Bankruptcy Filings Are Low, But Not Everywhere

Bankruptcy filings have dipped to their lowest rate since 1990, as previously blogged (ignoring anomalous statistical gyrations around the 2005 changes to the bankruptcy law). Over the past twelve months the bankruptcy filing rate per 1,000 persons has been 2.95, which is the first time it has been below 3.0 in almost 25 years. But, the filing rate is not that low everywhere.

Before discussing how the filing rates are different across the country, the usual question I get is why the bankruptcy filing rate has declined so much. The answer to that question is not this post. Rather, see many previous posts like here and here. The short version is that U.S. households are carrying less debt -- nothing more complicated than that. Along those lines see the post at Calculated Risk about the Fed's household debt service ratio being near record lows.

Although bankruptcy filing rates may be at historically low levels nationally, there are considerable differences in different parts of the country. The map above right breaks down by judicial district the filing rate per 1,000 persons over the last twelve months (Oct. 2013 - Sept. 2014). The map groups the districts into five "natural" groups or, in more technical speak, the filings rates by judicial district are grouped according to a k-means clustering where k = 5. The table to the right gives the individual numbers by district. If you are having trouble reading the map or table, your best options are (a) to have younger eyes than me, (b) to squint really hard, or (c) click on them to open a larger version in a pop-up box.

What jumped out to me immediately is where the map highlights how Memphis stands alone. The bankruptcy filing rate in the Western District of Tennessee is 10.82 per 1,000 persons. It so much higher than the next highest district (Middle District of Alabama at 6.83) that I triple-checked the data to make sure it was not just data-entry error. In the cluster analysis, adding the Western District of Tennessee to any other cluster makes that cluster less mathematically alike. It is truly a category of one.

Regional variations within a state or across state borders also show up when looking at filings by judicial district. For example, the Eastern District of Arkansas shares more with its neighbor in Tennessee than the rest of Arkansas. Similarly, one sees patterns across state lines -- the DC metro area, the Chicago-Milwaukee-Gary metroplex, and perhaps even the St. Louis metro area influencing the statistics in the Southern District of Illinois. At the other end of the distribution, the northeast, the Plains states, and Texas standout as low-filing areas.

The table shows the wide variation in bankruptcy filing rates from less than 1.0 per 1,000 in Alaska to the high of 10.8 per 1,000 in the Western District of Tennessee, as noted before.  On this blog, I have noted the variation at the state level (e.g., here), but digging down to the district level shows an even bigger spread as well as locally intense filing rates. For example, the eastern half of Michigan has a filing rate of 4.56 per 1,000 as compared to 2.78 per 1,000 in the western half. The Central District of Illinois appears more like Iowa, and the northern Plains than the reast of Illinois. The Northern District of Florida and the Middle District of Louisiana stand out as low-filing anomalies in high-filing regions.

A bigger question is why we are seeing these large differences. These data don't answer that bigger question and anything further would be speculation. But, hey, if speculation is good enough for 24-hour news networks, it is good enough for us.

Certainly, some will see economic factors at play -- returning to Michigan the economic woes in Detroit could explain the split across the state or the extreme nature of the housing crises in Nevada and south Florida could explain the high filing rates there. On the other hand, I have a difficult time coming up with economic reasons why, for example, North Carolina and South Carolina would be so different than Tennessee and Georgia.

My hunch is that the likely explanation to explain most of the variation is local culture and local legal culture. The idea of "local legal culture" was developed in separate works by Jean Braucher and Teresa Sullivan, Jay Westbrook, and Elizabeth Warren and is firmly entrenched in the bankruptcy literature as an explanation for differences we see in practices and filing rates across the country.

"Local legal culture" gets us only so far as an explanation. These limits are captured in some ideas contributed by my wiser co-author, Dov Cohen, to a book chapter we co-wrote in Broke. The term certainly fits with a common conception that many people have -- "local legal culture" is what the people in a local legal community "do;" it is their practices that define them. At the same time, "local legal culture" is problematic as an explanation for several reasons. First, the term "legal" can be problematic because the legal professionals in a community may be expressing cutural values that are part of a broader community consensus rather than anything unique to the legal community. As to "culture," we have few direct measures of the attitudes, values, and beliefs of persons in the community or professionals in the legal system. We have to be careful not to use "culture" as a short-hand and empty filler for variation we simply cannot explain. Indeed, Dov and I are working on some research we hope will better explain attitudes, values, and beliefs held around debt, and I better draw this post to a close to get back to work on it.

Many thanks to Epiq Systems for providing the bankruptcy data and especially many thanks to Margie Williams at the Federal Judicial Center who, so long ago that she probably thinks I forgot, sent me a spreadsheet that helped me assemble the U.S. Census population data by federal judicial district.