What's Wrong With The Land Value Tax?

Ever since I read this book review about Georgism, I haven't been able to stop thinking about it. I'd always heard from the economists I knew that the land value tax was a good idea that we could never really do because of politics, and I didn't understand why. But now, having gone down the Georgism rabbit hole, I understand far more acutely both why it would be a great idea, and why politics makes it extremely difficult to do.

The rabbit hole started with the book review. After reading it, I got really excited about the ideas of Georgism, and then I got to wondering -- how practical is it, really?

I found this article on strongtowns.org talking about "The Pennsylvania Land Tax Experiment." Turns out over a dozen towns in Pennsylvania have tried some variation on it, using a "graded tax" that taxes the total property value at a low rate, and the total land value at a higher rate. Wow! I thought. They actually did it! But the Strong Towns article glossed over the fact that some towns in Pennsylvania used to have a Land Value Tax, but repealed it later. Including Pittsburgh. That seems like a pretty big hole in the story. What went wrong?

This article on landvaluetaxguide.com has the whole story.

It is sort of because implementing it is hard. However, it's not for the reason you might think! The assessments themselves are easy enough. For one thing, you can easily calculate the land value of most houses by subtracting the sale price from the amount your insurance will pay out if your house burns down. And assessors are used to evaluating somewhat intangible things about a property, when there's no market data for years at a time (because no one has bought or sold the house). It's not that much more of an issue to evaluate the structures separately from the land.

No, it turns out the problem is local politics. City politicians and assessors always have the incentive to systematically underprice houses, in order to lower property taxes. This is benign enough most of the time; tax rates have to be a little higher to accommodate it, and it all more or less works out. But in Pittsburgh, this problem got way, way worse, essentially because the county that it's in reassesses home values very infrequently. One day they decided to finally get it together and re-price all the houses, and when they did, everyone was paying WAY more taxes than they were before! People got mad and blamed this on the graded tax, and in the resulting uproar, they repealed it.

My takeaways from this story are: 1) Land-value tax is actually pretty doable, 2) Home value reassessments are actually quite important, and need to happen regularly to maintain a healthy tax system (pointed glare at California and Prop 13). Also, more tentatively: 3) The incentives of land owners make it difficult to maintain an LVT. After all, land owners want to extract economic rents, and if they've been doing so already for hundreds of years, it's politically difficult to take that away.

This article also believes that part of the reason LVT was so successful early on in Pittsburgh, and why it was so unsuccessful later, was that everyone was on board. City assessors and officials were Henry George fans. They understood the purpose of LVT and wanted to make it a reality. Later on, everyone sort of forgot about it, County assessors came in and didn't really care about the ideals of Georgism, and everyone started wondering why they had this weird tax system in the first place. Why not just rely on income taxes?

Well, because income taxes are distortionary. And more to the point, Georgism was invented before income taxes. Imagine living in that halcyon time when no one quite knew what tax system would work best, when you could discuss and debate what would come next, before the creation of massive government bureaucracies around one single form of taxation, before everyone's minds calcified into believing that that form of taxation was the only sensible one!

... Oh well. Those aren't the times we live in. And it seems the land value tax is too good for this timeline.