My research interests are: Finance, Banking, Macroeconomics, Economic History, Computational Economics and Econometrics.
My research statement: link to latest version (August 2023)
(Also circulated as "Government Funding Advantage and Financial Repression" and "Convenience Yields and Financial Repression".)
with Bálint Szőke.
Link to latest version (March 2025).
Abstract: US federal debt plays a special role in the US economy and so gives the US government a funding advantage, often summarized by the spread between the yield on high-grade US corporate bonds and comparable US treasuries. Why? One reason is that government regulation (and/or repression) of the financial sector influences asset pricing and helps make long term US federal debt an endogenously ``safe-asset''. We study the mechanics, limitations, and macroeconomic trade-offs involved with generating a government funding advantage through restrictions on the financial sector. We show the government cannot choose all three of: (i) high funding advantage, (ii) a well-functioning financial sector, and (iii) fiscal policy that leads to systematic debt devaluation. We relate our theories to new US historical corporate and treasury yield curve data from 1860-2024.
with Clemens Lehner, Jack Shurtleff, and Bálint Szőke.
Link to latest version (April 2025).
Abstract: We estimate a historical funding cost advantage of the US government, as measured by the spread between yields on high-grade corporate bonds and treasuries. We construct a new dataset with monthly price, cash-flow, and rating information for US corporate bonds over the period 1860-2024. We deploy a Kernel Ridge regression to estimate US high-grade corporate and treasury yield curves making adjustments for tax treatment and time-varying embedded option values. A high-grade corporate to treasury spread emerged well before Bretton Woods with the introduction of the 1862-65 National Banking Acts. Previous estimates have mismeasured and exaggerated US funding advantage in the post-WWII period. In particular, funding advantage is negatively correlated with inflation and goes to zero during the Great Inflation in the 1970s-80s. We find little evidence that the US strategically exploits its monopoly power.
with Bálint Szőke. (Forthcoming in the ARFE Volume 17.)
Link to latest version (January 2025).
Abstract: US federal debt plays a special role in the US economy and so gives the US government a funding advantage, often summarized by the "convenience yield" on US debt. Why? One reason is that government regulation (and/or repression) of the financial sector influences asset pricing and helps makes long term US federal debt a "safe-asset". We study the macroeconomic consequences on government borrowing capacity, financial stability, and investment. We then test our theory using new historical data on US convenience yields going back to 1860 and data from the Eurozone sovereign debt crisis.
with George Hall, Thomas J. Sargent and Bálint Szőke. (Forthcoming in the Quarterly Journal of Economics Feb 2025 Issue.)
Link to latest version (January 2024). Slides.
Link to extended older version (July 2022).
Abstract: From a new data set, we infer time series of term structures of yields on US federal bonds during the gold standard era from 1791-1933 and use our estimates to reassess historical narratives about how the US expanded its fiscal capacity. We show that US debt carried a default risk premium until the end of the nineteenth century when it started being priced as an alternative safe-asset to UK debt. During the Civil War, investors expected the US to return to a gold standard so the federal government was able to borrow without facing denomination risk. After the introduction of the National Banking System, the slope of the yield curve switched from down to up and the premium on US debt with maturity less than one year disappeared.
Supplementary materials:
Github repository with bond data. This respository contains the data on prices, quantities, and descriptions of bonds and notes issued by the United States Federal government from 1776 to 1960.
Data description (February 2018). This document describes Pandas DataFrames and the spreadsheets underlying them that contain prices, quantities, and descriptions of bonds and notes issued by the United States Federal government from 1776 to 1960. It contains directions to a public github repository at which DataFrames and other files can be downloaded.
with George Hall, Thomas J. Sargent and Bálint Szőke.
Link to latest version (January 2024)
Abstract: US federal governments have confronted trade-offs among lowering borrow- ing costs, maintaining price stability, and maintaining financial stability. During the gold standard era, successive administrations prioritized decreasing government borrowing costs and keeping trend inflation low. Starting with FDR, the government prioritized financial and business cycle stability and was willing to use inflation taxes to lower its debt obligations and redistribute wealth between nominal creditors and debtors. Towards the end of the twentieth century, the government embraced financial deregulation and aggressive inflation targeting. We use our estimates for historical yields and inflation processes to indicate how those changing policy priorities affected or coincided with key macroeconomic correlations. The slope of both the US federal debt yield curve and the “Phillips curve” has changed signs as government priorities have changed.
with George Hall, Thomas J. Sargent and Bálint Szőke.
Link to latest version (August 2023).
Abstract: Estimating 19th century US federal bond yield curves involves challenges because few bonds were traded, bonds had peculiar features, government policies changed often, and there were wars. This paper compares statistical approaches for confronting these difficulties and shows that a dynamic Nelson-Siegel model with stochastic volatility and bond-specific pricing errors does a good job for historical US bond prices. This model is flexible enough to interpolate data across periods in a time-varying way without over-fitting. We exploit new computational techniques to deploy our model and estimate yield curves for US federal debt from 1790-1933.
Time line of major US monetary and financial events: Link to latest version (June 2024).
with Markus Brunnermeier.
Link to latest version (May 2025) Slides
Abstract: This paper studies strategic decision making by a private currency ledger controller facing competition from public money and/or other ledgers. It chooses a markup fee for making payments using its ledger and the settlement asset for credit contracts. A monopoly ledger controller can incentivize contract enforcement across the financial sector by threatening exclusion but can also extract rents through its pricing power. The emergent market structure bundles the provision of ledger and trading technologies. Regulation should balance better contract enforcement against higher ledger rent extraction.
with Joshua Weiss. (Resubmit at Journal of Monetary Economics.)
Link to latest version (April 2024).
Abstract: We introduce a new mechanism that eliminates self-fulfilling runs on a Diamond Dybvig intermediary without requiring deposit insurance. During a run, a depositor can take unliquidated intermediary assets in exchange for closing their account. We show our mechanism would have been useful in the most recent banking crisis. We also show our mechanism improves upon suspension if depositors repeatedly interact with the intermediary, there is aggregate return risk, and/or policy makers have limited commitment.
with Markus Brunnermeier.
Link to latest version (April 2024).
Link to older longer version (August 2023).
Abstract: Technological change has led to increased, but segmented, information collection. Tech platforms record the information about trading histories required for making of uncollateralized loans, whereas banks specialize in making the assessment of collateral quality required for collateralized lending. Current regulation hinders tech platforms from offering financial services while strategic behavior between tech platforms and banks impedes cooperation on information sharing and contract enforcement. We study how the government should design - and how lobbying efforts impact - information portability, which ultimately affects financial market segmentation and financial inclusion.
Link to latest version (December 20 2024).
Abstract: This paper studies the disruption of bank business credit during a financial crisis in a model with optimal long term contracting under agency frictions and a directed search market for bank funding. Banks commit to long term contracts with entrepreneurs but then face heterogeneous shocks to their cost of funds during a crisis. The optimal contract can be implemented using standard debt securities and a ``covenant'' that allows bankers with high funding costs to adjust debt terms once the entrepreneur has accumulated sufficiently many losses. This is consistent with empirical evidence from the recent financial crisis. In general equilibrium, the contracting frictions amplify the crisis by increasing the termination rate of projects and decreasing the financing rate. The model is extended to incorporate project heterogeneity and working capital. The frictions then skew the economy towards lower volatility projects and sub-optimally reduce project size.
Distinguished CESifo Affiliate Award (2019)
CICM Best Junior Scholar Paper Award (2019)
BFI Workshop on deep learning methods: Link to BFI talk (April 2024).
PASC 2024 Conference: Session MS6F - Advances of Deep Learning in Economics: Link to talk (June 2024).
Princeton Initiative 2024: Link to talk (September 2024).
with Zhouzhou Gu, Mathieu Lauriere and Sebastian Merkel.
Link to latest version (December 2024). SSRN. arXiv.
Link to supplementary appendix on the continuous time Krusell & Smith model.
Abstract: We propose and compare new global solution algorithms for continuous time heterogeneous agent economies with aggregate shocks. First, we approximate the agent distribution so that equilibrium in the economy can be characterized by a high, but finite, dimensional non-linear partial differential equation. We consider different approximations: discretizing the number of agents, discretizing the agent state variables, and projecting the distribution onto a finite set of basis functions. Second, we represent the value function using a neural network and train it to solve the differential equation using deep learning tools. We refer to the solution as an Economic Model Informed Neural Network (EMINN). The main advantage of this technique is that it allows us to find global solutions to high dimensional, non-linear problems. We demonstrate our algorithm by solving important models in the macroeconomics and spatial literatures (e.g. Krusell and Smith (1998), Khan and Thomas (2007), Bilal (2023)).
with Adam Rebei and Yucheng Yang.
Link to latest version (February 2025). SSRN.
Abstract: We develop a new method to globally solve and estimate search and matching models with aggregate shocks and heterogeneous agents. We characterize general equilibrium as a high-dimensional partial differential equation with the distribution as a state variable. We then use deep learning to solve the model and estimate economic parameters using the simulated method of moments. This allows us to study a wide class of search markets where the distribution affects agent decisions and compute variables (e.g. wages and prices) that were previously unattainable. In applications to labor search models, we show that distribution feedback plays an important role in amplification and that positive assortative matching weakens in prolonged expansions, disproportionately benefiting low-wage workers.
with Goutham Gopalakrishna and Zhouzhou Gu.
New version coming soon.
Abstract: How does the regulation of the financial sector impact household welfare and inequality? To answer this question, we build a macroeconomic model with banks, insurance/pension funds, heterogeneous households, asset market participation constraints, and endogenous asset price volatility. We develop a new deep learning methodology for characterizing global solutions to this class of macro-finance models. We show how asset price dynamics, financial stability, and wealth inequality all depend upon which investors are able to purchase assets in bad states of the world. In particular, allowing pension/insurance funds broad access to asset markets leads to greater stability at the business cycle frequency but exposes the household to other risks. Ultimately, the government faces complicated trade-offs between ensuring stability, lowering borrowing costs, and maintaining household equality.
with Yinshan Shang.
New draft coming soon.
We examine the role of green finance in an economy with both political and financial frictions. Investors with heterogeneous preferences for carbon reduction (``green'' and ``brown'' investors) voluntarily contribute to environmentally-focused financial intermediaries (``green'' funds) and also participate in political bargaining over environmental policy. Firms with heterogeneous production technologies choose carbon emission and output levels to maximize their market value. Green funds increase the transparency of firm emission levels, improve coordination amongst investors, and ultimately segment investor portfolios along their environmental preferences. This can increase demand for green stocks and so may incentive firms to reduce carbon emissions. However, it also potentially reduces political cooperation because green investors have a better outside option and brown investors are more exposed to environmental regulation. Risk in the political bargaining process leads to complex hedging behaviour by investors in the green finance market.
with Toru Kitagawa, José Luis Montiel Olea, and Amilcar Velez. Journal of Econometrics. (2020). Vol. 217, Iss. 1, pp. 161-175.
Abstract: This paper examines the asymptotic behavior of the posterior distribution of a possibly nondifferentiable function g(θ), where θ is a finite-dimensional pa- rameter of either a parametric or semiparametric model. The main assumption is that the distribution of a suitable estimator θn, its bootstrap approximation, and the Bayesian posterior for θ all agree asymptotically. It is shown that whenever g is Lipschitz, though not necessarily differentiable, the posterior distribution of g(θ) and the bootstrap distribution of g(θn) coincide asymptotically. One implication is that Bayesians can interpret bootstrap inference for g(θ) as approximately valid posterior inference in a large sample. Another implication—built on known results about bootstrap inconsistency—is that credible sets for a nondifferentiable parameter g(θ) cannot be presumed to be approximately valid confidence sets (even when this relation holds true for θ).
with Lawrence Uren. Journal of Money, Credit, and Banking. (2014). Vol. 46, Iss. 2-3, pp. 347-370.
Abstract: We use a standard New Keynesian model of a small open economy, extended to include a government sector, to investigate the Great Depression in Australia. A calibrated model with a fixed exchange rate regime, similar to the gold standard, does well in replicating the dynamics of output during the interwar period. We then ask to what extent shocks to the economy would have been moderated by adopting modern-day policies. We find that if policymakers had adopted a flexible exchange rate with a Taylor rule policy that output fluctuations during the Great Depression would have been moderated by up to 25%. Changes in government fiscal policy would also have moderated output fluctuations, but by a slightly smaller amount. Overall, we find that improved policy could have reduced output fluctuations by almost 50%.
with Markus Brunnermeier. In Niepelt, Dirk (24 November 2021). Central Bank Digital Currency: Considerations, Projects, Outlook. VoxEU.
with George Hall, Thomas J. Sargent, and Bálint Szőke.
Link to latest version (February 2018).
Abstract: This document describes Pandas DataFrames and the spreadsheets underlying them that contain prices, quantities, and descriptions of bonds and notes issued by the United States Federal government from 1776 to 1960. It contains directions to a public github repository at which DataFrames and other files can be downloaded.
NBER Asset Pricing (2022). Discussion of “Exorbitant Privilege Gained and Lost: Fiscal Implications” by Chen, Jiang, Lustig, Van-Nieuwerburgh, Xiaolan. slides
UCLA Fink Center Conference on Financial Markets (2023). Discussion of “Exorbitant Privilege Gained and Lost: Fiscal Implications” by Chen, Jiang, Lustig, Van-Nieuwerburgh, Xiaolan. slides
CBDC Webinar (2023). Discussion of “The Demand For Programmable Payments” by Kahn and van Oordt. slides
NBER Summer Institute (2023). Discussion of “A Theory of Payments-Chain Crises” by Saki Bigio. slides
YPFS Fighting a Financial Crisis Conference (2024). Discussion of "Two Centuries of Systemic Bank Runs" by Rustam Jamilov, Tobias Konig, Karsten Muller, and Farzad Saidi. slides
EFA (2024). Discussion of "Admissible Surplus Dynamics and the Government Debt Puzzle” by Pierre Collin-Dufresne Julien Hugonnier Elena Perazzi. slides