Beyond the average: patterns in UK price data at the micro level

Lennart Brandt, Natalie Burr and Krisztian Gado

The Financial institution of England has a 2% annual inflation fee goal within the ONS’ client costs index. However its 700 merchandise classes, we discover that only a few costs ever change by 2%. In actual fact, on a month-on-month foundation, solely about one fifth of costs change in any respect. As a substitute, we observe what economists name ‘sticky costs’: the value of an merchandise will stay fastened for an prolonged period of time after which alter in a single massive step. We doc the time-varying nature of stickiness by wanting on the share of worth modifications and their distribution within the UK microdata. We discover a seen discontinuity in price-setting within the first quarter of 2022, which has solely partially unwound.

Idea of sticky costs and associated literature

Understanding price-setting dynamics is crucial for central banks. Most structural fashions within the literature use a type of time-dependent pricing, below which companies hold costs the identical for fastened quantities of time (Taylor (1980)), or for random quantities of time such that there’s uncertainty concerning the exact size of the value spell (Calvo (1983)). One other approach of modelling sticky costs emphasises that companies is not going to simply have a look at the time that has handed since they final adjusted its worth, but in addition at how far their worth is from some desired worth degree. That is referred to as state-dependent pricing. Macroeconomic fashions don’t sometimes enable for time-variation within the diploma of stickiness or switching between pricing methods. Just lately, nonetheless, companies within the Resolution Maker Panel inform us that they’ve moved more and more away from time-dependent in the direction of state-dependent pricing. On this case, when there’s a massive shock affecting many companies, the shock results in an elevated frequency of worth modifications and so extra speedy pass-through to total inflation.

With a view to higher perceive the pricing behaviour of companies in occasions of huge inflationary shocks, we discover the pricing dynamics on the micro degree utilizing CPI microdata revealed by the ONS. We’re after all not the one ones who’ve been thinking about this kind of knowledge. Financial institution authors have been utilizing this knowledge set for various years. For instance, Bunn and Ellis (2011) doc stylised info about pricing behaviour from the UK microdata and the August 2020 Financial Coverage Report used CPI microdata to tell coverage. Elsewhere, Karadi et al (2020) use US microdata to analyse companies’ price-setting in response to modifications in credit score circumstances and financial coverage. Nakamura et al (2018) analyse the societal price of excessive inflation utilizing microdata from the Seventies and Nineteen Eighties, and Montag and Villar (2023) analyse the impact of extra frequent price-changes on mixture inflation throughout Covid. Relatedly, Davies (2021) finds that the distinction between the share of worth rises and worth cuts within the UK microdata is said to mixture inflation, specializing in price-setting in the course of the pandemic. And at last, authors of the FT’s Alphaville weblog have additionally been wanting into these knowledge (see right here and right here).

The information

The microdata spanning from 1996 till September 2023 is publicly obtainable and up to date month-to-month after every CPI launch. It accommodates the month-to-month worth quote knowledge underpinning the ONS’ CPI collection for over 700 gadgets with identifiers on the store, store kind, and area ranges. We clear the information which works out to about 30 million observations. When figuring out a worth change within the knowledge, which is in the end what issues for inflation, we attempt to be as exact as attainable almost about the product and the timing of the change. To that finish we solely depend the change if we discover the identical merchandise, in the identical area, in the identical store, in two precisely neighbouring months. For instance, if a bag of potatoes price £2 in January and £3 in March however was not recorded in February, slightly than imputing a worth we discard the remark since we can’t be certain through which month the change truly occurred.

Stylised info from the microdata

A quick have a look at the information lets us set up some stylised info. Chart 1 exhibits a decomposition of those month-on-month worth actions over all gadgets within the knowledge set. 4 key observations stand out:

Costs rise and fall on a regular basis, however the overwhelming majority of costs don’t change between months. In any given month, on common since 1996, round 80% of costs stay unchanged relative to the earlier month (blue line).

The share of costs rising (in inexperienced) has elevated notably since 2021 to an extent that has not occurred in earlier inflationary episodes within the pattern (excluding VAT modifications).

The share of costs falling (in crimson) has fallen considerably however stays steady since 2021, relative to historic common. The principle margin of adjustment has been within the share of worth will increase.

However, in latest months, whereas the share of costs rising has tapered off, it stays elevated relative to its historic common. 

Chart 1: Decomposition of worth actions month-on-month

Notes: The share of costs rising and falling replicate month-on-month modifications. Shares are seasonally adjusted utilizing the R package deal seasonal. Spikes in 2008, 2010 and 2011 are a consequence of UK VAT modifications (17.5% to fifteen% in 2008, improve to 17.5% in 2010 and improve to twenty% in 2011). The gray shaded space covers the time between March 2020 and July 2021 when the economic system (and knowledge assortment) was most affected by the Covid pandemic. Dashed strains present the 2011–19 averages. Newest remark: September 2023.

Sources: ONS and authors’ calculations.

To be clear, this chart just isn’t saying that 80% of merchandise by no means change costs. If the value of an merchandise remained fixed between December and January, and rose between January and February, it could transfer from the blue into the inexperienced class throughout this era. Equally, it could fall out of the inexperienced, into the blue or crimson, if from February to March the value once more remained fixed, or fell, respectively.

So, maybe surprisingly, this chart exhibits that month-to-month worth dynamics within the economic system are pushed by solely a comparatively small fraction of roughly 20% of all items and companies within the consumption basket. Additionally, we see that in the newest episode, the shift into rising costs has been largely out of the ‘no change’ class. Therefore, fewer costs are staying fastened, and extra are rising. It’s value noting that the latest up-tick within the shares of costs rising is just matched traditionally by these attributable to VAT modifications in 2008, 2010 and 2011, which nonetheless seem as one-off worth spikes slightly than a persistently greater share of worth rises, as in 2022.

If it’s a minority of whole merchandise whose worth modifications, you will need to take a more in-depth look. Chart 2 exhibits the distribution of costs modifications from 2019 by quarter (truncated at zero to exclude no-change observations). According to the rise within the inexperienced line in Chart 1, we observe that over 2021 and 2022 a variety of mass moved into the correct facet of the distribution, that’s the share of worth will increase, with the share of worth decreases being comparatively steady.

Chart 2: Evolution of the distribution of worth modifications by quarter 2019–23

Notes: The share of costs that didn’t change is excluded from these densities. The truncated densities are estimated in R through the Bounded Density Estimation package deal utilizing the boundary kernel estimator. Darker colors correspond to quarters through which year-on-year CPI inflation was comparatively excessive, lighter colors to quarters through which it was low. Every distribution represents month-on-month modifications inside the similar quarter. Newest remark: 2023 Q3.

Sources: ONS and authors’ calculations.

A notice on the chart: the distribution of worth modifications, when mixture inflation is at or shut to focus on, is roughly symmetric in logarithms. On this scale, a doubling (+100%) is equally distant from zero as a halving of the value (-50%). Resulting from gross sales, the doubling and halving of costs truly occurs frequently within the knowledge, which explains the bunching round these factors. Whereas these could also be a supply of seasonality within the knowledge, which can obscure the underlying dynamics, we don’t consider they’re essential for the general form of the distribution which we present right here.

In Chart 3, we zoom in on a few these densities to higher see variations of their form. They’re the densities corresponding to cost modifications within the third quarter of 2022 and 2023 alongside a median density over the pre-Covid interval.

Chart 3: Comparability of densities from 2022 and 2023 in opposition to a pre-Covid common

Notes: The share of costs that didn’t change is excluded from these densities. The truncated densities are estimated in R through the Bounded Density Estimation package deal utilizing the boundary kernel estimator. To match densities throughout time, they’re normalised to sum to the common share of costs falling and rising respectively inside the quarter. The yellow line exhibits the pointwise common density over the third quarters of the years 2011–19.

Sources: ONS and authors’ calculations.

We will see how, in comparison with this historic common – which we use as a stand-in for pricing behaviour when inflation was near the two% inflation goal – 2022 noticed an enormous variety of costs improve whereas there was little change within the behaviour of the decrease a part of the distribution. Within the newest knowledge, this mass of will increase has begun to subside, and, on the similar time, there’s a rising variety of costs outright falling on the month. Nonetheless, the modal worth improve (that’s, probably the most possible) continues to be elevated at about 6%, in comparison with roughly 3% on common throughout 2011–19).

Conclusion

To summarise, wanting on the micro degree of worth modifications, we discover a seen discontinuity in price-setting within the first quarter of 2022. Quite a lot of components, corresponding to the big rise in vitality costs in early 2022, in addition to supply-chain points following Covid lockdowns, seemingly contributed to this vital change in price-setting dynamics within the UK (relative to any latest historic precedent no less than). On the micro degree, companies’ pricing selections led to the emergence of a giant rebalancing within the distribution of worth modifications. All of a sudden, extra costs for a lot of totally different merchandise have been rising on the similar time. In comparison with the obtainable historical past for these knowledge, the latest interval is exclusive. Extra analysis will likely be wanted on the causes of this marked shift within the distribution of worth modifications, each at a micro and at a macro degree.

Within the very newest knowledge, there’s some proof that the distribution of worth modifications has certainly begun to return within the course of its historic common, although it’s too quickly to ascertain a pattern.

Lennart Brandt and Natalie Burr work within the Financial institution’s Exterior MPC Unit, and Krisztian Gado is a PhD candidate at Brandeis College.

If you wish to get in contact, please electronic mail us at [email protected] or go away a remark under.

Feedback will solely seem as soon as permitted by a moderator, and are solely revealed the place a full identify is equipped. Financial institution Underground is a weblog for Financial institution of England workers to share views that problem – or assist – prevailing coverage orthodoxies. The views expressed listed here are these of the authors, and are usually not essentially these of the Financial institution of England, or its coverage committees.