Conference Program

Neural-Network Models for Long-term Care Insurance Insolvency: An Enterprise Risk Management Approach (310169)

*Sebastain Awondo, University of Alabama 
Haileab Halifu, University of Tennessee 

Keywords: Neural-Network, Predictive models, Long-term care insurance, Insolvency

We develop and apply neural-network models to predict Long-term Care Insurance (LTCI) insolvency using firm-level accounting and financial time-series data and enterprise risk management framework. Long-term care services currently account for about 1 percent of the U.S gross domestic product and is expected to rise to 3 percent by 2050. According to the Department of Health and Human Services (HHS), at least half of elderly Americans will need long-term care at some point. Despite the growing need, the number of insurers offering LTCI coverage has decreased from over a 100 in 2004 to about dozen in 2020 due to insolvency caused by poor initial pricing assumptions and unexpectedly high insurer exposure to claims payments. Predicting the solvency status of LTCI assist state insurance regulators to make timely informed decisions on insurer rate increase filings and enhancement of insurer reserve adequacy to ensure their future solvency.