Illustration by Lluis Pino
“The choice we made (…) was the key to Germany’s success in its fight against unemployment. This new flexibility will be beneficial to the French economy and to workers who will profit from this increased competitiveness”
President Nicolas Sarkozy, 2012, Le Figaro.
“The first thing to do is to reform labour law – not to abandon our principles: regulation of working times, employment contracts and the minimum wage are fundamental – but it is true that it is too heavy, too complicated. I have two principles, flexibility and security. Flexibility for firms, to allow them to hire, to take into account variations in economic activity. Security for workers who face these transformations.”
President François Hollande, 2016, France 2.
“I am not afraid of the word: we have to endorse more flexibility, to adapt our labour laws to ongoing transformations.”
Presidential candidate Emmanuel Macron, 2017, Les Echos.
If you have steered clear of the daily chanting of the flexibility mantra, well done you! The idea that labour market regulations are causing unemployment (and, conversely, that trimming employment protection legislation will boost employment rates) is pervasive. From pundit discourse to bar talk, the idea is often asserted with the strength of self-evidence: of course labour market ‘rigidities’ prevent employers from hiring. Are you expressing doubts about the whole “flexibility rhetoric”? Get ready to attract suspicious looks: it probably means that you’re economically illiterate, but also backward-looking, resistant to change, in short: conservative.
I find this last idea particularly ironic, since policy proposals for more “flexibility” of the labour market are far from new. A quick check on Google Ngram viewer reveals that calls for more labour market flexibility have been going around for a while. Even the incredibly trendy concept of “flexicurity” – the idea that increased flexibility must be compensated by increased support for laid-off workers – and its associated ‘nice Scandi’ marketing blurb (“Look, the Danes are doing it, and they’re happy!”) aren’t all that new, by that metric.
% results in Google Books English corpus, 1980-2008 (Source: Google Books Ngram Viewer)
% results in Google Books French corpus, 1980-2008 (Source: Google Books Ngram Viewer)
In 1998, Economics Nobel Prize winner Robert Solow pointed to the “knee-jerk” character of repeated calls for labour market flexibility. Blaming labour market “rigidities”, he contended, had achieved the status of “a reflex exhibited by any central banker you might care to ask”. He identified the publication of the 1994 OECD Jobs Study—whose main assumption was that greater flexibility account for higher job creation in the US compared with Europe—as a turning point in the dissemination of the flexibility paradigm. The IMF and the World Bank went on to endorse the principle. All three institutions have repeatedly taken the position that increasing labour market flexibility would boost employment creation.
But what’s behind the recurring buzzword? What does flexibility entail? And why would it fuel employment? The recent labour market reform in France provides a compelling case study to explore the matter. It is a state-of-the art example of the craze: French television, newspapers, and radios were echoing the official rhetoric, calling for the labour market to be “liberated”, and for creative forces to be “set free”. The metaphor is strong: entrepreneurial energies in France were allegedly chained, nipped in the bud by a rigid set of labour market laws. Flexibility must come to the rescue. But how exactly would that boost employment?
Unpacking the flexibility logic
At the heart of the flexibility logic lies the idea that labour market regulations, through their influence on the incentives facing employers and workers, affect labour market performance (i.e., employment and unemployment rates). Different specifications of the particular mechanisms at play have been formulated; the figure below provides a comprehensive overview of the various rationales discussed in the literature. Regulations restricting the possibility for firms to adjust wages, working time, and more generally the allocation of workers to jobs (e.g. “hiring and firing” procedures) are thought to negatively affect labour market performance. Flexibilization reforms aim to remove restrictions in these three areas.
Reproduced from HM Treasury, 2003:11
The scholarship on the issue has evolved since the original formulation of the logic. A 2013 paper by IMF economists aims to “go beyond ritual invocation of the mantra”: it develops a nuanced take on the flexibility logic in a post financial crisis context. The authors distinguish between flexibilization reforms designed to enhance micro- and macro-flexibility. Micro-flexibility is “the economy’s ability to carry out the reallocation of workers to jobs needed for productivity growth”. In plain terms, this means the ability for firms to fire and hire workers according to their needs and the worker’s productivity. “Excessive protection or complex restriction on the separation process (…) leads to lower hiring and thus higher unemployment duration”, the authors assert. In other words, employers are put off by regulations limiting their future ability to fire workers and they hire less than they would in the absence of these rules. Loosening or suppressing laws regulating dismissals would ensure a more fluid reallocation process; risk-averse employers would be more inclined to hire, knowing that they can easily let go of workers in the future. Transformation of firing regulations would, therefore, be the first policy tool to rein in flexibilization reforms.
Macro-flexibility is the economy’s capacity maintain a low unemployment rate in the face of macroeconomic shocks”. It gives firms the ability to adjust their costs (through e.g. paying workers less, having them work more at the same pay rate, or reducing their shift) during economic downturns. Here, the authors argue, “the critical institutions is the structure of collective bargaining, and in particular the degree of coordination of social partners”. At this point the paper adopts a rather confused tone: the authors hesitate as to which level of negotiation (from national to firm-level) will make it easier to get employees to accept lower wages, longer (unpaid) hours or partial unemployment, should the economic situation require it. Nonetheless, the policy recommendation remains: finding the right level at which to negotiate specific agreements on working conditions and allowing for the efficient coordination of actors is key to ensuring that firms can weather adverse economic situations. With such a system in place, employers should not restrain from hiring. The second lever to pull in flexibilization reforms concerns bargaining level. “A combination of national and firm-level bargaining seems attractive”, the authors conclude—somewhat inconclusively.
The case of the recent French labour market reform
The economic theory linking increased flexibility to employment creation provides policymakers with two clear avenues for reform. Both of these have been diligently implemented in the latest episode of the flexibilization saga in France. This was acclaimed by many international observers: after all, France had the quintessential rigid labour market where it was “very difficult for companies to turn over underperforming employees”, so it was about time it was reformed—right?
Well, an empirically grounded answer to that question would have to be a resounding “no”. Before the new law was passed, French labour law already granted many possibilities for employers to dismiss employees. True, completely random or arbitrary firing is prohibited. But an employee can be laid off under many different circumstances. “Economic” dismissals include cases in which the company needs to cut costs to maintain competitiveness, to adapt to technological change, or to change the content of an employee’s job. Contrary to a well-entrenched belief, this type of firing procedure is widely practised and rarely leads to legal challenges. Dismissals for “personal reasons” include cases of professional negligence, but also of inaptitude, and incompetence. Commentators pointing to the “rigidity” of firing procedures in France often put forward cases of employees suing their employers after a dismissal; yet they often omit to explain that this is only possible in cases of unfair dismissals, i.e. if none of the justifications mentioned above applied. In addition, the recourse to this type of legal challenge is not significantly higher in France than in other European countries. These are the rules of the game, and paying attention to them gives a more balanced view of the situation.
Nonetheless, the new labour law contains a whole swath of measures designed to make firing easier. First, it creates a new type of contract, interestingly called the “Project-length permanent contract” (Contrat à Durée Indéterminée de projet). It is only “permanent” in the sense that it does not entail paying employees the bonus they normally get at the end of a fixed-term contract. Other than that, it’s pretty much fixed term – it ends when the “project” is completed. The creation of this new type of fixed term contracts was a long-time demand from the main employers’ association, representing large firms. These contracts provide employers with more regular occasions to part ways with a worker without compensating her either for terminating her permanent contract, or for the precariousness of a fixed-term contract.
In the same logic, the law introduces the possibility for collective contractual terminations, allowing employers to separate from several employees at the same time without having to fire them—and hence without having to justify firing them on any of the grounds outlined above. Former workers will be able to claim unemployment benefits. But collective lay-offs previously entailed more protection for workers (including the obligation for their former employers to provide support for job search or re-training).
Finally, legal compensations for unfairly fired employees have been capped. This amounts to putting a price tag on unfair dismissals. And the price is rather low: it ranges from the equivalent of a month salary for employees that have been employed for less than 2 years, to 20 months for someone that has been employed for 30 years.
The new law is also a textbook example of macro-flexibilization, with several transformations of the bargaining system. In the previous flexibilization episode carried by the Hollande government in 2016, working time arrangements had become a matter for firm-level agreements. The new law introduces the primacy of firm-level bargaining for virtually all subjects. Eleven issues are exempted, which will now be negotiated at the sector-level; this includes minima wages, fixed-time contracts, part-time work, the length of trial periods, professional training and gender equality. Many of these used to be a matter for centralized bargaining.
These reforms are very contentious and have stoked fierce opposition and social unrest (the fact that it took the form of an executive order did not help). Surely if the government is willing to weather a social tempest so early in its term, and to use such a controversial legislative vehicle for it, it must be pretty confident in the capacity of this strategy to effectively boost employment rates. Surely this is an approach that is backed by a lot of empirical evidence—right?
The evidence behind the mantra: do flexibilization reforms create employment?
In fact, considering the scarcity of evidence backing the theory that flexibilization reforms favour employment creation, the prevalence of this idea is astonishing.
Back in 1998, Robert Solow tested whether rigidities of the labour market actually led to higher unemployment. His conclusions went “squarely against the cliché that high and persistent European unemployment is entirely or mainly a matter of ‘labour-market rigidities’.” While it was “certainly possible that job creation could be inhibited by apprehensiveness about the working of the labour market”, this mechanism remained empirically unsubstantiated, he cautioned.
Of course this dates back to 1998. And since then has proof been gathered? Fast-forward a decade. In 2007, the International Labour Organization reviewed the evidence on flexibilization and employment creation, and concluded that it “did not provide much support for the view that greater flexibility results in higher employment.” Studies looking at the relationship between employment protection regulation and aggregate employment or unemployment find “surprisingly muddy” results, the authors argued (see also this 2006 review of the evidence by Olivier Blanchard). This is made clear in the following tables, combining these findings with those from other studies I came across in researching for this article.
Results confirming the expectations derived from the flexibilization logic are marked in green; results contradicting these expectations are marked in red. As explained above, expectations are not clear-cut when it comes to bargaining levels. The French reform clearly went in the direction of a decentralization of bargaining, and contains no measures to increase coordination between individual firms. Therefore we use as a theoretical benchmark the idea that decentralized (and not necessarily coordinated) bargaining favours labour market performance (e.g. results contradicting the logic applied in the French case are marked in red).
These are some of the most widely quoted pieces of research on the subject. And the picture that comes out of this table is not that of an overwhelming scientific consensus backing the flexibility theory. In fact, only a minority of studies support the idea that easy firing regulations are associated with lower unemployment, while none of them find bargaining decentralization helpful to reduce unemployment. Crucially, although some studies find a statistically significant correlation between lower labour market flexibility and higher unemployment, only one of the studies looking at consequences in terms of job creation (highlighted in yellow in the tables below) actually find a positive effect of increased flexibility, and its methodology is contested.
Studies looking at the effect of hiring and firing regulations on various measures of labour market performance
|Study||Effect of||On||Countries and period||Results|
|Scarpetta (1996)||Employment protection legislation||Unemployment||OECD countries, 1983-1993||Positive|
|Nickel (1997)||Employment protection legislation||Unemployment||OECD countries, 1983-1988 & 1989-1994||Non significant|
|Blanchard and Wolfers (2000)||Employment protection legislation||Unemployment after a shock (interaction of institutions and economic situation)||E15 countries, the US, Canada, New Zealand and Japan, 1960 to 1995||Positive (but ambiguous)|
|Dolado et al (2002)||Reduced employment protection for temporary workers||Employment||Spain (sectoral and regional data), 1984-2000||Ambiguous|
|OECD Employment Outlook (2004)||Employment protection legislation||Unemployment||OECD countries||Non significant|
|OECD Employment Outlook (2004)||Employment protection legislation||Employment||OECD countries||Negative (correlational)|
|Baker et al. (2005)||Employment protection legislation||Unemployment||20 OECD countries, 1960-1999||Non significant|
|Baccaro and Rei (2005)||Employment protection legislation||Unemployment||OECD countries, 1960-1998||Non significant|
|Autor and Houseman (2005)||Reduced employment protection for temporary workers||Long term employment||US (individual level, policy quasi-experiment, 1996)||No effect|
|Zijl (2005)||Increase in availability of temporary jobs||Duration of unemployment spells||The Netherlands||Negative|
|Monastiriotis (2006)||Flexible working arrangements||Unemployment||UK (regional level)||Positive|
|Bassanini and Duval (2006)||Employment protection legislation||Employment||OECD countries||Non significant|
|Amable et al. (2007)||Employment protection legislation||Joblessness||Panel of 18 countries, 1980 to 2004||Negative|
|Guell and Pertrongolo (2007)||Development of temporary contracts||Permanent employment||Spain||Very weak|
|Kugler and Pica (2008)||Increase in firing cost||Employment||Italy (firm level)||Negligible effect|
|Feldmann (2009)||Employment protection legislation||Unemployment||Panel of 73 countries, 2000 to 2003||Positive|
|Bernal-Verdugo et al. (2012)||Flexible hiring and firing regulations, lower firing costs||Unemployment level; change in unemployment||Panel of 97 countries, 1985-2008||Negative|
|Signoretto (2013)||Contractual termination||Employment creation||France (firm level)||Negative|
Studies looking at the effect of collective bargaining institutions on various measures of labour market performance
|Study||Effect of||On||Countries and period||Results|
|Calmfors and Driffill (1988)||Bargaining centralization||Employment||17 OECD countries; 1963-1985||Full centralization and full decentralization lead to reduced unemployment|
|Scarpetta (1996)||Bargaining centralisation and coordination||Unemployment||Full centralization and full decentralization lead to reduced unemployment (larger effect of centralization)|
|Nickell (1997)||Bargaining coordination||Unemployment||OECD countries, 1983-1988 & 1989-1994||Negative|
|Traxler (2003)||Bargaining centralization||Change in unemployment||20 OECD countries, 1970-1996||Ambiguous|
|Blanchard and Wolfers (2000)||Bargaining coordination||Unemployment||E15 countries, the US, Canada, New Zealand and Japan, 1960 to 1995||Negative|
|Baker et al. (2005)||Bargaining coordination||Unemployment||20 OECD countries, 1960-1999||Negative|
|Basselini and Duval (2006)||Bargaining centralization||Unemployment||OECD countries||Negative|
|Amable et al. (2007)||Bargaining coordination||Joblessness||Panel of 18 countries, 1980- 2004||Negative|
|Traxler and Brandl (2011)||Bargaining centralisation||Unemployment||18 OECD countries, 1980-2000||Insignificant|
|Cazes et al. (2012)||Bargaining centralisation and coordination||Unemployment||Meta-analysis||Negative|
|Bernal-Verdugo et al. (2012)||Bargaining centralization||Youth and long-term unemployment||Panel of 97 countries, 1985-2008||Insignificant|
Such a patchy mix of empirical evidence is in stark contrast with the wealth of theoretical arguments detailing the workings of the flexibility logic. The flexibility paradigm is largely based on an understanding of how humans should work, according to economic theory. Rational employers should be put off by restrictions on labour adjustments; conversely they should react to the incentive of easy-firing regulations by employing more people.
The problem is that there is nothing self-evident about the ways in which employers weigh various incentives and other considerations when making hiring decisions. In fact, there is actually some empirical evidence that employers are not primarily concerned with the stringency of firing processes when making employment decisions. 47% of the employers surveyed in this 2017 study by the French statistics authority declared they were facing “barriers to hiring”—and the issue of labour market regulations was not the most important concern of those who did. 28% mentioned ‘economic uncertainty’, and 27% listed the difficulty to find qualified employees as the primary obstacle to hiring. 18% mentioned labour costs (with an emphasis on high social security contributions); concerns with labour market regulations came last: 14% mentioned the legal risks associated with the firing process, and 10% mentioned the cost of dismissing an employee.
Looking at this data in conjunction with the results presented in the first table above, it seems that the continued obsession with facilitating firing might be misplaced. Similarly, the empirical evidence showing a positive effect of centralized and coordinated bargaining on employment is at odds with the current obsession with firm-level bargaining. The idea that flexibilization reforms facilitating firing and decentralized bargaining will boost employment is at best very poorly substantiated, at worst blatantly contradicted by existing evidence.
Still, flexibility might be worth trying. There are little guarantees that these reforms will work—but why not giving it a shot? To start with, a 30-years trial period seem like a fair amount of time after which to expect some results or reject a policy as inefficient. Beyond, there are good reasons to call for the end of what has actually been a rather harmful experimentation.
The dangers of the flexibility strategy
The bleak reality of a “flexibilized” life
Flexibilization of the labour market comes at a cost for the “flexibilized” workers. It facilitates the development of jobs that are temporary (hence its unclear effect on employment in the long run), low-paid, that offer little career advancement prospects, and no possibilities for upskilling. It is very unlikely to be conducive to rising living standards in the medium to long-term. Flexible patterns of employment are associated with adverse health effects. Individuals moving from unemployment into poor-quality jobs are more likely to suffer from chronic stress, and a deteriorated mental health. Even in contexts where increased flexibility has supposedly been matched with increased security for workers, as in the Danish “flexicurity” model, there are perverse effects at play: individuals with ill-health are more likely to be laid-off in Denmark, suggesting that health selection is at play.
These adverse consequences of flexibilization have to do with its very objective—a more fluid circulation of workers. In non-abstract terms, this means that they often have to go back to square one: job hunting. The necessity to very regularly rethink your whole life, readjust your projects, re-evaluate your hopes and ambitions is built into the flexibility logic. In a flexible labour market, workers are expected to frequently go through the whole process of finding a new job: routinely scrolling down pages of job ads; re-assessing their skills; rebranding and re-selling themselves. No matter what degree of financial security is provided: looking for a job is an anxiety-provoking experience. Increasing the frequency of job-searching episodes is bound to take a toll on your mental health. And nowhere in the world – not even in Denmark – are unemployment benefits unlimited; this means that the risk of financial destitution is always looming in the background. Coercive measures—such as those limiting the number of job offers an unemployed person can turn down—are very much part of the flexicurity package, too, adding threat and social stigma to the equation.
Flexibilized workers are also expected to bear the brunt of adjustment (mostly through a reduction in their income) in cases of economic downturn – i.e., at times when it is likely to be even more difficult and stressful. Remember the concept of “macro-flexibility”: in a flexible labour market, employers can easily adapt their labour cost in line with the evolution of economic demand. In a flexible world, individual workers, rather than employers, bear the risks associated with a slowing down of the economy. This is a fundamental transformation induced by flexibilization reforms, albeit one that is very rarely discussed: labour market flexibility implies a risk shifting from employers to individual workers.
This is at odds with the general discourse accompanying flexibilization reforms, which often lauds the courage of risk-taking entrepreneurs, celebrating them as the driving forces of the economy. You might expect such high praises to entail a sense of responsibility towards those without whom firms could not function, i.e. employees. On the contrary underlying the flexibility logic is a vision of employers as relieved from any form of responsibility, legal or moral, towards workers. The logic dictates that employers are able to get rid of their employees as soon as doing so makes rational sense. In the UK, such logic has brought back employment by the hour from the great progressive era of the 19th century. When employee’s living standards employees become merely an adjustment mechanism, rather than an objective to strive for alongside profitability, the notion of “enterprise” is arguably degraded. An entrepreneur’s success is rather less impressive if it depends on exploiting an army on precarious workers on zero-hours contracts. What’s striking is how little ambition, in fact, is ascribed to employers in the flexibilization paradigm.
A short-sighted economic strategy
Flexibilization reforms tend to result in a dualization of the labour market, between high-skill, highly paid, protected jobs, and low-skill temporary unprotected jobs. This phenomenon is largely documented (including for instance in the IMF study quoted above). This has led some to call for more flexibilization: suppressing employment protection further and for everyone would remove the divide between labour market insiders (those occupying the higher skilled protected jobs) and outsiders (the army of unprotected temps). This reasoning led the French government to introduce the “project-length contract”: thanks to that new contract, employers would reduce their recourse to temporary contracts. In other words, to prevent the recourse to temporary contracts, the government introduced… a new form of temporary contract.
Beyond its logical absurdity, this strategy is bound to fail. It does not make sense for employers to recruit a worker with highly specific skills, in whom they might have to invest firm-specific training, on a temporary contract. On the contrary, they will want to retain that employee, by offering her a competitive salary as well as a secure contract. From an employer’s perspective, using temporary contracts only makes sense in the case of relatively low-skilled, easily replaceable workers. Hence the skill divide observed in highly flexible labour markets.
For that reason, reforms favouring the recourse to flexible short-term contracts might be ill-advised in terms of overall economic strategy. Not only do these contracts hold limited prospect for individual skill-building, but relatedly, their large-scale use is bound to favour the development of low productivity industries. Consistently, three researchers from the International Labour Organization showed that employment tenure has a positive effect on productivity at the firm level. Servaas Storm and C.W.M. Naastepad also found a strong positive relationship between employment protection and productivity growth. Opting for flexibility is a bad strategic choice in the long run.
Why is the flexibility logic so resilient?
Flexibilization reforms have been around for a long time. Available empirical evidence is often disconfirming. Adverse consequences of flexibility are well-known. This begs the question: why are we still talking about labour market flexibility?
The answer to this question is likely to be multifaceted and complex – for one, the continuous lobbying of large employers’ association, such as the Medef in France, or CBI in the UK should be thoroughly explored. But the persistence of the flexibility logic could also partly be understood as the headlong rush of governments with their back against the wall. Government resort to the religious-like preaching of a policy dogma, in spite of all logic, when constraints on alternative policy options are thought to be unsurmountable.
Other options to deal with the issue of persistent unemployment do exist: governments could invest in the development of leading economic sectors; they could reorient public spending towards a social investment strategy. For instance, the billions of euros spent yearly in France on subsidizing low-quality, low paid jobs in the private care sectors could instead be used to create more decently paid jobs of better quality in childcare and elderly care. Governments could explore the job-creation prospect of a reduction of working time. In the short term, they could seriously consider whether forcing a low-skilled worker into a low-quality temporary job, with the adverse consequences on her health highlighted above, really makes more sense than providing her with a universal income, and allowing her to contribute through volunteering, or to invest in upskilling.
Reforming employment protection legislation is free; the common denominator between these alternative options is that they cost money. Implementing these solutions implies public spending, a notion that has become anathema in policymaking. Some of these solutions would entail the development of public sector jobs, a concept which has been banned from the realm of “realistic policy ideas”. Yet, there is nothing unrealistic about these solutions: a government could decide to use its budget to implement them. If necessary, it could decide to expand its budget through taxation to fund the jobs that are needed in childcare and elderly care. The problem is that uttering the words “taxation”, “spending”, “public sector” might relegate them to the fourth league of “unrealistic”, and “economically illiterate” politicians. These policies are are therefore crossed off the menu of “reasonable policy options”. To maintain their economic credibility, politicians are thus incentivized to keep adhering to the economic policy status quo, even when it is blatantly failing.
Flexibility is the labour market equivalent of sticking to a drug that has no proven positive effects and some known adverse side-effects, instead of funding research for a new drug, all because the first one is widely believed to be efficient, and it is cheap. This is an absurd way to go about solving any problem, medical or societal. In this situation, there is an urgent need to call out unsubstantiated mantras, and reintegrate in the list of realistic policies some options that have been arbitrarily kicked out of it. This, rather than chanting with the crowd, is what progressive policymaking is about.