Using Small Business Administration 7(a) loan data, Lynn Hulsey and Ken McCall of the Dayton Daily News found a pattern of defaults among some restaurant franchise owners, and few payments on some defaulted loans. They write:
“More than half of the 168,324 charged-off loans failed before 20 percent of the loan was repaid. More than one in three repaid only 10 percent or less of the loan. More than 7 percent did not reduce the principal on their loan at all.”
Getting data is rarely easy and the SBA data was no exception, Ken says. The reporters received about five sets of data from their FOIA requests before they got the information they needed. He suggests reporters do “integrity checks” by counting different fields and records. He also says to check for missing years and coding values.
The next obstacle: the data came in an Excel format.
“Excel does weird things to data types because it can’t specify what data is what,” he says.
He came up with what seems to be an easy solution – though I wouldn’t have thought if it. He exported the information into the Access program, saved it as a .csv file and imported it into MySQL,which handles a larger pool of data, he says.
Finally, they learned SBA lenders submit loan information without a standardized system, which meant the team had to clean up more than 1 million records, Ken says.
The result was a localized searchable database showing companies that defaulted on their loans. The analysis also allowed them to see the effects the recession had on defaults and lending.
Stay tuned for more stories generated by the database, Ken says.