Methodology

Sources

Metric definitions

Per-resident metrics divide a dollar amount by population for the same year. Shares divide a revenue component by total revenue.

Where the money comes from and where it goes

Each city page breaks revenue and spending into categories and lines, per resident, against the peer median. Every category lists the Census item codes behind it.

• Each side adds up exactly: the revenue categories sum to total revenue, and the spending categories sum to total expenditure, the same totals as the headline figures.
• Spending lines combine operating spending and construction. Where a line or category includes construction, the construction part is shown separately (“incl. $82 construction”).
• Sewer, trash, parking and recreation charges sit under “Fees & charges”, following the Census classification. Water, electric, gas and transit revenue is “Utility & transit charges”.
• “None reported” means the city filed nothing in that category for the year. It is a reported zero, not missing data.
• A withheld total withholds its whole side, and a withheld component withholds the category it belongs to.
• A city is marked a regional hub when its sales tax per resident is at least 1.5× the statewide median for the year. Visitors and shoppers who live elsewhere pay much of that tax, so revenue per resident reads high for that reason; the page says so wherever it applies.

Peer groups

Each municipality is compared with the five Colorado municipalities nearest to it on population, median household income and region. Peers are chosen separately for each fiscal year, so a peer group can change from one year to the next; trend charts mark the years it changed.

Nearness is a weighted distance. Population and income are compared on a log scale, so a town of 5,000 is as far from one of 10,000 as a city of 50,000 is from one of 100,000:

• Population: weight 1.5 on the difference in log population.
• Median household income: weight 1.0 on the difference in log income (American Community Survey, latest vintage at or before the fiscal year).
• Region: 0.3 is added for a different region within the same part of the state, and 1.0 for a different part of the state (Front Range, mountain and western, or plains).
• Budget size is not used. Matching on total revenue would pick peers partly on the figures being compared, so it carries no weight. A municipality needs reported revenue, population and income for the year to be eligible as a peer.

Peers can still differ in what they run. Municipalities that operate their own water, electric, gas or transit utilities are marked “Utilities”, and shopping and visitor hubs are marked “Regional hub”, wherever they appear in a peer table.

When too few comparable municipalities have data for a year, the peer benchmark is marked “Benchmark unavailable” rather than computed from a thin group.

Medians and percentiles

The peer median is the middle value of the peer group; the Colorado median is the middle value of every municipality with a figure for that year.

A percentile says what share of Colorado municipalities have a lower value. The 62nd percentile means higher than 62% of them. Percentiles and ranks are descriptive: a higher tax figure is not better or worse on its own, so we don’t color them as good or bad.

Statewide comparisons

The Census Bureau collects finances from every municipality only in Census of Governments years (FY2016 and FY2021), when about 225 Colorado municipalities report. In other years it surveys a sample of about 80 to 100, mostly larger cities. A Colorado median, rank or percentile therefore describes the municipalities that reported that year, not the whole state, and medians from a sample year and a full year aren’t comparable.

Every statewide figure on these pages states how many municipalities it covers (“Colorado median · 103 with a published figure”), and trend charts plot the statewide median only in full-coverage years. A year counts as full coverage when at least 200 municipalities report revenue per resident.

Trends

Dollar metrics show a compound annual growth rate (CAGR) over three and five years, written as “+3.1%/yr”. Metrics that are already rates or shares show a point change instead, written as “+1.2 pts”, because a percent change of a percent is hard to read. Charts break the line wherever a year is missing or withheld; nothing is interpolated. All dollar figures and growth rates are nominal: they are not adjusted for inflation, so part of any dollar growth reflects rising prices rather than more resources or service.

Fiscal years

Years are each government’s own fiscal year as reported to the Census Bureau, labeled FY2023 and so on. Demographics use the same year when available, otherwise the nearest earlier year, and the page says so.

Why the lag. The Census Bureau publishes each year’s finance survey well after governments close their books, usually more than a year later, so the latest year here trails the current one. Figures describe that fiscal year, not today’s budget. New years are added when the Census Bureau publishes them; each city page states its latest year and when the data was retrieved.

Quality rules

We would rather show nothing than show a wrong number. A value is withheld when it:

• breaks from the city’s own history — less than a third or more than three times its typical level;
• is extremely far from its peers; or
• is an outlier and the city has no history to check it against.

Withheld values stay withheld until they are verified against the city’s audited annual comprehensive financial report (ACFR). They are labeled “Withheld” with the reason, and are not shown anywhere on the page, including charts and peer tables.

Things to know

• Census figures are what each city reported to the Census Bureau. They follow Census classifications, which don’t always line up with a city’s own budget categories or funds.
• In years a city didn’t respond, the Census Bureau fills in estimates. Pages note when estimates make up most of a year’s revenue.
• Small towns are surveyed mainly in Census of Governments years (FY2016 and FY2021), so their pages show fewer years.
• Capital outlay counts construction only, not equipment or land purchases.
• Totals follow the Census definition, which includes city-run utilities (water, electric, gas, transit) and liquor stores. Cities that own their utilities show higher revenue and spending per resident for that reason alone; their pages are marked “Includes city-owned utilities”.

Debt

Debt outstanding is the total the municipality owed to bondholders and lenders as reported to the Census Bureau: long-term bonds, notes and loans plus short-term debt. Debt per resident divides it by population; debt to annual revenue divides it by total revenue for the same year.

• Utility and enterprise debt. The total includes revenue bonds for city-run water, sewer, electric and other enterprises. Those are usually repaid from the rates customers pay rather than from taxes, so cities that run their own utilities can carry more debt per resident without a heavier tax burden. Their pages are marked “Includes city-owned utilities”.
• Small towns. In a town of a few thousand residents one bond issue can move debt per resident by thousands of dollars in a single year, and the figure falls back as the bond is repaid. Read a small town’s debt across several years, not one.
• A balance at fiscal year end. Debt outstanding is what was owed on the last day of the fiscal year, not what was borrowed or repaid during it. A bond issued or retired just before or after year end moves the figure on its own.
• Debt to annual revenue. Dividing by revenue instead of residents relates debt to what the municipality collects in a year: 1.00 means debt equal to one year’s revenue. Municipalities with high revenue per resident, from utilities or visitors, can sit at a very different percentile on the two measures.

Public safety

The public safety pages benchmark police spending, staffing, reported crime and clearances for city staff and councils. They describe; they don’t grade. Every metric is neutral, nothing is colored as good or bad, and there are no rankings.

Sources

• FBI Crime Data Explorer — NIBRS data that Colorado agencies submit through the Colorado Bureau of Investigation (CBI). Offense counts and clearances follow FBI summary definitions.
• FBI police employee counts — sworn officers and civilian staff on the payroll as of October 31.
• Census government finance, function 62 (police protection) — current operations (E62), construction (F62) and, where reported, payments to other governments for police services.
• CDOT statewide crash listing — used for road metrics once they are published.

Definitions

• Violent crime is murder and nonnegligent manslaughter, rape, robbery and aggravated assault. Property crime is burglary, larceny-theft and motor vehicle theft; arson is excluded, following FBI summary counts. Both are shown per 1,000 residents.
• Clearance is offenses cleared by arrest or exceptional means during the year, as a share of offenses reported that year. Clearances can include offenses from earlier years, so a rate can exceed 100%; pages note it when that happens.
• Officers per 1,000 residents divides full-time sworn officers (as of October 31) by population.
• Police spending per officer divides fiscal-year police spending by sworn officers. Spending covers civilian staff, equipment and construction as well as officer pay, so it is not a salary figure.

Rules

• A year’s crime figures are withheld when the agency reported fewer than 12 months.
• A clearance rate isn’t shown when fewer than 10 offenses of that type were reported.
• Crime rates built on fewer than 10 reported incidents are flagged, because a few incidents move them a lot.
• Peers are the finance peers. When a crime year is newer than the latest fiscal year with peers, the latest fiscal year’s peers are used and the page says so.
• Crime and staffing rates divide by the Census population estimate for the same year, the same estimates the finance pages use. Each city’s calculation panel shows the population and its estimate year.
• Crime, clearances and staffing are calendar years, labeled with the year alone. Police spending is the government’s fiscal year, labeled with an FY prefix. When a fiscal year isn’t published yet, the latest earlier one is shown and labeled “latest available”. Each figure carries its own label, and tables group fiscal-year and calendar-year columns under separate headings.
• The Census Bureau stopped reporting police equipment purchases and payments to other governments after FY2021, so police spending from FY2022 on excludes them. Trend charts mark the change, and a five-year change that spans it carries a footnote.

Limits

• Reported crime reflects what is reported to police and how it is recorded, not only how much crime occurs.
• A city’s own published statistics may differ: CBI counts some offenses once per victim, while many cities count once per case.
• A clearance rate describes how many offenses were cleared during the year relative to offenses reported that year. It doesn’t on its own measure how effective a department is: clearances can include earlier years’ offenses, and agencies record clearances differently.
• Towns without their own police department are policed by another agency, such as the county sheriff, whose crime figures aren’t broken out by town. Their pages show police spending only.
• CDOT revises crash records for several years after the fact, and crashes on private property aren’t in the statewide listing.

Open data

The published figures are free to download as CSV: one file for all Colorado municipalities (revenue, spending, tax and debt per resident, every year) and one per city with every metric on its page. Each row carries the fiscal year, unit, peer and statewide medians, and its Census sources. Values are raw numbers as stored: dollars per resident, and shares from 0 to 1. Withheld figures appear as rows with an empty value and the reason.

Free to reuse under CC BY 4.0 (opens in a new tab). Please credit “Source: Orivian Municipal Index, orivian.ai/benchmarks”. See the data use terms.

Suggested citation: Orivian. (2026). Colorado Municipal Index [Data set]. https://orivian.ai/benchmarks