Guideline Adherence and Plea Trends, FY2015-2024

Within-guideline sentencing peaked in 2020, then fell sharply

The share of federal sentences imposed within the guideline range climbed steadily from 15.3 percent in fiscal year 2015 to a decade high of 27.9 percent in fiscal year 2020, then dropped back to 14.4 percent in 2021 and has hovered in the mid-teens since, reaching 17.5 percent in 2024. U.S. Sentencing Commission Individual Offender Datafiles, FY2015-2024. The 2020 peak and the 2021 reversal coincide with the pandemic-era disruption to federal court operations, which shifted both the volume and the composition of cases reaching sentencing. The broader picture is that for most years, fewer than one in five federal sentences landed squarely within the calculated guideline range, with the rest reflecting departures and variances.

That low within-range share is a defining feature of post-Booker federal sentencing. After the 2005 decision made the guidelines advisory, judges retained discretion to vary, and government-sponsored reductions for substantial assistance and fast-track immigration programs account for a large block of below-range sentences. U.S. Sentencing Commission Sourcebook of Federal Sentencing Statistics, 2024. The within-range figure is therefore best read as a measure of how often the advisory range and the imposed sentence align, not as a compliance score.

Average sentences rose as caseloads fell

The national average sentence moved in the opposite direction from the within-guideline share over the back half of the decade. After dipping to 43.3 months in fiscal year 2020, the average climbed each year to 57.7 months in 2024, the highest in the ten-year window. U.S. Sentencing Commission Individual Offender Datafiles, FY2015-2024. Over the same span the total caseload fell from a peak of 76,376 cases in 2019 to 61,546 in 2024. The combination suggests a docket that shrank while shifting toward more serious, longer-sentenced offenses, raising the average even as fewer cases were sentenced overall.

The guilty-plea rate stayed remarkably stable across the decade, ranging only from 67.8 percent in 2020 to 77.2 percent in 2022 and sitting at 72.1 percent in 2024. U.S. Sentencing Commission Sourcebook of Federal Sentencing Statistics, 2024. Federal cases overwhelmingly resolve by plea rather than trial, and that did not change materially over the period; the share never fell below two-thirds. The brief dip in 2020 again reflects pandemic-era court closures that delayed some dispositions rather than any structural shift toward trials.

What the ten-year arc shows

Taken together, the decade tells a story of a smaller but more serious federal docket. Caseloads peaked before 2020 and have not recovered, while the average term has risen to a ten-year high and the within-guideline share has settled below where it stood in the late 2010s. U.S. Sentencing Commission Individual Offender Datafiles, FY2015-2024. None of these series moves in lockstep, which underscores that headline averages reflect a shifting mix of offenses as much as any change in sentencing severity for a fixed offense. Offense-level and geographic detail behind these national figures is available in the companion offense category breakdown and state aggregation.

Why a ten-year window changes the reading

A single year of federal sentencing data invites a tempting but unreliable conclusion: that whatever the current average happens to be reflects how tough the system is right now. The ten-year window dismantles that reading. The average term in 2020 was nearly fifteen months lower than in 2024, yet no statute was rewritten to make sentencing harsher across the board in between. U.S. Sentencing Commission Sourcebook of Federal Sentencing Statistics, 2024. The shift came instead from a changing caseload, the pandemic disruption to court operations, and the slow recovery of a docket that now skews toward more serious offenses. Reading any one year in isolation would have produced a misleading story in three of the last five years. The decade view is the corrective: it shows the direction and the volatility of each measure, separates genuine trend from one-year noise, and makes clear that the federal sentencing average is a moving composite rather than a fixed dial. That is the discipline this site applies to every figure it publishes.