In recent years the question of overtime work has become increasingly relevant as part of the wider issue of the reduction in the working day. A direct relation between policies aiming at reducing working hours, and increases in overtime work neutralising their beneficial effects on employment, has been envisaged by those opposing such policies. We investigate this issue using microdata by the Bank of Italy. In Italy, the incidence of overtime work among male dependent workers is relatively high. In particular, we seek to ascertain if, for Italy too, the fear that a reduction in working hours could give rise to a substitution of overtime work for new jobs is legitimate. We estimate the probability of working overtime, together with equations for overtime hours of work, using different econometric models, both for cross-section (probit, tobit) and panel data (conditional fixed effects logit). Among several other variables, we control for wages and normal hours. We are particularly interested in exploring differences by sex. Overtime has always been studied over selected samples of male employees working in the private sector. Of course, focusing on workers who are most likely to work overtime will yield the result of a relatively large “substitution” effect. We show that extending the analysis to a more realistic labour market that includes female workers, this effect may become relatively modest for some specific policy measures. This result is robust across different sampling assumptions and model specifications, thus giving support to the hypothesis that the policies aiming at reducing the normal working day may have positive employment effects.