demand for cotton is -0.4, which is close to estimates of about -0.3 by Lowenstein (1952), Wohlgenant
(1986), and Waugh (1964) and in the range between Shui, Behgin, and Wohlgenant (1993), which found
an elasticity of -0.6 and Capps et al. (1997), which estimated an elasticity of -0.16.
The world price of cotton also has a large and significant effect on cotton mill use. This variable
is a strong indicator of the cost of imported cotton textile products. Higher world cotton prices raise the
cost of producing cotton in foreign countries, which translates into higher prices of cotton products
imported and higher U.S. mill consumption of cotton. It is important also to recognize that, because the
U.S. is not a small country in international trade of cotton, feedback effects may exist from changes in the
U.S. cotton price on the world cotton price. Therefore, the effect is included in simulations of the impact
of promotion and research on returns to cotton producers this feedback effect needs to be considered.
The elasticity of promotion is estimated to be 0.023 and the long-run elasticity estimates of mill
consumption with respect to research (the sum of current and lagged effects) is 0.35. These elasticities
imply that a 10 percent increase in promotion expenditure would lead to a 0.23 percent increase in cotton
demand, while a 10 percent increase in nonagricultural research expenditures would lead to a 3.5 percent
increase in cotton demand.
The results for the preferred model24 for export demand provided in Table 1 conform with other
studies of the cotton market. There is significant seasonality to U.S. cotton exports as indicated by highly
significant monthly dummy variables (not included in table), although the seasonal pattern of exports is
quite different from that of mill consumption. The export demand price elasticity for cotton is about -0.7,
which is just below the lower end of the range estimated for the export demand elasticity by Duffy,
Wohlgenant, and Richardson (1990) and is more elastic than domestic demand, as trade theory would
suggest. The effect of lagged exports is highly significant. It seems that the major factors contributing to
export demand are the domestic cotton price, the world cotton price, seasonality, and partial adjustment of
exports over time to these and other unobserved trade shocks, with none of the other variables having a
24An LM test (used because of the inclusion of a lagged dependent variable) reveals the presence of autocorrelation, so a
first-order autocorrelation correction was included in the model.
21