Department of Agricultural and Biological Engineering, University of Florida, Gainesville, Florida, United States of America.
Institute for Food and Resource Economics, University of Bonn, Bonn, Germany.
PLoS One. 2021 Jan 27;16(1):e0245867. doi: 10.1371/journal.pone.0245867. eCollection 2021.
Conventional empirical studies of foodborne-disease outbreaks (FDOs) in agricultural markets are linear-stochastic formulations hardwiring a world in which markets self-correct in response to external random shocks including FDOs. These formulations were unequipped to establish whether FDOs cause market reaction, or whether markets endogenously propagate outbreaks. We applied nonlinear time series analysis (NLTS) to reconstruct annual dynamics of FDOs in US cattle markets from CDC outbreak data, live cattle futures market prices, and USDA cattle inventories from 1967-2018, and used reconstructed dynamics to detect causality. Reconstructed deterministic nonlinear market dynamics are endogenously unstable-not self-correcting, and cattle inventories drive futures prices and FDOs attributed to beef in temporal patterns linked to a multi-decadal cattle cycle undetected in daily/weekly price movements investigated previously. Benchmarking real-world dynamics with NLTS offers more informative and credible empirical modeling at the convergence of natural and economic sciences.
传统的农业市场食源性疾病爆发（FDO）的实证研究采用的是线性随机模型，这些模型假设市场会对包括 FDO 在内的外部随机冲击进行自我修正。这些模型无法确定 FDO 是否会引发市场反应，或者市场是否会内在地传播爆发。我们应用非线性时间序列分析（NLTS），从美国疾病控制与预防中心（CDC）爆发数据、活牛期货市场价格和美国农业部（USDA）牛存栏量中重建了 1967-2018 年美国牛市场的 FDO 年度动态，并使用重建的动态来检测因果关系。重建的确定性非线性市场动态是内在不稳定的，不是自我修正的，牛存栏量推动了期货价格和牛肉 FDO，这些 FDO 的时间模式与以前研究的每日/每周价格波动中未检测到的数十年牛周期有关。使用 NLTS 对真实世界动态进行基准测试，可以在自然科学和经济科学的融合中提供更具信息量和可信度的实证模型。
