INTERACTION EFFECTS OF PROMOTION, RESEARCH, AND PRICE SUPPORT PROGRAMS FOR U.S. COTTON

Q^d + Q^d = Q^s
fd fx fd

Farm-level market clearance                                        (12)

where Q_i^k_j represents quantities for which k denotes quantity supplied (s) or demanded (d); i denotes
market level (retail [r], textile [t], or farm [f]); and j denotes domestic (d), export (x), or foreign (f). In
addition, P_rd is the retail price for domestic cotton products, P_rm is the retail price for imported cotton
products, Ag is generic promotion expenditures for cotton, Ab is branded advertising expenditures for
cotton, Af is advertising expenditures for man-made fibers, Z_r is a vector of demand factors in the retail
market other than advertising (e.g., income), P_td is the price of domestic intermediate cotton textiles, P_tm
is the price of imported intermediate cotton textiles, W_r is a vector of supply factors in the retail market
(e.g., retail wages, energy costs), P_cd is the domestic price of cotton at the mill level (net of Step 2
subsidies), P_cf is the price of foreign cotton fiber, R_a is agricultural cotton research expenditures (made by
both Cotton Incorporated (CI), the legal entity that implements the research and promotion activities
under the CRPP, and public institutions funding cotton research), R_t is nonagricultural research
expenditures made by CI, W_t is a vector of supply shifters for the cotton textile market (e.g., textile
wages, energy costs), Wf is a vector of supply factors for cotton producers (e.g., input costs, prices of
alternative crops), Z_x is a vector of demand factors for export markets, W_m is a vector of supply factors
for foreign cotton textiles, and Tf represents shifters of the supply of raw foreign cotton.¹⁰

Because our emphasis is on raw cotton producers and that is the market level for which data are
most readily available, partially reduced-form supply and demand equations at the farm level are
estimated. To estimate these equations while incorporating effects from the other levels of the market, the

¹⁰Both branded promotional expenditures for cotton-containing products and promotional expenditures for substitute
products such as man-made fibers are important to consider and were part of our theoretical model. However, in the
final model estimated empirically, these variables are not included. Promotional expenditures for man-made fibers
were not included because no data were available. Branded promotional expenditures were proxied using Levi
Strauss expenditures, but their inclusion revealed no significant impact on the coefficient for CI promotion. Because
the available data series for these expenditures was shorter than for the other demand variables (and was available
only quarterly rather than monthly for part of the series) and did not reveal significant interaction effects with CI
promotion, they were dropped from the final model.

<<   7   8   9   10   11   12   13   >>

More intriguing information