By Design
Utility Bills Household for the Fifth It Stopped Using
Portland General Electric's residential customers used 20% less electricity in 2024 than in 2000 and paid 141% more for it. The appliances worked. The savings were collected by the unit price.
By Hollis Carver Domestic Affairs · By Design 2026-08-15
You replaced the bulbs. You replaced the fridge when it went, and the new one draws less than half what the old one did. Perhaps you went further — a heat pump, a smarter thermostat, the whole payback spreadsheet. Then the bill arrived and it was larger than it had ever been, and you concluded that you must have done the sums wrong, or that you are somehow still wasteful.
You did not do the sums wrong. Every appliance in that spreadsheet performed as promised. The savings were real, they were delivered, and they were collected by someone else on the way to you.
The object
The object here is not a lamp or a compressor. It is the payback calculation — the small table on the box, or in the rebate brochure, that says this device costs more today and returns the difference in four years of lower bills.
That table has two variables. One of them is how much energy the device uses, which the manufacturer controls and generally states accurately. The other is what a kilowatt-hour costs, which the manufacturer does not control, cannot promise, and assumes will hold still.
It does not hold still.
What actually happened
Oregon PUC utility statistics for Portland General Electric’s residential customers, 2000 against 2024, with BLS CPI-U for the inflation line:
| 2000 | 2024 | Change | |
|---|---|---|---|
| Average monthly bill | $58.50 | $141.25 | +141% |
| Average consumption | 972 kWh/mo | 777 kWh/mo | −20% |
| Price per kilowatt-hour | 6.02¢ | 18.19¢ | +202% |
| General inflation (CPI-U) | — | — | +82% |
Read the middle row first, because it is the one that gets skipped. Households did the thing. Consumption per customer fell by a fifth over twenty-four years, across a period when the number of powered objects in an average home went up, not down. That is the efficiency programme working.
Now read the third row. The price of each unit tripled. Multiply it out and the rows agree with each other: 18.19¢ × 777 kWh is $141.34, within a rounding error of the stated bill, and 6.02¢ × 972 kWh is $58.51 against a stated $58.50. These are not three numbers that happen to sit near one another. They are the same number told three ways.
Against inflation, the 2000 bill of $58.50 should have become $106.57. It became $141.25. The difference is $34.68 a month, or about $416 a year, above and beyond what general price rises would have done to it.
Why it fails by design
A regulated utility does not earn by selling kilowatt-hours. It earns an allowed return on capital it has prudently invested — the rate base. Its revenue requirement is set to recover costs plus that return, and rates are then calculated to collect the requirement from however much electricity customers turn out to use.
Look at what that machinery does to conservation. If every household cuts usage by a fifth, the revenue requirement does not fall by a fifth; most of it is fixed plant, and the return on it is owed either way. So the price per unit is adjusted upward until the requirement is met. The efficiency is real, the kilowatt-hours genuinely go unburned, and the bill lands where the bill was always going to land.
This is not an accusation of bad faith. It is the design, working. The system was built to guarantee recovery of prudent investment, and it does that reliably. It was not built to let a household keep the savings it paid for, and nobody mentions this on the box.
The part where we do the division
It is customary at this point to produce the chief executive’s pay. Very well: Portland General Electric’s CEO total compensation went from $0.79M in 2000 to $7.37M in 2024, a ratio of 45:1 against a median employee’s $163,109, both figures on the SEC’s own total-compensation methodology.
Now divide, because a reader who does not will be misled, and a reader who does will catch us.
That $34.68 a month comes to $416 a year per household. Total CEO compensation of $7.37M is therefore equal to the entire above-inflation increase for about 17,700 households. Portland General Electric serves vastly more households than 17,700. Spread across the actual customer base, the chief executive’s entire pay packet accounts for a fraction of one percent of the gap on your bill.
Delete the salary altogether and you would not notice. The number is a symbol, and symbols are worth arguing about, but it is not the mechanism and printing it as though it were would be a lie of emphasis.
The mechanism is the rate base. Capital earns a return; more capital earns more return; the return is collected from customers whether they conserve or not. That is far less satisfying than a cigar and a top hat, and it is where the $416 actually lives.
What the money bought, stated fairly
The investment is not fictional, and a piece that pretended otherwise would deserve to be disbelieved:
- 475 MW / 1.9 GWh of battery storage energized in 2025, which the company says can cover roughly 300,000 homes for four hours and reduce reliance on expensive short-term power purchases.
- 1,015 MW of clean energy and storage contracted for 2027–28, about $1.03B of planned investment.
- 71 dynamic line-rating sensors across 11 transmission lines, plus AI-assisted planning the company credits with finding 80+ MW of near-term grid capacity.
- Wildfire mitigation exceeding $85M in 2024, with roughly $120M forecast for 2026.
Batteries are real hardware. Wildfire hardening is real work, and the alternative to paying for it is not “cheaper power”, it is a different and worse bill. In the 2025 rate case, only 2.5 of the 5.5 percentage points of residential increase were base-rate or capital related; the rest was power costs and mandated adjustments.
The honest question is not whether anything was built. It is who pays, and on what terms, and whether a household that reduced its own demand should carry the same share as one that did not. And it is worth noting that 2025 load grew 4.7%, driven partly by data centres and high-tech — which is to say the grid is being enlarged for demand that is not yours, on a bill that is.
What should exist
- Payback tables with a rate assumption printed on them. “Four-year payback” is not a claim about a device. It is a claim about a device and a tariff, and only one of those is in the box.
- Decoupling done in both directions. Several states break the link between utility revenue and volume sold, so conservation does not mechanically trigger a rate increase. Where that exists, the household keeps more of what it saved.
- A conservation dividend that survives the next rate case. If reduced demand defers capital spending, the deferral is a real saving and should appear on the bills that produced it, rather than being absorbed into the requirement and redistributed.
The fix, which is not a fix
There isn’t a three-dollar part for this one. Conservation still lowers your bill relative to not conserving — the household that did nothing since 2000 is paying far more than $141.25 — and that remains the correct thing to do. The appliances work.
What does not work is the promise implied on the box: that the savings are yours to keep. They are yours to keep only until the next rate case, at which point they become a smaller denominator.
Buy the efficient thing. Do it because it uses less. Do not buy it on the strength of a four-year payback calculated at today’s price of a kilowatt-hour, because that price is not a constant, is not yours, and has tripled once already while you were doing everything right.
Portland General Electric serves Oregon and is not Pacific Gas & Electric, with which it shares an unhelpful abbreviation. Figures for 2000–2024 are actuals from Oregon PUC utility statistics and BLS CPI-U; executive compensation is from PGE’s SEC proxy filings. Figures for 2025–2027 appearing in those filings are proposals or projections subject to regulatory approval, and are not used above as though they were settled.
