Ask the experts: Inflation metrics

Christopher Payne, a statistician, and head of consumer prices development at the Office for National Statistics, delves into the composition of inflation metrics and asks if we can really trust the data
by Dan Atkinson


About the expert


Christopher Payne is head of consumer prices development at the Office for National Statistics. He graduated as a statistician at the University of the West of England, and has a masters degree in statistics from the University of Southampton. He joined the ONS in 2012, and has been involved in price research since 2014, with special reference to inflation.

Quiescent for decades, inflation is once again at the top of the Western economic agenda. Governments have resorted to measures long thought discarded or discredited, not least huge subsidies to hold down energy bills. But can we trust the figures we are given? And what are they really telling us?

There are several inflation measures, both in the UK and internationally. Items are included or excluded, different weights are assigned to different elements, and so forth. Of all these measures, which is considered the most accurate?

This is really a question of what you are trying to find out. Accuracy refers to how well a particular metric measures what we are looking at. The Consumer Prices Index (CPI) in the UK follows international guidelines. EU countries produce an equivalent measure to CPI, as does the US Bureau of Labor Statistics. The most comprehensive measure in use in the UK is CPIH, which takes owner-occupiers’ housing costs into account. Because it is difficult to harmonise the measurement of these costs across countries, individual jurisdictions tend to develop measures to suit their own conditions.

Are there ongoing developments in terms of constructing new measures of inflation to sit alongside the established metrics?

Yes. We are working on what we call Household Costs Indices (HCIs). Put simply, this adjusts inflation calculations to take account of how people are paying for items in the index. Take student tuition fees, which most people pay through a student loan, so they do not impact household finances for quite some time after an individual starts working. Therefore, price changes would be experienced by households in a different way than that expressed by the CPI, for example. The HCIs project aims to look at the effect of inflation on different types of households. Then there is the Personal Inflation Rate calculator, available in its present form since August 2022 on our website, enabling people to work out, from inputting their spending patterns, what their own inflation picture looks like. These are complementary measures, not intended to replace the ones we have.

How important are housing costs for measuring inflation across the globe?

They are certainly significant in the UK, at 17% of the CPIH basket. Internationally, it varies a great deal. Eurostat in the EU doesn’t measure housing costs. One problem is that housing costs represent two things – the purchase of a service, i.e., somewhere to live, and the purchase of an asset. We try to separate the two.

To calculate the service element, we take the approach of working out what the equivalent rent would have cost, as do the US and the Netherlands, for example. Some countries use the simpler approach of tracking mortgage interest payments, but the problem here is that when interest rates are raised to tackle inflation, the inflation rate goes up. In Australia and New Zealand they use a ‘net acquisition’ approach that separates the cost of the land, which is taken as the asset value, from the cost of the dwelling, which represents the consumable element. Elsewhere, countries have measures that look a lot like CPIH.

How legitimate are measures of ‘core’ inflation that exclude volatile items such as food and fuel? Isn’t this just a way of ignoring items that are likely to rise in price, or is that overly cynical?

I think it probably is! As with all inflation measures, it comes down to what you want to look at. Core inflation is about examining longer-term trends in prices. There are different approaches to calculating core inflation. Some, as mentioned, omit volatile items. Another approach is ‘trimming’, which involves taking out the outlying price movements either up or down. Whatever approach is taken, it is important to remember that a core inflation measure is just a supplementary tool and does not replace the main metrics.

So-called hedonic adjustments take account of product improvements, thus lowering the rate of price increases for those products from what it could have been without the adjustment. An example is the modern television, with many more features and higher quality than those available even two decades ago. Is that right?

Yes, the hedonic regression is a statistical modelling technique that allows us to control for changes in the quality of high-tech goods in particular.

Is it not sometimes argued that there is no corresponding adjustment when product quality has declined, as it has arguably with flimsier refuse sacks, thinner shirts and trousers, and less robust footwear, to give just a few cases?

The short answer is that such adjustments are indeed made. Our price collectors who compile the statistics track quality changes in the items in the ‘basket’ that make up the inflation numbers. As mentioned, the improving quality of laptops, for example, is reflected in the adjustment to the effect of price changes so that, to an extent, the later models become different products. In the same way, our quality adjustment procedures can reflect declines in product quality, taking account of the fact that people are, in effect, paying more than they were for their refuse sacks, among other items. We can look at product quality and adjust for it accordingly. We adjust in a consistent, objective way, regardless of the direction of quality change so that we don’t bias the index one way or the other.

Are there any noteworthy developments in terms of data collection?

Yes, there is transformational work under way in the UK in terms of data sources, work that is of interest to statisticians in other countries. We are working with retailers to get access to their transaction data. At present, out of the basket of around 730 items, around 550 are collected manually, in effect, by sending our price collectors to gather the information. Working with retailers, we will be able to use barcode data giving us the actual prices paid and the number of items bought. It’s not only retailers, as we’ve also acquired data from AutoTrader and the Rail Delivery Group. It’s a very large research project involving new systems. But with new datasets come new problems. It is much harder to maintain a consistent sample over time with this sort of information source. Consumers need not worry that their shopping habits are being passed on to us, as individual data is not provided to us, and the identity of the retailers concerned will be protected if they wish. We hope to introduce some elements of this new data collection method this year, moving on to groceries next year. We have other, smaller-scale projects in hand to improve how we measure inflation, but this is the most important.

In global terms, how close are we to an international measure of inflation that is comparable across national frontiers? Does the International Monetary Fund (IMF) or any organisation calculate such a measure?

The EU has a measure covering all member states called the Harmonised Index of Consumer Prices (HICP). The HICP for each country is calculated by member states and compiled into an EU aggregate by its Eurostat branch based in Luxembourg. A single measure is needed not least because eligibility for membership of the single currency, the euro, depends in part on a track record on inflation that is standardised across the EU, thus comparable from one country to the next. Britain, then an EU member, adopted HICP, which it called the Consumer Prices Index, in 2003. The International Labour Organisation publishes an international manual for the collection of inflation data, but these are guidelines. The IMF and the 38-strong Organisation for Economic Co-operation and Development (OECD) jointly collect inflation statistics, the OECD from its own member countries, along with those in the accession process and partner countries, and the IMF from everyone else. These are CPI statistics, but the collection of the data is the responsibility of individual governments. In terms of an international, uniform inflation measure, perhaps collected directly by an IMF-type body, I am not aware of any work on such a project.

Even so, the IMF/OECD publications use the expression ‘Consumer Price Index’, and both organisations speak of coordinating and harmonising data. Isn’t that a step towards a worldwide inflation measure?

No. CPIs are in common use as a measure of price changes for goods that households buy or otherwise acquire to be used, directly or indirectly, to satisfy their needs and wants. But the construction of the indices, and the collection of the data, sits better with the individual nations rather than the IMF or OECD. As the IMF puts it, the information collected, which is a valuable course for data users throughout the world, is made possible by the cooperation of the countries concerned.

The EU has a role in determining the harmonised methodology for member states to follow (and it draws on input from member states).

For many years, the UK relied on a measure called the Retail Prices Index (RPI). Why was that abandoned and did other countries have anything similar?

In short, it was abandoned in favour of CPI because of significant shortcomings in the RPI’s methodology that led to an upward bias in the figures. It had already had to be refined when it was used as the inflation target for the Bank of England, with a supplementary measure called RPIX, to exclude mortgage interest payments. It lost its ‘national statistic’ status in 2014, but it is still collected – as we are legally obliged to do – given some payments and contracts are still calculated using the old measure. In terms of other countries, I suspect you would have to go back some way to find other countries using anything like RPI or the method it uses to average the raw price quote data, as we were the only nation still employing it before it was retired.

Doesn’t that make comparisons more difficult, both from one period to another and in terms of gauging inflation internationally?

While a continuous measure would be preferable, the CPI was introduced in 1996 and we don’t have the raw data to construct a CPI series back further. But it is possible for us to model CPI for, say, 1975 – the post-war peak year for inflation – to give a direct comparison with today.

The Royal Statistical Society states in a press release that “CPIH is a macro-economic indicator that is good for gauging the general performance of the economy, while RPI is intended to reflect changes in the cost of living.” For this reason, it says, RPI should be retained. What is your view?

All price indices – whether macro-economic or household – reflect how prices are changing. The RPI is a deeply flawed measure, often significantly over or underestimating inflation, and for this reason we discourage its use. For those wanting an index showing payments by UK households, the ONS is developing Household Cost Indices, which show how prices are rising for households, without the flaws found in RPI.

Seen a blog, news story or discussion online that you think might interest CISI members? Email
Published: 07 Mar 2023
  • Risk
  • International regulation
  • Compliance
  • Wealth Management
  • RPI
  • IMF
  • CPIH
  • Consumer Prices Index

No Comments

Sign in to leave a comment

Leave a comment