We analyze Chicago based daily wheat price volatility over more than 140 years using a novel data set of daily high and low futures prices starting in 1877. We identify five long-run regimes and find that volatility shifts between regimes are statistically more pronounced than fluctuations within regimes, even when conditioning on economic states. Historical volatility estimates derived from average commodity price data, a common practice in empirical studies, exhibit a regime-dependent upward bias between 0% and 22%. The magnitude of the bias and the importance of regimes potentially explain contradictory findings on volatility patterns in earlier studies.

我们使用自 1877 年开始的每日最高和最低期货价格的新颖数据集，分析了 140 多年来芝加哥的每日小麦价格波动。我们确定了五种长期机制，并发现从统计上看，机制之间的波动变化比机制内的波动更为明显，即使在以经济状态为条件时也是如此。根据平均商品价格数据得出的历史波动率估计（实证研究中的常见做法）表现出 0% 至 22% 之间的制度依赖性上行偏差。偏差的大小和制度的重要性可能解释了早期研究中关于波动模式的矛盾发现。