Bank of England shows jobs gains are concentrated in lower-paid work
ONS released today another upbeat set of employment numbers, as well as an underlying nominal earnings figure above 2% for the first time in approaching four years (see my colleague Richard Exell’s post).
But in the meantime the Bank of England issued a sobering analysis of the nature of recent employment gains. Figure 4.6 of today’s Inflation Report (below) suggest that recent jobs gains have been in types of work that pay less than the norm. The results lead into the inevitable discussion of weak productivity, and a downgrading of expectations into the future.
Chart 4.6 The changing composition of employment weighted on average wage growth in 2014
(see end of post for the full explanatory notes to this chart)
Each segment of the columns reflects an estimate of the contribution of the various employment characteristics to four quarter wage growth. So for the most recent figures, the Bank concludes:
… the changing composition of employment growth — including the mix of occupations, industries, ages and job tenures — could explain around 1 percentage point of the recent weakness in average annual earnings growth. (p. 37)
It is not news that compositional factors are important in recent pay outcomes, as the recent IDS report for the TUC showed. The Bank also emphasises (Chart 3.6) that “recent employment growth has been concentrated in lower-skilled jobs”. Nonetheless, the particularly striking facts of the Bank analysis above are first the scale of the present negative effect and second that for the first time since 1996 no compositional factors are acting in a positive direction.
It is a difficult to figure out the full implications of the analysis, particularly over time – but the suggestion is that the recent rapid growth in jobs is has been heavily skewed to lower paid industries, qualifications, ages, tenure, occupations etc.
Beyond the social implications of an apparently significant shift to lower earnings work, are the usual macroeconomic questions of cause and effect.
For the Bank, the shift in the composition of the labour force helps explain the very weak productivity outcomes, especially over the past two years (coinciding particularly with the shift to lower skilled work). Looking ahead they continue to predict a recovery in productivity growth, but it is delayed “as compositional effects drag on productivity over the next few quarters”. (Productivity growth in 2015 is revised down to ¼ from ¾ per cent and productivity growth in 2016 to 1¼ from 1¾ per cent, and then growth in 2017 in unchanged at 1¾ per cent. ) (The basic underlying reversion to business as usual also explains why the Bank sees recent weakness in GDP and CPI inflation at and potentially below zero inflation as temporary.)
But helping explain the reasons for recent unprecedented lows in wages growth does not necessarily resolve the productivity puzzle. It does not explain why the economy has shifted to this type of work. The implication is that the Bank thinks that increases in the supply of labour are concentrated in lower paid categories. The alternative is that this is where the economy is creating new work, and it is the demand for lower quality and lower paid work that is dictating outcomes. As we have argued (e.g. here), the government’s policies over the past five years have dictated low growth outcomes that the labour market has accommodated through reduced earnings growth. The Bank has helped to understand better how this accommodation has operated, but this can still be as effect not cause. The revival in wages would then be contingent on the conditions of demand, rather than the always-anticipated but never-yet-happened reversion of the conditions of supply.
Background to chart
Sources: Labour Force Survey and Bank calculations.
(a) Estimates are shown relative to their averages over 1995 Q2–2014 Q4. Estimates of the effect of individual and job characteristics are derived from a regression of these characteristics on levels of employee pay using Labour Force Survey data. The estimate of the total compositional effect is obtained by combining these estimates with changes in the composition of the labour force.
(b) Other includes gender, region of residence, whether working full-time and whether in public sector employment.