Turning the corner: understanding this week’s wage data

After nearly six years in which wage growth has been consistently outstripped by inflation, expectation is building that this week will mark an official end to the pay squeeze. As the chart below shows, the gap between the Office for National Statistics’ Average Weekly Earnings (AWE) and Consumer Price Index (CPI) inflation measures has been narrowing in recent months. When the February 2014 data is released later this week, inflation is expected to fall further below its 2 per cent target while nominal wage growth continues its gradual improvement, meaning the two lines may finally cross over.

Source: ONS, AWE

If so, it will undoubtedly constitute good news. And it will feed into the growing political debate about the state and likely future direction of living standards in the build up to next year’s election. But in thinking about quite how to interpret the data there are three important factors to bear in mind.

Survey coverage

First, different wage surveys tend to tell different stories. In addition to the monthly AWE series, the ONS collects wage data from the Labour Force Survey (LFS) and, on a yearly basis, the Annual Survey of Hours and Earnings (ASHE). Each source has its advantages and disadvantages, but key differences in approach mean that the results can sometimes contradict each other.

ASHE is generally considered to provide the most accurate estimate of gross pay because it is drawn from employer returns for individual salaries. Unlike AWE, which is derived from aggregate paybill data and facilitates the calculation of an average, this approach means that ASHE allows measurement of trends across the earnings distribution, providing an important indicator of the extent to which wage growth is shared or skewed. But it is less good than AWE at capturing bonus payments – an increasingly important source of remuneration for some workers. And, most crucially, it lacks the timeliness of the other two surveys: it will be December before we get the results for April 2014. Unlike the other two measures the LFS is based on employee-reported data, and so is considered less reliable, though it does allow wages to be linked to worker characteristics such as qualification level.

Importantly, all of these measures exclude the growing number of self-employed workers. Having increased by around 1 million between 2002 and 2014, the number of self-employed now stands at some 4.5 million. Yet the earnings of such workers are absent from all the pay surveys. Using less timely data from the Family Resources Survey we know that the typical incomes associated with this form of work fell sharply before and during the early stages of the economic downturn, but we have no real way of knowing what has happened over the last few years.

It is likely that some of the newly self-employed are higher skilled professionals choosing to work on a freelance basis following redundancy, but there is also the suspicion that the recent surge in self-employment reflects a shift towards low paid odd-jobbing rather than genuine entrepreneurial zeal. Either way, the absence of 15 per cent of all workers from these statistics makes it difficult to draw definitive conclusions.

Inflation measures

Secondly, the timing of the recovery in wages will depend on precisely which measure of inflation is used. While the CPI tends to get the most coverage – with the government choosing to switch the uprating of benefits, tax credits, pensions and tax thresholds to CPI in recent years – it takes no account of a range of housing costs. The omission of mortgage interest payments in particular is set to grow in importance in the coming years as interest rates start to rise.

Until recently, the main alternative to CPI was the Retail Prices Index (RPI). It covers a broader range of household costs, meaning that it was traditionally considered the best measure for gauging what was happening to living standards. But it has fallen out of favour, with the ONS deeming that it no longer qualifies as a National Statistic because the formula it uses for aggregating prices (the Carli index) tends to overstate inflation (though note, the government continues to use it in relation to index-linked gilts and bonds).

To bridge the gap between a CPI that excludes important elements of housing costs and an RPI that is no longer considered credible, the ONS has established two new inflation measures: RPI-J (which maintains the RPI coverage but uses a more reliable formula similar to CPI) and CPI-H (which adds an owner-occupied housing element to CPI).

Stitching together these different approaches to wage and inflation coverage, the chart below provides a range of real-terms average weekly wage measures. It shows the very broad range of stories that can be told about recent wage performances. Depending on your favoured metrics, it’s possible to argue both that average wages were falling in 2006 and 2007 – sometime before the financial crisis hit – and that they grew for a period between 2012 and 2013 – even as most declared that the pay squeeze continued.

Notes: Figures show weekly wages recorded in the LFS, AWE and ASHE surveys deflated by CPI, RPI and CPI-H.

Source: ONS

Variations around the mean

The final thing to bear in mind is that averages hide a multitude of divergent experiences. In recent decades, the most obvious variation has been across the pay distribution, with those at the top tending to do much better than those lower down. Those concerned with the experience of ‘typical’ workers have therefore tended to focus on median pay.

As the next chart shows however, the recent pay squeeze has been remarkably evenly felt across the distribution. Between 2008 and 2013, pay fell by 6 per cent (measured using ASHE and CPI), with very little variation by earnings decile.

Real-terms weekly wage growth by earnings decile: 2008-2013

CPI-adjusted

Source: ONS, ASHE

But there have nonetheless been winners and losers in the recent experience. For some – primarily those in the private sector and those who have remained in the same job over the course of the downturn – real wages have either fallen by less than average or have been rising year-on-year. The government has shown, for example, that the wages of people holding the same full-time job for at least a year grew faster than inflation in five of the last six years.

For others – those in the public sector, those affected by a period of unemployment and new entrants to the workforce – the situation has been far starker than the mean or indeed median would suggest. As the chart below shows, differences across age groups have been particularly marked.

Real-terms weekly wage growth by worker characteristic: 2008-2013

CPI-adjusted

Source: ONS, ASHE

Looking to the next few years, no one knows for sure how the historic gap between mean and median pay will pan out during economic recovery, but few would be surprised if it resumes. By way of illustration, consider the chart below. It converts the OBR’s latest projections for nominal mean wages into median, on the assumption that the ratio of growth between the mean and median reverts to the one which held over the period 1997-2007. It further adjusts these figures to produce real-terms projections on the basis of the OBR’s official figures for CPI and RPI inflation and for a constructed version of RPI-J (this assumes – imperfectly, but defensibly – that past relationships hold: holding constant the ratio between annual growth in the RPI and RPI-J in the years ahead, reflecting the relative stability of this ratio over the course of the history of the RPI-J).

Source: ONS, OBR and Resolution Foundation modelling

It shows average CPI-deflated wages growing steadily by 6 per cent between 2014 and 2018, fully restoring their pre-crisis peak. It also shows median RPI-deflated pay continuing to fall to the end of the forecasting horizon, ending the period some 12 per cent below peak. In between, any number of alternative positions can be supported.

Given the unprecedented nature of the squeeze of the last six years, any improvement in real-terms pay – on any measure – is a genuine cause for celebration. But deciphering exactly what this means for wages, never mind the wider issue of living standards, requires careful interpretation.