Assistant professor of finance at the University of Colorado Boulder (Leeds School of Business)
Visiting scholar at the Federal Reserve Banks of St. Louis and Philadelphia
Research interests: housing markets, household finance, insurance policy, mutual funds
Recently Published Articles
Medicaid and Household Savings Behavior: New Evidence from Tax Refunds with Jorge Sabat, Radhakrishnan Gopalan & Michal Grinstein-Weiss. Journal of Financial Economics (forthcoming, 2019)
Transparency, Investor Information Acquisition, and Money Market Fund Rebalancing during the 2011-12 Eurozone Crisis with Lawrence Schmidt, Allan Timmermann, & Russ Wermers. Review of Financial Studies (forthcoming, 2019)
The Effects of Health Insurance on Home Payment Delinquency: Evidence from the ACA Marketplace Subsidies with Radhakrishnan Gopalan & Michal Grinstein-Weiss. Journal of Public Economics, Volume 172, April 2019, Pages 67-83
with Stephen Billings and Lowell Ricketts *Updated October 2019
Hurricane Harvey submerged over 25% of Houston in August 2017. Using a treatment intensity difference-in-difference design on administrative credit data, we find that average treatment effects of flooding on household debt outcomes mask substantial underlying heterogeneity based on initial financial well-being, homeownership, and floodplain (i.e., flood insurance) status. Negative financial outcomes are concentrated among residents who entered the hurricane in a weaker financial position. For example, the average bankruptcy rate in high ownership areas that heavily flooded and have many financially constrained residents rose by 1.1 percentage points (or 30%) after Harvey relative to similar areas that did not flood. Being in a floodplain and, hence, having a higher likelihood of flood insurance, largely mitigates these negative effects. Using individual FEMA registrant and SBA loan data, we present evidence that our results may be explained, in part, by inequalities in access to federal disaster assistance and loans.
with Jorge Sabat
The rules of thumb offered by financial advisors regarding how much to hold in liquid reserves vary widely and usually imply far greater sums than low-income households save. This paper seeks empirically-grounded insights into the minimum liquidity buffer needed by the average low-income household. First, we document diminishing benefits to liquid savings in terms of the likelihood of experiencing financial hardship. Then, we formalize this relationship with a theory of poverty traps. Finally, to observed data, we fit a regression kink model with an unknown threshold (kink) point that must be estimated. Our key finding is that the threshold point is $2,467 with a 95% confidence interval of $1,814–$3,011 (in 2019 dollars) or roughly 1 month of income for the average low-income household – which is far less than the savings amounts implied by common rules of thumb (typically 3–6 months of income). Theoretical evidence suggests that financial advice based on an empirically-estimated threshold point is welfare enhancing for households with naive perceptions of their probability of experiencing financial problems.
with Stephen Roll, Michal Grinstein-Weiss, and Cynthia Cryder *Under review
This paper presents the results of an experiment testing the roles of a savings pre-commitment and different savings-focused choice architectures on the savings deposit decisions of 845,786 low- and moderate-income (LMI) tax filers. Results suggest that pre-committing to save at the start of the tax filing process can, among certain populations, dramatically increase savings rates. Among early tax filers, pre-commitment is associated with a 20.6 percentage point increase in savings deposits and a $418.86 increase in the amount deposited to savings. We observe more modest effects of pre-commitment on a general sample of tax filers. We also see strong evidence that choice architectures emphasizing savings strongly impact the deposit decisions of tax filers. The experiment also revealed cautionary evidence that the structure of pre-commitment can solidify decisions, making it then harder to later nudge those who opt-out of savings to change their minds. These findings may be broadly applicable to settings beyond the tax time moment -- such as to applications that seek to encourage particular behaviors (like work or exercise) on the part of its participants.
Health Insurance as an Income Stabilizer
with Nathan Blascak, Stephen P. Roll, and Michal Grinstein-Weiss *Under review
We evaluate the effect of health insurance on the incidence of negative income shocks using the tax data and survey responses of nearly 14,000 low income households. Using a regression discontinuity (RD) design and variation in the cost of nongroup private health insurance under the Affordable Care Act, we find that eligibility for subsidized Marketplace insurance is associated with a 16% and 9% decline in the rates of unexpected job loss and income loss, respectively. Effects are concentrated among households with past health costs and exist only for “unexpected” forms of earnings variation, suggesting a health-productivity link. Calculations based on our fuzzy RD estimate imply a $256 to $476 per year welfare benefit of health insurance in terms of reduced exposure to job loss.
Research in progress
Thirsty for Credit: Mortgage Lending During the Flint Water Crisis
with Nathan Blascak *Paper pending
For a property to qualify for a government-insured loan, it must meet certain safety standards, including having a continuous supply of potable water. While the intention is to protect homebuyers and reduce default, we ask whether these rules may have unintended consequences when violations become geographically concentrated. The Flint water crisis, which started in April 2014, offers a setting to test this hypothesis. We document a 20% increase in the rate of loan denials on purchase mortgages in Flint relative to border tracts coupled with a 30% decline in home prices. We tie a substantial portion of the decline in house prices to the denial rules. Our evidence suggests that the credit freeze was driven primarily by the water quality rules imposed by government mortgage insurers rather than by a decline in the credit quality of Flint borrowers or by lender perceptions of elevated default risk in Flint.
Get Out While the Getting's Good? A Test of First-Mover Behavior in Bond Funds
with Sean Collins and Xiaowen Hu *Paper pending
This paper addresses the concern that illiquidity in corporate bonds may generate a first-mover advantage among investors in bond funds, which could potentially magnify outflows from such funds, and thus also market shocks. Our key insight is that first-movers, if present in bond funds, should impose costs that significantly disadvantage remaining investors, particularly in less liquid categories of bond funds. In other words, there should be a “cost of remaining invested” (CRI) that is large enough to incent investors to redeem in the first place. We estimate CRI using daily data from 2010 to 2017. Although the data encompass several periods of turbulent bond markets, our results provide little evidence of a significant incentive for first movers to redeem. In aggregate, CRI is typically very small, less than one basis point on a daily basis, even for funds that hold less liquid bonds. Using a residual resampling bootstrap procedure, we develop a hypothesis test for the presence of first movers in funds. Only a handful of funds exhibit first-mover activity beyond that which would be expected by random chance. Since these funds are small, the amount of CRI that can be attributed to first-movers is economically inconsequential.