by F. Owen Irvine
No. 3, January 2005 - June 2005

Motivation for the Research
A number of explanations have been offered for the observed decline in GDP volatility since the mid 1980s. Valerie Ramey and Daniel Vine offered the hypothesis that a decline in the persistence of sales led to the decline in GDP volatility. Their theoretical production smoothing models show that a decrease in sales persistence leads to a decline in the variance of production relative to the variance of sales. They estimated equations explaining unit sales of motor vehicles that show that sales persistence declined in this industry after 1983.

This paper tests the Ramey-Vine hypothesis, first on sales date from the aggregate retail, wholesale, and manufacturing sectors, and then on sales data from many other industries. Finally, it explores reasons why sales persistence may have declined in some of these industries.

Research Approach
Using monthly, seasonally adjusted sales data for available 2- and 3-digit SIC manufacturing and trade industries from January 1967 through March 2001, the author estimates equations using the same empirical model employed by Ramey and Vine.

Key Findings

• The estimates confirm the Ramey-Vine findings for motor vehicle retailers, wholesalers, and manufacturers.
  
  
• In equations estimated for aggregate manufacturing, wholesaling, and retail sector sales, declines in sales persistence were not found, despite the fact that these manufacturing and trade sales aggregates each contain motor vehicle sectors.
  
  
• While sales variance decreased for most durable goods industries, outside of motor-vehiclerelated industries this reduction in variance came mainly through channels other than a reduction in sales persistence.
  
  
• For the nondurable goods industries, sales persistence is estimated to have increased significantly at the retail level but decreased at the wholesaler and, especially, the manufacturing level.
  
  
• For a number of industries outside motor vehicles, especially those in wholesaling and nondurable manufacturing, considerable evidence was 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 technology and the adoption of new inventory and production control systems.

Implications
There appears to be no indication of a decline in the persistence of the sales of the main goods sectors of the economy—manufacturing, wholesaling, and retailing. In this sense, the Ramey-Vine hypothesis—that declines in sales persistence are a major separate explanation of the decline in GDP volatility—is not supported.

However, the finding that sales persistence has declined in a number of industries seems to be consistent with the structural changes that have occurred as a result of the introduction of new information technology and the new relationships between upstream and downstream firms in supply and distribution chains. To prove that structural change is a major reason for the decline in GDP volatility will require further detailed modeling of the interactions among industries. Declines in sales persistence will probably be part of that story.

Full text of Working Paper 05-5