Employer-provided training and knowledge spillovers Evidence from Italian local labour markets

Following suggestions from theoretical and empirical literature on agglomeration and on social returns to education which emphasise the contribution of local knowledge spillovers to productivity and wage growth, this paper aims at uncovering the relationship between local human capital and training. Furthermore, we check the effects of other variables which measure some distinctive features of local labour markets, like the degree of specialization, average firms’ size, intensity of job turnover, economic density, employment in R&D activities and some other control variables. Our key-results are consistent with the prediction that training should be more frequent in areas where the aggregate educational level is higher. Moreover, interaction between local and individual human capital is positive and significant for those with an upper secondaryeducational attainment. These results have proved to be robust since they are not altered when different definitions of local human capital are adopted or different sub-samples are considered (with the exception of female workers). We also coped with the problem of omitted variables and spatial sorting, that could bias econometric results, by means of a two-step strategy based on instrumental variables.