Drawing out implications for government policy from research on the labour process is a delicate thing to do. This piece of research, as any other that undertakes a realistic account of the relationship between machines and people in industry, cannot stop at the technical or economic level. One soon runs up against, or better into, the political field. We called it the politics of production, which arise from employers’ need to obtain a certain behaviour pattern from their workers and from workers’ reluctance or resistance to comply. This was an important, but limited, step to take; limited because the issue of control in the workplace was not followed through and connected sufficiently with the question of workers’ unions and the role of the state in developing countries. In our empirical work much attention was given to the Brazilian experience; few can have doubts about the nature of the Brazilian regime, even though recent years have seen a decline in repression and a strengthening of the labour movement. Thus, the policy implications that follow cannot and do not assume a neutral, benevolent government. The suggestions might, however, be of some use to those working within the state apparatus trying to contribute to a more genuine concern with employment, training and working conditions in industry.