2018), you need to quickly—if not already—have
processes in place to begin collecting the additional
data on these loan applications. However, there are
still many unanswered questions that make this data
collection difficult. In April 2017, the CFPB released
proposed changes to the rule to increase clarity
within the currently published rule and commentary.
This publication has been useful to allow us to better
understand the intent of some of the data points as
we outline and write out data collection procedures;
however, until this is finalized and the 2018 HMDA
error codes are released, it makes it difficult to write
procedure on some of the nuances of the HMDA
data that we need to be collecting today for our
2018 final action loans.
Speaking of things that may be difficult, have
you obtained your Legal Entity Identifier (LEI) yet?
If not, you may want to make this a priority quickly,
as you may be surprised at the challenges you may
face completing the application.
As we have been working to implement the
necessary policies, data collection changes, and
trainings, we have identified as well as received
input from other compliance officers regarding some
technological challenges that you may also need to
consider as you are finalizing your 2018 HMDA plan.
The first and biggest of these challenges may
be your data collection systems. Has your LOS
provider already launched their HMDA solution?
Do you have multiple LOS’s for various channels or
products (HELOC, Reverse Mortgages, Brokered,
and Correspondent)? If you are working in multiple
systems, or have multiple lines of business, you get
to do HMDA implementation for each of these; do
you have the time and staffing available to test each
of these? What is your back-up plan if the system
is not ready to launch when you are? Is your loan
numbering convention different enough that you
won’t produce the same ULI on different files in
different systems ever? Do the online application
tools that integrate into your LOS have a plan to
be ready to launch into the 2018 data as quickly as
you need them or will you be collecting additional
data outside of the application process—how will
you manage that if the customer submitted an
application, but won’t return your phone call? Will
you be using your LOS to submit the HMDA report,
or do you need to be testing the integration of your
LOS with your HMDA data aggregation software?
Once your software is ready to gather the data,
how will you ensure proper documentation of the
data? Technology companies and some of our
investors are speculating that the Uniform Residential
Loan Application (URLA) will not be ready for use until
2019 or 2020. So how will you, as the compliance
officer in the back office, accurately document the
loan purpose when the current URLA does not ask the
questions needed to solve the loan purpose puzzle?
• Is this a purchase, refi or cash-out refi?
• Does the borrower currently have a dwelling-secured debt?
• Will the loan replace the dwelling secured-debt?
• Will any of the proceeds be used to purchase a
• Will any of the proceeds be used for home-improvement?
Other questions that are not addressed on the
current application include construction method
of the property and the subject property interest
when the home is a manufactured home. Will
you require a printed document to answer these
questions, or do you feel confident that just having
the information marked in your system will be
acceptable to a regulator or your investors? If you
are going to require the documentation, will you
produce your own addendum to the URLA and can
you tie it to the application so that the questions
are being addressed by the LO on every loan
One of the more challenging requirements is
identifying and reporting the “relied upon data.”
Identifying which data fields were truly “relied
upon” may be very difficult and possibly create
substantially more work for your underwriting
team. From the technology aspect, once the credit
decision is made, will your system capture and lock
down those identified “relied upon” data points?
What if information changes between the approval
and closing—for instance, if the hazard insurance
coverage comes in less than anticipated, updating