This paper presents a new stage-of-fabrication inventory model with deliver
y, usage, and stocks of input materials that distinguishes between gross pr
oduction and value added. It extends the linear-quadratic model of output i
nventories by adding the joint determination of input inventories. Empirica
lly, input inventories are more important than output inventories. Maximum
likelihood estimation of the decision rules yields correctly signed and sig
nificant parameter estimates using data for nondurable and durable goods in
dustries, but the overidentifying restrictions of the model are rejected. T
he value added specification dominates because adjustment costs on material
s usage are critical to fitting the data. (C) 2001 Elsevier Science B.V. Al
l rights reserved.