According to the production-smoothing buffer stock model, inventories of a firm’s

products allow it to supply unexpected demand without having to adjust output

immediately. When costs per additional unit produced are increasing, using

inventory to smooth production is efficient as long as the savings from not

adjusting production exceed the cost of holding inventory. Inventory acts as a

buffer stock, absorbing increases or decreases in demand while production

remains relatively steady.

If firms are smoothing production, then sales should vary more than

production. If inventories are used as a buffer stock, then high-frequency

changes in inventory should be in the opposite direction to sales. Empirical

research using aggregate data does not confirm this expectation, however;

production varies more than sales, and changes in inventory and sales tend to

vary in the same direction. Thus either the production-smoothing buffer stock

model is incorrect, or other factors ere preventing empirical confirmation of the

smoothing effect.

Most of the research that finds contradictions of production smoothing uses

seasonally adjusted aggregate data concerning inventory and sales. It is possible

that firms actually do use inventory to smooth production, but the research has

failed to detect signs of this activity because the data are too highly aggregated

over many firms. But even research at the level of individual companies has

failed to confirm the model.