We develop a method for decomposing firm performance to impacts coming from the inflows and outflows of workers and apply it to study whether older workers are costly to firms. Our estimation equations are derived from a variant of the decomposition methods frequently used for measuring micro-level sources of industry productivity growth. By using comprehensive linked employer–employee data, we study the productivity and wage effects, and hence the profitability effects, of the hiring and separation of younger and older workers. The evidence shows that the separations of older workers are profitable to firms, especially in the manufacturing ICT-industries. To account for the correlation of the worker flows and productivity shocks we first estimate the shocks from a production function using materials as a proxy variable. In the second step the estimated shock is used as a control variable in our productivity, wage, and profitability equations.