ECONOMICS AND MARKETING Panel Data Analysis of U.S. Cotton Yields for 2002-2011

Technological innovation in agriculture allows increased production while maintaining inputs at sustainable levels. Cotton yield increases, an annual average of 2.5% from 2002-2011, have been accompanied by acreage decreases. This research develops an aggregate U.S. cotton yield model based upon relevant variables identified in previous research. Results indicate that yield increases are attributable to technology and are not only due to acreage shifts that leave more productive land in cotton production. I ncreased crop yields are an indicator of technological innovation to agricultural production systems. That such innovation has occurred in cotton is indicated by the long-term increase in cotton yields in the U.S. since 1960. The increase has not been linear, however, exhibiting variable rates of increase interspersed with periods of stagnant and in some cases, declining yields. The most rapid period of yield increase began in 2002. National trends in crop yield increase are an aggregate of regional production trends which may be influenced by region specific factors. The inclusion of these regional factors in crop yield analyses could be important in understanding the trends. Previous research related to crop yields has focused on weather, technology, and land as factors influencing crop yields. Weather variables affecting yield are typically precipitation and temperature, or an index that incorporates these factors. Technology is most often included in yield models as a trend with a specified functional form. Land factors affecting yields include considerations of either soil characteristics or acreage quantities. Tannura, Irwin, and Good (2008) investigated the relationship between weather, technology, and corn and soybean yields in the U.S. Corn Belt. Analysis of multiple regression results showed that corn yields were particularly affected by technology, the magnitude of precipitation during June and July, and the magnitude of temperatures during July and August. The effect of temperatures during May and June appeared to be minimal. Soybean yields were most affected by technology and the magnitude of precipitation during June through August. Tests for structural change did not indicate a significant change in the technology trend for corn or soybeans since the mid-1990s. Choi and Helmberger (1993) estimated the sensitivity of corn, wheat, and soybean yields to changes in price and land idled. Yields were found to be insensitive to price changes. The research did not find significant evidence that land idling programs significantly affect crop yields. Foster and Babcock (1993) investigated how changes in federal tobacco policy affected levels and growth of flue-cured tobacco yields. The research used an index of available technologies that was derived from research-station data and that allowed distinguishing effects of new technologies and adoption decisions. The empirical results showed that tobacco yield levels and the responsiveness of yields to changes in available technology depend upon price effects of program design. Specifically, the 1965 drop in land rents and output price, resulting from the shift from acreage allotments to poundage quotas, decreased yield levels by 12 percent, In addition, the movement to poundage quotas decreased the responsiveness of yields to changes in available technology. These findings are consistent with the hypothesis that high land prices lead to high adoption rates of yieldincreasing technologies. The growth of yields declined from an annual rate of 4.32 to 0.5 percent because of a change in relative prices and a slowdown in the rate of increase of available technologies. Geigel and Sundquist (1984) reviewed the literature for models which develop specific relationships between climatic variables and crop yields. The authors found that most past modeling of crop yields had focused on short-term (intraseasonal) “weather” and not long-term “climatic” related variables. In order to fully explain changes in crop yields, these models have also tried to account for the impacts of changing production technologies.