In the first chapter,a harmonised linked employer-employee dataset is used to study the impact of firm-level agreements on the wage structure in the manufacturing sector in Belgium, Denmark and Spain. To our knowledge, this is one of the first cross-country studies that examines the impact of firm-level bargaining on the wage structure in European countries. We find that firm-level agreements have a positive effect both on wage levels and on wage dispersion in Belgium and Denmark. In Spain, firm also increase wage levels but reduce wage dispersion. Our interpretation is that in Belgium and Denmark, where firm-level bargaining greatly expanded since the 1980s on the initiative of the employers and the governments, firm-level bargaining is mainly used to adapt pay to the specific needs of the firm. In Spain, the structure of collective bargaining has not changed very much since the Franco period where firm agreements were used as a tool for worker mobilisation and for political struggle. Therefore, firm-level bargaining in Spain is still mainly used by trade unions in order to reduce the wage dispersion. In the second chapter, we analyse the impact of the bargaining level and of the degree of centralisation of wage bargaining on rent-sharing in Belgium. To the best of our knowledge, this is the first study that considers simultaneously both dimensions of collective bargaining. This is also one of the first papers that looks at the impact of wage bargaining institutions on rent-sharing in European countries. This question is important because if wage bargaining decentralisation increases the link between wages and firm specific profits, it may prevent an efficient allocation of labour across firms, increase wage inequality, lead to smaller employment adjustments, and affect the division of surplus between capital and labour (Bryson et al. 2006). Controlling for the endogeneity of profits, for heterogeneity among workers and firms and for differences in characteristics between bargaining regimes, we find that wages depend substantially more on firm specific profits in decentralised than in centralised industries , irrespective of the presence of a formal firm collective agreement. In addition, the impact of the presence of a formal firm collective agreement on the wage-profit elasticity depends on the degree of centralisation of the industry. In centralised industries, profits influence wages only when a firm collective agreement is present. This result is not surprising since industry agreements do not take into account firm-specific characteristics. Within decentralised industries, firms share their profits with their workers even if they are not covered by a formal firm collective agreement. This is probably because, in those industries, workers only covered by an industry agreement (i.e. not covered by a formal firm agreement) receive wage supplements that are paid unilaterally by their employer. The fact that those workers also benefit from rent-sharing implies that pay-setting does not need to be collective to generate rent-sharing, which is in line with the Anglo-American literature that shows that rent-sharing is not a particularity of the unionised sector. In the first two chapters, we have shown that, in Belgium, firm-level bargaining is used by firms to adapt pay to the specific characteristics of the firm, including firm’s profits. In the third and final chapter, it is shown that firm-level bargaining also allows wages to adapt to the local environment that the company may face. This aspect is of particular importance in the debate about a potential regionalisation of wage bargaining in Belgium. This debate is, however, not specific to Belgium. Indeed, the potential failure of national industry agreements to take into account the productivity levels of the least productive regions has been considered as one of the causes of regional unemployment in European countries (Davies and Hallet, 2001; OECD, 2006). Two kinds of solutions are generally proposed to solve this problem. The first, encouraged by the European Commission and the OECD, consists in decentralising wage bargaining toward the firm level (Davies and Hallet, 2001; OECD, 2006). The second solution, the regionalisation of wage bargaining, is frequently mentioned in Belgium or in Italy where regional unemployment differentials are high. In this chapter we show that, in Belgium, regional wage differentials and regional productivity differentials within joint committees are positively correlated. Moreover, this relation is stronger (i) for joint committees where firm-level bargaining is relatively frequent and (ii) for joint committees already sub-divided along a local line. We conclude that the present Belgian wage bargaining system which combines interprofessional, industry and firm bargaining, already includes the mechanisms that allow regional productivity to be taken into account in wage formation. It is therefore not necessary to further regionalise wage bargaining in Belgium.

Thèse numérisée par la Direction des bibliothèques de l’Université de Montréal / Le chapitre 1 passe en revue les modèles et les résultats pertinents à la formation formelle en milieu de travail, tandis qu'un bref survol des modèles qui prévoient une cor-rélation entre l'utilité alternative et le salaire des travailleurs est incorporé au chapitre 2. Au chapitre 2, nous analysons la corrélation entre les revenus de travail contempo-rains et les conditions sur le marché du travail, à savoir le taux de chômage, mesuré à différents moments pendant la durée du contrat de travail. Les données que nous utilisons proviennent du Panel Socio-économique Allemand, et comprennent la période 1984-1994. Contrairement aux résultats pour le marché américain, nous trouvons que l'état actuel du marché du travail est un important facteur même en contrôlant pour le taux de chômage en vigueur au début de la relation d'emploi. Les élasticités estimées varient entre 9 et 15 pourcent pour l'élasticité des revenus de travail par rapport au taux de chômage contemporain, et entre 6 et 10 pourcent par rapport au taux de chô-mage en début de contrat. Notamment, tandis que le taux de chômage régional affecte le niveau des revenus d'emploi, le taux national influence la variation dans les revenus. Ces résultats sont cohérents avec la présence simultanée de conventions collectives et de contrats individuels, tels les modèles de contrats implicites, qui expliquent une partie de la variance de revenus de travail et des mouvements de revenu à long terme. En plus de la variation régionale, nous étudions l'hétérogenéité des contrats selon certains caracté-ristiques des travailleurs et des emplois. En particulier, nous constatons que les contrats de travail diffèrent selon la taille de l'entreprise et le type de travailleur. Un travailleur dans une grande entreprise est remarquablement plus isolé de fluctuation du marché de travail qu'un travailleur dans toute autre taille d'entreprise, suggérant l'importance des marchés du travail internes pour ces firmes. Les chapitres 3 et 4 portent sur la mobilité des travailleurs tel qu'affecté par la forma-tion formelle en milieu de travail. Au chapitre 3, nous utilisons des données américaines provenant du National Longitudinal Survey of Youth (NLSY) pour examiner l'effet de la formation formelle en lieu de travail par rapport à la mobilité observée des jeunes travailleurs américains. Des modèles de durée paramétriques nous permettent d'évaluer l'impact économique de la formation sur le temps productif passé avec un employeur. Nos résultats sont cohérents avec la plupart des études précédents, qui trouvaient un im-pact positif et significatif. Cependant, la durée de la relation de travail nette du temps passé en formation n'augmente pas de manière significative. Nous procédons par la suite à l'analyse de la mobilité intrasectoriel et intersectoriel à fin de permettre l'inférence par rapport à la spécificité du capital humain acquis par la formation, soit du capital humain spécifique à la firme, soit spécifique à l'industrie, soit général. L'analyse économétrique permet de rejeter un modèle séquentiel de choix de secteur en faveur d'un modèle à risques concurrents. Nos résultats présentent de l'évidence forte en faveur de la spécifi-cité de la formation à l'industrie. La probabilité d'un changement de secteur d'activité suite à une séparation d'emploi décroît avec la formation reçue dans l'industrie présente, peu importe si celle-ci a été reçue du dernier employeur ou d'un employeur précédent. La probabilité de détenir un emploi suite à une séparation augmente avec la formation sur le tas. Ces résultats sont robustes à des variations du modèle de base. Finalement, le chapitre 4 étudie la mobilité des travailleurs allemands à la lumière d'un modèle de capital humain dont la spécificité est sectoriel. En outre, j'utilise et décris des données peu utilisées sur la formation formelle ayant lieu après la fin d'un apprentissage. Comparativement aux États-Unis, un plus grand nombre de travailleurs suit une formation annuellement, et ce, en dépit d'une incidence élevée d'apprentissage précédemment. Tandis que plusieurs autres études font uniquement une distinction entre capital humain spécifique à une seule firme et capital humain général, je montre que les travailleurs allemands ont une plus grande probabilité de trouver un emploi dans un secteur s'ils ont suivi une formation formelle dans ce même secteur. Ce résultat n'est cohérent ni avec la présence de capital humain spécifique à une seule firme, ni avec du capital humain complètement général. Conjointement avec des résultats semblables pour des travailleurs américains, ils suggèrent l'importance du capital humain spécifique à l'industrie. Par ailleurs, l'effet de la formation sur la mobilité semble sensible à l'état de la conjoncture, suggérant une interaction entre offre et demande plus complexe que celle décrit par la plupart des modèles théoriques.

In the first chapter,a harmonised linked employer-employee dataset is used to study the impact of firm-level agreements on the wage structure in the manufacturing sector in Belgium, Denmark and Spain. To our knowledge, this is one of the first cross-country studies that examines the impact of firm-level bargaining on the wage structure in European countries. We find that firm-level agreements have a positive effect both on wage levels and on wage dispersion in Belgium and Denmark. In Spain, firm also increase wage levels but reduce wage dispersion. Our interpretation is that in Belgium and Denmark, where firm-level bargaining greatly expanded since the 1980s on the initiative of the employers and the governments, firm-level bargaining is mainly used to adapt pay to the specific needs of the firm. In Spain, the structure of collective bargaining has not changed very much since the Franco period where firm agreements were used as a tool for worker mobilisation and for political struggle. Therefore, firm-level bargaining in Spain is still mainly used by trade unions in order to reduce the wage dispersion. <p>In the second chapter, we analyse the impact of the bargaining level and of the degree of centralisation of wage bargaining on rent-sharing in Belgium. To the best of our knowledge, this is the first study that considers simultaneously both dimensions of collective bargaining. This is also one of the first papers that looks at the impact of wage bargaining institutions on rent-sharing in European countries. This question is important because if wage bargaining decentralisation increases the link between wages and firm specific profits, it may prevent an efficient allocation of labour across firms, increase wage inequality, lead to smaller employment adjustments, and affect the division of surplus between capital and labour (Bryson et al. 2006). Controlling for the endogeneity of profits, for heterogeneity among workers and firms and for differences in characteristics between bargaining regimes, we find that wages depend substantially more on firm specific profits in decentralised than in centralised industries ,irrespective of the presence of a formal firm collective agreement. In addition, the impact of the presence of a formal firm collective agreement on the wage-profit elasticity depends on the degree of centralisation of the industry. In centralised industries, profits influence wages only when a firm collective agreement is present. This result is not surprising since industry agreements do not take into account firm-specific characteristics. Within decentralised industries, firms share their profits with their workers even if they are not covered by a formal firm collective agreement. This is probably because, in those industries, workers only covered by an industry agreement (i.e. not covered by a formal firm agreement) receive wage supplements that are paid unilaterally by their employer. The fact that those workers also benefit from rent-sharing implies that pay-setting does not need to be collective to generate rent-sharing, which is in line with the Anglo-American literature that shows that rent-sharing is not a particularity of the unionised sector. <p>In the first two chapters, we have shown that, in Belgium, firm-level bargaining is used by firms to adapt pay to the specific characteristics of the firm, including firm’s profits. In the third and final chapter, it is shown that firm-level bargaining also allows wages to adapt to the local environment that the company may face. This aspect is of particular importance in the debate about a potential regionalisation of wage bargaining in Belgium. This debate is, however, not specific to Belgium. Indeed, the potential failure of national industry agreements to take into account the productivity levels of the least productive regions has been considered as one of the causes of regional unemployment in European countries (Davies and Hallet, 2001; OECD, 2006). Two kinds of solutions are generally proposed to solve this problem. The first, encouraged by the European Commission and the OECD, consists in decentralising wage bargaining toward the firm level (Davies and Hallet, 2001; OECD, 2006). The second solution, the regionalisation of wage bargaining, is frequently mentioned in Belgium or in Italy where regional unemployment differentials are high. In this chapter we show that, in Belgium, regional wage differentials and regional productivity differentials within joint committees are positively correlated. Moreover, this relation is stronger (i) for joint committees where firm-level bargaining is relatively frequent and (ii) for joint committees already sub-divided along a local line. We conclude that the present Belgian wage bargaining system which combines interprofessional, industry and firm bargaining, already includes the mechanisms that allow regional productivity to be taken into account in wage formation. It is therefore not necessary to further regionalise wage bargaining in Belgium. <p> / Doctorat en Sciences économiques et de gestion / info:eu-repo/semantics/nonPublished