Why land tax can't be shifted onto tenants

Gavin R. Putland expands on his letter in Wednesday's Financial Review.*

Landlords hate land tax because they can't pass it on. So they campaign against it — by pretending that they can pass it on!

In its crudest form, their argument tacitly assumes that a land tax behaves like a sales tax. But there is no analogy between the two. A seller can shift a sales tax onto buyers by holding out for a higher price, and can afford to hold out because the tax isn't payable until the item is sold. But if a landlord tries to shift a land tax onto tenants by holding out for higher rent, the tax is still payable in the mean time, although there is no rental income with which to pay it. So the tax increases the pressure on the landlord to find a tenant. And you attract a tenant by reducing the rent, not by raising it!

Sometimes the “crudest form” of the argument is accompanied by a supply-demand graph with an upward-sloping supply curve, indicating that higher rents make landlords more willing to let, just as higher prices make sellers more willing to sell. The analysis of the graph need not be discussed further, except to say that it assumes that the act of supply triggers the tax. That assumption is not applicable to land tax, which is payable even if the owner fails to “supply” the land. But if you don't notice that assumption, or don't notice that it's not applicable, you will be taken in by the impressive visuals.

In a more sophisticated form, the argument alleges that because the land under owner-occupied housing is exempt from land tax, landlords can avoid the tax by selling to owner-occupants, reducing the supply of land for rent, hence raising rents. That reasoning is faulty on two counts.

1. To the extent that the tax reduces the supply of land for rent, it increases the supply for owner-occupation. This in turn draws demand out of the rental market and into the owner-occupation market, offsetting the effect on rents.
2. If the effect on rents is not fully offset, the result is that rents are higher than they would be if the land tax applied across the board — not that rents are higher than they would be if there were no land tax at all. Indeed, if there were no land tax at all, landlords would have less need to sell or let their vacant properties, so that both prices and rents would rise.

If “supply” of land is understood as selling or letting, the pressure on owners to sell or let their land amounts to an increase in effective supply. If on the contrary, the “supply” of land is understood simply as the quantity of land in existence, then the pressure to sell or let amounts to a reduction in a particular form of demand, namely speculative holding of vacant property. Either way, the effect of the tax is to reduce prices and rents.

Consider two adjacent sites, one of which you can own with no holding cost, and the other of which has a holding cost in the form of a land tax of (say) two percent of the market value per year. If you were a prospective buyer, for which lot would you be willing to pay a higher up-front price? If you were the owner of both, which one would you be keener to sell? If the two lots were owned by different landlords, which one would be keener to get a tenant?

Because land tax is levied on the value of the land, it cannot reduce the effective supply of land by forcing the land out of use; if it did, the land would then have no value on which to levy the tax.

If, on the contrary, the tax were on the value of land plus buildings — like municipal rates in Victoria! — it would discourage building, reducing the supply of accommodation and forcing up rents. But it has no such effect when levied on the value of land alone.

Which brings us to another fallacy by which critics of land tax pretend that it is passed on in rents: they argue that land tax reduces the supply of accommodation by discouraging investment in “real estate” or “property”, where the quoted term glosses over the distinction between buildings and land. Buildings indeed owe their existence to investors. But land is a gift of nature and would still exist if all the “investors” who have ever owned it had never been born. The existence of the land is the one thing that stays constant under all possible taxation systems. So the effect of a tax consists in its effect on the use of that given expanse of land. And the effect of land tax, as we have seen, is to encourage owners to generate income from the land — for which purpose they probably need to build on it, thus increasing the supply of accommodation.

What about contractual provisions? A contract purporting to make a commercial tenant liable for land tax will distress the tenant if the tax bill rises during the term of the lease and the landlord has not fully developed the site, so that the tenant cannot earn a commensurate increase in income. But that's because, during that term, the tenant is locked in and market forces are locked out. As soon as a lease is renegotiated, the market compensates for the contract and puts the tax burden back on the landlord. In the mean time, the distress of the tenant is extremely effective for generating adverse headlines about land tax, and for inducing tenants to go out to bat for their landlords by campaigning against land tax, as if the problem were with the principle of the tax rather than the implementation. That, no doubt, is the main reason why the law in most jurisdictions permits such contracts. But the effects of bad laws and restrictive trade practices must not be confused with the effects of market forces.

What is true of rents is also true of prices charged by businesses that own their premises. A tax on the values of buildings reduces the supply of business accommodation, so that there are fewer businesses, hence less competition for customers, hence higher prices. But a tax on the value of land has no such effect. On the contrary, a tax on the value of land encourages commercial owner-occupants to generate sales with which to pay the tax, and encourages commercial landlords to seek tenants. These effects mean more businesses and more competition for customers, hence lower prices.

Or consider two competing businesses, A and B, and suppose that A is on a site that is highly desirable and therefore of great value, while B is on a site that is barely viable and therefore of negligible value. For B, the land tax bill will be negligible and therefore will have negligible effect on prices, whether it is passed on or not. For A, the land tax bill will be high. But if A tries to recover the tax through prices, it will be undercut by B. It follows that the tax is not borne by customers, but merely takes away some of the locational advantage of A.

Or look at it this way. Competition between businesses tends to reduce the return on capital to the bare minimum. If a business site commands a market value, it does so because the return to capital on that site is so much higher than the minimum, so that the owner of the site can demand the surplus in return for permission to use the site. As long as the land tax amounts to less than the surplus (and it does; otherwise there would be no land value on which to levy the tax), it merely takes part of that surplus from the owner and does not cut into the return on capital for the occupying business.

While the effect of a land tax by itself is to reduce rents, that effect may of course be superimposed on the effects of other things.

In particular, if an increase in land tax is accompanied by a cut in taxes payable by tenants, the tax cut will increase the capacity of tenants to pay rent and will therefore tend to be competed away in higher rents. But this is a case of a benefit being shifted from tenants to landlords, not a cost being shifted from landlords to tenants.

A similar effect may occur when the revenue from land tax is spent. If the government derives revenue from land values, it has the ability and the incentive to invest in infrastructure that makes the surrounding land more desirable, hence more valuable. Obviously tenants will then offer higher rents for the affected land. But this is an effect of public expenditure, not an effect of taxation; it is another example of a benefit being shifted from tenants to landlords, not a cost being shifted from landlords to tenants.

Does this mean that the tenants pay for the infrastructure through their rents? Yes, but tenants do that regardless of what tax system is in place, and regardless of whether any of the rent flows through to the providers of the infrastructure. A land tax doesn't add to the tenants' contribution; it is a deduction from the rent as received by the landlords, not an addition to the rent as paid by the tenants. The effect on the rent is caused by the infrastructure, not by the tax, and would be the same if the infrastructure had been financed in any other way. But if it is financed by a land tax, the bill is paid by the beneficiaries, namely those whose land increases in value.

Similarly, if land tax were the only tax, tenants who don't own land would contribute to the cost of government through their rents, but their rents would not be higher in consequence of the land tax; the tax would be a deduction from the rent, not an addition to it.

But whatever else happens, a land tax pressures land owners to generate income from their land, strengthening the bargaining positions of those who can supply that income (tenants and buyers) relative to those who need it (landlords and sellers), thus making access to land more affordable.

Unfortunately, those who wish to assert that land tax is passed on in rents have bigger campaign budgets than those who wish to tell the truth. To make matters worse, the assertion can be made in a 9-second soundbite, but usually cannot be refuted in a 9-second soundbite. Of course, if the landlords' representatives pretend that the tax is passed on in its entirety, one need only say: “If that were true, why would they bother complaining?!” But as long as they only claim that the tax is at least partly passed on, the claim is shorter and simpler than the rebuttal.

To get around this problem, those who are sympathetic to land tax as a means of making accommodation more affordable might consider implementing it in such a way that it is not payable when a tenant is present. Obviously you can't pass on a tax if it isn't payable; and if it is payable, you can't pass it on to a tenant who isn't there. Equally obviously, such a tax reduces rents because landlords need tenants not in order to pay the tax, but in order to avoid it, and will therefore accept a rent cut rather than lose the tenant and incur the tax.

The most obvious solution of this sort is to impose a holding tax on unoccupied property. To be effective, the tax should be levied on the land value alone, and should be payable whether the land has a building on it or not. Such a tax could not be avoided or reduced by demolishing buildings, or by leaving buildings vacant. But unfortunately it could be avoided by installing a token tenant in a token building on an otherwise grossly underdeveloped site. And there would be disputes about what exactly constitutes occupation of property.

A more subtle and seamless method is to deem every property to be yielding a certain minimum income for the purpose of (e.g.) income tax and GST. The minimum deemed income should be specified as a certain percentage of the land value per annum, and should be set as such a level that if the property is genuinely rented or owner-occupied, it will be yielding more than the minimum deemed income, so that the deeming has no effect. If there's a tenant, there's no effect to pass on; and if there's no tenant, there's no one to whom to pass on the effect.

I think the last sentence fits into a 9-second soundbite.

__________

* Last modified April 28, 2012.