We learn this morning that the UK Prime Minister is writing to every household with a grim message on coronavirus – ‘Things will get worse before they get better’. He’s referring of course to the terrible effect of Covid-19 on people’s health. But the same can be said of the effect on people’s livelihoods. So how much worse might things get and how long might it be before they get better?
A times of economic crisis such as that triggered by the pandemic, attention is understandably focused on the severity or nature of the shock and specific policy interventions introduced to counter it. As important, however, is the broader underlying orientation of policy – the labour market policy model (LMPM) – which will ultimately determine how the current crisis affects people’s jobs and incomes.
There is no off-the-shelf definition of an LMPM. I think of it as the mix of polices that in different ways affect the behaviour of employees, people in work, and people without work who might reasonably be expected to find work. Ideally, during a time of economic crisis, the LMPM will act to encourage employers to cut wages or hours or work rather than jobs, encourage people in work to restrain wage demands, and ensure people without work remain willing and able to fill any job vacancies that are or become available.
In terms of specific policies this covers things like employment protection measures (i.e. the ease of ability of employers to hire and fire staff), individual employment rights, wage bargaining procedures, wage subsidies/tax credits, and active welfare measures that place job search conditions on jobless benefit claimants.
The way in which these types of policies interact, operating as an LMPM, not only determines how much unemployment an economy experiences in normal times but also makes a critical difference to the way in which we experience recessions. Labour markets with good LMPMs are often referred to as ‘flexible labour markets’, though one needs to be careful about the latter term since conceptually it can mean different things and has ideological connotations.
Guidelines on best practice began to spread internationally following the OECD Jobs Study in 1992. A good LMPM will operate to minimise the peak in unemployment during a recession and thereafter facilitate a rapid sustained fall in unemployment to at least the pre-recession rate. To achieve this, the LMPM must continually adapt to structural change in the labour market and if necessary be easily amenable to short-run policy adjustments. No recession is ever exactly the same as another and it’s important that one doesn’t persist with a rigid LMPM only capable of ‘fighting the last war.’
The UK’s LMPM has developed considerably in the last 40 years, during which the economy has suffered three major recessions, in the early 1980s, the early 1990s and the late 2000s.
Between January 1980 and March 1981 GDP contracted by 4.6% before starting to recover. At that time, employers’ ability to hire and fire was fairly weak but union power was still very strong, limiting wage restraint. In addition, the welfare system was almost totally inactive.
Large numbers of benefit claimants gave up looking for work, effectively detaching themselves from the labour market. This resulted in a massive pool of long-term jobless and cut the usable supply of people, thereby putting upward pressure on wages and causing mass unemployment to persist for years after the recession had ended.
The unemployment rate increased from 5.5% before the recession to a peak of 11.9% in the summer of 1984 and stuck at around that rate before staring to fall in the second half of the decade, to a low of 7% at the onset of the 1990s recession. The fall is normally attributed solely to the expansionary ‘Lawson boom’ but the developing LMPM played an important part, notably the introduction in 1986 of Restart job search interviews for claimants of unemployment benefit which led the way in activating the welfare system.
At the start of the 1990s after a decade of reform the LMPM was considerably different. Both employment protection and union power had been weakened but active welfare reform was still a work in progress.
The contraction in GDP of 2.5% between July 1990 and September 1991 was less than that suffered in the early 1980s recession. However, the enhanced ability of employers to fire staff outweighed slightly greater wage restraint. The result was what some economists at the time called ‘oversacking’ because the shake out of jobs was much bigger than the depth of the recession actually warranted.
On the plus side, however, once GDP began to recover employers’ enhanced ability to hire and fire, plus the emergence of a somewhat more active welfare regime, resulted in a much swifter and sustained fall in unemployment than experienced in the 1980s. Although unemployment peaked at a very high rate of 10.6% in the spring of 1993, the peak came only 3 years after the previous trough. The unemployment rate then fell almost continually and was still on a downward path in 2008 at the onset of the Great Recession when the rate stood at 5.2%
As its name suggests, the Great Recession was far worse than its two immediate predecessors, GDP contracting by 6.4% between April 2008 and September 2009. Within the LMPM, employment protection was slightly stronger than in 1990, union power remained limited and the welfare system had become much more active.
This combination resulted in a smaller shake out of jobs than in either the 1980s or 1990s recessions, with wage cuts and shorter hours predominant, and a correspondingly smaller rise in the unemployment rate, from 5.2% at the start of 2008 to 8% at the start of 2010. The main shortfall in the model at the time proved to be lack of adequate provision for dealing with young jobseekers. By dampening the shake-out of jobs the LMPM also served to depress recruitment activity, hitting young entrants to the labour market even harder than usual.
The Noughties version of the LMPM would almost certainly have enabled an earlier peak in total unemployment following the Great Recession and a rapid subsequent fall had a nascent jobs recovery in early 2010 not been curtailed by strict fiscal consolidation measures that began to be imposed on the economy later that year. Having dipped between the spring and autumn of 2010, the unemployment rate started to rise again reaching a peak of 8.4% at the end of 2011. There was then a de facto easing of fiscal policy with the acceptance that the initial deficit reduction timetable would need to be extended, allowing for higher public borrowing than originally planned which helped support aggregate demand.
From that point onward the LMPM began to facilitate the kind of swift and sustained fall in unemployment that one would otherwise have expected to see from early 2010 onward. This, incidentally, places talk of a remarkable post 2012 ‘jobs miracle’ in its proper perspective. An understanding of the Noughties version of the LMPM made the supposed ‘miracle’ entirely predictable a decade ago, and some of us are on record as saying so at the time.
What might we expect in the wake of the Covid-19 crisis? How recession proof is LMPM 2020?
In its essential features the LMPM appears as robust as it did when the Great Recession struck. But two issues in particular are worthy of attention.
The LMPM is slightly weaker on employment protection than in 2008, with employers able to fire regular employees with less than two years’ service without legal comeback (at present that’s around 7 million people). Moreover, the model has been largely permissive on the rise of the so-called ‘gig economy’ of casualised contract workers (employing around 6 million people, including the self-employed) which now represents a significant enclave of super (or maybe one might say uber) hire and flexibility within the labour market.
There is therefore a clear risk that the current LMPM leaves us more prone to a bigger shake-out of jobs than during the Great Recession. In view of this the government’s Covid-19 emergency adjuncts to the LMPM such as the Job Retention Scheme and Self-Employment Income Support Scheme might have been necessary notwithstanding the sheer scale and nature of the coronavirus crisis.
Another potential worry is the operation of the Universal Credit (UC) welfare system. While UC is in principle no more than the latest iteration of an activity focused benefit system, concern has been raised that in practice it is performing less well than its predecessor systems at helping claimants into work. Well before any of us had heard of Covid-19, some labour market policy experts were warning that UC was slowing flows from unemployment and economic inactivity to employment, leading to a considerable rise in the stock of UC claimants. With early DWP figures pointing to a surge in the inflow of UC claims since the Covid-19 lockdowns came into effect, possibly leading to system overload, urgent improvement to the activation aspects of the benefit would seem necessary.
At the time of writing, the projected UK peak in coronavirus cases is still several weeks away and the duration of the widespread social lockdown is unknown. There is also the distinct possibility of further intermittent lockdowns to combat subsequent waves of the virus in the absence of herd immunity or a suitable vaccine. With large swathes of the economy already effectively placed in a coma, the contraction in GDP is almost certain to be larger than during the Great Recession, and while it’s hoped the Covid-19 induced recession will be shorter this cannot be taken for granted.
At present I anticipate that measures such as the Job Retention Scheme and the Self-Employment Income Support Scheme should be enough to buttress us from some of the employment protection weaknesses built-in to the LMPM since 2010. This would suggest a doubling of the unemployment rate from 3.9% prior to Covid-19 to a peak of around 8% (similar to the Great Recession peak), bad enough yet not disastrous. Additional interventions might also prove necessary to address particular problems if they arise, such as differential impacts on different groups in the labour market, so as to avoid a repeat of the Great Recession when too great a burden was allowed to fall on young jobless people.
The fiscal cost of achieving this outcome will be jaw dropping but preferable to the combined economic and social cost of allowing the unemployment rate to soar into double digit territory.
How long unemployment remains at the peak rate before starting to fall will then depend in part on the length of the Covid-19 recession and in part on the operation of the LMPM.
While the current hire and fire features of the LMPM makes us more prone to a big shake-out of jobs during a recession they should also enhance prospects of an early rebound in hiring during the recovery. This bodes well for a swift post-recession jobs bounce back.
However, question marks over the operational effectiveness of UC raise the possibility that the jobs recovery might be slowed by bottlenecks in the system for moving jobless people off welfare and into work. This possibility ought to be considered and addressed with considerable urgency.
In any event, to support the LMPM economic policy makers will need to learn from the mistake of 2010 and adopt a slow and steady rather than fast and furious approach to the major fiscal consolidation that inevitably awaits us once Covid-19 has finally been defeated.