I am very happy Bob Lawless invited me to be a guest blogger on Credit Slips for the next few days. I have been reading Credit Slips for a long time and I really enjoy it.
In the next few days I would like to talk about some of my economics based research involving consumer bankruptcy in Canada. What I try to do in my research is to exploit some of the data available in the Canadian system to answer questions that may be harder to answer using US based data.
For example, in one of my recent papers I try to provide new evidence on a longstanding debate in the bankruptcy literature by exploiting a seemingly innocuous Canadian institution – the Canadian post code system.
The question involves the impact that your neighbors declaring bankruptcy have on your declaring bankruptcy. There are a variety of possible channels through which neighbors can influence individuals in the bankruptcy context. It is possible that neighbor’s behavior can influence perceptions of the social stigma of bankruptcy. If many of your neighbors are filing for bankruptcy, then you may feel that there is less stigma involved, which could encourage you to file yourself (“because everybody else is doing it”). Alternatively, it may be possible that by talking to your neighbors who have previously filed, you can learn about the process involved in bankruptcy (“look how easy bankruptcy is”), which may also encourage you to file.
Two very famous papers in the economics literature on bankruptcy Fay, Hurst and White (2002) and Gross and Souleles (2002) have tried to examine this question empirically by examining the impact of the number of bankruptcies in an individual’s “neighborhood” on the probability that the individual will file. An important concern with both these papers, however, is that because of US data limitations, they are forced to define “neighborhoods” as either US states or US bankruptcy court districts, both of which can contain many millions of individuals.
This is where the Canadian post code system comes in. In my research, I use data provided to me by the Canadian bankruptcy regulator, the Office of the Superintendent of Bankruptcy (OSB), which allows me to observe the Canadian postal code of every bankruptcy filing in Canada (I will talk more about the OSB this week).
Whether by accident or design, Canadian post codes are extremely tiny compared to US Zip Codes (and certainly compared to US States or bankruptcy court districts). On average there are only 13 households in each Canadian Post Code area, and often these areas take up less than a city block in terms of space. In other words, Canadian post code areas really can be thought of as “neighborhoods”.
The punch line of my study is that there is indeed a significant impact from the past bankruptcies of neighbors (as defined by the very small Canadian Post Codes) to the probability that an individual in the neighborhood will file. In other words, my findings support the neighborhood spillover findings of Fay, Hurst and White (2002) and Gross and Souleles (2002), but unlike these papers, I am able to really capture the idea of a “neighborhood”, rather than rely on less than satisfactory proxies for “neighborhood” such as US States or US bankruptcy court districts.
Another interesting element of this study is that my data, which is taken from individual credit card accounts, allows me to compare and contrast two very different kinds of default – bankruptcy and credit card charge-off. While both bankruptcy and credit card charge-off involve default, (the credit card contract is terminated), they have quite different costs and benefits.
In brief, bankruptcy allows for unsecured (e.g. credit card) debt to be discharged, but bankruptcy is publically disclosed through the legal system. On the other hand, credit card charge-off is not publically disclosed (thus affording more privacy), but at the cost of not having the credit card debt discharged.
These institutional differences allow me to generate new hypotheses about which of bankruptcy or charge-off defaulters will choose, based on the characteristics of the neighborhood they live in. I propose, and provide evidence for, the hypothesis that if a defaulter lives in a neighborhood with a large number of previous bankruptcies among the neighbors, then that individual will choose to default via bankruptcy rather than charge-off. This is because more neighborhood bankruptcies will lower stigma or provide more information about the process of bankruptcy.
On the other hand, I show that defaulters who live in low bankruptcy neighborhoods choose to default via charge-off rather than bankruptcy. This is consistent with the argument that low bankruptcy neighborhoods have higher levels of bankruptcy stigma, thus individual defaulters choose to default via charge-off in order to maintain more privacy about their default.