Economic models often assume that agents always know the market value of their assets. We use residential property tax assessment as a laboratory to test this assumption for housing. We first show that assessed market value (AMV) is a noisy proxy for transaction-based market values (TMV). Innovations in AMV are less volatile, are weakly correlated with, and lag innovations in TMV. An AMV-based, national-level house price index has shallower troughs and shorter peaks than its TMV-based counterpart. We merge in credit bureau data to test whether homeowners use AMVs, as signals of housing wealth, to make consumption decisions. Using local mass reassessments as an instrument, we find that AMV changes causally effect the likelihood that households take out a new home equity line of credit (HELOC) with a similar economic magnitude as TMV changes. A partial equilibrium calibration exercise suggests that innovations in AMV can explain approximately 1% of annual HELOC origination. Overall, our results suggest that homeowners do not fully know the value of their homes.

This paper studies the May 2023 change in the conforming mortgage upfront guarantee fee schedule. Consistent with incomplete pass-through, lenders raise rejection rates and sell fewer loans to the GSEs when fees rise. For small-dollar mortgages (SDM), pass-through is near zero and rejection rates are more sensitive to fee increases. This implies that the overall incomplete pass-through is partly driven by liquidity-constrained borrowers and that the inequality in mortgage access via higher rejection rates on SDMs is partly driven by lenders' inability to pass costs onto SDM borrowers. Without offsetting effects from fee cuts, fee hikes reduced aggregate mortgage origination by 1.4%.

This paper compares the careers of Ivy League athletes to those of their non-athlete classmates. Combining team-level information on all Ivy League athletes from 1970 to 2021 with resume data for all Ivy League graduates, we examine both post-graduate education and career choices as well as career outcomes. In terms of industry choice, athletes are far more likely to go into business and Finance related jobs than their non-athlete classmates. In terms of advanced degrees, Ivy League athletes are more likely to get an MBA and to receive it from an elite program, although they are less likely to pursue an M.D., a Ph.D., or an advanced STEM degree. In terms of career outcomes, we find that Ivy League athletes outperform their non-athlete counterparts in the labor market. Athletes attain higher terminal wages and earn cumulatively more than non-athletes over the course of their careers controlling for school, graduation year, major, and first job. In addition, they attain more senior positions in the organizations they join. We also find that athletes from more socioeconomically diverse sports teams and from teams that have lower academic admissions thresholds have higher career outcomes than non-athletes. Collectively, our results suggest that non-academic human capital developed through athletic participation is valued in the labor market and may support the role that prior athletic achievement plays in admissions at elite colleges.   

In 2021, the U.S. Treasury instituted hard caps to reduce Government-Sponsored Enterprise exposure to second-home and investment-property mortgages, leading to declines in their purchase of affected mortgages. The policy lowered credit supply to affected housing investors, with higher interest rates and lower originations. Bank and non-bank lenders display a similar supply response, suggesting that deposits offer no advantage in mortgage lending. Lenders adjust at the portfolio level by reallocating mortgage credit across local markets, suggesting that they manage credit provision market-by-market. Rental housing supply and condominium prices decline while rents increase, consistent with higher financing costs in the rental market from the policy.

This paper uses a large data set of single-family mortgage applications to study the relationship between borrower's age and application outcomes. Conditional on a rich set of observable characteristics, borrower's age is positively correlated with rejection probability and coupon rate. The result holds for young borrowers, implying that age, not just old age, is an important determinant of mortgage application outcomes. Survey data and indirect empirical tests suggest that the rejection probability results are partly driven by age-related mortality risk, the coupon rate results are partly driven by differences in shopping behavior, and taste-based age discrimination is unlikely to contribute. Policy implications are discussed. 

This paper proposes a non-pecuniary measure of career achievement, seniority. Based on a database of over 130 million resumes, this metric exploits the variation in how long it takes to attain job titles. When non-monetary factors influence career choice, assessing career attainment via non-wage measures, such as seniority, has significant advantages. Accordingly, we use our seniority measure to study labor market outcomes of VC-backed entrepreneurs. Would-be founders experience accelerated career trajectories prior to founding, significantly outperforming graduates from same-tier colleges with similar first jobs. After exiting their start-ups, they obtain jobs about three years more senior than their peers who hold (i) same-tier college degrees, (ii) similar first jobs, and (iii) similar jobs immediately prior to founding their company. Even failed founders find jobs with higher seniority than those attained by their non-founder peers. 

Among owner-occupied single-family homes in the United States that enjoy the same set of property tax-funded amenities and pay the same statutory property tax rate, owners of inexpensive houses pay almost 50% higher effective tax rates than owners of expensive houses because tax assessments are regressive with respect to house price. Using a near-national data set, I provide empirical evidence that assessment regressivity is partially caused by valuation methods that do not fully incorporate important pricing information such as neighborhood characteristics. Systematic sorting across neighborhoods implies that economically disadvantaged households disproportionately bear the cost of this valuation problem. The findings potentially have important implications for wealth inequality and other forms of ad valorem tax.