For empiricists, the goal of Left Foot Forward to provide “evidence-based blogging” is admirable.

Partisanship can influence how we interpret statistics, and also lead to political claims that opponents are being misleading with statistics.

In an article posted on 26th February 2018, the title states:

UK hit by worst cost of living crisis in 50 years, new data shows — and the Tories have ignored it

The paper being cited is ‘A Democratic Measure of Household Income Growth: Theory and Application to the United Kingdom’, by Dr Andrew Aitken and Prof Martin Weale.

What the paper is about

This Economic Statistics Centre of Excellence paper looks at how real average equivalised household incomes are measured, and particularly what we mean by:

  • Real: what is the measure of inflation that is being used?
  • Average: what average of the equvalised household incomes is being calculated?

Commonly-used measures such as the Office for National Statistics’ Consumer Price Inflation are ‘plutocratic’ (essentially meaning they weigh money the same), rather than ‘democratic’ (weighing households the same). The ONS is starting to move towards democratic measures of inflation, through the experimental Household Cost Indices.

The slightly higher average growth rate in the democratic price index accumulates over the years.

For averages, the arithmetic mean is typically used, but the paper discusses using the geometric mean instead. The paper goes through the economic theory of these measures.

Finally, the paper compares growth in the democratic geometric mean household income with the plutocratic arithmetic mean household income. The authors find that, in the decade up to 2015/16 in the United Kingdom, the former measure grew by 0.20% per year and the latter by 0.52% per year.

Despite the title making about a claim about the last “50 years”, the cited paper does not look at this time period.

Other Errors

There are numerous other errors in the article, and its publication on social media too.

Real household income grew by only 0.2 per cent a year in the decade to 2015–16, rather than the 0.52 per cent shown in official figures

Both figures are types of real average household income: different choices are made regarding the calculated average and the inflation measure.

The government’s official statistics for measuring living standards are incorrect because they’re skewed in favour of wealthier households.

This is not the “government’s official statistics”, since it is calculated by the Office for National Statistics (as seen in their section on household incomes), which is operationally independent of the government.

Furthermore, it would be too harsh to describe the measure as “incorrect”, since different statistics have different uses.

Tories massaging the figures? Never…

Again, this accusation is nonsensical, because:

  • It is calculated by the Office for National Statistics, and not the government (wearing whatever rosette);
  • As the paper itself implies, calculating a mean average of household income using a plutocratic inflation measure is standard practice among statistical offices.

The question of whether it should be common practice, or if alternate measures provide a better summary of household income growth, was the entire point of the paper.

Partisanship can lead to strange places, such as accusing the government of ‘massaging’ figures produced by an independent statistics office, based on a economic statistics discussion paper.

This blog looks at the use of statistics in Britain and beyond. It is written by RSS Statistical Ambassador and Chartered Statistician @anthonybmasters.

This blog looks at the use of statistics in Britain and beyond. It is written by RSS Statistical Ambassador and Chartered Statistician @anthonybmasters.