Toggle high contrast

Recession Report August 2009 - Pay and the recession

Issue date

Headline statistics

The latest labour market figures cover the period April to June 2009, and show that:

  • Unemployment reached 2,435,000 - this is 54,000 higher than last month's figure, 220,000 higher than the January - March quarter and 750,000 higher than the same quarter last year.
  • The unemployment total has now risen for 13 successive months. With the exception of one month, unemployment has been rising since the start of 2008 - before the recession is normally taken to have started.
  • Women's unemployment reached 942,000 - this is 19,000 higher than last month's figure 63,000 higher than the January - March quarter and 248,000 higher than the same quarter last year. Women's unemployment is highest it has been since May 1994.
  • Men's unemployment reached 1,492,000 - this is 34,000 higher than last month's figure, 157,000 higher than the Jan - Mar quarter and 502,000 higher than the same quarter last year. Men's unemployment is now the highest it has been since July 1996.
  • The unemployment rate reached 7.8 per cent. This is up 0.2 points on the previous months, 0.7 points on the previous quarter, 2.4 points on the same quarter last year. Rates were last this high in December 1996.
  • There were 28,933,000 people in work. This is 65,000 fewer than in the previous month, 271,000 fewer than in the previous quarter and 573,000 fewer than the same quarter last year. This is the biggest reduction since the current series of data began in 1992.
  • The working age employment rate was 72.7 per cent, down 0.2 percentage points on the previous month, 0.9 on the quarter and 2.0 points on the same period in 2008. This is the lowest rate since the second quarter of 1997.
  • Long-term unemployment will continue to rise for a long time. This month, for the first time in this recession, the number unemployed over 6 months passed the 1 million mark, reaching 1,002,000 - it has been below that level since July 1997. Except for one month, long-term unemployment has been rising since the start of 2008. 35.6 per cent of long-term unemployed people are women (357,000), compared with 38.4 per cent of all unemployed people.
  • There are now 533,000 people who have been unemployed for over 12 months and 232,000 who have been unemployed for over 24 months. These are the highest figures since June 1998 and July 2000 respectively. At one point (August 2004) there were just 119,000 people who had been unemployed over 24 months.

Labour market slack

During a recession, the rise in unemployment results from a deficiency of demand for labour - a 'slack labour market '. This slack shows up in other ways as well; the most noticeable is that there is an increase in the number of people who have got part-time jobs, but would have preferred to work full time - a year ago there were 685,000 people who said they were only working part-time because they couldn't get a full time job, today that figure has risen to 964,000.

Measuring labour market slack is difficult. One approach is what is known as the 'want work rate'. The want work level takes the number who are unemployed and adds the number in the Labour Force Survey who are 'economically inactive' (students, people looking after their family or home, long-term sick and disabled people) who say that they want a job. The 'want work rate' expresses this as a proportion of the population in much the same way as the figures for the employment and unemployment rate. In the latest figures, the want work level is about 4.5 million, and the rate is one-in-seven.

But that figure doesn't take into account involuntary part-time work. One response to this is to add the two figures together. We are reluctant to do this - adding together figures for people who don't have any work and people who don't have enough work ends up, in our view, not properly measuring what we are interested in. So we prefer to quote two sets of figures - the want work rate and the involuntary part-time figure which gives a rounded picture. We have also published an article on the Touchstone Blog[1], which further explains our approach.

Two measures of unemployment

A related issue has appeared in a lot of the media coverage of the latest data - the continuing difference between the figures that use the ILO measure of unemployment and the figures based on the numbers claiming Jobseeker's Allowance. A quarter of a century ago, the figure that appeared in the headlines was the 'claimant count', but in the 1980s Government interference in the unemployment figures made many journalists suspicious of any measures that could be controlled by Ministers, and tended to switch their reporting to the ILO measure which is more transparent. In the latest release, the ILO figure (for Apr - Jun) is 2,435,000, while the claimant count (for July) is 1,582,700.

ILO and claimant unemployment, Mar-May 1992 - Apr-June 2009

graph showing a downward trend

Some of the coverage of this issue seems to suggest that this large difference is a new development, but as the chart shows, it is not. As we have reported previously, our view is that it is significant that the divergence between the two measures began in 1996, at the time when the Jobseeker's Act 1995 came into effect, replacing Unemployment Benefit with Jobseeker's Allowance.

The contributory version of Jobseeker's Allowance is harder to qualify for than the Unemployment Benefit (UB) it replaced. It takes more National Insurance Contributions to qualify, and it only lasts for 6 months, as opposed to 12 months of UB. This means that unemployed people who have sufficient savings, or a partner in employment, and therefore do not qualify for the means-tested version of JSA, are less likely to qualify for contributory JSA in the first place, and disappear from the claimant count after six months, instead of 12.

In addition, the fact that the proportion of women in the workforce has risen means that the proportion of redundant workers who have a partner in employment has also risen, so the impact of JSA on the claimant count is more pronounced than it would have been if it had been introduced during a period when fewer women were in paid employment.

This probably explains why the ILO measure and the claimant count first began to diverge. But this is not the explanation that has been picked up in media reports; instead, they have concentrated on the phenomenon of middle class unemployed people deciding that JSA isn't worth claiming.

Now, it is true that most people have no idea how demeaning and dispiriting the process of claiming JSA can be. One can well imagine that many people who are at the receiving end of this process for the first time may well feel that, for £64.30 a week, it simply isn't worth all the trouble. Furthermore, the relative value of the benefit has declined over the last 13 years - as we have reported previously, simply to be worth the same fraction of average earnings as it was in 1997, JSA would have to be increased by £15 a week.

All this is likely to be having some impact. It may mean that unemployed people who have other resources may decide not to sign on. But we would want to see more research before accepting that this is an important factor. Most unemployed people simply don't have sufficient savings or a partner who is earning enough for them to have the luxury of deciding not to claim - the divergence is more likely to be a result of many unemployment people not being eligible to do so.

Vacancies and redundancies

The statistics for the number of vacancies registered with Jobcentre Plus are mixed at present. The monthly estimates increased slightly in the last month, with provisional figures for July showing an increase from 427,000 to 430,000. On the other hand, the three-month rolling average figures for May-July show a fall from 431,000 to 427,000; the vacancy ratio (the number of vacancies for every 100 employee jobs) fell from 1.7 to 1.6.

The redundancy figures show a slight improvement. In Apr - June the number of redundancies was 277,000, down from 286,000 in Jan - March. The redundancy ratio (the ratio of the redundancy level to the number of employees in the previous quarter) also fell, from 11.3 per thousand employees to 11.0. But these are still high figures: a year ago the redundancy level was 127,000 and the redundancy rate was 5.0.

Disabled people and the recession

In last month's Recession Report we reported on the recession and disabled people. Since then we have received new data that suggests that the picture painted in official data, of steady progress in the employment of disabled people, is a little too rosy. Nigel Meager, of the Institute for Employment Studies, has kindly pointed out to us the impact on the figures of using different definitions of disability.

The government figures use the Disability Discrimination Act definition, and people in this group, have seen their employment rate increase significantly in recent years. On the other hand, if you look at people who are classified as having a 'work-limiting disability' their employment rate actually fell between 2005 and 2008. There is an extended discussion of the significance of these figures at the Touchstone Blog[2] - in it, we conclude that, if we had had these data when writing last month's Report, we would have changed our analysis: while we would still have emphasised the good news that the disability employment and unemployment gaps (using the DDA definition) have continued to shrink during the recession, we would have added that we also need to know what has happened to people with a work-limiting disability to get a fuller picture.

Pay and the recession

In the second section of this month's Recession Report we consider how pay levels have been affected during the recession. During the downturn many employers' organisations have been keen to play up the impact that the downturn has been having on pay. For example, in March the CBI announced that the 'overwhelming majority'[3] of its members would freeze pay this year, and the British Chambers of Commerce (BCC) recently concluded that '58 per cent of companies intend to freeze wages in 2009'[4], and that 12 per cent would be cutting staff pay. There have also been increasing calls for public sector wage freezes.[5]

In this section of the report we therefore undertake an analysis of how pay has in fact been affected by the recession so far, and consider what the economic impacts of widespread pay restraint could be.

Earnings and inflation

The difference between pay rises and earnings growth[6]

Pay rises and earnings growth are two different measures of employees' income, with earnings growth generally slightly higher than median pay settlements. The reason for the gap is that the measure of average earnings growth captures the totality of changes in all elements of pay, such as overtime pay, progression increases and pay rises due to promotions or changes in the composition of the workforce (for example a reduction in low-paid unskilled jobs and a greater reliance on skilled workers would lead to an increase in average earnings across a workforce). In contrast, pay settlements, recorded by organisations such as IDS and Labour Research Department, measure consolidated increases in basic pay and performance-related pay rises, but do not take account of wider changes which impact on aggregate employee earnings.

During most of the last twelve years, growth in the Annual Earnings Index (AEI) has outpaced RPI inflation. ONS[7] note that this was made possible mainly through increases in productivity, enabling employers to pay higher wages while not increasing their prices by the same extent to finance their wage bill. But between September 2006 and November 2008 prices rose faster than earnings, reflecting high energy prices and the consequent impact on other commodities during this period. But as the downturn has taken hold house prices have fallen and credit flows have been severely reduced. As a result, RPI is now in negative growth, while earnings growth remains at around two per cent. It is however important to note that negative RPI has largely been driven by falling mortgage repayment costs, so when housing costs are removed, the gap between AEI and RPI narrows significantly.

AEI (three month average), RPI (all items) and RPI X (excluding mortgage payments) annual growth, May 1997 - June 2009

graph showing all items RPI falling

This pattern of negative growth in RPI is different to previous recent recessions, where inflation rates were much higher at the start of the economic downturns. But, as we highlighted in June's Recession Report[8], ONS have shown[9] that although recessions commonly cause falls in inflation, it is normal for real earnings to continue to grow. By deflating earnings by the RPI to chart real earnings over time, ONS showed that during the 1980s and 1990s real earnings growth was negative during the first two quarters of the recessionary periods, before returning to positive growth. Their analysis of data for the first three quarters of this recession shows that the current recession is also following this pattern. The graph above shows earnings at 2.5 per cent and inflation (less housing costs) at 1 per cent, so the economy has already returned to modest real earnings growth.

The recession and earnings

The Annual Earnings Index

ONS's Annual Earnings Index (AEI) measures earnings growth (not levels) and considers how current earnings compare to those of a year previously. Data are available including and excluding bonuses. ONS note that when bonuses are included, the series can be influenced by a small number of companies that pay high bonuses, and that by removing bonus payments the series becomes more stable, giving a better indication of trends in pay growth. In addition, the AEI series is seasonally adjusted. This removes effects such as the timing of annual pay settlements, seasonal patterns of part-time work and overtime payments.[10]

The latest earnings data are for the three months to the end of June 2009, and as discussed in the previous section the average increase excluding bonuses is 2.5 per cent. This is the lowest since the figures started distinguishing between increases with and without bonuses in 1997, which can be seen below.

Whole economy, seasonally adjusted earnings excluding bonuses, three month average, Q2 1997 - Q2 2009

graph showing a downward trend for quarter 2 of 2009

After showing negative growth of -0.3 per cent in the three months to March 2009, after the collapse of city bonuses, earnings including bonuses have been showing a year on year increase for the last three months.

Whole economy, seasonally adjusted earnings including bonuses, three month average, Q2 1997 - Q2 2009

graph showing a big decrease in quarter 2

Public sector earnings excluding bonuses are increasing faster than private sector earnings excluding bonuses. However, for significant periods since 1997 private sector earnings have been increasing faster than those in the public sector.

Change in public and private sector AEI (excluding bonuses), seasonally adjusted, three month average, Q2 1997 - Q2 2009


Average earnings growth remains stronger in services than manufacturing, although rates of increase in earnings in both sectors have declined during the recession.

Change in manufacturing and service sector AEI (excluding bonuses), seasonally adjusted, three month average, Q2 1997 - Q2 2009

graoh showing manufacturing excluding bonuses falling

Particularly in manufacturing, these declines in average earnings are likely to reflect increased the incidence of short-time working and involuntary part-time work.

The recession and pay

The IDS pay databank

The settlements recorded in the IDS databank include both the pay rises that are collectively bargained between unions and employers, and also non-negotiated pay awards. The data are gathered directly from employers and unions, and are based on asking about the percentage rise applied to wages or salaries in the annual pay review. Where the percentage rise varies (for example differing according to individual performance) the figure recorded is the average increase, or the increase received by most employees. The median settlement level is the point at which half of all settlements are above and half are below.

The latest analysis from IDS[11] covers the three months to the end of June, and shows that the median pay settlement remains at 2 per cent - the same as the last two months. The distribution of settlements shows that while a third are for freezes, half are worth 2 per cent or more. This has led IDS to conclude that this level is becoming the floor for pay rises. Their pay report notes that the latest EEF survey has shown similar findings, with companies with freezing pay or paying rises of 2 per cent or over.

Median pay settlement level, reproduced with permission of IDS

graph showing the median pay settlement

The IDS analysis also notes that if freezes are excluded from the analysis, the median settlement level is 2.5 per cent, which provides an indicator as to the decisions being taken by organisations not freezing pay.

The IDS analysis further demonstrates that pay settlements vary significantly across sectors. For example, while EEF report that in the three months to June two thirds of engineering deals were pay freezes[12], and British Airways pilots have accepted a 2.6 per cent pay cut, Heinz employees have received a 2.8 per cent increase, and staff at Imperial Tobacco have benefited from a 3.35 per cent award. IDS show that while there are some sectors that are very severely affected others are barely affected at all: a recent report IDS described this as 'a spectrum from the car and component sector at one end through to the oil and energy/utilities companies at the other', concluding that 'despite the gloom, it is not a universally bleak picture'.[13]


So, while some employers have had to resort to freezing or occasionally cutting staff pay, the majority have not had to take this tough choice. With inflation low and jobs scarce it is no surprise that pay settlements and earnings growth are lower than in recent years, but the median settlement remains at 2 per cent and earnings are still rising.

There are important reasons for maintaining earnings during a recession. Firstly, cutting earnings unnecessarily has been shown to have a negative impact on productivity and worker morale, which would make it even harder for businesses to return to growth and to take advantage of the recovery.

Furthermore, pay cuts and freezes would have a negative impact on economic demand. Cutting take home pay even more, in a context of the reductions we have already seen as a result of rising unemployment, could lead to an even deeper cuts in consumer spending - further aggravating the downturn.

The same arguments apply to public sector pay- as Professor David Blanchflower has recently argued[14], cutting spending could easily prolong the recession and limit recovery, as a result of lost demand.

Trade unions do recognise that in companies facing particular difficulties agreeing to go without, or limit, a pay increase, may be the best way to save jobs. But in general, but a uniform pay freeze would be disastrous for the economy and unfair to employees.

[1] Want Work Rates and the Recession, 12th August 2009.

[2] Disabled People and the Recession: new information, 29th July 2009.

[3] Sunday Times, Pay Rise Looms for Millions, 1st March 2009.

[4] British Chambers of Commerce, Majority of firms plan to freeze wages as recession continues to bite Press Release, 27th April 2009.

[5] For example, Bundred S, We've had years of growth - so let's not be afraid of cuts The Observer, 5th July 2009.

[6] This section draws upon the following article: Miller S (2005) The difference between pay settlements and earnings growth Labour Market Trends, February 2005.

[7] ONS (1999) Social Trends: Average earning index (AEI) and retail prices index (RPI): 1991-1999

[8] TUC Recession Report Special June 2009.

[9] ONS (2009) The impact of the recession on the Labour Market Newport: ONS.

[10] This section draws upon the following article: Miller S (2005) Ibid.

[11] IDS Pay Report, August 2009, 1030.

[12] EEF, Manufacturing pay deals chart course of downturn Press Release, 17th June 2009.

[13] IDS (2009) Focus on Recession London: IDS.

[14] Blachflower D, And next for Britain, the semi-slump The Guardian, Tuesday 14th July 2009.

Enable Two-Factor Authentication

To access the admin area, you will need to setup two-factor authentication (TFA).

Setup now