Economics of Covid-19 – why total hours of work could prove the key labour market statistic in this crisis

Each day brings yet sadder news. The UK death toll from Covid-19 is now just shy of 3,000 and will probably pass that mark when we receive the next update this afternoon. Particularly sobering is the constantly updated tally of deaths in the United States and globally published by Johns Hopkins University, the global total already above 50,000

The human scale of this relentless tide of grim health statistics is evidenced by the fact that staggering record-breaking figures on the number of people making claims for unemployment related welfare benefits haven’t monopolised the headlines in either the US or UK in quite the way they would at any other time.

As mentioned in an earlier blog, it will be a while before we start to see the effects of the Coronavirus crisis appear in the UKs published official labour market statistics. In this respect the US Bureau of Labor Statistics is ahead of the game. But the UK’s Office for National Statistics (ONS) has helpfully begun to issue snapshot data on the business impact of Covid-19.

The first weekly release published yesterday showed that of more than 3,600 business responding to a survey between 9 March and 22 March 27.4% reported they would be cutting staff numbers ‘in the short-run’. A similar proportion 28.5% had reduced staff hours, while approaching half (46.2%) were requiring staff to work at home in line with government advice on social distancing for non-essential workers. Some of these businesses were making a combination of such adjustments and one would expect some will also be making pay cuts or introducing pay freezes, though this is not covered in the survey.  

While these data are useful, they don’t come close to giving us the kind of numbers that will at some point be provided by the large-scale rolling Labour Force Survey (LFS). Assuming the ONS’ normal data collection process isn’t itself disrupted by social distancing – interviews are conducted face-to-face as well as by telephone – analysts will be digging into crisis relevant data from late spring onward. However, even when these data become available, the labour market impact of this particular crisis could prove trickier to read than usual.

One reason is that the impact is very abrupt. An emergency brake has been applied to large swathes of economic activity. We’re not experiencing a slowdown and slide into recession but a sudden sharp halt. There is not the kind of lag in the labour market impact one would normally expect but instead a punch in the gut that immediately takes the wind out of the sails. The bad news is that, among other negative effects, this causes a mega surge in job or income losses so big that it tests the ability of policy makers to respond. The less bad news is that the initial surge doesn’t necessarily mean a relentless wave of pressure on the system. More likely is a sudden bout of acute pain which doesn’t get noticeably worse the longer the crisis continues but will nonetheless be very tough to live with.     

Another oddity of the current crisis is widespread use of furlough procedures. An as yet unknown but almost certainly high number of employees will be retained in jobs but placed on furlough under the government’s emergency Job Retention Scheme (involving an 80% wage subsidy subject to a monthly cap of £2,500). It also looks as though some employers will retain staff in similar ways without participating in the government scheme.

Furloughing hasn’t tended to be a noticeable feature of UK employment practice during recessions: job cuts, shorter hours, pay cuts or some combination of these adjustments are generally more common place. But the significance of the furlough for our reading of labour market statistics is clearly demonstrated by a look at how the Job Retention Scheme will operate.

To be eligible, furloughed employees won’t be able to do any paid work for their employer or paid work elsewhere (they can if they wish do unpaid voluntary work). Despite them being ‘idle’ the LFS will count these employees as in employment. A person is deemed to be employed if he/she has done at least an hour of paid work in the week when surveyed or is temporarily away from work for a legitimate reason such as paid holiday, sickness absence or, in current circumstances, on furlough. Consequently, the headline employment figure obtained from the LFS will give a somewhat flattering picture of the effect of the crisis, as might the level of unemployment and economic inactivity insofar as these furloughed employees would otherwise lose their jobs.

The hit on the labour market will therefore be much better reflected by the slump in hours worked by furloughed employees, which will be in addition to cuts in hours worked by non-furloughed employees who keep their jobs and by self-employed people unable to find any or only a limited amount of work. But this is not the only reason why hours of work will prove to be an even more important statistic than usual in this particular crisis.

In a normal recession, hours lost through sickness absence tend to dip; throwing a sickie is not a good idea when jobs are on the line. But in this Covid-19 induced crisis, rates of sickness absence will rise significantly, the virus in this way having its most direct negative impact on the labour market. Offsetting this, however, are two other unusual features of the current crisis. Essential services, across both public and private sectors, are working even harder than usual and increasing hours as well as recruiting extra staff. And anecdotally, it appears that the shift to widespread working from home during the lockdown might be seeing people putting in extra hours because they’re spending less time commuting.

What kind of overall hours effect should we thus be looking at? In the last three major UK recessions (the early 1980s, early 1990s, and 2008/9) total hours worked fell by 5.3%, 5.2% and 4.1% respectively. The corresponding falls in total employment were 2.4%, 3.4% and 1.9%. Judging from what we can see so far about the current crisis and the initial employment policy response it looks as though the fall in hours will be at least on a par with the relatively large falls suffered in the 1980s and 1990s even if we manage to avoid job losses on quite the same scale. Either way, hours of work (and, more important still, what change in these mean for weekly earnings) could prove to be the key labour market statistic to look at when measuring the impact of Covid-19.