Online shopping

Spending and saving during the COVID-19 crisis: evidence from bank account data

Published on 29 October 2020

The measures taken to help reduce the spread of COVID-19, resulting from both policy and consumers’ changes in behaviour, have had major impacts on consumer spending patterns. In this briefing note, we explore how consumer spending has evolved, both during lockdown and in the recovery phase since.

The measures taken to help reduce the spread of COVID-19, resulting from both policy and consumers’ changes in behaviour, have had major impacts on consumer spending patterns. In this briefing note, we explore how consumer spending has evolved, both during lockdown and in the recovery phase since. We document how different sectors of the economy have fared, and look at how spending has varied with the prevalence of COVID-19 cases across local areas. We also analyse how the closure of many businesses has led to ‘forced saving’, and explore how this, together with income falls, has affected the extent to which richer and poorer households have accumulated or decumulated savings over the crisis.

Our analysis uses anonymised user data from the Money Dashboard (MDB) budgeting app. These track transactions of app users in real time, allowing us to build a detailed picture of their spending patterns and financial balances from before the crisis to the end of September. We combine this with health-authority-provided data on case prevalence in the population to look at the effect of case rates on spending.

Key findings

1. The recovery in consumer spending following the easing of lockdown restrictions has been very partial. Since the end of July, spending has stalled at around 90% of the level we would have expected in the absence of the pandemic. Spending initially fell by around a quarter but gradually recovered through May and with the reopening of non-essential retail and hospitality in June and July.

2. Local areas with low case counts have not recovered discernibly differently from those with high case counts. Since the very early stages of lockdown, consumers’ behaviour has borne very little relationship to being in an area with higher or lower prevalence of COVID-19 compared with the rest of the country. For example, overall spending in September was at 89% of its 2019 level, while in the local authorities with the fewest cases it was barely any different at 90%.

3. While the fall in total spending has yet to be unwound, changes in the composition of spending that we saw during lockdown have also proved persistent thus far. Spending on groceries rose during lockdown and remains higher than in 2019; the opposite is true for spending on restaurants, pubs, holidays and transport, in which the recovery after reopening stalled at around the end of July. One of the reasons why these persistent sectoral differences matter is the knock-on consequences for workers: for example, those in the still-struggling sectors are disproportionately low earners.

4. There has been a persistent shift in shopping and payment habits during the crisis, away from cash and towards online-only merchants. Even in August and September, cash use remained below half of normal levels, having plummeted after lockdown, whereas spending on online alternatives for goods bought in physical stores remained over a quarter higher.

5. ‘Forced saving’ – declines in spending on goods and services that were substantially affected or shut down by lockdown – has been significant across the income distribution, but greater for higher-income households. Spending inunaffected sectors and those with close alternatives to physical retail rose across the income distribution, but with proportionally larger increases for poorer households.

6. Higher income groups appear to have accumulated more savings than in previous years during the crisis, with falling spending from forced saving outweighing income falls. In contrast, the poorest fifth have seen an average £170 per month decline in their bank balances between March and September (£1,220 total and equivalent to 14% of pre-crisis income) relative to what we would expect in normal times, as income falls are not fully cancelled out by lower spending. This is likely to reflect both lower saving and higher debt.