This is the first piece in a new section of this site. The rule for the section is stricter than the rest: a named public dataset for every claim, every disparity stated with its denominator, the innocent explanation weighed before the damning one, and a fix at the end that a real office could actually carry out. No manufactured outrage. If the data doesn't support a sentence, the sentence doesn't run.
The dataset here is the Minnesota Sentencing Guidelines Commission's 2024 Sentencing Practices: Annual Summary Statistics for Felony Cases, a 109-page report the Commission's research staff publishes every year and almost nobody reads. It covers every felony sentence handed down in Minnesota in 2024, 14,193 of them with a recorded race, and it compares those people to the Census Bureau's estimate of the state's adult population. I read the whole thing. Here is what it says.
The disparity, with its denominator
Start with the number that should be on the front page of every newspaper in the state and isn't.
In 2024, Minnesota sentenced 201 white adults for felonies for every 100,000 white adults in the state. It sentenced 1,243 Black adults per 100,000. It sentenced 1,809 American Indian adults per 100,000. Hispanic adults were sentenced at 351 per 100,000; Asian adults at 164. The statewide rate was 316.
Put differently: a Black adult in Minnesota was 6.2 times as likely as a white adult to be sentenced for a felony last year. An American Indian adult was 9 times as likely.
The shares tell the same story. White Minnesotans are 81 percent of the adult population and 52 percent of the people sentenced. Black Minnesotans are 7.7 percent of adults and 30 percent of the sentenced. American Indians are 1.5 percent of adults and 9 percent of the sentenced. (The population percentages add to slightly more than 100 because the Census counts multiracial residents in more than one category; the Commission's own footnote says so.)
It gets sharper by geography. In the Fourth Judicial District, which is Hennepin County, Black residents are 14 percent of the population and 61 percent of the felony sentences. In the Ninth District, Bemidji and the northwest, American Indians are 6 percent of the population and 34 percent of the sentences. In the Eighth, around Willmar, Hispanic residents are 8 percent of the population and 18 percent of the sentences.
None of this is new, exactly. In 1981, the first full year of the guidelines, 82 percent of people sentenced were white. The Commission is careful to say that the state's changing demographics "do not entirely account for" the shift. Its own population-adjusted chart, going back to 2002, shows the Black, American Indian, and Hispanic rates "consistently higher than the total rate" every single year.
Now the part the outrage merchants skip
Here is where an honest reading gets uncomfortable for everyone.
If you assume the disparity is made in the courtroom, by a judge looking at a Black defendant and reaching for a longer sentence, the Commission's data says you are wrong, at least about 2024.
Minnesota's guidelines produce a presumptive sentence for every case, from a grid of offense severity and criminal history. The judge can follow it or depart from it. The report tracks both.
-
Departing downward on prison time. Among people who actually went to prison, judges gave a sentence shorter than the guideline called for to 24 percent of Black defendants and 13 percent of white defendants. American Indians: 14 percent. That is not a typo. When a Black defendant was sent to prison in Minnesota in 2024, the judge was almost twice as likely to cut the term below the guideline as for a white defendant.
-
Departing downward on prison itself. When the guideline called for prison, judges kept the defendant out of prison in 45 percent of white cases and 39 percent of Black cases. That gap runs the other way, six points against Black defendants. But look at who has the smallest break: American Indian defendants, at 31 percent, and the reason turns out to be geography more than race, which I will come back to.
-
Prison rates track the grid. The guidelines recommended prison for 46 percent of Black defendants and 34 percent of white defendants. Actual prison rates were 30 percent and 21 percent. The gap between recommended and actual was 16 points for Black defendants and 14 for white. In other words, the judges applied roughly the same discount to both groups. The disparity in who goes to prison was already sitting in the presumptive sentence before the judge said a word.
So the question is not "why do judges sentence Black Minnesotans more harshly." On this evidence, at the hearing itself, they don't. The question is why a Black adult is six times as likely to be standing in front of the judge in the first place, and why, once there, the grid recommends prison for 46 percent of them and 34 percent of white defendants.
Where the disparity is actually built
Two places, and the Commission's report can see one of them clearly and only the edge of the other.
First, the front door. The sentencing report starts at conviction. Everything before that, the stop, the arrest, the charging decision, the plea offer, is invisible to it. A six-to-one gap in who gets sentenced is mostly a gap in who gets arrested and charged, and the Commission has no data on that. Neither, as a practical matter, does the public. Minnesota's county attorneys do not publish charging decisions by race. This site found the same wall when we looked at cannabis-in-a-vehicle charges: the only reason anyone knows that 42 percent of those charged in Hennepin County were Black was a reporter's records request.
Second, the grid. The presumptive sentence comes from two inputs: how serious the offense is, and the defendant's criminal history score. Both push the Black presumptive prison rate to 46 percent.
Offense mix is real, and I am going to say so plainly because the section's rule is to weigh the innocent explanation first. The Commission's Table 9 breaks every offense down by race. Of 86 first-degree aggravated robbery sentences in 2024, 86 percent were Black defendants. Of 105 second-degree murder sentences, about two-thirds. Simple robbery, 64 percent. Those are offenses the grid rightly treats as serious, and they are not evenly distributed. On the other side of the table, first-degree criminal sexual conduct defendants were 50 percent white, criminal vehicular homicide 66 percent white, and soliciting a minor for sex 81 percent white. Anyone who tells you the sentencing gap is purely a story about bias has not read Table 9. A meaningful share of it is the offenses charged.
But the criminal history score is different, and it is the part of the machine nobody elected. Every prior conviction adds points; the points move you rightward on the grid toward a presumptive prison cell. A population that is arrested and charged at six times the rate accumulates history at something like six times the rate. The score then converts yesterday's enforcement disparity into today's presumptive prison sentence, automatically, with no one deciding anything. That is a legal rule producing a disparate result by design, and it is precisely the kind of thing this section exists to name.
The Commission is not blind to this. It publishes "demographic impact statements" on proposed sentencing legislation, so the Legislature can see in advance who a new penalty will fall on. The statute that created the Commission tells it to build guidelines "based on reasonable offense and offender characteristics." Whether a criminal history score that mechanically compounds an upstream disparity is a reasonable offender characteristic is a question the Commission has the authority to ask about its own grid, and has not.
The zip code problem, which is also a race problem
The American Indian numbers deserve their own paragraph, because they show how geography hides inside race.
American Indian defendants received the fewest breaks in the state: a 31 percent downward dispositional departure rate against 45 percent for white defendants, and a mere eight-point gap between the prison rate the grid recommended and the prison rate they got, against 14 points statewide. On its face that looks like judges being hardest on Native defendants.
Now look at the districts. In the Third, Sixth, and Ninth Judicial Districts (Rochester, Duluth, Bemidji), judges cut prison terms below the guideline in about 6 percent of executed sentences. In the Second and Fourth (St. Paul, Minneapolis), they did it in 30 and 31 percent. A five-to-one difference in judicial leniency depending on which courthouse you walk into. The Ninth District is where a third of the felony sentences fall on American Indians. The rural bench departs least, and Native Minnesotans are disproportionately sentenced in rural courts. Some of the "race" gap is a "county" gap wearing a different label. That does not make it better. It makes the fix different.
What I could not verify, and what this report cannot tell you
The report is convictions and sentences. It cannot separate charging discretion from underlying conduct; it does not contain arrest data; it does not say how many of the departures were prosecutor-agreed pleas versus judicial decisions over the state's objection (the Commission notes elsewhere that many downward departures come "with prosecutorial agreement or neutrality," but not by race). The population denominators are Census estimates for adults and, for the district comparison, residents age 15 and up, so the two tables are not perfectly aligned; I used each as the Commission presents it. I did not independently recompute any figure; every number above is the Commission's, and the report is linked below so you can check me.
What we can do
Three things, all within reach of the Attorney General's office and the Legislature, none of which requires anyone to agree about the cause first.
1. Open the front door to daylight. Require every county attorney to report charging decisions by offense, race, and disposition to a single statewide dataset, published annually alongside the Commission's sentencing report. The state already does this for sentences. The gap between arrest and sentence is the one place in the pipeline where the largest disparity is created and no data exists. An Attorney General cannot order 87 county attorneys to do this, but can build the template, publish the office's own numbers first, and put the bill in front of the Legislature with the Commission's report attached.
2. Run the impact statement on the grid itself. The Commission already scores proposed bills for demographic impact. It should score its own criminal history rules: how much of the 46-versus-34 presumptive gap is offense severity and how much is accumulated history points, and what happens to the gap if old convictions decay faster or custody-status points are trimmed. Then publish it. If the answer is "the gap is all offense mix," we will all have learned something. If it isn't, the Commission has the authority to change the grid without waiting for the Legislature.
3. Close the courthouse lottery. A defendant's chance of a break should not be five times higher in Minneapolis than in Bemidji. The Commission and the Judicial Council should set expectations for departure practice, publish district-level departure rates every year with names attached to courthouses, not just numbers, and give rural benches the same access to the sentencing alternatives that urban counties have built. When the Ninth District departs at 6 percent, the question to ask is not "are those judges biased" but "what would they need to have the option."
The honest concession, one more time: a real part of this gap is the offenses people are charged with, and no reform of the grid changes who committed an aggravated robbery. But a six-to-one sentencing rate is not explained by robbery counts, the criminal history score is a rule we wrote and can rewrite, and the state that invented sentencing guidelines in 1980 can be the state that finally measures what they do.
First the facts. Then the fix.
Sources
Minnesota Sentencing Guidelines Commission, 2024 Sentencing Practices: Annual Summary Statistics for Felony Cases (PDF read in full, 109 pages, downloaded from mn.gov/sentencing-guidelines on Sept. 6, 2026): Executive Summary (pp. 5–6); Figure 7 and Figure 8 and the accompanying text on population-adjusted sentencing rates (p. 16); Table 1, felony cases sentenced by sex, race/ethnicity, and judicial district compared to the 2024 estimated adult population, with the Census Bureau's July 1, 2024 estimates as the denominator (p. 18); the judicial-district comparison (p. 19); Table 2, incarceration rates by race (p. 23) and the accompanying discussion of presumptive versus actual prison rates (pp. 24–25); Table 3, dispositional departures by race and district (p. 34); Table 4, durational departures among executed prison sentences by race and district (p. 42); Table 9, offense type by race (pp. 53–56); Table 21, presumptive and actual imprisonment rates by race, 1978 and 1981–2024 (pp. 77–78). Minn. Stat. § 244.09, subd. 5 (verified verbatim at revisor.mn.gov): the guidelines "shall be based on reasonable offense and offender characteristics" and "shall be advisory to the district court." Minnesota Sentencing Guidelines Commission, Research & Data pages listing its Demographic Impact Statements and Data Library (mn.gov/sentencing-guidelines/research-data, viewed Sept. 6, 2026). This site's earlier reporting on cannabis-in-a-vehicle charging by race in Hennepin County, relying on Minnesota Reformer's records-based reporting (Dec. 9, 2025).
Arrest and charging data by race are not published statewide and are not asserted here. The share of downward departures that were prosecutor-agreed is not reported by race in the Commission's summary and is not asserted. Multipliers (6.2×, 9×) are simple ratios of the Commission's per-100,000 rates. The Commission's report notes a small difference between its total case count (14,229) and cases with a recorded race (14,193); the race tables use the latter. Corrections: campaign@madgettformn.com.