Three Essays In Applied Econometrics

Abstract

In this thesis, I use both structural- and reduced-form econometric models to analyze economic issues relating to the mortgage and labour markets. In chapter 2, I use a dynamic discrete choice model to quantify the impact of the Home Affordable Refinance Program (HARP) in the United States during the Great Recession, on mortgage default rates, as well as its monetary benefit to the government. I find that HARP led to a 71.6% reduction in the total number of defaults and has saved the government US$ 800 million. Furthermore, I also use the structural model to evaluate HARP under a counterfactual scenario of prolonged depression of house prices instead of the V-shaped recovery of house values we observed after the crisis. In chapter 3, I use a differences-in-differences (DiD) approach to establish the causal link between securitization and screening standards in the U.S. mortgage market. The estimation strategy exploits the discrete set of policy changes in the conforming loan limits that set the maximum loan size (conforming loans) that government-sponsored enterprises (GSEs) can purchase and securitize. I find that GSEs' purchase eligibility has increased the probability of loan origination by $\sim3%$. Chapter 4, co-authored with Jason Adams and David Ros'{e}, investigates the efficacy of the Ontario Works' (OW) active labour market programs (ALMPs) in helping welfare recipients find and retain employment using an instrumental variables (IV) approach. We find that structured job search activities and other skills-enhancing ALMPs reduce spells by between 1.6 and 3.1 months, whereas programs such as placement services increase spell lengths considerably, but lead to lower return rates to OW one and two years after the end of their welfare spells.