Bank Closures in the US: 2000-2020
The 2008 financial crisis in the United States saw the closing of numerous banks, some large and many small ones. In this project, I gather, clean, organize and visualize data on the bank closures that have taken place in the United States over the last 20 years. The visualizations in particular lead to important and interesting questions.
(Updated 5-22-2020)
Interact with the data: Tableau Dashboard
Download PDF
1. Retrieving the data
The main source data is the FDIC Failed Bank List. It is available to the public here. Initially, I located the data through Data.gov. The csv (comma-separated values) file contains information on all bank closures in the United States since October 1, 2000.
Here are the first 6 rows (the 1st one corresponds to column names):
'Bank Name', 'City', 'ST', 'CERT', 'Acquiring Institution', 'Closing Date'
'Old Harbor Bank', 'Clearwater', 'FL', '57537', '1st United Bank', '21-Oct-11'
'The Bank of Miami,N.A.', 'Coral Gables', 'FL', '19040', '1st United Bank', '17-Dec-10'
'Republic Federal Bank, N.A.', 'Miami', 'FL', '22846', '1st United Bank', '11-Dec-09'
'The Bank of Commerce', 'Wood Dale', 'IL', '34292', 'Advantage National Bank Group', '25-Mar-11'
'Prosperan Bank', 'Oakdale', 'MN', '35074', 'Alerus Financial, N.A.', '6-Nov-09'
In the Jupyter notebook (available here), I store the data in a database, from which I execute queries to sort, group and filter by various criteria. It would be just as effective to use a pandas DataFrame instead.
A second data source, also from the FDIC, will be used in Chapter 5.
2. Cleaning the data
Although this step is not necessary in a modern visualization program such as Tableau, it was important in Python to convert the date_of_closure format from '21-Oct-11' to '2011-10-21'. The way this was done was with the following SQL command:
ALTER TABLE bank_list ALTER COLUMN date_of_closure TYPE DATE using to_date(date_of_closure, 'DD-Mon-YY');
There were no empty fields and no duplicates. The data now looks like this:
3. Bank closures by date
First, we look at the number of closures per year. As we can see in the following plot, most of the failures occured during and shortly after the Great Recession of 2008-2009. The numbers stabilize later, around 2015.
4. Bank closures by state
When we look at the number of closures in each state we see that Georgia, with ranks 8th in terms of population, accounts for the largest number of closures. Texas (2nd in population) and New York (4th) are far from the top.
The number of closures in Georgia is surprising. Nobel-prize winner Paul Krugman wrote about this in an opinion piece for the New York Times in 2010. At the time, Mr. Krugman attributed this to the number of small banks in Georgia that took advantage of a lack of regulations to irresponsibly expand their lending practices:
And for all the concern about banks that are too big to fail, Georgia suffered, if anything, from a proliferation of small banks. Actually, the worst offenders in the lending spree tended to be relatively small start-ups that attracted customers by playing to a specific community. Thus Georgian Bank, founded in 2001, catered to the state’s elite, some of whom were entertained on the C.E.O.’s yacht and private jet. Meanwhile, Integrity Bank, founded in 2000, played up its “faith based” business model. It was featured in a 2005 Time magazine article titled “Praying for Profits.” Both banks have now gone bust.
It's important to note at this point that, so far, we've been tracking only the number of bank closures and not total assets. Next, we'll look at the size of these banks.
5. Bank closures by assets
For this section, we consider our 2nd dataset, which is available here via the FDIC. Here are the first 5 entries:
The data showed that there is a large difference between the biggest bank (Washington Mutual) and the rest: Washington Mutual had close to $307 billion in assets, while the 2nd largest bank was IndyMac Bank with $32 billion, almost 90% smaller.
6. Acquiring institutions
Here we take a look at what banks acquired the most failed institutions, keeping in mind that the data showed 31 banks did not get acquired.
Interestingly, 4 out of the top 5 acquiring institutions are based and operate mainly in the Southeast.
The bank that acquired the most dissolved institutions, State Bank and Trust Company was recently acquired by Cadence Bancorporation, effective January 1, 2019.
"Funded by private investment capital, State Bank and Trust was opened in 2005 to acquire failed banks in FDIC assisted transactions." *Problem Bank List, October 14, 2011*