Supporters of a mansion tax seem to have overlooked something - that such a tax would not be a tax upon the rich so much as upon the older rich.

I say this for a simple reason. The price of any asset should be equal to the discounted present value of the net benefits from it. Because a tax reduces the net benefits of a mansion, prices of them should thus fall. In other words, the tax should get capitalized into prices.

History suggests this is just what happens. Some research (pdf) at the IFS found that the replacement of rates by the poll tax in 1989 - the abolition of a tax on housing - caused a big rise in house prices. In inner London, an anticipated saving of ''471 per year on rates added almost ''6000 to the price of an average house (in 1985 prices).

That was not by any means full capitalization; the capitalization rate was only around one-third. But there are two good reasons why the capitalization rate might be higher for a mansion tax:

- In the late 80s there were doubts as to whether the shift from rates to poll tax would persist, given the unpopularity of the latter; such doubts were of course wholly correct. The risk of a reversion to property taxes would have dampened the boost to house prices. Because there's more chance of the mansion tax persisting - especially if it gets cross-party support - it should be more fully capitalized.

- There's more substitutability for mansions than for housing generally. A ''1.9m house is a decent substitute for a ''2m-plus one, whereas there are fewer close substitutes for home ownership generally. This means that potential buyers of mansions have the bargaining power to bid their prices down.

Insofar as the mansion tax is capitalized, it is present owners of them who lose two-fold; they have to pay the tax, but also see the value of their asset depreciate.

But house prices are not net wealth. A falling price of mansions is a benefit for those who would like to buy them. If the tax were 100% capitalized, the effect upon future buyers would be zero. Yes, they'd face a tax on their purchase, but they'd also pay a lower price for the Mansion. The two would net out.

And who are future buyers of mansions? They are the younger, slightly less rich - those who might well already own houses in the ''1-2 million bracket. Indeed, some of these could even benefit from such a tax. This would happen if the price of their desired house falls to just below ''2m, or if prospective buyers shift from ''2m to sub-''2m-priced houses, thus bidding up the price of the latter.

Now, I don't think all this is a clinching argument against a mansion tax. I favour it, insofar as it is a slippery slope towards a fuller, more intelligent, land value tax. But let's not kid ourselves that a Mansion tax is hitting the rich. Some of the rich might even gain from it.