Abstract: Standard narratives portray older workers as less productive and even obstacles to innovation, suggesting that their exit should free firms to invest in new capital and adopt new technologies. I show instead that the sudden loss of older workers reduces firms’ capital stock and delays technology adoption. Exploiting the 2014 German pension reform as a natural experiment, I use linked employer employee data to estimate firm responses to unexpected retirements. Affected firms reduce capital persistently and are less likely to adopt new technologies. Effects are strongest in small firms, in firms with outdated capital, and in firms with little in-house training programs. To explain these results, I develop a simple model in which incumbents train entrants and lower technology adoption costs by carrying firm-specific human capital about production routines needed for integration of new capital vintages into the firm´s operations. Their unexpected loss reduces the marginal product of capital and raises the cost of adopting new vintages. 

Abstract: This paper examines the determinants of cross-sectional and lifecycle inequality using a lifecycle earnings process model that incorporates earnings mobility and non-employment risks across birth cohorts and over time. We show that changes in unobserved skill prices and the variance of individual fixed effects across cohorts are the primary drivers of inequality. While non-employment risk contributes little to cross-sectional inequality, it is central to explaining lifecycle inequality. To explain the increase in the variance of fixed effects, we interpret these within a Roy model as realized productivity influenced by both ability and task choice. We provide evidence that technological change can amplify inequality beyond the canonical skill price channel by strengthening the mapping from ability to productivity through within-occupational task sorting.