Thursday, December 14, 2006
Inflation expectations - encouragingly low, or disturbingly high?
Posted by David Smith at 10:00 AM
Category: Thoughts and responses

The Bank of England has released its latest survey on inflation expectations. It shows that people believe prices have risen by 2.9% over the past 12 months and will rise by 2.7% over the next year. That is well above the 2% target but very close to the current CPI inflation rate. It is a long way from the 5%, 6% or 9% "true" inflation rates sometimes bandied around. Mind you, 72% of people expect a further hike in interest rates over the next 12 months.

Comments

I am a lowly engineer, so excuse me if my calculations below are incorrect.
My most major expense over the last 7.5 years has been paying off my 100,000 mortgage. It works out at around 260 per week (on top of the usual repayments).
The inflation figures only include the cost of servicing the mortgage interest payments.
If I were to purchase the same house now, my mortgage would be approx 200,000. So weekly payments would be 500 pm for 7.5 years. That equates to approx 13% per annum increase in my largest expense.
Hmmm, larger than 2.7% !
A large proportion of population are effectively working for the banks who are allowed to lend 5x earnings. This causes the "unrecorded" inflation, with seemingly little political will to change it.
Or am I just being interventionist ?

Posted by: Nick Thorne at December 15, 2006 09:16 AM

As for as the CPI is concerned, it's worse than you think - it doesn't even include mortgage interest payments. The RPI is better - it includes mortgage payments and should allow for the effect (larger mortgages) you describe, being based on the overall spending basket.

Posted by: David Smith at December 15, 2006 12:46 PM

RPI is fundamentally flawed as far as accurately measuring the impact of housing costs to the cost of living. RPI doesn't include mortgage payments, only mortgage interest payments, but, more so, it excludes the top 4% of households as measured by income. Consider this: the average house price in London is now 330,000 pounds. If one took out a mortgage for 90% of the purchase price, you would need a salary of 100,000 pounds to get the required mortgage, i.e. within the top 4% of households as measured by income. So RPI isn't evening including households that can afford to buy the average house in London. How is that an accurate measure of inflation?

Not that CPI is a better measure. The entire rationale for adopting CPI over RPIX as an inflation target by the BoE was to be consistent with the EU. This ignores the fact that the structure of the housing market in the UK is fundamentally different from other EU housing markets so that using the same methodology ends up measuring different things. Because mortgage interest payments are excluded from the CPI, the only place where major housing costs are reflected in the CPI is rent. 80% of British households live in their own home, so the rental component of CPI is based on only 20% of the population. However, more than half of renters do not live in private rentals, but in social housing which does not reflect market rents. This selection bias translates into an artificially low estimate of inflation. This is in contrast to Germany where more than 50% of households rent. A CPI calculation done in a country where 50% of households rent is not telling the public the same thing as a CPI calculation in a country where 20% of the population rents and more than half of those are subsidized so that they're paying roughly a quarter of comparable market rents.

Posted by: RichB at December 15, 2006 06:58 PM

Your points on the CPI are generally fair, though you're too high on owner-occupation (70%). You're unfair on the RPI. It includes mortgage interest payments on a typical repayment mortgage - the interest effect - and a depreciation component, the house-price effect. The fact that the RPI excludes the spending patterns of the top 4% of households is irrelevant. Apart from the fact that most such households will live in properties worth well over 300,000, so do many people on more modest incomes, who have traded up or just sat and watched their property rise. You can't just look at new entrants. A 100,000 earner with no housing equity would be highly unusual. In any case, what matters is the rate of change, for both these housing components.

No inflation measure can ever accurately reflect anybody's individual experience, let alone everybody's. As an example, the RPI includes mortgage interest payments and the depreciation element, but also rent. Most people don't rent and own at the same time.

Posted by: David Smith at December 15, 2006 07:30 PM

It is interesting to see how the housing weight in the RPI has increased. It was 72 (parts per 1,000) in the early 1950s, 158 during the house price lull of the mid-1990s, and is now 222.

Posted by: David Smith at December 15, 2006 07:35 PM

How is the housing weight determined? Surely the weight should not be arbitrary but reflect the volume of those particular transactions within the full spectrum of of trading in the reporting period?

Posted by: Mustafa Arif at December 18, 2006 01:48 AM

As with all components of the RPI, it is based on a survey of household expenditure. Even among owner-occupying households, there will be people who have zero or very small mortgages, right through to those with extremely large loans.

Posted by: David Smith at December 18, 2006 10:05 AM

That's fair enough for mortgage payments. But does RPI not take into account those who buy houses outright (or indeed the outright component of most mortgage sales)?

Posted by: Mustafa Arif at December 18, 2006 03:12 PM

Taking the RPI figures for 1996 to 2006 :
2.4,3.5,2.4,3.1,3.4,1.5,3,1.8,1.7,2.9,1.5

This is an 11 year period over which compound RPI gives an increase of 28.9% in total.

I quote from a recent BBC story: http://news.bbc.co.uk/1/hi/business/6084096.stm

"According to the Halifax, house prices have risen by an average of 187% across the UK since 1996.

The average UK house price has risen from 62,453 in the first quarter of 1996 to 179,425 in the third quarter of 2006 - an average increase of 10.6% a year, way ahead of income growth."

If the average house price had risen by RPI, it would cost 80,487. That is actually 97,937 less than the actual cost.

Therefore the buyer of the average house has to fork out an extra 97,937 compared to 11 years ago. Not to mention the hidden costs of the other extras, like life insurance, redemption payments, arrangement fees etc, and outright risk he/she is taking on against future interest rate movements.

The RPI weighting of 50/1000 points for mortgage interest equates to 5% of household income. That works out at about a 1:1 gearing ratio of household debt to income at a 5% interest rate.

If average household income is then say 30,000 pa. This means the average person has a 30,000 mortgage on a 179,000 house. Which just doesn't sound right.

In my experience, a graphical plot of "number of households" vs "debt levels" would be a bathtub shape, with a large number (I think I read it is 47% or so) of households with no mortgage debt. Then falling to a low number of households with moderate debt, then a large number with massive debt. Several of my friends (thirty somethings) are up in the 250,000, 400,000 and even 500,000 mortgage levels.

So, the RPI treatment of house prices really is giving a free lunch to half the population and crucifying the other half. If mortgage debt levels were a Gaussian distribution, it would be OK, but I don't think this is so.

When the social inequality grows too strong, I guess some political force will eventually come to bear on the issue, as exemplified by the current renewed debate on the green belt.

Posted by: Nick Thorne at December 21, 2006 08:14 PM

I have a more broad question regarding CPI as well as RPI. I am currently a high school student participating in an internship where analyzing thses two figures is important; however, I am a bit confused. Britain still publishes both the CPI as well as the RPI, so which is a better measure or more comparable and why? Furthermore, EU harmonized CPI comparison, so it one a standard but not as good in the UK for some reason? Any assistance that can be provided will be greatly appreciated.

Posted by: Zach Garippa at March 10, 2008 09:01 PM

As another angle on house price inflation, how do mortgage repayments compare over the years as a proportion of average household income since, say, 1996? This would provide an indication of how difficult payment are for many households, and partly why so many need two earners.

Posted by: Graham Miller at July 5, 2009 06:59 PM