We adopt an approach known as bright spots analysis to identify U.S. regions with surprisingly high corn yields given regional expectations, seasonal weather, and soil characteristics. These counties are regional ‘surprises’ that, by definition, achieve unexpectedly high levels of agricultural productivity. We then use multinomial logistic regression to identify the actionable factors—or the factors over which agricultural stakeholders can exert a certain level of control—that most strongly predict whether a county is a bright spot. We find that farmers in surprisingly productive regions spend an average of $17.6 more per acre on fertilizer, $12.4 more per acre on labor, irrigate 12% more of operated land, and receive $6.6 more per acre from government programs than those cultivating in less productive regions. We conclude by questioning whether and to what extent these attributes of productive regions can be managed for a sustainable future.

中文翻译：

---

美国玉米生产的亮点

考虑到区域期望，季节性天气和土壤特征，我们采用一种称为亮点分析的方法来识别玉米单产惊人高的美国地区。根据定义，这些县是地区性的“惊喜”，实现了意想不到的高水平的农业生产力。然后，我们使用多项式Lo​​gistic回归来确定可操作的因素（或农业利益相关者可以施加一定程度的控制的因素），这些因素可以最有力地预测一个县是否是一个亮点。我们发现，出乎意料地生产区的农民每英亩肥料平均要多支出17.6美元，每英亩劳动力要多支出12.4美元，要灌溉的耕地多12％，政府计划所获得的每英亩土地收入要比生产少的地区多6.6美元。 。