Stablecoins as Dry Powder: A Copula-Based Risk Analysis of Cryptocurrency Markets
Stablecoins as Dry Powder: A Copula-Based Risk Analysis of Cryptocurrency Markets
Authors: Elliot Jones, Toshiko Matsui, William Knottenbelt Date: 2026-03-24 Paper ID: arxiv:2603.23480
Summary
This paper investigates the systemic role of stablecoins in cryptocurrency markets by employing copula-based models to quantify the transmission of risk and volatility. The authors establish causal links between stablecoin activity (volume and upside volatility) and broader crypto market dynamics across multiple time horizons. Incorporating these stablecoin factors was shown to significantly improve forecasting accuracy and reduce risk in a volatility targeting strategy, framing stablecoins as crucial “dry powder” infrastructure. The findings suggest stablecoin metrics are vital, non-redundant inputs for financial risk assessment in the digital asset ecosystem.
Key Contributions
- Quantified the transmission of volatility and activity from stablecoin markets to broader cryptocurrency markets using copula-based methods.
- Demonstrated in-sample causality of stablecoin factors on cryptocurrency dynamics across daily, weekly, and monthly time horizons.
- Showed that incorporating stablecoin metrics significantly reduces Mean Squared Error in cryptocurrency forecasting models.
- Established the economic value of stablecoin factors by demonstrating reduced risk in a volatility targeting strategy for cryptocurrency portfolios.
Limitations
The study focuses on causality using in-sample analysis; out-of-sample predictive power and the robustness of the identified causal links under extreme market stress warrant further investigation.
Open Questions & Future Work
Limitations
The study focuses on causality using in-sample analysis; out-of-sample predictive power and the robustness of the identified causal links under extreme market stress warrant further investigation.
Links
Metadata & Links
- url
- https://arxiv.org/abs/2603.23480
- paper_id
- 2603.23480
- paper_source
- arxiv
- domain
- finance
- tags
- time-seriesforecastinganomaly-detectionfinance
- architectures
-
- datasets
-
- skill
- TimeSeriesSkill
- created_at
- 2026-03-25T21:18:08Z