Sunday, 26 February 2017

Wage Inflation and Unemployment in the UK

As a further follow up on the NAIRU discussions mentioned in my last two posts, I have looked at the historical relationship between wage inflation and unemployment in the UK. 

The first of the two graphs below shows annual change in average earnings against unemployment from 1973 to 2016.  The second graph shows the change in the change, what we might think of as the acceleration of wage inflation.  I have divided the points into three periods: 

1973 - 1979: the period before the Thatcher government
1980 - 1991: up to inflation targeting
1992 - 2016: the period of inflation targeting

The trend line covers the whole period.

 A few points are worth making here.

Both charts show quite a wide scatter.  Although both show a vague trend across the whole period, the correlation is not particularly strong in either.

Although the correlation is somewhat weak, the charts suggest that both the level of wage inflation and its acceleration are negatively correlated with unemployment.

There appears to be a somewhat better correlation if we look at the periods individually.  Each individual period generally continues to show the same direction of correlation.  However, the steepness declines over time, both in relation to wage inflation and its acceleration. 

For the period of inflation targeting, there appears to be very little correlation between unemployment and the actual level of wage inflation, whilst there is still a slight correlation with its acceleration.  This is probably the opposite of what I would have expected.

I addition to displaying the steepest relationship, the plot for the 70's also shows the greatest scatter.  This probably reflects various external factors such as the oil price shock and the social contract.

The only really clear conclusion here is that the relationship between these variables is a complex one, but it does appear that there is some connection between unemployment and not only, wage inflation, but also its acceleration.


  1. One thing that really struck me in SWL's second post about the NAIRU was where he wrote:- "‘b’ stands for a collection of slow moving variables. These could include a measure of union power, or how mobile labour was, or the degree of monopoly in the goods market."

    I read that as being a sobering assessment that any attempt by workers to shift returns to labour rather than to owners would probably raise the NAIRU and so condemn more people to unemployment. But your data doesn't seem to show a lower NAIRU after the unions were broken than there was in the 1970s when union power was at its strongest.

    1. I would say the change in that relationship between those periods was in large part due to those losing their jobs becoming structurally unemployed and not really able to cover the demand for labour that later pushed up wages. To me this just illustrates that even if we do want to use the NAIRU concept, we can't really view it as being a fixed point.

      Whilst, on the whole, I'd say that more unionisation does increase inflation pressure and that the reduction in unionisation plays a role in the low wage inflation of the most recent period, it is not a straightforward story.