got probationary terms were lieutenants, outranking him. Finally, Robertson observed that a true assessment of disparity should consider that the government did not charge many troopers, despite them engaging in the same overtime abuse as Robertson and Griffin.

The district court didn't buy what Robertson was selling, so it gave him a 36-month prison sentence.

On appeal, Robertson renews his disparity arguments.

And now he contends that the district court "not only refused to address unwarranted disparities but appeared affirmatively hostile to the suggestion that it should," a form of procedural error. To

56

prove the point, Robertson pulls these three quotes from the judge during the sentencing hearing:

• "I take issue, counsel, with a comparison to other judges and other cases. I've been in the criminal justice system as a lawyer and a judge for more than 30 years."

• "I -- I take issue [with the argument] that somehow [Robertson]'s being treated differently. I didn't have those other cases. I don't know what their evidence was. And I do know this. He took an oath, and he violated the oath . . . ."

• "[W]hether we treated other troopers lightly, the criminal justice system did, well, we didn't do us any favors. No favors. If it is wrong, it is wrong. There is no moral equivalence. It's wrong. It was wrong for those [troopers]. They weren't in my courtroom. I didn't see their trial. I don't know their evidence."

In Robertson's view, the district court's remarks showed how it refused to consider the need to avoid unwarranted disparities, rendering the sentence procedurally invalid.

And Robertson next argues that the sentence was also substantively unreasonable, an error flowing naturally from the procedural gaffe just proffered. He observes (again) that in eleven other Massachusetts police overtime cases, not even one resulted in more than three months of prison time; nine out of

57

eleven resulted in none. So the district court, in giving Robertson 36 months, imposed a sentence "twelve times harsher than the stiffest sentence in that group," making it substantively unreasonable -- especially given that many more fraudster troopers were not even charged. He bolsters the point now by drawing from a broader pool of cases than he did below, too. And, relying on data from the Sentencing Commission, Robertson also observes that the district court imposed a sentence fifty-percent higher than the national average for offenders with the same base-level offense and Guidelines sentencing range.

He makes one last point: his sentence was also substantively unreasonable (he says) because of the disparate treatment between him and his co-defendant Griffin. While Griffin received the "zero-point" offender reduction, the district court denied it for Robertson (as we detailed above), while also failing to address Robertson's lesser role in the offense.

The government counters at almost every turn. Starting with procedure: in the government's view, the quotes that Robertson highlighted show why there is no procedural error. After all, the transcript reflects that the court heard Robertson's arguments but was unpersuaded by them, and that is not procedural error. And the remarks go to a more fundamental flaw in Robertson's position, too; they show that Robertson failed to give the district court enough information to try to make the requisite "apples-to-apples"

58

comparison to other cases. That failure is fatal for both the procedural and the substantive parts of his disparity claim.

As for substance, the government retorts that judges may respond differently to the same sentencing arguments, and the district court hardly abused its discretion in doing exactly that by departing from other sentences but still imposing a below-the- Guidelines punishment. Finally, as for the purported disparity arising from the district court's decision not to apply the zero- point offender reduction for Robertson, it fails because Robertson was not identically situated to Griffin, as is necessary for a co- defendant unwarranted disparity claim. (The government doesn't make much of the national statistics that Robertson provided.)

The parties' arguments explained, we now take our best shot at this one. Starting with some framework: "A credible claim of sentencing disparity requires that the proponent furnish the court with enough relevant information to permit a determination" that he and other defendants "are similarly situated." United States v. Rodríguez-Adorno, 852 F.3d 168, 177 (1st Cir. 2017). More simply, the court must have enough information to compare them "apples to apples." Id. (internal citation omitted).

And that, right there, is where Robertson's claim falters. He stumbles out of the starting blocks because the information that he provided to the district court was too sparse, as shown by United States v. González-Rivera, 111 F.4th 150, 155

59

(1st Cir. 2024). There, the panel rejected a sentencing disparity claim precisely because the defendant failed to furnish enough information for the district court to make the necessary "apples to apples" comparison. Id. (internal quotation omitted). Although the defendant, in pursuing a disparity argument at the district court, "provided a barebones description of the facts of the various cases and chronicled the [proposed comparator] defendants' sentences," he "did not present even a sliver of pertinent information about the defendants' criminal histories, total offense levels, or the various guideline adjustments, up and down, to which they were entitled." Id. Without that information, "it was impossible for the district court to conclude that the cases alluded to by the appellant were substantially similar to the case at hand," so "any claim that the district court abused its discretion by not treating the proffered cases as fair congeners is doomed to failure." Id. at 155–56.

So too here. The table that Robertson provided the district court -- with just defendants' names, case numbers, and sentences, and his urging that they arose in similar overtime fraud contexts -- gave the court either the same or even less to work with than the descriptions found insufficient in González-Rivera. (As the reader will see below, we view Griffin's presentation more favorably.) Robertson's argument at sentencing was similarly brief on the point. With that in mind, we view the district

60

court's statement that it "didn't have those other cases" and didn't "know what their evidence was" not as defiance of § 3553(a)(6) (as Robertson submits) but rather an expression that Robertson had not given it enough to work with. And no procedural error arises from that.

Robertson's response -- that the cases obviously "could not be more similar" simply because they all involved Massachusetts law enforcement officers convicted based on overtime abuse -- misses the mark. Among other things, we can't find where Robertson told the district court:

• whether the specific charges were similar;

• whether the Guidelines ranges were similar;

• whether the loss calculations (which affect the Guidelines) were similar;

• whether the defendants' pretrial conduct, including cooperation with the government, was similar; or • whether the nature of the schemes (beyond just "overtime fraud") was similar.18

That's a big problem. See, e.g., Rodríguez-Adorno, 852 F.3d at 177 (rejecting a disparity claim where all that was provided was

18That is, aside from one point Robertson made below: "The government may respond that the other charged cases are distinguishable because those defendants pleaded guilty and those cases did not include document destruction," unlike this one. But that doesn't help his case either.

61

"a barebones list of the various coconspirators and their sentences," but the defendant "present[ed] no information" about the proposed comparators' "specific criminal involvement," "criminal history," "career offender status," or "cooperation (if any) with the government"); United States v. Casillas-Montero, 152 F.4th 306, 328-29 (1st Cir. 2025) (rejecting a disparity argument on similar grounds).

To boot, Robertson's position would put an unnecessary onus on district courts, requiring them to do the parties' work for them. If Robertson were right, defendants could just list cases that they say are similar; the sentencing court would then have to do the yeoman's work of looking at each docket, sifting through the filings, and figuring out how similarly situated the defendants in those cases are, all to avoid the risk of unwarranted disparities.

But that's not the law. A "credible" disparity claim "requires that the proponent furnish the court with enough relevant information to permit a determination that he and his proposed comparators are similarly situated." González-Rivera, 111 F.4th at 155 (quoting Rodríguez-Adorno, 852 F.3d at 177). Yet the natural consequence of Robertson's position is that, if the sentencing court fails to do that work for the defendant (and then fails to put, on the record, why the cases aren't similar enough to warrant the same sentence), reversible error occurs. Both

62

González-Rivera and the party presentation principle caution us to think otherwise.19 All that is to say: Robertson's general asseverations that the cases are similar is not enough to sustain a disparity argument, and he, not the court, bore the burden in making that point.20 A word about Robertson's unpublished, out-of-circuit authority on procedural error. For one, his reliance on United States v. Oba is not convincing. 317 F. App'x 698 (9th Cir. 2009) (unpublished). It seems (and we say "it seems" because the decision's rather brief) that the district court there wholly "did not address Oba's argument" regarding unwarranted disparities. Id. at 700. In contrast, the district court here did address the argument, though not in the way Robertson wanted: "I didn't have those other cases. I don't know what their evidence was." Meanwhile, the sentencing court that was reversed in United States

19See, e.g., United States v. Sineneng-Smith, 590 U.S. 371, 375 (2020) (explaining that, "in both civil and criminal cases, in the first instance and on appeal . . . we rely on the parties to frame the issues for decision and assign to courts the role of neutral arbiter of matters the parties present") (internal quotation omitted); id. at 376 (explaining that courts "do not, or should not, sally forth each day looking for wrongs to right").

20 On the chance that Robertson wanted to incorporate Griffin's more extensive discussion in his briefs below, we don't see him make that point anywhere. We'll again note, however, that Robertson's sentencing on April 30, 2024, came after Griffin's on April 26, 2024. So the sentencing court also knew the nuances that Griffin identified (which we'll discuss later) by the time it sentenced Robertson.

63

v. Killen said, specifically, that "sentencing disparity is not a recognized basis for a sentence to be imposed." 729 F. App'x 703, 717 (11th Cir. 2018) (unpublished). That obviously erroneous statement of law is miles away from what happened here, where the district court said it didn't have enough information to do the requisite comparison.

For the most part, the conclusion that Robertson failed to furnish enough information would end the matter. See González- Rivera, 111 F.4th at 156 ("Against this sketchy backdrop, any claim that the district court abused its discretion by not treating the proffered cases as fair congeners is doomed to failure."). Without sufficient information, the district court need not do a deeper dive into substantive disparities, so we won't try ourselves.

But we would be remiss if we did not pause here to comment on the district court's rather odd admonishment of Robertson's counsel for trying to argue the point. We weren't there and couldn't hear the tone behind what was said, but reading the cold page of the transcript, we're troubled by the court's words. We see no need for the district court to "take issue" with comparisons to other cases; in fact, it's the district court's duty under the law to think about how one sentence compares to another (so long as the "another" is properly presented). See 18 U.S.C. § 3553(a)(6).

64

And the district court's suggestion that its experience in the criminal justice system gave it a pass around that requirement is equally confounding. The judge's experience should have made it all the clearer that comparing cases and considering disparities (again, when properly presented) is part of its job. Those comments weren't an abuse of discretion, given our holding about the insufficiency of Robertson's "apples," but they should be avoided going forward lest they leave the wrong impression.

iv. Ordering Restitution

Lastly, Robertson contests the restitution order. That order required him (along with Griffin, who's jointly and severally liable for this one) to pay back $142,774.77, the aggregate amount of improper overtime paid to all members of TPS during the scheme's heyday.

Everyone agrees that we review the district court's decision for abuse of discretion, where we examine "the court's subsidiary factual findings for clear error and its answers to abstract legal questions de novo." United States v. Ochoa, 58 F.4th 556, 560-61 (1st Cir. 2023) (internal citation omitted).

Below, Robertson asked the district court to do two things: (1) direct the government to see what, if any, money has already been recouped, so as to avoid a windfall and (2) apportion only $32,180.50 to Robertson, leaving Griffin with the rest. The district court did neither, and in Robertson's view, disregarded

65

the usual restitution process in three ways: it ignored the government's initial burden, Robertson's attempts to clarify if a windfall was in play, and its own statutory authority and discretion to apportion restitution.

Some table-setting. The restitution order here was the result of the Mandatory Victims Restitution Act ("MVRA"), 18 U.S.C. § 3663A. That statute requires the district court to order a defendant convicted of "an offense against property . . . including any offense committed by fraud or deceit" to "make restitution to the victim of the offense." 18 U.S.C. § 3663A(c)(1)(A)(ii) (first quote); id. § 3663A(a)(1) (second quote). And, relevant to the first argument, the sentencing court must order that restitution "to each victim in the full amount" of their losses and "without consideration of the economic circumstances of the defendant." Ochoa, 58 F.4th at 561 (cleaned up). That's easy enough when there's only one defendant who's on the hook for everything. See id.

But it's harder when "more than one miscreant has contributed to the victims' losses." Id. In those cases, "the MVRA gives the court a choice between two options." Id. It can "apportion liability among defendants according to culpability or capacity to pay," or it can "make each defendant liable for the full amount of restitution by imposing joint and several liability." Id. (internal citation omitted). Between those two

66

options, a sentencing court "enjoys broad discretion." Id. So, as we previously put it, "the court may weigh an individual defendant's role in the offense when deciding whether to apportion restitution -- but it is generally free to decide the issue either way." Id. And in conspiracy prosecutions, this "freedom of choice is especially appropriate." Id.

We see that caselaw as plainly foreclosing Robertson's first argument, regardless of whether it was waived (as the government suggests). The district court could -- but was not required to -- apportion liability; its decision not to do so is one that we should hesitate to pick apart. See id. ("In making this choice, a sentencing court enjoys broad discretion."). Our previous cases haven't told us when we should do so, and without more from Robertson, we won't try to enumerate the circumstances. But we can say, rather comfortably, that this case isn't the right vehicle to figure that out.

That leaves Robertson's burden and windfall arguments.

"As a general matter, restitution orders should not generate windfalls." United States v. Carrasquillo-Vilches, 33 F.4th 36, 46 (1st Cir. 2022). And the "government bears the burden of proving a victim's actual loss by preponderant evidence." United States v. Naphaeng, 906 F.3d 173, 179 (1st Cir. 2018). But that's not all. We don't "demand absolute precision in the fashioning of restitution orders," and district courts aren't "expected to

67

undertake a full-blown trial" in determining what is owed. Carrasquillo-Vilches, 33 F.4th at 46 (first quote); Naphaeng, 906 F.3d at 179 (second quote).

The district court, in Robertson's sentencing, ordered from him the same amount that it had already ordered from Griffin, and it held them jointly and severally liable. We can infer from that decision that the district court thought the government had met its burden on those figures during trial or Griffin's sentencing. The court said as much, too: it explained at Robertson's sentencing that "I have two defendants in front of me at one trial, and the trial produced evidence that established a certain amount of money," meaning the amount in question for restitution. Given the court's reasonable reliance on the trial evidence, we don't see how we can hold that it erred in deciding that the government met its burden.

As for the windfall point, we see no error. Robertson wanted the government not only to prove its point but also prove a negative -- that the victims hadn't already recovered. His basis for doing so, in our view, wasn't grounded in much. His supplemental sentencing memo only included a paragraph about it, and he didn't include the article that he now cites before us. And even if he had, the district court's decision not to force the government to prove a negative, once it was already satisfied that the preponderance standard had been met, was not, in our view,

68

erroneous. Robertson provided only speculation about a windfall; he didn't lay out a plan to figure out if there had been one (aside from asking the district court to do that for him), nor did he seek a continuance to try and obtain that information on his own. As with other parts of restitution, the district court had some discretion in determining whether a windfall happened and whether more exploration into the issue was needed. Cf. Carrasquillo- Vilches, 33 F.4th at 46; Naphaeng, 906 F.3d at 179. In not buying Robertson's unsubstantiated concerns, it didn't err.

And with that settled, we see no reversible error in the district court's approach to restitution here.21 Thus concludes our ride with Robertson.

21 We'll note that Robertson also pointed out some possible misunderstandings on the district court's part that strike us as non-reversible. After Robertson protested the restitution calculation, the district court said his arguments were "premature," because the restitution process was only beginning (in its words, the process was on "day one"). But restitution should wrap up alongside sentencing because it is part and parcel of the sentence. See 18 U.S.C. § 3663(a)(1)(A) (explaining how the court may order restitution "when sentencing a defendant"); United States v. Sasnett, 925 F.2d 392, 399 (11th Cir. 1991) (explaining how that language "suggests that restitution must be ordered at the time of sentencing, and not thereafter"). Yet the court's simultaneous remarks about being satisfied by the trial evidence assure us there was not reversible error from this misunderstanding. We also decline to act on the court's remark that it could adjust the order later, but note that it was probably mistaken in suggesting so. See generally United States v. Harvey, 20 F.4th 71, 75-79 (1st Cir. 2021); 18 U.S.C. § 3664(j)(2) (delineating exceptions).

69

C. Griffin's Grievances

We now turn to Griffin's individualized grievances.

They can be put into three buckets: those related to Guidelines calculations, those related to the sentence he received, and those related to the forfeiture order. And we'll take them in that same order.

i. Calculating the Guidelines Griffin argues first that the district court, while calculating the Guidelines, erred in two big ways.22 First, it overestimated the "loss" (or really, the gain) that the fraud caused. Second, it improperly grouped the fraud arising from the overtime scheme with the fraud arising from BHS (to remind, the "Belmont Hill School"). In Griffin's view, each drove up the Guidelines offense level erroneously. But in our view, the district court got it right on both.

22Again, here's the math on the Guidelines. The court split the offenses into two groups. Group I, the overtime abuse and school scheme counts, had a base level of 7. With loss between $250,000 and $500,000, the district court added another 12 points. It added another 2 for Griffin's sophisticated means, another 2 because he abused his position of trust, and another 4 because he was an organizer or leader. So Group I's level was 27. Group II, the tax fraud, had a base level of 16 and no enhancements. The combined offense level, derived from Group I as the higher of the two, was 27. The district court reduced it by 2 points because Griffin was a "zero-point" offender, bringing his total offense level to 25. With no criminal history, that led to a Guidelines range of 57 to 71 months. The district court gave Griffin a sentence of 60 months, within that range.

70

a. Loss/Gain

Griffin first argues that the district court improperly jacked up the "loss" figure for Guidelines purposes. In his view, the district court should have figured out how much financial aid his family would have received from BHS if he hadn't filed fraudulent papers; the court's decision not to, Griffin posits, violates our caselaw apparently requiring a nexus between the loss and the fraud.

We'll start with some basics. Section 2B1.1(b)(1) has a table that directs a sentencing court "to increase a defendant's offense level based on the amount of loss attributable to the defendant's fraud." United States v. Kumar, 112 F.4th 30, 36 (1st Cir. 2024). And on the rare instance that "there is a loss but it reasonably cannot be determined," the court can use "the gain that resulted from the offense as an alternative measure." U.S.S.G. § 2B1.1(b)(1) n.(B). The government bears the burden to prove loss (or gain as its proxy) by a preponderance of the evidence. Kumar, 112 F.4th at 36. And the sentencing court has "considerable discretion" to figure out the loss (or gain) in a couple of ways: in determining what evidence is reliable and in making a "reasonable estimate" of the calculation. Id. Nothing more than that reasonable estimate is required from the district court. Id.

Below, Probation calculated the loss in the PSR as to Counts 7-10 (related to the BHS fraud) as $176,700. That figure

71

was all the financial aid that Griffin and his family received from 2013 to 2020. Griffin objected to the PSR; he argued that because there was testimony from school officials stating that they didn't know exactly how much the fraud altered the financial aid, the government failed to prove loss by a preponderance.

In response, Probation recognized that "the exact loss cannot be determined," but "the gain can be," because it was "secured based on fraudulent documentation submitted via Wire Fraud." And gain, in Probation's view, was an acceptable stand- in for loss here. While (in Probation's words and with its emphases) "some of the [BHS] monies may have been granted nonetheless, all of it was ultimately 'gained' based on fictitious representations."

The district court heard Griffin out on this argument during sentencing and ultimately sided with the government, rejecting Griffin's objection and adopting the PSR's reasoning. We see "neither clear nor legal error" with that decision. United States v. Jiménez, 946 F.3d 8, 14 (1st Cir. 2019).

For one, "although calculating gain as a substitute for loss is only appropriate in limited scenarios," we don't see Griffin pressing a direct "challenge to the use of gain as a measure of loss on this record under U.S.S.G. § 2B1.1." Id. He writes extensively about loss, references the "loss" provisions in the Guidelines, and cites cases about loss, but in the ten pages

72

of his brief on the issue, he only uses the term "gain" four times (all in passing). In none of those instances does he tell us how to think about gain or why the district court's use of "gain" (rather than loss) was wrong. So his arguments and the district court's reasoning (insofar as it adopted the PSR) seem like two ships passing in the night; that alone could leave us "no choice but to affirm." United States v. Henry, 848 F.3d 1, 7 (1st Cir. 2017) (holding as much when the defendant failed to challenge "the district court's alternative basis for denying the motion to suppress").

But even giving Griffin the benefit of the doubt and holding his loss arguments up against the district court's "gain" theory on the merits, we still think they fall flat. To his credit, it's true that "gain" poses a similar linguistic problem to the one he identifies with "loss." The Guidelines, in Table Note (B) of § 2B1.1(b)(1), use the term "the gain that resulted from the offense," and that language tees up the same type of causation analysis that we adopted in United States v. Alphas, 785 F.3d 775, 783 (1st Cir. 2015).23 There, we held that, in calculating "intended loss"

(importantly, defined in the Guidelines as "the pecuniary harm

23 The careful reader will note that Alphas was, in part, superseded by amendment to the Guidelines. See Carrasquillo- Vilches, 33 F.4th at 42. But we don't see that as weighing on its applicability here.

73

that was intended to result from the offense"), sentencing courts should determine "whether and to what extent legitimate claims" are "embedded in the fraud." Id. That was in large part because the phrase "results from" in the Guidelines "implies 'a requirement of actual causality,' which typically 'requires proof that the harm would not have occurred in the absence of -- that is, but for -- the defendant's conduct.'" Id. (quoting Burrage v. United States, 571 U.S. 204, 211 (2014)).

Because of the similarities in language just described, we think Alphas also applies to the "gain" context, so Griffin wasn't off base in relying on it. But there's still a problem for Griffin: the burden-shifting framework that Alphas laid out.

In Alphas, we explained, "in a case such as this -- where a defendant's claims were demonstrably rife with fraud -- a sentencing court may use the face value of the claims as a starting point in computing loss," or, as relevant here, gain. 785 F.3d at 784. After that, the "burden of production will then shift to the defendant, who must offer evidence to show (if possible) what amounts represent legitimate claims." Id. Only "after the record is fully formed, the sentencing court must determine the amount of loss that the government (which retains the burden of proof) is able to establish." Id. (cleaned up). But (and here's the kicker), "depending on the defendant's offer of proof, the court

74

might well conclude that the amount of loss is equal to the face value of the submitted claims." Id. (cleaned up).

Here, the court found that it would use the value of the financial aid that Griffin received as the figure for the gain.24 It drew that figure from the PSR and the government's materials. All that was fine under Alphas Step One.

It thus "fell to" Griffin "to establish which portion of those amounts represented legitimate claims" of financial aid. United States v. Rivera-Ortiz, 14 F.4th 91, 103-04 (1st Cir. 2021) (applying Alphas framework) (cleaned up). The district court even invited Griffin to make his case: "I will take any evidence that you wish to present that would call into the question the figures of the government."

In response, Griffin generally observed that the $177,600 figure wasn't accurate because of the timing of his children's attendance at the school. But he did not break down what he saw as a reasonable alternative, and when specifically

24True, the district court didn't say that the figure would function as the "starting point" in the exact way described in Alphas. But as far as we can tell, Griffin also didn't argue much about Alphas below, so we don't see it as error that the court didn't divine exactly the same framework out of thin air. Even Alphas said that, in calculating loss, we've traditionally "eschewed rigid rules and instead taken a pragmatic, fact-specific approach." 785 F.3d at 781 (cleaned up). With that in mind, what mattered more was that Griffin knew the court trusted the $177,600 figure to start, that he was then invited to cut it down to size, and that he failed to do so.

75

asked, he said that "the answer's zero because the government just wholly failed to meet its burden." Nor did his exhibits help clarify this figure, insofar as they primarily consisted of the financial aid directors (one who wasn't there at the relevant time) shrugging their shoulders and saying they couldn't be certain of how it would change things.

The district court didn't need to buy what Griffin was peddling for a couple of reasons. First, we've held that a very similar argument to Griffin's is a loser in the Alphas context. See Rivera-Ortiz, 14 F.4th at 103-04.25 We don't see much daylight between that rejected argument (quoted in the preceding footnote) and the one Griffin now offers.

Second, Griffin's alternative, that the loss figure should jump to zero, strikes us as unconvincing. "A loss of zero is presumptively wrong" for BHS, because "it does not even remotely approximate" Griffin's "wrongdoing" in filing false information

25 Here's the losing argument from that case fleshed out a bit more:

On appeal . . . Rivera does not point to any evidence delineating what portion of the benefits were legitimate, or even how he could establish such a figure. Rather, he argues that it remained the government's burden to show the amount of fraudulent, as opposed to legitimate, claims. But that contention is inconsistent with the framework outlined in Alphas.

Rivera-Ortiz, 14 F.4th at 103-04.

76

there for years. United States v. Parsons, 141 F.3d 386, 392 (1st Cir. 1998). And a zero-figure loss would run up against the purpose of the loss calculation: to serve "as a measure of the seriousness of the offense and the defendant's relative culpability." U.S.S.G. § 2B1.1 cmt. background.

So, considering Griffin's "failure to meet his evidentiary burden, we conclude that the court reasonably concluded that the amount of [gain] was equal to the face value" of the financial aid that Griffin received. Rivera-Ortiz, 14 F.4th at 104. All told, we see no error in the district court's approach in calculating the gain.

b. Grouping

Next, Griffin contests the district court's approach to its "grouping" of separate offenses in calculating his offense level. It's a technical argument that we'll break down piece by piece.

In Griffin's view, he was convicted of three "buckets"

of crimes. He sees Counts 1-6 as "TPS" counts, because they concern "conduct involving the Massachusetts State Police Traffic Programs Section and related monies." Counts 7-10 are the second bucket, related to fraudulent financial aid applications for BHS, where his children attended school during the relevant years. And Counts 11-21 are the "tax" counts, "strictly related" to tax issues about his business and personal tax returns.

77

And here's where Griffin sees the problem: the district court (by adopting Probation's recommendation) combined the TPS counts and the BHS counts into one "group," separating only the tax charges into a second group. That combined group created a higher aggregate loss number, driving up his overall Guidelines offense level (and thus his recommended range). Griffin argues that its approach was erroneous under the relevant Guidelines provision, § 3D1.2.

Before we go any further, it's worth walking through that provision. Under § 3D1.2, "[a]ll counts involving substantially the same harm shall be grouped together into a single Group." The provision then provides four times when counts involve "substantially the same harm," including three relevant here:

(a) When counts involve the same victim and the same act or transaction.

(b) When counts involve the same victim and two or more acts or transactions connected by a common criminal objective or constituting part of a common scheme or plan . . . .

(d) When the offense level is determined largely on the basis of the total amount of harm or loss, the quantity of a substance involved, or some other measure of aggregate harm, or if the offense behavior is ongoing or continuous in nature and the offense guideline is written to cover such behavior.

78

U.S.S.G. § 3D1.2(a)-(b), (d).26 But that's not all the relevant text. Right after those four times, the provision explains that offenses "covered by the following guidelines are to be grouped under this subsection," referring to Subsection (d), and the provision proceeds to list thirteen rows of Guidelines provisions. And after that, it "specifically" excludes all offenses under another eleven rows' worth of offenses for Subsection (d).

Griffin essentially thinks that Subsections (a) and (b)

should have guided the district court's analysis. In his view, it should have looked at whether the charges involved the "same victim and the same act or transaction" or a "common scheme or plan," rather than just if they were on the same row. And under Griffin's test, it would have been obvious that the TPS counts and the BHS counts can't be grouped together: they involve different victims (the DOT for the former, the school for the latter); different acts (filing false hours at work for the former, filling out a school's financial aid form falsely for the latter); and constitute different forms of harm (TPS monies better seen as a loss, while BHS monies better seen as a gain).

The government points instead to Subsection (d). No one disputes that the TPS counts (federal programs fraud conspiracy,

26No one argues that provision (c) was relevant, so we passed by it to save some keystrokes -- hence, the ellipses after part (b).

79

federal programs fraud, and wire fraud) and the BHS counts (wire fraud) are all covered by the same provision, § 2B1.1. And § 2B1.1 is one of the Guidelines "to be grouped" under Subsection (d), so the government posits that the TPS counts and the BHS counts must "be grouped" together to comply with § 3D1.2(d).

Importantly, we aren't writing on a blank slate here.

We previously held that two counts that were "listed in the same row" under § 3D1.2(d) were "automatically" to be grouped. United States v. Zanghi, 189 F.3d 71, 85 (1st Cir. 1999). By Zanghi's logic, this case is even easier: not only are the different counts under provisions listed in the same row, but they're even under the same provision, § 2B1.1. That seems to settle things in favor of the district court's approach.

Bolstering our conclusion that the district court correctly grouped the counts is the Sentencing Commission's commentary to § 3D1.2. As one example, it provides: "The defendant is convicted of five counts of mail fraud and ten counts of wire fraud. Although the counts arise from various schemes, each involves a monetary objective. All fifteen counts are to be grouped together." U.S.S.G. § 3D1.2, cmt. 6, ex. 3. We don't see much difference between Griffin's case and this example.

It's true that other courts have viewed the requirements of § 3D1.2(d) differently. See, e.g., United States v. Lenoci, 377 F.3d 246, 254 (2d Cir. 2004) (explaining that "the view that

80

we must group counts that fall under the same 'to be grouped' guideline has not [been] met with universal agreement"). And at oral argument, Griffin asked us to disavow Zanghi and instead adopt an approach closer to that taken in Lenoci.

Here's the problem with that argument: in our multi-panel circuit system, the law of the circuit doctrine controls, meaning that new panels are "bound by prior panel decisions closely on point." United States v. Barbosa, 896 F.3d 60, 74 (1st Cir. 2018) (cleaned up). And although there are "narrowly circumscribed" exceptions to that rule whose incidence are "hen's teeth rare," Griffin does not explain, based on our precedent, why they would be applicable here.27 Id. ("Unless a litigant can demonstrate that one of these exceptions applies to a prior panel decision, a newly constituted panel must continue to adhere to the earlier holding.").

27In his opening brief, Griffin didn't address the law of the circuit issue at all. And we can quickly dispense with the arguments that he unveiled for the first time at oral argument. See Piazza v. Aponte Roque, 909 F.2d 35, 37 (1st Cir. 1990) ("Except in extraordinary circumstances not present here, a court of appeals will not consider an issue raised for the first time at oral argument."). For one, Griffin emphasized that Zanghi was pre-Booker, issued amid the Guidelines' adolescence. For another, Griffin generally averred to more recent Guidelines commentary (in its introduction) about how the power and impact of one Guidelines determination should not drive sentencing. And for a third, Griffin argued that the alternative approach should look at the factors listed in Subsections (a) and (b) and explain the overlapping victims and harms. Even if right, none of those arguments come close to meeting the high standards set out in our caselaw. See Barbosa, 896 F.3d at 74.

81

Those challenges both settled, we see no error in the calculation of Griffin's sentence.

ii. Issuing the Sentence

Griffin, like Robertson, also takes aim directly at his sentence. And like Robertson, his argument is based on unwarranted disparities, so the place to begin is below.

In his sentencing memo, Griffin argued that similarly situated defendants received lesser sentences than the one that the government sought here. He detailed the prosecution of Massachusetts state troopers in overtime cases, noting that the government's preferred sentence was twenty-five times greater than the sentences received by all (save one) of the other troopers, his ideal comparators. And unlike Robertson, Griffin did make this point lucidly in his memo and developed it further in a five- page attached exhibit. But to his credit (and, as the reader will see shortly, also his doom), Griffin identified some key differences between this case and those: the addition of a conspiracy charge, the longer period alleged in the indictment, the role enhancement Griffin received, and his guilty pleas in the tax charges.

At sentencing, the parties debated the import of these comparator cases. The government went out of its way to distinguish them. Ultimately, the district court agreed with the

82

government that a prison sentence was appropriate; it gave Griffin a within-Guidelines sentence of 60 months.

Now on appeal, Griffin similarly argues that his 60-month sentence is unwarrantedly disparate based on the "wildly lighter" sentences imposed in other cases. (Unlike Robertson, he seems to pursue only a substantive claim.) As proof, he includes a two-and-a-half-page chart detailing what he casts as similar Massachusetts cases. Griffin does so to show that his sentence was "twenty-five times more" than the sentence received by all but two similarly situated officers. And he, like Robertson, argues that the government's decisions not to charge others should factor into the equation.

The government responds with a few quick points. First, the inquiry focuses on "national," rather than "district-based," disparities, so Griffin's comparisons to Massachusetts are incomplete. (The reader might think that the government would itself then frame this in national terms, but it doesn't.) Second, Griffin failed to provide the requisite information for us to make the apples-to-apples comparison now. Third, a closer look at the cases cited reveals key differences: his extra conspiracy charge, the greater length of the scheme, the differing Guidelines calculations, and his choice to go to trial rather than cooperate. Fourth, there is no precedent to support the proposition that a defendant can sustain a disparity claim based on uncharged conduct

83

for others. Finally, along with some other facts, Griffin's status as a leader in TPS heightened his culpability as compared to Robertson and others.

Now, our take. Unlike Robertson, Griffin has given us enough to try to make the apples-to-apples comparison. But the information that he furnished reveals that his case and his recommended benchmarks are more like apples and oranges, so his claim fails.

Recall that the disparity claim flops "if material differences between the defendant and the proposed comparators suffice to explain the divergence," and material differences include "dissimilar criminal involvement, criminal histories, or cooperation with the government, to name just a few." United States v. Romero, 906 F.3d 196, 211–12 (1st Cir. 2018) (cleaned up). Griffin himself highlighted some of these differences below. Taking him at his word:

• "No other defendant was charged with conspiracy so as to increase the loss figure by some multiple."

• "The period of time for the overtime fraud charged in the [proposed comparator] Quincy police officer case was 7 months. The period of time charged in virtually all the Troop E cases [other proposed comparators] was a year or occasionally two. Here, the time period was three years -- the most of any case," and the "three-

84

year conspiracy was multiplied by at least five (5)

participants to generate a loss figure more than five times any other defendant."

• "None of the four (4) lieutenants charged to date suffered a role enhancement of any kind, let alone four (4) levels."

• "The defendant here (Griffin) also pleaded to tax charges and what is in essence, financial aid fraud."

Alone, each seems material enough to justify a different sentence. Romero, 906 F.3d at 212; Freeman, 147 F.4th at 39 (explaining that "[n]one of the cited cases include tax evasion charges" and that inclusion of a leadership enhancement is a material difference). Together, it's all the clearer. And to boot, at least some of those cases involved cooperation, but Griffin chose to go to trial (albeit not on every count, as his partial guilty plea evinces). See Romero, 906 F.3d at 212 ("But unlike Guzman, Maldonado, Wallace, and Reynoso, Romero did not cooperate with the government."); see also Freeman, 147 F.4th at 39 (same result when co-defendants pled guilty but Freeman didn't). Those are plenty of material differences.28

28Griffin also relied on some state-law cases, both here and below. While we haven't yet expressly said that state prosecutions are poor comparators, it flows naturally from the caselaw's emphasis on needing close federal Guidelines ranges and the same federal charges to do the analysis. See Freeman, 147 F.4th at 38 n.35 (comparing federal Guidelines ranges); United States v.

85

To be fair to Griffin, he highlights those differences to make almost the opposite point: that his charges "are essentially similar, the conduct is essentially similar, and the criminal records are definitely similar," and that the government's reasons for pursuing a higher sentence are bogus. But in highlighting those distinctions he seems to foreclose his own disparity claim; the very disparities that he points out "suffice to explain the divergence" in the sentencing judge's decision. United States v. Demers, 842 F.3d 8, 15 (1st Cir. 2016).

Both the others' cooperation and Griffin's additional conduct "make the sentencing disparities reasonable." Romero, 906 F.3d at 212. And while we hear Griffin's points about the potential unfairness in the charging decisions and the government's strategy in pursuing this case in this way, compared

Bishoff, 58 F.4th 18, 26 (1st Cir. 2023) (rejecting sentencing disparity claim in part because co-defendants were charged with different federal gun-related offenses). Those federal issues aren't at play in state court prosecutions. And it's well-accepted that the emphasis is generally on "national" disparities, which we read to mean federal. See, e.g., Bishoff, 58 F.4th at 25 (explaining that § 3553(a)(6) is "primarily aimed at national disparities, rather than those between codefendants"). All that seems to preclude state prosecutions from being apples-to-apples. To wit, our judicial superiors in Washington have also emphasized that the uniformity the Guidelines seek to achieve is across federal courts, not all courts. See, e.g., Hughes v. United States, 584 U.S. 675, 684 (2018) ("A principal purpose of the Sentencing Guidelines is to promote uniformity in sentencing imposed by different federal courts for similar criminal conduct." (cleaned up)). So those three state comparators probably don't get us very far.

86

to those cases in different ways, we aren't really able to adjudicate that in the disparity context, given that it creates an "apples and oranges" scenario.

At sentencing, the government made a couple additional points to show why Griffin's conduct was distinct from his proposed comparators. One set of the comparator cases involved "a bunch of renegade troopers using ten different ways to steal money," rather than the coordinated Griffin-Robertson scheme. And the other troopers "only stole overtime," unlike Griffin who was also "running side businesses committing fraud and stealing money and defrauding a private school." We didn't see much of a rebuttal from Griffin, but those differences matter, too. Demers, 842 F.3d at 15 (explaining how "a sentencing disparity claim may easily be repulsed if material differences between the defendant and the proposed comparator suffice to explain the divergence").

Only one of Griffin's arguments is left standing.

Griffin posits that, because other troopers who committed fraud were not charged but Griffin's on the hook for five years in prison, there's an unwarranted disparity. That position doesn't find any support in law, as far as we can tell, and there's good reason why.

While there's obviously a disparity between Griffin and these uncharged troopers, it's not one that we are well-positioned to assess. The disparity analysis is closely tethered to holding

87

cases up against each other; if there's no charges, there's no case to compare it to. Or more colorfully, if Griffin's an apple and the other sentenced troopers are oranges, the uncharged troopers are salmon -- part of an entirely different food group.

With all that in mind, Griffin's sentence, like Robertson's, "rests on a plausible rationale" and "represents a defensible result." United States v. Ortiz-Pérez, 30 F.4th 107, 113 (1st Cir. 2022) (cleaned up). That's true even if it is longer than most, and so we affirm it.

iii. Ordering Forfeiture

Finally, Griffin sets his sights on the forfeiture order of $177,600, which the reader will remember was the full sum of financial aid that the Griffin family received from BHS between 2013 and 2020. Here's where we agree with him.

Again, we start with some table-setting. Under 18 U.S.C.

§ 981(a)(1)(C) (the civil forfeiture statute), "[a]ny property, real or personal, which constitutes or is derived from proceeds traceable to a violation of" the wire fraud statute is subject to forfeiture to the United States. The standards for determining "proceeds" vary depending on whether the forfeiture falls under § 981(a)(2)(A) or § 981(a)(2)(B). And while the different definitions of "proceeds" have generated some caselaw, Griffin seems to assume that the financial aid is indeed a form of "proceeds" (though neither he nor the district court tell us which

88

kind). He instead draws our attention to a statutory phrase not defined in the statute and, by our lights, not often argued about: "traceable to." See United States v. George, 886 F.3d 31, 41 n.6 (1st Cir. 2018) (declining to interpret "traceable to" because the "proceeds" question was clear).

Griffin argues that the government failed to prove that any or all of the $177,600 in financial aid that the Griffin family received was "traceable to" the offense. He relies on our decision in George, 886 F.3d at 42, where we explained that, to "undergird" a forfeiture order, the district court "ultimately had to determine by a fair preponderance of the evidence" that the proceeds "were either directly or indirectly obtained as fruit of the charged" crime. And he points to a pair of district court cases rejecting the government's proposed forfeiture when it failed to prove that the proceeds were indeed traceable. Applying all that here, he argues (as he did above in the Guidelines calculation) that the BHS witnesses couldn't verify exactly how much of the aid resulted from the fraud, so the forfeiture order is bunk.

The government responds simply that the aid was the "direct result" of his fraudulent applications, so they're obviously "traceable." And it argues that at least one of the district court cases relies on Eighth Amendment reasoning that's wholly irrelevant here.

89

On this point, Griffin's got the better argument. For one, the government bears the burden of proving, by a preponderance of the evidence, the amount of proceeds that are subject to forfeiture. That's straight from the statute. See 18 U.S.C. § 983(c)(1) ("In a suit or action brought under any civil forfeiture statute for the civil forfeiture of any property -- (1) the burden of proof is on the government to establish, by a preponderance of the evidence, that the property is subject to forfeiture."). And our fellow circuits have held the same. See, e.g., United States v. Asomani, 7 F.4th 749, 754 (8th Cir. 2021); United States v. 5208 Los Franciscos Way, 385 F.3d 1187, 1193 (9th Cir. 2004).

And the government didn't quite hold up its end of the bargain here. That's because it didn't carve out what portion of financial aid Griffin would have received, as distinct from what portion he wouldn't have sans the fraud. The district court's decision to adopt the government's figure in full and not to carve that out, we think, is reversible error. And we see that as true no matter which definition of "proceeds" applies.

We aren't alone in that approach. Consider the Seventh Circuit's decision in United States v. Hodge, 558 F.3d 630 (7th Cir. 2009). There, that court vacated a forfeiture order for all the revenues of a purported massage parlor and spa that served as a front for a prostitution business. Id. at 637. The forfeiture

90

statute, in the Seventh Circuit's view, covered "only income and assets obtained from the unlawful deeds," meaning the prostitution. Id. at 635 (citing 18 U.S.C. § 981(a)(2)(A)). But that led to an issue: "somewhere between 20% and 50% of the spa's customers never signed a second credit slip." Id. And from that, one could draw dueling inferences about which proceeds were traceable to the criminal conduct: it "could mean that a substantial fraction of the spa's business was the provision of lawful massages" or that "these customers paid the prostitutes in cash." Id. Yet the district court "did not attempt to sort this out." Id. That was error, the Seventh Circuit held, because "the difference matters." Id. Only those proceeds related to the illegal conduct were forfeitable. Id.

The same principle undergirded the affirmance of a forfeiture order in United States v. Torres, 703 F.3d 194 (2d Cir. 2012). That case (which also fell under § 981(a)(2)(A)) concerned a defendant who had understated her household earnings to get rental subsidies. Id. at 197. The PSR stated, and the defendant did not dispute, "that the sum she was excused from paying . . . because of her fraud was $11,724." Id. The forfeiture order was permissible because it represented "a discrete sum of money saved by Torres, exactly corresponding to the reduction she was granted on her rent, and available for her use as a direct result of her fraud." Id. at 201.

91

By contrast, consider the government's approach here.

It says that Griffin omitted material information from his financial applications and received a total of $177,600 in aid, and that all of it is thus forfeitable. And that's true, the government says, even though Griffin offered statements from BHS officials that they could not say whether the financial aid amount would have been different had Griffin provided true information.

But "the difference matters." Hodge, 558 F.3d at 635.

The approach taken by the government and accepted by the court falls short of the particularized approach affirmed in Torres, held necessary in Hodge, and rooted in principles already accepted in our caselaw. After all, "to undergird its forfeiture order, the district court ultimately had to determine by a fair preponderance of the evidence that the [financial aid] [was] either directly or indirectly obtained as fruit" of the wire fraud. George, 886 F.3d at 42 (cleaned up). By declining to determine what came from the fraud and what would have been doled out anyway, we don't think the district court did that.

And, as we explained earlier, we see that inquiry as necessary no matter which of the two definitions of "proceeds" applies.29 True, Hodge and Torres both applied the definition in

Frankly, we aren't sure which definition was applied here, 29

as it isn't evident from the district court's forfeiture order. Griffin doesn't address it either. The government, for its part, argues that the definition in § 981(a)(2)(B) applies. But that

92

§ 981(a)(2)(A). But we don't see that analysis as changing for § 981(a)(2)(B) either, because both definitions stem from § 981(a)(1)(C), requiring that the property subject to forfeiture must be "traceable to" the offense. If that money were coming into Griffin's pockets anyway, we struggle to see how it's "property" that "constitutes or is derived from proceeds traceable" to the offense, no matter how "proceeds" is defined. 18 U.S.C. § 981(a)(1)(C).

We'll pause here for a second. We recognize that, on its face, there's some incongruity between our holdings as to gain and loss under the Sentencing Guidelines above and forfeiture here. How could the government have met its burden up there as to the

only governs "cases involving lawful goods or lawful services that are sold or provided in an illegal manner." 18 U.S.C. § 981(a)(2)(B). Yet, at least in our view, what Griffin did was neither a "lawful good" nor a "lawful service" provided in an illegal manner.

Instead, the wire fraud perpetrated here (filing false applications to a school) probably fits more neatly into § 981(a)(2)(A), which applies instead to "cases involving illegal goods, illegal services, unlawful activities, and telemarketing and health care fraud schemes" (emphasis added). And that section defines "proceeds" as "property of any kind obtained directly or indirectly, as the result of the commission of the offense giving rise to forfeiture, and any property traceable thereto." Id. § 981(a)(2)(A).

Still, we're well aware that "the classification of an offense of conviction can have a profound effect on the amount that may be subject to forfeiture in a particular case," so we don't hold for certain that his conduct fit under 18 U.S.C. § 981(a)(2)(A). George, 886 F.3d at 40. Instead, we don't think it matters for the question we were asked to decide.

93

BHS financial aid, but not down here? Don't both apply a preponderance standard?

We get it. But let us explain ourselves, because we see a few differences that reconcile these outcomes. For one, the calculation process is different. Our caselaw requires that the "gain/loss" calculation be governed by Alphas' burden-shifting framework. Once the government made the initial showing that there was $177,600 paid out and the district court thought that number reasonable, the burden shifted to Griffin to knock that figure down for Guidelines purposes. And, for the reasons explained above, the district court reasonably didn't accept any of his arguments.

But in the forfeiture context, the government bears the burden of proving the particularities of forfeiture all the way through.30 See 18 U.S.C. § 983(c)(1) ("In a suit or action brought under any civil forfeiture statute for the civil forfeiture of any property -- (1) the burden of proof is on the government to establish, by a preponderance of the evidence, that the property is subject to forfeiture."); cf. United States v. Flete-Garcia, 925 F.3d 17, 38 (1st Cir. 2019) (noting, in a parallel context,

Except, of course, any "direct costs" under § 981(a)(2)(B), 30

which we've previously held that the defendant must show. United States v. Carpenter, 941 F.3d 1, 10 (1st Cir. 2019). Yet (at least by the parties' characterizations) this case doesn't concern "direct costs," so that's no matter.

94

that "there are significant differences between the proper method of calculating loss and the proper method of calculating restitution").

For another, these different results also make sense considering the different purposes they serve in the criminal justice process. Cf. Flete-Garcia, 925 F.3d at 37 (comparing purposes of Guidelines loss and restitution). The loss calculation serves "as a measure of the seriousness of the offense and the defendant's relative culpability." U.S.S.G. § 2B1.1 cmt. background. More practically, the loss calculation plays only one part of a larger, non-binding calculation that the district court must undertake. That's why we're alright with settling for a "reasonable estimate" (read: a rough proxy) in the loss context. United States v. Curran, 525 F.3d 74, 78 (1st Cir. 2008) (internal citations omitted).

Forfeiture orders play a more concrete role: they "help to ensure that crime does not pay," because they "at once punish wrongdoing, deter future illegality, and lessen the economic power of criminal enterprises." Kaley v. United States, 571 U.S. 320, 323 (2014) (cleaned up). And in stark contrast to the non-binding and sometimes abstract-feeling loss calculation, they do so in a very material way: the defendant must cough up exactly as much as is found to be forfeitable. See id. (explaining how "forfeitures are imposed upon conviction to confiscate assets used in or gained

95

from certain serious crimes.").31 Requiring a bit more precision from the government in this context seems natural.32 So we vacate the district court's order and remand for proceedings consistent with this opinion. We'll summarize what we mean: to revive the past forfeiture order (in full or in part), the government will need to show what financial aid specifically was "traceable to" (i.e., resulted from, as explained earlier in the opinion) the fraud. What wasn't "traceable to" the fraud (i.e., what Griffin would have gotten if he had submitted honest forms) can't be subject to forfeiture. And if the government can't separate wheat from chaff here in a way that the district court (with our guidance in mind) finds satisfactory, forfeiture is not appropriate.

31True, Kaley involved the criminal forfeiture statute. But we think its general observations about the role of forfeiture ring true for the civil forfeiture statute, too.

32 We'll also note the somewhat concerning context of this forfeiture order. Originally, the government sought both restitution and forfeiture for the financial aid amount. But BHS didn't want restitution, and it wouldn't say it would not have granted financial aid but-for the fraud, so the court didn't order any restitution on that point. Yet the government still pursued forfeiture of that money. It's worth remembering that "forfeiture and restitution are distinct remedies." United States v. Webber, 536 F.3d 584, 602 (7th Cir. 2008) (cleaned up). "Forfeiture serves no remedial purpose, is designed to punish the offender, and cannot be imposed upon innocent owners," United States v. Bajakajian, 524 U.S. 321, 332 (1998) (cleaned up), and we don't think that forfeiture of the full financial aid amount should be used to punish Griffin where BHS cannot say it would not have granted financial aid if he provided all true information.

96

PARTING THOUGHTS

All that's to say: we affirm as to everything but the forfeiture order against Griffin. On that, and that alone, we vacate and remand for more litigation consistent with this opinion.