Projected Filings for 2011

There will be 1,457,787.3 bankruptcy filings in 2011. Well, at least that is what the numbers say, although I'm not sure what will happen in that last .3 of a case.

That projection is based on a fairly simple model that uses monthly data from 2006 - 2010 on bankruptcy filings (again thanks to Epiq Systems for providing those data) as well as government data on revolving consumer credit (e.g,, credit card debt), nonrevolving consumer credit (e.g., car loans), and the unemployment rate. The model also accounts for cyclical monthly effects--bankruptcy fiings often spike in February and March and decline in November and December--as well as fixed effects from the The graph to the right shows the model's results. The solid black line shows the actual bankruptcy filings for 2010 through November. The dotted red line shows what the model would have predicted for 2010 while the solid red line shows the model's projections for 2011.

The important information is not the specific projection--1,457,783.3 bankruptcy filings. The specific projection almost surely will be off. Predictions, after all, are a dangerous business. Rather, the important information lies in the trend, and I have added a linear trend line to try to make it easier to see. Bankruptcies likely will decline slightly in 2011, maybe somewhere between 5-10%. A declining bankruptcy rate is not surprising given that unemployment also is expected to decline (at least according to some sources) and there some signs that access to consumer credit is easing slightly (meaning consumers will be able to borrow more easily to avoid bankruptcy if cash flow is tight). Perhaps most significantly, however, consumers have piled up less debt in the past three years meaning there is less debt to discharge and fewer incentives to file bankruptcy.

The mathematical computations show a great fit for the model, and the model does a fairly decent job of capturing the actual trend from 2010. But, the model does underperform in predicting the largest monthly increases and decreases. I am particularly concerned that the model substantially underpredicts the drop in filings for the most recent data in November, suggesting that something happened in November for which the model does not account. For example, one commenter suggested the November drop stemmed from the unique circumstances of the freeze in mortgage foreclosures in September and October. Any model is only as good as the data in the model. If we have unprecedented influences on bankruptcy filing rates, the model's predictions will not be valid. As a thought experiment to understand the point, consider what would happen to bankruptcy filing rates if tomorrow Congress raised the filing fee to $10,000. In this instance, my elegant mathematical would be predicting results from a real-world that no longer exists.

The overall trend in bankruptcy filing rates has been in decline. The second graph to the right shows the year-over-year changes in the daily bankruptcy filing rate. Since April 2009, the year-over-year change has been on a downward trend and now has become negative. Even if we knew nothing else other than the trend in the bankruptcy filing rate, we probably would predict a decrease in bankruptcy filings for the 2011 calendar year.

My prediction is based on total filings. The distinction between consumer and business cases is pretty thin for many self-employed persons, and more self-employed persons file bankruptcy than the official statistics would suggest. The count of consumer versus business filings depends a lot on the incentives filers have to self-identify as a business filer when completing the bankruptcy forms. When using time-series data for bankruptcy filings, the total filings figures produce more stable results. If you're interested in how many consumer filings there will be, consider that about 1 in 7 bankruptcy filers is self-employed and that about 1 in 25 to 30 bankruptcy filings formally identify as a business filer.

For similar reasons, I have not attempted to predict chapter 11 fiings separately although I recognize that is a data point of significant interest to many persons in the insolvency industry. The monthly figures on chapter 11 are very noisy and are very sensitive to how many affiliates a particular chapter 11 filer might have. It is problematic to use chapter 11 filings in a time series analysis.

What does it all mean? Declining bankruptcy rates are hardly bad news, but they are also not the sine qua non of an improved economy. Filing bankruptcy is a legal act with certain benefits and consequences. Not everyone who is financially distressed will benefit from filing bankruptcy. For example, persons with no income or assets to protect will not benefit from bankruptcy. Because bankruptcy wipes out past debts, it does not help households who are having trouble making ends meet but have not yet piled up the massive debts bankruptcy would wipe out.

What the numbers principally mean is that there will be less work for the lawyers who provide bankruptcy services and judicial resources might be stretched a little bit less thinly. The predicted decline, however, is very modest. Also, the prediction is on a national basis; local effects may be very different. The principal takeaway point is that we most likely have stopped seeing the big year-over-year increases in bankruptcy filings that were characteristics of the past five years.

Comments

[Alex Zarcone]

I'm thinking that bankruptcy filings will continue at a good clip. At the end of the day, business aren't hiring and with no jobs debts will continue to pile up.

Small businesses play a major role in the U.S. economy, employing half of all private-sector workers, according to the U.S. Small Business Administration. They have also historically started adding jobs more quickly after recessions than large companies. But they are not doing that this time around apparently. It makes for a sluggish recovery, if any, and fewer new jobs means more bankruptcies.

[Armil @ business listing on google maps]

I don't think it'll gonna happen. Besides, it was just a prediction by computation. So, nothing is exact. Let's just pray that it will not come.