This thesis analyses the impact on productivity, employment, overtime, earnings and costs of shorter working hours with particular reference to the 1989/91 campaign by the Confederation of Shipbuilding and Engineering Unions (CSEU). The events leading up to the CSEU campaign and the reasons for its success are investigated. One result of this union success was the end of national bargaining. The changing role of national bargaining, including why it became a casualty of reduced hours, is also examined. Research on earlier reductions in hours has tended to suggest that productivity rises as a result of reduced hours. A review of this research concludes that the productivity effect of reduced hours has been overstated. It also raises some important methodological issues. The thesis presents research on 20 engineering plants where the 37-hour week was introduced as a result of the CSEU campaign. A variety of managers and union representatives were interviewed. In addition there was a survey of engineering plants. This included plants with unchanged hours. Finally, the effect of unions on working hours in the whole economy is explored using a large data set. The research finds that when engineering hours were reduced hourly earnings typically rose so that weekly pay was unaffected, at least in the short-term. While measures to increase productivity were a feature of collective agreements on reduced hours, there is little evidence that productivity has been permanently increased by reduced hours. The productivity-increasing measures would in general have been agreed without reduced hours, albeit somewhat later in many cases. There is even less evidence that reduced hours have affected output and overtime than there is of a productivity effect. So, increased employment is left as the major consequence of reduced hours. The recession, which was at its most serious when reduced hours were implemented, had a much larger effect on employment. This makes the employment effect of reduced hours hard to observe as it mainly took the form of job retention. Increased costs may well mean that the employment effect of reduced hours is a little less than it would otherwise have been.