Reading the runes on Britain’s labour market

The UK labour market has shifted into different gears over the last decade. The first – and most dramatic – move happened early in 2009 when real pay began to fall for the first time in decades. Early in 2012 the jobs market bottomed out and five years of nearly uninterrupted employment growth began. 2015 saw real growth pick up for the first time in half a decade, while 2017 saw the return of the pay squeeze. Are we seeing the beginnings of another such shift?

Predicting such shifts are hard. To do so one must pick through a plethora of indicators and decide what (to borrow Nate Silver’s dichotomy) is signal and what is noise. Today’s labour market statistics offered up a bit of both.

First the noise. Employment increased by around 100,000 meaning that the country returned to its record high employment rate. On the face of it this may suggest that expectations of a slowdown in jobs growth are wrong. However a look at the single month employment figures suggest it may be too soon to dismiss the idea that employment has plateaued.  The chart below compares how the headline rate and the single month rate change month-to-month. Two things stand out. The first is that the monthly figure varies a lot more. The second is that the variation in the monthly figure has increased over the past year. Strip away this noise and the monthly rate could revert to the mean next month, bringing the headline rate back down again.

Today’s earnings stats appear to be sending a clearer signal. Real average weekly earnings growth remains at -0.5 per cent and has hovered around this level for over six months. But this is not a good guide for what will happen in future. There are two reasons for this; the first mechanical, the second substantive. Mechanically the ONS calculates real pay growth as the percentage change in the three month average of real pay compared to an average of the three months a year previously. This matters as over the past twelve months real pay has fallen from a high of £462 in November 2016. Real pay is falling because recent figures are set against this late 2016 ‘pay peak’. Once it falls out of the data we are likely to see real pay turn positive. This will happen even if pay does not grow for the next four months, as demonstrated in the figure below.

The more substantive reason why real pay growth is likely to strengthen over the coming months is because inflation appears to have peaked. As the impact of the post-referendum devaluation diminishes inflation is likely to move closer to 2 per cent. We project the squeeze to ease to between -0.5 and -0.1 per cent in January.

The long-term outlook for pay depends on the UK rectifying its dismal productivity performance. Today’s stats provided an (albeit faint) signal that this may be happening. Although employment rose the number of hours worked remained flat. Based on this and GDP projections we estimate that quarterly productivity growth will rise to roughly 1 per cent in Q4. This would be the highest recorded since the financial crisis and would translate into annual productivity growth to the end of 2017 of 1.3 per cent.

Will this uptick in productivity be maintained? Further labour market tightening suggests it might. Vacancies are at an all-time high and shortages in many sectors are likely to help boost pay and encourage capital investment, which should have a knock-on effect on productivity.

While such signals are encouraging, they could also turn out to be just noise. Productivity growth of 1.3 per cent would still be well below the pre-crisis average of 2.3 per cent. And while there are plenty of reasons to be optimistic, uncertainty surrounding the Brexit negotiations, or any indication that the UK is heading for a disruptive exit, could prove to be more important.

Turning points are easiest to spot after the event, but that doesn’t mean we shouldn’t try. A bit like rowing a boat, you spend most of the time looking backwards but you need to look over your shoulder once in a while, otherwise you may miss the waterfall.