Trade Inventories and (S,s)

Abstract

The paper presents empirical tests of the (S,s) model of inventory behavior using aggregate retail trade data. Estimation and testing are based on the probability distributions of inventories derived by Caplin [1985]. The excess volatility of retailers' demand over their consumers' demand, and the “forgetfulness” of inventories under (S,s) are emphasized. Test results indicate that the time series properties of deliveries and sales are consistent with (S,s) and not a quadratic cost model. Finally, when autoregressions of inventories are given an (S,s) rather than a stock adjustment interpretation, traditional empirical problems such as low speeds of adjustment are explained.