In the recent macroeconomics literature, discussion of inventories has revolved around a “production-smoothing” function for holding stocks. As Blinder (1986, p. 12) puts it: “A firm is said to smooth production if its production responds less to a sales [that is, a demand] shock than it would if it could not carry inventories.” At first glance this role for stockholding appears to be different from the consumption-smoothing buffering effect emphasized so far in this book. It also seems intuitively plausible. Because of unforeseeable fluctuations in the demand for final goods, a firm would have to alter its contemporaneous production schedule a great deal and at considerable expense, unless it held inventories of finished goods. By extension (so the argument goes), aggregate production should be less variable than aggregate final consumption, both for particular industries and for the economy as a whole. Nevertheless, this production-smoothing role of inventories, at least as formally modeled, appears to be contrary to the facts in the U.S. economy. As Blinder (1986) observes, a broad stylized fact is that aggregate current production is more variable than aggregate current final consumption. West (1986) also tests the model for particular industries, specifically food, tobacco, apparel, chemicals, petroleum, and rubber, and decisively finds final consumption to be less variable than production.

In the model and its variations here - by construction a model of inventories and production rationally atuned to the costs of irregular production - inventories smooth production, by Blinder's standard, even though they conform to the stylized fact of production (i.e., processing) being more variable than consumption.