HMDA Adoption Costs: Did You Say $2 Billion? As a Matter of Fact, We Did!

Link to PDF

Summary:

  • Two adjustments to ensure consistency and comparability with estimated annual HMDA compliance costs increase the CFPB’s estimated one-time HMDA adoption costs from $1.34 Billion to $2.12 Billion, an increase of 58%.
  • The CFPB’s effort to rationalize adoption costs through a “frame of reference” and suggested accounting treatment of adoption costs can only be described as analytical alchemy. 
  • The projected expenditure of approximately $2 Billion to provide the CFPB with 50 additional HMDA data fields indicates an unprecedented cost of $40 Million per data field.
  • The projected HMDA adoption costs raise a number of critical questions including (i) the reliability of CFPB financial models and related analytics, and (ii) the sufficiency of the estimated benefit of the new HMDA rules to consumers. 

Analysis:

It’s difficult, if not impossible, to argue with the CFPB’s projected one-time costs to modify processes in response to the new HMDA regulatory requirements.  Literally.

On one hand, the estimated one-time adoption costs seem to be comprehensive as they include updating software systems, training staff, updating compliance procedures and manuals, and overall planning and preparation time.  On the other hand, there isn’t much detail behind the numbers presented in Table 1 taken from the CFPB commentary accompanying the 2015 HMDA amendments.

Table 1 | Estimated HMDA Rule Change Adoption Costs

Institution Complexity Closed-End Only Closed-End and Open Open-End Only
High $800,000 $1,200,000 No Lenders is this category
Medium $250,000 $375,000 $250,000
Low $3,000 $3,000[1] $3,000

Based on the estimates shown in Table 1, the CFPB projects that the combined overall one-time impact of adopting the new HMDA rules is between $725.9 Million and $1.34 Billion.

As we said, we can’t argue with these numbers.  However, two adjustments seem appropriate to ensure consistency and comparability with the CFPB’s estimated annual HMDA compliance costs:

  1. Matching Estimated Cost with Lender Counts.  The CFPB’s analysis of recurring HMDA compliance costs under the new rules detailed the number of institutions by reporting profile.  We applied the estimated costs in Table 1 to the lender counts in the interest of accuracy, consistency, and comparability.  In the interest of conservatism[2], we used the lowest estimated cost for medium complexity institutions rather than attempt to derive a “weighted cost” to account for the mix between closed-end lenders and lenders that provide both closed-end and open-ended products.
  1. Labor Cost Adjustment.  As noted in our evaluation of the annual HMDA compliance costs under the new rules[3], the CFPB uses an hourly labor rate of $33 in their estimate of recurring annual compliance costs.  We believe an hourly labor rate of $50 is much more realistic as it reflects the cost of employee benefits, information technology, and corporate overhead.  The $17 difference in the hourly labor rate represents a variance of approximately 50%.

Adjusting labor costs requires an estimate of the portion of the CFPB’s projected adoption costs that are attributable to labor.  Based on our experience in compliance systems and compliance components on core processing platforms, the aggregate cost to update software for the new HMDA rules should not exceed $50 Million[4], or approximately 4%, of total projected costs in any known solar system, galaxy, or universe. In the interest of conservatism, our labor cost adjustment assumes that labor costs comprise 75% of the total estimated HMDA adoption costs[5] which translates into $335 Million generously allocated to updating software systems.

The results of our adjustments, presented in Table 2, indicate an estimated range of $1.184 Billion to $2.12 Billion for adoption costs.

Table 2 | Adjusted Estimated HMDA Rule Change Adoption Costs

Table2

We believe these adjustments are reasonable but it’s important to take our analysis for what it’s worth – an attempt to improve an estimate that is not particularly transparent.  Regardless, the magnitude of the cost estimate – whether $1.34 Billion or $2.12 Billion – leaves us speechless and demonstrates that the CFPB is adept at using “other people’s money”[6] to acquire the data they want to achieve their mandate.  In any event, lenders are covering the cost of providing the CFPB with the data that will make lenders’ lives more challenging. We’re frankly surprised at the lack of substantive “hue and cry”.

We’ll offer further – and final thoughts – on these numbers after addressing the CFPB’s efforts to temper the estimated adoption costs.  Specifically:

A. Frame of Reference.  The CFPB offers a creative “frame of reference” for its adoption cost estimate.  Specifically, the CFPB indicates that the $1.34 Billion in estimated adoption costs is 0.3% of approximately $420 Billion in total non-interest expenses reported by HMDA Respondents in 2012.  This analytical alchemy suggests that the projected costs are immaterial and, therefore, reasonable.  This frame of reference is faulty for several reasons including:

  • Four lenders account for 56%[7] of the $420 Billion.  Adjusting estimated adoption costs and non-interest expenses for these four lenders increases the rate for the remaining lenders from 0.3% to 0.8% of non-interest expenses.  Reasonably assuming that the top 29 lenders account for $79 Billion (18.8%) of the $430 Billion and adjusting the associated estimated adoption costs increases the rate for the remaining 6,221 lenders to 1.24%.
  • Measuring the adoption costs as a percent of non-interest expense rather than net income before taxes yields a less relevant “frame of reference”.  Further, using 2012 results (while annual compliance costs are based on 2013 data) yields a lower “frame of reference” estimate due to the strength of 2012 application activity compared to 2013.

B. CFPB Regulatory Accounting Principles.  Financial reporting rules and regulations, embodied in Generally Accepted Accounting Principles (“GAAP”), are promulgated by the Financial Accounting Standards Board and the Securities and Exchange Commission, respectively.  Despite this fact, the CFPB seeks to temper estimated HMDA adoption costs by suggesting that “financial institutions are expected to amortize these costs over a period of year” and then indicate that the annual upper bound cost of $1.34 Billion becomes $326 Million per year by applying a 7% discount rate and a five-year amortization period. Three issues with this approach immediately come to mind:

  • First, the application of a discount rate and a five-year amortization period is valid only to the extent that a lender borrows funds to cover adoption costs.  Such an approach adds $114.3 Million in interest costs over the five-year period to the $1.340 Billion adoption costs, raising the total cost to $1,634 Billion.[8]
  • Second, the capitalization and amortization of expenditures is generally accomplished when qualifying costs are incurred, capitalized, and then amortized over the benefit period.  This more likely approach in the event of adoption cost capitalization means the annual amortized cost would be $268 Million and not $326 Million.
  • Third, and most importantly, adoption costs are likely a period expense and not eligible for capitalization for reasons that include (i) the avoidance of impairment is different than creating future value, and (ii) difficulty in measuring the duration of any potential future benefit.

The CFPB’s accounting treatment suggests a basis of accounting – CFPB Regulatory Accounting Principles (“CRAP”) – that supersedes GAAP.  For the record, we recommend that mortgage lenders consult with their auditors in determining how to account for HMDA adoption costs.

The CFPB’s efforts to temper the adoption costs associated with the new HMDA rules is understandable if erroneous.  On more than one occasion in the commentary accompanying the amendments to Regulation C, the CFPB commented about the lack of feedback on the cost of the proposed HMDA rules[9].  This comment suggests that comments from mortgage lenders were prepared without the assistance of the Chief Financial Officer or other accounting professional.

Our view is that mortgage compliance professionals should pass along a copy of the final HMDA rules to their CFO or other accounting professional for an ex post facto assessment of the financial consequences of the new rule and, more importantly, the development of a plan to generate a Return on Investment on the HMDA adoption and annual compliance costs.

Perhaps the best way to evaluate the reasonableness of HMDA adoption costs is to evaluate them in the context of the what’s being “purchased”.  Appendix A, which summarizes the HMDA reporting model under the amended rules, indicates that the that the substance of the amendments and the estimated $2.11 Billion adoption costs, is to provide the CFPB with 50 additional data fields – at an average cost of approximately $40 Million dollars per data field. The staggering costs – absolute and per field – leads to one of two conclusions:

  • The CFPB’s estimates are wrong.  While likely the case, this conclusion calls into question the CFPB’s ability to measure – and justify – the costs and benefits of their actions in any instance[10].
  • The CFPB prepared the HMDA cost/benefit analyses to support the decisions reached based on other factors.

Either conclusion presents incredible challenges – and existential risks – to lenders.

* * 

Our next HMDA Insight will (finally) provide specific examples of how the current HMDA data model can drive a lender’s growth.

 * * *

“It is a capital mistake to theorize before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts.” ― The Adventures of Sherlock Holmes, A Scandal in Bohemia

 * * * *

 

Appendix A | HMDA Reporting Fields Pursuant to October 2015 Amendments

HMDA Data Fields

HMDA Data Fields

 

 * * * * * 

[1] “The Bureau believes that the additional one-time costs of open-end reporting will be relatively low for low complexity financial institutions” and “these institutions are less reliant on information technology systems for HMDA reporting and that they may process open-end lines of credit on the same system and in the same business unit as closed-end mortgage loans” thus “the Bureau estimates that the additional one-time cost created by open-end reporting is minimal and is derived mostly from new training and procedures adopted for the overall changes in the final rule.”

[2] As suggested by the use of words such as “consistency” and “comparability”, the primary author of this document is a CPA.  Therefore, the use of the word “conservatism” is in an accounting context and not a political context.

[3] See “What Cost HMDA Data Quality?  Try $425 Million” at www.mortgagetrueview.com.

[4] The primary author was involved in the implementation of significant changes to mutual fund compliance reporting protocols for more than 8,000 mutual funds involving the adoption of XBRL and multi-jurisdictional reporting.  The aggregate industry-wide cost of updating compliance systems (including feeds from core platforms) was less than $50 million.

[5] Thus, the CFPB HMDA adoption cost estimates where multiplied by 1.375% (50% hourly rate differential x 75% of gross adoption costs estimated to be labor-related.

[6] Without expressing a political view on the merits of building a wall to protect our borders, perhaps the suggestion that another sovereign nation will pay for a wall to isn’t so crazy.

[7] Specifically, JPMorgan ($65 Billion), Wells Fargo ($49 Billion), Bank of America ($71 Billion), and Citibank ($51 Billion).

[8] The use of a 7% discount rate indicates another flaw in the analysis and calls into question the CFPB’s analytical acuity.  The Fed Funds rate available to financial institutions engaged in mortgage lending is currently 50 basis points.  Using a discount rate of 50 basis points would produce a significantly lower annual amortized cost estimate.

[9] See, for example, Home Mortgage Disclosure (Regulation C); Final Rule, 12 CFR Part 1003, page 6624 which states that “no commenter provided specific estimates on the potential one-time costs of reporting open-ended lines of credit…”

[10] It also calls into question the CFPB’s analytical acuity.

What Cost HMDA Data Quality? Try $425 Million

Link to PDF Version

Summary:

  • Analysis of the CFPB’s HMDA compliance cost estimates in the October 2015 HMDA amendments indicates the CFPB is projecting the compliance burden to increase 16% for lenders with more than 20,000 applications and between 100% and 200% for all other lenders.
  • Analysis indicates that data management and validation costs represent approximately 87% of estimated aggregate annual costs of $489 Million.  Representing approximately $425 million in aggregate, HMDA data management and validation costs can be managed – and reduced – through advanced analytical protocols.

Analysis:

Part VII.F.1 of the October 2015 HMDA Amendments contains the CFPB’s “concise, high-level overview of the benefits and costs of the final rule”.  While the 33 pages comprising Part VII.F.1. are admittedly high-level and, in many cases, riveting, the three pages of cost analysis provided as Part IX pursuant to the Paperwork Reduction Act are much more concise and, therefore, useful.  The following Table from Part IX is particularly interesting:

Table1

Table 2 above makes it clear that the CFBP anticipates that in order for a lender to meet the new HMDA reporting requirements, the average Tier One Respondent will need approximately five Full-Time Equivalents (FTEs), the average Tier Two Respondent will need approximately 75% of one FTE, and the average Tier Three respondent will require approximately 4 week of effort.  As shown in the following table, the expectations represent a significant increase over the CFPB’s estimate of the current, or “baseline[1]”, HMDA reporting burden:

Table2

The preceding tables indicate that somewhere between the CFPB’s assessment of the current – or historical – HMDA reporting “baseline” and issuance of the final rules, the reporting burden for Tier One lenders increased nominally while the reporting burden for Tier Two respondents doubled and the burden for Tier Three respondents almost tripled.  The tables above frame two fundamental (and possibly existential) questions for mortgage lenders – what are the costs associated with the CFPB’s estimated burden hours and what Tier am I in?

HMDA Compliance Costs

The following chart[4] quantifies the costs associated with the CFPB’s estimated HMDA reporting burden[5]:

   Chart 1 | Annual HMDA Compliance Cost Estimate (Click to enlarge)

Chart 1_ANnual HMDA Compliance Cost Estimate

Chart 1 indicates the estimated average annual HMDA compliance cost, exclusive of amortized adoption costs[6], is $529,400 for Tier One Respondents, $71,700 for Tier Two Respondents, and $8,650 for Tier Three Respondents with an estimated aggregate annual cost estimate of $489 million.  The CFPB indicates that that these are averages and should not be construed to indicate that every lender in a specific tier will incur the indicated average cost.  We agree with this position but we believe it is particularly true for Tier One as noted below.

Reporting Tier Applications Counts

This bring us to the critical question facing each mortgage lender – what HMDA Reporting Tier am I in? The answer to this question requires identifying the number of applications associated with the respondents in each tier.  Our estimate of applications in each tier, presented in the Chart below requires, some explanation. Specifically:

  • The CFPB’s estimate of 6,250 HMDA filers for 2013 includes an estimate of new filers.  We don’t have the information needed to form our own estimate so our analysis applies the new HMDA reporting provisions to the 2013 HMDA data submitted by 7,190 respondents.
  • We determined pro forma 2013 HMDA respondents based on whether or not the reporting threshold was met for each of the two preceding years and then categorize the respondents into Tiers are set forth in Part VII.F.1.
  • Our analysis determined that 5,527 of the 2013 HMDA respondents meet the reporting threshold.

We noted a number of issues in preparing the Chart below that call into question whether or not the major benefits enumerated by the CFPB in Part VII.F.1 can be achieved.  Among these issues:

  • 217 of the 5,527 respondents that qualified based on 2011 and 2012 originations do not report any applications in 2013.
  • 6 respondents each reporting more than 1,000 applications in 2013 do not meet the origination threshold in 2012.  These 6 respondents result in the omission of 332,564 applications.  One of the 6 lenders reported 315,608 applications, essentially all of which were purchased loans.
  • 38 respondents each reporting more than 1,000 applications in 2013 do not meet the origination threshold in 2011.  These 30 respondents result in the omission of 150,796 applications with one lender accounting for 36,197 applications.
  • 29 respondents each reporting more than 1,000 applications in 2013 do not meet the origination threshold in either 2011 or 2012.  These 29 respondents result in the omission of 181,740 applications with one lender accounting for 96,897 applications.

Here is table presenting the number of applications by Tier:

Chart 2 | Annual HMDA Compliance Cost Estimate with Applications by Tier (Click to enlarge)

Chart2_Annual HMDA Compliance Cost Estimate with Applications by Tier

This Chart indicates that:

  • the average Tier One costs are for a lender with approximately 26,000 applications;
  • the average Tier Two costs are for a lender with approximately 1,100 applications; and
  • the average Tier Three costs are for a lender with approximately 65 applications.

As noted above, the calculated compliance costs per tier are based on average applications per tier.  Therefore, a lender’s actual compliance cost will likely vary; however, the degree of variance is (i) nominal for Tier Three due to the number of estimated hours and application standard deviation (101) and (ii) significant for Tier One based on the number of estimated hours and application standard deviation (27,186).

Finally, the increase in compliance hours over the baseline model are most likely attributable to data management and validation activities as baseline reporting and submission, audit, and examinations charges are relatively fixed.  This means that approximately 87% ($425 million) of HMDA compliance costs – the management and validation of data – are controllable and any savings drop to the lender’s bottom line!

* *

Our next HMDA Insight will take a deeper dive into the numbers and outline implications and solutions.

* * *

“It is a capital mistake to theorize before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts.” ― The Adventures of Sherlock Holmes, A Scandal in Bohemia

* * * *

[1] The baseline was set through “a cost-accounting, case-study methodology consisting, in part, of interviews with 20 financial institutions of various sizes, nine vendors, and 15 government agency representatives).  See Federal Register, Volume 80, No. 208 pages 66271 through 66273.

[2] The number of Upper Bound Tier 3 respondents (921) is less than the number of Lower Bound Tier 3 respondents (3,943) which identifies The Upper Bound as the most likely reporting scenario.  This conclusion is further confirmed by the fact that the Tier 2 Upper Bound burden hours (1,434) is less than the Tier 2 Lower Bound burden hours (1,619).

[3] Tier 1 Upper Bound burden segment differential of 476 hours (11,034 hours – 10,558 hours) is attributable to quarterly reporting requirements imposed on lenders reporting more than 60,000 applications exclusive of loans purchased.

[4] Our cost estimate is based on an hourly rate of $50.  CFPB baseline costs are based on an hourly rate of $33 which represents the national average hourly wage for compliance officers based on the May 2014 Bureau of Labor Statistic’s National Compensation Survey.  We believe increasing this rate to $50 is necessary to incorporate non-wage employee costs including benefits, payroll taxes and allocated overhead such as financial accounting and controls, insurance, office space and information technology costs.

[5] The number of respondents in the CFPB’s Table 2 did not foot to the indicated totals.  We assigned differences to Tier 3.

[6] The adoption costs associated with the new HMDA rules will be addressed in a separate HMDA Insight.

 

 

Improving HMDA Data Quality: Is HMDA Data Quality Good Enough to Justify Enforcement Actions?

Link to PDF Version

Summary:

  • Regulators use of HMDA data to evaluate mortgage lending activity can result in significant fines and penalties to mortgage lenders[1].
  • Mortgage lenders use of HMDA data released by the Federal Financial Institutions Examination Council (“FFIEC”) can prevent, detect, and correct potential issues using advanced analytics and peer benchmarking. 
  • Publicly available HMDA data, however, is not updated for late and/or revised filings thus calling into question regulatory actions based the original data and compromising the effectiveness of mortgage lenders’ responsible conduct protocols.[2]

Analysis:

Our last HMDA Insight[3] highlighted the delay in releasing HMDA data based on a comparison of the HMDA file modification dates to the public release date for those files.  A year-over-year comparison of the HMDA file modification dates for the period 2010 – 2014 shows that the HMDA files were not changed after their release to the public.  This indicates that the HMDA data was not subsequently updated to incorporate late and/or revised HMDA filings.

The likely justification for not issuing revised HMDA data files is that the number of applications involved is immaterial.  While there are no public comments – or supporting data – from regulators to substantiate this justification, our analysis shows that the scope of incomplete and/or inaccurate HMDA data is sufficient to merit the issuance of revised HMDA files.  Specifically:

  • Our analysis of HMDA data for the period 2010 -2014 (Table 1) identified six cases indicative of incomplete data. Our review of the result found instances of non-compliance with HMDA reporting requirements.  In one case the lender was unaware that their filing, consisting of more than 25,000 applications, had not been accepted by the FFIEC and was not included in the FFIEC’s public HMDA file. [4]
  • Mortgage TrueView’s LenderScore [5] algorithms found that aberrant B Scores were attributable to inaccurate data involving more than 25,000 applications per year over the five-year period.

A recent discussion with regulators indicated the most likely reason HMDA data files aren’t revised, at least in the case of incomplete data, is the lack of regulatory analytics (such as those summarized in Table 1) to identify HMDA non-filers.  This rather surprising admission raises the possibility that regulatory actions based on HMDA data lack credibility.

It’s unclear if the credibility of regulatory actions is further compromised by revised HMDA filings.  If the database used by regulators is not updated for revised HMDA filings, then regulatory actions likely lack credibility.  If regulator databases are updated for revised HMDA files, the results are likely credible; however, mortgage lenders would be justified in arguing that they didn’t have the information needed to responsibly, and proactively, address the issues that led to a regulatory action.

Table 1 | Incomplete HMDA Data Cases and Assessment (2010 – 2014)

Case

Results Comment

Assessment

  1. 2010 Respondents reporting more than 100 applications with no reported 2011 applications

7 Respondents

All but one of the 7 respondents reported applications in 2012, 2013, and 2014 Probable that results indicate missing LAR from the identified Respondents.
  1. 2011 Respondents reported more than 100 applications with no reported applications in 2012.

7 Respondents

All but one of the 7 respondents reported applications in 2013 and 2014.  In addition, all but one reported applications in 2010. Probable that results indicate missing LAR from the identified Respondents.
  1. 2012 Respondents reported more than 100 applications with no reported applications in 2013.

4 Respondents

All four respondents reported applications in 2010, 2011, and 2014. Probable that results indicate missing LAR from the identified Respondents.
  1. 2013 Respondents reported more than 100 applications with no reported applications in 2014.

186 Respondents

All but 10 of the 186 respondents reported applications in 2012.The Respondent reporting the most applications was an acquisition. Unable to fully differentiate between Respondents that stopped originating mortgages and those that may have failed to submit a LAR until release of 2015 HMDA data.
  1. 2014 Respondents that reported more than 100 applications in 2014 with no reported applications in 2013.

10 Respondents

All Respondents reported applications in at least one other year with four (4) of the 10 reporting in 2010, 2011, and 2012. Probable that results indicate missing LAR from the identified Respondents.
  1.  2014 Respondents reporting more than 100 applications with no reported applications in any of the prior periods.

133 Respondents

The top Respondent based on application count was involved in an acquisition.There are 16 other respondents that reported more than 1,000 applications in 2014. Unable to differentiate between new Respondents and those that may have failed to submit a LAR in one or more of the preceding periods.

 * * *

“It is a capital mistake to theorize before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts.” ― The Adventures of Sherlock Holmes, A Scandal in Bohemia

 * * * *

[1] Examples include http://files.consumerfinance.gov/f/documents/201606_cfpb_bancorpsouth-consent-order.pdf (Accessed on July 27, 2016), http://files.consumerfinance.gov/f/201511_cfpb_hudson-city-consent-order.pdf (Accessed on July 27, 2016), http://files.consumerfinance.gov/f/201310_cfpb_consent-order_washington-federal.pdf (Accessed on July 27, 2016) and http://files.consumerfinance.gov/f/201310_cfpb_consent-order_mortgage-master.pdf).  (Accessed on July 27, 2016).

[2] See “Responsible Business Conduct: Self-Policing, Self-Reporting, Remediation, and Cooperation” located at http://files.consumerfinance.gov/f/201306_cfpb_bulletin_responsible-conduct.pdf. (Accessed on October 2, 2015).

[3] See HMDA Insights: Earlier Release of HMDA Data, Volume 7, July 18, 2016.

[4] The mortgage lender provided us with a copy of the data and it was loaded into Mortgage TrueView. The results presented in Table 1, therefore, exclude this instance of incomplete data

[5] See LenderScores.com.

Earlier Release of HMDA Data

Link to PDF Version

Summary:

  • The annual Federal Financial Institutions Examination Council (“FFIEC”) HMDA data plays an important part in assisting mortgage lenders prevent, detect, and correct potential issues through advanced analytics and peer benchmarking.  As a result, the FFIEC HMDA data is an important element in a mortgage lender’s “responsible” governance, risk management, and compliance “conduct.”
  • The annual Federal Financial Institutions Examination Council (“FFIEC”) HMDA data file is finalized approximately two months before it is released to the public.
  • Delaying the release of this data to the public compromises the efficiency and effectiveness of responsible conduct protocols.

Analysis:

The following chart summarizes the “Hold Days” (i.e., the number of days between the date the data file was last modified and the date the file was released to the public) for the three most recent annual FFIEC HMDA data files:

Chart 7.1

                                                     Source: FFIEC website.

The delay, averaging 71 days, is likely due in part to providing the FFIEC time to prepare the analysis included in the Press Release issued in connection with the release of the HMDA data.  Delaying the release of the HMDA data places mortgage lenders at unnecessary risk by delaying its use in their preventative, detective and corrective governance, risk management, and compliance controls.Source: FFIEC website.

Releasing the HMDA data earlier does not prevent the FFIEC from issuing their analysis at a subsequent date.

* * *

“It is a capital mistake to theorize before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts.” ― The Adventures of Sherlock Holmes, A Scandal in Bohemia

 


 

© 2016 Mortgage TrueView, Inc.  | All Rights Reserved.  This material may not be used for any commercial purpose in whole or part without the express written permission of Mortgage TrueView, Inc.

 

 

HMDA Data Validation Issues: Gender|Race|Ethnicity

Link to PDF Version

Summary:

  • The Federal Financial Institutions Examination Council (“FFIEC”) performs syntactical, validity, and quality edit checks before accepting a Respondent’s Loan/Application Register (“LAR”).
  • Mortgage TrueView HMDA Insight Dashboards[1] show that FFIEC edit checks are not sufficient to prevent HMDA Respondents from filing erroneous and/or inaccurate data that, at best, calls into question the quality of a Respondent’s data or, at worst, adversely impacts the evaluation of a lender’s lending profile.

Analysis:

The following simple examples are illustrative of the shortcomings associated with FFIEC edit checks:

Example 1.  Respondents are required to provide at least one, but not more than five, Applicant Race indicator(s) for each application.  Allowable indicators consist of 1 (American Indian or Alaska Native), 2 (Asian), 3 (Black or African American), 4 (Native Hawaiian or Other Pacific Islander), 5 (White), 6 (Information not provided by applicant in mail, Internet, or telephone application), or 7 (Not Applicable).

FFIEC HMDA Validity Error Check V310 tests to ensure that the Applicant Race field includes at least one defined value and Validity Error Check V480 tests to ensure that if two or more Applicant Race values are included, the same value is not reported more than once (for example, 1,1, 2,2, etc.).  Chart 1 shows, however, that there is not an FFIEC HMDA edit check to ensure that the Applicant Race fields are not populated with all available race values (i.e. 1,2,3,4,5).

         Chart 1 | Applications with Applicant Race Fields Reporting 1, 2, 3, 4, and 5 by Action Taken Category

6.Chart1

While reporting 1,2,3,4,5 might be a valid racial profile, further analysis of the this reporting pattern suggests that such instances may not always be indicative of a multi-racial applicant. For example, analysis of the 2014 results indicate:

  • 43.51% of the reported instances are associated with denied applications. This indicates that the coding may be attributable to the disposition of the application within regulatory timelines.
  • One Respondent accounts for 228 (approximately 50%) of the comprehensive racial instances with 112 of the 228 (approximately 50%) of those instances associated with denied loans.

Example 2. Respondents are required to provide at least one Applicant Gender indicator, one Applicant Race indicator, and one Applicant Ethnicity Indicator.  Such indicators may report that the field is “Not Applicable” (Gender code 4, Race code 7, and Ethnicity code 4).  The HMDA reporting instructions indicate that “Code 4 for Ethnicity and Code 7 for Race…can only be used when the Applicant…is not a natural person or when applicant information is unavailable because …the loan has been purchased…”.

Chart 2 indicates those instances where the Applicant Gender code is reported as either Male or Female but the related Race and Ethnicity codes are reported as Not Applicable. This suggests that either the reported Gender code is incorrect or the reported Race and Ethnicity codes are incorrect.

 Chart 2 | Applications with Gender = Male or Female with Race and Ethnicity = NA

6.Chart2

With regards to 2014, approximately 72% of the non-compliant usage is associated with Loans Purchased (and approximately 50% of the non-compliant usage is reported by one Respondent) suggesting that the reported Gender code should be “Not Applicable” or that the Ethnicity and Race codes should be reported as other than “Not Applicable”.

Unfortunately, space does not permit more exposition on this topic but we have other examples we’ll share in connection with future insights.

* * *

“It is a capital mistake to theorize before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts.” ― The Adventures of Sherlock Holmes, A Scandal in Bohemia

* * * *

[1] Mortgage TrueView dashboards can be accessed by 2015 HMDA Survey participants using their complementary access credentials.  Non-participants may obtain credentials by sending a request to hmda@mortgagetrueview.com.


© 2016 Mortgage TrueView, Inc.  | All Rights Reserved.  This material may not be used for any purpose in whole or part without the express written permission of Mortgage TrueView, Inc.

 

 

HMDA Reporting Rates: Gender|Race|Ethnicity

Link to PDF Version

Summary:

  • Mortgage TrueView HMDA Insight Dashboards[1] show that the rate at which net adjusted applications[2] include applicant gender, race, and ethnicity rates vary in a given MSA compared to the overall rate of 85.60% (Chart 1)

Chart 1 | Gender, Race, and Ethnicity Rates by MSA/MD [Net] : Top 50 MSAs/MDs 5.Chart1

  • Gender, race, and ethnicity reporting rates and related variances show that it is essential to understand the MSA- and lender-specific gender, race, and ethnicity reporting rates to properly evaluate a lender’s lending profile on a direct and comparative basis. 

Analysis:

Table 1 presents 2014 Action Taken and Aggregate gender, race and ethnicity data rates.

5.Table1

These overall results indicate that:

  • As expected, the gender, race, and ethnicity reporting rate is the highest for approved loans and the lowest for incomplete loans.
  • The gender, race, and ethnicity reporting rates for denied and withdrawn rates are comparable, raising the issue of whether or not a written request for withdrawal was received before a credit decision was made.
  • The gender, race, and ethnicity reporting rates for incomplete and withdrawn applications highlights the challenge in evaluating a lender’s fair lending profile for these applications.

Table 2 highlights gender, race, and ethnicity reporting rates, exclusive of Puerto Rican[3] MSAs, for the indicated Action Taken categories.

Table 2 | Gender, Race, and Ethnicity Reporting Rates for Selected Segments [2014]

5.Table2

This table demonstrates a number of key issues including:

  • The evaluation of a lender’s fair lending profile is more problematic – and perhaps impossible – in MSAs/MDs where the overall rate of gender, race, and ethnicity reporting is below a standard threshold.
  • Lenders with gender, race, and ethnicity reporting rates below the average for an MSA are likely to attract the attention of regulators.  Perhaps more important, lenders lack relevant demographic information to grow their business.
  • Lenders with gender, race, and ethnicity reporting rates above the average for an MSA face a risk/reward paradox – higher gender, race, and ethnicity reporting rates result in greater scrutiny of their lending while the lending profile for lenders with lower gender, race, and ethnicity reporting rates is less certain therefore lowering the likelihood that the heightened scrutiny will yield a favorable reaction from regulators and other stakeholders.

 * *

Our next HMDA Insight will provide further insight in gender, race, and ethnicity reporting by presenting instances of data issues not detected through HMDA validity and quality edit checks performed by the Federal Financial Institutions Examination Council at the time lenders submit their data.

* * *

“It is a capital mistake to theorize before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts.” ― The Adventures of Sherlock Holmes, A Scandal in Bohemia

* * * *

[1] Mortgage TrueView dashboards can be accessed by 2015 HMDA Survey participants using their complementary access credentials.  Non-participants may obtain credentials by sending a request to hmda@mortgagetrueview.com.

[2] Net adjusted applications exclude Loans Purchased and the reclassification of loans with Denial Reason Code 7 as Incomplete Action Taken.

[3] Gender, race, and ethnicity reporting rates for Puerto Rican MSAs range from 95% to 99%.


© 2016 Mortgage TrueView, Inc.  | All Rights Reserved.  This material may not be used for commercial purpose in whole or part without the express written permission of Mortgage TrueView, Inc.

 

 

HMDA Denial Reason Codes: Trends and Opportunities

Link to PDF Version

Summary:

  • Mortgage TrueView HMDA Insight Dashboards[1] indicate that the percentage of denied applications providing a Denial Reason Code (“DRC”) has dropped approximately 16% – from approximately 80% to 68% – during the period 2010 – 2014[2] (Chart 1).

Chart 1 – Denied Applications with at least one Denial Reason Code 

4.Chart1

  • Analysis confirms that the DRC reporting trend shown in Chart 1 is driven by lenders that are supervised by either the Federal Reserve System or the Department of Housing and Urban Development.
  • The paradox of lower DRC reporting prior to mandatory[3] DRC reporting identifies a strategic opportunity for those lenders not currently required to provide a DRC.  Such lenders have an opportunity to report DRCs prior to January 1, 2020, to provide a more comprehensive understanding of denial activity for the overall market and themselves. 

Analysis:

Table 1 details the drop in DRC reporting rates among lenders by supervising regulatory agency.  Overall, the percentage of denied applications reporting at least one DRC averaged approximately 80% during the period 2010 – 2012, dropping to 77% in 2013 and 68.00% in 2014.

4.Table1

The drop in DRC reporting presented in Table 1 is attributable to a several factors including (i) changes in market share between DRC reporting lenders and non-DRC reporting lenders, and (ii) lower levels of  voluntary DRC reporting by lenders not obligated to report DRCs.

Table 2 highlights the trend in DRC reporting for a representative group of HUD-supervised lenders.  Among other things, this table shows that some lenders “refined” their view on reporting DRCs over the five-year period but, in the end, all the lenders in Table 2 reached the conclusion to no longer report DRCs in 2014.

4.Table2

For those lenders reporting DRCs, Table 3 shows that the use of collateral-related DRCs as the primary DRC have declined from 28.97% in 2010 to 21.75% in 2014.  Applicant-related DRCs have increased from 56.63% to 68.66% due primarily to an increase is use of the “Credit History” DRC (from 25.24% to 35.58%) and the “Debt-to-income” DRC (from 25.45% to 27.70%).  The increase in applicant-related DRCs is, to some extent, due to decreases in the “Other” DRC.

4.Table3

The trend in the use of DRC “Other” – from 14.40% in 2010 to 9.59% in 2014 – is of interest in light of the provision in §1003.4(a)(16) that applications denied on the basis of “Other” include further details in free-form text field.  This requirement for further information is likely to reduce – if not eliminate – the number of applications denied on the basis of “Other”.

Table 3 highlights how DRC reporting rates can help lenders evaluate their lending profile through comparison of their denial activity to the broader market.  For example, while an individual lender demonstrating the trends in Table 3 might be subjected to questions about their lending profile, this lender-specific pattern in the context of market results such as those shown in Table 3 reduces the lender’s risk for adverse determinations regarding their lending profile[4].

The absence of DRCs makes it more difficult for lenders to benchmark their lending activity.  This fact suggests that lenders may be better served to include DRCs in their HMDA filings even if providing such information is not yet required.  An indication by a lender as to the basis for denial of an application will provide essential context to any regulatory determinations of a lender’s lending activity.

 **

Our next HMDA Insight provides insight into gender, race, and ethnicity reporting rates and discuss how this information can be used in evaluating lending activity.

***

 “It is a capital mistake to theorize before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts.” ― The Adventures of Sherlock Holmes, A Scandal in Bohemia

****

[1] Mortgage TrueView dashboards can be accessed by 2015 HMDA Survey participants using their complementary access credentials.  Non-participants may obtain credentials by sending a request to hmda@mortgagetrueview.com.

[2] This trend is confirmed by the results of Mortgage TrueView’s 2015 HMDA Survey which show that the 2015 DRC reporting rate was 67%, down only slightly from 2014’s 68%.

[3] See 12 C.F.R. 1003.4(a)(16).

[4] Benchmarking also allows a lender to identify and leverage comparative market advantages.  For an example of how benchmarking allows for a comparison of mortgage lenders, visit https://lenderscores.com


© 2016 Mortgage TrueView, Inc.  | All Rights Reserved.  

This material may not be used for any commercial purpose in whole or part without the express written permission of Mortgage TrueView, Inc.

 

 

HMDA Action Taken Rates: Adjusting for Denial Reason Code 7

Link to PDF Version

Summary:

  • Mortgage TrueView’s HMDA Insight Dashboards[1] show that the 10 largest lenders reporting Denial Reason Codes (“DRCs”) demonstrate over-weighted use of certain DRCs indicating an adjustment is required to properly state Incomplete Action Taken rates.

Chart 1 | DRC Distribution : 10 Largest DRC Reporting Lenders3.Chart1

Analysis:

Table 1 presents the DRC reporting profile for the 10 largest lenders reporting DRCs and indicates that two DRCs – Insufficient Cash and Credit Application Incomplete – are significantly over-weighted.

3.Table1

Further analysis indicates:

  • Both Insufficient Cash and Credit Application Incomplete DRCs are over-weighted for 2010, 2011, 2012 and 2013.
  • Insufficient Cash (Table 2) and Credit Application Incomplete usage rates (Table 3) are disproportionately used by approximately half of the 10 largest DRC reporting lenders.

3Table2

3Table3

The reason(s) for the highlighted rates – representing significant usage rates for the indicated Respondents – shown in Tables 2 and 3 cannot be fully substantiated based on available HMDA data.  However, research suggests that the rates may be due to operational protocols regarding the disposition of applications within regulatory timelines. This view is substantiated by the fact that the Incomplete Action Taken rates – an analogue of Credit Application Incomplete –  for the lenders highlighted in Table 3 are below the overall average rate.

This analogue prompts a critical question – how does an application denied on the basis of an incomplete credit application substantively differ from an application reported as Incomplete Action Taken?  Statutorily, an application’s Action Taken status is reported as Incomplete if (a) the lender sent a written notice of incompleteness to the applicant pursuant to Section 1002.9(c)(2) of Regulation B and (b) the applicant did not respond to the request.  Guidelines for the use of DRC 7 are less declarative and burdensome.

Despite the statutory guidelines, Credit Application Incomplete and Incomplete Action Taken are substantively the same and, as indicated in Table 3, reclassification of loans denied due to an incomplete credit application to Incomplete Action Taken is reasonable and necessary.  The impact of this reclassification is shown in Chart 2.

Chart 2 | Adjusted Denied and Incomplete Rates3.Chart2

* *

Our next HMDA Insight provides additional insight into how Denial Reason Code activity can enhance the evaluation of a lender’s lending profile.

* * *

It is a capital mistake to theorize before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts.” ― The Adventures of Sherlock Holmes, A Scandal in Bohemia

* * * *

[1] Mortgage TrueView Dashboards can be accessed by 2015 HMDA Survey participants using their complementary access credentials.  Non-participants may obtain credentials by sending a request to HMDA@mortgagetrueview.com.


© 2016 Mortgage TrueView, Inc.  | All Rights Reserved.

This material may not be used for any commercial purpose in whole or part without the express written permission of Mortgage TrueView, Inc.

 

 

HMDA Multi-Dimensional Dynamic Benchmarking

Link to PDF Version

Summary:

  • Dynamic benchmarking based on an advanced analytical business intelligence platform combines multiple dimensions – including geography, applicant profiles, and application volumes – to provide the most meaningful view of a lender’s compliance, risk, and strategic profile.

Analysis:

Our last HMDA Insight highlighted the impact of 2014 Loans Purchased on the St. Louis, MO-IL MSA.  Table 1 summarizes this MSA by presenting both gross (i.e., all reported) and net (i.e., net of Loans Purchased) applications.  This table provides the basis for a basic case study regarding multi-dimensional dynamic benchmarking.

2.Table1

Table 1 shows that:

  • Lenders with fewer than 100 applications account for approximately 80% of all lenders (and less than 10% of all applications) measured on both a gross and net basis.  
  • Lenders with more than 750 applications account for approximately 5% of all lenders (and approximately 66% of applications on a gross basis and 60% of applications on a net basis).

While it is common to evaluate a lender’s lending activity against all lenders in the relevant market, Table 2 presents a 2014 benchmark comprised of comparable application volumes.  A benchmark based on application volume enhances comparability as lenders of comparable volumes are likely more similar in terms of governance, risk management, and compliance protocols as well as operational and strategic profiles.  The Table 2 Benchmark, however, is problematic in view of the fact that the Respondents report 18,804 Loans Purchased accounting for approximately 75% of all Loans Purchased (which represents approximately 30% of the benchmark applications). 

2.Table2

Table 3 presents a benchmark comprised of the Top 20 Lenders based on Actioned Applications.  The Table 3 Benchmark replaces six of the eight lenders in the Table 2 Benchmark that reported significant levels of Loans Purchased (leaving Respondent 2 and 10 who report 2,973 and 1,028 Actioned Applications, respectively).

2.Table3

The “replacement respondents” reduce the number of Total Applications from 64,175 to 57,368 with Loans Purchased reduced from 18,804 to 8,007.  The number of Actioned Applications increases from 45,371 to 49,361.  

Five of the six replacement respondents are “in market” or market-based lenders, replacing lenders who are based in other markets as further “relevant” dimensionalization.  As shown in Table 4, the Table 3 benchmark is comprised of 11 in market-based lenders accounting for approximately 53% of the benchmark applications. 

2.Table4

Table 4 also shows that the Table 3 Benchmark can be further adjusted to more meaningfully evaluate top lenders in the St. Louis, MO-IL MSA by including all in-market lenders reporting more than 750 applications.  The resulting adjusted benchmark is comprised of 27 of the top 30 lenders (and 18 market-based lenders representing 67% of all benchmark lenders) accounting for approximately 58% of benchmark Actioned Applications.  The bottom-line is that comparing a lender’s in-market underwriting activity to other in-market underwriting activity results in the most meaningful benchmarking.

With the right advanced analytical Business Intelligence platform – Mortgage TrueView – the Adjusted Table 3 Benchmark can be further dynamically dimensionalized based on Denial Reason Codes, applicant attributes, and other HMDA data elements (and enterprise data elements) providing even more relevant insight into a lender’s lending profile. 

* *

Our next HMDA Insight will focus on Denial Reason Codes and the need for another adjustment to Action Taken Rates.

 * * *

“It is a capital mistake to theorize before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts.” ― The Adventures of Sherlock Holmes, A Scandal in Bohemia


© 2016 Mortgage TrueView, Inc.  | All Rights Reserved.

This material may not be used for any commercial purpose in whole or part without the express written permission of Mortgage TrueView, Inc.

 

 

HMDA Action Taken Rates : Adjusting for Loans Purchased

This is the first in a series entitled HMDA Insights : Capitalizing on New Perspectives prepared to provide mortgage lenders with advanced analytical insight to enhance governance, risk management and compliance activities in a challenging – and changing – environment.

Link to PDF Version

Summary:

  • Excluding Loans Purchased (i.e., Action Type 6) in evaluating a lender’s lending profile is necessary as the purchase decision is distinct from the origination decision.
  • Mortgage TrueView HMDA Insight Dashboards[1] show that differences between the aggregate Action Taken Rates[2] inclusive of all Action Taken codes (i.e., Gross Rate) and aggregate Action Taken Rates excluding Loans Purchased (i.e., Net Rate) are significant (Chart 1).

   Chart 1 | Comparative Action Taken Rates | 2010 – 2014 

Chart1

  • While the overall results are noteworthy, it is essential to evaluate – and understand – the results in the MSAs/MDs where a lender conducts business.

Analysis:

Insight1_Table1

Loans Purchased by all HMDA Respondents and the relationship of such applications to Actioned Applications (i.e., Action Taken Codes 1-5,7 and 8) (“Loans Purchase Rate”) are significant (Table 1).

Mortgage TrueView Insight Dashboards show MSA/MD-specific 2014 Loan Purchase Rates ranging from 1.42% to 37.41% (with the range for the top 20 MSAs based on Actioned Applications ranging from 13.30% to 26.57% as shown in Chart 2).  

 

      Chart 2 | Loans Purchased as a Percent of Actioned Applications [Top 20 MSAs/MDs] : 2014

Insight1_Chart2

Mortgage TrueView Insight Dashboards show that the St. Louis, MO-IL MSA has the highest 2014 Loan Purchase Rate – 26.57% – among the top 20 MSAs based on Actioned Applications.  Table 2 presents the Top 20 lenders (based on Total Applications) in the St. Louis, MO-IL MSA and shows their respective Loans Purchased Rate.  Table 2 shows that the St. Louis, MO-IL MSA Loans Purchased Rate of 26.57% is primarily attributable to 8 (highlighted) lenders that have Loans Purchased Rates significantly higher than the average Loans Purchased rate for the St. Louis, MO-IL MSA.

Insight1_Table2

Four of the eight lenders – Respondents 4, 8, 14, and 19 – show significant Loans Purchase Rates with Respondent 19 reporting 1 Actioned Application and 1,359 Loans Purchased.

Despite significant Loans Purchase Rates, two of the eight lenders – Respondents 2 and 10 – also show significant Actioned Applications, 3,973 and 1,028, respectfully.

It is clear from Table 2 that including Loans Purchased as Approved Applications significantly misstates the lending activity for the highlighted lenders.  By extension, including Loans Purchased misstates the lending activity for all lenders and provides an inaccurate view of lending activity for the overall market.

* *

Our next HMDA Insight will focus on how adjusting Action Taken rates makes a difference in establishing a more meaningful benchmark for purposes of evaluating lending activities.

* * *

“It is a capital mistake to theorize before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts.” ― The Adventures of Sherlock Holmes, A Scandal in Bohemia

* * * *

[1] Mortgage TrueView Dashboards can be accessed by 2015 HMDA Survey participants using their complementary access credentials.  Non-participants may obtain credentials by sending a request to hmda@mortgagetrueview.com.

[2] Approved includes loans originated, applications approved not accepted, and preapproval requests approved but not accepted.  Denied loans include applications denied by the financial institution and preapproval requests denied by financial institution.


© 2016 Mortgage TrueView, Inc.  | All Rights Reserved.

This material may not be used for any commercial purpose in whole or part without the express written permission of Mortgage TrueView, Inc.