This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies.Find out more here


To discuss how you and your organisation can get more involved with The Work Foundation, please contact us.

Call 020 7976 3575 or email


Ian Brinkley
Economic Advisor
Ian Brinkley

What is wrong with the average earnings index?

Authors: Ian Brinkley

31 October 2014

The TUC recently organised a seminar looking at why there seemed to be a big discrepancy between settlements data which has been showing median increases of about 2 per cent, and the average weekly earnings (AWE) index which on the latest figures was showing regular pay going up at just 0.7 per cent. We had the usual solid and authoritative report from Income Data Services (IDS), as well as expert contributions from the Office for National Statistics and the Bank of England.

As the report pointed out, the two measures are measuring different things in different ways, so we might expect different results. However, we would normally expect settlements to be lower than average earnings because the latter will include progression and other payments not captured in the annual settlement figure. This was broadly true until 2008, but since then and especially since 2010 the two measures have swapped round.

One possible explanation is compositional change, which is more likely to impact on the average weekly earnings figures because of the way they are constructed. It was pointed out that there has been an increase in employment in some low paying industries and a decrease in some high paying industries.  Compositional change can indeed be important, but I don’t really buy this as an explanation, especially from 2010 onwards.

Firstly, recessions always remove lots of full time and permanent posts and recoveries put them back. I can’t see much difference between 2010 and 2014 and 1993 and 1997 – if anything full time and permanent jobs have recovered a bit faster than they did in the 1990s. It would be interesting to see if the old average earnings index also dived below settlements in the 1990s recession and recovery. If it did, then we might conclude the divergence in measures is cyclical and the settlements and average earnings measures they will move back into line in due course.

Secondly, despite the mythology of this being a low wage job recovery, the vast majority of new jobs have been in the top three occupational categories, with strong growth in professional occupations. If the share of higher skill and better paid jobs in the economy is going up year on year it is hard to see how this could depress the average earnings index.

I also don’t buy the contention that there has been a significant shift of employment towards low wage private service industries.  The broad sector break-down in the AWE distinguishes between the low wage service sectors of wholesale, retail, and hospitality and the high wage financial and business service industries.  Between March 2008 and June 2014 the number of employees in these low wage industries (SIC codes G and I) went down by nearly 100,000 and the number of employees in higher wage business service industries (SIC codes K to N) went up by over 500,000.

These are of course broad brush figures and more detailed analysis may reveal a different picture. Some business service jobs are poorly paid. Moreover, we have also seen significant increases in employment in both education and health care. Neither on average is a low wage sector but areas such as social care do include many low wage jobs. However, it would be difficult to square the overall increase in higher wage employment with big increases in low wage work in education and health. 

Thirdly, the bit of the AWE which is misbehaving the most is high wage services, as the chart below shows.  It would be nice to think that the fall reflects a more realistic view of appropriate reward levels at the top of the finance sector.  There may have been a temporary fall-back in pay levels in the immediate aftermath of the banking crisis but I fear the continuing weakness in 2013-2014 has more to do with the short term restructuring of reward packages for top bankers in anticipation of tax rises and restrictions on bonuses. The complexity of pay in the finance sector helped kill off the old average earnings index and may also be causing the AWE difficulties. Expect that red line in the chart to keep on climbing as City expectations about pay move back towards pre-recession levels.

There may also be a more straight-forward reason. The flexible labour market means that there is plenty of slack and most employers can fill most of their posts without significantly increasing pay. Pay is holding up for those who stay in work but has been driven down for new entrants and for those who lose their job and can only get back into work by accepting lower pay. And with productivity performance so dire, there is little scope for big pay rises even if employers were moved towards generosity.

Perhaps we should reflect whether we either have too much numerical flexibility in the labour market or too little functional flexibility (work practices associated with higher productivity) in the workplace. Reducing numerical flexibility is risky. It might secure higher wages for incumbents but at the price of reducing employment opportunities. We will get better and more sustainable results from focusing on potential solutions for low workplace productivity and providing more and better progression opportunities for the low paid.