This paper investigates the historical behavior of initial unemployment claims (ICSA) in the United States (U.S.) during all the recession periods and epidemic diseases such as Severe Acute Respiratory Syndrome (SARS), Middle East Respiratory Syndrome (MERS) and COVID-19 since 1967 by applying statistical methods based on long range dependence and fractional differentiation. Using unit root tests (ADF, PP and KPSS) we discover that the original time series is stationary I(0) and the subsamples are non-stationary I(1). Finally, to analyze the original time series as well as the several periods corresponding to the recessions that occurred in U.S. and the three epidemic diseases, we use AIC and BIC criterion to fit the best ARFIMA model. We conclude that the results display long memory with a degree of integration strictly below 1 (d ​< ​1) for the COVID-19 episode and for the rest of the subsamples, except for the original time series and the 2nd subsample. Thus we can conclude that the impacts will be transient and with long lasting effects of shocks and expecting to disappear on their own in long term. Finally, we use a methodology proposed by Bai and Perron to estimate structural breaks not being necessary to know the time of the breaks in advance. The results are similar to those obtained previously.

本文调查了美国 (US) 在所有经济衰退时期和流行病如严重急性呼吸系统综合症 (SARS)、中东呼吸综合症 (MERS) 和 COVID-19 以来的历史行为。 1967 年通过应用基于长程相关性和分数微分的统计方法。使用单位根检验（ADF、PP 和 KPSS），我们发现原始时间序列是平稳的 I(0)，而子样本是非平稳的 I(1)。最后，为了分析原始时间序列以及与美国发生的衰退和三种流行病对应的几个时期，我们使用 AIC 和 BIC 准则来拟合最佳 ARFIMA 模型。我们得出结论，结果显示长记忆，集成度严格低于 1 (d < 1) 对于 COVID-19 剧集和其余子样本，原始时间序列和第二个子样本除外。因此，我们可以得出结论，这些影响将是暂时的，并且具有长期持续的冲击影响，并有望在长期内自行消失。最后，我们使用 Bai 和 Perron 提出的方法来估计结构性中断，而不需要提前知道中断时间。结果与先前获得的结果相似。我们使用 Bai 和 Perron 提出的方法来估计结构性中断，而不需要提前知道中断时间。结果与先前获得的结果相似。我们使用 Bai 和 Perron 提出的方法来估计结构性中断，而不需要提前知道中断时间。结果与先前获得的结果相似。