Transferring local human capital: Geographical proximity and perceived employees' productivity

While there is a wealth of literature dealing with the spatial nature of knowledge and its transferral, I argue that the underlying mechanisms have not been sufficiently understood. Existing research relating the geography of inflows to firm productivity does not adequately address firm and individual level heterogeneity. Some issues that are not tackled are the following. First, knowledge differentness might be the most important for the most skilled employees that can be expected to have the greatest impact on firm's productivity. Second, distant job matches might be a selective sample: greater spatial mobility enables finding better matches based on non-spatial human capital. Third, nonlocal job matches might be different along several dimensions and it is not clear where exactly the benefits of different knowledge are realised and where the bottlenecks occur. In this paper I fill this gap by making an attempt to disentangle the reasons behind different productivity levels of local and nonlocal employees. I employ the framework of Simon and Warner [Journal of Labor Economics, 10(3), 1992). I look into wage growth between jobs, tenure and wage growth on the job to infer about the initial certainty about employee's productivity, the quality and riskiness of job matches. I argue that different productivity levels can be determined by nonlocal employees being outsiders to local working culture, business practices, knowledge base and networks. In addition, since job matches made over long geographical distances are costly and parties might be unwilling to commit so such job matches unless they have much trust in their quality, distant job matches might be a selective sample where the non-spatial human capital is matched better than in local job matches. I expect the selectivity to drive up wage growth between jobs, tenure and wage growth on the job. Not possessing local knowledge, on the other hand, is expected to drive them all down. A dataset is used that is based on register data for the Netherlands 2008-2012 linked to survey data. Preliminary findings show that in general, nonlocal employees experience a higher wage growth between jobs. They have shorter tenures and on-the-job wage growth that is not different in a statistically significant way from that of the local employees; those findings together indicate worse job matches. Yet if being nonlocal is instrumented, all of those effects become statistically not significant.