Crop Yields: Random, Cyclical, or Bunchy?

This paper investigates whether changes in weather have random effects on crop yields. The data are yield series of several crops for the United States, individual major producing states, and major producing areas. Statistical tests show little evidence of nonrandomness in these series. There is almost no evidence of cycles or bunchiness prior to 1932, none for national average yields or the weighted average for major producing areas. Since 1932 there is some evidence of autocorrelation; however, this appears to reflect the uneven application rate of high yielding inputs, such as fertilizer and hybrid seed.