Stablecoins under Stress in a National Economy: Transaction-Level Evidence from Austrian Crypto-Asset Service Providers

Battle Count: 57

The paper is primarily a financial stability and regulatory monitoring study rather than a trading strategy paper. However, it provides valuable insights for quantitative trading: (1) the structural separation of retail vs institutional flows reveals different adjustment margins during stress, which could inform regime detection; (2) the finding that stablecoins do not act as uniform safe havens challenges common portfolio allocation assumptions; (3) the two-tiered USDC redemption mechanism creates predictable institutional arbitrage opportunities during peg stress; (4) the speed of flow responses (within days, no jurisdictional lag) informs liquidity risk models; (5) the concentration of volume in warm/cold wallet interactions (0.77% of transfers, 57.8% of volume) reveals market microstructure relevant for execution. The event-study framework and abnormal flow detection could be adapted for signal generation, but the paper does not propose or test trading strategies.

The methodology involves multiple complex components: (1) parsing full Bitcoin and Ethereum ledgers to extract transactions from 11.98 million asset transfers; (2) constructing directed graphs and computing network statistics (PageRank, eigenvector centrality, betweenness) for hub wallet identification; (3) implementing tracing algorithms for chained movements across same-CASP addresses with transaction matching; (4) applying multi-input clustering heuristics for Bitcoin counterparties; (5) computing chain-wide activity distributions and z-score thresholds for classification; (6) running event-study OLS regressions with bootstrap inference across multiple asset-group-flow combinations; (7) handling UTXO vs account-based model differences. The primary barrier is data access (confidential regulatory registry), but the analytical pipeline itself is substantial. The event-study component is standard econometrics; the blockchain data processing and network analysis are more technically demanding.

The core dataset (address-to-entity registry from Austrian FMA) is confidential and pseudonymized by agreement with Austrian financial authorities. The exact number of addresses is withheld. External data sources (Coingecko API, Google Trends, Carpentier-Desjardins et al. 2025 crypto crime dataset on Zenodo) are publicly available. The authors reference a codebase and documentation for tracing algorithms. Robustness checks with alternative thresholds and windows are provided. However, the primary regulatory registry data cannot be independently accessed or replicated by other researchers.