Tuesday, 30 December 2008

Taxes on rental values don't prevent property price bubbles

Many Land Value Taxers reckon that a high tax on annual site-only rental values would prevent bubbles in land and property prices. On the basis of actual evidence, I strongly suspect that this is not true.

1. I base this on the real-life example of Business Rates (also known as National Non-Domestic Rates), which are the closest thing we have to LVT in the UK (it's a progressive property tax with a few tweaks). Broadly speaking, NNDR are calculated as 47.1 per cent of the annual rental value of the whole property. In other words, if you are a tenant paying £10,000 rent to the landlord, you also have to pay £4,710 in NNDR.

2. Economic theory tells us that as the supply of business premises is fixed in the short or medium term but demand is price-elastic, the tax is borne by the landlord - in other words if it were scrapped, the landlord in this example would just increase the rent to £14,710, leaving the tenant no better or worse off. There are plenty of real life examples in the UK that bear this out, i.e. successive governments have declared certain areas exempt from NNDR from time to time, and all that happened was that landlords put up rents, or the market value of owner-occupied business premises increased.

3. The British Retail Consortium (who represent tenants rather than landowners) is perfectly aware of this phenomenom, which is why they recommended replacing NNDR with Land Value Tax in their submission to Sir Michael Lyon's Review of Local Government Finance at para. 6.43:

Land Value Tax (LVT) has a number of advantages. These include not distorting behaviour in the same way as taxes on income and profits do, LVT’s potential effectiveness in incentivising the efficient use of land (as all land would incur a charge even when it was not being used for productive activity) and taxing land values could also enable local governments to profit from some of the increase in value as a result of a prosperous local economy.

4. For simplicity therefore, let's assume that the landlord charges the tenant an inclusive price of £14,710 in rent and pays £4,710 NNDR. Unless the landlord is a tax exempt body (pension funds, Crown estates, CofE Commissioners etc) then there's income tax or corporation tax to pay as well, after deducting maintenance and interest costs, but let's ignore that for now. That equates to a tax of 32% on the total rental value.

5. The split between the rent paid for the location and for the building depends on what type of premises and where. In the case of used car lots or petrol stations, with very little in the way of buildings or improvements, NNDR is more or less the same as LVT, which is why (at least in East London), many of these sites were converted to residential use in response to the increase in residential property prices. Conversely, the split for a well maintained office block in a not-so-desirable area may be 90% for the building and only 10% for the location.

6. For simplicity let's assume that on average, one-third to one-half of the rent that businesses pay is for the location and the rest is for the buildings and improvement. Thus it wouldn't make much difference overall if NNDR were replaced with a tax of somewhere between 64% and 100% of the site-only rental value of each plot, payable by the landlord.

7. In the medium term, wages, business profits, share prices and rents all increase in line. Per Treasury figures (Excel, Table C4), over the last few years, total revenues from NNDR were as follows:
2001-02 £17.9 billion
2002-03 £18.5 billion
2003-04 £18.4 billion
2004-05 £18.7 billion
2005-06 £19.8 billion
2006-07 £21.0 billion
2007-08 £21.4 billion

which is a compound growth rate of 3% nominal (slightly less than I would expect, but hey).

8. Research by CB Richard Ellis (Powerpoint, Slide 16) shows that between January 2004 and July 2007, rents had risen by about 10% (also about 3% compound) but capital values had risen about 45% (about 10% per annum compound) - since when capital values have fallen by about 20%, but that's a different story.
9. To summarise so far, NNDR are pretty much the same as an LVT of between 64% and 100% of site-only location values, but they did not dampen the bubble in commercial property prices in the slightest; according to The Nationwide, the average UK house price increased by 'only' 37% over the same three-and-a-half-year period.

10. As we know, capital land values are merely a balancing figure between the value of a finished building and the cost of building (or replacing) it. It is futile to argue whether a bubble in property prices leads to a bubble in land values or vice versa. We also know that a half-way sophisticated investor looks at net present values and compares yields on different types of investments. We also know that property price bubbles and credit bubbles go hand in hand. For a constant annual rental stream of £10,000, if the landlord can borrow at (or earn interest on his cash at) 7% he is prepared to pay up to £142,857 for the building - but if interest rates fall and landlords can now borrow at (or only earn interest on his cash of) 5% (and expect this state of affairs to continue for the foreseeable future), he is now prepared to pay up to £200,000 for the same building. It is only mug investors who buy on the basis of expected capital growth.

11. Ignoring the 3% annual increase in rents and NNDR, it is therefore pretty much a truism to say that NNDR, as a fairly high % tax on location/land values has little or no dampening effect on bubbles in commercial property prices.

12. Which, to summarise yet again, is why I think NNDR (along with all other property- or wealth related taxes, such as Council Tax, Stamp Duty Land Tax, Inheritance Tax, Capital Gain Tax, TV licence fee etc) should be replaced with a tax on capital location/unimproved land values. Assuming that landlords and investors expect a return of 5% on the capital cost of land, a tax of 10% on the capital site-only land value would collect about two-thirds of the annual rental value thereof (the tax wouldn't apply to the buildings and improvements of course, for the reasons outlined by the BRC at point 3. above) and be a pretty straight swap.

13. Under such a tax, capital values could and would not increase by a ludicrous 45% in three-and-a-half years. Going back to CB Richard Ellis' figures (link above), for a £100 capital investment in UK commercial property in January 2004, rents would have been about £7.25 and the NNDR bill would have been £3.41. The LVT bill under the system I suggested in point 12. would have been £100 capital value less two-thirds relating to buildings/improvements = location value £33 x tax rate 10% = £3.33 (i.e. pretty much the same).

14. By mid-2007, capital values had risen to £145 and rents had risen to £8. Under my LVT suggestion, capital values simply would never have risen this far - LVT would have risen to £7.10 (£145 less indexed cost of buildings and improvements £74 x 10%). This would have made the return on investment negative (once other costs are taken into account) so the more sophisticated investors would have bailed out by 2005 or so. Yes, we know that there are mug investors who pile in just as a bubble is about to burst, who are prepared to accept negative rental yields in the hope of making capital gains, but firstly there's not much we can do to educate people like this, and secondly, the capital losses they will suffer from buying at £145 in 2007 are considerably higher than they would have been had the bubble in capital values been dampened by LVT and peaked at (let's say) £130 in 2005.

15. You may say "To hell with the property investors, why should I care?". It is important though - had the property price bubble been dampened then the credit bubble would also have been dampened, and the property price crash and credit crunch would, to some extent have been averted and we wouldn't now be staring at the wrong end of a recession, or even a Second Great Depression.

Here endeth.


Lola said...

For any quasi commercial propery - including BTL - capital value is a function of rental income. Your research that shows 10% rental increase vs 45% capital 'value' increase, shows incontrovertably just how far property values have to fall to get them back to fair value. It's even worse than that though because achievable rentals are also falling.

I feel a touch of schadenfreude coming on.

Mark Wadsworth said...

Only a touch? Not a Big Fat Slice Of, with Side Order Of Malicious Glee? That makes you a better human being than me.

Nick von Mises said...

Great post, as always.

Simon Fawthrop said...

I presume all these pre-pac (sp?) agreements that are in place before high street chains go in to adminitration include a large reduction in rents? This will drive capital values lower and faster than we have seen or expected.

Mark Wadsworth said...

It's called 'pre-pack'. Per the summary here, they aren't necessarily an opportunity to reduce rents, but they are an opportunity to close down unprofitable branches, which gives you a lever at least.

Lola said...

There must be a deeper, or perhaps longer term interaction between NNDR and rents.

Tax on yields, dividends, interest, rent, depresses retunrs. In other words the investor expects a certain return as compensation for the use of his capital and the risk to which it is exposed. In your example of rents rising by the same amount as the tax NNDR relief in designated special areas means that those investors are getting a windfall gain in rental income and capital growth in excess of the expected risk return for that class of property investment. Given time other investors would see these excess returns and move into the market (asuming that planning restrictions were flexible enough) and the supply would go up and the rental price go down.

LVT is still a capital tax and although seemingly better than the current system will depress returns and hence wealth creation as much as any other tax does.

One of its good points that leaps out of your analysis is that it would discourage wanton speculation and reward longer term professional investors, hence reducing the likelihood of bubbles.

But, LVT MUST be accompanied by national sound money policies. Officially sanctioned counterfeiting has to stop.

Mark Wadsworth said...

Lola, it is quite clear that taxes on income and production and output depress or deter economic activity. VAT and Employer's NIC are the worst of all because they have to be paid even if the business is making losses (unlike corporation tax).

But that's the point about LVT, however high the rate, it doesn't reduce the amount of land available, because land is in fixed supply. LVT quite probably would encourage more efficient use of land (see example of Harrisburg) so if anything LVT increases the amount of buildings (which would not be taxable, obviously).