The post sitting on the front page of Hacker News this week is titled “Does Georgism work? Five years later,” it lives at astralcodexten.com, and as I write it holds 355 points and 259 comments. Read the title again. Five years later. Which means the original argument is now old enough to have a track record — and the follow-up, at whatever length, is still a follow-up to an argument rather than a report on a program.
So I went looking for the program. I could not find one. Not a full-strength land-value tax running anywhere at scale in this country, with published parcel-level numbers, an appeal docket, and a bored municipal bond analyst who has opinions about its assessment ratio. If such a place exists, it has been remarkably discreet about it.
That absence is the real subject, and it is not the subject either side of this argument wants discussed.
The Comment Section Is the Dataset
Here is the awkward shape of the Georgist revival: it is a revival of talk. The 2021 post, the 2026 post, the 259 comments underneath — that is the corpus. None of it is weak as reasoning. The deadweight-loss case is genuinely good. So is the observation that you cannot manufacture more land, and therefore cannot manufacture less of it by taxing it, which is the sort of point that annoys you because you didn’t get there first.
But reasoning is not evidence. The Georgist case has been argued to a draw at the level of theory for generations, and the empirical shelf next to it is nearly bare, because an empirical shelf requires somebody to have done the thing. The most concrete artifact in the canon remains a thought experiment about a meadow. Meanwhile the 259 commenters are, by and large, people who can produce nine hundred words on incidence and cannot name a single jurisdiction’s assessment ratio.
Assessors Are the Binding Constraint
The standard explanation for why land-value taxation never scales is the landlord lobby. That explanation is lazy and, I suspect, mostly wrong. The binding constraint is that the tax requires the least-funded office in American local government to perform the hardest valuation job in American local government: separating the value of the dirt from the value of whatever sits on the dirt.
Assessors are decent at buildings. Buildings have comparables, permits, square footage, a sales history. Dirt is where they go to war.
A two-year appraiser trainee, waiting out a board-of-revision hearing in a county office outside Columbus, put it better than any of the 259 comments did: “A house takes me an afternoon. The ground underneath it takes a year, and we still lose the appeal.”
Understand what that means politically. A land-value tax converts every assessment dispute into a budget event, because the assessment is the revenue base. Politicians will tolerate an inaccurate property tax. They will not tolerate a tax whose inaccuracy is the headline number in next year’s budget. That is a governance problem, not an economic one, and no amount of deadweight-loss algebra touches it.
The Homeowner Subsidy the Right Won’t Name
Now the part that should make readers who agree with me uncomfortable.
Exclusionary zoning is a transfer program — probably the largest one in American life — paid by renters and the young to households that already own. The political right has spent two generations developing an eloquent vocabulary of property rights and has no word for this, because the “property” being defended is a legal permission to exclude the next guy, not a bundle of use rights. Defending a permission to exclude is not defending property. It is defending a cartel.
A land-value tax is the only levy that claws that transfer back without inventing a new agency. It charges you for what your neighbors’ growth makes your ground worth, and charges you nothing for improving your own house. If you actually believe in the distinction between earned and unearned, you should want this tax, and you should be mildly embarrassed that the urbanists and the rationalist bloggers got to it before you did. The tax revolt of the last half-century was, on this reading, a mistake — a popular one, but a mistake.
A Test Someone Could Actually Run
Pick a mid-size city with a shrinking population and flat land values. Not a boomtown — the politics invert in a shrinking city, because a land tax there raises revenue from absentee holders of vacant parcels who are not voters in any ward. Publish every parcel’s land value and improvement value separately, before and after, at the address level. Pre-register the outcome measures: building permits per thousand residents, vacancy rate, dispersion in assessment ratios, and whether the assessor’s office survives its first appeal cycle intact. Run it ten years, not five.
Then the follow-up post writes itself, with numbers in it, and the commenters can argue about results instead of about a meadow.
The 2021 post gave us an argument. The 2026 post gave us 259 comments. If the 2031 post gives us 259 more, the question in the title will have been answered — just not the way the title meant.