Why we should tax improved land values

Tax is complex, and it's natural that people would change their mind when they come into contact with new arguments and new evidence. In the past I've argued that we should only tax unimproved land values. After all, taxing improved land values reduces the incentive to improve land. But I've changed my mind: we have to tax value added, and we do it on everything else, so there is no reason to have a special land preference.

I have always wondered why land value tax enthusiasts were so obsessed with VAT. They have a very specific animus against it, as will surely be demonstrated in many twitter replies to this very piece. Nearly all taxes reduce value-added.

Taxes on capital income reduce the incentive to add value by saving and building up capital. Taxes on labour income reduce the incentive to add value by working, and indeed by gaining experience, qualifications, and skills. Transactions taxes reduce all of this and reduce the value added when something moves from someone who values it less to someone who values it more.

The only exceptions are firstly Pigovian Taxes that balance out externalities—they're more like prices than levies, and make people take into account the cost of their behaviour to others—and secondly lump sum taxes.

Lump sum taxes are those that are extremely difficult to avoid. There's a general rule that the more elastic the response to a given tax, the more inefficient it is; each pound changes more behaviour. Well the response to lump sum taxes is the least elastic—the most inelastic—of the lot. Examples are a tax on height, a poll tax, and a land value tax. Reclaiming land is possible but very difficult, and essentially impossible in the most valuable parts of cities. All of these are highly efficient because none of the incentives around adding value are changed.

Now the first two seem obviously unfair to most people. I think the third is unfair too, for reasons Sam explains here:

If buyers know that they’ll have to pay £10,000 in tax on a piece of land they valued at £100,000, they’ll only be willing to pay £90,000 for that land. The tax lowers the returns from land ownership which is reflected in the value of the land. The current owners of land are the ones who bear the full cost of future tax bills.

It's best seen as a one-off raid on current property owners, or alternatively a nationalisation of all land. Technically a one-off raid on current kebab shop owners is efficient too—if the tax only relates to past behaviour then there can be no response to it. But not hitting people for retrospective (legal) investment decisions seems like a basic threshold for a civilised, and indeed prosperous, society.

A more compelling candidate for the "single tax" base is consumption, which is what the value-added tax hits. But we currently tax consumption of property values at a far lower rate than most other goods, especially as it gets pricy. Expensive flats are massively tax-advantaged because of the intentionally-regressive council tax banding system.

Yes, taxing consumption reduces the total value added by business and commerce in society, but we have to fund government programmes somehow. If we don't want to discourage economic activity with our taxes then we need to slash government spending. Since wealth is just stored-up claims on future consumption, raising consumption taxes is also a one-off tax on wealth. However, a switch to a single consumption tax would not require higher total tax take, just the same roughly-two-fifths taken in a clearer way. And it is less arbitrary, by hitting everything, rather than a single asset class.

What taxing consumption doesn't do is expropriate particular people for arbitrary reasons, disincentivise saving and investment compared to consuming now, or arbitrarily encourage lifestyle, consumption, or employment choices—many of the taxes on the books do.