Research
In my research, I use structural models and causal inference to study financial markets and household finance. I also use these methods to study AI adoption and evaluate how AI systems behave in high-stakes domains.
Economics & Finance ▼
Publications
Refinancing Cross-Subsidies in the Mortgage Market
with Jack Fisher, Alessandro Gavazza, Lu Liu & Tarun Ramadorai
Journal of Financial Economics, 158:103876 (2024)
🏆 Jensen First Prize: Best Corporate Finance Paper, Journal of Financial Economics 🏆
JFE Journal Version. Bank of England. CEPR.
Abstract & links ▼
In household finance markets, inactive households can implicitly cross-subsidize active households who promptly respond to financial incentives. We assess the magnitude and distribution of cross-subsidies in the mortgage market. To do so, we build a model of household mortgage refinancing and structurally estimate it on rich administrative data on the stock of outstanding mortgages in the UK. We estimate sizeable cross-subsidies from relatively poorer households and those located in less-wealthy areas towards richer households and those located in wealthier areas. Our work highlights how the design of household finance markets can contribute to wealth inequality.
Macroprudential Policy, Mortgage Cycles and Distributional Effects: Evidence from the UK
with José-Luis Peydró, Francesc Rodríguez-Tous & Arzu Uluc
Review of Financial Studies, 37(3), 727-760 (2024)
RFS Journal Version. Bank of England. BSE. CEPR.
Abstract & links ▼
We analyze the distributional effects of macroprudential policy on mortgage cycles by exploiting the U.K. mortgage register and a 2014 15% limit imposed on lenders' high loan-to-income (LTI) mortgages. Constrained lenders issue fewer and more expensive high-LTI mortgages, with stronger effects on low-income borrowers. Unconstrained lenders strongly substitute high-LTI loans in local areas with higher constrained lender presence, but not high-LTI loans to low-income borrowers—consistent with adverse selection problems—implying lower overall credit to low-income borrowers. Consistently, policy-affected areas experience lower house price growth postregulation and, following the Brexit referendum (negative aggregate shock), better house price growth and lower mortgage defaults for low-income borrowers.
Previous version circulated as Macroprudential policy and the housing market: Evidence from the UK .
Cross-Border Effects of Regulatory Spillovers – Evidence from Mexico
Journal of International Economics, 126:103350 (2020)
JIE Journal Version. Bank of England.
📰 Media: CentralBanking.com
Abstract & links ▼
This paper studies the spillover of a macroprudential regulation in Spain to the Mexican financial system via Mexican subsidiaries of Spanish banks. The spillover caused a drop in the supply of household credit in Mexico. Municipalities with a higher exposure to Spanish subsidiaries experienced a larger contraction in household credit. These localized contractions caused a drop in macroeconomic activity in the local non-tradable sector. Estimates of the elasticity of loan demand by the non-tradable sector to changes in household credit supply range from 1.2–1.8. These results emphasize cross-border effects of regulations in the presence of global banks.
Handbook Chapters
Distributional Consequences of Borrower-Based Macroprudential Tools
with José-Luis Peydró, Francesc Rodríguez-Tous & Arzu Uluc
Chapter in Research Handbook of Macroprudential Policy. Bank of England.
Abstract ▼
This paper provides an overview of evidence from a range of country-specific studies on the effectiveness of borrower-based macroprudential tools which limit household leverage at the borrower level. Most studies find these measures effective in breaking the self-enforcing loop between household leverage and house prices. These measures have beneficial effects in terms of lower defaults and less volatile house price dynamics during periods of economic distress. Their effects are heterogeneous across borrower types, with stronger impacts where leverage requirements are higher, such as among first-time buyers. Studies point to restrictions on household leverage having downstream effects on job search, location choice, homeownership, and exposure to income shocks. Looking ahead, further research is required to conduct a comprehensive cost-benefit analysis of these measures, to adapt them to increased use of technology in financial intermediation, and to examine their broader societal effects, including political outcomes and mental health.Working Papers
Behavioral Lock-In: Housing-Market Taxation with Reference-Dependent Agents
with Cristian Badarinza, Tarun Ramadorai & Juhana Siljander
📰 Media: Financial Times
Abstract ▼
We study the aggregate implications of reference dependent and loss averse preferences in the housing market. Motivated by micro evidence, we embed optimizing homeowners with these preferences into a dynamic search and matching equilibrium model with rich heterogeneity and realistic constraints. We assess the model using large and granular administrative data tracking buyers and sellers in the UK housing market; the predictions match regional and time variation in price growth and transaction volumes. The model shows that behavioral frictions in a decentralized market can link nominal quantities with real outcomes; and reveals that the distribution of potential nominal gains in the housing market is a key policy-relevant statistic.Interest-rate Fee Substitution: Credit Facilitation in Segmented Markets
with João Cocco & S. Lakshmi Naaraayanan
Abstract ▼
We study the aggregate implications of reference dependent and loss averse preferences in the housing market. Motivated by micro evidence, we embed optimizing homeowners with these preferences into a dynamic search and matching equilibrium model with rich heterogeneity and realistic constraints. We assess the model using large and granular administrative data tracking buyers and sellers in the UK housing market; the predictions match regional and time variation in price growth and transaction volumes. The model shows that behavioral frictions in a decentralized market can link nominal quantities with real outcomes; and reveals that the distribution of potential nominal gains in the housing market is a key policy-relevant statistic.Housing-Consumption Channel of Mortgage Demand
with Gabriel Ahlfeldt & Nikodem Szumilo
Abstract ▼
We quantify the consumption channel in mortgage demand according to which households borrow more in response to house price increases since housing and non-housing consumption are imperfect substitutes. To identify the consumption channel, we take a structural approach to mortgage demand and supply, exploiting exogenous variation in house price growth and a unique matched house price-mortgage data set. We feed our estimate of the elasticity of mortgage borrowing with respect to house prices of 0.82 into a counterfactual analysis of the general equilibrium of housing and mortgage markets. We find that without the consumption channel, mortgage borrowing in the UK would have increased by over 50% less since the 1990s and house price growth would have been lower by 31%.Housing Booms and Credit Misallocation
📰 Media: CentralBanking.com
Abstract ▼
This paper studies the effect of asset bubbles on economic growth in the presence of financial constraints and heterogeneous projects. I consider an economy with two sectors which differ in their productivity and level of financial constraints. In this framework, asset bubbles in credit markets raise interest rates and lower aggregate investment productivity by directing financial resources away from the sector with higher productivity. Further, they may lead to steady states with low investment productivity and welfare, creating bubbly growth-traps.Selected Work in Progress
Star Charities, Director Networks and Firm Performance
with Saleem Bahaj, Andrew Blake, Angus Foulis, Andy Haldane & Gábor Pintér — draft available on request
Abstract ▼
We use a unique dataset on the population of UK company directors and their firms to study a novel, unexplored territory for social interactions among firm managers: charities. The most central charities comprise of highly skilled directors, and firms associated with these directors manage up to 60% of total assets in the UK economy. We show that a firm whose director joins one of these star charities subsequently experiences faster growth and increased employment. These effects are stronger for younger and smaller firms, and for firms with younger directors. The key mechanism points to skill diffusion: joining a star charity improves the quality of the director's network of co-directors. Our results indicate a pecuniary benefit from associations with star charities via increased social capital.AI Research ▼
Research applying causal inference methodology to AI model internals and evaluation.
Working Papers
Fair Outputs, Biased Internals: Causal Potency and Asymmetry of Latent Bias in LLMs for High-Stakes Decisions
causal inference mechanistic interpretability LLM evaluation
with Marcus Buckmann
Pre-print: arXiv:2605.15217
Work in Progress
Measuring the Effect of AI Adoption on Firm Productivity using Microdata
causal inference AI adoption firm productivity