Essential Differences Between Transaction Tax and Property Tax

In 1569 the Duke of Alba, governing the Netherlands for Philip II, asked the States General for three new taxes. The first, the Hundredth Penny, was a one-time levy of one percent on all property. The second, the Twentieth Penny, took five percent of every sale of land and buildings. The third, the Tenth Penny, took ten percent of every sale of movable goods, every time they changed hands. The States grumbled but granted the Hundredth Penny. The other two they refused, then bought off for two years with a lump sum, and when Alba tried to collect the Tenth Penny anyway in 1571, the shopkeepers of Brussels closed their shutters. Bakers stopped baking and brewers stopped brewing. The next spring the Sea Beggars took Brill, and town after town in Holland and Zeeland went over to the revolt.

The provinces were not refusing to pay for government. They had just agreed to pay a tax on what they held. What they would not accept was a tax on what they did with each other, collected at every exchange, by an officer standing at every counter. They seem to have understood something that modern tax debates mostly lose: a tax on property and a tax on transactions are different in kind, not just in rate or base. One is a charge for something received. The other is a charge for permission to deal.

This essay sets out that difference, then walks through the many forms transaction taxes take.

Two agreements

Every tax sits on top of some agreement between people, and the first question about any tax is which agreement it is attached to and who the parties to that agreement are.

Land is held under an agreement with everyone else. When a person holds a parcel, every other person in the community agrees to stay off it, to respect its boundary, to treat the holder's claim as settled, and to back that claim with courts and, in the end, with force. The holder did not make the land, and the exclusive possession they enjoy exists only because others refrain. That is the land agreement. It has two sides: the holder, and the community whose forbearance makes the holding real.

An exchange is an agreement between two people. A baker sells a loaf to a customer. A carpenter hires out a week's work. Two neighbors trade a horse for a cart. The parties are the two people who meet, and the agreement is complete when they shake hands. Nobody else's forbearance is required beyond what the law of property and contract already guarantees to everyone.

A property tax is attached to the first agreement. It is paid by one party to the other: by the holder, to the community that keeps the agreement whole. A transaction tax is attached to the second agreement. It is paid by the parties to someone who is not a party at all.

What the property tax pays for

The oldest land records and the oldest land taxes are the same documents. William the Conqueror's survey of 1086, the Domesday Book, recorded who held what land in England, how many plough teams it supported, what it was worth, and what it owed in geld. The record that settled a holder's claim and the record that fixed their tax were a single book. When Napoleon ordered a parcel-by-parcel cadastre of France in 1807, the same pairing held: surveyors measured every field, the register fixed its boundaries and its owner, and the land tax was assessed from it. In much of Europe today the cadastre that defends your boundary and the roll that bills your land tax are one office.

This is not a coincidence of bureaucracy. It reflects what the holder is buying. A land title is only as good as the community's willingness to record it, survey it, defend it in court, and refuse to recognize anyone else's claim. In 1858 Robert Torrens persuaded South Australia to go a step further: the government would register titles and guarantee them, compensating anyone who lost land through an error in the register. A Torrens title is a promise by the whole community, backed by a fund, that this parcel belongs to this person. It is a service, and a valuable one.

How valuable shows most clearly where it is missing. Hernando de Soto's The Mystery of Capital (2000) documented the enormous quantity of land and housing across Latin America, Africa, and Asia held outside any formal title, sometimes for generations. The holders often had real possession, but they could not borrow against it, sell it to a stranger, or defend it against a powerful claimant. Land with an intact agreement behind it is worth far more than the same land without one. The difference is the value of the agreement.

The property tax is the payback for that value. The holder receives exclusive use of something nobody made, recognized and defended by everyone else. The tax returns part of that value to the people who provide it. This is Henry George's argument for collecting land rent, and it is why a land value tax can be defended as a price rather than a taking: the land's value comes from the community's presence and the community's promises, and the tax hands part of it back.

Two consequences follow. First, the payback is real and continuous. The agreement is kept every day the holder holds the land, so an annual charge matches an annual benefit. Second, the payback is local. The agreement is kept by neighbors, recorded in the county, enforced in the local court. It is no accident that the property tax has always been the natural revenue of towns and counties, and that the United States Constitution made a federal tax on land awkward by requiring direct taxes to be apportioned among the states by population. The land agreement belongs to the place where the land lies.

The strongest form of this argument applies to land, not to buildings. A tax on improvements charges a person for having built something, which is an activity, and in that respect the building part of a property tax behaves more like a transaction tax than a land tax. That is one of the reasons Georgists have long argued for shifting the property tax off buildings and onto land alone. But even the ordinary combined property tax keeps its essential character: it is a recurring charge on a holding, paid to the community that secures the holding.

What the transaction tax pays for

Now consider the baker and the customer. The loaf costs four dollars. A sales tax of nine percent adds thirty-six cents. What do the parties receive for the thirty-six cents that they did not already have?

The usual answer is "government": roads, police, schools. But they would receive all of that whether or not they traded. The baker's oven, the customer's walk to the shop, and the courts that would enforce a broken contract are all already paid for, or could be paid for, by other means. The exchange itself adds nothing that the tax collector supplies. The tax collector is a third party who did not bake the bread, did not want it, and did not arrange the meeting, and yet takes a share of the agreement.

This is the essential shape of every transaction tax: a third party inserted between two people who were about to cooperate, extracting a carrying charge on the activity. Its effects follow from the shape.

It discourages exactly what makes a society wealthy. Every exchange happens because both sides expect to be better off. A tax on the exchange shrinks the gain to both, and some exchanges that would have been worth making no longer are. Economists call the lost exchanges deadweight loss, the same waste To Subsidize Is To Tax describes. The trades that do not happen leave no record, so the cost is invisible.

It compounds along the chain of production. A loaf of bread is the end of a long series of exchanges: grain to miller, flour to baker, labor to both, fuel, rent, equipment. A tax charged at each step falls again and again on the same final product. Adam Smith described the Spanish alcabala in The Wealth of Nations: a tax that had once stood at ten percent and later fourteen, charged on every sale of every kind of property and repeated every time the property was sold. He reported the view of the Spanish economist Gerónimo de Ustáriz that it had ruined the manufactures and agriculture of Spain. It does not take a high rate to do damage when the rate is charged a dozen times.

It reaches into private life. A tax on exchange cannot be collected without knowing about exchanges, so it needs a witness at every one: the stamp on every document, the receipt for every sale, the report of every wage, the record of every bank transfer. The land is visible from the road. The transaction is visible only if someone is watching. That point matters enough to get its own section below.

And it carries no natural limit. A land tax is bounded by the land's rent: it cannot take more than the land yields without the land being abandoned. A transaction tax is bounded only by the amount of activity that will tolerate it, and activity tolerates a great deal before it stops altogether. The record of transaction taxes is a record of rates and bases that kept growing.

Sand in the gears

There is a further cost that is harder to measure and easy to feel. An economy full of transaction taxes is heavy. Every exchange carries a surcharge, a form, a registration, a record to keep, a rate to look up. Every business has to know which of thousands of taxing jurisdictions a buyer is in and what each charges on which goods. Every hire comes with withholding and reporting. Every sale of a house comes with a transfer tax and its paperwork. None of these alone is large. Together they make dealing with other people slower, costlier, and more tiring than it needs to be.

Ronald Coase argued in "The Nature of the Firm" (1937) that the cost of making transactions, of finding a partner, agreeing on terms, and enforcing the deal, explains why some activity happens inside firms rather than between independent people in a market. When dealing is expensive, people stop dealing and do things in-house or not at all. A transaction tax is a transaction cost created by law. It pushes in the same direction: fewer exchanges, larger organizations, less independent trade.

This runs directly against what technology is doing. Almost everywhere else, the friction of exchange is falling fast. India's Unified Payments Interface, launched in 2016, now carries well over ten billion payments a month, most of them free and instant. Brazil's central bank launched Pix in 2020, and within a few years it was used by most of the adult population. American securities markets cut the time to settle a trade from two days to one in 2024. A seller can find a buyer on the other side of the world in seconds. Software, and now AI agents acting for people and firms, can make and settle thousands of small exchanges where a person once made one. We should expect markets to keep getting faster, finer-grained, and lighter.

A transaction tax grows heavier exactly as markets get lighter. Its burden scales with the number of exchanges, not with the value created, and much of its cost is compliance that does not shrink with the size of the sale. A fixed cost of registering, calculating, collecting, and reporting is a nuisance on a large sale and prohibitive on a tiny one. When the Supreme Court held in South Dakota v. Wayfair (2018) that states may require out-of-state sellers to collect their sales taxes, small online sellers suddenly faced the prospect of tracking the rates and rules of thousands of jurisdictions they had never seen. The faster and smaller exchanges become, the larger the share of each one the transaction tax and its paperwork consume, and the more the tax becomes the main friction left in an otherwise frictionless market.

A land tax has no such problem. It does not care how often the holder trades, how quickly, in what amounts, or through what machines. It is charged once a year on something that does not move. An economy that raised its revenue from land could let its markets run as fast as technology allows.

Questionable constitutional worth

The English-speaking world's constitutional history begins, in more than one sense, with a transaction tax. The Stamp Act of 1765 required that legal documents, newspapers, pamphlets, licenses, ship's papers, and even playing cards be printed on stamped paper bought from royal distributors. It was a tax on paperwork, which is to say on the recorded agreements of daily life, and it produced the Stamp Act Congress, the boycotts, and the slogan that taxation without representation is tyranny. Within a decade the Townshend duties and the tea duty, both taxes on goods at the moment of trade, had turned protest into war.

The Constitution written after that war reads, in places, like a document wary of transaction taxes. It allows duties, imposts, and excises, but requires them to be uniform throughout the country, so the federal government cannot pick out one region's trade to charge. It forbids any tax on exports from a state. It forbids the states from taxing imports and exports without the consent of Congress, so that no state can make itself a toll booth on its neighbors' commerce. Within three years of ratification, the first federal excise, on whiskey, produced the Whiskey Rebellion of 1794: farmers in western Pennsylvania who distilled their grain because whiskey was the only form in which it could be carried over the mountains found themselves paying a tax on the only way they had to trade.

The courts later found the deeper problem. In McCulloch v. Maryland (1819), Maryland had imposed a stamp tax on the notes issued by the Bank of the United States, a charge on each instrument of exchange. Chief Justice Marshall struck it down with the line that "the power to tax involves the power to destroy." A charge on an activity, laid by a party who does not share in it, is a tool for suppressing the activity, and the Court saw it that way. The same reasoning ran through the twentieth century. In Grosjean v. American Press Co. (1936), the Court struck down Huey Long's two percent tax on the advertising receipts of Louisiana's larger newspapers, the ones that opposed him. In Murdock v. Pennsylvania (1943), it struck down a license tax on door-to-door solicitation as applied to Jehovah's Witnesses distributing religious literature: a state may not charge for the exercise of a right. In Minneapolis Star v. Minnesota Commissioner of Revenue (1983), it struck down a use tax on the paper and ink consumed by large newspapers.

These cases did not hold that transaction taxes are unconstitutional in general. Courts uphold sales taxes, excises, and payroll taxes every day. What they show is that whenever a transaction tax lands on an activity the constitution names, speech, the press, religion, interstate commerce, the courts recognize at once what it is: a third party's charge for permission to act. The only thing protecting ordinary trade from the same analysis is that ordinary trade has no clause of its own.

A constitution in the older sense is the agreement among people about what government is for and what it may do. On that understanding the property tax fits easily: the community keeps the land agreement, and the holder pays for it. The transaction tax fits badly. It is hard to find the clause in any social compact under which two people agree that a stranger may stand between them and take a share whenever they deal with each other.

Even California's own constitution half-admits this. Proposition 13, passed in 1978, capped the property tax and allowed local governments to impose special taxes by a two-thirds vote, with two exceptions: no new ad valorem tax on real property, and no "transaction tax or sales tax on the sale of real property." The voters who wrote that clause wanted to stop the sale of a home from becoming a new occasion for taxation. The irony is that Proposition 13 itself did exactly that. By freezing assessments until a property is sold, it turned California's property tax into a tax on the act of sale. Two identical houses on the same street pay wildly different taxes, and the difference depends entirely on when each last changed hands. The payback for the land agreement, the thing that justified the property tax, was cut loose from the land and attached to the transaction.

The witness at every exchange

A property tax needs to know very little. It needs the parcel, its boundary, and its value, and all three are already in a public register or visible from the street. The assessor has no reason to ask what the holder grows, whom they invite to dinner, or what they buy. The land agreement is public by nature, and so is the tax on it.

A transaction tax cannot work that way. To tax exchanges, the collector must know about exchanges: who sold, who bought, what, when, where, and for how much. A sales tax requires every seller to keep records for the state. A value-added tax requires an invoice at every stage, each one naming both parties. An income tax requires every employer, bank, broker, and platform to report what it paid to whom. Collection is not an extra step added to the tax. Collection is surveillance, because the thing taxed is private activity and there is no way to measure private activity without watching it.

This is not a new observation. When Robert Walpole proposed in 1733 to move the duties on wine and tobacco from the ports to inland excise, with excise officers entitled to inspect warehouses and shops, the protest ran under the slogan "Liberty, Property, and No Excise," and Walpole withdrew the bill. Thirty years later the cider excise of 1763 let excise officers enter private houses to gauge the cider kept there. It was in opposing that tax that William Pitt the Elder made the famous claim that the poorest man in his cottage may bid defiance to all the forces of the Crown. The tax collector, Pitt argued, could not cross that threshold.

The American Fourth Amendment grew from the same root. In 1761 James Otis argued in Boston against the writs of assistance, general warrants that let customs officers search any house, shop, or ship for smuggled goods on which duties had not been paid. Those writs existed to enforce transaction taxes. John Adams, who watched the argument, later wrote that the child Independence was born then and there. The amendment that came out of that history protects people "in their persons, houses, papers, and effects" against unreasonable searches, and requires warrants that describe the place and the things to be seized. It was written against the habits of tax collectors at the border.

What the founders feared was an officer at the door. What transaction taxes require today is subtler and much larger: a record of every exchange, held by the bank, the employer, the payment network, or the platform, and available to the government without anyone entering a house. The courts have largely allowed it. The Bank Secrecy Act of 1970 requires banks to keep records of their customers' transactions and to report large cash deposits and suspicious activity. In United States v. Miller (1976), the Supreme Court held that a person has no reasonable expectation of privacy in records held by their bank, since they handed the information to a third party. In Smith v. Maryland (1979), it extended the same reasoning to the numbers a person dials. Under this third-party doctrine, almost everything a transaction tax needs to know falls outside the Fourth Amendment, simply because the transaction had a second party.

The doctrine is under strain. In Carpenter v. United States (2018), the Court held that the government needs a warrant to obtain weeks of a person's cell phone location records, even though a phone company holds them. Chief Justice Roberts wrote that such records give an intimate window into a person's life, and that the old doctrine did not fit the age of digital records. The same is true of a complete record of someone's purchases. In 2016 the Internal Revenue Service sought, through a "John Doe" summons, the account records of every American customer of the Coinbase cryptocurrency exchange over three years, and a federal court eventually ordered the company to produce records on some fourteen thousand of them. Congress lowered the threshold at which payment apps must report a seller's receipts to the IRS from twenty thousand dollars to six hundred in 2021, then restored the old threshold in 2025 after years of delays. Each step is defended as necessary to collect taxes already owed, and each is true on its own terms. That is the point: a tax on transactions always has a reason to see more.

Artificial intelligence changes the scale of what can be seen. A single receipt says little. Millions of receipts, joined with location data, bank records, and employment reports, and analyzed by a model trained to find patterns, say a great deal. A decade ago a retailer's statistical model could infer that a customer was pregnant from what they bought, before they had told their family. Purchase records reveal health conditions, religious practice, political affiliation, debts, relationships, and daily movements. Tax agencies are adopting the same tools: the IRS has announced it uses machine learning to select returns for audit, and tax administrations around the world are building systems to match every reported transaction against every other. Proposals for central bank digital currencies, money issued as a ledger entry by the state itself, would make every payment visible to the issuer by design.

None of this needs bad intent. A tax that falls on activity creates a standing demand for knowledge of activity, and as the cost of storing and analyzing that knowledge falls toward zero, the demand will be met. The Fourth Amendment's authors drew the line at the door of the house because that was where the collector stood in their time. In ours, the collector stands inside every payment. A society that wants to keep any part of private life private should count this as a cost of transaction taxes, and a serious one.

The contrast with the land tax is complete. A tax on land asks nothing about what anyone does. It can be assessed from maps, sales records, and a walk down the street. It leaves the holder's activities, purchases, and associations entirely their own. Of all the ways to raise public revenue, it is the one that needs the least surveillance.

The types of transaction tax

Transaction taxes go by many names, and part of their political success comes from the fact that they rarely look like the same thing. Grouped by the agreement they sit on, they fall into a handful of families.

Taxes on buying and selling

General sales taxes. A percentage of the price of retail sales, collected by the seller. The alcabala of Castile was one; the Tenth Penny would have been another. The modern American version began when Mississippi adopted a two percent retail sales tax in 1932, in the depth of the Depression, and spread to almost every state within a generation. Sales taxes are popular with governments because they are collected in small amounts at the moment the buyer has money in hand, and because the seller, not the government, does the collecting.

Value-added taxes. A sales tax collected in pieces at every stage of production, with each business paying tax on its sales and claiming a credit for the tax it paid on its inputs. Maurice Lauré designed the modern version for France in 1954, and some 170 countries now use one, usually under the name VAT or GST. The credit mechanism prevents the tax from compounding, which is its great technical virtue over the alcabala. Its great political virtue is that it is nearly invisible: the tax is buried in the price, and every business in the chain becomes an unpaid collector who must keep records for the state.

Turnover and gross receipts taxes. A tax on a business's total sales, with no deduction for what it paid its suppliers. Washington's Business and Occupation tax, adopted in 1933, is the longest-running American example; Germany's turnover tax before 1968 was the classic European one. Because each stage pays on the full value of what it sells, including what was already taxed at the stage before, these taxes compound in exactly the way Smith described, and they push firms to merge their suppliers in-house simply to avoid being taxed at each step.

Excise taxes. A charge on the sale of particular goods, usually per unit rather than by price: so much per gallon of fuel, per pack of cigarettes, per proof gallon of spirits. Excises are the oldest transaction taxes still in common use, and the ones most often defended by appeal to something other than revenue. A fuel excise spent on roads is sometimes described as a user fee, and a tobacco excise as a charge for the harm smoking does. Those arguments are not empty, and they are discussed below. But the whiskey excise of 1791 was neither: it was a charge on frontier farmers for turning their grain into something they could carry.

Taxes at the border

Tariffs, customs duties, and imposts. A tax on an exchange between a buyer on one side of a border and a seller on the other. The tariff was the federal government's main source of revenue until 1913, and it has a special place in the history of the transaction tax because it is so easily disguised as something other than a tax. A tariff is paid by the buyer, in higher prices, to a government that took no part in the sale, and it is often defended as a protection for producers who did not take part in it either. The Townshend duties and the tea tax were tariffs. So was the tax that touched off the Rebelión de las Alcabalas in Quito in 1592, when the Spanish Crown extended the alcabala to its American colonies and the city's council and townspeople rose against it.

Taxes on paper

Stamp duties and documentary taxes. A tax on the writing of a document: a deed, a contract, a bill of exchange, a share certificate, a license. England has collected stamp duties since 1694, and the Stamp Duty Reserve Tax still takes half a percent of most purchases of British shares. The Stamp Act of 1765 belongs here. Documentary taxes are transaction taxes in their purest form, since they charge not for any good or service but for the act of recording an agreement.

Real estate transfer taxes. A tax on the sale of land and buildings, usually paid when the deed is recorded. Alba's Twentieth Penny was one. In the United States most states and many cities levy one; Toronto doubled its own on top of Ontario's in 2008. These deserve particular attention because they are so often confused with property taxes. They are not. A property tax is a recurring charge on holding land, paid for the agreement that secures the holding. A transfer tax is a one-time charge on changing who holds it, paid at exactly the moment the community does least: recording a new name in a register that already exists. Transfer taxes discourage people from moving, keep land in the hands of those who no longer need it, and tie up housing that would otherwise change hands.

Mansion taxes. Transfer taxes that apply only above a price threshold. New York's has taken one percent of residential sales over a million dollars since 1989. Los Angeles's Measure ULA, passed in 2022, takes four percent of sales over five million dollars and five and a half percent over ten million, on the whole price, not just the part above the threshold. Its collections fell well short of its projections in its first years because many high-value properties simply stopped selling, or sold just under the line. A tax on the act of selling is avoided by not selling, and land that does not sell is land whose use does not change.

Taxes on money moving

Financial transaction taxes. A small percentage of the value of trades in stocks, bonds, currencies, or derivatives. James Tobin proposed a tax on currency transactions in 1972 to slow speculation, and the idea is still called a Tobin tax. Sweden tried a tax on equity trades in 1984, at half a percent, then doubled it. Trading volume moved to London, revenue fell far short of projections, and a later tax on bond trades raised almost nothing. Sweden repealed the whole scheme by 1991. France has taxed purchases of large French companies' shares since 2012, and a European Union proposal for a common financial transaction tax has been under negotiation for more than a decade without being adopted.

Bank debit taxes. A tax on every withdrawal or transfer from a bank account. These have mostly been Latin American experiments: Argentina's impuesto al cheque, Venezuela's, and Brazil's CPMF, introduced in 1997 as a "provisional" contribution to health spending, which lasted more than ten years and reached 0.38 percent of every debit before the Brazilian Senate let it lapse in 2007. Debit taxes reach every payment in an economy, which is why they raise money so easily, and why they push people toward cash and away from banks.

Taxes on work and income

Payroll taxes. A charge on the wages paid when an employer hires a worker, usually split between the two. In the United States they are called contributions and tied to social insurance, but in form they are a tax on the agreement to work, charged to both parties by a third. They make every hire more expensive and every off-the-books arrangement more attractive.

Income taxes, at the edge of the category. An income tax is levied on a person rather than a sale, but nearly every dollar of income is the result of an exchange, and the tax falls on the gain from exchange. The United States Supreme Court held in Pollock v. Farmers' Loan & Trust Co. (1895) that a tax on income from property was a direct tax requiring apportionment, and it took the Sixteenth Amendment in 1913 to authorize the federal income tax as it now exists. Whether income tax belongs in this list is arguable. What is not arguable is that it shares the transaction tax's central feature: it charges people for producing and trading rather than for holding something the community provides.

Capital gains taxes. A tax on the profit from selling an asset, collected only when it is sold. The gain may have accrued over decades, but the tax waits for the transaction. The result is the familiar lock-in effect: owners hold on to assets they would otherwise sell, because selling triggers the tax and holding does not.

Taxes on gifts and deaths

Estate, inheritance, and gift taxes. A tax on the transfer of property without a price: at death, or as a gift during life. The American federal system calls these "transfer taxes," and the name is accurate. They differ from the other taxes in this list because one side of the agreement receives something without giving anything in return, and that has given them a defense other transaction taxes lack. But the shape is the same: a third party taking a share when property passes from one person to another.

Taxes on permission

Licenses and privilege taxes. A charge for permission to carry on a trade, an occupation, or a business at all: a business license tax, an occupational license fee, a gross receipts charge on "the privilege of doing business." Where the fee pays for an inspection that actually protects the customer, it can be something close to a price. Where it is simply a percentage of receipts, it is a sales tax by another name, and where it is charged for the exercise of a protected right, Murdock shows what the courts will make of it.

New and hidden forms

Digital services and platform taxes. A tax on the revenue large online platforms earn from transactions between their users, such as France's three percent digital services tax of 2019. These are gross receipts taxes on intermediaries, in practice passed on to the buyers and sellers who meet on the platform.

Room, ticket, and telecom taxes. Hotel occupancy taxes, admissions taxes, rental car taxes, and the long list of charges on telephone bills. Each is a sales tax on a particular kind of sale, usually chosen because the buyer is a visitor who does not vote locally.

Reassessment-on-sale property taxes. Proposition 13's acquisition-value system, described above. It is formally a property tax and substantively a transaction tax, since what determines the bill is the date of the last sale. It shows that the line between the two kinds of tax is not drawn by name. It is drawn by what triggers the charge.

Where the line blurs

Not every charge connected to a transaction is a pure carrying charge, and the distinction is more useful if its edges are stated plainly.

A toll on a road is a price, not a tax, because the party collecting it supplied the road, as Land as Road argues. A fuel excise spent entirely on roads comes close to the same thing, a rough price for road use collected at the pump. The test is whether the third party contributed something to the transaction it charges. When it did, the charge is not an insertion but a participation.

An excise that prices a real harm, smoke in a neighbor's lungs, carbon in a shared atmosphere, can be defended as a charge for the harm rather than for the trade. But then the real base of the tax is the harm, not the sale, and a tax on the sale is only a rough proxy for it. A charge measured directly on the harm is usually better.

The property tax has its own weak edge. When it falls on buildings rather than land, it charges for activity, and when assessments rise faster than a long-settled household's ability to pay, it can force people from homes they have held for decades. These are real defects. The answer to them is to tax land rather than buildings and to let the tax follow the land's rent, not to abandon the recurring charge for one that fires at every sale.

What follows

The distinction can be put in a few working principles.

Tax what is held, not what is done. A recurring charge on held land pays for the community's continuing work of securing the holding. A charge on exchange pays for nothing the parties receive.

Keep the payback local. The land agreement is kept by neighbors, recorded in a local register, and enforced in a local court. Its revenue belongs to the same place. This is part of why land revenue suits municipal self-government so well.

Judge a tax by its trigger, not its name. A property tax that fires at sale is a transaction tax. A fuel excise spent on the road it is collected for is a price. The question is always what event causes the charge and what the payer receives in return.

Let exchange be light. As payments and markets grow faster and finer, a tax charged per exchange becomes the main drag left on them. Revenue drawn from what is held, not from how often people deal, lets markets move at the speed technology allows.

Prefer taxes that need no watching. A tax base that can be measured without recording people's private exchanges is worth a great deal in an age when every record can be kept forever and analyzed by machine. Where transaction records must exist, they should be collected for the narrowest purpose, kept no longer than needed, and reachable only with a warrant.

Treat every transaction tax as a burden to justify. When a third party wants a share of an agreement between two people, the burden is on the third party to say what it contributed. Most transaction taxes cannot say, and the revolts of the Netherlands, Quito, and the American colonies suggest that people have long recognized that.

The shopkeepers of Brussels in 1571 were not anarchists. They had just consented to a tax on what they owned. What they closed their shutters against was the idea that the state should stand at every counter and take its share each time a loaf of bread crossed it.

Sources

• Geoffrey Parker, The Dutch Revolt (1977), on Alba's Hundredth, Twentieth, and Tenth Penny.
• Adam Smith, An Inquiry into the Nature and Causes of the Wealth of Nations (1776), Book V, ch. 2, on the Spanish alcavala.
• Gerónimo de Ustáriz, Theórica y práctica de comercio y de marina (1724).
• Bernard Lavallé, Quito et la crise de l'alcabala, 1580–1600 (1992).
• Domesday Book (1086).
• Robert Torrens, The South Australian System of Conveyancing by Registration of Title (1859); Real Property Act 1858 (South Australia).
• Hernando de Soto, The Mystery of Capital (2000).
• Henry George, Progress and Poverty (1879), Book VIII.
• Stamp Act 1765 (5 Geo. III c. 12).
• Constitution of the United States, Art. I, §§ 8–10.
• McCulloch v. Maryland, 17 U.S. 316 (1819).
• Pollock v. Farmers' Loan & Trust Co., 157 U.S. 429 (1895).
• Grosjean v. American Press Co., 297 U.S. 233 (1936).
• Murdock v. Pennsylvania, 319 U.S. 105 (1943).
• Minneapolis Star & Tribune Co. v. Minnesota Commissioner of Revenue, 460 U.S. 575 (1983).
• California Constitution, Art. XIII A (Proposition 13, 1978).
• Los Angeles Measure ULA (2022).
• James Tobin, "A Proposal for International Monetary Reform," Eastern Economic Journal (1978).
• Peter Umlauf, "Transactions Taxes and the Behavior of the Swedish Stock Market," Journal of Financial Economics (1993).
• Maurice Lauré, La taxe sur la valeur ajoutée (1952).
• Paul Langford, The Excise Crisis: Society and Politics in the Age of Walpole (1975).
• William Pitt the Elder, speech against the cider excise (1763), as reported in Lord Brougham, Historical Sketches of Statesmen (1839).
• John Adams to William Tudor, 29 March 1817, on James Otis and the writs of assistance.
• Constitution of the United States, Amendment IV.
• Bank Secrecy Act, Pub. L. 91-508 (1970).
• United States v. Miller, 425 U.S. 435 (1976).
• Smith v. Maryland, 442 U.S. 735 (1979).
• Carpenter v. United States, 585 U.S. 296 (2018).
• United States v. Coinbase, Inc., No. 17-cv-01431 (N.D. Cal. 2017).
• American Rescue Plan Act of 2021, § 9674 (Form 1099-K threshold).
• Ronald H. Coase, "The Nature of the Firm," Economica 4 (1937).
• South Dakota v. Wayfair, Inc., 585 U.S. 162 (2018).
• National Payments Corporation of India, UPI monthly statistics; Banco Central do Brasil, Pix statistics.
• U.S. Securities and Exchange Commission, Shortening the Securities Transaction Settlement Cycle (T+1), final rule (2023).
• Charles Duhigg, "How Companies Learn Your Secrets," New York Times Magazine (2012).

---

A note on how this piece was written: the subject, the distinction it draws, and the positions it takes are mine. Much of the research, the examples, and the sentences were drafted by an AI model working from that direction, and then edited by hand. I'd rather say that plainly than have a reader guess at it.