Better ask first if climate has changed in the Eastern US.
A note on the the long-difference method – Burke and Emerick (2013)
Burke andEmerick (2013) study if corn and soybeans growers in the Eastern US have adapted to climate change from 1980 to 2000. Instead of climate they consider five-year weather and crop yield averages from 1978 to 1982 and from 1998 to 2002. For each county they calculate the differences of average yields between the five-year averages centered on 1980 and 2000 and regress it on the difference between 1980 and 2000 of the average number of degree days below and above 29 °C during April-September. They find that the coefficient of degree days above 29 °C is negative and significant, as expected. However, the coefficient is not significantly different from the coefficient estimated using a traditional panel model with fixed effect. It thus seems that the response function of yields is the same whether it is estimated using weather fluctuations or longer term temperature changes. They argue this is evidence of lack of adaptation.
I argue instead that this is just what one would expect to observe. Because climate has not changed in the Eastern US. Burke and Emerick (2013) captures noisy weather signals rather than a stable climate pattern.
See here for a longer discussion, references to the scientific literature and maps of climate patterns in the Eastern US.