# Archived Bankruptcy Filing Data

(files were found via archive.org of the uscourts.gov website)

Main site archived in May of 2005. These are mostly the F-2 Tables, which has a default of twelve-month filings preceding. There are a few F-2: Three Tables, which would be quarterly level filings. Keep in mind that the Family Farmers Bankruptcy Act of 1986 was signed into law on 1986-10-27 and into effect on 1986-11-26. This would make the start of Chapter 12 for the quarter ending on 1986-12-31 and be the full value of 1986 Chapter 12.

Original files:

• Twelve-month period ending on March 31 from 1986-03-31 until 2003-03-31.
• Twelve-month period ending on June 30 from 1983-06-30 until 2002-06-30 (plus the quarter ending 2003-06-30 and national values for twelve-month period ending on June 30 since 1960).
• Twelve-month period ending on September 30 from 1987-09-30 until 2003-09-30.
• Twelve-month period ending on December 31 from 1990-12-31 until 2002-12-31.

I have attempted to break these files apart so that each F-2 table is its own .pdf. The naming convention is adopted similar to current files f2 and f2_three tables. The f2_XXXX-YY-ZZ.pdf naming convention indicates that an F-2 twelve-month table ending on the quarter of XXXX-YY-ZZ is contained in the .pdf. A F-2 three-month is referenced as f2_three_XXXX-YY-ZZ.pdf.

Year March June September December
1983 twelve
1984 twelve
1985 twelve
1986 twelve twelve
1987 twelve twelve twelve
1988 twelve twelve twelve
1989 twelve twelve twelve
1990 twelve twelve twelve twelve
1991 twelve twelve twelve twelve
1992 twelve twelve twelve twelve
1993 twelve, missing 11th Circuit twelve twelve twelve
1994 twelve twelve twelve twelve
1995 twelve twelve twelve both
1996 both both both both
1997 both both both both
1998 both both both both
1999 both both both both
2000 both both both both
2001 both both both, missing 10th and 11th Circuit in twelve both
2002 both both both both
2003 both both both both

(We already have these beginning on 2001-03-31 from currently active USCourts.gov, but it’s still a nice double-check)

In addition, there are explicit quarterly level files which appears to be what PACER charges over \$600 for plus some more.

# Backsolving Quarterly Data

At this point, this becomes a recursive problem in attempting to determine the quarterly values prior to 1995-09-30. Let’s start this off with realizing what is entailed in each 12-month file (the subscript A refers to 12-month files while the subscript Q refers to the quarterly version of the data):

${March}_{t, A} = {March}_{t, Q} + {December}_{t-1, Q} + {September}_{t-1, Q} + {June}_{t-1, Q}$ ${June}_{t, A} = {June}_{t, Q} + {March}_{t, Q} + {December}_{t-1, Q} + {September}_{t-1, Q}$ ${September}_{t, A} = {September}_{t, Q} + {June}_{t, Q} + {March}_{t, Q} + {December}_{t-1, Q}$ ${December}_{t, A} = {December}_{t, Q} + {September}_{t, Q} + {June}_{t, Q} + {March}_{t, Q}$

We now have a system of equations and we mostly have all of the 12-month values going back until 1989-12-31, our first missing values. Knowing that we are first missing 1995-09-30, that is our first value we want to solve for, which can be done as such:

${September}_{1995, Q} = {June}_{1996, A} - {June}_{1996, Q} - {March}_{1996, Q} - {December}_{1995, Q}$

We now know the value for 1995-09-30, we can use this in order to solve for 1995-06-30 and continue recursively for each of the other quarters in this generalized fashion (rearranging the above equations):

${June}_{t-1, Q} = {March}_{t, A} - {December}_{t-1, Q} - {September}_{t-1, Q} - {March}_{t, Q}$ ${March}_{t, Q} = {December}_{t, A} - {December}_{t, Q} - {September}_{t, Q} - {June}_{t, Q} +$ ${December}_{t-1, Q} = {September}_{t, A} - {September}_{t, Q} - {June}_{t, Q} - {March}_{t, Q}$ ${September}_{t-1, Q} = {June}_{t, A} - {June}_{t, Q} - {March}_{t, Q} - {December}_{t-1, Q}$

Which will continue going until we run into the 11th Circuit issues in 1993-03-31, which compromises the ability to figure out quarterly values for the 11th Circuit. Then there is an abrupt stop in 1989-12-31 due to missing the 12-month value.

We can alternatively start from the beginning of our data to build up on each other. For instance, we know that the quarter ending in 1986-12-31 was the first instance of Chapter 12 although the Chapter was not officially available until 1986-11-26. Effectively, the 1986-12-31 filing quarter can be simply lumped into the 1987-03-31 quarter and we move forward.

${March}_{1987, A} = {March}_{1987, Q}$ ${June}_{1987, A} = {June}_{1987, Q} + {March}_{1987, Q}$ ${September}_{1987, A} = {September}_{1987, Q} + {June}_{1987, Q} + {March}_{1987, Q}$ ${December}_{1987, A} = {December}_{1987, Q} + {September}_{1987, Q} + {June}_{1987, Q} + {March}_{1987, Q}$

The 1987-12-31 twelve-month filings missing in our data, which makes the quarter of 1987-12-31 unidentified and thus leads to the following values also unidentified:

${March}_{1988, A} = {March}_{1988, Q} + \hat{December}_{1987, Q} + {September}_{1987, Q} + {June}_{1987, Q}$ ${June}_{1988, A} = {June}_{1988, Q} + {March}_{1988, Q} + \hat{December}_{1987, Q} + {September}_{1987, Q}$ ${September}_{1988, A} = {September}_{1988, Q} + {June}_{1988, Q} + {March}_{1988, Q} + \hat{December}_{1987, Q}$