The Wires, Not the Electrons: Where Your Electric Bill Actually Went

If you live in New England, you've watched it happen on your own bill: the "delivery" line — the part that pays for poles, wires and the utility itself, not the electricity — has grown relentlessly, even in years when the electricity itself got cheaper. I wanted to know why, so I built an analysis pipeline over ten years of federal data: every US utility's revenue, sales and customer counts from EIA-861 filings, state and national prices from the EIA open-data API, hourly demand from EIA-930, generator additions and retirements from EIA-860M, and interconnection-queue data from Lawrence Berkeley National Laboratory.^[1–5] Everything computed below is reproducible from those public files.

Here is what the data says — and it is not the story usually told.

First, the national picture

US residential electricity went from 12.55 to 17.30 cents per kilowatt-hour between 2016 and 2025 — a 37.8% increase, with 87% of it arriving after 2020.^[1] (All prices in this article are nominal; one Massachusetts legislator's real-terms estimate puts the state's delivery increase at +36% after inflation, still far above other household costs.^[13])

But the increase was not evenly split between the two halves of a bill. In restructured states, EIA data separates the wires charge from the energy charge, and the two diverged sharply between 2016 and 2024:

The wires got expensive twice as fast as the electrons. (Caveat: this split covers the ~7% of residential sales reported as separated service, concentrated in restructured states — a strong signal for those states, not a national decomposition.)

It is not because we're using more power

The most counterintuitive finding in the whole dataset: prices did not rise where demand grew. Across the seven organized US electricity markets, the correlation between 2016–2024 demand growth and price growth is −0.28 — negative. The three biggest price risers all shrank:

Texas grew by nearly a third — more than any other market — and had a below-median price rise. New England used 8–9% less electricity in 2024 than in 2016 and got the second-largest increase in the country. Whatever is driving these bills, it is not data centers or demand growth in these regions. (Seven data points can't prove causation; but the sign is the opposite of the popular story.)

New England, utility by utility

Because sales are falling, a per-kWh rate rises mechanically as the denominator shrinks. So the honest metric is delivery revenue per residential customer per year — immune to volume effects. From the utilities' own EIA-861 filings, 2016 → 2024:^[2]

Consumption per customer is flat to falling everywhere. This is a price story, not a usage story. The average New Hampshire Eversource household now pays roughly $500 a year more for delivery than in 2016 — about $84 a month for the wires alone before a single kilowatt-hour of energy is billed.

What did the money buy?

The federal reliability metric SAIDI counts the minutes per year an average customer spends without power (excluding major storm days). If doubled delivery charges bought a transformed grid, it would show here:^[2]

Five of six utilities improved — a real gain, typically 20 to 60 fewer outage minutes per year. But buying roughly a quarter less outage time for 40–100% more money is a poor trade, and Central Maine Power's customers paid 48% more for a grid that got worse. Storm-inclusive outage minutes, where "resiliency" spending should show up, have not trended down at all.

So if the money didn't primarily buy reliability, where did it go? The answer has three parts, and all three are documented.

Part 1: The rate-base machine

A regulated utility doesn't profit by selling electricity. It earns a regulator-authorized percentage return — historically around 9–10% — on the capital it has invested (the "rate base"). That means the way to grow profit is not to sell more or operate better; it is to build more. Economists Averch and Johnson described this incentive in 1962;^[16] a 2026 study in the Journal of Public Economics measured it: each extra percentage point of allowed return produces 3–4% more capital assets with no matching improvement in operating costs, at an estimated $7 billion a year in excess costs to US consumers.^[15]

Eversource's own SEC filings show the machine running:^[6]

The identical ratio is not a coincidence; it's the formula. Earnings track rate base one-for-one by construction. Annual capital spending rose from $1.98bn to $4.16bn, and the forward plan pre-announces the earnings growth to match. An investment-industry utilities outlook states the business model in one heading: "Record investment (rate base growth) leads to EPS growth."

Two facts cut against the crude "gouging" version, and they matter: state regulators have been granting less than utilities ask (Massachusetts cut NSTAR's requested return from 10.50% to 9.80% in 2022, and National Grid's to 9.35% in 2024, trimming its revenue request by 40%^[13]); and Eversource's earnings per share rose 61%, not 98%, because it issued new shares to fund the building spree. The problem is not the percentage. It is the ever-growing base the percentage is applied to — and the fact that ratepayers, not shareholders, carry it.

Part 2: Transmission — the part with no brake

Distribution rates at least go through adversarial rate cases. Transmission does not. New England transmission runs on FERC "formula rates": the utilities calculate their own revenue requirement annually, file it informationally, and collect it. Regional documents show where that leads:^[9–12]

• Over $9 billion was invested in New England transmission from 2016–2024, and 93% of currently planned transmission spending (~$6bn) is classified as "asset condition" work — replacement of existing equipment — a category that receives "effectively no oversight," in the words of a Connecticut state agency white paper.^[9]
• One regional analysis found that not a single asset-condition project has ever been denied or modified.^[9]
• In 2025–26, consumer advocates from five states alleged Eversource misclassified a $360.6M project (X-178) specifically to avoid the scrutiny that new projects get;^[10] Senator Blumenthal wrote FERC about the practice;^[12] ISO-NE is now creating an independent reviewer — starting 2027.
• ISO-New England's own external market monitor reports the region has the highest transmission costs of any US system operator, more than double the average of other RTOs.^[9]

Part 3: The policy programs hiding in your delivery charge

Here is the part most people don't know: a large and fast-growing share of the "delivery" line is not wires at all. It is state energy programs that the legislature chose to fund through electric bills rather than taxes.

National Grid's Massachusetts residential tariff — the actual filed rate document — shows the anatomy.^[14] Between August 2017 and May 2026, the stated total delivery rate went from 11.479 to 22.786 ¢/kWh (+98.5%). Inside it:

• Net metering recovery (paying solar owners' credits): 0.424 → 2.451 ¢ — +478%
• Residential assistance (discount-rate subsidy): 0.483 → 2.389 ¢ — +395%
• SMART solar program: 0 → 0.935 ¢ (didn't exist in 2017)
• Energy efficiency (Mass Save): 2.083 → 2.713 ¢
• Plus new lines for grid modernization, smart meters, EV programs, winter relief…

Explicitly named policy programs grew from roughly a quarter of the itemized rate to about 44%. The state's own Department of Public Utilities, opening a formal investigation of delivery charges in December 2025 (D.P.U. 25-200), found that public-policy charges grew more than any other component of the bill over the past decade.^[13],[19]

This also resolves the paradox that bothers everyone: if solar is so cheap now, why doesn't my bill fall? Three reasons, all in the data. First, New England's grid physically runs on gas — 57% gas, 4% solar in 2024 — so cheap Texas and California sunshine never reaches these wires.^[1] Second, the early cost of the transition was locked in at early prices: Massachusetts' SMART program paid 2018-vintage rooftop solar 31–39 ¢/kWh on ten-year contracts — two to three times the retail price — recovered through everyone's delivery charge.^[13] Third, revenue decoupling (Massachusetts since 2008) guarantees the utility its approved revenue regardless of sales; when customers conserve or go solar and sales fall, rates are trued up the next year to compensate.^[16] Conservation is good policy — but under decoupling it cannot, by design, lower the utility's take.

The honest verdict

Is it greed, waste, or corruption? The record supports something more specific than any of those words:

• A structural incentive to over-build, working as designed. Profits doubled because rate base doubled, and rate base doubled because everyone in the chain — utility, investors, even regulators approving individual projects — is rewarded for capital spending. The peer-reviewed estimate of the excess this creates is $7bn/yr nationally.^[15]
• A regulator-confirmed overcharge on transmission, the one segment with no real oversight: fifteen years of excessive returns, $1.5bn ordered refunded, 93% of ongoing spending still effectively unreviewed.^[7],[9]
• Policy costs quietly loaded onto the delivery line — approaching half of a Massachusetts delivery rate — which is a legislative choice about where to collect money, not utility pricing, but lands on the same line of the same bill.^[13],[14]

And what the evidence does not support: the version where a monopoly simply raised its price and pocketed it. Authorized returns went down. The shareholder's slice of a Massachusetts delivery bill is about 1.4–1.5 ¢/kWh — roughly 6% of the delivery rate.^[13] You could eliminate utility profit entirely and the delivery charge would still have nearly doubled. The money went into capital you now pay a return on for decades, into programs your legislature chose to fund this way, and — for transmission — into returns a federal regulator has ruled were too high all along.

One myth worth killing while we're here: the widely repeated claim that Connecticut cut Eversource's return as punishment for the Tropical Storm Isaias response is wrong as usually told — that cut was rescinded in a 2021 settlement. (The utility whose return actually was cut is United Illuminating.) The real, standing penalties are smaller: a $28.4M Isaias fine, ~$32M of storm costs disallowed, $15M returned over billing failures, and a Massachusetts AG settlement of up to $45M over irregular tree-work invoices.^[6],[13]

You are not imagining it — and you're not alone

Since late 2025 this has broken into the open: Boston.com asked why the delivery fee is so high,^[17] WBUR traced how Mass Save became a political fight,^[18] the Massachusetts DPU opened its investigation,^[19] the governor announced a $5.8bn ratepayer-relief plan, and think tanks from PowerLines to the Clean Air Task Force to Columbia's energy center have converged on the same conclusion nationally: the bill crisis is a delivery crisis, driven by a record wave of utility capital spending.^[20–23] Investor-owned utilities plan $1.4 trillion of capex through 2030.^[20] Under the current regulatory model, every dollar of it becomes rate base — and every dollar of rate base becomes a bill.

Every dollar of capital spending becomes rate base — and every dollar of rate base becomes a bill.

The fix isn't mysterious, and pieces of it are already moving: independent review of "asset condition" transmission projects (coming to New England in 2027), competitive bidding instead of formula rates, returns tied to performance instead of capital deployed, and an honest public debate about whether climate programs belong on electric bills or state budgets. What made this investigation possible is that the evidence was public all along — in EIA filings, SEC reports, tariff sheets and FERC dockets. It just needed to be joined together.

Sources & data

Computed findings (price trends, per-customer revenue, reliability, demand, generation mix) are derived from the public datasets [1–5]; the full pipeline, tests and methodology notes are available on request. External claims cite the documents below. Prices are nominal unless noted. Correlations are descriptive, not causal claims.

1. US Energy Information Administration, Open Data API v2, electricity/retail-sales and electric-power-operational-data — eia.gov/opendata
2. EIA Form 861 annual utility filings (sales, revenue, customers, reliability) — eia.gov/electricity/data/eia861
3. EIA-930 Hourly Grid Monitor (demand by balancing authority) — eia.gov/electricity/gridmonitor
4. Lawrence Berkeley National Laboratory, Queued Up 2026 edition (interconnection queues through 2025) — emp.lbl.gov/queues
5. LBNL, 2024 United States Data Center Energy Usage Report — eta-publications.lbl.gov
6. Eversource Energy SEC filings (10-K, 8-K), CIK 0000072741 — SEC EDGAR
7. FERC Opinion No. 594 (Docket EL11-66, 19 Mar 2026); coverage: Utility Dive, "Eversource, Avangrid ask FERC to stay $1.5B refund decision" and "FERC declines to stay $1.5B in refunds"
8. Hartford Business Journal, "Eversource cuts earnings guidance, vows legal fight after FERC slashes transmission returns"; Utility Dive, "New England transmission owners ask FERC for increased ROE"
9. NH Office of the Consumer Advocate, "The $500,000,000 Question: 'Asset Condition' Transmission Projects"; CT DEEP transmission oversight white paper (Feb 2025); NESCOE asset-condition statements — nescoe.com; Maine Office of the Public Advocate, joint advocates' statement
10. Utility Dive, "Eversource misclassified $385M transmission project to avoid scrutiny: ratepayer complaint"
11. Utility Dive, "Eversource, others may be capitalizing on lax reviews for some transmission projects: Maine officials"
12. Sen. Richard Blumenthal, letter to FERC on asset-condition projects, 8 Aug 2025
13. Massachusetts D.P.U. dockets 22-22 (NSTAR rate case, 2022), 23-150 (Mass. Electric rate case, 2024), 25-200 (investigation into delivery charges, opened 15 Dec 2025) — DPU File Room; real-terms delivery estimate from state Sen. William Brownsberger's published analysis
14. National Grid (Massachusetts Electric) residential R-1 delivery tariff summaries, Aug 2017 and May 2026 — nationalgridus.com rates
15. K. Dunkle Werner & S. Jarvis, "Rate of return regulation revisited," Journal of Public Economics 259 (2026); working-paper version: Energy Institute at Haas WP-329
16. H. Averch & L. Johnson, "Behavior of the Firm Under Regulatory Constraint," American Economic Review 52(5), 1962; on decoupling: Mass. D.P.U. 07-50-A (2008)
17. Boston.com, "Why is the delivery fee so high on electric and gas bills?" (Nov 2025)
18. WBUR, "How an energy efficiency program in Mass. became so politicized" (Jun 2026)
19. Greenfield Recorder, "Massachusetts investigates utility charges" (Dec 2025)
20. PowerLines, "Utilities are planning to spend $1.4 trillion on capital expenditures through 2030"
21. Inside Climate News, "Why electricity bills are so high" (Mar 2026)
22. Clean Air Task Force, "A data-driven look at rising U.S. electricity costs and policy solutions" (Mar 2026)
23. Columbia University Center on Global Energy Policy, "The Effects of Load Growth on Electricity Prices in the United States"