Manufacturing Productivity with Worker Turnover
70 Pages Posted: 25 Sep 2018 Last revised: 22 Nov 2019
Date Written: November 18, 2019
We find that rapid worker turnover significantly disrupts the productivity of responsive manufacturers. Our study uses a uniquely rich dataset drawn from China-based FATP (final assembly, testing, and packaging) facilities that produce millions of units of consumer electronic goods weekly yet exhibit high worker turnover exceeding 300% annually. The data cover the firm's weekly production plans, 52,214 workers' compensations and assignments, and assembly station productivity. To study managerial prescriptions, we extend the classical production planning problem to include endogenous worker turnover as an Experience-Based Equilibrium and use advances in reinforcement learning and approximate dynamic programming to estimate and simulate our model. Our empirical analyses exploit instrumental variables, including the firm's demand forecasts as ``demand shifters''. We find that turnover's impact on yield waste is conservatively $146-178M, and that a well-calibrated wage increase reduces the manufacturer's variable production costs (including wages) by up to 21%, or $594M for the product we study. The wage increase reduces the firm's reliance on a larger workforce and overtime to hedge against yield disruptions from turnover; it stabilizes a leaner workforce and improves both production reliability and flexibility. In settings where performance depends on workers repeating known tasks in coordinated groups, our results suggest that firms responsively matching supply to demand can pay a steep price for a disruptively turnover-prone workforce.
Keywords: Data-driven workforce planning, Empirical operations management, Employee turnover, Experience-Based Equilibrium, Production planning, Productivity, Stochastic optimization, Structural estimation
Suggested Citation: Suggested Citation