The problem that is Imputed Rent and hence GDP

Over the lifespan of this website I have explained quite a few problems with our main measure of economic well-being and growth called Gross Domestic Product or GDP. This time I will focus on the problems and issues caused by a rarely discussed issue called imputed rent. This concept skulks away in the back ground partly because it is from the income version of GDP and the main figure is the output version. For those who are not aware of the state of play there are 3 ways of measuring GDP which are output, expenditure and income. Output is the most commonly used and if you here GDP mentioned then invariably that is what is meant but not every version is as for example Japan has a more expenditure based calculation.

There have been two roads which have led me to the Income GDP version. They are that the American numbers were a better guide post credit crunch to economic activity than the output version, and my interest in the housing sector reflected in this instance by the rent issue. Sadly such numbers are restricted access in the UK as a problem in particular occurred in the late 1980s under the then Chancellor but now Lord Lawson. In theory the 3 versions are supposed to come to the same answer but back then the variation was wide enough for him to order our statisticians to prioritise the output numbers and “adjust” the other versions. I can give you an example from Portugal of how the 3 numbers can vary as a while ago when I was looking at the data the divergence was 4%. It makes you think about those who discuss 0.1% changes does it not?!

What is Imputed Rent?

The story starts here.

In the national accounts, owner occupiers are deemed to be unincorporated businesses producing housing services, which they then consume.

Are “deemed to be”! So here is the first issue which is that it does not actually exist. After noting that let us press on.

The principle involved is to impute a rental value for an owner-occupied property, which is the same as the rental that would be paid for a similar property in the private rented sector. The imputed rent methodology calculates rent for owner occupiers and rent-free dwellings.

Why is this done. The US Bureau for Economic Analysis explains.

 The largest imputation in the GDP accounts is that made to approximate the value of the services provided by owner-occupied housing.  That imputation is made so that the treatment of owner-occupied housing in the GDP is comparable to that of tenant-occupied housing, which is valued by rent paid.  That practice keeps GDP invariant as to whether a house is owner-occupied or rented.

Their explanation is from 2006 when Imputed Rent was already 6.2% of GDP and the largest imputation which combined were 14.8% of GDP. It then argues this.

Without imputations, the GDP story is incomplete and can be misleading.

The other side of the argument is that including things which do not exist – owner occupiers do not receive Imputed Rent – is misleading.

Measurement of Imputed Rent

As it does not exist it cannot itself be measured and the only route to it is to measure actual rents. This poses its own problems in practical terms as this from the UK ONS demonstrates.

Imputed owner occupier rent is calculated from an average rent per room being multiplied by the total number of rooms in owner-occupied dwellings. Rent per room is calculated from Actual Rental (see section 02.4.1) and number of rooms rented (based on Living Cost and Food survey – LCF).

In the UK they will have some idea of the number of rooms but there will be errors in those numbers. However the main issue is whether we have numbers for rents which are reliable. I am sure that there are issues in every country but the UK has had particular problems and this is linked to my articles on the CPIH measure of inflation which includes rents. My view is that this has been a shambles illustrated by the way that the UK establishment had to abandon its rental estimates because they were in disarray.

You might think that a complete change to the actual rental numbers would have a big impact on Imputed Rent. In fact they seemed to sail through it pretty much unscathed as all sorts of other adjustments were made to provide the same answer. Or as Kylie would put it.

I should be so lucky
Lucky, lucky, lucky

As the luck quotient rose the credibility one fell.

Upwards Revisions

Back in the 2013 Blue Book the UK ONS decided the Imputed Rent numbers had been too low.

There are upward revisions to the level of total annual HHFCE (national concept) in all years from 1997 to 2011. The largest revisions, of just under 2% of total HHFCE, are in 2008 to 2011.

HHFCE is Household consumption and increasing it by 2% is a big deal and it was Imputed Rent that did it. Actually it more than did it as looking at 2010 will explain. UK household consumption and hence GDP rose by £17.1 billion of which the rise in Imputed Rents was £33.6 billion. The difference was a rise in estimates of repairs of £12.7 billion and some smaller items such as smuggling.

The New Economics Foundation weighs in

Just over a year ago the NEF gave an idea of scale.

Inclusion of how much home-owners would pay if they actually rented boosted UK GDP in 2014 by £158bn – a 8.9% share

We also got an idea of the scale of the housing and Imputed Rent boom.

A growing proportion of GDP is nothing more than earnings from property. 12.3%  of the UK’s measured GDP in 2014 was rent and “imputed rent”…….Since 1985, rent and imputed rent have almost doubled as a share of GDP, from 6.2%.

Ch-ch-changes

In the last few days and weeks the situation has changed again and let me show how.

these changes will have a substantial impact both on imputed rental itself and on total current price GDP.

Okay how? I summarised it thus on the Royal Statistical Society website.

For those who have not looked at the numbers then nominal UK GDP has been revised up by at least £50 billion in each of the years 1997 to 2006 due to Imputed Rent and then by a declining amount up to 2011. To give an idea of scale VAT fraud is considered a big deal but changes to it top out at £2.1 billion in 2011.

The official view on the changes is as shown below.

Although this improved the series for the most recent period, bringing it in line with the CPIH, it also led to a discontinuity (which has now been removed in the new method).

The discontinuity peaked in 2010 and I would tell you by how much but the link to the numbers on the official ONS site take you to a page which does not exist. Friday’s update tells us this.

In 2014, annual real GDP growth has been revised up by 0.3 percentage points from 2.9% to 3.1%,

Not the strongest grasp of mathematics there I think! Anyway there was yet another change to Imputed Rent as it added 0.1% to economic growth in that year (and in 2012 too).

Comment

You are perhaps waiting for an idea of scale so let me help out from the last quarter of 2015 when Imputed Rentals in the UK reached £43.2 billion in current price terms compared to £24 billion a decade before. That is a lot for a number which not only has theoretical issues in terms of its concept but the way we have tried to measure it has been very flawed as otherwise we would be needing all these “improvements” would we?! There was an obvious problem here in a nation the size of the UK.

The LCF data are based on around 400 households’ rental prices per quarter,

So whilst I welcome the efforts to improve the quality of the UK data on rents – which also feeds into the inflation numbers – there is a clear problem with what we have been told in the past. This feeds into less confidence in what we are being told now. At a time of house price booms this poses more than a few questions for the UK economic landscape and as for the Imputed Rent numbers well they continue to sing along with Jeff Lynne and ELO.

You took me, higher and higher
It’s a livin’ thing,
It’s a terrible thing to lose
It’s a given thing
What a terrible thing to lose.

Oh and this whole episode provides another critique to nominal GDP targeting.