According to Dwyer (2014), a land value tax should be based on the highest and best use. I interpret this as not taking into account the permissible use, meaning that people owning land in a zone that permits uses much lower than highest and best use face a significant disadvantage, because they would not be able to sell their land at a value compensating them for the land value tax. Here is an excerpt explaining it:
There has been some confusion in the literature on why such taxes cannot be shifted. The inability to shift a land value tax is not because it is a lump-sum tax (which it is not), nor because it is uniform in rate in all tax jurisdictions (which it need not be), but because the tax does not discriminate between different uses of land. The land value tax should be levied according to potential use, determined by the competitive market price, not based on its current use. A land use tax can be shifted, and so can a tax based on current rental income, which happens when assessors improperly assign a value to land according to its current use. In those cases, the tax can be shifted, and it is no longer neutral. One technical feature of land values that explains certain puzzles is that the optimal assessed value should always be based on the second-highest bid for a parcel, not the highest. The value of any site is always determined by the value that other users are deprived of. This guarantees that a visionary who sees a use for land not seen by others can still make an extraordinary profit from that insight. A final note about this section is the surprising conclusion that a revenue-neutral shift from existing taxes to land value taxes might not result in any lowering of land prices. That is, the general equilibrium result (the effect on aggregate land prices) might be the opposite of what is expected by a partial equilibrium result (the effect of land value taxes on a single site). Since tax capitalization works in both directions, the removal of existing taxes should raise land prices, and the imposition of a land value tax should lower them. The result should logically be little or no change in aggregate land prices. Of course, the price of some locations would rise and others fall, but the net effect should be no change, although there are complexities, discussed in Section 4.2. The net effect on land prices is important in considering questions of compensation, which are discussed in a later chapter. It means the transition to a land value tax system would be much smoother than has often been anticipated.
My question is, if LVT is assessed based on highest and best use (which it apparently need to be to ensure that efficiency conditions holds), doesn't that mean that planning gain (the gain in value from rezoning land from for example farmland to residential land) is not captured and would need a separate planning gain tax?
If LVT was based on highest and best permissible use, it would completely capture planning gain since the land value would increase, meaning the land value tax would increase meaning that this tax increase would be capitalised into a lower possible selling price through the net present value approach. Is this not ideal? Or did I misinterpret the highest and best use?
