This thesis examines the theoretical and empirical predictions of the effects of reductions in hours of work on wages and employment, the economic efficiency arguments for such reductions and the related issue of the determination of hours of work in a bargaining framework. The conventional approach, assumes that workers will want to maintain their incomes in the face of reductions in hours of work per period. This is difficult to justify theoretically, when hours and union/worker utility are taken properly into account. Rather, unions and workers that desire reductions in hours of work are likely to opt for the same or even a reduced hourly wage, leading to a significant employment effect of any such measure. We show that this result is true in labour markets and economies where unions determine or bargain over the wage and in models where firms set the wage because it affects their workers' productivity. It is also true when firms demand positive overtime, when such models account properly for the long-run movement of hours of work. The thesis examines also the determination of hours, employment and wages in a bargaining framework and shows that the employment effects of unionism are likely to be overestimated when no account is taken of the hours determination procedure. This allows us also to provide an economic rationale for union behaviour regarding reductions in hours of work and maximum hours legislation and determine the conditions under which reductions in standard hours of work can increase union utility and firm profits. Finally, empirical evidence is provided with a test of the relationship between the hourly wage and weekly hours of work, using aggregate data. We use a large number of variables and different estimation techniques to avoid simultaneity. Our results suggest, in line with the theoretical predictions, that changes in hours of work have no effect on the hourly wage.