Working Paper: NBER ID: w25507
Authors: Richard B. Freeman; Wei Huang; Teng Li
Abstract: Using administrative data from a major Chinese insurance firm that raised its sales targets and rewards for insurance agents in a highly non-linear incentive system, we examine the effects of the changes on productivity, workers gaming the system, and the division of benefits from the new system between the firm and workers. We find that while the steeper incentive system creating the bunching distortions on which theories of non-linear incentives focus and other gaming behaviors by workers, the productivity increases dwarfed those costs. The magnitude and division of the productivity benefits improved the well-being of both the firm and workers. The firm gained about two-thirds of the higher net output, making the change profitable to it. Labor turnover fell, which suggests that the greater pay for workers from their one-third of the benefits exceeded the non-pecuniary cost of extra worker effort. The key to the success of non-linear incentives appears to rest more on its inducing workers to increase output than on its distortionary effects, suggesting that greater attention be given to the first order effects of motivating workers to produce more than to its incentivizing some distortionary behavior, which it does.
Keywords: nonlinear incentives; worker productivity; firm profits; quasiexperiment
JEL Codes: J00; J22; J3; M5; M52
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Increased likelihood of gaming the system (C70) | Selling fewer non-incentivized products (D49) |
| Increased likelihood of gaming the system (C70) | Higher rates of withdrawn commissions (G24) |
| Increased income for agents (L85) | Improved overall welfare (D69) |
| Increased net output for the firm (D21) | Increased profitability for the firm (L21) |
| Steeper nonlinear incentive system (J33) | Increased productivity among insurance agents (G52) |
| Increased productivity among insurance agents (G52) | Increased life insurance commissions (G52) |
| Steeper nonlinear incentive system (J33) | Increased life insurance commissions (G52) |
| Increased life insurance commissions (G52) | Increased income for agents (L85) |
| Steeper nonlinear incentive system (J33) | Increased net output for the firm (D21) |