A Blessing in Disguise? DeFi Exploits and Short-Horizon Responses in U.S. Commercial Paper Spreads

Battle Count: 74

The paper is primarily an academic finance/regulatory economics study rather than a trading strategy paper. However, it has moderate relevance to quantitative trading in several ways: (1) The documented flight-to-quality pattern in AA CP spreads around DeFi exploits could inform short-term fixed-income trading strategies; (2) Understanding the liquidity-recycling channel helps model cross-asset correlations between crypto and traditional money markets; (3) The market segmentation insights (prime vs. government MMFs, AA vs. A2/P2 CP) are relevant for relative-value strategies in short-term funding markets; (4) The event-study methodology and GIV framework could be adapted for other operational risk events. However, the effect sizes (2-3 bps) are small and the event frequency is low, limiting direct trading applicability.

The paper employs multiple sophisticated econometric techniques: stacked event studies with declustering, Jordà local projections with Newey-West HAC inference, granular IV construction and 2SLS estimation, covariate-adaptive Monte Carlo placebo tests, threshold regression (Hansen), gradient boosting ML, cross-asset DiD with regulatory segmentation, and frequency-aligned state-dependence regressions. Replicating the full analysis requires expertise in applied econometrics, access to multiple data sources (FRED, DeFiLlama, Rekt, Fed EFA), and careful implementation of the event-dating and declustering algorithms. The theoretical appendices (robust control, global games) add conceptual complexity but are not directly estimated.

Core data sources are publicly available: FRED (CP spreads, T-bill rates, VIX, DXY), DeFiLlama Hacks Dashboard, Rekt Database, and Federal Reserve EFA MMF holdings. However, the paper does not provide a GitHub repository or explicit code. The event list construction (top 50 exploits, declustering rules, event dating conventions) requires careful replication. The GIV instrument construction and covariate-adaptive placebo test involve algorithmic choices that would need code for full reproducibility. Monthly MMF holdings data from Fed EFA is accessible but requires specific query knowledge.