2018 Peer Update

The CFPB/FFIEC has just released the 2018 National HMDA data set and transmittal info we were all waiting for at 3:00pm EST 8/30/2019. As mentioned in the earlier Start Page article below, the data will be converted and released in two flavors. The first and most useful to all clients will be the Peer (classic) format which will basically take the 2018 data and create the older 2017 format and data fields. Since the data will be exactly the same as years past, it will work with your current Comply versions with no changes as well as the peer website. Since this data will take a little massaging to get into that format and because of the pending hurricane bearing down on Central Florida, we are hoping to release the data by next Thursday, Sept. 5th.

For those needing or wanting to look at the extended peer data, that will take longer.  First we still need to figure out how to deal with banded fields, and the size of the peer data will be over three times bigger than the previous peer data (nearly 30G estimated vs 10G). Most of the extra data will be of little to no use for summary analysis, so for those waiting to look at the entire file peer-wise, that will be around the end of October. We will have to update Comply to be able to import the new fields and the next release will be Comply 19, but as mentioned earlier, 98% of your peer analysis can be done with the "classic" format which will be available next week.

Feel free to contact us if you have any questions or comments.

The FFIEC/CFPB recently released the 2018 HMDA peer data to their Modified LAR page. If you would like to look up your data or see other individual institution data you can find it here https://ffiec.cfpb.gov/data-publication/modified-lar/2018.

In years past, RATA has been able to download the data for all institutions, manipulate the data and release it for use in Peer2Peer and DataMine very quickly. The reason we were able to do this was that the data always contained the same fields and they were in the same basic format. This year however, that is not the case. We want to inform our users about what we are seeing with this new data and new format, the challenges we are running into, and the plan(s) we have in place to get you this data as quickly as possible.

Phase One and Challenge One:

Through 2017, the public data was in the same format with all of the same fields, which we will refer to as the "classic format" for the remainder of this article. We were able to convert that data very quickly because we did not have to make programming changes in Comply, DataMine or the HMDA Peer2Peer website to utilize the data. With the addition of the 2018 fields and the way the government is creating data buckets for certain fields, the format is no longer consistent, and we will refer to this as the "new format". No one really knew exactly how the CFPB was going to release this year's data or even which fields would be released until the data was actually posted, so we could not make any prior accommodations to accept the new data format, essentially placing us in a holding pattern.

To get you the basic data quickly so that you could start using it for your analysis, we were planning to convert the new format to the classic format so that it will work in the current version of DataMine and the Peer2Peer website. Unfortunately, as is often the case, we have run into a roadblock. Last year, the Respondent ID was a part of the submission and the downloaded files were named with the Respondent ID and bank name, however this year the downloaded files are named with the LEI and bank name, since Respondent ID is no longer reported. Because the transmittal information is not released with all of the reporters' pertinent information, in the past we matched the Respondent ID with the Transmittal file from the national dataset for the prior year's reporting to get the necessary information. This year, without the Respondent ID, we are unable to perform that lookup since they used LEI which provides no way to cross reference back to last year's transmittal info. Once the transmittal information can be obtained, it will allow us to cross reference the LEI to the RSSD ID, which then gets us to the Respondent ID.

Without the transmittal info, there is no way to create peer groups or identify institutions. We have sent a couple of emails to the CFPB requesting the transmittal data early, or at least a reference table, but have not heard back yet. They say "the national data will be released later this summer". Once we are able to get the data into the classic format it will look and work exactly as it has in previous years to perform your basic analysis. We think this will give you most of what you need because the new fields are going to be tricky to analyze and we're not really sure if most of them will really be used. If any of you have ideas on other ways we might be able to get the information we need, please feel free to let us know.

Phase Two and Additional Challenges:

Once we have released the new data in the classic format, we then have to architect how we are going to handle the new data going forward. The CFPB has created buckets for some of the data (Age, DTI and Total Units) that use ranges like <25, 25-35 etc. which are not numeric values and will not work with our existing fields. We also need to add all of the new fields to DataMine and reformat all of the previous year's data to make them consistent. All of these changes will take design, programming, QA testing and a new release of Comply. Without knowing the full repercussions of these changes, a time estimate is hard to provide, although we are anticipating a couple of weeks to complete once the actual national data is released by the CFPB. Just as a reference, last year they released the full national file on May 7th. Hopefully they will be somewhere in that general timeframe again. We will keep you posted as we learn more.

Conclusion:

We wanted to let you know that we are aware of the data release, we are actively working on what we can, and we will have the data out to you as quickly as possible so you can start working with. If you need something more quickly you can go to the CFPB Modified LAR page linked above to download your data or other individual banks by Name only. Hopefully you won't need to look for "First National Bank", and if you do, good luck. 

The good news is that next year we'll be able to provide the public data within a couple of days after release, like usual, as the fields will already be set up and we'll have the LEIs for cross referencing. 

Please feel free to call us to discuss any of this and also if you have any ideas or suggestions. We are eager to help in any way we can.