Consumer prices in light of the COVID-19 crisis

The COVID-19 pandemic has caused huge shocks to both demand and supply of goods and services, which could result in respectively disinflationary and inflationary pressures. Together with the lockdown, loss of income, deep uncertainty among the population and risk-avoidance behaviour, it has resulted in a dramatic shift in consumption patterns. This article analyses inflation developments in Belgium during the Spring lockdown and soon after it.

The cut-off date for completing this article was 16 October 2020. In the meantime, new measures to fight the pandemic have been implemented, which are not taken into account in the article.

Due to the restrictions on the movement of people and the closure of non-essential stores and services, some prices could not be polled normally in order to construct the consumer price index. In Belgium, Statbel has decided to temporarily suspend the field price collection. However, the need for local price collection had already been greatly reduced in Belgium thanks to the switch to big data for several goods and services (scanner data, web scraping, administrative data). Since 2014, the share of field price collection has declined from about 64 % of the basket weight to about 32 % in 2020.

Following Eurostat recommendations, as regards goods and services that could no longer be bought, the last available prices were carried forward by Statbel should no seasonal pattern be observed in the monthly pricing (e.g. restaurants, hairdressers, electricians, jewellers, etc.). When monthly price changes are marked by a seasonal pattern (e.g. hotels, travel, flowers, etc.), imputed prices have been obtained by applying the same monthly price change as the previous year. In April, price imputations were necessary for 24 % of the consumption basket. In May, this was the case for 17 %, while in June, only 4 % of the prices were imputed.

Along with the outbreak of the COVID-19 crisis, oil prices collapsed in March and April 2020, causing total inflation to fall. Core inflation (i.e. total inflation excluding energy and food) remained stable throughout most of the year, but the postponement of the summer sales exerted a strong upward impact in July and a downward impact in August on year-on-year core inflation.

Food inflation has been rising since the end of 2019, a trend which continued during the Spring lockdown and the following months. The upward movement can be explained by both COVID- and non-COVID-related factors. As regards the former, the crisis has caused some disruptions in the supply chain, with an upward impact on prices. In order to discourage hoarding behaviour, the government decided to impose a ban on promotions in supermarkets, starting on 18 March. Promotions were only fully allowed again from 4 May onwards. Non-COVID-related factors include the particularly low food price growth in 2019 and – regarding unprocessed food – high pork prices due to swine fever and meteorological factors that mostly affected fruit prices in early 2020.

The consumption pattern of households has drastically changed. Among other restrictions, restaurants and bars were closed between 14 March and 8 June; this was also the case for non-essential shops (e.g. clothing) although in a more gradual manner in March with a re-opening in May. Cultural and recreational activities were cancelled. Hence, consumers were forced to save, and, in relative terms, spent a much larger part of their budget on food.

We construct a consumer price index based on the HICP principles, but with weights more in line with the actual consumption during the lockdown. For this, we use a sample of debit card data, from which we have data for the total amount of monthly transactions per category of shops. Our weights vary on a monthly basis as of February 2020: the relative weight of food increased considerably in April. By contrast, the relative weights of most non-food sub-categories – such as clothing, restaurants and bars, etc. – showed a V-shape: they declined from February to reach a low point in April 2020, after which they went up again.

We have opted for a Fisher index, with weights that correspond to changes in spending patterns from the current month compared to the reference period (December 2019). We obtain a somewhat higher inflation rate than the officially published figure; over the months from February to July 2020, the average difference between the two series amounts to 0.2 of a percentage point (the average inflation rate was estimated at 0.7 %, against 0.5 % for the official one).

While the consumer price index is constructed for an average Belgian household, households may differ in many respects, such as their level of income. However, to a certain extent, the pandemic has forced all consumers to have a more similar consumption basket, as differences in spending patterns across households have been fading out.

Still, on the basis of the 2018 Household Budget Survey (HBS), we calculated the national inflation rate in a simplified way, per income quartile. The share of electricity, gas and heating oil (three categories that have registered declining inflation rates since the beginning of 2019) being the largest for the lowest income categories, and the share of food (with higher inflation rates in 2020) being the largest for higher-income families, the lowest quartile has displayed somewhat lower total inflation rates in 2020.

As regards consumer perceptions and in a forward-looking manner, their expectations too, the NBB surveys give an indication of the changes in prices evolution perceived over the past 12 months and expected for the next 12 months. At the time of the Spring lockdown, households’ inflation expectations and measured inflation diverged. While total inflation fell sharply from January to May 2020 in Belgium, households expected a sharp rise in prices during the April survey.

The less consumed goods were those for which prices rose the least rapidly (such as some services) or even fell (fuels) while the more consumed goods were those for which prices grew the most (food). In April, the higher proportion of households expecting a strong price increase could be explained by an overreaction to changes in food prices. Several studies have shown that, when forming inflation expectations, households weigh food prices much higher than their actual share in expenditure. However, after the lockdown, consumer expectations have been trending downwards and, in that respect, they seem to be disconnected from inflation perceptions, which relate to the past 12 months. Households' uncertainty about overall inflationary pressures may have grown, as their expenditure patterns have been greatly transformed.