Thursday 2 June 2016

Rents and GDP

In my (even if I do say so myself) excellent paper on our current economic predicament, available here, I briefly discuss the problems with including housing rentals in GDP.

When a building is built, the construction should add to GDP. This could count, for GDP purposes as consumption, or could be considered an investment and depreciated over several years of consumption. One house is produced, so the construction cost of one house should be added to GDP.

Instead what happens is on construction the house counts as investment and is added to GDP. Then all future rental paid (minus depreciation) counts as consumption and is also added to GDP in addition to the original building cost. This extra addition to GDP is simply the payment of a royalty to the owner of the land due to a state given monopoly right over this land; a transfer of wealth from non-landowners to landowners. It is not indicative of any productivity and should not count as the product of the country.

I argue below that since 2008, the fact that rental prices have gone up has led to a misstatement of real GDP growth. The official figures have total nominal GDP growth at 22.7%. I argue that a more correct value is 19.6%. In other words 2.9% of the UK's GDP recovery since 2008 is an accounting trick. In real (rather than nominal) GDP terms this means maybe a quarter of real GDP growth since 2008 is illusory, and a product of the transfer from those without property to those who own. All that had happened is that the royalty payments paid for use of land have increased. There is no increase in production.

The idea of imputed rents shows how ridiculous it is. If I buy a house, built 100 years ago, then should my simply living in it count as a part of the product of the country in this current year? If rental prices across the country go up does that mean that my house is producing more? It is a stock, not a flow. And yet by my simply living in the house, the imputed rent is added to GDP.

Why is this a problem? Because increases in rental payments are actually detrimental to production. As I discuss when talking about Osborne's zero probability of hitting his deficit target, rental and interest payments drain money from workers and those with a high marginal propensity to consume, and give to those with a low marginal propensity to consume. As described in this article, an economy paying high rentals and high interest payments, especially one where the corporate sector is able to pay high dividends, will much reduce disposable income to those who spend.

The reduction in demand for real goods and services actually genuinely produced in the economy is a natural consequence of an economy that pays too much to savers and not enough to workers and others who will consume more. And when rental payments are included in GDP, this is hidden.

What is the consequence of this? Using OECD figures (HT NEF), one can see that rentals in the UK have risen a lot recently. While the economy has nominally grown around 23% since 2008, rental payments have gone up by 53%. Rent and imputed rent is now 12.6% of GDP even though nothing is produced. In 2008 this figure was 10.2%. In the 1990s it was 8-9%.

And this means that an economy that has actually nominally grown by 19.6% in terms of productivity has official nominal GDP growth of 22.7%. While those who don't own property are being squeezed, the economy looks like it is doing OK thanks to the rent extraction that is actually damaging the economy.

Note that real GDP has only grown 8%.  Without knowing the construction of the inflation basket it is difficult to be exact, but it is possible that over a full quarter of the real GDP growth since 2008 comes only from rental increase.

This, in my opinion, is something that should have more attention.

NOTE - I previously wrote that counting housing as investment and then consumption is double counting. But thanks to Diane Coyle (who wrote the excellent book, 'GDP: A Brief but Affectionate History') for pointing out that housing investment is depreciated, so it is not officially double counting. However, the cost of depreciation is much smaller than that of the rental payments.