Monday, 4 November 2019

Killer Arguments Against LVT, Not (473)

Physiocrat reminded me to dismantle the Scottish Land Commission - Investigation of Potential Land Value Tax Policy - Options for Scotland - Final Repor.

A lot of it is fairly positive, but they are in a muddle on valuations, making it seem far more difficult than it really is, probably deliberately or perhaps out of intellectual laziness.

From page 30:

Land needs to be valued. This should ideally be undertaken using the comparison approach i.e. by analysing market evidence of comparable land sales. However, evidence of undeveloped land may be scarce.

The alternative is to use an approach whereby evidence of the value of land and buildings sold or rented as an ‘entity’ is analysed to extract the value of the land.

This is not 'an alternative', this is how it's done. The site premium of the few plots of bare land in urban areas is inferred from this (assuming in same area with same planning).

Undertaking this can be problematic as the ‘residual’ method, whereby build costs and other adjustments are subtracted from the total value of the development to arrive at a ‘residual’ land value, can produce confounding results...

We should be using rental values, not selling prices, but for the initial valuations, it's good enough.

... For example, take two dwellings side-by-side. One is three-storey and developed to highest and best use (market value = £1m, build and other costs = £0.5m, so land value = £0.5m), the other is two-storey (market value = £0.7m, build and other costs = £0.3m, so land value = £0.4m).

The land value (and therefore the LVT) of the first property is higher. The relationship between property value and build cost is penalising the development of land to highest and best use, which is counterintuitive as far as a land value tax is concerned.

Initial valuations can be based on actual use, as this is a good proxy for 'optimum permitted use'. At the time any building was built, it probably was the optimum permitted use.

That might change over time, but only gradually; if nobody's applied for change of use on his own plot, we can reasonably assume that the current owner considers this still to be the optimum. If somebody build a larger home ab initio; or bought a smaller home, knocked it down and built something bigger; or bought a smaller home and built an extra storey, that is clearly his opinion of optimum use.

Whose opinion is a better indicator? Nobody knows. Does it matter? No. Valuations are always going to be a bit arbitrary, what matter is consistency of approach.

1. We shouldn't ignore the past and what's already there, so we could just assess the smaller house plot at £400,000 and the larger house plot at £500,000 and have done with it.

Similarly, we shouldn't ignore demolition costs and hassle (as the report does) and practicalities. What is the optimum permitted use of the smaller home? It's probably "leave it as it is".

If you want a larger home on that street but none are for sale, you could buy a smaller one for £700,000, knock it down at a cost of £50,000; build a larger home for £500,000, you've ended up spending £1.25 million for a home worth £1 million, which ain't going to happen. So £400,000 is still a fair assessment.

2. Or maybe you can just buy a smaller home and spend £150,000 on an extra storey; half of the value uplift from £700,000 to £1 million. In which case, if it's a mixed street with equal numbers of easily extendable smaller and larger/already extended homes, it makes sense to assess them all at £450,000. If there are ninety smaller homes and only ten larger ones, assess them at the weighted average of £410,000.
Whichever method is chosen, it will be good enough for initial assessments (I am heartily indifferent), and the % rate would simply be based on the total revenue required to replace existing taxes are replaced, so not huge £££ amounts. Over time, we can be a it more sophisticated; use rental values not selling prices, and so on.


Physiocrat said...

One way to discover residential land rental values is to use web sites like Zoopla and find the 'floor level' rents. These will be in places like the north-east; a three bed semi at Seaham, County Durham, is about £650 pcm. A similar property in BN2 is around £1500 pcm and London NW4 around £2000. A reasonable assumption would be to take the £650 pcm figure as a base level to cover the cost of supplying the building, with anything in excess being location ie land value.

The £650 figure seems to be widespread; it applies in such diverse locations as Ardrossan, Carlisle, Merthyr Tydfil - residential land values can be assumed to be zero over a large tract of the country. When different searches are done, location value starts to appear at places such as Truro, Llandudno and Swansea which are desirable for one reason or another.

Given the size of the residential and commercial rental markets, there is a large and solid body of rental evidence these days.

Using rental values avoids several problems which apply when assessments are on selling price. The first is that LVT is perceived as a wealth tax, which it is not, since land is not wealth. The second is that land prices are reduced by an amount equal to the capitalisation of the tax payable; the tax cuts into its own base, which is like sitting on a log and sawing it off the main trunk. This is an important issue at the initial valuation, because land in commercial use is subject to the UBR, which is several times higher than land in residential use subject to Council Tax. This leads to an apparently low figure for commercial land values, from which the conclusion is drawn that a high proportion of the LVT burden would fall on house owners.

A third difficulty is the fact that land price may include an element of hope value on the assumption that consent for development could be granted in the future. In the case of rental value assessments, that value is the value today ie it is current use value or the value an the assumption that the site has been developed in accordance with any planning consents already granted; there is no doubt on the matter.

The Scottish report, which uses the word 'challenging' thirty times in 119 pages, provides an excellent set of arguments against the use of selling prices for LVT assessment.

Bayard said...

£650 pcm is a bit high for a base value, e.g. Merthyr Tydfil is within commuting distance of Cardiff and Seaham benefits from being by the sea. You can rent a three-bed semi in Neath for £550 pcm, which I'd suggest is a better base figure.

Physiocrat said...

@Bayard. Perhaps. It needs local knowledge. It would be a worthwhile exercise to assemble the information on a map.

Mark Wadsworth said...

Ph, agreed.

B, I'd go with £550 pcm baseline. You can find cheaper than that, in areas where location value is negative.

Ph, Anthony Molloy and I did all this for a sample of 2,000 postcode districts. Easy to establish that baseline is £550-ish. As s national tax, that's the only figure you need.

benj said...

Our present tax system is so far off an ideal land tax, the commissions concerns are a beating up a strawman.

I believe they had an excellent submission showing how reformulating Council Tax gets us most of the benefits of a proper LVT, without fussing about valuations etc.

Once that was in place, we could, if we wished, worry about the cost/benefits of getting nearer to an ideal land tax.

Mark Wadsworth said...

B, yes they did. But whoever wrote it was clearly biased, so they ignored it.