Here is a claim I will make without hedging, after twenty years of consumer-protection law: there is not one person in America who reads their entire cell phone contract. Not one. And the person who did read it start to finish and then signed it anyway would be a fool, because the terms are not negotiable — the only choice on offer is take it or go without a phone.
Yet in a courtroom, you are presumed to have read every word. That presumption is treated as ordinary, unremarkable law. It is a fiction, everyone in the building knows it is a fiction, and almost nobody in a position to say so will say it out loud.
So let's say it out loud, with the data.
One or two people out of a thousand
Researchers at NYU tracked the actual browsing behavior of 48,154 monthly visitors to the websites of 90 software companies — not what people said they did, what they did. The finding: only one or two out of every thousand shoppers even clicked to open the license agreement.
And the ones who opened it? Median time on the page: 34 seconds. Average, 59 seconds. Nearly half spent under 30 seconds. The agreements themselves averaged 2,277 words — eight or nine minutes of honest reading.
So the "informed minority" that the entire legal defense of boilerplate depends on — the idea that a few careful shoppers police the terms for everyone else — is, empirically, two people in a thousand giving it half a minute.
A separate experiment built a fake social network and watched 543 people sign up. Seventy-four percent skipped the privacy policy entirely. Of those who didn't skip it, average time spent reading was 73 seconds — for a document that, at normal adult reading speed, would take 29 to 32 minutes. The terms of service got 51 seconds against a 15-to-17-minute read.
Ninety-seven percent agreed to the privacy policy anyway. And 98 percent missed the clauses the researchers had planted in it — including one giving the company the right to share their data with the NSA and their employer, and another requiring them to hand over their first-born child as payment for the service.
Those people were not lazy. They were rational. Which brings us to the arithmetic nobody wants to do.
It isn't laziness. It's unreadable, and it's on purpose.
Law professors ran standard linguistic readability tests on the sign-up contracts of the 500 most popular websites in the United States. The consumer-materials benchmark is an eighth-grade reading level, because that is roughly where most American adults read — the same standard the FDA and NIH apply to informed-consent forms.
The results: the median contract scored at a grade level of 14.9 — college sophomore, junior. Four hundred ninety-eight out of 500 — 99.6 percent — failed the readability benchmark. On the second test they ran, the same 99.6 percent failed. In 70.4 percent of these contracts, the average sentence ran longer than 25 words.
The authors' summary is that these agreements read like articles in academic journals — documents written for specialists, which is a peculiar way to write something you insist a bus driver has read and understood.
And here is the asymmetry at the center of the whole thing, in their words: "the duty to read is unilateral: although consumers are presumed to read contracts, there is no general duty on suppliers to provide consumers with readable contracts."
Read that twice. You are obligated to read it. Nobody is obligated to make it readable. That is not a bug that crept into consumer law. It is the architecture — and it is exactly what enables companies to take advantage of Americans, because a term nobody can find and nobody can parse is functionally a term nobody agreed to, enforced as though they did.
"But there are disclosures." Here's what disclosure is actually worth.
This is where defenders of the system retreat: fine, but the important terms have to be disclosed. Congress has been passing disclosure mandates since 1968 on exactly that theory.
It happens that we have an unusually clean measurement of what disclosure is worth, and it comes from the same statute, the same data, and the same research design — so there's no comparing apples to oranges.
The Credit CARD Act of 2009 did two different things. Title I substantively limited fees — it banned certain charges outright. Title II required better disclosure — most notably a table on every statement showing what you'd pay if you only made minimum payments, and what you'd save by paying the balance in 36 months.
Economists studied both, using 160 million credit card accounts, comparing consumer cards (covered by the law) against small-business cards (not covered). Published in the Quarterly Journal of Economics, the results:
- The fee restrictions saved consumers $11.9 billion a year.
- The disclosure mandate moved half a percentage point of accounts to the 36-month payment — and by the authors' own upper-bound estimate in the working paper, saved consumers about $71 million a year.
Same law. Same data. Same method. Banning the practice was worth roughly 170 times more than telling people about it.
That is not an isolated result. When British regulators tested minimum-payment nudges on 183,441 cardholders, they found "precise null effects" on debt — and in a follow-up survey, 96 percent of people still underestimated how long repayment would take. The nudges reduced confusion measurably. Underestimation remained, in the authors' words, "overwhelmingly common."
The government's own auditors reached the same place years earlier. When the GAO tested real cardholders on the legally mandated credit-card disclosures — the famous "Schumer box" — only about half could use them to identify the default interest rate or what would trigger it; asked to find the default rate in the actual cardmember agreement, 3 of 12 could do it. The FTC's testing on mortgage forms was worse: with the legally required disclosures in hand, half of borrowers could not identify the loan amount, two-thirds did not realize their loan carried a prepayment penalty, and nearly nine in ten could not identify the total up-front charges. And the FTC noted its test form was better than what the law required — meaning the real-world numbers are worse.
Even the regulators have admitted it in writing. In the Federal Register, federal banking agencies stated that consumer testing "identified the limitations of disclosure," that "disclosure alone would not be effective in preventing the harm," and that "disclosure alone was insufficient to protect consumers." The Chairman of the Federal Reserve said it plainly in 2008: "improved disclosures alone cannot solve all of the problems consumers face."
One honest counterexample, because the rule of this series is that inconvenient evidence goes in the text. A field experiment at payday lenders found that reframing the cost in accumulating dollars rather than an annual rate cut future borrowing 11 percent over four months. Disclosure designed by people who study how humans actually decide can work. The generic legal boilerplate we mandate by the ream mostly does not — and even those researchers called themselves "agnostic as to the overall sufficiency of better disclosure policy."
The scholars got there first, and they were blunt about it
Two law professors — one at Chicago, one at Michigan — spent a hundred pages in the University of Pennsylvania Law Review documenting this across consumer lending, informed consent, and contract formation. Their conclusion: "the empirical history of mandated disclosure is a history of failure." And they identified why it can't be patched. There are two problems, and the second one is the one nobody legislates for:
Overload — any single disclosure grows past what a human can hold, and it grows inevitably, because once you're listing the calories you can't argue the sodium is irrelevant. Accumulation — lawmakers write disclosures one at a time, but you receive them all at once, competing with each other and with your actual life. In their words: "One disclosure by itself may seem trivial, but en masse disclosures are overwhelming. Even if disclosees wanted to read all the disclosures relevant to their decisions, they could not do so proficiently, and practically they could not do so at all."
They also made a point that ought to matter to anyone who claims to be on the side of working people: "Mandated disclosure not only helps the rich more than the poor, but also perversely and unintentionally obligates the poor to subsidize the rich." The person with a lawyer, an accountant, and an unhurried afternoon can use disclosure. The person working a double shift pays for a system built around the assumption that they will read it.
Two fairness notes, because these authors deserve accurate treatment. They deliberately declined to prescribe a replacement — they called that "too large a question" for the article, so nobody should recruit them as advocates for any particular alternative. And a pair of NYU professors writing in the Harvard Law Review made the narrower, more careful version of the argument I'd actually adopt: the point is not that disclosure is always wrong, but that "there should be no presumption or precommitment in favor of choice-preserving regulatory options over others." Put the substantive rule on the table and compare it honestly, instead of reaching for a disclosure because it's the politically cheapest thing in the drawer. That debate is live — Cass Sunstein published a response defending choice-preserving approaches — and reasonable people are on both sides of it.
But note what is not in dispute: nobody in that literature argues the disclosures are working as promised.
And what it buys the company: a clause you never saw, waiving a right you didn't know you had
The most consequential term in most consumer contracts isn't the price. It's the arbitration clause — the one that quietly moves any future dispute out of court and out of any class action.
The federal government studied how well that works as "agreement." Among consumers whose credit-card contracts contained an arbitration clause, 93 percent either didn't know whether they could sue, or believed they could — when they could not. Nine in ten people had bargained away their day in court and did not know it.
And the practical result, from the same study: across six consumer financial markets over two years, arbitration produced $172,433 in affirmative relief to consumers. Class actions over five years produced about $2.2 billion. That is not a comparison of two roads to justice. It's a comparison of a road and a wall.
Minnesota already tried. Our law has never produced a single reported decision.
Minnesota has a plain-language statute. In the same law review article that measured those 500 unreadable contracts, the authors list state plain-language laws in a footnote and describe ours this way: Minn. Stat. § 325G.31, "requiring nearly all consumer contracts to be 'clear[ly] and coherent[ly]' written, but failing to provide objective criteria."
A peer-reviewed law review, naming Minnesota's statute and its precise defect. We told companies to write clearly, never said what "clearly" means, and never attached a number — so nobody can ever be shown to have violated it.
Here is the proof, and it's the fact that convinced me this article was worth writing. A search of the case databases returns no reported Minnesota decision construing § 325G.31. Not one, in forty-five years. A law can be on the books since 1981, cover nearly every consumer contract in the state, and generate exactly zero cases — which is what a statute looks like when it is unenforceable by design rather than by neglect.
The rest of the Act is aging the same way. Its coverage ceiling — the contract value above which it simply stops applying — hasn't been raised since 1984. Mortgages are carved out. Attorney fees in a class action are capped at $10,000, which is to say no lawyer will ever bring one. And § 325G.35 gives the Attorney General a power almost nobody has heard of: for a $50 fee, a company can submit a contract for state certification that it complies. That tool has sat there, unused, for four decades.
One thing Minnesota did get right, and recently. When the Eighth Circuit — the federal appeals court covering Minnesota — vacated the FTC's "click to cancel" rule in July 2025, six days before it took effect, the federal protection against subscription traps reverted to a regulation written for book clubs in 1973. But Minnesota had passed its own click-to-cancel law, in force since January 1, 2025. When Washington's rule died, Minnesotans kept theirs.
That is the whole argument for state consumer protection in one episode — and the whole argument for writing state laws with teeth, because the ones without teeth just sit there.
What we can do
Put a number in the statute. Readability is measurable — the same standardized tests the researchers used are built into word processors. Amend § 325G.31 to require consumer contracts to meet an objective readability score, with the key terms — price, term length, automatic renewal, cancellation, fees, arbitration — stated in a bounded summary at the top. A rule with a number is enforceable. "Clear and coherent" is a wish, and forty-five years without a single reported case is the proof.
Modernize the rest of the Act. Raise a coverage ceiling that hasn't moved since 1984. Fix a fee cap that guarantees no one will ever bring the case. And use § 325G.35 — the certification power the state has held unused since the Reagan administration.
Make the duty mutual. If the law is going to presume you read it, the law should require that it be readable. Those two propositions belong in the same sentence or neither belongs in a courtroom.
And stop treating disclosure as the finish line. The single best-measured lesson in modern consumer finance is that prohibiting an abusive practice is worth orders of magnitude more than describing it in a box. When the choice is between a new disclosure and a substantive rule, the evidence says take the rule. When someone offers a disclosure instead of a rule, understand what's being traded away.
I have spent my career on the far end of these contracts — reading, for one client at a time, what an entire company wrote knowing nobody would. The presumption that you read it is not a small technicality. It is the load-bearing fiction underneath a great deal of what gets done to ordinary people, and Minnesota is perfectly capable of being the state that says so first.
Nobody reads the fine print. Everybody knows it. It's time the law admitted it.
Sources
Yannis Bakos, Florencia Marotta-Wurgler & David R. Trossen, Does Anyone Read the Fine Print? Consumer Attention to Standard-Form Contracts, 43 J. Legal Stud. 1 (2014) — the 48,154-visitor browsing study and the one-to-two-per-thousand access rate. Jonathan A. Obar & Anne Oeldorf-Hirsch, The Biggest Lie on the Internet, 23 Info., Commc'n & Soc'y 128 (2020) — the NameDrop experiment, reading times, agreement rates, and the planted clauses (the "first-born child" term was a research instrument in a fictitious service's terms, not a real company's). Uri Benoliel & Shmuel I. Becher, The Duty to Read the Unreadable, 60 B.C. L. Rev. 2255 (2019) — readability testing of the 500 most popular U.S. sign-in-wrap contracts (median Flesch-Kincaid 14.9 against an 8.0 benchmark; 498 of 500 failing both tests; 70.4% with average sentences over 25 words), the unilateral-duty formulation, and the footnote characterizing Minn. Stat. § 325G.31. Sumit Agarwal, Souphala Chomsisengphet, Neale Mahoney & Johannes Stroebel, Regulating Consumer Financial Products: Evidence from Credit Cards, 130 Q.J. Econ. 111 (2015) — the $11.9 billion annual saving from fee restrictions and the finding of "a small increase in the share of accounts making the 36-month payment value but no evidence of a change in overall payments"; the ~$71 million upper-bound estimate for the disclosure provision is from the authors' earlier NBER Working Paper 19484 (Sept. 2013) and is cited to the working paper, not the published article. Paul Adams et al., Do Nudges Reduce Borrowing and Consumer Confusion in the Credit Card Market?, 89 Economica S178 (2022) — the 183,441-cardholder null result and the 96 percent underestimation finding. U.S. Government Accountability Office, GAO-06-929 (2006) — cardholder testing on credit-card disclosures. Federal Trade Commission Bureau of Economics, Lacko & Pappalardo, Improving Consumer Mortgage Disclosures (June 2007) — the 819-consumer mortgage-form testing, including the note that the tested form exceeded legal requirements. Unfair or Deceptive Acts or Practices, 74 Fed. Reg. 5498 (Jan. 29, 2009) — the agencies' statements on the limits of disclosure; Chairman Ben S. Bernanke, statement of May 2, 2008. Marianne Bertrand & Adair Morse, Information Disclosure, Cognitive Biases, and Payday Borrowing, 66 J. Fin. 1865 (2011) — the 11 percent reduction and the authors' own hedge. Omri Ben-Shahar & Carl E. Schneider, The Failure of Mandated Disclosure, 159 U. Pa. L. Rev. 647 (2011) — the "history of failure" conclusion (p. 746), the overload and accumulation framing (pp. 686–90), and the distributive finding (pp. 741–42); the authors expressly decline to prescribe a regulatory alternative (p. 742) and are not cited here as endorsing one. Ryan Bubb & Richard H. Pildes, How Behavioral Economics Trims Its Sails and Why, 127 Harv. L. Rev. 1593 (2014) — the argument that regulatory instruments should be compared without a presumption favoring choice-preserving tools (p. 1601); Cass R. Sunstein, Nudges vs. Shoves, 127 Harv. L. Rev. F. 210 (2014), is the published response defending choice-preserving approaches. Consumer Financial Protection Bureau, Arbitration Study (2015) — the finding that 93 percent of consumers subject to an arbitration clause did not know or wrongly believed they could sue, and the comparison of affirmative arbitration relief ($172,433 across six markets over two years) to net class-action relief (approximately $2.2 billion over five years). Minnesota's Plain Language Contract Act, Minn. Stat. §§ 325G.29–.36, including the operative standard at § 325G.31 (unamended since 1981), the coverage ceiling last raised in 1984, the attorney-fee cap, and the Attorney General's certification power at § 325G.35 — verified against raw statutory text at revisor.mn.gov; a search of case databases returned no reported Minnesota decision construing § 325G.31. Custom Communications, Inc. v. FTC, 142 F.4th 1060 (8th Cir. July 8, 2025), vacating the FTC's Negative Option Rule days before its compliance date; Minnesota's own automatic-renewal law, Minn. Stat. §§ 325G.56–.63, effective Jan. 1, 2025.
Statutory and regulatory language above was taken from the primary sources, not from secondary summaries. Corrections: campaign@madgettformn.com.