BenchRecon Data Study · Federal
Below the guidelines 50.8% of the time, but who's driving it?
Across 376,349 federal sentencing records (USSC, FY2020–FY2025), 45.9% of sentences land within the guideline range and 50.8% land below it. But 29.5% of that below-range rate is government-sponsored , prosecutor-driven fast-track and cooperation motions, and only 21.3% reflects the court's own discretion. At the extremes, the two rates rank districts in a sharply different order.
We analyzed every federal individual-offender record in the U.S. Sentencing Commission's public datafile for FY2020 through FY2025, 376,349 cases, and split the below-range sentences by mechanism: those driven by the prosecutor (substantial assistance §5K1.1, fast-track §5K3.1, government departures and variances) versus those driven by the court (downward departures and below-range variances the bench chose on its own). The split is what actually tells you something about a district's judges.
Aggregate analysis. Reproducible from public USSC data. No individual is named.
A high below-range rate is not judicial leniency
The most important thing this study shows is what a raw below-range rate does not tell you. A district that sentences below the guidelines 84.3% of the time looks dramatically lenient, until you split that number by who drove it. Southern District of California is below the guidelines 84.3% of the time, but only 4.2% is judicial. The remaining 80.1% is government-sponsored fast-track, the district's heavy early-disposition immigration docket generates enormous volumes of §5K3.1 government variances that have nothing to do with how the bench views any individual case.
The government-sponsored below-range rate is prosecutor-driven: it reflects DOJ charging policy, the availability of fast-track programs, and the volume of cooperation agreements in a given district. The judicial rate, downward departures and below-range variances the court chose on its own, is the number that reflects the district's bench. The two rates rank districts very differently.
The 376,349 records span FY2020–FY2025. Above-range sentences (3.3%, n=12,563) account for the remaining share.
Total below-range, top 10 districts
Sorted by the combined below-range rate (government + judicial). The judicial column shows how much of that rate is the court's own discretion. In border districts with heavy fast-track immigration dockets, the judicial rate is a small fraction of the headline.
| District | n | Total below % | Judicial % |
|---|---|---|---|
| Southern District of California | 18,355 | 84.3% | 4.2% |
| Northern District of California | 2,523 | 82.1% | 38.2% |
| District of Oregon | 2,558 | 80% | 14.3% |
| District of Nevada | 1,806 | 78.3% | 21.2% |
| Eastern District of Wisconsin | 1,780 | 78.1% | 42.2% |
| Southern District of New York | 6,243 | 77.9% | 56.9% |
| District of Utah | 4,261 | 76.5% | 9% |
| Eastern District of New York | 3,589 | 75% | 43.1% |
| Southern District of Ohio | 3,394 | 73.1% | 22.1% |
| Middle District of Tennessee | 1,704 | 72.7% | 28.1% |
Only districts with at least 1,000 FY2020–FY2025 records with a coded SENTRNGE value are shown. “Total below %” = (gov-below + judicial-below) / n. “Judicial %” = (downward departure + below-range variance the court chose) / n.
Judicial below-range ranking
When the government-sponsored share is stripped out, the ranking changes completely. These are the districts where the bench itself is most, and least, likely to sentence below the guidelines on its own initiative.
Highest judicial below-range rate, top 12 districts
| District | n | Judicial % | Gov % |
|---|---|---|---|
| Southern District of New York | 6,243 | 56.9% | 21% |
| Northern District of Illinois | 3,528 | 50.7% | 16.6% |
| Eastern District of Michigan | 3,916 | 47.5% | 15% |
| District of Connecticut | 1,853 | 44.3% | 26.7% |
| Eastern District of New York | 3,589 | 43.1% | 31.9% |
| Eastern District of Wisconsin | 1,780 | 42.2% | 35.9% |
| Southern District of West Virginia | 1,421 | 40.1% | 10.5% |
| Central District of Illinois | 1,461 | 39.1% | 22.8% |
| Northern District of California | 2,523 | 38.2% | 43.9% |
| Central District of California | 5,671 | 38% | 33.2% |
| District of Minnesota | 2,038 | 37.1% | 30.3% |
| Southern District of Iowa | 2,481 | 37.1% | 23.5% |
Lowest judicial below-range rate, bottom 8 districts
| District | n | Judicial % | Gov % |
|---|---|---|---|
| Eastern District of Oklahoma | 1,146 | 12.3% | 18.4% |
| District of North Dakota | 1,950 | 11.8% | 49.7% |
| Eastern District of Texas | 4,895 | 11.8% | 21.2% |
| Southern District of Georgia | 2,361 | 11.3% | 19.8% |
| Western District of Texas | 45,362 | 10.3% | 5.3% |
| District of Utah | 4,261 | 9% | 67.5% |
| District of Arizona | 25,896 | 6.1% | 62.3% |
| Southern District of California | 18,355 | 4.2% | 80.1% |
“Judicial %” = (SENTRNGE 5 downward departure + SENTRNGE 8 below-range variance) / n. “Gov %” = (SENTRNGE 2, 3, 4, 7) / n. Every district shown has n ≥ 1,000. Sorted by judicial below-range rate descending (top 12) and ascending (bottom 8).
The district that rewrites the story: Southern District of California
Southern District of California is the single district whose raw below-range rate most dramatically overstates judicial discretion. It ranks near the top of any total-below list, 84.3% of its 18,355 FY2020–FY2025 cases are sentenced below the guidelines.
But split it by mechanism: 80.1% is government-sponsored and only 4.2% is judicial. The 80.1% government share reflects the district's heavy early-disposition immigration program, border districts running §5K3.1 fast-track dockets generate enormous volumes of government-sponsored variances that compress the total sentence length across thousands of cases simultaneously. That is a DOJ prosecutorial program, not a signal about how the district's judges sentence any individual defendant.
The judicial rate, 4.2%, is what actually reflects the bench. It ranks Southern District of California among the lowest-discretion districts in the country once the fast-track share is removed.
Florida's three federal districts
Florida's three federal districts land within a few points of each other on the judicial below-range rate (all near 30%), yet their total below-range rates spread wider , the difference is how much each leans on government-sponsored departures. The split matters for defense strategy: the judicial rate, not the headline total, is what reflects a bench's independent willingness to sentence below the range.
| District | n | Total below % | Gov % | Judicial % |
|---|---|---|---|---|
| Middle District of Florida | 7,067 | 50.8% | 20% | 30.7% |
| Northern District of Florida | 1,608 | 53.3% | 24% | 29.3% |
| Southern District of Florida | 8,005 | 45.9% | 13.6% | 32.3% |
Within-range rate by fiscal year
The national within-range share has fluctuated over this window. The FY2021 and FY2022 figures reproduce the USSC's own published Sourcebook within-range percentages to the decimal, that match is the data's grounding check, confirming the SENTRNGE code map is correctly applied.
| Fiscal Year | Within-range % | Grounding note |
|---|---|---|
| FY2020 | 50.4% | |
| FY2021 | 42.8% | Reproduces USSC Sourcebook 42.8% |
| FY2022 | 41.9% | Reproduces USSC Sourcebook 41.9% |
| FY2023 | 42.4% | |
| FY2024 | 45.7% | |
| FY2025 | 51.5% |
Methodology & limitations
Source & method
- Source: U.S. Sentencing Commission, Commission Datafiles, individual-offender records (public). ussc.gov/research/datafiles. Snapshot analyzed: 2026-06-26.
- Population: 376,349 FY2020–FY2025 individual-offender records with a coded SENTRNGE value.
- SENTRNGE code map: 0 = within range; 1 and 6 = above range; 2, 3, 4, 7 = government-sponsored below range (§5K1.1 substantial assistance, §5K3.1 fast-track, government departure, government variance); 5 and 8 = judicial below-range (downward departure and below-range variance the court chose). Verified against the USSC Variable Codebook; ground-checked against USSC published Sourcebook within-range figures for FY2021 (42.8%) and FY2022 (41.9%).
- District floor: only districts with n ≥ 1,000 coded-SENTRNGE records are shown in ranking tables (80 districts meet this threshold in this window).
- Reproducible: a published verification script re-queries the live USSC datafile and re-checks every figure on this page against it.
What the data does NOT show
- The government-sponsored below-range rate reflects DOJ charging policy, fast-track program availability, and cooperation-agreement volume in a district. It is not a signal of judicial philosophy or bench leniency. Some border districts running §5K3.1 early-disposition immigration programs (e.g. S.D. California and D. Arizona) have government-sponsored rates that dominate their below-range total, driven by those programs rather than by the bench.
- The judicial below-range rate is a district-level descriptive. It is not controlled for offense mix, criminal-history category, or cooperation, all of which affect whether a court has discretion to go below range. Districts with different offense compositions are not fully comparable even on the judicial rate.
- Rates describe populations of cases, not any individual sentence, and not the outcome in any specific matter. Do not infer anything about a specific judge, defendant, or case from these aggregate district figures.
- This is the federal guidelines population only. This study identifies no individual.
Cite this analysis
Journalists and researchers, please link to this page as the source.
BenchRecon, “Federal below-guideline sentencing by district: government vs. judicial downward departures” (USSC FY2020–FY2025 individual-offender data). https://benchrecon.com/data/federal-below-guideline-sentencing
The underlying figures come from the public USSC datafile and are reproducible by re-running the published verification script against it.
Build the below-range profile for your district
BenchRecon's Sentencing Snapshot turns this district-level signal into a case-specific comparables brief, the below-range rate, median, and distribution for your charge, criminal-history category, and district, cited to the underlying USSC records, ready to file as a §3553(a) exhibit.
Common questions
- What does this study measure?
- The share of federal sentencing records that fall within, above, or below the guideline range, FY2020–FY2025 (USSC data, n=376,349). Nationally, 45.9% land within range and 50.8% land below it. The key finding is that the 50.8% below-range rate splits into 29.5% government-sponsored (prosecutor-driven) and only 21.3% judicial, and those two rates rank districts very differently.
- What is the difference between government-sponsored and judicial below-range sentences?
- A below-range sentence can be driven by the prosecutor or by the court, and the distinction matters enormously. Government-sponsored departures and variances (SENTRNGE codes 2, 3, 4, 7) include substantial-assistance motions (§§5K1.1, where the prosecutor rewards cooperation), fast-track early-disposition programs (§5K3.1, used heavily in border immigration dockets), government departure motions, and government variances. These reflect DOJ charging policy and program availability, not the bench’s own philosophy. Judicial/non-government departures and variances (SENTRNGE codes 5 and 8) are the court’s own downward departures and below-range variances, the signal that reflects the district’s bench.
- Does a high total below-range rate mean a district's judges are lenient?
- Not by itself, and this study is specifically designed to show why not. Southern District of California sentences below the range 84.3% of the time overall, but only 4.2% is judicial. The rest is government-sponsored fast-track driven by the district’s heavy immigration early-disposition docket. Looking at the judicial rate alone, 4.2%, produces a completely different picture of that district’s bench than the raw 84.3% headline.
- How were the SENTRNGE code categories verified?
- The SENTRNGE code-to-category map was verified against the USSC Variable Codebook. As an independent ground check, the study confirms that code 0 (within guideline range) reproduces the USSC’s own published Sourcebook within-range percentages to the decimal: 42.8% in FY2021 (Sourcebook: 42.8%) and 41.9% in FY2022 (Sourcebook: 41.9%). This match confirms the code map is correct, if the codes were misread, the reproduced within-range share would not match.
- Why are only districts with 1,000+ records shown?
- A percentage computed from a small denominator is unstable, a single case can move it by several percentage points. The 1,000-record floor ensures each district rate rests on a stable sample. Districts below that floor are omitted from all ranking tables.
- Is this every federal case?
- It is the USSC individual-offender datafile for FY2020–FY2025, restricted to records with a coded SENTRNGE value (blank SENTRNGE records, which lack disposition-type information, are excluded from the denominator). It covers original sentences under the guidelines; it is not a census of every federal disposition. Each district row carries its own record count (n).
- Can I cite or reuse this?
- Yes, please link to this page as the source. The underlying data is the public USSC datafile, and every figure is reproducible by re-running the published verification script against it.