Publications in Refereed Journals
Who benefits the most? Risk pooling in Mortgage Loan Insurance: Evidence from the Canadian Mortgage Market (with Babak Mahmoudi and Chenggang Zhou), Real Estate Economics, August 2022, https://doi.org/10.1111/1540-6229.12405
Abstract: This paper evaluates the effect of mortgage loan insurance (MLI), an essential macroprudential tool available to policymakers, on housing affordability, household leverage, and the overall welfare of the economy. A dynamic model of the housing market with heterogeneous households and competitive housing and mortgage markets is constructed and is calibrated to Canadian data. Households are in finitely-lived and differ in income and cash-on-hand. They optimally choose to rent or own, housing services, and mortgage debt, in addition to consumption and savings. We find that relaxing the mandatory nature of MLI required for mortgages with a loan-to-value ratio of 80% or more dampens demand for housing to purchase and puts downward pressure on house prices. Some of the households with low income and low asset holdings can no longer afford a house; therefore, the aggregate homeownership rate drops. In contrast, demand for rental units increases, and rents go up. Compared to a baseline calibration with mandatory MLI, the Canadian housing nance system would feature lower leverage among households and significantly lower default rates.
Possible Income Misstatement on Mortgage Loan Applications: Evidence from the Canadian Housing Market (with Babak Mahmoudi), Real Estate Economics, January 2020, DOI: 10.1111/1540-6229.12310, pp.1-19.
Click here for Online Appendix
FMA 2018 semifinalist for the best paper award.
Media: Financial Post
Abstract: We construct a measure of Possible Income Misstatement (PIM) for first-time home-buyers by quantifying the gap between growth in incomes reported on mortgage applications and growth in incomes reported on tax files from 2004 to 2014 in Canada. Using a two-stage least square framework to correct for the endogenous nature of house prices and PIM, we find robust evidence that part of the observed dispersion in PIM is caused by house price variation. This suggests borrowers have a greater incentive to misstate income in high-priced markets. We report evidence that markets with a tendency for income misstatement also had higher default rates.
The Impact of Bankruptcy Reform on Consumer Insolvency Choice (with Jason Allen), Canadian Public Policy, Volume 44 Issue 2, June 2018, pp. 100-111.
Previous Version: Bank of Canada Staff Working Paper 2016-26.
Abstract: We examine the impact of the 2009 amendments to the Canadian Bankruptcy and Insolvency Act on consumer insolvency decisions. The amendments were successful in steering debtors out of Division I debt restructuring and into the more cost-effective Division II debt restructuring. Although total insolvencies remained flat after the amendments, they led to a significant substitution out of bankruptcies and into debt restructuring. The extent of substitution greatly depends on regional and individual circumstances. For example, generous asset exemptions under bankruptcy at the provincial level discourage debt restructuring, whereas home ownership encourages it. Our results show, therefore, that the impact of national bankruptcy policies can have sizably different impacts at the regional level.
Other Publications
Incentives for Rental and Owner-Occupied Housing in Canada, CMHC Research Insight, January 2018.
Working Papers
Asset Purchase Programs during the COVID-19 Pandemic: Evidence for Canada’s Insured Mortgage Purchase Program (with Babak Mahmoudi and Joshua Burslem).
Abstract: In response to the COVID-19 pandemic, the Government of Canada introduced the Insured Mortgage Purchase Program (IMPP) along with other programs. Under this program, the government, through the Canada Mortgage and Housing Corporation (CMHC), has purchased up to $150 billion in insured mortgages. The program was aimed at injecting liquidity into the financial system and providing stable financing for banks and mortgage lenders during the coronavirus crisis. We study the impact of the program and provide evidence of the spillover effect on the other asset classes.
The effects of on-campus living on academic performance and student satisfaction: Evidence from all university students in Toronto (with Shelagh McCartney, Cynthia Holmes, and Ximena Rosenvasser).
Abstract: Many universities provide on-campus housing, but only a minority of the student population can access it. Consequently, an increasing number of students are living in off-campus housing either by themselves or by remaining in their familiar homes, farther away from the university. Using over 700,000 student records from a five-year period covering all the university students of a large metropolitan city, this study examines the benefits and multi-year effects of on-campus housing. This analysis is complemented by single-year, self-reported data on student satisfaction, off-campus living situations, and commuting time. Findings indicate that living on campus has significant positive effects on Grade Point Average, particularly for first-year, lower quartile students, and that familial living situations of commuting students also have a significant influence on students’ university satisfaction.
Mind the Gap: Misreporting of Academic Performance in Student Surveys (with Shelagh McCartney, Cynthia Holmes, and Ximena Rosenvasser).
Abstract: This paper examines the accuracy of self-reported academic performance by comparing NSSE survey responses with administrative grade records from universities in Toronto. Using over one million linked student records, we construct a measure of reporting discrepancy and analyze its magnitude, direction, and predictors. The results suggest that the misreporting is quite common – most students had a discrepancy. The overall average error is 0.45, indicating many of those who do misreport are tend to overreport their academic performance. Grade misreporting is both common and skewed toward inflation and is systematically associated with student characteristics. Quantile regression results show that lower-performing students are significantly more likely to overreport, indicating meaningful variation across the grade distribution. We also examined nonresponse to the NSSE grade question and found that missingness is nonrandom and correlated with gender, academic engagement, and actual grade. These findings raise concerns about the reliability of self-reported academic data and underscore the importance of validating survey responses against institutional records.
Immigrants’ Insolvency Choices.
Abstract: Are insolvent immigrants more likely to restructure their debt or file for bankruptcy? This study investigates the variation in insolvency choices between immigrants and non-immigrants. First, I show the propensity of debt restructuring relative to bankruptcy is higher in neighborhoods with a higher share of immigrants. I confirm my findings using a generalized propensity score with neighborhoods' share of immigrants as a continuous treatment. Next, I establish the link between income uncertainty and insolvency choice. More specifically, I show for borrowers with positive assets, a higher chance of unemployment increases the marginal benefit of a proposal. To the extent that the neighborhood share of immigrants is a proxy for individual immigration status, my analysis highlights, income uncertainty as a channel through which immigrants' insolvency choice is different from non-immigrants and suggests the importance of self-insurance against higher income risk.