Sales persistence and the reduction in GDP volatility

A number of explanations for the observed decline in GDP volatility since the mid-1980s have been offered. Valerie Ramey and Daniel Vine (2003a, 2003b) in a couple of recent papers offer the hypothesis that a decline in the persistence of sales is an explanation for the decline in GDP volatility [Ramey, V.A., Vine, D.J., 2003a. The Effects of Sales Persistence on Volatility in a Production Smoothing Model. Draft dated November 18; Ramey, V.A., Vine, D.J., 2003b. Tracking the source of the decline in GDP volatility: an analysis of the automobile industry. Unpublished Working Paper, presented on the ASSA program, San Diego, January 2004]. Their models show that a decrease in sales persistence leads to a decline in the variance of production relative to the variance of sales. They provide econometric evidence that the persistence of unit automobile sales has declined at both the aggregate and model level. This paper explores reasons why sales persistence may have declined and then tests the Ramey-Vine hypothesis with monthly chain-weighted sales data from 2- and 3-digit manufacturing and trade industries. The estimates confirm the Ramey-Vine findings for motor vehicle retailers, wholesalers, and manufacturers. For a number of industries outside of motor vehicles, especially those in wholesaling and nondurable manufacturing, considerable evidence is found of declines in sales persistence. These declines seem to be consistent with changes in supply and distribution chains that have occurred as the result of the introduction of new information, inventory, and production control systems. However, in equations estimated for aggregate manufacturing, wholesaling, and retail sector sales, declines in sales persistence are not found.