17 Comments

• CP 10th Nov '12 - 3:35pm

  How odd that the Office of National Statistics has used its own made-up measure of inequality, rather than the international standard (the gini coefficient) used by organisations from the OECD to the World Bank.

  The world does look nicer through rose-tinted glasses, but I thought the ONS would give clear information.

• CP 10th Nov '12 - 4:24pm

  The data is also very skewed because
  – it uses the Consumer Prices Index (CPI) as the measure of inflation, which doesn’t include housing costs , rather than the RPI. Housing costs have risen fast and have taken an increasingly large part of peoples wages, so it is very important to talk about RPI in a discussion of inflation and wages.
  That is one of the reasons the LibDems had such a strong focus on housing at the autumn conference.

  – it doesn’t include part time work, or full time workers who have had hours cut.
  – Does this data take into account the increasing problem of wages that are below minimum wage, such as unpaid internships?

• Peter Andrews 10th Nov '12 - 7:06pm

  The use of CPI rather than RPI does unfortunately render all of the rest of the data useless

• Mark Inskip 10th Nov '12 - 7:15pm

  @CP the different between RPI and CPI is not as straightforward as you claim. Its not that CPI excluding all housing costs, it excludes owner-occupier housing costs, but don’t forget that ONS is also developing a version of CPI which includes onwer-occupier housing costs because of the other issues with the RPI measure.

  For example;
  CPI includes all households whereas the RPI leaves out the richest 4% and pensioners who rely on state benefits for at least 75% of their income.

  CPI generally uses the geometric mean whereas RPI uses arithmetic averages . As a consequence CPI better captures the behaviour of consumers who buy less of a product when it gets more expensive. And that is why arithmetic averaging is not permitted under European rules for the CPI on methodological grounds.

• Mark Pack 10th Nov '12 - 8:11pm

  Peter: I found Tim Harford’s piece on the flaws in the way the RPI is calculated rather persuasive on the flaws in using the Carli index approach for the RPI (still on iPlayer http://www.bbc.co.uk/iplayer/episode/b01lv7yn/More_or_Less_The_great_playing_field_sell_off/ starting at c.19:15 in). There’s also more at http://ftalphaville.ft.com/2012/09/18/1166061/rip-old-rpi-and-hello-a-happier-chancellor/ – though I’ve not fully digested that piece.

  As you’re so firm in your view that CPI is inferior to RPI, I wondered what your view is on this issue and why you think the RPI version is better?

• Tim Leunig 10th Nov '12 - 8:36pm

  If only all statistics were presented this clearly! Of course there are other things to look at – hours worked, for example, number of workers per household, but I thought this was a very well presented explanation of what has happened to full time wages in this period.

  (Nb If apples go from £1 to £2 a kg, and then back to £1, the RPI shows that prices are higher at the end than at the beginning. That is clearly wrong.)

• Matthew Green 11th Nov '12 - 11:35am

  The political message is that, based on the measures used, wage inequality grew most seriously in the Tory years 1986-97. The Blair/Brown years more or less held inequality the same, with improvements at the low paid end thanks to the minimum wage. This diverges from Lib Dem political rhetoric a bit. It would be interesting to compare to the US.

  I’m glad to see from the comments that the serious flaws in the methodology used for RPI are starting to creep into wider political awareness. This measure really must be put into the dustbin – I undersand that its distortions are getting worse. But if the same measure is being used across the whole spectrum, it shouldn’t affect measures of inequality. Or have I missed something?

• Peter Andrews 11th Nov '12 - 12:39pm

  Ok so having listen and read Mark’s links it seems there are serious flaws to RPI that I was not aware of but CPI does not fully reflect increases in housing costs which given housing costs are a major area of household expenditure means using CPI for these figures is not satisfactory either. It is misleading to say that wages have on average risen 10% in real terms (made up figure) if housing costs have risen 30% and this more than offsets the increase in wages as in that case overall people would actually be worse off but these figures would not reflect that.

  So it seems that unfortunately we are left in a situation where we do not actually have a satisfactory inflation measure which fully includes housing costs

• Mark Inskip 11th Nov '12 - 1:47pm

  @ Peter Andrews
  Its not that CPI doesn’t including any housing costs. It includes rental costs but does not include owner-occupier housing costs. The ONS has already announced that from March 2013 it will also start to report a new index CPIH, intended to incorporate owner-occupier costs.

• Peter Andrews 11th Nov '12 - 4:32pm

  I am aware of that and its a pretty big hole in a measure of inflation given how much house prices and therefore mortgage costs have increased over the last 3 decades.

  Anecdotal I know but my Dad bought his house 30+ years ago when he took a job as a lecturer at Leeds University. There is zero chance of someone taking the same job now being able to afford to buy a similar house in the same area.

• Mark Inskip 11th Nov '12 - 8:32pm

  @Peter Andrews
  If you look at the Halifax’s housing affordability index then for Q3 2012 it was 4.04, 25 years ago in Q3 1987 it was 3.92 which is only a 3% difference. In between there have been some significant peaks, e.g. 4.99 in Q2 1989 and 5.86 in Q2 2007 and troughs such as 3.11 in Q1 1999. However the BoE base rate was over 5% in 1999 and is only 0.5% today. Hence the changes in affordability aren’t as clear cut as you anecdotally observe.

• CP 11th Nov '12 - 10:39pm

  Mark – its correct that the CPI underestimates price changes. Therefore if you are trying to compare it to wages it is even more important to use the RPI. If you can no longer buy a car because they are too expensive, you cannot say that your standard of living is similar. There are some problems with the RPI too.

  Housing has got much more complicated. Now almost 50% of 1st time buyers are using shared ownership schemes (part rent-part buy). If CPI only includes rental costs, it will underestimate those housing costs by half. nb. not sure if/how the rental costs are included for CPI.

  The problem with the measure of inequality used here is that it is based on only 2 data points (top 1% and bottom 1%), and these are at the 2 extremes of the data. It ignores 98% of the population! (whereas the gini coefficient takes 100% of the population into account.).

  I agree that the presentation is very clear. However the data is certainly *ahem* skewed. Thanks for posting this Mark , it has generated a good debate!

• Mark Inskip 12th Nov '12 - 8:13am

  @CP
  ” its correct that the CPI underestimates price changes. ”
  That is not correct. CPI excludes owner occupier housing costs, CPIH intended to address this point.

  As Tim as pointed above, if apples go from £1 to £2 a kg, and then back to £1, the RPI shows that prices are higher at the end than at the beginning. CPI does not. RPI therefore gives a higher inflation figures than CPI for apples in this instance. Are you claiming CPI is underestimating price changes in this instance and that RPI is correct?

• Mark Inskip 12th Nov '12 - 8:15am

  @CP
  “Now almost 50% of 1st time buyers are using shared ownership schemes (part rent-part buy). ”
  That’s not correct.
  Last week Lloyds TSB press released a survey in which they claimed “Almost half (46%) of first time buyers are considering turning to shared equity and shared ownership schemes to help them onto the property ladder” on the back of an announcement of new shared equity and shared ownership products they are launching.

• Steve 12th Nov '12 - 9:21am

  Halifax’s affordability index is meaningless as it doesn’t take into account wage inflation.

  If you keep receiving a large nominal increase in pay every year then the gross percentage of your income that pays the mortgage over the lifetime of the mortgage decreases fairly rapidly. I would much rather have bought a house in the ’70s when interest rates and wage inflation were high rather than now .

  There is no historical positive or negative correlation between interest rates and house prices for this reason. The best way to measure affordability, therefore, is to simply relate the average house price to the average wage and on such a measure houses are currently overpriced. Hardly a surprising result given prices keep falling (unless you’re living in prime central London).

• Mark Inskip 12th Nov '12 - 8:50pm

  @Steve
  You state;
  “Halifax’s affordability index is meaningless as it doesn’t take into account wage inflation.”
  then you state;
  “The best way to measure affordability, therefore, is to simply relate the average house price to the average wage”

  The Halifax’s affordability index is calculated by dividing the Halifax seasonally adjusted standardised average house price by average earnings.

  How can it be “meaningless” and then “the best way”?

• Steve 14th Nov '12 - 1:15pm

  @Mark Inskip

  Apologies. My statement “Halifax’s affordability index is meaningless as it doesn’t take into account wage inflation.” should have read referred to your definition of affordability.

  I include House Price, Average Wage, Interest Rate and Wage Inflation, which effectively reduces to House Price and Average Wage if the effect of interest rates and wage inflation cancel each other out.

  You include House Price, Average Wage and Inflation without taking into account the erosion of debt by wage inflation.