Skip to main content

Research

Working Papers

Treasury Bill Shortages and the Pricing of Short-Term Assets (co-author: Adrien d'Avernas)

R&R at the Journal of Finance

Abstract: Since 2008, yields on money market assets have become increasingly associated with changes in the supply of short-term Treasuries. This paper documents and rationalizes these developments in a model with heterogeneous banks subject to a liquidity management problem and regulation. The combination of large amounts of excess reserves and more stringent regulation prevents banks from intermediating liquidity to shadow banks. Hence, the pricing of reserves disconnects from the pricing of money-like assets, making them free to react to variations in Treasury bills. Our model accurately predicts post-crisis time series for short-term yields and quantities at the Federal Reserve's reverse repo facility.

Intraday Liquidity and Money Market Dislocation [NEW DRAFT COMING SOON] (co-authors: Adrien d'Avernas and Damon Petersen)

Abstract: This paper investigates the pricing of repurchase agreements (repos) within the new post-crisis regulatory framework. We find that new liquidity regulation prevents banks from using intraday credit provisions from the Fed. As a consequence, reserve-rich banks—rather than the Fedare the marginal provider of liquidity to money markets. In this new regime, intraday liquidity can suddenly become scarce and constrain the supply of repo, leading to sharp increases in repo rates. These spikes in repo rates are more likely when the supply of Treasury debt financed by shadow banks is large and settlement volumes are high. 

Can Stablecoins be Stable? (co-authors: Adrien d'Avernas and Vincent Maurin)

Abstract: This paper proposes a framework to analyze the stability of stablecoins -- cryptocurrencies designed to peg their price to a currency. We study the problem of a monopolist platform earning seignorage revenues from issuing stablecoins and characterize equilibrium stablecoin issuance-redemption and pegging dynamics, allowing for various degrees of commitment over the system’s key policy decisions. Because of two-way feedback between the value of the stablecoin and its ability to peg the currency, uncollateralized (pure algorithmic) platforms always admit zero price equilibrium. However, with full commitment, an equilibrium in which the platform maintains the peg also exists. This equilibrium is stable locally but vulnerable to large demand shocks. Without a commitment technology on supply adjustments, a stable solution may still exist if the platform commits to paying an interest rate on stablecoins contingent on its implicit leverage. Collateral and decentralizing stablecoin issuance help stabilize the peg.

Central Banking with Shadow Banks (co-authors: Adrien d'Avernas and Matthieu Darracq Pariès)

Abstract: This paper investigates how the presence of shadow banks affects the ability of central banks to offset a liquidity crisis. We propose an asset pricing model with heterogeneous banks subject to funding risk. While traditional banks have direct access to central bank operations, shadow banks rely on the intermediation of liquidity from traditional banks. In a crisis, this intermediation stops due to lack of collateral and shadow banks are left without lender-of-last-resort. Traditional instruments are not sufficient to fully mitigate the crisis. Opening liquidity facilities to shadow banks and purchasing illiquid assets is then necessary to further boost asset prices and tackle the crisis.

A Solution Method for Continuous-Time General Equilibrium Models (co-author: Adrien d’Avernas and Damon Petersen) PYTHON LIBRARY

Abstract: We propose an algorithm capable of solving in a fast and standardized way a general class of continuous-time asset pricing models, including heterogeneous agent models. These models typically require to solve for a system composed of a Hamilton-Jacobi-Bellman equation for each agent, coupled with a system of algebraic equations. The resolution of such a system of PDEs is a tedious problem as approximation errors tend to amplify and create explosive dynamics. We rely on a Finite-Difference algorithm and show how using a Brocot-Tree decomposition as advocated by Bonnans, and al. (2004) allows for fast and stable convergence in settings with up to two endogenous and stochastic state variables. 

 

Work in Progress

The Bright Side of Transparency: Evidence from Supervisory Capital Requirements (co-authors: Nordine Abidi, Livia Amato, and Ixart-Miquel Flores). Draft available on demand.

A Model of High Risk Premia Stagnation (co-authors: Adrien d'Avernas and Valentin Schubert)