Save before you spend: a capital-funding strategy for Middlesex

The position, in numbers

Every town owns a stock of expensive, long-lived assets that wear out on a predictable schedule: a fire engine roughly every 20 to 25 years, a heavy plow truck every 10 to 15, a building every 40 or more. The question is never whether they will need replacing, but how the town will pay when each one does. In Vermont, where the property tax is effectively the only local revenue lever, a town has three choices: raise the full cost in a single year (a tax spike), borrow and repay with interest, or set money aside in advance so the cash is waiting.

Middlesex has, historically, chosen to borrow. Its heavy equipment rides on rotating five-to-fifteen-year bank notes, and its most recent purchase — a Case CX145D excavator authorized by voters in March 2024 — was financed with a $182,000 Community Bank note at 5.97 percent (2025 Town Report, debt schedule; FY2025 audited financial statements, Note 15). The town has never used the state's subsidized equipment-loan program, which lends at 2 percent (Vermont Treasurer, Municipal Equipment & Vehicle Loan Fund).

The town is not without reserves. Its FY2025 audit records roughly $299,298 across five capital funds — an Asset & Equipment fund of $100,463, a Town Hall building fund of $93,729, and smaller bridge, garage and resurfacing funds — and the treasurer reports about $368,000 in capital money-market accounts. Middlesex adopted a formal capital-improvement process only recently and now budgets $126,000 a year into it — about 5.6 percent of its $2.26m FY2027 municipal budget.

That is not enough, and the town knows it. A budget guide estimates $3m–$4m of capital needs over the next five years — a new town garage, town hall repairs, and overdue vehicles — against roughly $300,000 in reserves. The capital plan already lists a 2018 dump truck and a 2014 loader as "Due 2025/2026." In November 2024, voters rejected a $2.5m town-hall renovation bond, 610 to 534.

Reserves cover about a tenth of the five-year need

Estimated $3–4m of capital needs against roughly $300,000 in reserves.

And then came the water. The floods of July 2023 and July 2024 forced roughly $9m of emergency road and bridge reconstruction, pushing the town's long-term debt from about $2.14m to $7.16m in a single year and consuming the $200,000–$400,000 unassigned "rainy-day" balance the town had historically carried (FY2025 audit, Note 15; VTDigger, 26 Feb 2025). The town bridged its delayed FEMA reimbursements with Bond Bank climate-recovery loans priced as low as 0.5 to 1.3 percent — a reminder that when the state lends, it lends cheaply.

One year of floods more than tripled the town's debt

Long-term debt, before and after the July 2023 & 2024 emergency road and bridge work (+$5.02m, about 3.3×).

The selectboard has already signaled the direction of travel. At the March 2024 town meeting its chair put it plainly: "in the past interest rates were low and it was easier to borrow for our equipment; now interest rates are higher. The plan now is to use the [Asset & Equipment] fund to bring down debt." The question is not whether to change course, but how far and how fast.

What the standard-setters recommend

The recognized authority on municipal finance is the Government Finance Officers Association, a 30,000-member professional body whose guidance functions as the sector's reference. Its 2019 advisory, Strategies for Establishing Capital Asset Renewal and Replacement Reserve Policies, recommends that governments adopt a written policy and accumulate the cost of an asset over its useful life — illustrated by "annually setting aside 20 percent of a five-year asset's replacement costs so that funding is available when replacement is necessary" (GFOA, Renewal & Replacement Reserve Policies).

Crucially, GFOA does not rule debt out entirely, and an honest strategy for Middlesex should not pretend otherwise. But its framework is built on the useful life of the asset: short-lived, recurring items — "fleet, technology and smaller facility needs" — belong in cash and reserves; only long-lived, lumpy assets such as buildings are candidates for borrowing, and even then the debt term must never exceed the asset's life. Borrowing is never free: interest is a real, permanent charge on the tax rate, and for a depreciating machine it buys nothing the town could not buy more cheaply by saving first.

The practical upshot is a two-tier rule: pre-fund the rolling stock, and fall back on debt only for the buildings too large to save for in time. That is precisely the pattern the region's towns have settled into.

What the neighbors actually do

A review of twenty comparable Vermont towns — rural municipalities of roughly 1,000 to 3,600 people, each researched from its own town reports, budgets and audits and independently fact-checked — shows that full pre-funding is neither rare nor exotic, and that the two-tier rule is close to universal.

What towns vote into capital reserves each year

Share of the municipal budget, 20 comparable Vermont towns. Middlesex sits in the bottom third; the disciplined pre-funders cluster at 10–15%.

Six of the twenty fully pre-fund their major equipment, and several are smaller or poorer than Middlesex. Arlington (population 2,457, budget $2.05m — nearly a statistical twin of Middlesex) votes roughly $241,500 into nine separate per-asset reserve funds every March, and bought a $459,711 fire engine outright in cash, carrying essentially no debt. Tunbridge (population 1,337) holds about $1.23m in reserves — some 55 percent of its annual budget — on a smaller grand list than Middlesex's; asked whether the town ever needs a loan to buy a dump truck, its treasurer answered: "no… We are a debt free town." Guilford puts a flat $225,000 a year into a capital fund, bought a $298,000 grader in cash, and its FY2025 audit states flatly that "the Town had no long-term debt."

Vernon shows the endgame. Two pooled funds built since 1990 and 2000 hold roughly $4.67m; the town has zero long-term debt. In FY2021–22, investment income and returned funds covered all but $15,155 of a $224,351 capital requirement, leaving almost nothing to be raised from taxes.

Where the flywheel ends up: Vernon's reserves nearly paid for themselves

Of a $224,351 capital requirement, investment income and returned funds covered all but $15,155 — leaving almost nothing to raise from taxes.

Two features of the survey bear directly on Middlesex's choices. First, the two-tier norm holds everywhere: even the model pre-funders bond their buildings.

The building exception. In November 2024, East Montpelier passed a $4.95m town-garage bond, 880 to 712 — while paying cash for every truck. That reframes Middlesex's rejected $2.5m town-hall bond: borrowing for a building is the normal tool, not a failure of discipline.

Second, the going rate for equipment pre-funding is 10 to 15 percent of the municipal budget — and Middlesex, at 5.6 percent, sits in the bottom third of its peers, above only the pure borrowers. The full spectrum, ranked by the share of the municipal budget each town votes into capital reserves each year (each figure links to the town's own report, so it can be checked at source; shares are approximate — see the report's underlying notes):

Shares are rounded and approximate: towns differ in fiscal year and in which funds they count as capital reserves. * towns vote a high share but much of it repays 2 percent state loans rather than accumulating as cash; † towns sit on unusually large grand lists; ‡ marks the least-certain figures. Berlin is omitted (it funds capital partly from a 1 percent local-option tax; share n/a).

The table makes the affordability objection hard to sustain. The towns that fully pre-fund at 12–13 percent — Tunbridge, Guilford, Arlington, Duxbury — are no richer than Middlesex; grand-list value per resident in Tunbridge, Duxbury and Arlington is at or below Middlesex's. They simply set the annual line higher and defended it every year.

The economics: when saving beats borrowing — and when it does not

An honest case must concede that pre-funding is not always the cheaper path. The comparison turns on three rates. Today, a Vermont town can earn roughly 3.5 to 4.4 percent on reserves in money-market accounts and CDs; it can borrow from the Vermont Bond Bank at roughly 3.5 to 4.5 percent; commercial bank equipment notes run 5 to 6 percent — Middlesex's own 5.97 percent excavator being the cautionary case; and the state's Municipal Equipment & Vehicle Loan Fund lends at just 2 percent, or 0 percent for equipment two or more towns buy jointly, capped at the lesser of $150,000 or 75 percent of the price (29 V.S.A. ch. 61 §1602).

Middlesex pays 5.97% for equipment the state lends at 2%

Cost of borrowing by channel. The dashed line marks what reserves themselves can earn (~4%).

Three conclusions follow. Against commercial bank notes, the reserve wins outright — a town earning 4 percent instead of paying 6 swings roughly ten points of spread on every dollar of equipment, and Middlesex is on the wrong side of it. Against subsidized state credit, borrowing can be the smarter play — when the Loan Fund lends at 2 percent while reserves earn 4, borrowing is arbitrage in the town's favor; but the Fund's $150,000 cap now covers under a third of a modern fire engine. And inflation punishes waiting — fire-apparatus prices have roughly doubled since 2020, from about $500,000 to $1m, with delivery backlogs now running two to four-and-a-half years. A town that defers replacement is not saving money; it is paying tomorrow's inflated price after running an unreliable asset for years.

The synthesis is not "never borrow." It is the hybrid the academic literature also favors: keep a reserve large enough to buy routine equipment in cash and to cover the down-payment and uncovered balance on the biggest items; borrow the rest through the cheapest available channel; and let the reserve earn interest between purchases.

What Middlesex should do

1. Reduce interest costs

Stop borrowing at commercial-bank rates for equipment. The single cheapest change the town can make is to route any equipment it still must finance through the state's 2 percent Loan Fund rather than a 5-to-6 percent bank note. Where a shared machine is practical, a joint purchase with a neighboring town carries 0 percent interest — Danville and Ryegate already co-own a roadside mower on exactly this logic. Buy routine rolling stock in cash as the reserve matures, eliminating interest entirely — the Tunbridge, Arlington and Guilford model. For buildings, use the Vermont Bond Bank, not a commercial lender: its pooled loans match the term to the asset's life and price well below commercial rates. Middlesex's own books prove the point — its Bond Bank climate-recovery loans carry rates of 0.5 to 3.5 percent, against 4.95 percent on a commercial line of credit.

2. Anticipate and smooth tax increases

Adopt a full cost-over-useful-life capital plan and publish it. The mechanism is arithmetic: list every vehicle, building and major machine with its replacement cost and expected life, divide one by the other, and set that as the annual contribution — the GFOA method, as East Montpelier and Vernon apply it:

Publishing the schedule in the town report — as Duxbury does with its rolling sixteen-year plan — means voters see each purchase coming years ahead, converting surprise into routine.

Raise the capital contribution from 5.6 percent toward the 10-to-15 percent peer norm — but phase it, and fund much of it from money the town already spends. Reaching the benchmark means lifting the annual contribution from $126,000 to roughly $230,000–$340,000. That gap need not fall entirely on the rate: Middlesex already spends about $200,000 a year — 9 percent of the budget — on debt service, and several notes retire over the next few years (the Kenworth in 2027, the fire-station bond in 2029). As each note is paid off, its debt-service line can be redirected into reserves without any new tax at all. At Middlesex's grand list, one cent on the municipal tax rate raises about $24,000, so the residual gap is a few cents, not a cliff.

The path to the peer norm — without a new tax

As notes retire (Kenworth 2027, fire-station bond 2029), ~$200k/yr of debt service can redirect into reserves. Shaded band = the 10–15% peer norm ($230k–$340k).

Use "the reserve pays the loan" as the on-ramp. During the transition, while old notes are still outstanding, Middlesex can do what Starksboro does openly — "Municipal loans are used to purchase large items and the reserve fund is used to make the loan payments." Berlin executed exactly this pivot, retiring its equipment notes early and then buying with cash. The tax line stays level throughout; only the destination of the money changes, from interest to savings. (For the biggest items, a reserve-funded down payment against a low-cost loan is the standard structure — Readfield, Maine bought a $623,000 pumper with 44 percent reserves and a 56 percent ten-year bond.)

3. Build resilience to disasters and surprise expenses

Rebuild the unassigned reserve the floods consumed, and give it a written policy. The town's historical $200,000–$400,000 rainy-day balance is gone. GFOA's guidance is to define a minimum balance and a replenishment rule for when it is drawn down.

Create a dedicated disaster reserve — the single strongest lesson of 2023–24. Tunbridge established a Disaster Preparedness fund by town vote in 2020; when its own flood hit, the pre-funded reserve absorbed the cost without borrowing. A Middlesex disaster fund need not be large; its purpose is to cover the cash-flow gap between spending and FEMA reimbursement — the very gap that forced $7m of emergency borrowing.

Protect the reserves from raids. A Vermont reserve fund can, under 24 V.S.A. §2804, be redirected "for other purposes" by a simple majority at any warned meeting — and the risk is real:

Why reserves need armor: Moretown's fell 78% in two years

Under 24 V.S.A. §2804 a reserve can be redirected by simple majority at any warned meeting. Moretown's capital reserve fell from $434,000 to $97,000.

The armor is governance: hold many narrow, single-purpose funds rather than one large pot — Arlington runs fifteen — and re-present the replacement schedule to voters every year so contributions are defended in the open. New Hampshire towns go further, placing reserves with independent elected trustees and requiring a two-thirds vote to repurpose a fund; Francestown, a town of 1,610, used a fire-truck reserve first opened in 1955 to buy an $875,000 engine for cash in 2025.

The bottom line

Middlesex does not need to invent anything. The legal machinery already exists in 24 V.S.A. §2804; the method is published by GFOA; the cheaper credit is sitting unused at the state Treasurer's office; and the template is on display in towns a short drive away, several of them smaller and less wealthy. The town has already built a capital plan and stated its intent to shift from borrowing toward saving.

What remains is a decision — to set each asset's annual contribution at cost divided by useful life, to raise the capital line toward the peer norm as old debt retires, to route unavoidable borrowing through the state's low-cost channels, and to hold a written, raid-resistant disaster reserve against the next flood. None of it is free, but the alternative — paying nearly 6 percent on equipment, absorbing tax spikes when assets fail, and borrowing $7m in an emergency because the rainy-day fund was empty — is considerably more expensive. The constraint, as the survey of twenty towns makes clear, is not feasibility. It is the decision to begin.

All figures are drawn from the towns' own published reports, budgets and audits; the underlying research is a verified survey of twenty comparable Vermont towns with per-town sourcing. Principal standards and mechanisms: GFOA renewal & replacement reserve policies; 29 V.S.A. ch. 61 (state equipment loan fund); 24 V.S.A. §2804 (reserve funds); Middlesex FY2025 audit.