Production smoothing and the bullwhip effect

摘要

The paper deals with demand signal processing technique namely, production smoothing, which is a production stabilization technique that mitigates the tolls of production variability namely overtime fees, strained toolsets and idle capacity. Another smoothing technique, the bullwhip can create disorder in the supply chain by amplifying the demand fluctuations up a supply chain. The present research work has tried to develop a new production smoothing metric, which estimates how much more variable production would be if firms did not moderate it. This has been done by benchmarking production to production, rather than to demand, thereby eliminating the bullwhip bias. A structural estimation model has been developed to explain the same. The model incorporates the martingale model of forecast evolution and the generalized order-up-to policy. The firm optimizes over the impulse response functions that govern its demand signal processing (DSP), its transformation of demand forecast revisions into production forecast revisions. This showed that when a firm faces linear-quadratic costs, each set of model primitives has a unique DSP signature. The DSP signature has been reverse engineered to estimate the firm's marginal costs. The empirical context was a sample of 162 car models produced by 20 auto manufacturers. The results showed that auto production would be 22% more variable without deliberate production stabilization amid a robust bullwhip effect: auto production is 220% as variable as auto sales. (46 refs.)