Lepori Gabriele M, Morgan Sara, Assarian Borna A, Mishra Tapas
University of Southampton, Southampton, UK.
University of Southampton, Southampton, UK; University Hospital of Southampton, NHS, Southampton, UK.
Soc Sci Med. 2024 Feb;342:116538. doi: 10.1016/j.socscimed.2023.116538. Epub 2023 Dec 29.
The relationship between economic activity and suicides has been the subject of much scrutiny, but the focus in the extant literature has been almost exclusively on estimating associations rather than causal effects. In this paper, using data from England and Wales between January 1, 1997 and December 31, 2017, we propose a plausible set of assumptions to estimate the causal impacts of well-known macroeconomic variables on the daily suicide rate. Our identification strategy relies on scheduled macroeconomic announcements and professional economic forecasts. An important advantage of using these variables to model suicide rates is that they can efficiently capture the elements of 'surprise or shock' via the observed difference between how the economy actually performed and how it was expected to perform. Provided that professional forecasts are unbiased and efficient, the estimated 'surprises or shocks' are 'as good as random', and therefore are exogenous. We employ time series regressions and present robust evidence that these exogenous macroeconomic shocks affect the suicide rate. Overall, our results are consistent with economic theory that shocks that reduce estimated permanent income, and therefore expected lifetime utility, can propel suicide rates. Specifically, at the population level, negative shocks to consumer confidence and house prices accelerate the suicide rate. However, there is evidence of behavioural heterogeneity between sexes, states of the economy, and levels of public trust in government. Negative shocks to the retail price index (RPI) raise the suicide rate for males. Negative shocks to the growth rate in gross domestic product (GDP) raise the population suicide rate when the economy is doing poorly. When public trust in government is low, increases in the unemployment rate increase the suicide rate for females.
经济活动与自杀之间的关系一直是诸多审视的主题，但现有文献的重点几乎完全集中在估计关联上，而非因果效应。在本文中，我们利用1997年1月1日至2017年12月31日期间英格兰和威尔士的数据，提出了一组合理的假设，以估计知名宏观经济变量对每日自杀率的因果影响。我们的识别策略依赖于定期的宏观经济公告和专业经济预测。使用这些变量来构建自杀率模型的一个重要优势在于，它们能够通过观察经济实际表现与预期表现之间的差异，有效地捕捉“意外或冲击”的因素。倘若专业预测是无偏且有效的，那么估计出的“意外或冲击”就“近乎随机”，因此是外生的。我们采用时间序列回归，并给出了有力证据，证明这些外生的宏观经济冲击会影响自杀率。总体而言，我们的结果与经济理论一致，即那些降低估计的永久收入、进而降低预期终身效用的冲击，会推动自杀率上升。具体而言，在总体层面上，消费者信心和房价的负面冲击会加速自杀率上升。然而，有证据表明，在性别、经济状况以及公众对政府的信任程度方面存在行为异质性。零售价格指数（RPI）的负面冲击会提高男性的自杀率。当经济表现不佳时，国内生产总值（GDP）增长率的负面冲击会提高总体自杀率。当公众对政府的信任度较低时，失业率的上升会提高女性的自杀率。
