Three facts about the 1999–2019 U.S. housing cycle present a puzzle: during the boom, income growth and mortgage growth correlate negatively across ZIP codes within metros but positively across metros, and metros with the worst busts recover fastest. I develop a unified credit expansion theory explaining within- and cross-metro patterns across the prior, boom, bust, and recovery periods, with new “double differences” hypotheses. Using a gravity-model instrumental variable, I show credit expansion in private-label mortgages, rather than government-sponsored enterprise mortgages, drives the boom– bust across metropolitan areas. This view, largely independent of speculation, explains the puzzle alongside long-run local fundamentals.

Presentation: ASU PhD Seminar, ASU Browbag,  2024 Eastern Finance Association, Georgia Institute of Technology, 2024 Econ Graduate Students' Conference at WashU, 2025 Financial Management Association Annual Meeting, 2025 Southern Finance Association Annual Meeting, 2026 Southwestern Finance Association Annual Meeting, 2026 Econometric Society North American Summer Meeting. 

Empirical business cycle studies using cross-country data usually cannot achieve causal relationships, while within-country studies mostly focus on the bust period. We provide the first causal investigation into the boom period of the 1999-2010 U.S. cross-metropolitan business cycle. Using a novel research design, we show that credit expansion in private-label (non-jumbo) mortgages causes a more pronounced boom (2000-2006) and bust (2007-2010) cycle in house-related industries in high-net-export-growth metros than in low-net-export-growth metros. Thus, our results are consistent with the credit-driven household demand hypothesis. Most importantly, our unique research design enables us to conduct the most comprehensive tests on theories (hypotheses) regarding the business cycle. We show that the following theories (hypotheses) cannot explain the 1999-2010 U.S. business cycle: the speculative euphoria hypothesis, the real business cycle theory, the collateral-driven credit cycle theory, the business uncertainty theory, and the extrapolative expectation theory. 

Presentation: 2026 American Economic Association Annual Meeting

This paper provides the first causal evidence that credit supply expansion caused the 1999-2010 U.S. business cycle mainly through the channel of household leverage (debt-to-income ratio). Specifically, induced by net export growth, credit expansion in private-label mortgages, rather than in government-sponsored enterprise mortgages, causes a more pronounced boom (1999-2005) and bust (2008-2014) cycle in household leverage in high-net-export-growth metropolitan areas than in low-net-export-growth ones. In addition, a stronger household leverage cycle creates a stronger boom-and-bust cycle in the local economy, including housing prices, residential construction investment, and house-related employment. Thus, our results are consistent with the credit-driven household demand channel (Mian and Sufi, 2018). Furthermore, we present multiple pieces of evidence against the corporate channel, as emphasized by other business cycle theories (hypotheses). 

Presentation: ASU Ph.D. Seminar, Georgia Institute of Technology, 2024 Financial Management Association Annual Meeting

This paper establishes the causality between export growth and increased corporate innovation in quantity and quality in US public firms by using a novel gravity model-based instrument from international economics. Strong empirical evidence shows up in various measures of innovation, including the number of patents, number of citations, scaled number of citations, and average number of citations. In addition, this paper uncovers three mechanisms through which export increases firm innovation: (1) export growth increases sales and Research & Development; (2) export growth induces experienced inventors to reallocate towards industries with more export growth and new-generation inventors to start their careers in these industries; (3) export growth also increase institutional ownership percentage and concentration in these industries, which have been shown to increase innovation via reducing career risk faced by managers. 

Traditional business cycle theories focus on either (1) the banks’ amplification role in the business downturn or (2) the banks’ passive role in the business cycle via the corporate investment cycle. This paper shows two crucial departures. First, I demonstrate that banks proactively expand excessive credit through mortgages, credit cards, auto loans, and securitization, creating an economic boom. Second, I show that banks’ proactive credit expansion for household consumption, rather than for corporate investment, contributed to the Great Recession. To achieve causal evidence, I use an IV from international economics that explores (1) exogenous timing of the invention of the Copula formula (the key technology in financial securitization), and (2) geographic differences in exposure to net export growth. Consistent with my new credit expansion theory, I show that after the invention of the Copula formula, banks in metropolitan areas with higher net export growth proactively expand more credits, leading to booms in household consumption and related housing and nontradable sectors. Ultimately, the bust is more severe in these metropolitan areas, as evidenced by bank losses, household bankruptcies, and contractions in the housing and nontradable sectors.