Commonwealth · Method

How the numbers are built.

Every figure in Commonwealth is built by script from public records, and every method is published here so anyone can check it. Dollars carry their stage: enacted for budget earmarks, paid for Comptroller payments.

Data through FY2026; built October 19, 2026.

The Fairness Index

No single definition of a fair share is right, so the index uses four and shows each one.

Equal. Every resident counts the same. A town’s expected share is its share of the state’s population.
Need. Each resident is weighted from 1 in the least needy towns to 3 in the neediest, using the town’s standing on poverty, median household income and tax base (equalized valuation per resident).
Burden. Each program’s money is split by what that program pays for: road miles for road programs, pupils for school aid, road-stream crossings for culvert grants, town-owned hazard dams for dam repair, and so on. Where a town has no figure for a measure, that program is left out of its burden score rather than guessed.
Effort. Each resident is weighted by how much the town already raises from its own residents: residential and open-space property tax (DLS assessed values and tax rates, averaged over FY2021–FY2025), less the share of homes that are seasonal, divided by residents’ income (ACS per-capita income times population). The rate is capped at the 1st and 99th percentiles.

For each lens, the town’s actual dollars are divided by what that lens would give it from the same pool of money. A ratio of 1 scores 50; half the expected amount scores 25 and twice the expected amount scores 75. To keep a very small town’s score from resting on a tiny number, every ratio is pulled slightly toward 1 by adding, to both sides, what a 250-resident town would get at the statewide rate. The headline score is the average of the four lens scores.

Chapter 70 school aid is 61% of all state aid in the index ($57.10 billion over ten years), so on all aid the four lenses mostly reflect how school aid is distributed. The discretionary score has no Chapter 70 in it.

The blend tracks the Equal lens closely. Across the 351 towns, the rank correlation between the blended score and the Equal (per resident) score is 0.98 on all aid and 0.996 on discretionary money, so a town’s blended rank is close to its per-resident rank. That is why every town page, and the table on the home page, shows the four lens scores beside the blend.

The Effort lens is shown as built. In practice the effort rate rises with median household income (rank correlation 0.24) and with school spending above the required minimum (0.37), and falls with poverty (−0.26). A town’s Effort score therefore reflects its spending choices and tax base as well as how heavily its residents are taxed.

The sensitivity range repeats the calculation under 231 weightings, with each lens between 10% and 40% in 5-point steps. A town is marked below 45 or above 55 under every weighting when its whole range sits there, between 45 and 55 when its whole range stays inside that band, and weighting-dependent when the range crosses into or out of the band.

All state aid means formula aid, competitive grants and enacted earmark clauses that name the town (the first of three measures), paid to the town government and its school districts. For Chapter 70 the index uses DESE’s gross aid by town, and for unrestricted aid, state-owned land, exemptions, libraries and veterans benefits it uses the cherry-sheet amounts, because the Comptroller records those payments after the state deducts charges such as charter tuition and transit assessments. Other formula programs and all competitive grants are Comptroller payments (paid). Charter and school-choice tuition are left out so they are not counted twice. Discretionary money means competitive grants and those earmarks. Because the expected amounts add up to the actual pool, the average town does not score exactly 50: a few large towns above their share are offset by many small towns below it, or the reverse.

The published scores were computed on October 18, 2026.

State money by class

Payments come from CTHRU, the Comptroller’s statewide spending data: grants and aid paid from FY2017 through FY2026, by appropriation and payee town. Each appropriation is classed as formula, competitive, earmark, shared revenue, transfer to a state entity, pass-through contract or unknown, by a written rule with the evidence recorded. About 81% of the dollars can be placed in a town. Regional school district payments are split among member towns by enrollment. Town pages show competitive grants and earmark-account payments to the town government and its school districts, the same payees as the Fairness Index; the other classes except formula are totaled in a note. Per-resident figures divide the ten-year total by residents in the American Community Survey 2020–2024.

For cherry-sheet programs, including Chapter 70 and unrestricted general government aid, CTHRU records the cash paid after the state deducts charges such as charter and school-choice tuition and transit assessments. Net figures would make charter tuition and transit assessments look like aid withheld; for Boston, ten-year Chapter 70 is $74 million net against $2.26 billion gross. Town pages and the Fairness Index therefore use gross formula aid: Chapter 70 as DESE attributes it to each town, cherry-sheet amounts for unrestricted aid, state-owned land, exemptions, libraries and veterans benefits, the special education circuit breaker from DESE, and other formula programs as paid.

CTHRU also records payments from accounts classified as earmarks. Those accounts can carry prior-year money and statewide-program money, and many enacted earmarks are paid through larger accounts, so the paid figure can be larger or smaller than the enacted clauses that name a town. Town pages show both and label each; the earmark section counts enacted budget text.

Three ways to count a local earmark

A local earmark is a “not less than” clause in an enacted budget that sends money to a particular place. This project looks at fairness by place, and there are three ways to decide which clauses belong to a town. They give different totals, so the site reports all three and labels every earmark figure with its measure.

Names the town, $485 million over ten years, FY2017–FY2026. Clauses that direct money to the city or town by name. This is the money closest to the municipal budget, which is why the Fairness Index uses it. District money on legislator pages uses it too.
Placed in the town, $718 million. Adds $268 million for organizations located in the town, placed by their address. State facilities and rate add-ons that name a town, $34.4 million, count on the third measure instead. The expected-and-actual model, the peer comparison and the list of largest items on town pages use this measure.
Everything placed, $811 million. Adds $93.5 million for state facilities located in the town (state hospitals such as the Pappas Rehabilitation Hospital for Children in Canton, state universities and community colleges, courts, sheriffs, state parks and beaches, the MBTA, state police) and MassHealth rate add-ons for providers there. The largest is the $31 million FY2026 Pappas clause.

All three are small beside the $5.44 billion in enacted “not less than” clauses over the ten budgets; most of that sets floors for statewide programs. The Losers Club, 35 municipalities, had nothing placed in them on any measure.

Earmarks

An earmark is a “not less than $X shall be expended” clause in Section 2 of an enacted General Appropriations Act. The project parsed all ten enacted budgets, FY2017 through FY2026, from the Legislature’s website: 10,040 clauses. Vetoed text that was not overridden is dropped, so every dollar shown is enacted.

A clause is placed in a town when it names the town (“city of”, “town of”, or the bare name with a false-positive filter) or when its recipient organization is based there. Clauses naming several towns are split equally. Three in four clauses can be placed in a town, but only 15% of dollars, because the largest clauses fund statewide programs. A hand check of 50 random clauses found 46 placed correctly; the other four were fixed.

Clauses that fund state facilities and MassHealth rate add-ons are identified by a written text rule (the recipient named at the head of the clause is a state hospital, university or college, court, sheriff, state park or beach, the MBTA or the state police; or the clause sets a provider rate). They count on the third measure, everything placed, and each town page shows them with their amounts.

Earmarks are counted where the money lands, not by who sponsored them. The Losers Club is the project’s name for the municipalities with no placed earmark in any of the ten budgets.

What the town carries

Burden measures come from the agencies that keep them: local road miles from MassDOT’s Chapter 90 apportionment and Road Inventory; bridges from MassDOT and the FHWA National Bridge Inventory; road-stream crossings from the North Atlantic Aquatic Connectivity Collaborative; dams from the DCR Office of Dam Safety inventory; land and state-owned land from MassGIS; seasonal housing from the American Community Survey.

Bridges are scored, not counted. Only structures the town owns count. Each structure’s replacement cost is its deck area times the FHWA 2025 Massachusetts unit cost ($981 per square foot, or $1,306 on the National Highway System). The burden score weights that cost by condition, age and use; those weights are judgment and are stated so they can be argued with. The replacement cost does not depend on them.

Condition, by the lowest of the deck, superstructure, substructure or culvert ratings: poor (0–4) counts 100% of replacement cost; fair (5) 50%; fair (6) 25%; good (7–9) 10%; not rated 50%.
Age: from 0.8 for a new or rebuilt structure to 1.2 at 100 years or more.
Use: from 0.8 to 1.2, by traffic, truck traffic, detour length and road class relative to all Massachusetts structures.
A structure on a town line is shared equally by the towns it touches. State- and federally owned bridges, including the Cape Cod Canal bridges, never count against a town.

The burden index takes each competitive program in the award catalog, matches it to the measure it pays for (road miles for Complete Streets, crossings for culvert grants, hazard dams for dam repair, shoreline for seawalls, floodplain acres for resilience grants, housing units for Green Communities, jobs for MassWorks, cost-burdened households for housing, residents for libraries) and asks what each town would have received if the program had split its money in proportion. Actual ÷ that amount is the index; 1.0 is proportional. A second set of plausible denominators leaves the ranking close (rank correlation 0.94). Sixty-seven western towns have no digital FEMA flood map; their flood grants are left out of their index rather than guessed.

State land and growth. PILOT-eligible acres, their FY2026 valuation and the FY2026 payment in lieu of taxes come from the Division of Local Services state-owned land reports; the payment follows assessed value, so PILOT per acre (statewide $117 in FY2026) is a comparison, not the formula. The share of land area divides eligible acres by the town’s land area. New growth is the assessed value of new construction and improvements DLS certifies each year, summed FY2017–FY2025 and divided by residents (ACS 2020–2024).

Costs and the 2½% limit

Each town page carries a panel on state aid, fixed costs and the 2½% limit. It uses the same Division of Local Services build as section 8, “The squeeze,” of the Commonwealth report (October 19, 2026), which gives the statewide figures and the full method; the panel shows one town’s line of those tables.

Share of spending. Gross cherry-sheet state aid as entered on the tax rate recap, divided by the town’s general-fund expenditures as filed (Schedule A), for FY2006, FY2016 and FY2025. Net subtracts the charges the state assesses on the same cherry sheet (charter and school-choice tuition, transit, county and other assessments), itemized from FY2010. The median town’s share was 14.1%, 12.2% and 11.5%. A regional member’s cherry sheet carries no Chapter 70 for pupils in the regional district, while its assessment to the district is in its spending, so its share understates state support for schools; the panel says so and gives the Chapter 70 paid on its behalf.
Against prices. FY2006 aid grown by the Boston-area CPI-U (up 55.1%) and by the federal price index for state and local government purchases (up 74.5%), fiscal-year averages, set against FY2025 aid. In FY2010 gross aid exceeded the itemized cherry sheet by $313 million statewide, consistent with school-building reimbursements that ended by FY2024; 170 town pages carry a flag where that gap was more than 2% of the town’s FY2010 aid.
Fixed costs. Health insurance and all general-fund fixed costs, FY2016 to FY2025, against 2.5% a year (24.9% over nine years). A year is squeezed when fixed costs grew by a larger percentage than the town’s allowed levy growth (2.5% plus new growth, as a share of the prior year’s levy limit). A one-year change beyond 30% in either line is flagged on the page: such swings usually mark costs moving between budget lines, into or out of trust funds, or into the Group Insurance Commission.
Earmarks. Enacted local earmarks in the nine budgets FY2017–FY2025, on the three measures, set against the rise in the town’s health insurance. Pooled over the 342 towns that reported health insurance in both years: 28.2%, 42.3% and 45.2%.
Limits. Spending stops at FY2025; most towns had not filed FY2026 by October 19, 2026. Where the source has no figure, the page prints “not reported.” Dollars are nominal except in the price comparison. The figures are associations and do not measure cause.

Schools

Chapter 70 aid, foundation budgets, required local contributions and enrollment come from DESE’s Chapter 70 files, FY2017–FY2026. For towns in regional districts, aid and contribution are DESE’s attribution to the member town; how each region splits its assessment is set by its own agreement and is not published statewide. DESE town numbers are not the Department of Revenue’s town codes for 33 towns, so every DESE file is joined by town name through a checked crosswalk.

The required contribution’s share of household income divides the contribution per household by the town’s median household income (ACS 2020–2024). Real changes use the CPI-U fiscal-year average in FY2026 dollars. Real aid per pupil compares FY2017 and FY2026 attributed aid per resident pupil.

Expected and actual

A statewide model predicts each town’s ten-year discretionary dollars (enacted earmarks plus competitive grant payments) from factors a fair allocation could reasonably follow: population, road miles, jobs, poverty, income, tax base, Chapter 70 share, population band, the presence of hospitals, colleges and nonprofits, county seat and Gateway City status, with gross formula aid per resident as a control. These factors explain about 55–60% of the differences between towns. Earmarks in the model are placed in the town, the second measure; state facilities and MassHealth rate add-ons, on the third measure only, are not in it.

A town above or below its expectation is not evidence of anything improper. A single large project, a regional facility or an organization headquartered in town can put a town far from the line for ordinary reasons.

Peers

Each town’s ten peers are its nearest neighbors on standardized log population, the need score, income need and tax-base need. The wider class is the 30 nearest; the best and lowest performers in it are ranked on earmark dollars placed in the town per resident, the second measure. This method weights need heavily, which is why Boston’s peers should be read loosely.

Legislator pages

Each person who served in the House or Senate in the 189th through 194th General Courts, which enacted the FY2017–FY2026 budgets, has a page. A member who served in both chambers has one page. The pages carry facts only: seats, posts and stipends, the district’s money, sponsored budget amendments, bills, recorded votes and campaign receipts. No score or grade is computed, and associations are not evidence of cause.

Seats. From the Legislature’s member records; 2011 district lines through the 192nd General Court (the FY2023 budget), 2021 lines after.
Stipends. Additional compensation in the Comptroller’s payroll data (CTHRU 9ttk-7vz6), by calendar year, 2016–2026.
District money. Enacted earmarks in clauses that name a place in the district (the first of the three measures), by session, per resident per budget, ranked among the chamber’s districts. This is where money lands, not money attributed to the member, and leadership’s base-text money that names no place is never assigned to anyone. Clauses carried by a Ways and Means chair’s corrective amendments are counted where they name a place in the district; each page shows those dollars in their own column, with the per-resident figure and rank without them beside the main one. Each page also shows the district on the other two measures, placed in the district and everything placed, with the same per-resident ranking; those include any corrective-clause money.
Amendments. Every amendment filed to the House and Senate budgets, FY2017–FY2026. The adoption rate is set beside the chamber’s rate for amendments of the same ask size, because small asks pass far more often than large ones. House outcomes are mostly inferred by matching amendment text to the consolidated amendments; the page says how many. Enacted items are listed when an enacted clause matches by text an amendment the member filed. Corrective and technical amendments filed in a Ways and Means chair’s name carry leadership’s money; they are shown on a separate line and left out of the chair’s own counts, asks, adoption rate and items.
Votes. Selected roll calls on the Legislature’s own rules, audits, public records, term limits and pay, with the project’s plain description of each question (for rules-package roll calls, the question as recorded). The headline counts yes and no votes; not voting and present are shown apart. Rules-package adoptions and public-records passage votes are marked “contested meaning” because reform advocates were on both sides.
Campaign money. Itemized receipts of $250 or more reported to the Office of Campaign and Political Finance, grouped by the contributor’s employer or organization. Contributions are receipts from individuals, committees and PACs, and unions and associations; candidate loans, transfers, refunds, interest and aggregated small receipts are reported on a separate line. Employer fields that are a job title or a bookkeeping note are counted as not reported. No individual contributor is named.

Limits

Earmark dollars are enacted text, not payments. A held or unspent earmark counts in full.
Payments are placed by payee town. A payment to a regional body or a statewide organization is placed where the payee is, or not placed at all.
Population is a single ACS 2020–2024 value for all ten years.
Representation shows current members and 2021 district lines. Earlier years used the 2011 lines.
Where a value is missing, the page says “not available.” Nothing missing is shown as zero.