I've been taking a look at labour market data, trying to get a handle on reasonably solid and comparable statistics. Lining up several countries at once seemed like an overwhelming proposition, so I kept things down to France and the UK. The sharp contrast between the two is often cited. In fact, it's often from London (whether from pundits or politicians like the PM-to-be, for instance) that Paris gets lectures about the need for "reform". The two countries are similar in population size and demographic profile, making the comparison handier.
National statistics agencies (in this case INSEE and the NSO) carry out an annual Labour Force Survey (LFS). This is a household survey in which the statistics depend on answers to questions - not, in particular, on claimant counts of the officially unemployed kept by employment agencies. You can go to ILO Statistics for a portal to these (and those of other countries) but you'll only see what you'll see on the national websites. OECD collates national LFS data and standardises it into comparative tables. Eurostat features the EU Labour Force Survey, which is carried out by national statistics agencies according to EU standards, and processed by Eurostat. The so-called "key" numbers are similar whichever of these sources you look at. But, when you want to find and compare more detailed data, you find there are slight variations arising from dates and methodology, and above all that what is available - free to the public - differs considerably in organisation and presentation. So in the end I used EU LFS raw data, cross-checking with INSEE and NSO and drawing on their tables wherever appropriate. The year for which it was easiest to get full data plus analysis was 2004, work on 2005 not being quite completely available yet. The numbers are not out of date, because the situation hasn't substantially changed since.
Disclaimer: Though not a statistician, I'm interested in stats here, and how they are used/abused. Lest anyone should think I'm making light of real-life situations - of the problems of the unemployed in France, of the chronically sick in Britain, for example - I'm not. Neither am I expressing an opinion on any of the numerous issues that might be thrown up: debating the pros and cons of part-time work, for example, or prolonged studies for young people, is off topic. I'm not even trying to knock New Labour, or rather, I'll be trying not to. Every time I come near such an issue, I'll put up the warning SPATS : Simply Paying Attention To Statistics.
The Unemployment Rate is the percentage of the unemployed in the labour force, also called active population. The active population is made up of those who have a job (full or part-time, even extremely part-time), plus those who don't but are looking for one; in other words, all those who are considered as being on the labour market.
Labour Force Surveys apply ILO definitions both of :
unemployed : Persons aged 15 years and over who were without work during the reference week, had taken steps to find work during the last month and were available for work within two weeks;
and of
employed : Persons aged 15 years and over who state that they were active or worked for at least one hour during the reference week (the week preceding the survey).
NB: UK labour statistics run from 16+, but this is standardised in EU stats to 15+. Since over 95% of 15-year-olds are at school in both France and the UK (16 is the official school-leaving age in both countries), the inclusion of this cohort has no effect on the unemployment / employment figures. (I spent some time stripping 15-year-olds out of the French stats and saw this did not change the overall picture. In what follows, 15-year-olds are included.)
The UR is not the percentage of unemployed in the total working-age population, just in the part of it labelled "active". Another measure, (call it the "Employment Rate"), is that of those who have a job as a percentage of the total working-age population. Another, called "labour force participation rate" (Participation Rate) is the percentage of the active population (employed + unemployed) in the total working-age population.
If you take out the active population, there's a leftover. I say "leftover" because this group is defined by exclusion: it's made up of those who fit neither the definition of an employed person, nor that of a jobseeker. "Inactive" is a default position. It can cover a variety of circumstances, some of which certainly include work, like looking after a home and children or being a full-time student, or may preclude work because of short or long-term sickness or disablement. An Inactivity Rate can be calculated as the percentage of inactive persons in the total working-age population.
So the working-age population is considered as being divided between the economically inactive and the economically active. The latter are divided between employed and unemployed.
As mentioned above, the working-age population begins, in the stats here, at 15. It runs to 64. Retirement is in fact at 65 for men in the UK, 60 for women; 60 for both sexes in France. It's possible to find some statistics 16-59/64 for the UK, and to take out the 60-64 group for France, but it's easier to use the standardised 15-64 numbers from Eurostat (or the OECD). As with 15-year-olds at school, the effect of including retired people in the working-age population is simply to swell the size of the inactive population in those age-groups.
This is what the working-age populations of France and the UK look like (in thousands):
Extremely comparable. To appreciate the similarity and symmetry, (again, these are actual population levels), see this graph of the major labour market age groups:
The "youth" segment is 15-24, and the "senior" 55-64. The mid-section, from 25-54, is that of prime-age workers. In each of the two countries, this prime-age working population is made up of about 24 million persons. This is the heart of the labour market.
This vitally important group was mentioned in France is not in decline... in this way:
according to the OECD, 87.6% of men aged 25-54 worked in 2004 in France, as opposed to 87.3% in the US. Yet the unemployment rate for that category then was 7.4% in France, and 4.4% in the US.
The numbers (in fact, they should read 86.7% in place of 87.6%, 86.3% for 87.3%, and 4.6% for the American UR, but the point remains unchanged) are in Table C of the Statistical Annex, OECD Employment Outlook 2005. Along with them are the numbers for the UK:
87.5% of men between 25 and 54 were in employment, while the UR for this group stood at 3.8%.
So the employment rate was similar (86.3% for the US, 86.7% for France, 87.5% for the UK), but the UR was in stark contrast: low for the US and UK with 4.6% and 3.8% respectively, relatively high for France at 7.4%.
Here's another look at the employment rate, this time from EULFS:
This is a picture of employment for the working-age population, both sexes (which is why the percentage is lower than for males alone, since more men than women have jobs [SPATS]!). The mid-section, from 25-54 (we'll get back to the younger and older sections later) shows very similar characteristics in the two countries. In terms of numbers, our two prime-age populations of 24 million have about 19 million employed among them. Yet when we look at the UR for the prime-age group:
there's a surprising difference. Intuitively, what we understand when we see this is that a much greater proportion of people of prime working age are out of work in France than in the UK. Yet the two populations are otherwise homogenous, which is puzzling.
However, we can ask this question: is a greater proportion of people of prime working age part of the inactive population in the UK than in France?
The answer is yes, across the board. Don't compare the percentages literally here, though: the inactivity rate is a percentage of the total population, while the UR is a percentage of part of the population, the active part or labour force. This sets out the comparison in a different way:
Based on the inactivity rate (percentage of inactives in the total population) and the percentage of unemployed in the total population, this graph compares the prime-age population not in employment in each country. The proportion of the population not in employment is similar for all age groups but the youngest and oldest (and even then the difference is slight), but the internal composition differs: the ratio of inactivity to unemployment is considerably higher in the UK.
By way of explanation, we might pick this out from the Eurostat Guide to Labour Force Surveys:
In recent decades the borderline between the labour force and what is termed the "economically inactive" population has become increasingly blurred, due to the increasing incidence of part-time and temporary work and the ease with which large numbers of persons (particularly women and young persons in the final stages of their education) repeatedly enter or leave the labour force.
Education is unlikely to keep many prime-age people out of work, but we might wonder if it's traditional in British society for women to go out to work less, for example, thus increasing the inactive population. But EULFS gives the employment rate for women 15-64 as 65.6% (UK) against 57.4% (FR), so that one won't fly, on the contrary.
Wikipedia takes us a little further : Unemployment - Wikipedia
... the definition of unemployment relies on the distinction between inactive and unemployed, a quite subjective measure which can be easily manipulated by policies that do not change the situation of the labor market, but decrease unemployment by shifting people from unemployed to inactive status.
Or, from (France is not in decline...)
Clearly, the line between "unemployed" and "inactive" is not drawn in the same way in each country.
Women working as homemakers (and classified "inactive" [SPATS]!), or persons in full-time education, form major cohorts of the inactive population, but another component is made up of those who are prevented from working by sickness and/or disability. In 2005, I quoted OECD evidence in OECD says ET was right about UK unemployment that shows a high rate of UK sickness/disablement benefits [SPATS]!, especially compared to countries, like France and Germany, habitually noted for their high UR:
Click to enlarge
 Inactivity because of illness or disability
As a percentage of population in each age group, 2003 |
OECD has not updated this since, but there has been little or no substantial change.
| Incapacity Benefit [SPATS]! |
In fact, the problem was already known and recognized by the British government, since a Green Paper (Pathways to Work) about it was published in 2002. Here's a quote:
The number of working-age people claiming incapacity benefits is now over 2.7 million. That is greater than the combined total of lone parents and unemployed people on benefit. About 7.5 per cent of the working-age population now receive these benefits. These numbers have more than trebled since the 1970s.
The Financial Times has commented on the matter (Martin Wolf should read his own paper?). Here's Simon Briscoe in 2004:
Generous disability benefits encouraged more than 500,000 men of working age to leave the labour market over the course of the 1990s according to a Bank of England report.
The Bank says this "remarkable decline" in the male employment rate was particularly noticeable in relation to the least-educated who were under threat of losing their jobs. Those with low-skill levels appeared to have chosen to claim long-term incapacity benefit rather than attempt to re-enter the labour market.
David Webster, a research fellow at Glasgow University, said the paper was important as it showed "the UK has not had the labour market success which the British economic establishment has been claiming; unemployment is far higher than the official statistics indicate; and official unemployment statistics misrepresent the geographical distribution of unemployment". He suggested that policymakers should reassess labour market policies. <...>
The drift to sickness benefits went largely unnoticed in the 1990s as the expansion of employment and sharp decline in unemployment masked divergent trends in labour market participation.
While many women entered the labour market, this analysis shows a decline in participation among men - almost exclusively unskilled. At the same time this group reported increasing long-term illness.
The decline was most pronounced among men aged between 25 and 54, with early retirement explaining "very little of the change".
The Bank of England working paper Briscoe refers to says :
Trends in the UK labour force participation remain puzzling. The 1990s witnessed a rapid expansion of employment in the United Kingdom and an associated decline in unemployment to levels last seen in the 1970s. Yet over the same period, the aggregate participation rate was flat... The decline in labour force participation has been most pronounced among prime-age men <...> very few of these workers entered retirement; the largest flow was into long-term sickness. At the same time, there was a sharp increase in the number of such men claiming disability benefits.
And, conclusion after analysis: The decline in participation is almost exactly matched by a rise in disability benefit rolls.
A 2004 social and geographic study of the phenomenon by researchers from Sheffield Hallam University The Diversion From `Unemployment' To `Sickness' Across British Regions And Districts (a study I referred to in my 2005 diary, and that was summarised by Detlef here), explains:
It is not widely recognised that in the UK two separate benefit systems provide support to non-employed adults of working age.
The first relates to `unemployment'. Since 1996 this has taken the form of Jobseeker's Allowance (JSA). To claim JSA a person must demonstrate that they are available for work and looking for work, and they must `sign on' once every two weeks. For most claimants, including all those claiming for more than six months, JSA is means-tested (on the basis of household income).
The other benefit system relates to `sickness'. Since 1995 this has taken the form of Incapacity Benefit (IB). IB is paid to non-employed adults of working age who have health problems or disabilities. About two-thirds of IB claimants actually receive Incapacity Benefit. The remaining third, with insufficient National Insurance (NI) credits to qualify for IB itself, are counted as `NI credits only' claimants and in most cases actually receive means-tested Income Support with a disability premium. Importantly, Incapacity Benefit is not means-tested except for a small number of new claimants with substantial pension income. Also, although Incapacity Benefit payments start at almost the same rate as JSA they increase after six months and again after twelve months. The disability premium payable to `NI credits only' IB claimants also makes this worth more than JSA.
The workings of the benefits system may seem a long way removed from the measurement of unemployment. The point is however that for many of the longerterm jobless who have health problems, the differential in benefit payments creates an incentive to claim IB rather than JSA.
This is a fairly long-term phenomenon, as the tripling, almost quadrupling, over a quarter of a century shows; it has been happening in other countries too, and has simply been particularly marked in the UK. New Labour is not "to blame" for it either, as this Bank of England graph demonstrates:
 The rise is even sharper before 1997 |
However, the Blair/Brown government has always been perfectly aware of the facts, and has now built policy to bring about a reverse movement from inactivity to activity that is designed to reduce the total number on Incapacity Benefit by at least a million over the coming years. Apart from the Green Paper quoted above, here are some BBC articles that discuss the problem and government reaction to it:
The phrase "long-term jobless" of course lets the cat out of the bag. According to EU LFS 2004 (Eurostat), the total French long-term unemployment rate stood at 3.8%, while the UK was credited with just 0.9%. This apparent success camouflages the uncomfortable fact that a large number of the long-term unemployed in Britain are on Incapacity Benefit and not available for work "within two weeks" - while the French benefits system, in which claiming sickness offers no palpable advantage except in the case of permanent disablement, tends to keep the long-term unemployed signing on as jobseekers, and therefore answering the survey questions in such a way as to be (correctly) categorised as unemployed.
The extent of the UK problem can be seen here:
The graph shows the share, for the prime-age population, taken up by people on IB in the inactive population. For the 40-54 group, the proportion is surprisingly high, around 50%.
What is the effect of this? Supposing a French-type benefits system were in application in the UK, what would the UR look like? The UK government has suggested that 90% of recipients of IB would like to work. Applying 90% to 2.7 m would suggest a drop of 2.4 m. Perhaps more realistically, the government has stated its intention to reduce (over several years starting in 2008) the number of IB beneficiaries by at least a million. A careful appraisal was carried out, catchment area by catchment area, in Beatty and Fothergill's Sheffield study referred to above. The authors estimated the diversion from unemployment into sickness-related benefits, (across the whole working-age population and not just prime-age), at 1.1 m persons.
In the 2004 aggregate working-age population of 38 m, a transfer of 1.1 m from inactivity to the unemployed category would result in (my calculation):
| Total Population | In Employmt | Employmt Rate | Unem- ployed | UR | Particip. Rate | IB | Other Inactives |
| 38 m | 27.6 m | 72.7% | 1.4 m | 4.7% | 76.2% | 2.7 m | 6.3 m |
| 38 m | 27.6 m | 72.7% | 2.5 m | 8.3% | 79.2% | 1.6 m | 6.3 m |
Meanwhile, the French UR (EULFS) for 2004 was 10.3%. Higher than 8.3%, but in a much more similar range.
I'm not attempting to say that UK unemployment is the same as French (or to deny the importance of unemployment in the French economy, [SPATS]), but to point out the volatility of the UR metric and its susceptibility to influence by factors such as benefit and classification systems. In this simulation, a 12% fall in the inactive population leads to a 77% increase in the UR. This may only be a rough calculation, but it's based on serious statistics and on considerable discussion of long-term unemployment issues in the UK in the media, by academics, by organisms like the OECD, and by government.
Beyond the BBC articles and other sources already linked to above, see also:
Labour Market Trends April 2002: Sickness and Unemployment
and this study (hat-tip to Detlef) :
Inactivity, Sickness and Unemployment in Great Britain: Early Analysis at the Level of Local Authorities
There is so much the UR doesn't say. Here we have looked at two blocs of prime-age labour of similar size (24 m) and a similar employment rate - but a contrasting UR that doesn't tell the whole story. Not just the whole human story, the history and geography and sociology of labour market disadvantage, (see Beatty & Fothergill op cit for an overview), but not even the proper statistical story.
In the second part of this diary, we'll look at the "Youth" and "Senior" segments of the working-age population.
:: :: :: :: :: :: :: ::
Previous European Tribune writing on employment/unemployment :
What's really wrong with the Eurozone labour market? by Colman June 2005
The Party line by Jerome a Paris Aug 2005
Comparing unemployment statistics by Colman Sep 2005
OECD says ET was right about UK unemployment by afew Oct 2005
European Unemployment by TGeraghty Nov 2005
Youth unemployment by Jerome a Paris Jan 2006
Graphic statistics by Alexandra in WMass Jan 2006
More employment statistics by Jerome a Paris Jan 2006
The Inverted Example of Spinning Jobless Statistics by DoDo Feb 2006
Actual facts about the French labor market by Jerome a Paris March 2006
European Employment: Some Good News by TGeraghty Sep 2006
What's unemployment, again? by Jerome a Paris Jan 2007
Fascinating employment and unemployment numbers by Jerome a Paris Feb 2007
CS Monitor: stats are only good when they suit us by nicta March 2007
(I'm sure there are more, please add links in comments and I'll edit the list).
