Risk Assessment of Synthetic GICs: A Quantitative Framework for Asset Liability Management

In collaboration between Valerian Capital Group, the Actuarial Science Department of the University of Lyon, and CNAM University in Paris, this research provides insights into the risk management of stable value wraps, particularly from the insurer’s perspective.

The initial objective of the paper is to establish a quantitative framework enabling insurers to assess tail risks effectively. Studies have shown that when the risk is very deep in the tail, cognitive bias is to either neglect or become overly concerned by them. The history of wraps supports this: some providers have exited the market partially as a result of misjudgments regarding its risks, among other reasons, while others perceive these products as almost zero risk. Consequently, a more objective and quantitative approach to evaluating tail risks is critical for the industry. This research aims to provide a method to assess the risk associated with stable value wraps and also facilitates comparisons of risk across different transactions.

Building on this objective, the paper introduces an asset-liability model. The asset model is designed to integrate the two most critical aspects of any fixed income fund: duration and yield. On the other hand, the liability model is driven by observed participant behaviors and incorporates two significant components: disintermediation risk, and a regime-switching cash flow trend component influenced by various factors, including the plan’s structure, employer characteristics, participant demographics, etc.

The paper draws several conclusions. First, it highlights two key scenarios for stress testing by insurers: a yield inflationary scenario, reminiscent of post-2022 conditions, and a yield spike mean-reverting scenario, similar to the 2008 financial crisis. Additionally, it addresses the escalation of risk with the duration of held assets, despite being deep in the tail. Furthermore, the paper discusses regulatory implications, noting that the authors’ interpretation of the NAIC’s reserve and RBC language results in zero capital and reserves under normal and moderate stress conditions. This finding contrasts with the risk metrics estimated through a Monte Carlo model used in the study, suggesting that despite the risk’s low probability, it is not zero. Such findings also differ from the capital requirements if the product were provided in other regulatory regimes such as Solvency II or the Swiss Solvency Test.

For a comprehensive exploration of the research, including detailed methodologies and further discussions, please refer to the full paper.