Patrick Harker, WFFI Academic Director
Who manages the variance? Operational concentration and the regulation of financial infrastructure | October 1, 2026
In July 2026 the United Kingdom placed four cloud providers under direct supervisory oversight as its first designated critical third parties. Comparable regimes are in operation or under construction across the European Union and the United States. I argue that all of them act on the wrong half of the problem. Expected systemic damage from operational failure is governed by a mean and a variance. Markets already manage the mean, because every institution cares about its own uptime. Nobody manages the variance, which is set by who fails together, and concentration and cross-layer alignment are variance rather than mean. Five recommendations follow, each using authority supervisors already hold. I also take on the strongest objection directly: in core processing, no bank can realistically run two vendors. The answer is not to force dual sourcing. It is to move the instrument to market structure.
Loretta Mester, WFFI Senior Fellow
How Do Large, Sophisticated Cryptocurrency Trades Impact Broader DeFi Market Dynamics? | September, 2026
Using transaction-level data for Bitcoin (BTC), Ethereum (ETH), and Wrapped Bitcoin (WBTC) matched to public “Whale Alert” signals, we examine how large-scale cryptocurrency trades shape market microstructure. Specifically, we test whether whale transactions alter the composition of active non-whale participants, trigger leader-follower herd behavior, and elevate short-term return volatility. We document a stark divergence between networks: Whale alerts heavily reshape (native) BTC participation by mobilizing previously inactive retail and institutional investors, whereas ETH and WBTC participant profiles remain highly stable on the Ethereum platform. Furthermore, non-whale investors exhibit strong leader-follower behavior in the BTC market — buying and selling in tandem with whale direction. This pattern is largely absent in the Ethereum ecosystem, except among the largest non-whale tranches. These behavioral dynamics directly mirror market stability: Whale alerts induce a brief 24-hour spike in BTC volatility (most acutely following WBTC alerts) but coincide with compressed volatility on the Ethereum platform. Additionally, these asymmetric market responses persisted across Ethereum’s transition from proof-of-work to proof-of-stake. These findings indicate persistent informational and structural asymmetries between large and small digital-asset investors.
Co-authored with Keith Hazen and Julapa Jagtiani of the Federal Reserve Bank of Philadelphia.
Itay Goldstein, WFFI Academic Director
The Fragility of Semi-Liquid Private Credit Funds | September 24, 2026
We study fragility in semi-liquid private credit funds, which have expanded rapidly and now manage over $300 billion in assets. These funds perform liquidity transformation by holding loans that are far less liquid than those held by traditional open-end mutual funds while allowing investors to redeem at NAV through quarterly repurchase offers, typically capped at 5% of shares outstanding. We show that cash buffers and contractual loan repayments are insufficient to meet repeated 5% quarterly redemptions; inflows tend to decline when outflows rise; and net outflows are met with loan sales and two forms of leverage: external borrowing and delayed payments through repurchases payable. Strategic complementarity arises, as redeeming investors impose liquidation and leverage costs on the remaining investors. Overall, redemption gates do not eliminate run-like fragility, raising concerns about hidden leverage in private markets, the design of retail access to private credit, and the allocation of liquidity risk to households.
Co-authored with Chuck Fang, professor of finance at Drexel University and Yao Zeng, professor of finance at the Wharton School.
Patrick Harker, WFFI Academic Director
Critical Infrastructure Concentration and Systemic Financial Fragility | August 18, 2026
The financial sector runs on a shared digital infrastructure: a small number of cloud platforms, core banking software vendors, payment and market utilities, data and identity services, and a growing layer of artificial-intelligence model providers. As institutions outsource more of their information systems to this common layer, the concentration of the layer itself becomes a source of system-level risk.
We develop an analytical model in which a continuum of financial institutions chooses provider portfolios from a market with endogenous concentration, and in which providers fail through both idiosyncratic and common-cause channels. We prove that equilibrium diversification falls strictly short of the social optimum whenever a joint-failure externality is present, with provider concentration strictly above the planner’s when entry responds; we derive comparative statics on the resulting wedge, and we characterize optimal governance. Following Weitzman (1974), we compare a price on concentration with a quantity rule that requires minimum portfolio diversity. The quantity rule typically dominates under empirically plausible conditions, because the marginal social damage from concentration is steep near the concentrated region while the marginal private cost of diversification is comparatively flat. We also study how the provider market responds, which can amplify or attenuate the value of regulation, and we derive a robust-optimization characterization of optimal portfolios when the common-cause failure structure is uncertain.
The paper connects the information systems literature on IT outsourcing and cloud concentration to the financial-networks, pecuniaryexternality, and operations research literatures, and it speaks directly to the critical third-party and digital operational resilience regimes now being implemented by the Financial Stability Board, the European Union, and national regulators: the United Kingdom’s critical third parties regime entered live operation in July 2026 with the designation of four cloud providers, the first concrete instance of the oversight architecture the model evaluates. The model is deliberately kept simple enough that every result is available in closed form, with the general-model counterparts collected in a separate technical companion; the analysis is foundational, and we are explicit about the limits of the model for setting regulatory parameters directly.
Itay Goldstein, WFFI Academic Director
The Limits of AI Trading | August 14, 2026
We show that powerful autonomous learning does not ensure trading dominance in competitive markets, even when humans hold systematically biased beliefs. We characterize how information structure and market sophistication bound reinforcement-learning traders’ performance relative to humans. Humans receive private signals, potentially including soft information unavailable to algorithms, and differ in their ability to extract information from prices and anticipate others’ trading, as captured by level-k reasoning. AI traders learn demand schedules from prices and realized profits and may also receive private signals. Despite heterogeneous initial conditions and persistent differences in learning parameters, all AI traders converge to a common probabilistic trading policy, generating algorithmic herding, which amplifies the equilibrium-price response to aggregate AI demand and compresses profits. Without an informational advantage, AI is outperformed by the most sophisticated humans because it learns from market data generated by average, rather than frontier, human sophistication. Even the least sophisticated humans can outperform AI when their private information is either very precise or very imprecise. AI sophistication rises with that of the humans who compete with and train it.
Co-authored with Winston Wei Dou, professor of finance at the Wharton School and faculty director of the Wharton AI in Finance Lab, Zigang Li, finance PhD candidate at the University of Toronto, and Liyan Yang, professor of finance at the University of Toronto.
Patrick Harker, WFFI Academic Director
Speed, Fraud, and Bank Liquidity: The Operational Architecture of Real-Time Payment Systems | August 13, 2026
A real-time payment system settles each payment within seconds, around the clock, with finality. The defining design choice is to drive the settlement window to zero. This paper shows that the zero window shuts off two stabilizers at once. A universal settlement delay buys time to screen fraud before settlement, and it restores the netting that real-time settlement takes away from banks, so one design primitive acts on a fraud channel and a bank-liquidity channel together and the welfare-optimal architecture must weigh it against both. The fraud side adds two suspension tools, targeted and system-wide, by analogy to speed bumps and circuit breakers in equity markets; the liquidity side adds the precautionary-buffer and risk-taking channel documented for Brazil’s Pix. The optimal delay is strictly positive and strictly larger than fraud alone would justify, it rises with deposit runnability, the two suspension tools partition rather than reinforce, the system-wide threshold loosens with concentration, and the pure realtime corner is dominated through two independent channels at once. A calibration disciplined by the instant-payments literature shows the optimal delay is measured in seconds and is roughly twice what fraud alone would justify, and the result is an instance of stability-induced fragility.
Itay Goldstein, WFFI Academic Director
Payments, Reserves, and Financial Fragility | May 30, 2026
We develop a theory of payments and fragility, emphasizing a conflict between money’s payment and non-payment functions (e.g., store of value). Payments and production occur smoothly when reserves are abundant and money’s payment functions dominate. Otherwise, delays and disruptions arise due to dynamic coordination. Payments and production may exhibit history-dependence, generating fragility in that past delays cause future disruptions without fundamental shocks. The theory explains delays and disruptions in modern bank payments and U.S. dollar-based cross-border transactions. Improving payment technologies may not eliminate this fragility. Our theory offers new insights into the implications of quantitative tightening (QT) and the evolution of payment systems over time.
Co-authored with Ming Yang, professor of finance at University College London and Yao Zeng, professor of finance at the Wharton School.
Patrick Harker, WFFI Academic Director
The Tightening Trap: Detection, Capacity, and Strategic Routing in Multi-Institution Fraud Prevention Networks | April 20, 2026
Faster payments equals faster fraud.” This saying, now common among payments executives, captures a central tension in modern financial payment systems: the same innovations that enable instant, irrevocable fund transfers also compress the window for fraud detection and create ways for sophisticated adversaries to attack the system. This paper develops a framework for analyzing fraud detection in payment systems in which fraud detection algorithms generate demand for human review and fraudsters strategically route attacks across institutions. The core finding is a tightening trap: tightening detection at one institution can increase aggregate system losses. This finding has direct implications for regulatory design. Detection-only mandates, such as requirements to improve model accuracy or lower false-negative rates, can lead to unintended consequences when they push review operations toward saturation without a corresponding investment in capacity. Effective fraud governance requires standards on both detection and operational resilience. The paper provides a computational framework for analyzing these tradeoffs and offers concrete design principles for regulators, payment system operators, and financial institutions involved in real-time payment systems.
Max Harris, WFFI Senior Fellow
Stablecoins and the Dollar: Historical Parallels and Future Risks | January 3, 2026
Stablecoins have burst onto the scene, offering the potential for cheaper, faster, and smarter payments — perhaps even strengthened dollar dominance. Yet, for all the novel technology, stablecoins present a host of age-old risks that caution against such boosterism. This paper draws parallels to 19th-century monetary history, focusing on the chaotic free banking era (1837-1863) and the more stable national banking era (1864-1913), and finds that the current regime under the GENIUS Act is closer to the former than the latter. The paper also considers additional risks that could undermine trust in stablecoins, including illicit finance and corruption.
Co-authored with Kenneth Rogoff, professor of finance at Harvard University.