The effect of coronavirus on consumer spending
The coronavirus is set to have a very large negative impact on global consumer spending. Over 40% of consumer spending in advanced economies is linked to ‘social situations’ and is vulnerable to lockdowns and other social distancing. Some of this spending will be delayed, and some lost entirely. The extent of the losses depend how long restrictions stay in place. Our baseline forecasts look for a fall in consumption of 5.5% in the EU 2020, with GDP declining around 5%. With longer lockdowns annual declines in consumption and GDP could easily exceed 10%.
The coronavirus is set to have a large negative impact on world growth, with particularly big effects on consumer spending. This is because a large share of consumer spending normally takes place in crowded areas and other ‘social situations’. This spending is therefore highly vulnerable to people avoiding crowds due to health concerns and to lockdown policies which close shops and transport systems and keep people at home.
Especially at risk is ‘discretionary’ spending in sectors like recreation and culture, hospitality, and transport. These sectors account for over 40% of consumption in the UK (Chart 1) and similar shares elsewhere in Europe. The likelihood is that a large chunk of spending in these sectors will either be postponed or lost altogether as a result of the pandemic. For example, new car purchases or clothing purchases are likely to be delayed, while public transport journeys related to work and much spending on restaurants may be lost altogether.
High frequency data already point to very large declines in spending in these vulnerable sectors. Restaurant spending in the advanced economies has collapsed almost entirely, airlines have seen demand dry up and Citymapper data shows a massive slump in urban movements across Europe – of 90% or so in most cases – which in turn points to a collapse in discretionary spending (Chart 2).
What might the aggregate impact on consumption be? We can estimate this if we adapt the framework and assumptions used by Keogh-Brown et al. in their study of the economics of flu pandemics. They assume consumer spending on ‘core’ areas such as food and drink and utilities is relatively unaffected while for discretionary sectors like clothing, furniture, hospitality, transport, and recreation large shares of spending – between 50% and 100% - are either postponed or lost (Table 1).
To get an estimate of how much consumer spending might decline over different time horizons, we then need to flexa number of other assumptions such as the duration of lockdowns, the pattern of recovery in spending after outbreaks come under control, and (crucially) the ‘consumer shock factor’, i.e. the share of consumers whose behavior changes. The stricter the lockdown, the higher the latter will be.
Even relatively short-lasting lockdowns produce quite large effects. A three-week lockdown affecting 50%-90% of a population would cut consumer spending in the three-month period featuring such a lockdown by 5%-8%, a six-week lockdown would lower it by 9%-16%, and a 12-week lockdown would slash it by 18%-32% (Table 2). With consumption typically 70% or so of GDP in advanced economies, GDP impacts would also be very large.
Full-year effects depend both on the length of lockdowns and on how quickly postponed consumption recovers after restrictions ease. Most models, and the evidence from past outbreaks (e.g. SARS), suggest much of the postponed consumption reappears quickly, i.e. in the quarter after outbreaks are contained. But even fairly quick recoveries can still imply big full-year losses. In some stylized scenarios, an initial 18% slump in consumption still implies a full-year loss of 9%, even if spending recovers to pre-pandemic levels after four quarters. If the recovery takes eight quarters, the full-year loss is an enormous 14% (Table 3).
Consumer spending may also be negatively affected by the financial market effects of the coronavirus outbreak. Since mid-February, European stock markets have dropped sharply in value, by 20-25%. This may generate negative wealth effects whereby households spend less and save more to rebuild balance sheets (as in Japan in the 1990s). Such effects might even stretch years into the future unless markets rebound dramatically.
Another risk is from ‘second-round’ effects relating to layoffs and business failures. Early data already point to steep rises in unemployment in the US and UK, which will lead to reduced consumer incomes and spending in affected groups. This underlines the importance of effective government schemes to backstop consumer incomes and tide businesses over.
Set against these negative effects, there will be some positive offsets. It appears likely that consumption of home delivered goods, online entertainment such as video games and film and TV streaming services may rise. But these additional purchases will be swamped by the deep declines in spending already seen in discretionary sectors.
Taking all these factors into account, our baseline forecasts predict a decline in consumer spending in the EU of 5.5% in 2020. The impact is concentrated in Q2 where we expect to see around a 10% quarter-on-quarter drop. This translates into a forecast GDP decline of just under 5% for this year. This assumes the outbreak comes under control during May and June, with consumer spending then staging a strong rebound in Q3 and Q4 of this year (Chart 3). But as our scenarios earlier showed, should the virus take longer to contain, it is easy to see consumer spending (and GDP) dropping at a double-digit rate in 2020.