Working time reduction and employment

Abstract

In recent years, in some European countries where unemployment has been stubbornly high, interest has re-emerged in the policy intended to increase employment by reducing standard hours worked per time period of full-time workers, that is, the number of hours that are determined exogenously- at government level or by broad national-level collective bargaining agreements- beyond which an overtime premium must be paid. The basic intuition underlying this policy is that reducing normal working time per worker would encourage firms to hire additional employeesas they seek to replace the labour services currently provided by those workers putting in longer standard hours. This idea is usually referred to as “work-sharing”. However, from a theoretical point of view, this strategy has been severely criticized. There are five main sorts of objections to the idea that reducing standard hours of work can create new jobs. Perhaps the most commonly heard of these is that work-sharing arguments are invalid because they commit the so-called lump of labour fallacy. This critique refutes the idea that there is a fixed total amount of work thatcan be redistributed at will, so that reductions in normal working time automatically lead to the lost time being worked by new employees. On the contrary, employment is determined by demand and supply factors, or in a more modern context by price and wage setting behaviour. Reductions in standard working time may affect both price setting and wage setting, and the final impact on employment is uncertain. The second argument against standard working time reduction as anemployment strategy is that employers may choose not to create new jobs in response. Instead they may simply cut production along with working hours, or maintain production without new recruitment by expanding the use of overtime, bringing under-utilised workers into full operation, or improving employee productivity. A third argument is that shorter scheduled working time would attract more people into the workforce, namely those who in the past have been unable orunwilling to work longer hours, such as women caring for children and students. This implies that even if the number of employed rose, the number of unemployed might not fall, as the concomitant increase in participation rates limits the impact of a reduction in normal hours of work on unemployment. A fourth argument is whether the labour services of existing workers could bereplaced easily from the ranks of the unemployed. It is well known that the pool of non-employed people tend to possess notably different skills than the pool of employed people. This mismatch in skills may make it difficult for firms to find new workers for some jobs and thereby limit the effectiveness of a work-sharing policy through shorter normal hours of work for increasing employment while accentuating any negative impact on productivity and output. A fifth and more serious argument against work-sharing is that higher labour costs caused by reducing standard working time may lead to lower employer demand for labour, impaired competitiveness and/or lower investment, and thus to higher, not lower, unemployment. Indeed, reductions in normal hours of work can increase labour costs in a number of ways. If standard working time is reduced with full or partial wage compensation – that is, without making proportionate cuts in wages – hourly wage costs for employers rise. Also, to the extent that overtime attracts wage premia,a cut in scheduled working hours can increase wage costs insofar as employers replace lost working time by increasing the use of overtime. Further, given constant costs per employee for recruitment, training, and other fixed costs, an increased number of employees doing the same amount of work means increased administrative costs for employers both relative to the total cost of each employee and in aggregate. And unless taxes and social security contributions per employee are reduced along with normal working time, these too can raise non-wage labour costs. Finally, successin reducing unemployment may itself lead to higher labour costs because having fewer unemployed people competing for jobs may lead to higher wages for those in employment. The main objective of this doctoral dissertation is to shed light on the issue of shorter standard working hours as a policy instrument for reducing unemployment, or equivalently for increasing the number of people in paid employment. Thus, in this study, I will provide a survey of the most pertinent theoretical and empirical contributions to the relevant literature on the employment effects of normal workingtime reductions. In addition, I shall seek to offer some new theoretical and empirical evidence on the efficacy of this policy in reducing unemployment. Specifically, the structure of this dissertation is as follows. Chapter 1 provides the background to the growing importance of standard working time reductions in Europe over the past 20 years. Also, it offers a number of policy views of tradeunions, employers and governments on the wisdom and effectiveness of such policy as a means of stimulating new jobs. It then presents an outline of some recent experiences with statutory or collective agreed reductions of working hours in Western Europe. Chapter 2 presents the existing theoretical evidence of the employment effects of cuts in normal working hours by considering four main types of models delineated in the literature. These are: labour demand models with predetermined wages, collective bargaining models, where the employment level and/or the hourly wage are determined through the bargaining process, efficiency wage models, which are basedon the assumption that there exists a direct and increasing relationship between the wage paid by firms and the level of effort provided by workers, and search-matching models, where the complex process of search and matching up of workers and jobs is taken into account. Empirical evidence relating to the theory of chapter 2 is reviewed in chapter 3. A large part of that chapter deals with research into the estimated effects on employment, total (or overtime) hours and wages of a change in standard hours of work. Two main aspects of this work are considered. The first derives essentially from regression analyses, although the work here is further subdivided into analyses that have incorporated aggregate data (national- and industry-level) and those that are tested using more disaggregated data (firm/plant-, and individual-level). The second type of modelling approach involves simulation exercises carried out on large-scale macroeconometric models and a review is undertaken of the findings from a wide variety of European studies. The other topics featured in the chapter concern the labour productivity effects and the labour supply responses of working hours changes, and work is examined, from a variety of countries, that has looked into these issues. The subject matter of chapter 4 is to examine the relationship between hoursworked and unemployment in the context of a shirking type efficiency wage model, where involuntary unemployment occurs owing to imperfect monitoring of workers’ effort. In particular, I have used Shapiro and Stiglitz’s model (1984) extended to allow for endogenous effort and a distinction between employees and hours. It is shown that, when the number of hours worked is determined exogenously, work-sharing through shorter hours of work could indeed lead to a reduction of unemployment, although such a reduction comes at the expense of lowering the wage earnings of those already employed, the effort per worker, the output level and firms’ profits. Inthe presence of such unpleasant implications, it is expected that either employers or employees will be resistant to any reduction in the length of working time, despite any sympathy they may feel with the growing numbers of unemployed. Moreover, when hours are determined endogenously, it is found that a work-sharing policy has no impact on unemployment. In chapter 5, I investigate the married women’s labour supply due to an exogenous restriction on the standard hours worked by husbands. It is well known that economic analysis suggests that if the husband experiences unemployment, the wife responds to this transitory event by entering the labour force, and eventually employment. This has been referred to as the “added worker effect” and may beexplained by both an income and a cross-substitution effect. However, the incidence of the added worker effect may result not only from unemployment spells faced by the husband, but also from compulsory constraints on his working hours, especially if these are not accompanied by full wage compensation. I confront this hypothesis by empirically studying the consequences on married women’s labour supply of the exogenous husbands’ working week reduction in France from 39 to 35 hours, which was first applied to large firms in 2000, by a natural experiment. Using data forFrance from the European Community Household Panel (E.C.H.P.), I show that an exogenous reduction in husbands’ standard working hours increases the wives’ probability to join the labour force, when they were outside of the labour market in previous years. Moreover, this effect is found to be more conspicuous in low income families and for wives with relatively low human capital. Conversely, I find no significant effect on the number of hours worked by wives who are already in the labour market and are either affected or not by this policy. Finally, chapter 6 discusses the scope for achieving increases in employment through decreases in standard working hours and summarizes the circumstances that need to be satisfied in order to have a positive and sustainable effect on employment.