rdamarapati / credit-analytics

1 stars 1 forks source link

Mapping HMDA -> "Respondent-ID" with FDIC "ID RSSD" #2

Closed rdamarapati closed 11 years ago

rdamarapati commented 11 years ago

'HMDA Reporter Panel' file is the universe of all institutions that are reported under HMDA for the activity year.

HMDA Reporter Panel File Format- http://www.ffiec.gov/hmdarawdata/FORMATS/2011HMDAReporterPanel.pdf

HMDA Reporter Panel Data File - http://www.ffiec.gov/hmdarawdata/OTHER/2011HMDAReporterPanel.zip

Respondent RSSD ID - Format is right-justified and zero filled in HMDA Reporter Panel ID RSSD - Format is just a number and is not zero filled in FDIC files.

So, we need to make the adjustment for the leading zeros when we compare and match the 'ID RSSD' in FDIC with 'Respondent RSSD ID' in HMDA files

sharathchandra92 commented 11 years ago

Got it.

So the 2nd coloumn of the LAR files has the RSS ID without the 0s filled and the 2nd coloumn of reporter panel has the RSS ID with the 0s filled.

Approach 1. Remove all the filled 0s from Reporter panel data. And compare. Approach 2. Make the RSS ID in LAR files to be a 10 digit number, thereby adding 0s to it. Computationally very intensive, when we scale it to entire country and across different years.

We'll try the 1st approach.

rdamarapati commented 11 years ago

2nd Column of FDIC file is the RSS ID without zeros. We should match this 2nd column of FDIC file with 'Respondent RSS ID', I think it is 2nd or 3rd LAST column on the Reporter Panel Data.

2nd approach is recommended, these are government assigned numbers and they have defined them as 10 digit numbers. So we need to pad FDIC RSS ID with zeros and then compare. FDIC should have actually released these files after padding.