Navigation – Plan du site

L’expérience du chômage : quelles leçons pour la France ?

Unemployment experience: some implications for France
Richard Layard
p. 23-28


Le taux de chômage dépend des caractéristiques de l’offre sur le marché du travail. La question principale réside dans la façon dont sont traités les chômeurs. Dans les années 1990, la plupart des pays ont connu une période d’essor économique, mais un niveau élevé de vacance d’emplois s’accompagnait d’un niveau élevé de chômage persistant. Certains pays comme le Danemark, les Pays-Bas ou le Royaume-Uni ont alors pris des mesures décisives pour activer leurs chômeurs, en leur apportant davantage d’aides mais en exigeant aussi d’eux une réponse active. Dans la période de hausse de 2000, ils ont eu des taux de chômage inférieurs de 3 à 4 points. Mais durant la même période, d’autres pays comme la France et l’Allemagne ont fait moins pour augmenter la conditionnalité des bénéfices : ils ont eu des taux de chômage identiques dans les années 2000 à ceux des années 1990. Une baisse soutenue du chômage aux niveaux d’avant la crise demandera des politiques d’activation plus importantes des demandeurs d’emploi. Certains devront accepter des emplois moins attractifs, car le chômage diminue le bien-être et même un moins bon emploi vaut mieux que pas d’emploi du tout.

Haut de page

Texte intégral

1The new OECD Job Strategy is a delight to read for us old hands, especially if we can assume (without anyone exactly saying so) that the most important Pillars in the strategy are those that come first. I welcome the change from the previous strategy, where demand and supply sides were jumbled up, to a strategy where (after macroeconomic stability) supply-side issues figure at the outset. So let me focus mainly on the supply side of the labour market and, especially, on the problem of mobilising the unemployed in France. I will then say a little about wage flexibility, which should be central to the demand side, and about skills.

Mobilising the unemployed

2If we look across countries or over time, we can see clearly that in the medium term, the demand for labour responds to the effective supply. Before the Pilgrim Fathers landed in New England, there were no jobs there, but after they landed, jobs sprang up. Or, to take a less obvious example, the labour force in Britain has grown by 212 percent since 1851 (see Mitchell, 1992; Officeof National Statistics, 2005); over the same period, the number of jobs has grown by 212 percent. So – ignoring the business cycle – a market economy always provides more jobs, if there are more people “effectively” seeking work. The issue is how to increase the “effective” supply of labour.

3There are two big issues here. The first is how to mobilise unemployed people so that they are more effective fillers of vacancies. The second is how to get more people into the labour force. I will start by taking the labour force as given and talking about unemployment. The issue here is: How many people can be employed before inflation starts increasing and economic recovery comes to an end? The problem is that as the demand for labour expands, it becomes increasingly difficult for employers to fill their vacancies. So the number of unfilled vacancies increases, and wages get bid up until inflation starts increasing. At that point we have reached the lowest sustainable level of unemployment.

4How much unemployment there is at that point depends mainly upon how effective a system we have for mobilising the unemployed to fill the available vacancies. That is the view which a number of us formed some twenty years ago, (see for example: Layard, Nickell, Jackman 1991; Oecd, 1990) and I can illustrate the point from the history of what has happened since then, since it provides a perfect real-world experiment to test the theories which we were proposing in the 1980s.

A natural experiment

5Around 1990 there was a European boom. Vacancies in most countries rose to levels as high as in the previous boom that had ended in 1980. But unemployment remained high in every country. In some countries this paradox of high vacancies and high unemployment led to a public debate in which it became an accepted proposition that it mattered how unemployed people were treated. It was realised that if unemployed people were paid for doing nothing without being expected to fill the vacancies that arose, there would be higher unemployment. So over the 1990s some countries tightened their benefit regimes and at the same time introduced the principle that after a certain period every unemployed person must be offered activities which he or she must accept as a condition of receiving state support. Three of these countries were Denmark, Britain and the Netherlands, and the results by the year 2000 were remarkable (see Figure 1). In Denmark, unemployment was cut by more than half with no increase in vacancies, and in Britain and the Netherlands unemployment at given vacancies fell by 3 or 4 percentage points.

6By contrast, France and West Germany made little changes in their policies towards the unemployed. As a result, the European boom of 2000-2001 produced only small reductions in unemployment, while vacancies rose to record levels (see Figure 2). It is not often that politicians offer researchers such a neat natural experiment for testing their theories.

7So the enduring high level of unemployment in France and West Germany cannot be blamed on lack of demand, because we can see what happened when demand was high. It is due to supply-side problems in mobilising the unemployed.

Features of a good system

8An effective system for doing this must be helpful but also firm. There is a huge variety of systems within the OECD, as you can find in a wonderful OECD publication called Labour Market Policies and the Public Employment Service, and the lessons there are that an effective system needs four characteristics.

91. Benefits must be paid from the same office that provides job-search assistance and monitors job search.

102. The unemployed person must attend the office in person regularly. (It is every two weeks in Britain, and more often in Switzerland. But I believe that in France and Germany it is nearer to three months.) The unemployed person should have a personal adviser who provides active help – which requires a much better staffing ratio than exists in France and Germany today.

113. After not too long a period of unemployment a person should be willing to accept any job that is available.

124. There should be a maximum period before a person is offered some forms of activity: he must then accept one of these offers. The massive turnover in the labour market should make it possible to secure offers of regular jobs for most workers, if the office tries hard enough. But there must always be some fall-back type of activity within the office’s control.

Guaranteed activity

13So which forms of activity is it most useful to provide? The most useful in descending order are:

  1. – Regular job.

  2. – Subsidised employment with regular employer.

  3. – Subsidised employment on a project.

  4. – Off-the-job training.

14Many US studies done by the Manpower Development Research Corporation and others show that “work first” leads to higher employment rates over the subsequent five years than does “train first” (Gueron, Hamilton, 2002).

15As regards subsidised work we should reject most of the supposed findings about the so-called substitution effects. When employers are asked how many extra jobs they created as a result of the subsidy, this is generally a very small fraction of the jobs subsidised. But this overlooks the fundamental aim of the subsidy which is to mobilise the hard to place, thus releasing for other new vacancies the easy-to-place workers who would otherwise have taken the job given to the subsidised workers.

16In Germany and France there have been important reforms in recent years but further legislation is needed and, above all, speeding up. One problem is always the attitudes of those who work in the benefit offices and the employment offices. Unless those who are administering the reform believe in it, it will not work. It took the British employment service some ten years to believe that what they did would really affect how many people were in work. To change the mindset is essential.

A bad job is better than no job

17There is of course always the controversy over whether a bad job is better than no job. My answer is a firm yes. From the new science of happiness we now know enough of the causes of human happiness to make some quite firm statements. When a person becomes unemployed his welfare falls for two reasons – first the loss of income, and second the loss of self-respect and sense of significance (the psychic loss). The pain caused by the loss of self-respect is (we find) at least as great as the pain which a person would feel if he lost half his income (Layard, 2005). So unemployment hits with a double whammy – the loss of income hurts, but so does the loss of self-respect. That is why unemployment is so devastating, and why we would much prefer it if people were in work.

18But people also have strong feelings about what kind of work they do. Their job satisfaction depends on their income from work but also on the other qualities of the work: the amount of autonomy, job security, human contact, quiet, and stress. Research on job satisfaction tells us how much this matters. But does a bad job bring less happiness than being unemployed? The longitudinal research evidence shows clearly that ex post most people who get jobs feel better than people who remain unemployed (even if ex ante they did not always expect this). Moreover, when a person works there are also gains to the taxpayer (lower benefits and more taxes) and higher profits to employers.

19Thus the key issue for French unemployment is how unemployed people are treated. It is a great pity that the government recently gave more priority to the issue of employment protection, where, as the Jobs Strategy shows, the evidence provides no clear message about its impact on unemployment. It is also a pity that in earlier years there was a misplaced belief that cuts in hours of work or earlier retirement would reduce the equilibrium level of unemployment. There was never any reason to expect this, and all that these measures achieved was a reduction in the amount of work done.

Inactivity: the sick and disabled

20Unemployed people are not of course the only group whose behaviour is affected by how they are treated by the benefit authorities. Another important group is the sick and disabled. As the Employment Outlook shows, the number of working age people on disability benefits in most countries exceeds the number on unemployment benefit. The numbers have been rising in most countries, even though the health of the population has been improving. But their numbers too can be reduced by a more positive approach to their problems.

21A high proportion say they would like to work. So why not help them? In Britain there has been one of the most clear-cut social experiments that I have ever heard of. In seven employment districts there have been so-called Pathways to Work pilots, with other matched districts acting as controls. In the pilot districts every new person coming on to what we call incapacity benefits receives an hour-long interview with an employment advisor each month from months three to eight of his spell on benefit. If he returns to work, he receives an extra £40 a week for the first year of his employment. The effect has been a 45% increase in the numbers who leave benefit within the first nine months (Adamet al., 2006). The lessons are obvious.

Mental illness and employment

22There is one snag: the policy has no effect on people who are mentally ill, as opposed to physically disabled. In Britain we have more mentally ill people on incapacity benefits than the total number of people on unemployment benefits. They mostly have either clinical depression or chronic anxiety disorders like agoraphobia. Unlike the physically disabled, they are not able to work unless they are cured of their illness. Modern evidence-based psychological therapies, especially CBT (cognitive behavioural therapy), can lift at least a half of them out of their illness and a simple cost-benefit analysis shows that if such treatment was offered to mentally ill people the Exchequer would fully cover its cost – through helping some people off incapacity benefits and helping others to avoid coming on to them (Layardet al., 2006) (See Table 1).

23I did not see mental illness figuring large in the OECD’s analysis. This may be because few people understand the size of the problem. Epidemiological surveys of the population show that in advanced countries at least one in six of the population would be diagnosed as mentally ill. It is a major reason for non-participation in the labour force, and it ought to become a standard dimension in all labour market analysis. (See Tables 2 and 3).

Table 1: Cost-ben...Table 1: Cost-benefit analysis of provision of cognitive behavioural therapy (11 sessions) to persons suffering from clinical depression or chronic anxiety disorders (average values per patient treated)

Table 1: Cost-benefit analysis of provision of cognitive behavioural therapy (11 sessions) to persons suffering from clinical depression or chronic anxiety disorders (average values per patient treated)

Table 2: Employme...Table 2: Employment rates and benefit rates for people aged 16-64 (%)

Table 2: Employment rates and benefit rates for people aged 16-64 (%)

Source: Psychiatric Morbidity Survey, 2000. As before, phobia only includes agoraphobia and social phobia.

Table 3: Days per...Table 3: Days per year lost due to sickness absence per person of each type in employment

Table 3: Days per year lost due to sickness absence per person of each type in employment

Source: Psychiatric Morbidity Survey, 2000.

The demand side : wage flexibility

24I have focussed so far on labour supply as a key issue throughout Europe. But there are of course problems on the demand side when there is a mismatch between the pattern of labour demand and labour supply, either by region, skills or age. The most obvious is the problem of regional imbalances, leading to high unemployment in East Germany, Southern Italy and Southern Spain. One element in a solution is migration, but the more one thinks of well-being as the ultimate object of public policy, the more one is aware of the costs of migration – the cost in disrupted families, in broken communities left behind, and in higher turnover and higher crime in the receiving region (see Layardop. cit., p 179-180). So how can demand for labour be increased in high employment regions? Since employer subsidies are very expensive, the key solution has to be relative wage flexibility. The new Jobs Strategy recommends this in Pillar C where there are “uncoordinated sectoral agreements” but it should apply whatever the system of agreements.

Skills and apprenticeship

25Finally, skills (Pillar D) – vital above all for a more equal society. I was delighted to see the emphasis on apprenticeship. As earlier OECD reports have shown, countries which use apprenticeship as the major route to non-university skills have lower youth unemployment and better transitions from school to work (Oecd, 2001). Britain’s Secretary of State for Education is trying to guarantee an apprenticeship for every school-leaver who wants one. I hope he succeeds.


 *    *

26Let me end by congratulating the OECD again on its new strategy and DARES for putting on this conference. I hope it can indeed contribute to better policies for helping the French unemployed. As I read the evidence, France needs a radical re-think about how unemployed people are treated. The central idea must be of rights and responsibilities. The right to unemployment benefits and job-search assistance must be complemented by the responsibility to look hard and to accept work when it is available. A system which effectively applies this idea must have the following elements:

  1. – one office only, providing both benefits and job-search assistance;

  2. – frequent attendance at this office;

  3. – a requirement to accept work that is available;

  4. – after one year a guarantee of activity, which must be accepted.

Haut de page


Adam S. et al. (2006), “Early quantitative evidence on the impact of the Pathways to Work pilots”, Department of Work and Pensions Research Report No 354.

Gueron J.M., Hamilton G. (2002), “The Role of Education and Training in Welfare Reform”, Policy Brief No 20, The Brookings Institution, Washington DC.

Layard R. (2005), Happiness-Lessons from a New Science, Penguin, p. 67.

Layard R., Nickell S., Jackman R. (1991), Unemployment-Macroeconomic Performance and the Labour Market.

Layard R., Clark D., Knapp M. Mayraz G. (2007), “Cost-Benefit Analysis of Psychological Theory”, in National Institute Economic Review, No 202, October.

Mitchell B. (1992), International Historical Statistics: Europe 1750-1988, Basingstoke: Macmillan.

Officeof National Statistics (2005), Labour Market Trends, Table A. ll, March.

Oecd (1990), Labour Market Policies for the 1990s.

Oecd (2001), Knowledge and Skills for Life: First Results from PISA 2000, Paris, OECD.

Ocde (2006), Perspectives de l’emploi de l’OCDE. Stimuler l’emploi et les revenus, disponible sur le site :

Haut de page

Table des illustrations

Fichier image/, 156k
Fichier image/, 174k
Titre Table 1: Cost-benefit analysis of provision of cognitive behavioural therapy (11 sessions) to persons suffering from clinical depression or chronic anxiety disorders (average values per patient treated)
Fichier image/, 4,4k
Titre Table 2: Employment rates and benefit rates for people aged 16-64 (%)
Légende Source: Psychiatric Morbidity Survey, 2000. As before, phobia only includes agoraphobia and social phobia.
Fichier image/, 6,2k
Titre Table 3: Days per year lost due to sickness absence per person of each type in employment
Légende Source: Psychiatric Morbidity Survey, 2000.
Fichier image/, 2,7k
Haut de page

Pour citer cet article

Référence papier

Richard Layard, « L’expérience du chômage : quelles leçons pour la France ? », Travail et Emploi, 118 | 2009, 23-28.

Référence électronique

Richard Layard, « L’expérience du chômage : quelles leçons pour la France ? », Travail et Emploi [En ligne], 118 | avril-juin 2009, mis en ligne le 07 mars 2011, consulté le 15 octobre 2019. URL : ; DOI : 10.4000/travailemploi.4174

Haut de page

Droits d’auteur

© La documentation française

Haut de page
  • OpenEdition Journals