DeFi Lending Stress Meets Duration Risk: Modeling the Link Between Aave Health-Factor Regimes and Treasury Rate Conditions

Authors

  • Guanzheng Zhao Financial Engineering, Stevens Institute of Technology, NJ, USA Author
  • Ziwei Wen Enterprise Risk Management, Columbia University, NY, USA Author
  • Jeffrey Chen Information Management, UIUC, IL, USA Author

DOI:

https://doi.org/10.69987/JACS.2024.40208

Keywords:

Decentralized finance, DeFi lending, Aave, health factor, Uniswap, FRED-QD, Treasury rates, duration risk, regime modeling, forecasting

Abstract

This paper tests whether measured DeFi lending stress co-moves with U.S. Treasury duration conditions. The empirical design uses fixed-release datasets: the Aave V2 Health Factor Dataset released in 2022, the Uniswap Daily Transaction Indices by Network released in 2023, and the FRED-QD 2022m12 quarterly macroeconomic vintage. The final aligned panel contains 54 weekly Aave observations from 4 June 2021 through 25 June 2022. Aave account-level data are aggregated into weekly collateral, debt, health-factor, risky-borrower, and liquidation-zone measures. Uniswap Ethereum daily volume and concentration measures are converted into rolling seven-day controls, and FRED-QD Treasury variables are matched by calendar quarter. A standardized Aave stress score is built from the risky borrower share, liquidation-zone share, debt-collateral ratio, and inverse borrower median health factor, then split into low-, medium-, and high-stress terciles. In the aligned panel, the Aave stress score correlates 0.677 with the 10-year Treasury rate and 0.616 with the quarter-over-quarter 10-year rate change. The risky borrower share correlates 0.736 with the 10-year rate. Regime contrasts show that high-stress Aave weeks have an average risky-borrower share of 20.95%, compared with 7.07% in low-stress weeks. The same high-stress weeks have an average 10-year Treasury rate of 2.42%, compared with 1.57% in low-stress weeks. HAC-adjusted regressions estimate that a one-percentage-point higher 10-year Treasury rate is associated with 12.95 percentage points more risky borrowers and 10.37 percentage points more liquidation-zone borrowers. In expanding-window forecasts, a duration-market ridge model lowers one-week-ahead Aave stress-score RMSE from 3.45 for the historical mean and 1.34 for AR(1) to 1.06. The results identify an empirical association, not a causal proof, between DeFi lending stress and Treasury duration conditions.

Author Biography

  • Jeffrey Chen, Information Management, UIUC, IL, USA

     

     

     

Downloads

Published

2024-02-22

How to Cite

Guanzheng Zhao, Ziwei Wen, & Jeffrey Chen. (2024). DeFi Lending Stress Meets Duration Risk: Modeling the Link Between Aave Health-Factor Regimes and Treasury Rate Conditions. Journal of Advanced Computing Systems , 4(2), 93-105. https://doi.org/10.69987/JACS.2024.40208

Share