We deflated all data series denominated in dollars using the CPI prior to use in estimation,
including all promotion and research expenditure data.19 CPI data are available online from the U.S.
Bureau of Labor Statistics (2001). For all of the models, we used the series for all U.S. urban consumers.
4. Econometric Estimation
4.1 Equation specifications
Mill-level demand for cotton was modeled as described by (15). In this specification, per capita
mill consumption of cotton is a function of the price of cotton, prices of substitute fibers, demand and
supply shifters of domestically produced textile products, and demand and supply shifters of foreign-
produced textile products. The demand and supply shifters of foreign-produced textile products represent
the impact of the price of imported textile products on mill-level demand for cotton.
The model was estimated in linear form and the following variables were included (the
corresponding variable name from (15) is included in parentheses):
MILLUSEt (Qfdd) = U.S. per capita raw cotton used by mills (pounds per person)
Mi,t = monthly dummy variables (Mi =1 for ith month, 0 otherwise) for i =
1,...,11 where December is the reference month with its effect
represented by the intercept
PCOTTONt (Pcd) = real U.S. raw fiber equivalent price of cotton (cents/lb)
PPOLYt (Zr) = real U.S. raw fiber equivalent price of polyester (cents/lb)
PRAYONt (Zr) = real U.S. raw fiber equivalent price of rayon (cents/lb)
DTEXWt (Wt) = real domestic wage in U.S. textile manufacturing industry ($/hour)
WPCOTt (Pcf) = real A Index of world cotton price (cents/lb)
DECIt (Wt, Wr) = U.S. real energy cost index (1982-84=100)
DPIt (Zr) = U.S. per capita real disposable income ($1,000/person)
FGDPt (Zr) = real GDP of OECD countries, excluding U.S. (billions of $)
SAGPROMt (Ag) = seasonally adjusted CI real promotional expenditures ($)
19Another option for deflating the promotional expenditures would be to use a media cost index. However, using a
media cost index implies that the question of interest is the effectiveness of a given quantity of promotion, whereas
the question we are interested in for this study is the effectiveness of a real dollar of promotion. The opportunity cost
of using a dollar for promotion is that it cannot be used for other activities (e.g., nonagricultural research) or returned
to those paying assessments, so that is the relevant comparison. Unless the rapid increase in the costs of promotion
relative to the general price level in recent years has been accompanied by an equal or greater increase in its
effectiveness relative to other CI activities, the rapid media cost increases may suggest that promotion is becoming
relatively less attractive over time as the real price of a unit of promotion becomes more expensive. This point is not
captured by the use of a media cost index to deflate expenditures.
16