The Bankruptcy Project

This document briefly describes some of my research idea around the institution of bankruptcy, consumer and corporate.

1. Measuring Bankruptcy Judge Ideology with Application to Bankruptcy Outcome
   1. This project has two parts:
      1. A NLP component that measures judicial ideology from text, applied to bankruptcy judges
      2. A causal inference component that estimates the effect of ideology
         1. Rule of law or rule of judges?
         2. Need to review: Manipulating Random Assignment: Evidence From Consumer Bankruptcies in the Nation’s Largest Cities, Morrison, Pang, and Zytnick
         3. A heterogenous treatment effect analysis
            1. Conditional treatment estimation from Athey et al
               1. Generalized Random Forests, Athey, Tibshirani, and Wage
            2. A selection into chapter 13 approach
               1. The Chapter 13 and not filing at all choice is not observable in the data – this will probably be a robustness check
               2. An Interval Estimation Approach to Sample Selection Bias, ArXiv preprint, biostatisticians provides bounds based on selection
   1. Additional literature review: the CS literature
   2. Need to figure out more about the mechanics about judge assignment and chapter choice
2. Bankruptcy Judges and the Credit Market
   1. Judges with different ideologies/characteristics plausibly enforce bankruptcy standards differently
   2. Under different judges, bankruptcy applicants faces different probability of receiving bankruptcy protection. Are these differences there? If the differences are there, are they priced in?
   3. On the consumer side, we have loan level data from freddie mac. On the corporate side, we can use data from compustat.
      1. Similar papers:
         1. Bankruptcy Law and the Cost of Credit: The Impact of Cramdown on Mortgage Interest Rates, JLE 2014, Goodman and Levitin
            1. Loan level data, Diff-in-Diff using a 1993 supreme court decision
            2. We can do staggered Diff-in-Diff use a series of events when a conservative judge is elected to the local bankruptcy court
         2. Busy Bankruptcy Courts and the Cost of Credit, working paper, Müller
            1. Compustat data, Diff-in-Diff using a 2005 a drop in consumer caseload induced by BAPCPA
      2. Need to do literature review
         1. What drives local credit market?
         2. Connection to the marco-economics
3. British Bankruptcy Reforms and the Credit Market
   1. There was a series of bankruptcy reform in British history, for example, 1883 and 1887
   2. There might be cross country comparison and, if we have loan level data, a micro study.
   3. This could be horribly confounded.
4. Bankruptcy Outcome and Voting Behavior