How much would school closures cost the economy? - Capital Economics
Global Economics

How much would school closures cost the economy?

Global Economics Update
Written by Vicky Redwood
Cancel X

School closures are one of the main ways in which government attempts to manage the spread of the coronavirus will dent economic activity. While parents can work around the closures to some extent, there is nonetheless likely to be a fairly sizeable hit to output via this channel in the affected countries.

  • School closures are one of the main ways in which government attempts to manage the spread of the coronavirus will dent economic activity. While parents can work around the closures to some extent, there is nonetheless likely to be a fairly sizeable hit to output via this channel in the affected countries.
  • So far, 29 countries have shut schools nationwide (China, Japan and Iran being the biggest ones to do to). A further 20 countries have implemented localised school closures. (For a list, see here.) Closures are likely to become increasingly widespread.
  • The most important economic impact of this will be the direct reduction in output due to parents taking time off work to look after their children. There will also be a reduction in demand if workers get paid less as a result, although in practice employers and governments are likely to offer support to employees in this position. There might also be sectoral shifts in demand – for example, cafes and gyms near workplaces will suffer, whereas people might spend more on food from supermarkets to eat at home.
  • On the face of it, the impact on output could be very big, particularly in countries where both parents tend to work. The OECD average is 61% and in Scandinavian countries, the share is as high as 80%. (See Chart 1.) Information on the actual number of working parents in countries is harder to come by, but labour market data in the US, UK and euro-zone indicate that between 15% and 20% of the workforce could need to take time off work to look after dependent children. Let us also assume that schools are closed for four weeks. A fifth of the workforce off work for that period has the potential to knock 6% off GDP in that quarter, or 1.5% off annual GDP.
  • In practice, the effect is likely to be smaller than this for several reasons. The first is that one of the parents might only work part-time. This will mean that the reduction in overall hours (and therefore output) will be less than the number of workers affected. Chart 2 shows that around a quarter of mothers work part-time in OECD countries on average, with the proportion much higher in the Netherlands and Switzerland.
  • Second, some people might be able to rely on other people to look after their children, including grandparents, nannies/au-pairs or clubs. Surveys in the US and UK show that around half of families where both parents work use such informal childcare. Indeed, there might be a partially offsetting boost to output if more clubs etc were put on to deal with the shutdown. That said, this care might now be unavailable. If schools get closed, then we assume that daycare/clubs get closed too. And if old people are advised to self-isolate, grandparents will in many cases be unable to care for children. Meanwhile, nannies/au pairs may not be able/willing to do extra hours.

Chart 1: % of households with children in which all adults work, 2018

Chart 2: % of working women in part-time work, latest

Source: OECD

Sources: Eurostat

  • The third mitigating factor is that some people might be able to do at least some work from home while looking after (i.e. turning on Netflix for) children. Admittedly, there are some types of work which cannot be done at home. On average, only 20% of people in OECD countries have ever worked from home. (Chart 3 shows the figures for women; the data for men are similar.) But companies are now ramping up their work-from-home facilities. Parents might also shift their working hours and work in the evening after children have gone to bed instead.
  • Fourth, even if people staying at home cannot get any work done, their co-workers might be able to cover their work to a certain extent. Academic research puts this so-called “elasticity of output” at around 0.8, meaning that a 1% drop in labour input leads to a smaller 0.8% drop in output. That said, if school closures are also accompanied by workplace closures, there will be less scope for co-workers to compensate.
  • Fifth, parents might just take their annual leave during the school closure, rather than later in the year. Indeed, delaying the school closures until the school Easter holidays – when parents might already have time off planned – could help in that respect. So although output would still take an immediate hit, it would be made up later in the year.
  • And finally, there is a question mark over whether any of this would actually show up in the economic data, given the difficulties of measuring of output in the services sector.
  • Chart 4 brings all of this together by showing how some assumptions about these various factors affect the impact of school closures. As we said earlier, the assumption that schools shut for four weeks and that a fifth of the workforce is affected has the potential to reduce quarterly GDP by 6%. (The bar on the far left in the chart.) Accounting for part-time work would reduce that effect to 5%. If we then assume that reliance on alternative childcare, working from home and co-workers picking up the slack each reduce the effect by a further fifth, then we end up with a hit to quarterly output of 2.5%.
  • If none of this were made up, it would knock 0.6% off annual GDP growth. The actual hit to annual GDP growth would be a bit smaller than this if workers made up some of this output when they returned to work and/or shifted the timing of their annual leave so they worked more later in the year instead.
  • Obviously these are all very broad-brush estimates and depend on both the assumptions used, the length of the school closures and the country in question. For example, the more advanced economies, which have a higher share of working women, might seem more vulnerable to school closures, but they might also be better-equipped for working from home. Meanwhile, the impact will be smaller in countries with relatively low birth rates, such as Japan.
  • Nonetheless, the potential effect of school closures seems sizeable and is one reason why we are reducing our economic forecasts further for many countries. We had already cut our forecast for global GDP growth from 2.9% to 2% and now a figure of more like 1.5% is looking likely.

Chart 3: % of women that sometimes work from home, 2015

Chart 4: Effect of 4 Week School Closure (% of Quarterly GDP)

Source: OECD

Sources: Refinitiv, Capital Economics

Vicky Redwood, Senior Economic Adviser, +44 20 7808 4989,