Blog

AllEarningsLabour market

The Great Wage Slowdown in Britain: Who are the winners and the losers?

The Great Wage Slowdown
Figure 1 reports trends in real median hourly wages of all employees and median weekly earnings of full-time employees by gender. From 1997 to 2008, median wages grew slowly (by historical standards) but steadily. After the Recession hit in 2007/8, median wages began to steadily decline in real terms. This period—which we are currently in—has become known as the ‘Great Wage Slowdown’. By 2015, real median wages were at roughly similar levels as they were a decade prior, eliminating most of wage growth accumulated 1997 to 2008. Much has been made about this median figures in the press. But as any statistician will tell you, ‘averages are only averages’. In this blog post, I present some evidence around this trend, focusing on occupations—that is very detailed classification of the sorts of jobs people do (roughly 350 occupations). This enables an analysis on which sections of the labour market account for stagnating earnings. The findings may surprise some.

Figure 1. Median wages and earnings trends by gender
Fig 1
Source: Annual Survey of Hours and Earnings. All employees on an adult rate whose earnings are not affected by absence. Weekly pay for full-time employees only.
Notes: Earnings deflated by 2015 CPI. Vertical dashed line indicates 2008.

Occupations and the stagnation of wages
Occupations relate to the shifts in the aggregate wage structure in two main ways: First, by shifts in the occupational structure. For instance, we are often told that there has been a proliferation in low wage occupations, and a ‘hollowing’ out of middle-income occupations. This could provide one route to why median earnings are stagnating—the ‘median occupation’ is becoming a low-paid one as middle-paying ones disappear. Second, the wages of occupations themselves could be stagnating or falling. We are often told that previously good jobs are now bad jobs due to a bidding down of their wages such that already low-pay or middle-paying occupations pay worse now than they did a decade or so ago.
To shed light on these sorts of claims, Figures 2 and 3 examine what has happened to the occupational employment structure (Figure 2) and occupational wage structure (Figure 3) before and after the Recession in 2007/8. Figure 2 shows that in the run-up to the Recession there was indeed a continuation of the longer-term trend (beginning in the early 1980s) of a decline in middle-paying occupations. In the post-Recession era, shifts in the occupational structure were rather flat by contrast. This implies that a shifting structure of the labour market cannot account for stagnating wages as is sometimes claimed.

Figure 2. Changes in the occupational structure
Fig 2
Source: Annual Survey of Hours and Earnings. All employees on an adult rate whose earnings are not affected by absence.
Notes: Earnings deflated by 2015 CPI. Each data point represents an occupation. Trend line is fractional polynomial weighted by occupation size in the base period.

Turning to shifts in the occupational wage structure, we can see that in the run-up to the Recession, almost all occupations received positive—and quite equitable—wage growth. If anything, wage growth was tilted slightly in favour of low-paying occupations—likely those most affected by the introduction of and successive hikes in the minimum wage. In the post-Recession era, the trends are rather different. Almost all occupations suffered real wage declines—with the big exception of low-paid occupations whose wage growth was close to zero (but not negative). What is less well-known, is that real wage declines were tilted against the highest paid occupations the most.

Figure 3. Changes in the occupational wage structure
Fig 3
Source: Annual Survey of Hours and Earnings. All employees on an adult rate whose earnings are not affected by absence.
Notes: Earnings deflated by 2015 CPI. Each data point represents an occupation. Trend line is fractional polynomial weighted by occupation size in the base period.

Is inequality increasing?

The net effect of these trends in the occupational employment and wage structures was a slight fall in overall wage inequality. As can be seen in Figure 4, which shows several indicators of wage inequality, inequality between the bottom 10 per cent of earners and the median has fallen (the P50-P10). This is because the low-paid occupations saw relatively strong wage growth prior to the Recession, and their earnings barely fell after it—while those in occupations above the lowest-paid ones saw their wages fall in the post-Recession era. At the same time, those in the highest paid occupations saw their wages fall the most after the Recession—more than those in low-paid and middle-paid occupations. In other words, the bottom tail of the aggregate wage structure has compressed slightly, while the previously growing differential between the median and the top 10 per cent has levelled out (the P90-P50) because, as we saw, many high-paying occupations saw their real wages decline more than almost all other occupations. Further analysis reveals that these shifts in the occupational wage structure almost entirely from these shifts in the aggregate wage inequality. The changing composition of employment had barely any effect.

Figure 4. Wage inequality trends
Fig 4
Source: Annual Survey of Hours and Earnings. All employees on an adult rate whose earnings are not affected by absence.
Notes: Earnings deflated by 2015 CPI.

Conclusions
In contrast to stories of soaring inequality, wage inequality in contemporary Britain has been rather flat. This is because the biggest losers in the Great Wage Slowdown have been the highest-paying occupations whose wages have declined the most and the biggest winners—if you can call them that—are the lowest-paid ones—likely because of the wage floor in the form of the minimum wage. But let’s be clear—there are no real winners here—wages are not growing for everybody. But it is true to say that differential wage declines have halted inequality trends somewhat.

Back to blog