The UK’s great wage stagnation: beware aggregates and averages
1 December 2016
SUGGESTED
Society and Culture
Economic Theory
Labour Market
I like Paul Johnson and the work he leads at the Institute for Fiscal Studies. I read his column in The Times each week with great interest.
However this week he repeats a point with which, as he notes, he has become particularly associated – that UK real wages (measured by median earnings) have stagnated since the financial crisis, and look set to continue to do so. I don’t dispute this, but I’m not at all sure the right conclusions are being drawn.
Mr Johnson will be aware that the labour market is in constant flux. New people are entering and leaving the workforce all the time. When we compare snapshots of median earnings in 2008, say, to those for 2016, we are looking at two very different workforces. Several million people have exited through retirement, looking after children full-time, going to college, emigration, illness, incapacity and death. Another several million have entered the workforce from school and college, immigration (including return immigration by people who have been working abroad), return from unemployment or inactivity, and so on. They are very likely doing different jobs from those who have left the workforce. Moreover those who have been employed throughout the period are also likely to be doing different jobs from those they were doing in 2008. Every three months somewhere of the order of 6-700,000 people change their jobs in the UK.
Since 2008 the number in employment has risen by well over 2 million, and the employment rate has risen by more than two percentage points. This all adds to the mix.
On average, the jobs held in 2016 do not pay very much better than those in 2008. But so what? They are different jobs, held by different people. If fewer low-paying jobs had been created, median real earnings would be higher, as logically would productivity, that other much-quoted target variable. But we would have more unemployed and inactive people, while GDP per head of the population would be much the same: only now those in work would be supporting more non-workers, either through family transfers or taxpayer-funded benefits.
What happens to median earnings tells us little about what is happening to individuals. As I said, they change jobs. Such job changes are overwhelmingly voluntary, and typically lead to pay increases: for many that is the motivation for the move. Even those who remain in the same post are often on pay scales which they move up either automatically (in the public sector) or on promotion, even if there is no change to overall pay rates.
The ONS published a fascinating set of figures a couple of years ago that illustrated very clearly that a concentration on workforce median earnings can be very misleading about what is happening to individuals.
Their figures show that between 2013 and 2014, median earnings grew by just 0.1% – the sort of thing Paul Johnson draws attention to. However, if you calculate median earnings for those who were in work in both years, ignoring new entrants and leavers, the median rose by 4.1%, a very much larger figure.
The point is that, in a dynamic labour market with large numbers of new entrants and leavers, and with frequent job changes, what happens to median earnings for the whole workforce tells us next to nothing about what is happening to those continuously in work. The ‘stagnation’ story has been wrongly linked to individuals, and completely swallowed by both left and right – it lies behind Mrs May’s ‘just managing’ trope, of course.
In principle a faster increase in whole-workforce median earnings could be associated with lower pay increases for those continuously in work. Or an actual fall in median real earnings could be compatible with increased earnings for them. This paradox is an example of the dangers of reading too much into aggregates and averages, something I have pointed to on a number of occasions.
But it’s not just an academic point, a piece of intellectual trivia. The danger is that this misreading of the data will lead to policies, such as pushing up the National Living Wage and demonising firms which pay low wages, or display large gender pay gaps, which incentivise firms to start paying more to fewer workers. Then we will be like France, with higher wages and higher productivity, but also higher unemployment – and probably higher taxes.
Would Paul Johnson prefer this?
However this week he repeats a point with which, as he notes, he has become particularly associated – that UK real wages (measured by median earnings) have stagnated since the financial crisis, and look set to continue to do so. I don’t dispute this, but I’m not at all sure the right conclusions are being drawn.
Mr Johnson will be aware that the labour market is in constant flux. New people are entering and leaving the workforce all the time. When we compare snapshots of median earnings in 2008, say, to those for 2016, we are looking at two very different workforces. Several million people have exited through retirement, looking after children full-time, going to college, emigration, illness, incapacity and death. Another several million have entered the workforce from school and college, immigration (including return immigration by people who have been working abroad), return from unemployment or inactivity, and so on. They are very likely doing different jobs from those who have left the workforce. Moreover those who have been employed throughout the period are also likely to be doing different jobs from those they were doing in 2008. Every three months somewhere of the order of 6-700,000 people change their jobs in the UK.
Since 2008 the number in employment has risen by well over 2 million, and the employment rate has risen by more than two percentage points. This all adds to the mix.
On average, the jobs held in 2016 do not pay very much better than those in 2008. But so what? They are different jobs, held by different people. If fewer low-paying jobs had been created, median real earnings would be higher, as logically would productivity, that other much-quoted target variable. But we would have more unemployed and inactive people, while GDP per head of the population would be much the same: only now those in work would be supporting more non-workers, either through family transfers or taxpayer-funded benefits.
What happens to median earnings tells us little about what is happening to individuals. As I said, they change jobs. Such job changes are overwhelmingly voluntary, and typically lead to pay increases: for many that is the motivation for the move. Even those who remain in the same post are often on pay scales which they move up either automatically (in the public sector) or on promotion, even if there is no change to overall pay rates.
The ONS published a fascinating set of figures a couple of years ago that illustrated very clearly that a concentration on workforce median earnings can be very misleading about what is happening to individuals.
Their figures show that between 2013 and 2014, median earnings grew by just 0.1% – the sort of thing Paul Johnson draws attention to. However, if you calculate median earnings for those who were in work in both years, ignoring new entrants and leavers, the median rose by 4.1%, a very much larger figure.
The point is that, in a dynamic labour market with large numbers of new entrants and leavers, and with frequent job changes, what happens to median earnings for the whole workforce tells us next to nothing about what is happening to those continuously in work. The ‘stagnation’ story has been wrongly linked to individuals, and completely swallowed by both left and right – it lies behind Mrs May’s ‘just managing’ trope, of course.
In principle a faster increase in whole-workforce median earnings could be associated with lower pay increases for those continuously in work. Or an actual fall in median real earnings could be compatible with increased earnings for them. This paradox is an example of the dangers of reading too much into aggregates and averages, something I have pointed to on a number of occasions.
But it’s not just an academic point, a piece of intellectual trivia. The danger is that this misreading of the data will lead to policies, such as pushing up the National Living Wage and demonising firms which pay low wages, or display large gender pay gaps, which incentivise firms to start paying more to fewer workers. Then we will be like France, with higher wages and higher productivity, but also higher unemployment – and probably higher taxes.
Would Paul Johnson prefer this?
1 thought on “The UK’s great wage stagnation: beware aggregates and averages”
Comments are closed.
SHARE
Very interesting, as the author points out, this is likely to be used as a justification for yet more intervention in wage-setting.
2 points to raise: Firstly, I think that in any discussion of wages, we need to add in a discussion of costs of living, which as the IEA’s own work has frequently pointed out, are driven up by government interventions (including the NLW/min wage itself) – particularly in terms of housing, childcare, transportation, energy and taxes. It does seem to me that even substantial increases in wages is rapidly outdone by rising costs imposed through government-imposed supply constraints – most evidently in housing. This needs repeating ad nauseam.
Secondly, to what extent do these figures incorporate benefits? As we know, any discussion of wages is futile without also assessing the benefit system. And surely many ‘JAMs’ (who I assume obtain substantial portions of their incomes from this system?) are unable to become ‘managing quite comfortably’ (acronym please?) because of the impact of marginal tax rates, combined with point 1?