If every ad had to be 100% true, advertising would not die; it would split. Claims that can be checked (price, ingredients, test results, delivery time) would become far more valuable, because buyers could finally trust them. Claims that can't be checked (miracle cures, "100%" boasts, fake urgency) would vanish, and so would a large part of what India's ASCI flags today, where 56% of ads it examined in 2024-25 were misleading. The hard part is enforcement: the penalty for lying must exceed the profit from lying, which in one documented case was $105 million against a $4 million settlement.
- ASCI examined 7,199 ads in 2024-25: 98% needed changes and 56% were misleading.
- Kellogg's false claim earned an estimated $105 million in 32 months; the class-action fund was $4 million.
- Lying stops paying only when p × F > G: catch probability times fine beats the gain.
- Ski resorts reported 23% more fresh snow on weekends, until a phone app let skiers check.
- Emotion and puffery survive: "the best coffee in town" is not a claim anyone can test.
Every country already says ads must be true. India's Consumer Protection Act, 2019 bans misleading ads, ASCI's code since 1985 asks for ads to be honest and truthful, and the US Federal Trade Commission has policed deception for over a century. So the real question is: what if that rule were enforced perfectly, on every ad, every time?
From 1921, Listerine was advertised as helping to prevent colds and sore throats. In 1972 the FTC began reviewing the claim and found the mouthwash no more effective against colds than warm water. Its 1975 order told Warner-Lambert to stop the claim and to run a correction in its next $10 million of Listerine advertising, about a year's budget: Listerine will not help prevent colds or sore throats or lessen their severity.
The company fought it on free-speech grounds. A federal appeals court upheld the order in 1977, and in April 1978 the US Supreme Court declined to hear the case. A commissioner later noted that the false claim had run in a 51-year campaign, and that this was the only adjudicated FTC case in more than two decades where corrective ads were ordered.
TIME: Business: Taking It Back (17 April 1978)How much advertising is untrue today?
Nobody measures the untrue share of all ads, but India's self-regulator gives a sharp lower-bound picture: of 7,199 ads ASCI scrutinised in 2024-25, 98% required some modification and 56% were found misleading. These are flagged ads, not a random sample, so the true share across all advertising is lower; the point is where the problems cluster.
See the data as a table
| Share of cases (%) | |
|---|---|
| Offshore betting | 43 |
| Realty | 24.9 |
| Personal care | 5.7 |
| Healthcare | 5.23 |
| Food and beverage | 4.69 |
| All other sectors | 16.48 |
Source: ASCI Annual Complaints Report 2024-25 press release (28 May 2025) · All other sectors is the remainder to 100%. 94.4% of the ads processed were digital.
Three details from the same report matter for our thought experiment:
- 3,347 ads were for things the law does not allow to be advertised at all, including 3,081 for offshore betting and 233 that may breach the Drugs and Magic Remedies Act, 1954 (ASCI annual report, 2024-25).
- 59% of advertisers changed or withdrew the ad quietly when contacted, without a fight (ASCI annual report, 2024-25).
- In the first half of 2025-26 the load grew: 6,117 ads investigated, 98% needing changes, 97% of violations on digital platforms and 78.9% on Meta alone (ASCI half-yearly report, 2025-26).

Why do companies lie if they can get caught?
Companies lie when the expected gain beats the expected penalty. This is Gary Becker's economics of crime (1968) applied to ads: a rational firm makes a false claim if the gain \( G \) exceeds the probability of being caught \( p \) times the fine \( F \).
Worked case: a breakfast cereal
- The gain. Rao and Wang estimate that Kellogg's claim that Frosted Mini-Wheats was clinically shown to improve children's attentiveness by nearly 20% brought in about \( G = \$105 \) million of extra revenue over 32 months (January 2008 to September 2010).
- The penalty paid. A 2013 class-action settlement fund of \( F = \$4 \) million (Rao and Wang).
- Deterrence check. Even with certain detection, \( p = 1 \): \( p F = \$4 \text{M} < \$105 \text{M} \). The lie paid about 26 times over.
- Fine needed. To deter at \( p = 1 \): \( F > \$105 \) million. At a realistic \( p = 0.3 \): \( F > 105/0.3 = \$350 \) million.
Calculated in this post with these example inputs: the $105 million gain estimated by Rao and Wang and an assumed catch rate of 0.3, which is an illustration, not a measured figure. $105 million ÷ 0.3 = $350 million.
A note on that \( G \): it is revenue, not profit, so the profit gain was smaller. Using revenue makes the case conservative on the fine side; the conclusion, that the settlement was far below the gain, holds at any plausible margin.
The authors used FTC settlements against four products (Kellogg's Frosted Mini-Wheats, Dannon Activia, DanActive and Airborne) as natural experiments, with Nielsen Homescan data on 40,000-60,000 households. After the false claims stopped, monthly revenue fell 12-67%, a loss of $0.40-3.82 million a month per product, even after controlling for price and ad spending. The drop came mostly from the least loyal buyers, the people the claim had pulled in. For Frosted Mini-Wheats they estimate a $105 million revenue gain over 32 months. The limit: four US products; the effect may differ for categories where buyers can test claims themselves.
Read the paper (free)Use this to check your own category. Move the catch rate and the fine.
Does lying in an ad pay?
What does "100% true" even mean for an ad?
Only some ad claims are true or false at all. Economists sort products by how buyers can check them, following Phillip Nelson (search and experience goods, 1970) and Darby and Karni (credence goods, 1973). Courts add a fourth bin, puffery: vague praise that no reasonable person reads as a fact.
| Claim type | Example | Can a buyer check it? | Under a 100% rule |
|---|---|---|---|
| Search attribute | "1.5-tonne, 5-star AC at ₹38,990" | Before buying | Survives; easy to police |
| Experience attribute | "Lasts 8 hours on one charge" | After using | Survives if tested; lies get punished by reviews |
| Credence attribute | "Boosts immunity", "100% natural" | Rarely, even after | Needs third-party proof or disappears |
| Puffery | "The taste India loves" | Not a fact | Survives, untouched |
| Emotion, story | A grandmother, a festival, a song | Not a claim | Survives, and gains share |
What happens when lies become easy to check?
When buyers can verify claims cheaply, exaggeration falls fast, even without a regulator. The cleanest evidence comes from an unlikely place: ski-resort snow reports.
US and Canadian ski resorts publish daily "fresh snow" reports, which the authors compared with government weather data for the 2004-05 to 2007-08 seasons. Resorts reported 23% more new natural snow on Saturday and Sunday mornings (1.59 vs 1.29 inches), with no matching weekend effect in government measurements. Resorts with more to gain from weekend visitors exaggerated more. When an iPhone app feature let skiers share conditions in real time in 2008-09, exaggeration fell sharply, especially where phone reception was good. The limit: one industry with one easy-to-measure quality.
Read the paper (free)The unravelling theorem
Economists Sanford Grossman and Paul Milgrom showed in 1981 that when quality claims are verifiable, silence itself becomes a confession. Here is the logic, worked:
- Set up. Sellers' quality \( q \) is spread uniformly from 0 to 1. Any seller can prove its \( q \) at no cost, but may stay silent.
- Guess a cut-off. Suppose only sellers above \( t \) disclose. Buyers then value a silent seller at the average of the rest: \( t/2 \).
- Check the incentive. Any seller with \( q > t/2 \) does better by disclosing, so the cut-off is not stable unless \( t = t/2 \).
- Solve. The only fixed point is \( t = 0 \): everyone except the very worst discloses.
So a 100% truth rule does more than remove lies. It forces disclosure, because once claims are checkable, the brand that says nothing is assumed to be the worst. That is where the rule's real winners come from.
Make lies checkable and silence becomes the new lie: the brand that will not show its numbers is assumed to have bad ones.

Which brands and categories would win or vanish?
The rule rewards brands with a measurable edge and kills categories whose product is the claim. In India, the line is already being drawn case by case. In June 2026 the CCPA fined Storia Foods ₹1 lakh because its "100% Tender Coconut Water" was water plus 9.6% coconut water concentrate, and it said an ingredient panel cannot cure a misleading front label.
Likely to vanish
- Miracle remedies that only work as claims: 82% of healthcare ads ASCI scrutinised in April-September 2025 broke the Drugs and Magic Remedies Act.
- "100%" and "natural" boasts that rely on fine print.
- Fake scarcity and urgency, the kind of dark pattern Indian regulators now penalise.
- Undisclosed influencer posts: ASCI found 76% of top Forbes-listed digital stars breaking disclosure norms in 2025.
Likely to win
- Brands with test data: safety ratings, lab results, energy stars, warranty terms.
- Price leaders, because price is the most checkable claim of all.
- Story-led brands, because emotion is outside the rule and their rivals lose their louder tools. The Amul girl and the Hamara Bajaj films never needed a factual claim.
- Brands that already survived a truth test; the Cadbury worm crisis of 2003 and the Maggi ban of 2015 show what proof costs after trust breaks.
Sensitivity: what if enforcement is imperfect?
- At ASCI's 2024-25 pace, a complaint takes about 16 days to resolve. A claim earning ₹1 lakh a day still banks ₹16 lakh before it stops.
- If the catch rate is 20% and the penalty is ₹1 lakh, as in the June 2026 Storia order, any false claim earning more than ₹20,000 of profit pays.
Calculated in this post with these example inputs: an assumed 20% catch rate and the ₹1 lakh penalty reported by LiveLaw for the Storia order. 20% × ₹1 lakh = ₹20,000 of expected penalty.
- If platforms check claims before an ad runs, \( p \) jumps toward 1 and the fine needed falls to the gain itself.
A world of 100% true ads is possible for factual claims, and impossible to define for puffery and emotion, which are not true or false. Making it real takes three changes: fines set above the gain from lying (in the Kellogg's case, above $105 million rather than $4 million), checks before ads run so the catch rate nears 1, and certified proof for credence claims like health benefits. Once claims are checkable, the unravelling theorem does the rest: honest brands disclose, and silence starts to look like guilt.
Trust is also the fault line in the AI era; I covered the cost of cheap synthetic ads in AI slop ads and the trust bill, and the century of Indian ads that built that trust in the history of advertising in India. My own approach to claims in campaigns is on my work page.
What's new
- CCPA fines Storia ₹1 lakh over "100% Tender Coconut Water" made from 9.6% concentrate LiveLaw, Jun 2026
- ASCI: 97% of ad violations now on digital, 76% of top influencers break disclosure rules ASCI, Nov 2025
- ASCI annual report: offshore betting and realty top violators, 56% of ads misleading ASCI, May 2025
What if next?
- What if AI wrote every ad on Earth? A machine can write a million claims a day; who checks them is the whole game.
- What if a brand could only say one word, ever? If that word had to be true, which would you pick?
- What if every customer were an AI agent? Agents already penalise what they cannot verify, before any law does.
Questions people ask
Is false advertising illegal in India?
Yes. The Consumer Protection Act, 2019 defines misleading advertisements and lets the Central Consumer Protection Authority order changes and impose penalties; a June 2026 order fined a juice brand ₹1 lakh for "100%" claims. Health claims can also breach the Drugs and Magic Remedies Act, 1954. ASCI handles complaints through self-regulation.
What is puffery in advertising?
Puffery is vague, subjective praise such as "the best taste" that no reasonable buyer would read as a measurable fact. Regulators usually allow it. Once a claim becomes specific and testable, like "twice as effective", it must be proven.
Do false ad claims actually increase sales?
The best evidence says yes. Rao and Wang found that when the FTC stopped false health claims for four products, monthly revenue fell 12-67%, mostly among less loyal buyers who had been drawn in by the claim.
What is corrective advertising?
It is an order to run ads that correct an earlier false claim. The best-known case is Listerine, which in the late 1970s had to state, in about $10 million of advertising, that it would not prevent colds or sore throats.
Sources
- ASCI: Annual Complaints Report 2024-25 press release (2025) (source for 7,199 ads, 98%, 56%, sector shares and 16-day turnaround)
- ASCI: Half-Yearly Complaints Report 2025-26 press release (2025) (source for 6,117 ads, 97% digital, 82% DMR Act, 76% influencers)
- Rao and Wang: Demand for "Healthy" Products: False Claims in Advertising (2015) (source for the $105M gain, $4M settlement and 12-67% revenue loss)
- Zinman and Zitzewitz: Wintertime for Deceptive Advertising? NBER WP 17829 (2012) (source for the 23% weekend snow effect)
- TIME: Business: Taking It Back (1978) (source for the Listerine order, $10 million and 1921 start)
- FTC: Statement of Commissioner Orson Swindle (1999) (source for the 51-year campaign and rarity of corrective ads)
- LiveLaw: CCPA fines Storia ₹1 lakh (2026) (source for the 9.6% concentrate finding)
