How far will UK unemployment rise?

The planned closure of the original job retention scheme (CJRS) at the end of October has focused attention again on the prospects for unemployment.  A ‘top down’ approach (based on GDP) suggests that UK unemployment could rise to more than 3 million, or 10% of the workforce. But a less pessimistic ‘bottom up’ approach suggests it will be capped at about 2¼ million, or less than 7%. Either way, talk of ‘many millions’ of job losses is hopefully still wide of the mark.

UK unemployment remains relatively low despite the slump in economic activity. However, it is normal for the labour market to lag behind the rest of the economy. The usual process of catching up could also now be made worse by the additional hit from the second wave of coronavirus and the changes to government subsidies that protected jobs during the first lockdown.

The headline measure in the UK counts someone as ‘unemployed’ if they do not have a job but have been actively seeking work within the last four weeks and are available to start work within the next two weeks.

On this basis, some 1,364,000 people were unemployed in the three months to February, before the pandemic struck, or about 4% of the economically active population. The latest official data suggest these figures rose only slightly, to 1,522,000 and 4½%, in the three months to August.

However, these figures surely understate the deterioration in the labour market, for three main reasons. First, they are an average over three months, at a time when the figures in each individual month have probably been rising.

Unfortunately, the ONS has temporarily suspended publication of the timelier single-month estimates while it updates the weighting methodology. But based on the previous trend, it seems likely that unemployment was already around 1¾ million by the end of August, or 5%, even on the basis of the official numbers.

(This is at least lower than the figure of 7% for September suggested recently by the Resolution Foundation, which uses a different methodology and is not directly comparable to the official measure. It is also simply implausible that unemployment has jumped this much in a single month on a like-for-like basis.)

Second, people who have lost their job but are no longer actively looking for work (perhaps because they do not expect to find any) do not count as unemployed. This has already been reflected in an increase of nearly half a million in the number of people who are economically inactive.

Third, the millions of people who are on furlough still count as employed, even if they are not doing any work and would already have been made redundant without the government subsidies. The number who might still lose their jobs is probably the biggest uncertainty.

So, how might we attempt to forecast what happens next?

One approach, from the ‘top down’, is to look at the usual relationship between economic activity and jobs. This is what the Bank of England’s Monetary Policy Committee (MPC) did in August, when it predicted that unemployment could rise to around 7½% (by 1 million to 2½ million) by the end of the year.

This applied a standard rule of thumb – known as an ‘Okun coefficient’ – which suggests that each 1 percentage point (pp) fall in GDP relative to what would otherwise would have happened could add 0.6pp to the unemployment rate.

At the time, the MPC was assuming that GDP would still be 5.4% below February levels at year-end, which implies a gap of at least 6% compared to the pre-Covid path. 6% times 0.6 is 3.6%, which would be enough to lift unemployment from 4% to around 7½%.

The assumption about GDP, alas, now seems far too optimistic. GDP in August was still 9.2% lower than it was in February. The gap probably closed further in September, but the recovery is likely to stall over the remainder of the year. Even without a ‘double dip’ recession, the gap between actual GDP and the pre-Covid path could widen to 10% over the coming months. This would be consistent with a rise of at least 6% in the unemployment rate – also to 10%.

That in turn implies an increase in the number of people who are unemployed from 1½ million (in the three months to August) to between 3¼ and 3½ million, or a jump of up to 2 million. To put that in context, UK unemployment hasn’t been above 3 million since the mid-1980s. The peak after the last global recession was around 2½ million in 2010, or 8%.

However, forecasts of as many as three million unemployed still seem too pessimistic. For a start, 0.6 is a relatively high number for the Okun coefficient, which varies significantly across time and between countries. Other studies have suggested a figure of around 0.3 for the UK, which would suggest that the unemployment rate will end the year at around 7%, rather than 10%.

What’s more, this is clearly an unusual recession where the normal rules are less likely to apply. Otherwise, unemployment should already be far higher, given that economic activity fell by 25% in just two months between February and April. Instead, despite reporting a relatively large decline in GDP in the second quarter, the UK saw the smallest drop in employment of any major European economy.

In part, of course, this reflects the extraordinary measures that the government and Bank of England have taken to protect businesses, jobs, and incomes, of which the furlough scheme is only one part. (The IMF has praised these responses as ‘one of the best examples of coordinated action globally.)

But, in addition, hopes that the downturn will only last as long as the Covid restrictions are in place have probably encouraged firms to keep staff employed, while the UK’s relatively flexible economy is better than most at creating new jobs to replace any that are lost.

Indeed, the labour market was already showing some signs of life over the summer – at least before the latest tightening in Covid restrictions. The media tends to focus on announcements of redundancies and job losses (though even these have been fewer than many fear, especially given the statutory notice period). But there has been some good news too:

  1. new hiring is also picking up, as seen in the improvements in surveys of recruiters and in the number of online job ads;
  2. experimental PAYE data suggest there was actually a small increase (of 20,000) in payroll employment in September;
  3. households’ own perceptions of ‘job security’ were the least negative in October for seven months.

In the meantime, the number of people on furlough has already fallen sharply. The official HMRC data here are not very timely, but the latest ONS business survey suggests 7.5% of the workforce (or about 2.1 million people, based on the 28.3 million on payrolls in September) remain on partial or full furlough.

The key question is therefore how many of these people might still lose their jobs. Most surveys suggest figures in the hundreds of thousands, not the millions.

For example, CFOs surveyed by Deloitte expected to retain 82% of their furloughed staff when the original CJRS ended. The implication that 18% are likely to be lose their jobs could be an under-estimate, since the sample mainly comprises larger businesses who are less at risk, and Covid restrictions have since been tightened further. Or it could turn out to be too pessimistic, given that the government has announced a replacement for the CJRS will go a long towards protecting more jobs.

Either way, even 25% of 2.1 million would ‘only’ be 525,000: still a huge and painful number, but far fewer than the ‘many millions’ some fear. What’s more, some of these people (or others) should be able to find work elsewhere in the economy. Overall, this could cap net job losses from current levels at no more than half a million.

You can get similar numbers by looking at the jobs at risk in the most vulnerable sectors. End-August data from HMRC show that 592,800 jobs were still furloughed in ‘accommodation & food services’ and 218,400 in ‘arts & entertainment’. These figures have probably fallen since then, but let’s say 800,000 remain on furlough in ‘hospitality’ overall. Even if half these people subsequently lose their jobs, that’s still less than half a million in total.

Finally, some people who lose their jobs may be foreign workers who return to their home country rather than remain unemployed in the UK.

So, let’s add an estimate of half a million additional job losses (net) to where unemployment probably was in August, taking it from 1¾ million to 2¼ million, or 6¾% of the workforce. That’s my best guess at where UK unemployment will peak in the coming months. To put that in context, unemployment in the euro area had already risen to 8.1% in August.

To be clear, this relatively optimistic forecast is not a reason for complacency, or to dismiss the broader economic and social costs of lockdown. Any increase in unemployment is a worry and, without the combination of policy support and (just as importantly) the relative flexibility of the UK labour market, things could indeed be a lot worse.

But the spectre of a return to 1980s-style mass unemployment can at least be banished. Indeed, most people should still be pleasantly surprised at how low unemployment remains.

PS. This blog was written before the announcement of a new nationwide lockdown (in England) on 31st October. However, the original job retention scheme has also been extended for at least another month. Overall, I continue to expect the UK unemployment rate to be capped below 7%.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s