Scholarlink

<p>The global onset of the COVID-19 pandemic in 2020 has had far-reaching impacts on health, economy, and, notably, household finance. This paper aims to assess and compare the financial vulnerability of United States (U.S.) households before and during the pandemic, while also exploring potential demographic factors associated with financial vulnerability using data from the 2019 and 2022 Survey of Consumer Finances (SCF). First, the hierarchical clustering method is employed to identify financially vulnerable households. Subsequently, supervised machine learning techniques, namely eXtreme Gradient Boosting (XGBoost), and Logistic regression are applied to examine the associations between the demographic variables at both the reference person and household levels, and financial vulnerability. Results unveil a slight increase in the percentage of financially vulnerable households during the pandemic (24.7%), compared to the pre-pandemic period (21.2%). Notably, homeownership, race and education emerge as the most important factors associated with household financial vulnerability during the pandemic. The Logistic regression analysis underscores that the pandemic intensifies the probability of financial vulnerability among reference persons who are female, aged 35 to 44, and with lower levels of education and financial literacy, as well as families not owning a home and having multiple children.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Household Financial Vulnerability Before and During the COVID-19 Pandemic: An Exploration Through Machine Learning Approach

  • 期刊论文

  • 发表日期：

    2026年1月9日

查看全文
  • Kexin Meng,
  • Jing Jian Xiao

摘要

The global onset of the COVID-19 pandemic in 2020 has had far-reaching impacts on health, economy, and, notably, household finance. This paper aims to assess and compare the financial vulnerability of United States (U.S.) households before and during the pandemic, while also exploring potential demographic factors associated with financial vulnerability using data from the 2019 and 2022 Survey of Consumer Finances (SCF). First, the hierarchical clustering method is employed to identify financially vulnerable households. Subsequently, supervised machine learning techniques, namely eXtreme Gradient Boosting (XGBoost), and Logistic regression are applied to examine the associations between the demographic variables at both the reference person and household levels, and financial vulnerability. Results unveil a slight increase in the percentage of financially vulnerable households during the pandemic (24.7%), compared to the pre-pandemic period (21.2%). Notably, homeownership, race and education emerge as the most important factors associated with household financial vulnerability during the pandemic. The Logistic regression analysis underscores that the pandemic intensifies the probability of financial vulnerability among reference persons who are female, aged 35 to 44, and with lower levels of education and financial literacy, as well as families not owning a home and having multiple children.