The pandemic's seismic economic disruption has suddenly and radically changed consumers' spending patterns around the world, leaving official inflation data unmoored from what they are experiencing.
Surges in price growth in many major economies are worrying investors and central bankers. But part of that trend can be blamed on Covid-19 — with important consequences for how we interpret official data.
Inflation is measured using a basket of goods and services which is meant to represent what people typically purchase. Items are given weights proportional to the amount spent on them.
Across the globe over the past year as the pandemic took hold, people stopped spending on restaurants, airfares and other lockdown-restricted activities practically overnight.
As a result, some countries' real-world experience of inflation differed from the official headline rate by as much as 0.89 percentage points in a single month last summer according to research by the Harvard professor Alberto Cavallo. US inflation may have been underestimated by 0.5 percentage points over the course of 2020 according to Cavallo's estimates.
The dislocation is set to continue as major economies shift back to more normal consumption patterns, economists have warned — making it harder to interpret the official measures of price change which comprise central banks' primary policy target.
The details of the US and eurozone inflation calculations are different, but both methods result in the same challenge.
The US calculation mismatch
The US Bureau of Labor Statistics uses survey data that lags real-world changes by two years. So since the pandemic began, official inflation data has been calculated using weightings that reflect the pre-pandemic world.
The mismeasurement is particularly acute in two categories.
Spending on groceries increased 29 per cent in March last year according to credit and debit card purchasing data from Opportunity Insights. The surge in demand pushed prices up 2.7 per cent month-on-month.
By contrast, transport spending declined 70 per cent in April last year and prices dropped 7 per cent from the start of the year. The effect has lasted: spending on transport was still 25 per cent below pre-pandemic levels in May this year.
According to economists, the net effect is that the official calculation of the consumer price index underestimated the price changes the population experienced in their day-to-day lives.
"You can easily see why CPI was a poor measure of inflation during the pandemic recession," said Miguel Faria e Castro, an economist at the Federal Reserve Bank of St Louis. "This recession entailed an unprecedented shift in the composition of household consumption."
As the US economy reopens and lockdowns end, the effect has reversed.
For example, a decline in production of cars and trucks over the past year has pushed up used vehicle prices. Since Americans have spent less on transport since the start of the pandemic than the CPI weighting suggests, April's official inflation reading — which recorded a big jump — was probably an overestimation.
"In April, what caused a third of the jump was actually in used cars and trucks. That was a big driver of it," said Cavallo, who has used real-time spending data from Opportunity Insights to reweight items in the CPI basket to more accurately reflect recent price changes.
"The CPI basket should have put less weight on [transport], and [then] you get a lower inflation reading in the month of April," he said.
The European reset problem
Unlike the US, eurozone inflation weightings are updated annually in January. Changes in consumption patterns during the pandemic mean this year's calculation will give little weight to the prices of items such as petrol, hotels and restaurants, hit by lockdowns last year.
As a result, even if European consumers return to pre-pandemic spending patterns as the economy normalises, official statistics will underestimate spending in these areas.
Carsten Brzeski, economist at ING, said that "interpreting this year's inflation data will not be an easy task" as pandemic-era spending creates distortions not only during lockdown periods "but also later this year with [consumers] returning to normal consumption patterns".
This means official inflation statistics are "less accurate", said Gregory Claeys at the economic think-tank Bruegel. "We are going to have the same problem as last year, but reversed," he warned.
The effect could be significant. For example, the weighting of recreation and personal services and of restaurants and hotels fell by 4 and 3 percentage points respectively in 2021 compared with the previous year. On average over the past 24 years these spending areas have only changed by 0.06 percentage points annually.
The issue is already visible in the case of petrol: both price and consumption volumes fell last year. Petrol prices are now rising again, but its influence on the overall inflation calculation will be 13 per cent smaller than last year — and that effect will last until the eurozone inflation basket is re-set again in January. The same trend could happen with hotels and cinemas.
The European Central Bank calculated that inflation experienced by consumers last year was 0.2 percentage points higher than official CPI between April and August, and the change of weights this year pushed the inflation rate up by 0.3 percentage points in January.
Katharina Utermöhl, senior economist at Allianz, said policymakers "will need to take note of this and test alternative measurement methods to get a better sense of actual inflation dynamics" — as these statistical glitches continue to play havoc with CPI in the months ahead.
Written by: Brooke Fox, Valentina Romei and Oli Elliott
© Financial Times