Sunday 11 March 2012

Fractals

There are some cunning formulas in maths which repeat themselves again and again, if you plot a graph of the results, it does not matter how much you enlarge it, the area always looks the same.

A similar story applies to UK land vales, taking as a proxy the average recent selling price of a semi-detached house in 2,630 postcode sectors (excl, the top five per cent). The ratio between the most expensive and the cheapest is more than 5:1; (if we included the top five per cent the ratio would be 50:1):If you take larger postcode districts, such as Birmingham or Leeds, you get a similar picture; but the ratio between the lowest and the highest prices is much smaller at about 4:1:This is as to be expected for large towns. The gradient for smaller towns (Swindon, which is the median town in so many respects) is much flatter. Here is the chart for all 18 inhabited Swindon postcode districts, the ratio here is only about 2:1.The ratio
six postcode districts which make up Swindon urban area can be split up into nineteen postcode sectors (these are roughly the same size as a local council ward), if we compare these, the ratio is 3:1.If we take the median postcode district within Swindon urban area, which is SN5, and plot a chart of the six postcode sectors within it, the chart is relatively flat with a ratio of 1.3:1:So the smaller the area we are looking at, the flatter the ratio between the most expensive and the cheapest.

No doubt you could sub-divide any one of those postcode sectors and you might find that the price ratio between the nicest streets and the least desirable streets is 1.5:1; and that on any particular street, the ratio is also 1.5:1 and so on, ad infinitum, but adjusted for the fact that more expensive house = bigger garden, the underlying land values expressed per square yard within any postcode sector would be almost flat.
------------------------------------------------
All of which leads me to believe that a postcode sector (or a local council ward, which is the same size) is a fairly optimal area for establishing [relative] land values for LVT purposes. Provided the tax is set at no more than 75% of the average rental value within that very small area, the least desirable couple of streets would be paying (say) 90% of the rental value in LVT and the lucky owners of the few most desirable plots would be paying (say) 60% of the rental value in LVT, which is perfectly acceptable for tax raising purposes.

So be it. There will always be winners and losers in the great lottery of land values, and remember also that marginal rates of income tax in the UK are anything between 0% and nearly 100%, so a range between 60% and 90% is pretty narrow i.e. accurate.

2 comments:

Anonymous said...

It doesn't matter for your post but...

Fractals are not (necessarily) the same as you zoom in. They are (or can be) self similar -- in particular places. What is magic about them is that they contain infinite detail. That detail is often self-similar.

A great many curves other than fractals are self-similar. Exponentials are the most relevant one for your post though.

Or that other famous self-similar curve... the straight line :-)

Mark Wadsworth said...

OP, yes, higher maths at an abstract level is not my forté. The only other example I could think of was holograms. Allegedly, if you cut one in two, you end up with two smaller whole holograms and not two halves of one picture. But that wouldn't have been such a cool post title.