An Updated Review of the New and Revised Data Points in HMDA

0

CONSUMER FINANCIAL PROTECTION BUREAU | AUGUST 2020

An Updated Review of the

New and Revised Data

Points in HMDA

Further Observations using the 2019 HMDA Data

1

This is another in an occasional series of publications from the Consumer Financial Protection

Bureau’s Office of Research. These publications are intended to further the Bureau’s objective of

providing an evidence-based perspective on consumer financial markets, consumer behavior,

and regulations to inform the public discourse. See 12 U.S.C. §5493(b).

1

This report was prepared by Feng Liu, Young Jo, Akaki Skhirtladze, and Laura Barriere.

 

Table of contents 
Table of contents ............................................................................................................ 2 
 Introduction .............................................................................................................. 4 
 Open-end and Reverse Mortgage Flags ............................................................... 9 
1  Open-end Line of Credit Flag ...................................................................9 
2  Reverse Mortgage Flag ........................................................................... 10 
3  Separating Reverse Mortgages from Forward Mortgages and Lines of 
Credit....................................................................................................... 11 
 Expanded and Revised Demographic Information ........................................... 15 
1  Age........................................................................................................... 15 
2  Expanded Race and Ethnicity Fields and Reporting of Disaggregated 
Categories................................................................................................ 17 
3  Visual Observation of Race, Ethnicity and Sex ...................................... 21 
 Property Type ......................................................................................................... 23 
 Loan Purpose and Characteristics ...................................................................... 24 
1  Business or Commercial Purpose Flag .................................................. 24 
2  Loan Purpose ......................................................................................... 26 
3  Loan Term .............................................................................................. 29 
4  Introductory Rate Period ........................................................................ 31 
5  Non-Amortizing Features ...................................................................... 34 
6  Prepayment Penalty Term ..................................................................... 36 
7  Submission of Application and Initially    Payable Flags ....................... 37 
 Applicant/Borrower and Property Characteristics ............................................ 41 
1  Occupancy Type ...................................................................................... 41 

 

2  Property Value ....................................................................................... 44 
3  Loan Amount and Conforming Loan Flag .............................................. 45 
4  Credit Score ............................................................................................ 46 
5  CLTV ....................................................................................................... 53 
6  DTI .......................................................................................................... 57 
7  Manufactured Home Secured Property Type ........................................ 60 
8  Manufactured Home Land Property Interest ....................................... 62 
9  Number of Affordable Units for Multifamily Loan ............................... 63 
 Pricing Outcomes and Components ................................................................... 65 
1  Interest Rate ...........................................................................................65 
2  Rate Spread ............................................................................................ 70 
3  Total Loan Costs or Total Points and Fees ............................................. 72 
4  Origination Charges ................................................................................ 75 
5  Discount Points and Lender Credits.......................................................78 
 Miscellaneous ........................................................................................................ 83 
Appendix A:  Tables .................................................................................................. 85 
Appendix B:  Figures .............................................................................................. 252 
 

4

1. Introduction

The Home Mortgage Disclosure Act (HMDA) is a data collection, reporting, and disclosure

statute that was enacted in 1975. HMDA data are used to assist in determining whether financial

institutions are serving the housing needs of their local communities; facilitate public entities’

distribution of funds to local communities to attract private investment; and help identify

possible discriminatory lending patterns.

2

Institutions covered by HMDA are required to

annually collect and report specified information about each mortgage application acted upon

and mortgage purchased during the prior calendar year.

3

The data include the disposition of

each application for mortgage credit; the type, purpose, and characteristics of each home

mortgage application or purchased loan; the census-tract designations of the properties; loan

pricing information; demographic and other information about loan applicants, including their

race, ethnicity, sex, and income; and information about loan sales.

4

In the Dodd-Frank Wall Street Reform and Consumer Protection Act of 2010 (DFA), Congress

amended HMDA to require the reporting of 13 new data points (Mandated Data Points): Age;

Total Points and Fees; Rate Spread for all loans; Prepayment Penalty Term; Property Value;

Introductory Rate Period; Non-Amortizing Features; Loan Term; Application Channel; Credit

Score; Mortgage Loan Originator Identifier; Universal Loan Identifier; and Property Address.

5

The DFA also granted the Bureau authority to use its discretion to require reporting of

additional data points.

In 2015, the Bureau issued a rule (2015 HMDA Rule) amending Regulation C, HMDA’s

implementing regulation, to include new data points. The 2015 HMDA Rule included the

Mandated Data Points discussed above. The 2015 HMDA Rule also included 14 additional data

points the Bureau issued pursuant to its discretionary authority under the DFA (Discretionary

Data Points): Origination Charges; Discount Points; Lender Credits; Mandatorily Reported

Reasons for Denial; Interest Rate; Debt-to-Income Ratio; Combined Loan-to-Value Ratio;

Manufactured Home Secured Property Type; Manufactured Home Land Property Interest;

2

For a brief history of HMDA, see Federal Financial Institutions Examination Council, “History of HMDA,” available

at www.ffiec.gov/hmda/history2.htm.

3

The 2019 HMDA data, which are the subject of this Data Point article, cover mortgage applications acted upon and

mortgages purchased during calendar year 2019.

4

See https://s3.amazonaws.com/cfpb-hmda-public/prod/help/2019-hmda-fig.pdf for a full list of items reported

under HMDA for 2019.

5

With respect to the last three listed data points, the DFA states that these shall be reported “as the Bureau may

determine to be appropriate.”

5

Multifamily Affordable Units; Automated Underwriting System; Reverse Mortgage Flag; Open-

End Line of Credit Flag; and Business or Commercial Purpose Flag.

The 2015 HMDA Rule also revised several preexisting data points. Among other changes, the

2015 HMDA Rule replaced Property Type with Construction Type and Total Units, added two

enumerations (“cash-out refinance” and “other purpose”) to Loan Purpose, and split the “non-

owner occupied” category of Occupancy Type into “second residence” and “investment

property.” In addition, under the 2015 HMDA Rule, applicants have the option to self-identify

their race/ethnicity in disaggregated sub-categories (for example, Indian or Chinese are sub-

categories under Asian) and financial institutions must report such detail, where applicable.

Financial institutions must also report, where applicable, whether the race, ethnicity, and sex of

applicants were collected based on visual observation or surname.

Finally, the 2015 HMDA Rule made changes in Regulation C’s coverage requirements. First,

reporting of open-end lines of credit became mandatory for reporters that meet certain loan

volume thresholds. Second, the transactional-coverage definition eliminated the previous

requirement to report unsecured loans made for home improvement purposes and now requires

reporting of consumer-purpose loans secured by a dwelling even if not made for one of the

previously enumerated purposes.

In May 2017, Congress passed the Economic Growth, Regulatory Relief, and Consumer

Protection Act (EGRRCPA) that granted certain HMDA reporters partial exemptions from

HMDA reporting. Under the partial exemptions, these institutions are not required to report any

of the Mandated Data Points other than age and are not required to report any of the

Discretionary Data Points for eligible transactions. Specifically, HMDA reporters that are

insured depository institutions or insured credit unions and that originated fewer than 500

closed-end mortgages in each of the two preceding years qualify for this partial exemption with

respect to reporting their closed-end transactions. HMDA reporters that are insured depository

institutions or insured credit unions that originated fewer than 500 open-end lines of credit in

each of the two preceding years also qualify for this partial exemption with respect to reporting

their open-end transactions. The insured depository institutions must also not have received

certain less than satisfactory examination ratings under the Community Reinvestment Act of

1977 (CRA) to qualify for the partial exemptions. The Bureau issued an interpretive rule in 2018

to clarify which transactions and which data points are covered by the partial exemptions.

6

In particular, the interpretive rule clarifies that Denial Reasons -- which had been an optional data point and was

made mandatory by the 2015 HMDA Rule -- reverts to an optional data point for partially-exempt transactions and

that institutions are not required to report Rate Spread -- which previously had been required with respect to certain

loans -- with respect to any partially exempt transactions.

6

As a result of all these changes, starting with the data collected in 2018 and reported in 2019, the

HMDA data differ significantly from the HMDA data of previous years both in terms of the

applications and loans reported and the data points required with respect to those applications

and loans. The Filing Instructions Guide (FIG) for HMDA Data Collected in 2019 provides

specifications for the new data points, some of which are reported under multiple data fields.

7

With respect to the public disclosure of HMDA data, in the 2015 HMDA Rule the Bureau

interpreted HMDA, as amended by the DFA, to require that the Bureau use a balancing test to

determine whether and how HMDA data should be modified prior to its disclosure to protect

applicant and borrower privacy while also fulfilling HMDA’s public disclosure purposes. In

December 2018, the Bureau issued final policy guidance (Policy Guidance) describing

modifications the Bureau intended to apply to the HMDA data before the Bureau, on behalf of

the FFIEC, made the data available to the public on the loan level.

8

The Bureau has announced

that it intends to address these privacy and disclosure issues through a legislative rulemaking,

which will provide the Bureau with an opportunity to reconsider the Policy Guidance following

notice and comment.

In accordance with this Policy Guidance, the following data fields are excluded from the 2019

public loan-level HMDA data: Universal-Loan-Identifier or Non-Universal-Loan-Identifier;

Application Date; Action Taken Date; Property Address; Credit Score Relied On in Making the

Credit Decision; Mortgage Loan Originator Nationwide Mortgage Licensing System and Registry

(NMLSR) identifier; Result Generated by the Automated Underwriting System; Free-form Text

Fields for Race, Ethnicity, Name and Version of Credit Scoring Model, and Reason for Denial;

and Name of the Automated Underwriting System. The Bureau also modified the public loan-

level 2019 HMDA data to reduce the precision of most of the values reported for the following:

Loan Amount; Age; Debt-to-Income Ratio; Property Value; Total Units; and Multifamily

Affordable Units.

In August 2019, the Bureau issued a Data Point article titled “Introducing New and Revised Data

Points in HMDA – Initial Observations Based on New and Revised Data Points in 2018 HMDA

Data.” The goals of that Data Point article were to introduce the new and revised data points in

the 2018 HMDA data and to provide some initial observations about the nation’s mortgage

market in 2018 based on those new and revised data points. Because 2019 is only the second

year those new and revised data points were collected under the 2015 HMDA Rule, it is possible

7

Available at https://s3.amazonaws.com/cfpb-hmda-public/prod/help/2019-hmda-fig.pdf Available at

https://s3.amazonaws.com/cfpb-hmda-public/prod/help/2019-hmda-fig.pdf

8

Available at

https://files.consumerfinance.gov/f/documents/HMDA_Data_Disclosure_Policy_Guidance.Executive_Summary.FI

NAL.12212018.pdf

7

that they are still relatively new to the public and many HMDA data users. Therefore, the

Bureau is issuing this article by building on the last article, with updates using the 2019 HMDA

data. The goal of this article is to help the public become more familiar with the new and revised

data points in the 2019 HMDA data and to provide some initial observations about the nation’s

mortgage market in 2019 using those new and revised data points.

The information contained in this article is not intended to be in-depth and comprehensive, but

rather offered as an initial set of findings from the 2019 HMDA data. Through this exercise, the

Bureau hopes to provide the public with a roadmap for the new HMDA data, as researchers,

government agencies, community groups, financial institutions, and others may use these new

data for various other purposes.

As in last year’s article, the focus of this article is on cross-sectional analyses, i.e. using the data

contained in one year’s loan application registrar (LAR) to explore various patterns and

relationships between different data fields to provide some initial observations. To the extent

some of those patterns or relationships might have changed significantly over the last year, this

article will highlight such changes in comparison to the observations from last year’s article.

Otherwise, the majority of the analyses in this article are limited to the data collected in 2019

and reported in 2020.

In tandem, the Bureau published a Data Point article titled “2019 Mortgage Market Activity and

Trends” on June 24th

9

, together with the publication of the static loan-level 2019 HMDA data

file that consolidates data from individual reporters.

10

That Data Point article mainly focuses on

historical comparison of trends and patterns using the data points that have been collected and

reported before the significant changes to HMDA reporting requirements took place starting

with the 2018 HMDA data. The Bureau views that article and this current article, as well as the

“Introducing New and Revised Data Points in HMDA – Initial Observations Based on New and

Revised Data Points in 2018 HMDA Data” article from last year as complementary. In a way,

this article not only updates last year’s article with the 2019 data, it also provides updates to the

June article by focusing on new and revised data points and cross-sectional analyses.

Both this article and the June Data Point article use the static loan-level 2019 HMDA data file

published on June 24th. This data file reflects modifications to the reported HMDA data to

protect applicant and borrower privacy. This data file and the two Data Point articles reflect the

data as of April 27, 2020. Though this static, consolidated loan-level file will not be changed, the

Bureau will separately provide updates to the consolidated loan-level 2019 HMDA data

9

Available at https://www.consumerfinance.gov/data-research/research-reports/data-point-2019-mortgage-market-

activity-and-trends/

10

Available at https://ffiec.cfpb.gov/data-publication/snapshot-national-loan-level-dataset/2019

8

(dynamic data) to reflect any resubmissions or late submissions. In addition, this article uses

some non-public HMDA data in its analysis and findings. Therefore, the results using the

dynamic data may differ from those reported in this Data Point article. However, we expect that

the dynamic data will produce substantially consistent results.

As the Bureau has acknowledged, collecting and reporting the new and revised data points

posed significant systems and operational challenges.

11

Consequently, while the Bureau has

taken customary steps to ensure the accuracy of the data presented in this article and released to

the public, such as excluding data that likely contain errors, there may be some anomalies and

non-material errors in the 2019 HMDA data.

For exposition purpose, the article groups the new and revised HMDA data points into seven

major categories: Open-end and Reverse Mortgage Flags; Expanded or Revised Demographic

Information; Property Type; Loan Purpose and Characteristics; Applicant/Borrower Credit

Characteristics and Property Characteristics; Pricing Outcome and Components; and Others.

These groupings, though natural from the perspective of most data users, do not reflect any

regulatory requirements.

12

The remainder of this article is organized as follows: For each grouping, we will discuss each

new or revised data point. For each data point, we will first explain the definition, basic

reporting requirements, and allowable enumeration or values under the 2015 HMDA Rule and

2019 FIG. We will also note any modifications applied to the data point before public disclosure

of the loan-level 2019 HMDA data. The article then provides some basic observations using the

2019 data. Where appropriate, the article will provide context to help data users better

understand the limitations of such data points, especially if one or a few data points are to be

used in isolation. Although this article is structured to introduce each new or revised data point

in a specified order, in many instances the interaction of multiple data points is examined prior

to the formal introduction of some of the data points. In such instances, readers can refer to the

formal definition of the not-yet-introduced data points in later sections.

11

See Bureau of Consumer Fin. Prot., “CFPB Issues Public Statement On Home Mortgage Disclosure Act Compliance”

(Dec. 21, 2017), https://www.consumerfinance.gov/about-us/newsroom/cfpb-issues-public-statement-home-

mortgage-disclosure-act-compliance/.

12

It is also possible that different HMDA data users or readers of this article may find different ways of grouping the

new/revised HMDA data points that are more relevant to them. Again, the grouping in this article is for exposition

purpose only and is entirely non-binding.

9

2. Open-end and Reverse

Mortgage Flags

2.1 Open-end Line of Credit Flag

The 2015 HMDA Rule changed the reporting of open-end lines of credit (LOC) from optional to

mandatory. Specifically, institutions that originated at least 100 open-end LOCs in each of the

two preceding calendar years and met other reporting criteria would have been required to

report data on open-end LOCs beginning with data collected in 2018 and reported in 2019. In

2017, the Bureau temporarily increased the open-end LOC reporting threshold to 500 for

calendar years 2018 and 2019.

13

The 2015 HMDA Rule also added a new data point consisting of

a flag for open-end LOCs to distinguish these from closed-end mortgage records. The open-end

LOC Flag is one of the Discretionary Data Points discussed in the introduction section of this

article. The open-end LOC flag is among the data points that institutions that qualify for an

EGRRCPA partial exemption are not required to report. It has an allowable value of 1 for “open-

end line of credit,” 2 for “not an open-end line of credit,” and 1111 for “exempt.”

In the 2019 HMDA data, 959 financial institutions reported about 2.17 million LARs for open-

end LOCs

14

, down from 1029 reporters with about 2.33 million open-end LOC LARs in 2018.

The total number of applications for open-end LOCs was about 2.14 million, including about

1.07 million associated originations for which the open-end LOC flag is reported to be 1. In

comparison, the number of open-end LOC originations reported was about 1.15 million in 2018.

13

In October 2019, the Bureau published a final rule that extended this temporary threshold of 500 open-end LOCs

through the 2020 and 2021 calendar years. In April 2020, the Bureau published a final rule that set the permanent

open-end LOC reporting threshold at 200 open-end LOCs in each of the two preceding years, beginning in calendar

year 2022. These changes do not affect the data collected in 2019 and analyzed in this report.

14

Note that under the EGRRPA the open-end LOC itself is among the partially-exempt data points. If a financial

institution that reports its open-end LOC transactions is eligible for partial exemption for its open-end LOC

transactions, such transactions would not be identified as open-end LOC in LAR if it chooses to report its open-end

LOC flag as “1111” using the partial exemption. Per the 2018 HMDA Rule, a financial institution was required to

report its open-end LOC transactions in 2019, if it originated more than 500 open-end LOCs per year in both 2017

and 2018 (assuming it also met other reporting criteria). Meanwhile, under the EGRRPA, a financial institution was

eligible for partial exemption of its open-end LOC transactions in 2019 if it originated less than 500 open-end LOCs

in each of 2017 and 2018. Given the large overlap between the open-end LOC reporting threshold and partial-

exemption criteria, the Bureau believes that most of the open-end transactions reportable under the Regulation C in

2019 were not eligible for partial-exemption. Consequently, the number of open-end LOC transactions reported but

not affirmatively flagged as open-end due to the partial exemption in the 2019 HMDA data is likely immaterial for the

purpose of the analyses presented in this article.

10

Table 2.1.1 lists the top 25 open-end LOC lenders by origination volume in 2019, their institution

type, number of open-end applications, number of open-end originations, number of open-end

purchased loans, assets, and their respective market share in terms of their reported open-end

originations relative to the total volume of open-end originations in the 2019 HMDA data. In

total, the top 25 open-end lenders accounted for about 573,000 open-end originations, or 53.6

percent of all open-end originations reported under HMDA. All the top 25 open-end lenders are

depository institutions or credit unions with the exception of one non-depository institution that

specializes in reverse mortgages. The two largest open-end reporters in 2018 HMDA data, Bank

of America and Wells Fargo, remained the top two open-end lenders in 2019 data, and most

other top open-end reporters from 2018 data remained in the top 25 list as well.

Table 2.1.2 breaks down the open-end LOC reporters by size category. Overall, 885 HMDA

reporters reported at least one open-end LOC origination, compared to 956 reporters that

reported open-end LOC originations in 2018. Specifically, 413 reporters originated fewer than

100 open-end LOC originations, 54 reporters originated between 100 and 199, and 118 reporters

originated between 200 and 499 open-end LOCs.

15

Together, the 585 reporters with an open-

end origination volume below 500 accounted for about 57,000 originations, or 5.4 percent of all

reported open-end originations. Under the open-end reporting threshold established by the

2019 HMDA Rule, which extended through 2021 the temporary threshold of 500 open-end

LOCs in each of the preceding two calendar years, those 585 reporters would not be required

under Regulation C to report their 2020 open-end lending activities.

2.2 Reverse Mortgage Flag

The 2015 HMDA Rule added a data point that flags whether the loan or application is for a

reverse mortgage. The reverse mortgage flag is one of the data points that institutions that

qualify for the EGRRCPA partial exemption are not required to report. It has an allowable value

of 1 for “reverse mortgage,” 2 for “not a reverse mortgage,” and 1111 for “exempt.”

In the 2019 HMDA data, 151 financial institutions reported approximately 75,500 reverse

mortgage LARs, down from 168 reverse mortgage reporters with 90,300 LARs in 2018. The total

number of applications for reverse mortgages is about 55,100, including approximately 34,800

15

Note that the temporary HMDA open-end reporting threshold of 500 originations is based on the origination

volumes for the two years preceding the HMDA activity year. Specifically, for the 2019 HMDA data that are collected

in 2019 and reported in 2020, a lender would be required to report its open-end lending activity if it originated at

least 500 open-end LOC in both 2017 and 2018, assuming it also met other reporting criteria. Therefore, it is possible

that some lenders with open-end LOC origination volume exceeding 500 in both 2017 and 2018 originated fewer than

500 open-end LOC in 2019 but were nevertheless required to report the 2019 data under the HMDA reporting

requirements. On the other hand, it is also possible that some of the reporters opted to report their open-end lending

activities even though they were not required to report.

11

reverse mortgage originations with a flag value of 1, up from about 33,000 reported reverse

mortgage originations in 2018.

Table 2.2.1 lists the top 10 reverse mortgage lenders by origination volume in 2019, their

institution type, applications, originations, purchased loans, assets, and their market share in

terms of their reported reverse mortgage originations relative to the total volume of reverse

mortgage originations in 2019 HMDA data. In total, the top 10 reverse mortgage lenders

accounted for approximately 31,700 reverse mortgage originations, or 90.9 percent of all reverse

mortgage originations reported under HMDA, up from 84.5 percent in 2018.

Table 2.2.2 breaks down the reverse mortgage reporters by size category. Overall, for 2019, 112

HMDA reporters reported at least one reverse mortgage origination, and 90 reported fewer than

100 reverse mortgage originations.

2.3 Separating Reverse Mortgages from

Forward Mortgages and Lines of Credit

Table 2.3.1 cross-tabulates the reported values for the Reverse Mortgage Flag against the

reported values of the open-end Flag for the 2019 HMDA data. As shown in the table, about 77.7

percent of reverse mortgage originations are structured as open-end LOCs and 22.3 percent are

closed-end. Similarly, about 74.7 percent of all reverse mortgage LARs are structured as open-

end and 25.3 percent are closed-end.

Reverse mortgages are different from traditional forward mortgages and LOCs in terms of their

intended purpose, characteristics, and customer base. Therefore, the remainder of this article

separates reverse mortgages from other forward transactions by grouping all LARs into three

transaction types: closed-end mortgages excluding reverse mortgages, open-end LOCs excluding

reverse mortgages, and reverse mortgages. The closed-end mortgages excluding reverse

mortgages are transactions with an open-end LOC flag reported as 2 (not an open-end LOC) and

reverse mortgage flag reported as 2 (not a reverse mortgage). The open-end LOCs excluding

reverse mortgages are transactions with an open-end LOC flag reported as 1 (open-end LOC)

and reverse mortgage flag reported as 2 (not a reverse mortgage). Reverse mortgages are

transactions with a reverse mortgage flag reported as 1 (reverse mortgage).

Open-end LOCs secured by dwellings (excluding reverse mortgages) are commonly known as

home equity lines of credit, or HELOCs. Due to the partial exemption granted under the

EGRRCPA, about 470,000 LARs have either the open-end flag or reverse mortgage flag reported

as 1111 (Exempt), up from about 403,000 such records in 2018. For most of the discussion

12

regarding transaction types, we have not included those records reported as exempt. They

account for only a small fraction of all LARs.

Table 2.3.2 shows the distribution of transaction type by action taken for closed-end mortgages,

HELOCs, and reverse mortgages of LARs.

16

The denial rate for HELOC applications is significantly higher than for closed-end mortgages.

Excluding purchased loans, preapprovals, and applications that are withdrawn or closed for

incompleteness, the denial rate for HELOC applications in the 2019 HMDA data was about 41.7

percent. In comparison, the denial rate for closed-end mortgage applications was 17.1 percent.

17

About 27.1 percent of the reverse mortgage records reported under HMDA are purchased loans,

and none of the reverse mortgage records have a code indicating a preapproval request denied

or preapproval request approved but not accepted. In total, there were about 34,800 reverse

mortgage originations reported. The denial rate for reverse mortgage applications, excluding

purchased loans, preapprovals, and applications that are withdrawn or closed for

incompleteness, was about 17.3 percent.

Table 2.3.3 shows the distribution of closed-end, HELOC, and reverse mortgage originations by

race/ethnicity, neighborhood income, and geography.

18

The table indicates that HELOC

borrowers are more likely than closed-end borrowers to be non-Hispanic White, be in high-

income tracts, and live in metropolitan areas. In particular, 72.1 percent of HELOC borrowers

are non-Hispanic Whites, compared to 59.4 percent for closed-end mortgage borrowers.

Approximately 11.7 percent of HELOC borrowers live in low- or moderate-income census tracts,

compared to 16.7 percent of closed-end borrowers. In addition, 46.4 percent of HELOC

borrowers live in high-income tracts, compared to 39.3 percent of closed-end borrowers. A

slightly higher percentage of HELOC borrowers live in a metropolitan statistical area (90.9

16

For brevity, we have removed the phrase “excluding reverse mortgage” from “closed-end mortgage” and “open-end

LOCs” from this point on. Unless it is specifically stated otherwise, for the rest of the article, “closed-end mortgages”

refer to “closed-end mortgages excluding reverse mortgages” and “HELOCs” refer to “open-end LOCs excluding

reverse mortgages.”

17

Only 0.8 percent of all 2.12 million HELOC records reported under HMDA are purchased loans, and none of the

HELOC records contain an indication for a preapproval request denied or preapproval request approved but not

accepted.

18

Note that in Table 2.3.3 the sums of total originations across the neighborhood income rows and the sums across

geography rows are slightly smaller than the sums across “borrower race and ethnicity” rows, because there are a

small percentage of records that did not report census tracts and hence for which we could not assign the

neighborhood income category. Similarly, there are a small percentage of records that did not report county or state

code, therefore, we could not determine whether they are in a metropolitan statistical area, micropolitan statistical

area, or rural area. Such records are omitted in relevant tabulation accordingly. In general, within this article, total

sample size may vary across tables because of differences in sample universe and in missing values across data points.

For more information, see the note section of each table.

13

percent), compared to 90.1 percent of closed-end borrowers. Only 3.1 percent of HELOC

borrowers are in rural areas, slightly lower than that of closed-end borrowers (3.8 percent).

19

Non-Hispanic Whites make up a higher percentage of 2019 reverse mortgage borrowers (74.7

percent) than they do of closed-end or HELOC borrowers. The share of reverse mortgages in

low- or moderate-income tracts is higher than for closed-end and for HELOC borrowers. The

rural share for reverse mortgages is slightly higher than for HELOCs but lower than for closed-

end mortgages.

As shown in Table 2.3.4, all the HELOC records reported their loan type as conventional. An

overwhelming majority of reverse mortgage originations (90.3 percent) and purchased loans

(96.7 percent) reported their loan type as Federal Housing Administration (FHA) insured,

because the Home Equity Conversion Mortgage (HECM) insured by FHA is the dominant

product in the reverse mortgage market. There are about 3,400 conventional reverse mortgage

originations (up from about 2000 in the 2018 data) and 676 conventional reverse mortgage

purchased loans reported, representing the niche non-HECM reverse mortgage products.

Combining the transaction type (closed-end, open-end, reverse mortgage), loan type reported

under HMDA (conventional, FHA, VA, RHS/FSA)

20

, conforming loan status based on loan

amount reported, and the conforming loan limits published by the Federal Housing Finance

Agency (FHFA), all single family LARs can be grouped into seven categories: 1) Conventional

Conforming; 2) Conventional Non-conforming or Jumbo; 3) FHA; 4) VA; 5) RHS/FSA; 6)

HELOC; and 7) Reverse Mortgage. These categories are referred to as “Enhanced Loan Type” in

the rest of this article. The conventional conforming loan is a closed-end forward mortgage (i.e.

excluding reverse mortgage) transaction whose loan type is reported as conventional and whose

loan amount is below the conforming loan limit, making it eligible to be purchased by Fannie

Mae or Freddie Mac (collectively known as Government Sponsored Enterprises, or GSEs). The

conventional non-conforming, or jumbo loan is a closed-end forward mortgage transaction with

its loan type reported as conventional and a loan amount above the conforming loan limit,

making it ineligible to be purchased by the GSEs. The FHA, VA, and RHS/FSA loans follow the

definition of loan types under HMDA, and are restricted to closed-end loans excluding reverse

mortgages. HELOCs are forward open-end LOC transactions, regardless of their Loan Type

reported under HMDA. Reverse mortgages are transactions identified as reverse mortgages,

19

In this article, rural areas are defined as areas that are located outside of any metropolitan statistical area or

micropolitan statistical area.

20

Conventional means “not insured or guaranteed by FHA, VA, RHS, or FSA”; FHA stands for Federal Housing

Administration insured; VA stands for Veteran Affairs guaranteed; RHS or FSA stands for USDA Rural Housing

Service or Farm Service Agency guaranteed.

14

regardless of its reported Loan Type or whether they are reported as an open-end or closed-end

transaction.

Table 2.3.5 shows the number of originations, mean and median income of borrowers, mean

and median loan amounts, percentage of originations that are for home purchase, percentage of

originations that are for refinance, and percentage of originations that are secured by first lien

for each enhanced loan type. HELOC borrowers generally have a higher income than borrowers

of all other enhanced loan types other than jumbo loans. The median income of HELOC

borrowers is approximately $107,000 and their mean income is approximately $149,000. In

contrast, the median income of reverse mortgage borrowers is approximately $28,000, and the

mean is approximately $32,000, which are the lowest among borrowers of all enhanced loan

types, perhaps reflecting the unique design of reverse mortgages to help income-constrained

seniors convert home equity into cash income.

The median loan amount for HELOCs (for which the loan amount is defined as the maximum

amount HELOC borrowers can draw) is approximately $75,000, and the mean is approximately

$113,000, lower than the loan amount of all other enhanced loan types. The median loan

amount for reverse mortgages is about $163,000, and the mean is $226,000.

About 7.2 percent of HELOC originations had a loan purpose reported as home purchase, and

28.3 percent of HELOC originations are secured by a first lien. All reverse mortgages are

secured by a first lien. About 6.7 percent of reverse mortgage originations had loan purpose

reported as home purchase.

21

Known as Home Equity Conversion Mortgage (HECM) for Purchase insured by the HUD, some reverse mortgages

allow seniors, age 62 or older, to purchase a new principal residence using loan proceeds from the reverse mortgage.

15

3. Expanded and Revised

Demographic Information

The DFA and 2015 HMDA Rule added or revised a number of data fields and data points to

gather additional demographic information regarding applicants and borrowers. Demographic

information now includes age, race, ethnicity, and sex.

3.1 Age

The DFA, as implemented by the 2015 HMDA Rule, added age as a new data point that

institutions must report. Age is one of the Mandated Data Points as discussed in the

introduction section of this article. Where applicable, the age of the applicant or borrower, or

age of the first co-applicant or co-borrower, is to be reported in years. The EGRRCPA’s partial

exemptions from reporting certain data points for certain transactions do not apply to age, that

is, institutions eligible for a partial exemption under the EGRRCPA must still report age for all

covered transactions. Age is binned into the same ranges in the public loan-level 2019 HMDA

data as in the public 2018 HMDA data: below 25, 25 to 34, 35 to 44, 45 to 54, 55 to 64, 65 to 74,

and 75 and above. In addition, the publicly released data contain a flag indicating whether the

reported age is 62 years or older.

Table 3.1.1 shows the age distribution of mortgage borrowers for each enhanced loan type.

22

The

median age of all borrowers in the 2019 HMDA data is 45 years, one year younger than the

median age of all borrowers in 2018 HMDA data, and the mean is 46.5. RHS/FSA loan

borrowers tend to be much younger than other borrowers, with a median age of 31, and a mean

age of 34.3. In addition, 19.1 percent of RHS/FSA borrowers are younger than 25 and 42.9

percent of RHA/FSA borrowers are between 25 and 34 years old. Not surprisingly, the reverse

mortgage borrowers are much older than borrowers with other loan types. The median age of

reverse mortgage borrowers is 73, and the mean is 74.1. For reverse mortgage borrowers, 9.7

percent are between the ages of 55 and 64, 46.0 percent are between 65 and 74, and 44.4

percent are 75 or older. The median borrower age for both conventional conforming and jumbo

loans is 45, but a slightly larger percentage of conventional conforming loan borrowers belong to

the youngest age bins relative to the jumbo loan borrowers. Specifically 2.7 percent of

22

There are two separate age fields starting in the 2018 HMDA data: the age for borrower/applicant, and the age for

co-borrower/co-applicant. For brevity of explanation, throughout this article we have only used the age of

borrower/applicant.

16

conventional conforming borrowers are younger than 25, and 21.8 percent are between 25 and

34. In contrast, only 0.1 percent of jumbo borrowers are younger than 25 and only 13.4 percent

of jumbo loan borrowers are between 25 and 34. The median age for FHA borrowers is 39 (6

years younger than the median age of conventional loan borrowers), and their mean age is 41.1.

The median age for VA loan borrowers is 46, and their mean age is 47.8. The median age of

HELOC borrowers is 54 and their mean age is 54 as well. Overall, the age profile of HELOC

borrowers is older than that of closed-end mortgage borrowers, though still younger than

reverse mortgage borrowers.

Table 3.1.2 shows the age distribution of mortgage borrowers (excluding reverse mortgages) by

race and ethnicity.

23

The median age of Hispanic White borrowers is 41 and their mean age is

42.8, making them on average the youngest group of borrowers among the listed race/ethnicity

groups. For Hispanic White borrowers, 4.6 percent are younger than 25, and 25.7 percent are

between 25 and 34 years old. Asian borrowers are the second youngest group, with a median age

of 42 and a mean age of 43.1. The median age of Black borrowers is 46 in 2019, one year younger

than the median age of Black borrowers in 2018, with a mean age of 47.5, the same as last year.

The median age of non-Hispanic White borrowers is also 46, similarly one year younger than

last year, with a mean age of 47.2.

Table 3.1.3 restricts the sample to closed-end mortgages excluding reverse mortgages and with

loan purpose limited to home purchase, and shows the age profile of borrowers by race and

ethnicity. Among the borrowers of closed-end home-purchase mortgages, 7.1 percent of

Hispanic White borrowers are younger than 25 and 33.2 percent are between 25 and 34 years

old, with the median Hispanic White borrowers’ age at 38. Overall the age distribution of

Hispanic White borrowers for home-purchase loans is younger than all other racial/ethnic

groups. Also, the median age of Black home-purchase loan borrowers is 41, and the median age

of non-Hispanic White home-purchase loan borrowers is 39. While 29.9 percent of Black

23

Consistent with the approach taken in the past Federal Reserve Board HMDA Bulletins and the past CFPB Data

Point Articles on HMDA data, throughout this article, with the exception of Sections 3.2 and 3.3, applications are

placed in one category for race and ethnicity. To keep the historical consistency, only the first digit of the reported

race and ethnicity, and only the first ethnicity reported in 2018 HMDA data, are used. The application is designated as

“Joint” if one applicant was reported as White and the other was reported as one or more minority races or if the

application is designated as White with one Hispanic applicant and one non-Hispanic applicant. If there are two

applicants and each reports a different minority race, the application is designated as two or more minority races. If

an applicant reports two races and one is White, that applicant is categorized under the minority race. Otherwise, the

applicant is categorized under the first race reported. "Missing" refers to applications in which the race of the

applicant(s) has not been reported, or is not applicable, or the application is categorized as White, but ethnicity has

not been reported. The “Other” group consists of applications by American Indians or Alaska Natives, Native

Hawaiians or other Pacific Islanders, and borrowers reporting two or more minority races.

17

borrowers who take out a home-purchase loan are younger than 35, 37.2 percent of non-

Hispanic White home-purchase loan borrowers are below 35.

Figure 3.1.1 depicts a binscatter plot of denial rate by applicant age of different enhanced loan

types, restricted to single-family, owner-occupied, first-lien applications with action taken code

values equal to 1 (originated), 2 (approved but not accepted), or 3 (denied).

24

The denial rate for

HELOCs generally decreases with age, except at its oldest groups as depicted in Figure 3.1.1. The

denial rates of most closed-end enhanced loan types generally are upward sloping with age, with

the exception of RHS/FSA loans that become more or less flat for applicants of older age groups.

The youngest age group also tends to have higher denial rates than the age groups that are

slightly older, as shown by the up-ticking left tails for most of the closed-end enhanced loan

types in Figure 3.1.1.

Table 3.1.4 shows the denial rates by enhanced loan type for applicants aged 62 or older. The

denial rates for applicants aged 62 or older are higher than the denial rates for applicants

younger than 62 for all enhanced loan types other than for HELOCs and reverse mortgages. It is

important to note that Figure 3.1.1 and the denial rates shown in Table 3.1.4 do not control for

any credit characteristics. Subsequent sections will examine how some credit characteristics of

applicants and borrowers vary with age.

3.2 Expanded Race and Ethnicity Fields and

Reporting of Disaggregated Categories

The new HMDA data include expanded reporting of race and ethnicity to allow for more detailed

categories. In the past, ethnicity was reported under one field for applicants and co-applicants,

whereas in the new HMDA data it is reported with up to five fields. Additionally, multiple free-

form text fields were added to allow applicants to provide and reporters to fill in race and

ethnicity of applicants and co-applicants that are not included among the standard

enumerations. Free-form text fields used to report race and ethnicity are excluded from the

public loan-level 2019 HMDA data. The EGRRCPA’s partial exemptions from reporting certain

data points for eligible transactions do not apply to race and ethnicity.

24

Binscatter plots are a convenient way of observing the relationship between two variables, especially useful when

working with large datasets, such as the entire HMDA LAR data. To generate a binned scatterplot, binscatter groups

the x-axis variable into equal-sized bins, computes the mean of the x-axis and y-axis variables within each bin, then

creates a scatterplot of these data points. The equal-sized bins are calculated for each enhanced loan type separately

for Figure 3.1.1.

18

Sections 3.2 and 3.3 focus on how the race and ethnicity fields were reported in the 2019 HMDA

data. These two sections present the 2019 data as they were reported by financial institutions.

Therefore, the presentation differs from how race and ethnicity are categorized in the rest of this

article, as well as in previous HMDA reports published by the Federal Reserve Board and the

Bureau, which combine certain race and ethnicity categories for brevity of exposition.

25

For

consistency and simplicity, the rest of this article uses the same aggregate race and ethnicity

categories that were used in the previous HMDA reports.

The applicant’s race data field for previous HMDA filings included seven categories: code 1

(American Indian or Alaska Native), code 2 (Asian), code 3 (Black or African American), code 4

(Native Hawaiian or Other Pacific Islander), code 5 (White), code 6 (Information not provided

by applicant in mail, internet, or telephone application), and code 7 (Not applicable). An

additional category, code 8 (No co-applicant), was included in the co-applicant’s race data field.

An applicant (or co-applicant) was able to select, and a reporter was able to provide, up to five of

these categories.

Under the 2015 HMDA Rule, two of the race categories were further disaggregated to allow for

applicants and co-applicants to self-identify using more detailed race categories. Seven

additional categories were added under code 2 (Asian): code 21 (Asian Indian), code 22

(Chinese), code 23 (Filipino), code 24 (Japanese), code 25 (Korean), code 26 (Vietnamese), and

code 27 (Other Asian). Four additional categories were added to code 4 (Native Hawaiian or

Other Pacific Islander): code 41 (Native Hawaiian), code 42 (Guamanian or Chamorro), code 43

(Samoan), and code 44 (Other Pacific Islander). The self-identification of the racial categories is

optional for applicants and co-applicants. However, if an applicant or co-applicant applies in

person and declines to provide any race or ethnicity information, the financial institution is

required to collect and report aggregate race and ethnicity information based on visual

observation or surname.

Reporters could populate up to five fields for the race of applicants and co-applicants. Table

3.2.1 presents the distribution of an applicant’s race in the first field.

26

By this field, in the 2019

25

Specifically, previous HMDA reports combined race and ethnicity of applicants and co-applicants, which resulted

in seven categories: Asian, Black or African American, Hispanic White, Non-Hispanic White, Other minority, Joint,

and Missing. See note section of Table 2A of the Data Point article published by the Bureau on June 24, 2020, titled

“Data Point: 2019 Mortgage Market Activity and Trends” for more information on how race and ethnicity are defined

for the remainder of this report. The article is available at https://www.consumerfinance.gov/data-

research/research-reports/data-point-2019-mortgage-market-activity-and-trends/.

26

The table presents what was reported in the first data field for race. Code 6 indicates a case where an applicant did

not provide information and a reporter could not determine race/ethnicity/sex based on visual observation or

surname because the application was not submitted in person. Code 7 indicates that an application was likely

submitted by a non-natural person, such as an LLC. Some observations were missing any enumeration and thus were

labeled as “missing” in the tables. Code 6, code 7, and missing data are lumped into one category under “Not available

or missing.”

19

HMDA data, 64.2 percent of applicants were reported as White, 6.7 percent as Black or African

American, 4.7 percent as Asian, 0.7 percent as American Indian or Alaska Native, and 0.2

percent as Native Hawaiian or Other Pacific Islander. A small share of applications (0.7 percent)

reported detailed categories in the first field, which is examined further in Table 3.2.3.

Table 3.2.2 presents the number of distinct races selected by the first reported race of an

applicant.

27

The vast majority of applicants selected only one race, with the exception of

applicants who selected American Indian or Alaska Native (in which case only a modest majority

selected one race). Among applicants who selected White in the first field, 99.8 percent selected

only one race. Similarly, among those who reported Native Hawaiian or other Pacific Islander,

Black, or Asian in the first field, 87.6 percent, 97.5 percent, and 94.3 percent respectively

reported one race. In contrast, among those who reported American Indian or Alaska Native,

59.7 percent selected one race, 35.4 percent selected two races, and the remaining selected three

or more races.

Table 3.2.3 presents the number and percentage of an applicant’s race in the second field

conditional on the race reported in the first field. Most applicants who populated two or more

race fields selected an aggregate race first and then a more detailed race afterwards. About 60

percent of those with Asian reported in the first field selected one of the detailed Asian

categories in the second field. Out of 825,767 applicants for whom Asian was reported in the

first field, 17.9 percent reported Asian Indian, 14.5 percent reported Chinese, 8.4 percent

reported Filipino, and 6.3 percent reported Vietnamese in the second field. About 34 percent of

applicants for whom Asian was reported in the first field had the second field as not applicable

or missing. A slightly larger percentage (48 percent) of those who reported Native Hawaiian or

Other Pacific Islander in the first field left the second field as not applicable or missing. On the

other hand, a small percentage of applicants reported a detailed Asian or a detailed Native

Hawaiian or Other Pacific Islander category in the first field, most of whom then had the second

race field blank. For instance, out of 55,751 applicants who had “Asian Indian” selected in the

first race field, 95.7 percent had the second race field as not applicable or missing.

The new HMDA data allowed ethnicity to be reported at a more detailed level as well. Previous

HMDA data allowed only two categories for ethnicity: code 1 (Hispanic or Latino) and code 2

(Not Hispanic or Latino). In addition to these two categories, the new HMDA data allowed

reporting of more detailed Hispanic or Latino categories: code 11 (Mexican), code 12 (Puerto

Rican), code 13 (Cuban), and code 14 (Other Hispanic or Latino). Consistent with race

27

The disaggregated categories of Asian and Native Hawaiian or Other Pacific Islander are aggregated for this

analysis to avoid duplicate counting. For example, without the aggregation, if an applicant selected Asian in the first

field and Chinese in the second field, the total count of populated race fields would be two. With aggregation, the

number of reported race for this observation is one, which is how it would be counted within Table 3.2.2.

20

information, the self-identification of ethnicity was optional for applicants and co-applicants.

Furthermore, reporters could populate up to five ethnicity fields for both applicants and co-

applicants. Table 3.2.4 presents the distribution of the applicant’s ethnicity reported in the first

field. Code 3 (Information not provided by applicant in mail, internet, or telephone application),

code 4 (Not applicable), and missing data are lumped into one category under “Not available or

missing.” Nearly ten percent of applicants reported Hispanic or Latino, and 67.9 percent

reported Not Hispanic or Latino in the first field. A small share of applicants reported detailed

ethnicity in the first field.

Table 3.2.5 is comparable to Table 3.2.3 and shows the number and percentage of an applicant’s

reported ethnicities in the second field conditional on the ethnicities reported in the first field.

Out of 1.6 million applicants who selected Hispanic or Latino in the first field, 27.6 percent

selected Mexican, 6.2 percent selected Puerto Rican, 2.7 percent selected Cuban, and 8.9 percent

selected other Hispanic in the second field. Similar to race, most applicants who reported

disaggregated ethnicity did so by selecting an aggregated ethnicity in the first field and detailed

ethnicity afterwards.

Table 3.2.6 shows how many of the ethnicity data fields were populated. Table 3.2.6 differs from

Table 3.2.2 in that the former table counts the number of reported ethnicity fields and the latter

table counts the number of reported races. For example, if an applicant reported Asian in the

first race field and Chinese in the second race field, Table 3.2.2 presents this as reporting one

race, Asian. On the other hand, if an applicant reported Hispanic in the first ethnicity field and

Mexican in the second field, Table 3.2.6 presents this as reporting two fields. In short, Table

3.2.2 shows how many applicants reported multiple races, while Table 3.2.6 shows how

extensively ethnicity fields were used. For most applicants, only one field of ethnicity was used

(94.4 percent). Only about six percent used two ethnicity fields.

One of the new features of HMDA data was to allow applicants and co-applicants to fill in race

and ethnicity information in free-form text. More specifically, three free-form text fields for race

and one free-form text field for ethnicity were added to allow applicants and co-applicants to fill

in information that was not listed among the standard enumerations. The first text field for race

was reserved for detailed “American Indian or Alaska Native or Principal Tribe.” The second and

the third text fields for race were reserved for detailed “Other Asian” and “Other Pacific

Islander” respectively. The text field for ethnicity was reserved to fill in “Other Hispanic”

information. Free-form text fields used to report race and ethnicity are excluded from the public

loan-level 2019 HMDA data. These free-form text fields were sparsely populated. About one

percent of the applicants filled in the free-form fields for race or ethnicity. Those applicants who

used the free-form text fields generally did so to report a more detailed race or ethnicity that was

not available in the standard enumerations. For example, an applicant would report code 2

(Asian) in the first reported race field and fill in “Cambodian” in the free-form text field.

21

Consistent with the 2018 HMDA data, the top five free-form entries for race were “Cherokee,”

“Pakistani,” “Indian,” “Cambodian,” and “Hmong.” The top five free-form entries for ethnicity

were “Colombian,” “Dominican,” “Salvadorian,” “Brazilian,” and “Peruvian.”

3.3 Visual Observation of Race, Ethnicity and

Sex

One of the new data fields first reported in the 2018 HMDA data was an indicator of whether the

race, ethnicity, or sex of applicants and co-applicants were determined by visual observation or

surname. Reporters had an option to choose code 1 (Collected on the basis of visual observation

or surname), code 2 (Not collected on the basis of visual observation or surname), or code 3

(Not applicable). For co-applicants, an additional code was included --code 4 (No co-applicant).

The self-identification of race, ethnicity, and sex is optional for applicants and co-applicants.

However, if an applicant or co-applicant applies in person and declines to provide the

information, the financial institution is required to collect and report aggregate race/ethnicity

and sex information based on visual observation or surname.

Table 3.3.1 presents the number and share of records for which race and ethnicity of applicants

and co-applicants were determined by visual observation or surname. Approximately four

percent of applicant’s and about two percent of co-applicant’s race were determined by visual

observation or surname. A similar share of applicant’s and co-applicant’s ethnicities were

collected by visual observation or surname. These shares have declined slightly from 2018, when

about five percent of applicant’s and slightly over two percent of co-applicant’s race and

ethnicity were determined by visual observation or surname.

Tables 3.3.2 and 3.3.3 show the share of race and ethnicity determined by visual observation or

surname given that the race and ethnicity information were reported in the first field. The

disaggregated race and ethnicity categories are aggregated in these tables because HMDA

reporters use visual observation or surname as a basis to collect and report only aggregate race

and ethnicity data. Table 3.3.2 indicates that about four percent of values for race information

were reported this way, with the lowest share for American Indian or Alaska Native (3.5 percent)

and the highest share for Native Hawaiian or Other Pacific Islander (5.4 percent). Table 3.3.3

shows that about five percent of ethnicity information of those whose first reported ethnicity

was Hispanic or Latino was determined by visual observation or surname.

22

Table 3.3.4 presents the distribution of sex for applicants and co-applicants. Reporters selected

among code 1 (Male), code 2 (Female), code 3 (Information not provided by applicant in mail,

internet, or telephone application), code 4 (Not applicable), and code 6 (Applicants selected

both male and female). Reporters also had the option of code 5 (No co-applicant) for the co-

applicant field. The share of reported male applicants (56.0 percent) was twice as large as the

share of reported female applicants (27.6 percent). Less than 10,000 applicants reported sex as

both male and female.

Table 3.3.5 shows the share of applicants for which sex was determined by visual observation or

surname by the reported sex of applicants. Approximately five percent of male and female

applicant’s sex was determined by visual observation or surname.

23

4. Property Type

In the HMDA data prior to 2018, the property type was defined in a single data point indicating

whether a property was a one-to-four-family home, a manufactured home, or a multifamily

home. Starting with the 2018 HMDA data, this information is now captured by two data points.

The first data point, “Construction Method,” indicates whether the property is site-built (code 1)

or a manufactured home (code 2). The second data point “Total Units” specifies the number of

individual dwelling units related to the property securing the covered loan or, in the case of an

application, proposed to secure the covered loan. Total units are binned into the following

ranges in the public loan-level 2019 HMDA data: 5 to 24; 25 to 49; 50 to 99; 100 to 149; and 150

and over. To map these two data points to the previous definition of property types, site-built

single-family homes (one-to-four-family homes) are equivalent to properties whose construction

method is reported to be 1 (site-built) and whose total units are less than or equal to four.

Manufactured homes are equivalent to properties whose construction method is reported to be 2

(manufactured home). Site-built multifamily homes are equivalent to properties whose

construction method is reported to be 1 (site-built) and whose total units are greater than four.

Table 4.1 shows the re-classified property type by action taken code in the 2019 HMDA data. In

total, there are about 16.9 million LARs for site-built single-family properties, 53.8 percent of

which are originations. The data includes 582,000 manufactured home LARs, including 178,200

manufactured home originations, and 67,100 multifamily LARs, including 53,400 multifamily

loan originations.

For site-built single-family loans or applications, the overwhelming majority are for a single

unit. As shown in Table 4.2, 97.5 percent of all single-family LARs are for one unit, 1.8 percent

are for two units, 0.4 percent are for three units, and 0.3 percent are for four units. Among site-

built single-family originations, about 97.7 percent are for one unit, 1.6 percent are for two units,

0.4 percent are for three units, and 0.3 percent are for four units.

The overwhelming majority (98.2 percent) of manufactured home originations are for one unit.

There is a very small percentage of manufactured home originations for more than one unit,

including 1,125 loans for two units, 239 loans for three units, 118 loans for four units, and 1,704

loans for more than four units.

As shown in Table 4.4, among the 53,419 multifamily originated loans (which is up from 50,562

in 2018), about 65.1 percent have between five and 24 units, 12.7 percent have between 25 and

49 units, 9.2 percent have between 50 and 99 units, 4.1 percent have between 100 and 149 units,

and 8.8 percent have more than 150 units. Not shown in Table 4.4, the mean number of units

for multifamily originated loans is about 50, and the median is 14.

24

5. Loan Purpose and

Characteristics

The 2015 HMDA Rule added a number of new data points, and expanded the enumeration of

certain pre-existing data points, to allow users of the data to differentiate between types of

applications and loans based on their purpose and certain core features, such as the term of the

loan, fixed vs. adjustable rates, fully amortizing vs. balloon, interest-only or other non-

amortizing features. This section discusses those data points.

5.1 Business or Commercial Purpose Flag

The 2015 HMDA Rule added a flag for whether the loan or the application is primarily for a

business or commercial purpose. Business or Commercial Purpose Flag is one of the

Discretionary Data Points as discussed in the introduction section of this article. The Business or

Commercial Purpose Flag is among the data points that institutions that qualify for the

EGRRCPA partial exemption are not required to report. It has allowable enumerations of code 1

(primarily for a business or commercial purpose), code 2 (not primarily for a business or

commercial purpose), and code 1111 (exempt).

Table 5.1.1 presents the distribution of the business or commercial purpose flag by action type

for all LARs regardless of property type in 2019. There are about 524,000 LARs identified as

primarily for business or commercial purposes, up from 462,000 in 2018. They make up about

3.0 percent of all LARs. Among the originated loans, there are about 329,000 loans primarily

for business or commercial purposes (up from 289,000 in 2018), about 3.5 percent of all

reported loan originations. Among the purchased loans, there are about 30,000 loans primarily

for business or commercial purposes (down from 35,000 in 2018), or about 1.3 percent of all

reported purchased loans.

Among the originated loans, Table 5.1.2 breaks out the business or commercial purpose flag by

property type. About 3.1 percent of site-built single-family home originations, or 284,000 loans,

are primarily for business or commercial purposes. About 5,000 manufactured home loans, or

2.8 percent of manufactured home originations, are primarily for business or commercial

purposes. On the other hand, most site-built multifamily home loans are primarily for business

or commercial purposes (75.8 percent), 23.8 percent of site-built multifamily loans reported

“Exempt” for the commercial/business purpose flag, and only 0.4 percent of site-built

25

multifamily loans are affirmatively identified as not primarily for business or commercial

purposes.

The remainder of this article focuses on site-built single-family home applications and

transactions, except for the discussion of three new data points: Manufactured Home Secured

Property Type and Manufactured Home Land Property Interest, which are data points only

applicable for manufactured homes, and Multifamily Affordable Units. All statistics reported for

the rest of the article are for site-built single-family loans and applications, unless noted

otherwise.

As shown in Table 5.1.3, about 254,000 closed-end conventional conforming loans (or 4.8

percent), and 13,000 jumbo loans (or 3.6 percent) are primarily for business or commercial

purposes; about 13,000 HELOCs, (or 1.3 percent) are primarily for business or commercial

purpose. Only a small fraction of FHA and VA loans are reported primarily for business or

commercial purposes.

Loan Purpose is a data field that is separate from the Business/Commercial Loan Purpose

Flag.

28

As shown in Table 5.1.4, in 2019, about 20.8 percent of originated loans (about 1,400

loans) that had loan purpose reported as “NA” are primarily for business or commercial

purposes. However, they only account for about 0.5 percent of all primarily business or

commercial purpose originations. Among all primarily business or commercial purpose

originations, 56.2 percent (about 160,000 loans) reported their loan purpose as home purchase,

and 37.8 percent as refinance (including 15.9 percent for cash-out refinance).

Table 5.1.5 shows that about 95.5 percent of originated loans that are primarily for business or

commercial purposes (271,000 loans) are for investment properties.

29

Loans that are primarily

for business or commercial purposes account for about 49.6 percent of loans for investment

properties in 2019, up from 45.3 percent in 2018. About 4.1 percent of originated loans that are

primarily for business or commercial purposes are listed as being secured by the principal

residence, and 0.4 percent are reported as being secured by the second residence.

Table 5.1.6 breaks down originations by race and ethnicity and primarily for business or

commercial purpose flag. Approximately 41.7 percent of all single-family business or

commercial primary purpose originations (118,000 loans) had race and ethnicity reported as

28

Loan purpose is a pre-existing data point that was modified by the 2015 HMDA Rule. It will be the subject of

discussion in the next subsection (Section 5.2).

29

Occupancy status is a pre-existing data point that was modified by the 2015 HMDA Rule to break out investment

property, second residence and principal residence. It will be the subject of discussion in Section 6.1.

26

missing.

30

This may be because a large proportion of those loans were taken out by non-natural

persons (for example, a corporation, partnership, or trust) for which the race and ethnicity are

reported as not applicable.

5.2 Loan Purpose

The 2015 HMDA Rule revised the enumeration of the Loan Purpose data point by adding two

new reporting options: “cash-out refinance,” and “other purpose.” The revised loan purpose data

point has the following allowable values: code 1 (home purchase), code 2 (home improvement),

code 31 (refinancing), code 32 (cash-out refinancing), code 4 (other purpose), and code 5 (not

applicable). Importantly, the 2015 HMDA Rule also modified the definition of reportable

transactions. Under the rule, home improvement loans unsecured or secured by some collateral

other than a residential dwelling, as well as all agricultural-purpose loans and LOCs, are no

longer reportable. On the other hand, reporting of open-end LOCs becomes mandatory for

lenders that exceed the open-end threshold and meet other applicable criteria. This has strong

implications for the reporting of loan purpose, as some transactions not for the purposes of

home purchase, home improvement, or refinance, but secured by dwellings are now reportable

under HMDA and have their loan purpose listed under “other purpose,” while all home

improvement loans not secured by dwellings are dropped from the HMDA coverage.

Table 5.2.1 shows the tabulations of loan purpose by action type for all site-built single-family

LARs in the 2019 HMDA data. Of originated loans, 47.9 percent are for home purchase, 5.9

percent are for home improvement, 24.5 percent are for non-cash-out refinance, 16.2 percent

are for cash-out refinance, 5.3 percent are for “other purpose”, and 0.1 percent are reported as

not applicable in 2019.

31

In comparison, among all originated loans in 2018, 56 percent are for

home purchase, 7.5 percent are for home improvement, 14.1 percent are for non-cash-out

refinance, 15.6 percent are for cash-out refinance, and 6.7 percent are for “other purpose.” In

2019, there are about 485,000 originated loans with a loan purpose of “other.” These loans

would have not been reported under HMDA prior to the 2015 HMDA Rule, because their loan

purposes do not fall into the categories of home purchase, home improvement, or refinance.

Among purchased loans, 61.4 percent are for home purchase, only 0.3 percent are for home

improvement, 17.4 percent are for non-cash-out refinance, 12.6 percent are for cash-out

30

Note that within this article, to be consistent with the approach taken in the past Federal Reserve Board HMDA

Bulletins and the CFPB Data Point Articles on HMDA data, with the exception of Sections 3.2 and 3.3, "Missing" in

race/ethnic categorization refers to applications in which the race of the applicant(s) has not been reported or is not

applicable or the application is categorized as White but ethnicity has not been reported.

31

For the purpose of this article, non-cash-out refinance transactions are HMDA records that have loan purpose

reported as code 31 (refinancing) and not code 32 (cash-out refinancing).

27

refinance, 0.5 percent are for other purpose, and 7.8 percent have a loan purpose reported as

not applicable.

32

In comparison, among purchased loans in 2018, 69.3 percent are for home

purchase, only 0.8 percent are for home improvement, 7.1 percent are for non-cash-out

refinance, 10.8 percent are for cash-out refinance, 0.5 percent are for other purpose, and 11.5

percent have a loan purpose reported as not applicable. Purchased loans account for 91.4

percent of LARs for which the loan purpose is reported as not applicable.

Table 5.2.2 shows the loan purpose distribution for each closed-end forward enhanced loan type

among all site-built single-family originations in 2019. Home-purchase loans account for 53.7

percent of all jumbo loans, while home-purchase loans account for 51.0 percent of conventional

conforming loans. For RHS/FSA loans, 97.6 percent are for home purchase, with the remaining

2.3 percent for non-cash-out refinance. For FHA loans, 67.8 percent are for home purchase, the

second highest among all enhanced loan types. Not shown in Table 5.2.2, 5.0 percent of HELOC

originations are for home purchase, and 35.1 percent of HELOC originations are for home

improvement. The share for other loan types for home improvement is in the low single-digits.

In contrast, the share of home improvement loans in years prior to 2018 typically accounted for

a much higher percentage of all HMDA originations. Such difference in terms of the shares of

transactions reported for home improvement purpose between the HMDA data for 2018 and

2019 and prior to 2018 is most likely because beginning in 2018 home improvement loans that

are not secured by a dwelling are excluded from HMDA coverage per the 2015 HMDA Rule.

Compared to 2018, the share of loans for home purchase had dropped significantly for most

enhanced loan types, especially for jumbo, conventional conforming and FHA loans whose

home-purchase loan shares were at 71.8 percent, 61.5 percent, and 78 percent in 2018

respectively. This is mostly a result of rising shares of refinance loans.

In 2019, the shares of non-cash-out refinance for conventional conforming and jumbo loans are

at 25 percent and 30.6 percent, respectively. The share of cash-out refinance among

conventional conforming loans is 18.9 percent, and the share of cash-out refinance for jumbo

loans is 13.1 percent. The cash-out refinance share of VA loans is 18.5 percent, and the non-cash-

out share is 32.7 percent. The cash-out refinance share of FHA loans is 14.1 percent, and the

non-cash-out share is 17.0 percent. Overall, compared to 2018, the rise of the share of non-cash-

out refinance was prominent across all enhanced loan types. In particular, in 2018, the shares of

non-cash-out refinance for conventional conforming, jumbo, VA, and FHA loans were 14.2

percent, 14.7 percent, 7.8 percent, and 6.3 percent, respectively.

32

Note that the purchased loans that reported “other” as the loan purpose would have not been reported under

HMDA prior to the 2015 HMDA Rule, which took effect in 2018, because their loan purpose does not fall into the

home purchase, home improvement, or refinance categories. In addition under Regulation C, for purchased covered

loans, where origination took place prior to January 1, 2018, a financial institution complies with § 1003.4(a)(3) by

reporting that the loan purpose reporting requirement is not applicable. See comment 4(a)(3)-6.

28

Table 5.2.3 shows the distribution of loan purpose by race/ethnicity, borrowers’ age group, tract

income and metro/rural status for closed-end forward mortgages. In 2019, about 63.4 percent of

the loans taken out by for Hispanic White borrowers are for home purchase, followed by 60.1

percent for Black borrowers, 56.2 percent for Asian borrowers, and 53.3 percent for non-

Hispanic White borrowers. As in 2018, Asian borrowers are the least likely to take out loans for

cash-out refinance in 2019. Only 11.5 percent of Asian borrowers’ loans are reported as cash-out

refinance in 2019, the same share as in 2018. At 30.0 percent, the share of Asian borrowers’

loans for non-cash-out refinance, however, is larger than that of Black, Hispanic White, and

non-Hispanic White borrowers in 2019, which was not the case in 2018 when the non-cash-out

refinance share among Asian borrowers trailed slightly behind all other groups. The rise of

refinance share, especially non-cash-out refinance share, was particularly prominent among

Asian and non-Hispanic White borrowers.

Even though the shares of home purchase loans decrease in 2019 compared to 2018, just like in

2018, within 2019 data the share of borrowers taking out loans for home purchase decreases

monotonically with age. Approximately 91.8 percent of borrowers younger than 25 take out

loans for home purchase. The share of home-purchase loans is 73.4 percent for borrowers

between the age of 25 and 34, and 54.7 percent for borrowers between the age of 35 and 44. The

share of home-purchase loans drops to 31.2 percent for borrowers 75 years of age or older.

Conversely, the share of refinance loans (including both cash-out and non-cash-out refinance

loans) generally increases with age. The share of home improvement loans also increases with

age until the borrowers reach the 55 to 64 age range, and then it drops slightly for the next two

age ranges. These numbers are likely driven by people gradually moving into home ownership as

they age, and existing home owners seeking to refinance and make home improvements.

In high-income census tracts, 51.5 percent of borrowers take out loans for home purchase. This

is lower than the share of borrowers in middle-income tracts (54.6 percent), and the share of

borrowers in low/moderate-income tracts (57.7 percent). Similarly, borrowers living in

metropolitan statistical areas are less likely than borrowers in micropolitan statistical areas and

borrowers in rural areas to take out loans for home purchase, with home-purchase loan shares

of 53.5 percent, 58.4 percent and 56.8 percent for these three geographic areas, respectively.

Table 5.2.4 shows the distribution of loan purpose by lien status for all closed-end mortgage

originations. Properties secured by a first lien account for 97.7 percent of all home-purchase

mortgages in 2019 (down slightly from 98 percent in 2018), 98.2 percent of non-cash-out

refinances, and 98.1 percent of cash-out refinances (up from 95.4 percent and 97.3 percent in

2018, respectively). In contrast, only 48.4 percent of home improvement loans and 54.0 percent

of closed-end mortgages that report their loan purpose as “other purpose” are secured by a first

lien.

29

As shown on Table 5.2.5, the median loan amount of cash-out refinance loans is lower than the

median loan amount of non-cash-out refinance loans for all enhanced closed-end loan types

with meaningful volumes in 2019

33

. The opposite was true of 2018 when the median loan

amount of cash-out refinance loans was higher than that of non-cash-out refinance loans for all

enhanced closed-end loan types except for jumbo loans. Loans for home improvement and

“other purpose” have the lowest median loan amounts among conventional conforming loans.

5.3 Loan Term

The DFA, as implemented by the 2015 HMDA Rule, added Loan Term as a new data point that

must be reported. Loan Term is one of the Mandated Data Points as discussed in the

introduction section of this article. Loan term under Regulation C is defined as the number of

months after which the legal obligation will mature or terminate, or for applications would have

matured or terminated. It is among the data points that institutions that qualify for the

EGRRCPA partial exemption are not required to report.

In total, the 2019 data include over 445 distinct values of loan terms. Table 5.3.1 lists the top 20

most common terms reported for originated closed-end mortgages excluding reverse mortgages.

The dominant loan term of closed-end mortgages is 360 months (30 years), accounting for 80.5

percent of all closed-end mortgage originations, followed by 180 months (15 years) which

accounts for 9.4 percent of closed-end originations in 2019 (rising from 8.9 percent in 2018).

Additional commonly reported loan terms include 240 months (20 years), 120 months (10

years), and 60 months (5 years), accounting for 3.4 percent, 2.1 percent, and 0.8 percent,

respectively. Together, the top 20 most common loan terms account for 98.7 percent of all

closed-end originations.

Table 5.3.2 lists the top 20 most common loan terms reported for HELOC originations. For

HELOC originations the most common loan term is 360 months (30 years), accounting for 46.3

percent, followed by the loan term of 300 months (25 years), accounting for 18.6 percent.

Approximately 8.0 percent of HELOC originations have a loan term reported as 361 months.

This extra month difference is likely due to how the first month of credit is counted, and for

underwriting and pricing matters, it is not materially different from a HELOC term of 360

months (30 years). The other common loan terms for HELOCs are 240 months, i.e. 20 years

(6.7 percent), 120 months / 10 years (5.8 percent), 480 months / 40 years (4.3 percent), 180

33

A very small number of RHS/FSA loans (55 exactly) reported cash-out refinance for loan purpose, most likely due

to reporting errors. They are excluded from Table 5.2.5.

30

months / 15 years (3.8 percent), and 60 months / 5 years (2.2 percent). Together the top 20

most common loan terms account for 98.8 percent of all HELOC originations.

Reverse mortgages have no defined loan terms, as reverse mortgages have no maturity date and

generally only terminate when borrowers die, refinance, or move out.

Table 5.3.3 examines the five most common loan terms for closed-end originations by loan

purpose, race/ethnicity, borrowers’ age group, neighborhood income, and geography.

34

Of all

home-purchase loans, 91.1 percent have a term of 30 years. In contrast, 73.4 percent of cash-out

refinance loans and 70.5 percent of non-cash-out refinance loans have 30-year terms in 2019. It

is worth noting that in 2018 about 57.2 percent of non-cash-out refinance loans had 30-year

terms, while the share of 30-year terms for cash-out refinance loans was only slightly lower in

2018 than in 2019, at 73 percent. The significant rise of the use of 30-year mortgages for non-

cash-out refinance, coupled with the rise of non-cash-out refinance overall as discussed in

Section 5.2, shows that not only are larger numbers of borrowers taking out non-cash-refinance

loans in 2019, they are also locking in long terms given the low interest rate environment.

Consequently, the share of loans with 15-year terms also dropped from 19.2 percent in 2018 to

14.3 percent in 2019 for non-cash-out refinance loans, while approximately 15.5 percent of cash-

out refinance loans have 15-year terms in 2019, up slightly from 15.1 percent in 2018. Among

home improvement loans and loans reported as “other purpose,” only 25.1 percent and 31.9

percent, respectively, are reported as having a 30-year term, percentages much lower than those

of home-purchase and refinance loans.

For closed-end loans to non-Hispanic White borrowers, 80.0 percent have a 30-year term. In

comparison, 80.7 percent of Asian borrowers, 85.8 percent of Black borrowers and 86.6 percent

of Hispanic White borrowers take out loans with a 30-year term. The percentage of borrowers

taking out 30-year term mortgages decreases with age until the borrowers are 65 years or older.

For instance, 93.2 percent of borrowers younger than 25 and 90.2 percent of borrowers between

25 and 34 years old obtained mortgages with a 30-year term. The share of borrowers obtaining

30-year closed-end mortgages drops to 72.9 percent for borrowers between 55 and 64 years old,

then rises again with age, with 76.7 percent of borrowers between age 65 and 74 years old and

79.2 percent of borrowers 75 years or older taking out 30-year loans.

The variation in the shares of borrowers taking out 30-year term mortgages in the high-income

census tracts, middle-income tracts and low/moderate-income tracts is limited. On the other

hand, borrowers in rural areas are less likely to take out 30-year term mortgages than borrowers

in micropolitan statistical areas, who in turn are less like to do so than borrowers in

metropolitan statistical areas. The share of borrowers obtaining 30-year term mortgages in

34

Each of which accounts for more than one percent of all closed-end mortgage originations.

31

metropolitan areas, micropolitan areas, and rural areas is 81.2 percent, 75.7 percent and 70.5

percent, respectively.

Table 5.3.4 examines the seven most common loan terms for HELOC originations by

race/ethnicity, borrowers’ age group, neighborhood income, and geography.

35

At 51.4 percent,

Black HELOC borrowers are less likely than other borrowers to take out HELOCs for a 30-year

term. Unlike for closed-end loans, the percentage of HELOC borrowers taking out 30-year term

loans increases consistently with age. HELOC borrowers living in high-income census tracts are

more likely to take out 30-year term HELOCs than borrowers in low/moderate-income tracts or

middle-income tracts, and HELOC borrowers in rural areas are less reliant on 30-year term

HELOCs than borrowers in micropolitan statistical areas and metropolitan areas.

Table 5.3.5 shows the distribution of common loan terms for each enhanced loan type, excluding

reverse mortgages. RHS/FSA loans are reported as almost exclusively 30-year term loans. The

30-year mortgages make up 95.5 percent of all FHA loans and 93.4 percent of VA originations in

2019, each down slightly from 2018 when 30-year term loans accounted for 96 percent of FHA

loans and 94.4 percent of VA loans. The shares of jumbo loans and conventional conforming

originations with 30-year terms are 89.5 percent and 74.5 percent, respectively. For

conventional conforming loans, 12.8 percent are 15-year terms, 3.0 percent are 10-year terms,

and 1.1 percent are 5-year terms. For HELOC originations, 30-year term loans account for only

46.3 percent.

5.4 Introductory Rate Period

The DFA, as implemented by the 2015 HMDA Rule, added the Introductory Rate Period data

point to the reporting requirements. Introductory Rate Period is one of the Mandated Data

Points as discussed in the introduction section of this article. It is defined as the number of

months, or proposed number of months in the case of an application, until the first date the

interest rate may change after closing or account opening. For fixed-rate mortgages, this data

point is reported as “NA”, i.e. not applicable. The introductory rate period is among the data

points that institutions that qualify for the EGRRCPA partial exemption are not required to

report.

Most loans or applications reporting an introductory period (other than not applicable or

Exempt) are adjustable-rate mortgages, commonly known as ARM loans, including Hybrid

ARMs that offer a fixed rate for a predetermined period and then adjust periodically for the rest

35

Each of which accounts for more than one percent of all closed-end mortgage originations. For Table 5.4.3, the

HELOCs with reported loan term equal to 360 months and 361 months are combined into 30-year term.

32

of the loan term. Also, some loans have an introductory rate period after which the interest rate

resets to a predetermined fixed rate in what is known as a “step-rate product.” For simplicity, all

loans and applications with introductory rate period reported as not applicable are referred to as

fixed-rate mortgages, and all loans and applications with a positive number reported for the

introductory rate period are referred to as ARM loans, acknowledging that such nomenclature

may blend “step-rate products” or other non-standard non-fixed-rate products with traditional

ARM products.

Table 5.4.1 shows the share of fixed- and adjustable-rate originations for loans and LOCs that

did not report introductory rate period as Exempt. Among these originations, fixed-rate

mortgages make up 94.5 percent of conventional conforming loans, but only 58.3 percent of

jumbo loans. RHS/FSA loans are exclusively fixed-rate, and fixed-rate mortgages also make up

99.8 percent of FHA loans and VA loans. On the other hand, only 19.4 percent of non-exempt

HELOC originations are fixed-rate loans, and 80.6 percent of HELOCs are adjustable-rate loans

in 2019, up from 77.1 percent in 2018. Among non-exempt reverse mortgage originations, 48.0

percent are fixed-rate, and 52.0 percent are adjustable rate. In contrast in 2018, 58.2 percent of

reverse mortgage originations were fixed rate.

In terms of race and ethnicity, as shown in Table 5.4.2, Asian borrowers are the most likely to

take out adjustable-rate mortgages for closed-end loans at 13.9 percent, compared to 5.7 percent

of non-Hispanic White, 2.6 percent of Black, and 2.5 percent of Hispanic White closed-end

mortgage borrowers who take out adjustable-rate loans. The share of borrowers taking out

adjustable-rate mortgages generally increases with age. Only 3.0 percent of closed-end

borrowers younger than 25 take out ARM loans, while the share of closed-end borrowers taking

out ARM loans rises to 6.8 percent for borrowers 75 years of age or older.

ARM loans account for 8.1 percent of all closed-end mortgage originations in high-income

census tracts, while they only account for 4.5 percent of closed-end loans in middle-income

tracts and 4.2 percent in low/moderate-income tracts. Borrowers in rural areas are more likely

than borrowers in micropolitan or metropolitan statistical areas to use ARMs. In rural areas, 6.6

percent of closed-end mortgages are ARMs, compared to 6.3 percent in micropolitan statistical

areas and 5.8 percent in metropolitan statistical areas.

These patterns are consistent with the 2018 HMDA data, even though the overall shares of

borrowers using adjustable-rate mortgages have dropped across the board in 2019 compared to

2018.

Counting only non-partially exempt complete applications—i.e. the applications whose action

types show either approval or denial—Table 5.4.3 shows that the denial rates for fixed-rate

mortgages are higher than the denial rates for ARMs among conventional conforming, jumbo,

33

and HELOC applications, but lower than the denial rate for ARMs among reverse mortgage

applications.

36

Among the loans that reported a numerical introductory rate period, there are over 160 distinct

values of introductory rate periods. Table 5.4.4 lists the top 20 most common introductory rate

periods reported for originated closed-end forward mortgages, excluding reverse mortgages.

Together, they account for about 98.6 percent of all 454,000 adjustable-rate closed-end forward

loans, excluding reverse mortgages. In 2019, 84 months (7 years) and 60 months (5 years) are

the two most common introductory rate periods, accounting for 30.5 percent and 29.5 percent

of all adjustable-rate loans respectively, followed by an introductory rate period of 120 months

(10 years) with a 24.5 percent share. The next most common introductory rate period at 36

months (3 years) only accounts for 4.0 percent of all adjustable-rate mortgages. Table 5.4.5

regroups some reported introductory rate periods that are close to the most common traditional

ARM values and presents the most common regrouped ARM introductory rate periods by

common loan terms for closed-end mortgages. As shown in Table 5.4.5, the regrouping confirms

that the most common introductory rate periods among the closed-end ARMs are five years and

seven years, closely followed by ten years, and then followed further down by three years, and

then others.

ARMs with the same introductory period could have different loan terms, but the most common

terms for ARM products remains 30 years.

Most HELOCs (80.6 percent) are adjustable-rate, as discussed previously. Of the 835,000

adjustable-rate HELOCs, 574,000 (68.7 percent) reported a one-month introductory rate

period. The interest rate of these HELOCs immediately goes into float after the first month.

About 127,000 (15.2 percent) of adjustable-rate HELOC originations have a twelve-month

introductory rate period, and another 74,000 (8.8 percent) have an introductory rate period of

six months. Table 5.4.6 lists the top 20 introductory rate periods for HELOC originations that

reported a positive introductory rate period. Together the top 20 introductory rate periods

account for 99.95 percent of all adjustable-rate HELOC originations.

Of the 17,778 originated adjustable-rate reverse mortgages, 98.0 percent had an introductory

rate period of 12 months (72.4 percent) or 13 months (25.6 percent), with another 0.6 percent

reporting an introductory rate period of 3 month. (Not shown in a table.)

36

Note that as shown in Table 5.4.1, only about 2,600 or 0.2 percent of FHA single-family closed-end mortgages are

ARMs, in comparison to about 1,104,000 or 99.8 percent of FHA fixed-rate single-family closed-end mortgages. The

denial rate for FHA ARM loans is higher than the denial rate for FHA fixed-rate mortgages. Similarly, only about

1,200 or 0.2 percent of VA loans are ARMs. The denial rate for VA ARM loans is higher than that of VA fixed-rate

mortgages in 2019.

34

5.5 Non-Amortizing Features

The DFA, as implemented by the 2015 HMDA Rule, added Non-Amortizing Features as a data

point to be reported. Non-Amortizing Features is one of the Mandated Data Points as discussed

in the introduction section of this article. It requires HMDA reporters to indicate whether the

contractual terms of a loan or an application includes or would have included any of the

following: (1) a balloon payment, (2) interest-only payments for a period of time, (3) a

contractual term that would cause the covered loan to be a negative amortization loan, or (4)

any other contractual term that would allow for payments other than fully amortizing payments

during the loan term. Such information is reported through four relevant data fields: balloon

feature, interest-only payments, amortization, and other non-amortizing features. Each of these

four fields is a flag, with 1 indicating that the relevant amortization feature applies, and 2

indicating no such feature applies. These four data fields are among the data points that

institutions that qualify for the EGRRCPA partial exemption are not required to report. The

code 1111 for each of these fields represents “Exempt” from the reporting requirements.

Table 5.5.1 shows the tabulation of the four non-amortizing feature flags for originated closed-

end mortgages and HELOCs, respectively. There are about 225,000 originated loans that

include a balloon payment, about 122,000 of which are closed-end loans, and 103,000 of which

are HELOCs. Loans with balloon payments make up about 10.0 percent of all HELOC

originations, higher than the 1.6 percent of closed-end originations that have a balloon payment.

There are about 705,000 originated loans that have an interest-only feature, of which about

158,000 are closed-end loans, and 547,000 are HELOCs. A little more than half of HELOCs

(52.8 percent) feature interest-only payments. In contrast, only 2.1 percent of closed-end

mortgages are interest-only loans. There are only about 3,500 loans or lines of credit with

negative amortization features, approximately 2,300 of which are HELOCs. About 43,000 (or

4.2 percent) of HELOC originations are reported with “other non-amortizing features,” while

only 11,000 closed-end originations are associated with other non-amortizing features.

Among the closed-end mortgages, Table 5.5.2 examines the distribution of the four non-

amortizing features by enhanced loan types. Balloon loans account for 2.1 percent of

conventional conforming mortgages and 2.5 percent of jumbo loans. Loans with an interest-only

feature account for 1.9 percent of conventional conforming, and 14.7 percent of jumbo loans.

Only a tiny fraction of FHA loans is reported to have a balloon feature or an interest-only

feature. The same pattern exists for VA loans. Similarly, only a tiny fraction of RHS/FSA loans

are reported to have a balloon payment.

Table 5.5.3 presents some selected characteristics of the borrowers and loans by different non-

amortizing features for closed-end mortgages. Among balloon loans, 65.1 percent are for home

purchase, while 54.8 percent of non-balloon loans are for home purchase. Consequently, the

35

share of balloon loans for refinance (28.2 percent) is lower than that of non-balloon loans (43.3

percent), and furthermore the share of cash-out refinance is lower for balloon loans (7.9

percent) than non-balloon loans (18.1 percent). In total, 6.7 percent of balloon loans are for

home improvement or “other” purpose, compared to 1.9 percent of non-balloon loans.

37

The median interest rate of balloon loans is 5.75 percent, higher than the median interest rate of

non-balloon loans at 4.0 percent. Balloon borrowers have higher median income ($94,000)

than the median income of non-balloon borrowers ($90,000). The median credit score,

Combined Loan-to-Value Ratio (CLTV), and Debt-to-Income Ratio (DTI) of balloon borrowers

(725, 75.6, and 36.8, respectively) are all lower than those of non-balloon borrowers, (742, 80

and 37.5, respectively).

For loans with an interest-only feature, 65.1 percent are for home purchase. In contrast, 54.7

percent of loans that are identified as not interest-only are for home purchase. The median

interest rate for interest-only loans is higher than that for loans that are not interest-only, at 4.5

percent, compared to 4.0 percent. Interest-only borrowers have much higher incomes than

other borrowers. The median income of interest-only borrowers is $194,000 per year, compared

to the median income of borrowers with loans that are affirmatively reported as not interest-

only at $90,000. The median credit score of interest-only borrowers is also 27 points higher, at

768, compared to 741 for borrowers with loans that are reported as not interest-only. The

median CLTV on interest-only loans is 73.8 percent, lower than the median CLTV of non-

interest-only loans which is 80 percent. The median DTI of interest-only borrowers is also

slightly lower than that of borrowers with loans that are reported as not interest-only, at 35.8

percent compared to 37.5 percent.

Table 5.5.4 shows the distribution of balloon feature and interest-only feature loans by race and

ethnicity for closed-end mortgages. Approximately 0.9 percent of non-Hispanic White

borrowers take out loans with a balloon payment. In contrast, 0.6 percent of Asian borrowers,

0.8 percent of Black borrowers, and 0.9 percent of Hispanic White borrowers take out balloon

loans. Non-Hispanic White borrowers also are more likely than minorities to take out interest-

only loans. Approximately 1.8 percent of loans taken out by Non-Hispanic White borrowers are

interest-only. In comparison, 1.0 percent of Asian borrowers, 0.7 percent of Black borrowers,

and 0.7 percent of Hispanic White borrowers take out interest-only closed-end mortgages.

Table 5.5.5 shows the distribution of balloon and interest-only features by borrowers’ age groups

for closed-end mortgages. The share of borrowers taking out interest-only loans generally

increases with age. While only 0.5 percent of borrowers younger than 25 take out interest-only

37

The share of loans for home improvement or “other” purpose can be calculated from Table 5.5.3 by using the

formula: (100% - share of home-purchase loans – share of refinance loans).

36

loans, this share steadily increases until ages 55 through 64. For borrowers between the ages of

55 and 64, 2.1 percent take out loans that involve interest-only payments. This share dips

slightly for borrowers in the 65 to 74 age group, but rises again for borrowers 75 or older, to 2.2

percent.

Table 5.5.6 shows the distribution of balloon features and interest-only features by whether the

property is located in a metropolitan statistical area, micropolitan statistical area, or rural area,

again limited to closed-end originations. The table shows 1.4 percent of loans in metropolitan

statistical areas have balloon features in 2019. In contrast, 2.0 percent of loans in micropolitan

statistical areas and 6.1 percent of loans in rural areas carry balloon features. In comparison, the

share of loans in rural areas with balloon features was 3.6 percent in 2018.

5.6 Prepayment Penalty Term

The DFA, as implemented by the 2015 HMDA Rule, requires the collection and reporting of the

existence of a prepayment penalty term. Prepayment Penalty Term is one of the Mandated Data

Points as discussed in the introduction section of this article. It is defined as the term, in

months, of any prepayment penalty of a loan or an application. The prepayment penalty term is

among the data points that institutions that qualify for the EGRRCPA partial exemption are not

required to report.

In total, about 269,000 single-family originated loans reported a prepayment penalty term in

2019, down from 338,400 in 2018. Table 5.6.1 shows the breakdown of loans with or without

prepayment penalty terms by the enhanced loan types. About 23,200 originated conventional

conforming loans are reported carrying a prepayment penalty term, which account for only 0.5

percent of all conventional conforming originations. There are about 1,100 originated jumbo

loans that are reported to carry a prepayment penalty term, accounting for only 0.3 percent of

all jumbo originations. Loans with prepayment penalties are non-existent for FHA, VA, and

RHS/FSA loans. A prepayment penalty term is much more common among HELOCs. There are

244,600 HELOC originations that carry a prepayment penalty term. They account for about

23.9 percent of all HELOC originations in 2019, down from 28.4 percent in 2018. Prepayment

penalty terms are not applicable to reverse mortgages.

Table 5.6.2 shows that among closed-end mortgages, 0.3 percent of Asian borrowers, 0.5

percent of Black borrowers, 0.6 percent of Hispanic White borrowers, and 0.5 percent of non-

Hispanic White borrowers have loans with a prepayment penalty term. The percentage of

borrowers taking out loans with a prepayment penalty term increases with age. For instance, 0.1

37

percent of borrowers younger than 25, 0.2 percent of borrowers between the age of 25 and 34,

and 0.4 percent of borrowers between the age of 35 and 44 have loans with a prepayment

penalty term. This percentage increases to 0.7 percent for borrowers between the ages of 65 and

74, and 0.9 percent for borrowers older than 74. A slightly higher percentage of loans in rural

areas have a prepayment penalty term than those in micropolitan statistical areas and

metropolitan areas, at 0.7 percent, 0.6 percent and 0.4 percent, respectively.

Table 5.6.3 shows for certain loan features, borrowers’ demographics and geography of HELOCs

with and without a prepayment penalty term. Of the HELOCs with adjustable rates, 26.8

percent have a prepayment penalty term, compared to 12.0 percent of HELOCs with a fixed rate.

HELOCs with balloon features are less likely than HELOCs without balloon features to carry a

prepayment penalty term, at 16.9 percent compared to 24.7 percent. Similarly, HELOCs with

interest-only payments are slightly less likely to have a prepayment penalty term (22.4 percent)

than HELCOs without interest-only payments (25.64 percent). HELOCs reported with “other

non-amortizing features” do not have a prepayment penalty term.

Furthermore, 32.2 percent of Asian HELOC borrowers have a prepayment penalty term on the

LOCs they took, a much higher rate than all other race/ethnicity groups. Just like closed-end

mortgages, the percentage of HELOCs that are reported to have a prepayment penalty term

increases with the borrowers’ age. Unlike the closed-end mortgages, the shares of HELOC

borrowers with a prepayment penalty term are about the same across metropolitan statistical

areas, micropolitan statistical areas, and rural areas.

Table 5.6.4 shows the three most common prepayment penalty terms for closed-end mortgages

and open-end mortgages respectively for originated loans or LOCs that reported a positive

prepayment penalty term. For both closed-end loans and open-end LOCs, prepayment penalty

terms of 36 months, 24 months, and 12 months are the most common prepayment term, in that

order, and account for most of the originated loans or LOCs with a prepayment term.

5.7 Submission of Application and Initially

Payable Flags

The DFA, as implemented by the 2015 HMDA Rule, requires reporting of the application

channel of the covered loan or application. Application Channel is one of the Mandated Data

Points as discussed in the introduction section of this article. The application channel is

reported through two separate data fields: (i) whether the applicant or borrower submitted the

application directly to the reporting institution (“Submission of Application”), and (ii) whether

the obligation arising from the covered loan was, or, in the case of an application, would have

been, initially payable to the reporting institution (“Initially Payable”). This data point is one of

38

the data points that institutions that qualify for the EGRRCPA partial exemption are not

required to report. The Submission of Application data field has the following allowable codes:

code 1 (submitted directly to the reporting institution); code 2 (not submitted directly to the

reporting institution); code 3 (not applicable); and code 1111 (exempt). The Initially Payable data

field has the following allowable codes: code 1 (initially payable to the reporting institution),

code 2 (not initially payable to the reporting institution); code 3 (not applicable); and code 1111

(exempt).

The common terms for lending channels include retail, wholesale, correspondent, and broker

channels. However, none of these terms are formally defined in Regulation C. To understand

how the Submission of Application and Initially Payable data fields help characterize the

application channels from the reporters’ perspective, it is important to keep in mind how to

determine who reports a loan or an application under HMDA. In general, the key to determining

who reports HMDA data on wholesale-correspondent or wholesale-broker loans or applications,

is determining which entity makes the credit decision on the application.

38

For example, a

wholesale-correspondent lender with delegated underwriting authority would make the credit

decision, and hence report the loan or application under its name for HMDA purposes if that

lender also meets all relevant coverage criteria under Regulation C. Later, this wholesale-

correspondent lender could sell this loan to another lender, who may report the same loan as a

purchased loan, if that lender meets all coverage criteria. Alternatively, if the wholesale-

correspondent lender did not have delegated underwriting authority and did not make the credit

decision, this loan would be reported as an originated loan by the second lender, but never

reported by the first (wholesale-correspondent) lender regardless of whether the first lender is a

HMDA reporter and regardless of whether the first lender closes the loan in its name. Given

these reporting qualifications, the chart below illustrates examples of how the Submission of

Application and Initially Payable data fields in combination could align with general application

channels in common terms from the HMDA reporter’s perspective for originated loans.

Chart: Classification of Application Channels

Initially Payable

Yes

No

38

The rest of the discussion uses the term “wholesale” as the opposite of “retail,” comprising of both correspondent

and broker channels. In this section, the term “wholesale-correspondent” refers to correspondent channel in a

lender’s wholesale business separated from retail business; and the term “wholesale-broker” refers to broker channel

in a lender’s wholesale business separated from retail business. Some lenders in the industry may use “wholesale” in

reference to only its broker channel, or correspondent channel, or both. In general, a broker would not meet all the

relevant coverage criteria to be a “financial institution” as defined by § 1003.2(g) in Regulation C, and therefore would

not be a reporter under HMDA.

39

Directly

Submitted

Yes

The reporter made the credit

decision and the loan was

closed in the reporter’s name.

The reporter likely originated

the loan in its retail channel

but could participate in the

wholesale-correspondent

channel of another lender with

delegated underwriting

authority.

The reporter made the credit

decision pursuant to delegated

underwriting authority. The loan

closed in the name of another lender.

The reporter belongs to wholesale

channel of that lender.

No39

The reporter made the credit

decision without delegating its

underwriting authority.40 The

loan was closed in the

reporter’s name. The reporter

originated the loan in its

wholesale-correspondent or

wholesale-broker channel.

The reporter made the credit

decision without delegating its

underwriting authority. The loan was

not closed in the reporter’s name.

The reporter originated the loan in its

wholesale- correspondent channel.

Table 5.7.1 breaks down the number of originations reported in the 2019 HMDA data for each

application channel as defined by these two fields for different enhanced loan types.

Approximately 84.2 percent of all conventional conforming originations were directly submitted

to and initially payable to the reporting institution. Only 1.4 percent of conventional conforming

loans were directly submitted to but were not initially payable to the HMDA reporter. In

contrast, 10.5 percent of conventional conforming loans were not directly submitted to but were

initially payable to the reporting institution. Another 3.9 percent of conventional conforming

loans were neither directly submitted to nor initially payable to the reporter, but nevertheless

were reported as originated loans by the reporter who made the credit decision. The share of

loans directly submitted to and initially payable to the HMDA reporters make up 84.7 percent of

jumbo loans, 78.9 percent of FHA loans, 82.0 percent of VA loans, and 77.0 percent of RHS/FSA

loans.

39

It is also possible that the reporter made the credit decision on a covered loan or application through the actions of

an agent. For the purpose of this illustrative chart, such cases are generally similar to the cases in which the reporter

made the credit decision without delegating its underwriting authority.

40

About 15.9 percent of reported FHA loans and 13.2 percent of RHS/FSA loans were not directly

submitted to the reporting institution but were initially payable to the reporter, which are higher

than the shares of other closed-end mortgages that were not directly submitted to but were

initially payable to the reporter.

Among HELOCs, 97.8 percent of originations are from applications that were directly submitted

to the reporting institution and initially payable to the reporting institutions. About 62.3 percent

of reverse mortgages were directly submitted to and initially payable to the reporter, and 28.1

percent were not directly submitted but were initially payable to the reporter.

Overall, of all the reported HMDA originations in 2019, about 84.8 percent were directly

submitted and initially payable to the reporting institution, making it the most important

channel for reported loan originations among HMDA reporters. Loans that were not directly

submitted but were initially payable to the reporter account for about 10.3 percent of all

originations, ranked as the remote second most used channel. In contrast, in 2018 about 86.8

percent of all reported HMDA originations were directly submitted and initially payable to the

reporting institution, and 7.6 percent of all originations were not directly submitted but were

initially payable to the reporter.

Table 5.7.2 presents the distribution of closed-end originations channels by race/ethnicity,

borrowers’ age groups, and geography. Approximately 72.4 percent of Asian borrowers have

loans that were directly submitted and initially payable to the reporting institutions, compared

to 83.5 percent for Black borrowers, 78.5 percent for Hispanic White borrowers, and 84.5

percent for non-Hispanic White borrowers. The percentage of borrowers using the directly-

submitted, initially payable channel is higher for older age groups in general. More than 86

percent of borrowers aged 65 or older take out loans through the directly-submitted, initially

payable channel compared with younger groups. Additionally, 84.9 percent of borrowers

between the ages of 55 and 64 utilized the directly-submitted, initially payable channel. Nearly

86.0 percent of borrowers in rural areas and 87.0 percent of borrowers in micropolitan

statistical areas take out a loan through the directly-submitted, initially-payable channel,

compared to 82.8 percent of borrowers from metropolitan statistical areas who use the directly-

submitted, initially-payable channel.

Table 5.7.3 shows the denial rates for complete applications by application channel for each

enhanced loan type. For instance, the denial rate for the directly-submitted, initially-payable

channel of conventional conforming loans is 13.5 percent, higher than the denial rates for the

three other channels in the conventional conforming market.

41

6. Applicant/Borrower and

Property Characteristics

The 2015 HMDA Rule added or revised a number of data points that provide additional

information about the property securing, or for which the applicant is seeking, a mortgage loan,

including information about the property value and the applicant’s interest in the property on

which a manufactured home will be located. The 2015 HMDA Rule also added data points that

provide additional information about mortgage applicants, including credit scores and DTIs.

This section discusses these and other related new and revised data points.

6.1 Occupancy Type

Occupancy type is a data point that has long existed under HMDA. In the past, the occupancy

type was defined as “owner-occupied as a principal dwelling” or “not owner-occupied.” The 2015

HMDA Rule revised the enumeration of occupancy type to include the following applicable

codes: code 1 (Principal Residence), code 2 (Second Residence), and code 3 (Investment

Property).

Table 6.1.1 presents the distribution of occupancy type by enhanced loan type for originated

loans or lines of credit. About 4.68 million or 88.1 percent of conventional conforming loans are

secured by principal residences in 2019 (up from 86.2 percent in 2018), whereas 3.8 percent of

conventional conforming originations are secured by second residences in 2019 (down from 4.2

percent in 2018), and about 433,000 or 8.1 percent of conventional conforming loans are for

investment properties (down from 9.6 percent in 2018). Among jumbo loans, 87.5 percent are

for principal residences, 8.0 percent are for second residences, and 4.5 percent are for

investment properties in 2019. In contrast, the share of jumbo loans for principal residences,

second residences and investment properties were at 86.3 percent, 8.6 percent and 5.1 percent

in 2018, respectively. About 99.9 percent of FHA loans and 99.6 percent of VA originations are

for principal residences in 2019. A very small fraction of FHA loans is for investment properties.

All RHS/FSA loans are for a principal residence. All reverse mortgages are secured by principal

dwellings.

41

About 96.8 percent of HELOCs are secured by principal residences, 1.4 percent are

secured by second residences, and 1.8 percent are secured by investment properties.

41

Except for about 0.1% of reverse mortgages that are reported for investment properties. The Bureau is continuing

to research whether this is due to reporting errors.

42

Table 6.1.2 presents selected characteristics of loans by different occupancy type for

conventional conforming and jumbo loans separately.

Among conventional conforming loans, 50.8 percent of loans secured by a principal residence

are for home purchases. By contrast, 76.8 percent of conventional conforming loans secured by

second residences and 56.9 percent of the conventional conforming loans secured by investment

properties are for home purchases. Among conventional conforming loans, the median interest

rate is 4.0 percent for both loans secured by principal residences and loans secured by second

residences, and 5.0 percent for loans secured by investment properties. The median property

value collateralizing conventional conforming loans is $320,000 for principal residences,

$312,000 for second residences, and $240,000 for investment properties. The median loan

amount is $228,000 for conventional conforming loans secured by principal residences,

$225,000 for second residences, and $161,000 for investment properties.

Borrowers taking out conventional conforming loans for second residences report higher

incomes than borrowers taking out loans for principal residences. The median borrower income

for conventional conforming loans secured by second residences is $158,000, while for principal

residences it is $92,000. The median income of borrowers taking out conventional conforming

loans secured by investment properties is lower than that of second residence borrowers, but

higher than that of principal residence borrowers, at $130,000.

The median credit score of borrowers taking out conventional conforming loans secured by

principal residences is 754; for second residences, it is 776; and for investment properties, it is

763. The median CLTVs for conventional conforming loans secured by principal residences and

second residences are both 80 percent. The median CLTV for investment properties is 75

percent. The median DTI for borrowers of conventional conforming loans of all three occupancy

types are similar, with the DTI for principal-residence borrowers at 36.0 percent, for second-

residence borrowers at 35.8 percent, and for investment-property borrowers at 37.2 percent.

In 2019, the median interest rates, property values, loan amounts, and credit scores for

conventional conforming loans are all higher than 2018, and the median DTIs are lower than

2018 across all occupancy types, while the median CLTVs remain unchanged. Nevertheless,

within the conventional conforming loans of the same year, the patterns remain the same in

terms of the medians: overall, among conventional conforming loan borrowers, borrowers for

second residences have higher incomes and credit scores and take out larger loans than

borrowers of loans of the other two occupancy types. Borrowers for investment properties have

higher incomes and credit scores than the borrowers for principal residences, but they take out

smaller loans, have lower CLTVs on their properties, and pay much higher interest rates than

applicants borrowing for principal residences and second residences.

43

The same patterns generally exist among jumbo loan borrowers in terms of the medians. Jumbo

loan borrowers for second residences have significantly higher incomes and higher credit scores

than borrowers of the other two occupancy types. But unlike for the conventional conforming

loans, the median loan amount of jumbo loan secured by investment properties ($960,000) is

larger than the median loan amount of jumbo loans secured by second residences ($838,000).

In terms of the medians, jumbo loan borrowers for investment properties have slightly lower

credit scores than jumbo loan borrowers for principal residences. They take out larger loans

than borrowers of principal and second residences, but their property values are higher and

consequently are less leveraged in terms of the CLTV. Jumbo loan investment property

borrowers pay much higher interest rates than borrowers for principal residences and second

residences.

Table 6.1.3 breaks down occupancy types by race/ethnicity, age, neighborhood income, and

geographic locations for all conventional loans (including both conventional conforming and

jumbo originations). Among all racial/ethnic groups, Asians are the most likely to take out

conventional loans for investment properties. About 12.9 percent of conventional loans for Asian

borrowers are for investment properties, compared to 7.9 percent for Black borrowers, 7.5

percent for Hispanic White borrowers, and 5.6 percent for non-Hispanic White borrowers. Non-

Hispanic White borrowers are the most likely to take out loans for a second residence among all

racial/ethnic groups. Approximately 4.5 percent of non-Hispanic White conventional loan

borrowers take out loans for second residences, compared to 2.4 percent for Black borrowers,

2.3 percent for Hispanic White borrowers, and 2.8 percent for Asian borrowers. Approximately

84.3 percent of Asian borrowers’ conventional loans are for principal residences, lower than the

principal residence shares of all other groups (excluding loans where the race/ethnicity is

missing).

42

The share of conventional loan borrowers taking out loans for principal residences initially

decreases with age, falling from 97.5 percent for borrowers younger than 25 to 84.3 percent for

borrowers between the ages of 55 and 64. However, this share rises again for borrowers 65 or

older, with the principal residence share at 85.8 percent among borrowers between the ages of

65 and 74 and 87.7 percent for borrowers 74 or older.

The share of conventional loans secured by investment properties is 16.1 percent in

low/moderate-income census tracts, higher than the share for middle-income tracts (7.4

percent) and high-income tracts (5.0 percent). Conversely, the share of conventional loans

42

In our categorization of race and ethnicity, the “missing” category includes both the applications for which the race

and ethnicity are not reported and the applications under which the race and ethnicity are not applicable. In the

latter, the borrowers are non-natural persons and the share of investment property among them is generally high.

44

secured by principal residences is 80.7 percent in lower/moderate-income tracts, lower than the

shares in middle-income or high-income tracts.

The share of conventional loans secured by principal residences is 73.1 percent in rural areas;

15.2 percent of loans in rural areas are for second residences, a much higher share than in

micropolitan and, particularly, metropolitan statistical areas, which feature 10.3 percent and 3.3

percent shares, respectively. Unlike in 2018, loans for investment properties are relatively more

common in rural areas (at 11.7 percent) than in metropolitan and micropolitan statistical areas,

both at 7.8 percent in 2019. In contrast, in 2018, 9.3 percent of loans in metropolitan and

micropolitan statistical areas and 7.6 percent of loans in rural areas were for investment

properties.

Table 6.1.4 shows the action type by occupancy type for conventional conforming and jumbo

LARs. It is noticeable that the origination rates are higher for loans secured by second

residences than those for other occupancy types, for both conventional conforming and jumbo

loans.

6.2 Property Value

The DFA, as implemented by the 2015 HMDA Rule, requires lenders to report the values of the

properties securing the covered loans or, in the case of applications, the proposed covered loans.

Property Value is one of the Mandated Data Points as discussed in the introduction section of

this article. The reported values are the values relied upon in making the credit decisions.

Property Value is one of the data points that institutions that qualify for the EGRRCPA partial

exemption are not required to report. Property Value is entered in numeric form except for “NA”

values, which are entered if the requirement to report property value does not apply, or

“Exempt,” which is entered if the reporter is exempt under the EGRRCPA from reporting this

data point for the transaction. Property value is disclosed in the public loan-level 2019 HMDA

data as the midpoint for the $10,000 interval into which the reported value falls.

43

Table 6.2.1 lists the mean and median property values for properties securing the originated

loans for each enhanced loan type. The median property value securing conventional

conforming loans is $313,000, while the median property value securing jumbo loans is

significantly higher at $1,170,000. The median property value securing RHS/FSA loans is the

lowest among all enhanced loan types at $148,000. The median value of properties securing

FHA loans is higher than that of RHS/FSA loans but lower than that of other loan types, at

43

For example, for a reported loan amount or property value of $117,834, the Bureau would disclose $115,000 as the

midpoint between values equal to $110,000 and less than $120,000.

45

$224,000. The median value of properties securing VA loans is $280,000, $344,000 for

HELOCs, and $351,000 for properties securing reverse mortgages. Mean property values are

higher than the median values but show the same patterns across enhanced loan types.

Compared to 2018, the median property value of each enhanced loan type has increased.

Table 6.2.2 further breaks down the median value of properties by enhanced loan type, loan

purpose, occupancy type, and lien status for closed-end originations. The median property value

of cash-out refinances is generally lower than that of non-cash-out refinance loans in 2019,

which is the opposite of the cases in 2018 except for jumbo loans. Also, in 2019, both the median

property values of cash-out and non-cash-out refinance loans are higher than that of home-

purchase loans for each enhanced loan type, while in 2018 the pattern of differences between

the median property values of home purchase loans and refinance loans vary across enhanced

loan types.

The median property value of second residences securing jumbo loans is $1.25 million,

compared to the median property value of jumbo loans for principal residences at $1.15 million;

the median property values of principal- and second-residences securing conventional

conforming loans differ by only $8,000. Investment properties have lower median values than

principal residences and second residences for all loan types except jumbo loans.

6.3 Loan Amount and Conforming Loan Flag

Loan Amount is a data point that has long been reported and disclosed under HMDA. Prior to

the 2015 HMDA Rule, loan amount was rounded to the nearest thousand dollars, and it was

disclosed to the public at the loan-level without modification. The 2015 HMDA Rule requires

financial institutions to report in dollars the exact amount of the covered loan or the amount

applied for. Loan amount is disclosed in the public loan-level 2019 HMDA data as the midpoint

for the $10,000 interval into which the reported value falls.

The public loan-level 2019 HMDA data also contain a flag indicating whether the reported loan

amount exceeds the annual maximum principal loan balance for a mortgage eligible to be

acquired by Fannie Mae and Freddie Mac (the “GSE Conforming Loan Limits”) at the time of

application or origination. This is a field derived in preparing the public dataset from the

reported loan amount or amount applied for and the GSE Conforming Loan Limits published by

the Federal Housing Finance Agency (FHFA).

Throughout this Data Point article, analyses relating to loan amount use the exact amount as

reported by the reporter. This Data Point article uses the GSE conforming loan flag and loan

type reported in the HMDA data to identify the conventional conforming loans and applications.

46

6.4 Credit Score

The DFA, as implemented by the 2015 HMDA Rule, requires lenders to report information on

the credit scores of applicants and co-applicants. Credit Score is one of the Mandated Data

Points as discussed in the introduction section of this article. Credit scores are reported in four

standard data fields plus two free form text fields: Credit Score of Applicant or Borrower; Credit

Score of Co-applicant or Co-borrower; Name and Version of Credit Scoring Model for Applicant

or Borrower; Name and Version of Credit Scoring Model for Co-applicant or Co-borrower;

Conditional Free Form Text Field, if code 8 (Other credit scoring model) is chosen for Name and

Version of Credit Scoring Model for Applicant or Borrower; and Conditional Free Form Text

Field, if code 8 (Other credit scoring model) is chosen for Name and Version of Credit Scoring

Model for Co-applicant or Co-borrower. Institutions that qualify for the partial exemption under

the EGRRCPA are not required to report any of the credit score information fields. Credit score

and free form text fields used to report the name and version of credit scoring models are

excluded from the public loan-level 2019 HMDA data.

6.4.1 Name and Version of Credit Scoring Model

Lenders are required to report the names and versions of the credit scoring models used to

generate the credit scores relied upon in making credit decisions regarding

applicants/borrowers and co-applicants/co-borrowers, if applicable. The 2015 HMDA Rule and

2019 FIG allow the following standard enumerations for the name and version of credit scoring

models: code 1—Equifax Beacon 5.0; code 2—Experian Fair Isaac Risk Model v2; code 3—

TransUnion FICO Risk Score Classic 04; code 4—TransUnion FICO Risk Score Classic 98; code

5—Vantage Score 2.0; code 6—Vantage Score 3.0; code 7—More than one credit scoring model;

code 8—Other credit scoring model; code 9—Not applicable; code 10—No co-applicant. Codes 1,

2, 3, and 4 are all variations of FICO scores that are calculated and named by different consumer

reporting agencies based on generic and proprietary FICO formulas and credit information at

each of the three major consumer reporting agencies.

Table 6.4.1a shows the frequency distribution of the reported name and version of credit scoring

models for the borrowers. Approximately 29.2 percent of originated loans that reported this

information reported Equifax Beacon 5.0 as the model relied on for the borrower’s score, 24.1

percent reported Experian Fair Isaac Risk Model v2, and 25.8 percent reported TransUnion

FICO Risk Score Classic 04. Vantage Scores, the main alternative in the marketplace to FICO

scores, account for 0.3 percent of all originated loans that reported the borrower credit scoring

models and versions, including Vantage Score 2.0 and Vantage Score 3.0. Another 4.5 percent

reported “More than one scoring model” and 6.3 percent reported “Other credit scoring model.”

A closer examination of the Conditional Free Form Text Field, if “Other credit scoring model” is

47

chosen, indicates that an overwhelming majority of those filling in this free form text field

named some other variation of FICO scoring models and versions not listed in the standard

enumeration of the 2019 FIG, most commonly TransUnion FICO Risk Score Classic 08, Equifax

FICO Score Beacon 09, Experian FICO Risk Model 08, and FICO Risk Score Classic 09.

Table 6.4.1b shows the frequency distribution of the reported name and version of credit scoring

models for the co-borrower. Approximately 52.4 percent of applicants do not have co-applicants

or co-borrowers, and 25.2 percent reported this field as not applicable. Similar to the borrower

credit score model and version field, Equifax Beacon 5.0 is the most commonly reported

model/version for co-borrowers, followed by TransUnion FICO Risk Score Classic 04 and

Experian Fair Isaac Risk Model v2. Vantage Scores similarly account for a small fraction of

credit scoring models used in reported loan originations. Examination of the Conditional Free

Form Text Field reveals that an overwhelming majority of those filling in the free form credit

score model/version text field for co-borrowers used FICO 9.

6.4.2 Credit Score Values

The credit scores are reported as numbers with a special code 7777 indicating “it is not a

number,” code 8888 indicating “NA,” code 9999 indicating “no co-applicant” and code 1111

indicating “exempt.”

Different credit scoring models may add complexity to the analysis. Because the credit decision

process of mortgages commonly requires pulling credit scores from more than one credit

reporting agency, and the final credit score used could be any of the credit scores pulled based

on industry guidelines and common practice

44

, for tractability, the analyses in this article treat

all variations of credit scoring models equally, except for Vantage Score 2.0, which has a

different range than FICO scores and Vantage 3.0, and hence is omitted from the analyses.

45

Furthermore, the analyses combine the credit score for the applicant/borrower with the credit

score for the co-applicant/co-borrower by taking the lower of the two credit scores when both

are reported.

Table 6.4.2 shows the mean and median credit scores of originated loans by enhanced loan type.

It also shows the 5th percentile, 25th percentile, 75th percentile and 95th percentile in 2019. The

44

For example, see Fannie Mae Selling Guide describing Fannie Mae’s requirements for credit scores available at

https://www.fanniemae.com/content/guide/selling/b3/5.1/01.html, or Freddie Mac’s selling and servicing

requirements on selection and use of credit scores available at

http://www.freddiemac.com/learn/pdfs/uw/credit_scores.pdf.

45

For the analysis presented in this article, all credit scores with a valid value between 300 and 850 under the

reported credit scoring models, other than VantageScore 2.0 that has a valid score range between 501 and 990, are

used. The Bureau is continuing to research the implications of credit scores by different credit scoring models.

48

conventional conforming borrowers’ median credit score is 756 and their mean is 747, with the

5th percentile at 659, the 25th percentile at 715, the 75th percentile at 786, and the 95th percentile

at 809. Conventional jumbo loans have the highest mean and median scores among closed-end

mortgages, with a mean score of 765 and a median of 774. The 5th percentile of jumbo loan

borrowers’ credit score is 700 (meaning that 5 percent of borrowers have scores at or below 700

and the remaining 95 percent of borrowers have scores above 700), the 25th percentile is 746,

the 75th percentile is 791, and the 95th percentile is 808. FHA borrowers have the lowest mean

and median scores among closed-end mortgages, with a mean score of 668 and a median of 663.

The bottom 5th percentile of FHA borrowers’ credit scores is 599, the 25th percentile is 637, and

the 75th percentile is 695. The 95th percentile of FHA borrowers’ scores is 757. RHS/FSA loan

borrowers have mean and median scores higher than FHA borrowers, at 696 and 691, but

slightly lower than VA loan borrowers, whose mean credit score is 709 and median credit score

is 709 as well. The mean credit score for HELOC borrowers is 764 and the median is 773, both

very close to those of jumbo loans, and higher than those of all other closed-end enhanced loan

types. Reverse mortgage borrowers have a mean credit score of 729 and a median credit score of

749.

46

The last column of Table 6.4.2 reports the standard deviation of the credit scores.

Notably, in comparison, as reported in the 2018 Article, in 2018 the median credit score of

conventional jumbo loans was 771 and the mean was 762, the median credit score of

conventional conforming borrowers was 750 and their mean is 742, all lower than the values in

2019. On the other hand, the mean and median credit scores for FHA and RHS/FSA loans in

2019 remained largely unchanged from those in 2018.

Figure 6.4.1 provides complete histograms of the distribution of credit scores for originated

loans by enhanced loan type. Each bar depicted in the figures covers a credit score bin of 10

points. The reference line marks the credit score at 620, a common benchmark below which

borrowers are regarded as subprime. The patterns shown in Figure 6.4.1 are consistent with the

description provided above, but such a figure shows more details. For instance, one can see from

Figure 6.4.1 that credit scores for jumbo loans are more concentrated on the higher end with a

longer and steeper rising curve before its peak than other enhanced loan types; the peak of the

credit score distribution for FHA loans is near 640, to its right the histogram has a long

downward slope, and a not-insignificant percentage of FHA borrowers have credit scores below

620. The distribution of credit scores for VA borrowers is much flatter (i.e. more evenly

distributed) than the score distribution for other enhanced loan types.

46

According to the 2015 HMDA Rule, the lenders would only report credit scores if they were relied upon in the credit

decision. Note that of a little more than 32,000 reverse mortgage originations, only about 2100 had credit score

reported under HMDA. The mean and median credit scores of reverse mortgage borrowers shown in Table 6.4.2 are

based on those whose credit scores are reported, and should be interpreted with caution.

49

Table 6.4.3 provides the median credit scores of different enhanced loan types, broken down by

loan purpose, occupancy type, and lien status. Among conventional mortgages, the borrowers of

cash-out refinance loans have median credit scores lower than non-cash-out borrowers for both

conventional conforming and jumbo loan types. The median credit score of borrowers of home-

purchase loans is slightly higher than borrowers of non-cash-out refinance loans for jumbo

loans, but the median credit score of borrowers of home-purchase loans is lower than borrowers

of non-cash-out refinance loans for conventional conforming loans. Borrowers of loans secured

by a second residence have higher median scores than borrowers of principal residences for both

conventional conforming and jumbo loans. Borrowers of loans secured by a subordinate lien

have lower median scores than borrowers of loans secured by first lien for both conventional

conforming and jumbo loans.

Table 6.4.4 breaks down the median credit scores of different enhanced loan types by

race/ethnicity, borrowers’ age group, neighborhood income, and geography in 2019. Asian

borrowers have the highest median credit scores overall and across most enhanced loan types.

Their overall median credit score is 763 (up from 759 in 2018). Black borrowers have the lowest

overall median credit score, at 694 (up from 691 in 2018). Across each enhanced loan type, the

median credit score of Black borrowers is also the lowest in comparison to other racial/ethnic

groups. The overall median credit score for Hispanic White borrowers is 714 (up from 710 in

2018), the second lowest among all racial/ethnicity groups. Similarly, the median credit scores

for Hispanic White borrowers are lower than non-Hispanic White and Asian borrowers and

higher than Black borrowers for most enhanced loan types, except for FHA loans and RHS/FSA

loans in which the median credit score of non-Hispanic White borrowers are slightly lower. The

median credit score overall for non-Hispanic White borrowers is 752 (up from 748 in 2018),

lower than Asian borrowers but higher than Black and Hispanic White borrowers.

The oldest borrower age groups generally have higher median credit scores than the youngest

borrower age groups among conventional loan borrowers. But the variations in median credit

scores across different age groups among FHA, VA, and RHS/FSA loan borrowers are quite

limited and not monotonic. The median credit score of borrowers from high-income tracts is

higher than that of borrowers from middle-income tracts across all enhanced loan types, and the

median credit score of borrowers from middle-income tracts in turn is higher than that of

borrowers from low/moderate-income tracts, overall and across all enhanced loan types except

for FHA and jumbo loans. The median credit score of borrowers from metropolitan statistical

areas is higher than that of borrowers from micropolitan areas, who in turn have a median credit

score higher than borrowers from rural areas overall, and that pattern is generally true of

different enhanced loan type except for jumbo loans for which the median credit scores are very

similar across metropolitan, micropolitan and rural areas.

50

Among all applications, Figure 6.4.2 presents the histogram of credit score distribution

separately for each enhanced loan type. Again, the size of each bar represents a score bin with a

range of 10 points. The vertical reference line drawn in these figures corresponds to a credit

score of 620. Overall, the credit score profile of applicants for FHA loans is to the left of the

credit score profiles of all other loan types signifying that the scores skew lower, and the credit

score profile of HELOC borrowers is to the right of other enhanced loan types. There are some

big drops (bunching) of credit scores at 620 among applicants for conventional conforming

loans, FHA loans, VA loans, and RHS/FSA loans. Some other bunching points exist as well, such

as at 580, 600, and 640 for FHA applications, 640 for RHS/FSA applications, and 680 and 700

for jumbo loan applications. Such bunching possibly implies that some potential applicants with

a credit score below certain thresholds were either discouraged by the lenders from applying or

on their own avoided applying for a mortgage in anticipation of the high likelihood of rejection.

Figures 6.4.3.1 to 6.4.3.6 show for each of the enhanced loan types except for reverse mortgages,

the distribution of credit score among all applicants, grouped by race and ethnicity.

As depicted by Figure 6.4.3.1, among all applicants for conventional conforming loans, the

distribution of scores has the longest left tail for Black applicants, indicating a larger share of

applicants at the lower end of the credit score spectrum. Particularly, there is a relatively larger

percentage of Black applicants for conventional conforming loans who have credit scores below

620. The overall profile of Black applicants of conventional conforming loans is to the left of

other groups, indicating that their scores skew lower than for other racial and ethnic groups.

Hispanic White conventional conforming loan applicants’ credit score profiles are similar to

those of Black applicants, but slightly to the right, i.e. towards relatively higher credit scores.

The “Other” group (including Native American and Hawaiian Islander) who applied for

conventional conforming loans also have credit score profiles similar to Hispanic White and

Black applicants. Asian applicants’ credit score distribution concentrates on a higher credit

score range than other groups, and only a small percentage fall below 620. Non-Hispanic White

applicants’ profiles are largely similar to those of Asian applicants, though the non-Hispanic

White profile has a lower peak in the high score range, indicating that a smaller share of these

applicants have scores at the high end of the range.

{Figure 6.4.3.2 presents histograms of the credit score of applicants for jumbo loans by race and

ethnicity. Similar to the conventional conforming market, Black applicants’ score distribution

features a longer left tail than other groups, with a relatively larger percentage of Black

applicants’ credit scores falling below 620. The overall profile of Black applicants of jumbo loans

is also flatter compared to that of other groups, indicating a smaller share of applicants with

higher scores and a tendency towards the lower end of the credit spectrum. Hispanic White

jumbo loan applicants’ credit scores have a smaller tail below 620 than that of Black applicants,

and their overall profile is slightly to the right of Black applicants. The “Other” group (including

51

Native American and Hawaiian Islander) who applied for jumbo loans have credit score profiles

similar to Hispanic White applicants. The Asian, non-Hispanic White, and “Joint” applicants’

credit score distributions more heavily concentrate in higher credit score ranges.

The divergence in credit score distributions between different racial/ethnic groups is much

smaller among applicants for FHA loans than the divergence in conventional markets, as

depicted by Figure 6.4.3.3. Overall, each group’s credit scores are more narrowly concentrated,

with peaks near 650, and each has a noticeable percentage of applicants with a credit score

below 620.

Figure 6.4.3.4 shows that the credit score distributions for different racial/ethnic groups are less

divergent among applicants for VA loans than for applicants for conventional loans, but still

more dispersed than the score distributions for FHA applicants. The left tail of the score

distribution is larger for Black and “Other” applicants than non-Hispanic White applicants. The

left tail of Hispanic White applicants’ credit score distribution is slightly larger than for non-

Hispanic White applicants. The credit score distribution of Asian applicants’ peaks to the right

of non-Hispanic White applicants.

Figure 6.4.3.5 similarly demonstrates that a relatively larger share of Black applicants’ credit

scores for RHS/FSA loans are below 620 than for non-Hispanic White applicants. RHS/FSA

Hispanic White applicants’ credit score distribution is similar to that of the non-Hispanic White

applicants. The credit score distribution of Asian applicants is more symmetric than other

groups and peaks to the right of non-Hispanic White applicants.

Figure 6.4.3.6 shows the histogram of credit scores of HELOC applications by race and ethnicity.

The distribution of credit scores for Black applicants is to the left of all other groups. The credit

score distributions of Hispanic White applicants and “Other” applicants are slightly to the right

of Black applicants. The score distributions of Asian and non-Hispanic White applicants (as well

as Joint applicants) are noticeably more concentrated in the higher score range than the score

distribution of Black, Hispanic White, and Other applicants.

Credit scores are widely used in credit decisions and are among the most significant factors in

mortgage underwriting and pricing. HMDA data have consistently shown that denial rates for

Hispanic White, Black, and Native American applicants generally are higher than denial rates

for non-Hispanic White and Asian applicants.

47

Credit scores were not collected in the HMDA

47

As examples, see the CFPB Data Point articles titled “2019 Mortgage Market Activity and Trends”, available at

https://www.consumerfinance.gov/data-research/research-reports/cfpb-data-point-2019-mortgage-market-activity-

and-trends/ published on June 24, 2020, and the CFPB Data Point Article titled “2018 Mortgage Market Activity and

Trends”, available https://www.consumerfinance.gov/data-research/research-reports/introducing-new-revised-data-

points-hmda/ published on August 30, 2019.

52

data prior to 2018. As demonstrated above, the 2019 HMDA data show that the credit scores of

Black and Hispanic White applicants, on average, are lower than those of non-Hispanic White

and Asian applicants’ overall and across all enhanced loan types. Additionally, there are higher

percentages of Black and Hispanic White applicants whose credit scores fall on the low end of

the distribution and fall below the common underwriting cutoff points. These new data make it

possible for users of non-public HMDA data to analyze denial rates and pricing differentials

after controlling for credit scores (and other variables discussed in this article).

To demonstrate the importance of credit scores in underwriting decisions, Figure 6.4.4 creates a

binscatter plot relating the credit scores to the denial rates for all loan types except reverse

mortgages. The sample is limited to first-lien, principal-residence, and site-built single-family

properties. To create this graph, within each enhanced loan type, the credit scores of all

applicants with complete applications (HMDA action code equal to 1, 2, or 3) are grouped into

20 equal sized bins, i.e. each bin contains the same number of applicants. The average credit

score of applicants for a particular loan type in a credit score bin is shown on the horizontal axis,

and the average denial rates for these applicants of that loan type and that score bin are shown

on the vertical axis. Figure 6.4.4 demonstrates that, on average, the denial rate decreases with

the credit score for each enhanced loan type.

48

Credit scores, though important, are not the only factors used in lenders’ underwriting and

pricing decisions. Analyzing the denial decisions of mortgage underwriting should not be based

on bivariate analysis alone that only examines the relationship between the underwriting

decision and one single credit risk factor. In general, a multivariate approach, typically in the

form of multivariate statistical regression, should be used to explore the relationship between

credit outcomes and the applicants or borrowers’ characteristics, by controlling for relevant

factors, such as applicants’ credit characteristics, product features, underwriting and pricing

policies of lenders, and many others. However, such analyses would require additional

information, some of which is not available in HMDA data, and further, more sophisticated,

analyses may be needed that are beyond the scope of this introductory article to 2019 HMDA

data.

To illustrate how bivariate analysis could provide important insight, but alone may not provide a

complete picture and may even be misleading when viewed in isolation, Figure 6.4.5 creates a

binscatter plot relating the denial rates to credit scores of applicants for conventional

conforming 30-year fixed-rate mortgages for different racial/ethnic groups. The sample is

restricted to home purchase, first lien, and principal residence. A visual examination of the

48

This is with the exception of some right tails in the very high score ranges which slightly fluctuates and becomes

slightly upward sloping. The average denial rates in such high score ranges are generally very low and slight upward

sloping could be driven by idiosyncrasies.

53

figure demonstrates that, while denial rates are inversely correlated with credit scores on

average, among the applicants for conventional conforming 30-year fixed-rate mortgages for

home purchase, secured by principal residences and first liens, Black and Hispanic White

applicants are on average denied at a higher rate than non-Hispanic White applicants, even if

they are within the same credit score range.

However, a bivariate analysis alone, such as the one presented in Figure 6.4.5, may potentially

mask other factors which may interact with credit score and race/ethnicity. Figures 6.4.5 and

6.4.6, viewed together, illustrate both the relevance and the limitations of simple bivariate

analysis.

Figure 6.4.6 shows the relationship between credit scores and CLTV for different groups using

the same sample as the one underlying Figure 6.4.5, i.e. limited to applicants for conventional

conforming 30-year fixed-rate mortgages, for home purchases, secured by a first lien and

principal residence. As Figure 6.4.5 shows, for applicants within the same credit score range,

Black and Hispanic White applicants on average have higher CLTVs than non-Hispanic White

applicants. Given that CLTV is another important factor in underwriting decisions, this

additional observation may help partially explain the differences in denial rates between

different groups based on the credit score alone. It is beyond the scope of this article to assess

how much of the disparities in denial rates could be due to the differences in credit scores, or

CLTVs, or myriad other factors, all of which could be correlated among themselves. However, as

Figures 6.4.5 and 6.4.6 illustrate, such issues are highly complex, and one factor or a limited set

of factors alone could not lead to definite conclusions and should be viewed with caution. In

summary, HMDA data show that non-Hispanic White and Asian applicants are often denied at a

lower rate than Blacks, Hispanic Whites, and other minorities. Many underwriting factors now

available in HMDA data such as credit score and CLTV explain some of these disparities, but

data on other factors and detailed lender-level information on underwriting policies and

products that HMDA data do not include are needed to fully understand these disparities.

e d

6.5 CLTV

The 2015 HMDA Rule added combined loan-to-value ratio (CLTV) as a new data point starting

in the 2018 HMDA data. CLTV is one of the Discretionary Data Points as discussed in the

introduction section of this article. Reporters are required to report the ratio of the total amount

of debt secured by the property to the value of the property relied upon in making the credit

54

decision as a percentage.

49

CTLV is one of the data points that institutions that qualify for the

EGRRCPA partial exemption are not required to report. A reporter would report “NA” if the

requirement to report CLTV does not apply to the covered loan or application, or “Exempt” if the

reporter is exempt under the EGRRCPA from reporting this data point for the transaction.

Table 6.5.1 shows some summary statistics of the CLTVs of originated loans for different

enhanced loan types in 2019. The median CLTV for conventional conforming loans is 79.4

percent (down slightly from 80 percent in 2018). Their 5th percentile is 36.4 percent (meaning

that 5 percent of loans have a CLTV at or below 36.4 percent and the remaining 95 percent have

CLTVs above that level); their 75th percentile is 89.7 percent; and their 95th percentile is 97

percent. The median CLTV for jumbo loans is 76.5 percent (down from 79.5 percent in 2018).

Their 5th percentile is 43.1 percent, their 25th percentile is 65.2 percent, their 75th percentile is 80

percent, and their 95th percentile is 90 percent. The median CLTV for FHA loans is 96.5 percent.

Their 5th percentile is 73.9 percent, their 25th percentile is 90.0 percent, their 75th percentile is

96.5 percent, and their 95th percentile is 100.4 percent. The median CLTVs for VA loans and

RHS/FSA loans are both 100.0 percent. The median CLTV for HELOC originations is 71.1

percent; their 5th percentile is 19 percent, their 25th percentile is 50 percent, their 75th percentile

is 80 percent, and their 95th percentile is 90 percent. The median CLTV for reverse mortgages is

49.6 percent, lower than that of both HELOCs and closed-end mortgages.

50

The last column of

Table 6.5.1 reports the standard deviation of the CLTVs.

CLTV may vary significantly between home-purchase loans and refinance loans. Table 6.5.2a

presents the median CLTVs of different enhanced loan types by race/ethnicity, age,

neighborhood income, and geography for closed-end home-purchase loans; Table 6.5.2b

mirrors Table 6.5.2a, presenting the same information for closed-end refinance loans (including

both cash-out refinance and non-cash-out refinance loans).

49

The 2015 HMDA Rule did not add loan-to-value ratio (LTV) as a new data point. One can theoretically calculate the

LTV from the loan amount and the property value in HMDA data by taking the ratio of the two. However, such LTV

calculation may be subject to three constraints. First, the loan amount on the note reported under HMDA may be

different from the loan amount used for LTV calculation by the lenders per their underwriting and/or pricing policies.

Especially for FHA, VA, and RHS/FSA loans, the upfront mortgage insurance premium or funding fees are often

financed through the loan and the financed amount is added to the mortgage note, while for qualifying purposes FHA,

VA or RHS/FSA programs typically exclude such financed insurance premium or funding fees from its LTV and CLTV

calculation. Second, different lenders may use different rounding rules for LTV that they rely on. Third, for users of

public HMDA data, the loan amount and property values are both disclosed at the mid-point of 10,000-dollar

intervals, which leads to a loss of precision when trying to divide the loan amount by property value in order to derive

LTV.

50

According to the 2015 HMDA Rule, the lenders would only report CLTVs if they were relied upon in the credit

decision. Note that of a little more than 34,800 reverse mortgage originations, only about 1,900 had CLTV reported

under HMDA. The mean and median CLTV of reverse mortgages in Table 6.5.1 are based on those whose CLTVs are

reported and should be interpreted with caution.

55

As shown in Table 6.5.2a, the median CLTVs of Black and Hispanic White home buyers taking

out conventional conforming loans are 95 percent and 91 percent respectively, while the median

CLTVs for both Asian and non-Hispanic White conventional conforming loan home buyers are

80 percent. The median CLTV for home-purchase jumbo loans is 80 percent for every

racial/ethnic group except for Black borrowers whose median CLTV in 2019 is 81.7 percent. The

median CLTV for each racial/ethnic group of home-purchase FHA borrowers is 96.5 percent

and the median CLTV for each group of home-purchase VA borrowers is 100 percent. The

median CLTV for each racial/ethnic group of home-purchase loan borrowers among RHS/FSA

borrowers is at or slightly over 100 percent.

The median CLTV for home-purchase loans generally decreases with age for conventional

conforming loans (except for the 45-54 and 55-64 age groups who have the same median CLTV

at 80 percent). The median CLTV for jumbo home-purchase loans stays at 80 percent for age

groups younger than 64 and drops to 75 percent for borrowers 75 years or older. The median

CLTVs for non-conventional home-purchase loans (FHA, VA, RHS/FSA loans) do not vary with

age, with the median CLTV highly concentrated near the program limits for government

mortgages. Among conventional conforming home-purchase loans, the median CLTV for loans

in low/moderate-income census tracts (90 percent) is higher than that of middle-income tracts

(85 percent), which is in turn higher than that of high-income tracts (80 percent). There is little

variation in median CLTV by census tract income within all other enhanced loan types. The

median CLTVs in rural areas are slightly lower than that in metropolitan statistical areas and

micropolitan areas among conventional conforming and jumbo loans.

Table 6.5.2b presents the information similar to Table 6.5.2a for closed-end refinance loans.

Overall, the median CLTV of refinance loans is much lower than for home-purchase loans within

any given enhanced loan type. The median CLTV for Black borrowers who refinanced using

conventional conforming loans is 72 percent. This is only 1.3 percentage points higher than the

median CLTV for non-Hispanic White borrowers who refinanced using conventional

conforming loans with a median CLTV at 70.7 percent. This finding is in sharp contrast to the 15

percentage point gap between the median CLTV of Black home buyers and non-Hispanic White

home buyers using conventional conforming loans to finance their home purchases. The median

CLTV for Hispanic White refinance conventional conforming loan borrowers is 70 percent,

slightly lower than that of non-Hispanic White borrowers. In comparison, the median CLTV of

Hispanic White borrowers for home-purchase conventional conforming loans is higher than that

of non-Hispanic White borrowers who take out home-purchase conventional conforming loans,

as shown in Table 6.5.2a. The median CLTV for Asian conventional conforming refinance

borrowers is 69 percent, lower than that for each other racial/ethnic group. The median CLTV of

Black refinance jumbo loan borrowers (at 75 percent) is higher than other groups.

56

The median CLTVs for FHA, VA, and RHS/FSA refinance loans are all significantly lower than

for purchase loans in respective government loan programs, and there is little dispersion among

different racial/ethnic groups in these programs.

The median CLTVs for refinance conventional conforming loans and for refinance jumbo loans

both generally decrease with age.

The variation of the median CLTVs for refinance loans are generally small across different

income tracts and urban/rural areas within each enhanced loan type.

Figures 6.5.1a and 6.5.1b show histograms of CLTVs for conventional conforming loans for

home purchase and refinance, respectively. The CLTVs for conventional conforming home-

purchase loans are clearly bunched at 80 percent, 90 percent, 95 percent, 97 percent, and a few

other less pronounced values. The CLTVs for conventional conforming refinance loans have a

peak at 80 percent and are distributed more or less smoothly to the left of it (with a few minor

peaks at 75 percent, 70 percent, and 60 percent, for instance) and have a small right tail with

localized peaks at 85 percent, 90 percent, and 95 percent.

Figures 6.5.2a to 6.5.2b show the histograms of CLTVs for jumbo home-purchase loans and

jumbo refinance loans respectively. The CLTVs for jumbo home-purchase loans bunch most

prominently at 80 percent, with a number of localized bunching points to either side. The

distribution of CLTVs for jumbo refinance loans is largely similar to that of CLTVs for

conventional conforming refinance loans, with a peak at 80 percent, a wide left tail and several

other minor bunching points.

Figures 6.5.3a through 6.5.3b feature histograms for the CLTVs of FHA home-purchase and

refinance loans. The CLTVs for FHA home-purchase loans are heavily bunched at 96.5 percent,

with over 60 percent of FHA home-purchase loan borrowers making the minimum 3.5 percent

down payment under the FHA program. Another 10 percent have CLTVs at 98 percent and

about three to four percent have CLTVs over 100 percent. About 23 percent of FHA refinance

loans had CLTVs of 85 percent (down from about 31 percent in 2018). There is another small

mass (about 10 percent) of FHA refinance loans with CLTVs at or slightly over 96.5 percent, and

another about 7 percent of FHA refinance loans have CLTV at 80 percent.

Most VA home-purchase loans have CLTVs at 100 percent, as is shown in Figure 6.5.4a.

Similarly, there is a bunching point at CLTVs of 100 percent for VA refinance loans, as is shown

in Figure 6.5.4b. The remaining CLTVs are distributed mostly smoothly to the left (with the

exception of two localized peaks around 90 percent).

57

The CLTVs of RHS/FSA home-purchase loans rise smoothly until they spike at 100 percent as

depicted by Figure 6.5.5a. The distribution of CLTVs for RHS/FSA refinance loans is more

dispersed than home-purchase loans but still peak near 100 percent, as shown in Figure 6.5.5b.

Compared to closed-end mortgages, the CLTVs of HELCOs are much more dispersed. The

CLTVs of HELOC originations have a very wide and mostly smooth rising tail until it spikes near

80 percent as depicted by Figure 6.5.6. The CLTVs of 90 percent and 85 percent are two other

relatively common values for HELOCs.

6.6 DTI

The Debt-to-Income Ratio (DTI) is one of the data points that was first reported in the 2018

HMDA data. DTI is one of the Discretionary Data Points as discussed in the introduction section

of this article. A reporter is required to report DTI as a percentage, which reflects the ratio of an

applicant’s or borrower’s total monthly debt to total monthly income relied upon in making the

credit decision.

51

DTI is one of the data points that institutions that qualify for a partial

exemption under the EGRRCPA are not required to report. Reporters enter “NA” if the

requirement to report DTI does not apply or may enter “Exempt” if they are eligible for a partial

exemption under the EGRRCPA.

DTI is binned into the following ranges in the public loan-level 2019 HMDA data: less than 20

percent, greater than or equal to 20 percent and less than 30 percent, greater than or equal to 30

percent and less than 36 percent, greater than or equal to 50 percent and less than 60 percent,

and greater than or equal to 60 percent. Reported DTI greater than or equal to 36 percent and

less than 50 percent is disclosed in the public loan-level 2019 HMDA data without modification.

The discussion of the values of DTI in this article uses the DTI values as reported in 2019 HMDA

data rather than the partially binned values in the publicly released data to provide the public

greater insight.

Table 6.6.1 presents basic summary statistics of reported DTI for originated loans of different

enhanced loan types in 2019. The median DTI for conventional conforming loans is 36 percent

51

Note the DTI required to be reported by HMDA corresponds to what is also commonly known as the “back-end

DTI” that is calculated by using the applicant’s or borrower’s total monthly debt, including the mortgage debt or

housing expenses plus other debts such as credit card debts and car loans, divided by gross income. There is another

type of DTI, known as “front-end DTI” that lenders often also rely on in making the credit decisions. The front-end

DTI is calculated by using the applicant’s or borrower’s housing expenses, including their monthly payments on

mortgage principal, interest, insurance and tax, but excluding other debts such as credit card debts and car loans,

divided by gross income.

58

(down from 37 percent in 2018). Their 25th percentile is 28 percent (meaning that 25 percent of

these loans have a DTI at or below 28 percent and the remaining 75 percent of loans have DTIs

higher than 28 percent), their 75th percentile is 43 percent, and their 95th percentile is 49

percent. The median DTI for jumbo loans is 34 percent (down from 36 percent in 2018). The

25th percentile of jumbo loan DTIs is 27 percent, the 75th percentile is 40 percent, and the 95th

percentile is 47 percent. The median DTI for FHA loan borrowers is 44 percent (unchanged

from 2018); the 25th percentile is 37 percent, which is near the median DTI for conventional

conforming as well as jumbo loan borrowers. The 75th percentile of DTI for FHA borrowers is 50

percent, and the 95th percentile is 56 percent. The DTI distribution for VA borrowers is similar

to that of FHA borrowers. The median DTI of VA borrowers is 41 percent (down from 42 percent

in 2018). The 25th percentile is 33 percent, the 75th percentile is 48 percent, and the 95th

percentile is 58 percent. The median DTI of RHS/FSA borrows is 36 percent (unchanged from

2018), the 75th percentile is 40 percent, and the 95th percentile is 44 percent. Among HELOC

borrowers, the median DTI is 35.6 percent, the 25th percentile is 26.7 percent, the 75th percentile

is 43.0 percent, and the 95th percentile is 53 percent. The last column of Table 6.6.1 reports the

standard deviation of the DTIs.

Table 6.6.2 features median DTIs for different enhanced loan types by loan purpose, occupancy

status, and lien status for closed-end originations. The median DTIs of home-purchase loan

borrowers are slightly higher than those of non-cash-out refinance loan borrowers within each

enhanced loan type. The median DTIs of cash-out refinance loan borrowers are also slightly

larger than those of non-cash-out refinance loan borrowers within each enhanced loan type. The

median DTI of borrowers for second residence loans is somewhat lower than that of borrowers

for principal residences as well as borrowers for investment properties, for jumbo borrowers.

The median DTI of borrowers for loans secured by first liens is equal to borrowers for loans

secured by subordinate liens for jumbo loans, but is slightly lower within the conforming loan

space.

Table 6.6.3 presents the median DTI for different forward enhanced loan types by

race/ethnicity, age, neighborhood income, and geography. The median DTI for non-Hispanic

White borrowers is lower than those for Asian, Black, and Hispanic White borrowers, across all

enhanced loan types. The median DTI for non-Hispanic White borrowers is 35 percent for

conventional conforming loans, 34 percent for jumbo loans, 43 percent for FHA loans, 40

percent for VA loans, and 36 percent for RHS/FSA loans. Hispanic White borrowers’ median

DTI for conventional conforming loans is 40 percent, higher than that of all other racial/ethnic

groups among conventional conforming loan borrowers. Black borrowers’ median DTI is 39

percent for conventional conforming loans and 36 percent for jumbo loans. The median DTI for

Asian borrowers is 38 percent for conventional conforming loans and 36 percent for jumbo

loans. The median DTIs show limited variation in age and neighborhood income for each

enhanced loan type. For each enhanced loan type, the median DTIs are slightly lower for

59

borrowers in micropolitan areas and rural areas than in metropolitan statistical areas, with the

exception of jumbo loan borrowers whose median DTIs in micropolitan areas is the same as in

metropolitan areas.

Figure 6.6.1 is a histogram of the DTI distribution of conventional conforming loan borrowers.

Each bar represents an increment of one percentage point of the DTIs. For ease of reading,

Figure 6.6.1 includes three vertical reference lines at 43 percent, 45 percent, and 50 percent due

to bunching at these levels. DTI is a criterion for determining whether a loan is a qualified

mortgage (QM) under the Bureau’s Ability-to-Repay and Qualified Mortgage Rule, although

under the Temporary GSE Exception conventional loans that are eligible for purchase or

guarantee by one of the GSEs can obtain QM status regardless of DTI. There is bunching at the

DTI level of 43 percent, which is the boundary for QMs that are covered by the rule and do not

fall within the Temporary GSE Exception. This bunching may be due to conventional loans that

are not eligible for GSE purchase and for which lenders seek to obtain QM status, or because of

requirements by certain lenders on the maximum DTI they would accept that coincide with the

maximum QM DTI limit even for GSE-eligible loans. There is a very small percentage of

conventional conforming loans with DTIs greater than 50 percent.

Figure 6.6.2 shows the histogram of DTI distribution of jumbo loan borrowers. Similarly, Figure

6.6.2 includes three vertical reference lines at 43 percent, 45 percent, and 50 percent. There is

heavy bunching at a DTI level of 43 percent, matching the QM maximum DTI limit of 43

percent. However, there is still some percentage of jumbo loans originated with a reported DTI

greater than 43 percent. There are another two bunching points for those jumbo loans with DTI

greater than 43 percent, at 45 percent and 50 percent, respectively. There is only a very small

percentage of jumbo loans with DTI greater than 50 percent.

Figure 6.6.3 is a histogram of DTIs among FHA loans. The four vertical reference lines added in

the figure are at 43 percent, 45 percent, 50 percent, and 57 percent. There is no visual evidence

of bunching at 43 percent for FHA borrowers. There are three bunching points at 45 percent, 50

percent, and 57 percent. There is only a tiny percentage of FHA loans with DTIs greater than 57

percent.

The distribution of DTIs among VA borrowers is much smoother and more symmetrical than

that of other closed-end mortgages, as demonstrated in Figure 6.6.4. It peaks at 43 percent

(though not prominently) and has two additional minor bunching points at 50 percent and 60

percent.

The distribution of DTIs among RHS/FSA borrowers has a bunching point and peak at 41

percent and largely drops off at 46 percent, as shown in Figure 6.6.5.

60

The distribution of DTIs among HELOC borrowers is smooth to the left of (i.e. in the DTI range

lower than) 40 percent and has bunching points at 43 percent, 45 percent, 50 percent, and 55

percent. It peaks at 43 percent, coinciding with the maximum DTI limit of 43 percent for the

general QM category, even though the QM DTI limit does not apply to HELOCs. Its small right

tail extends to 66 percent and then drops off.

DTI is one of the factors often considered when lenders make underwriting decisions. Figure

6.6.7 shows a binscatter plot linking the denial rates and reported DTIs for complete

applications, separated by enhanced loan types. Figure 6.6.7 demonstrates that the relationship

between the denial rates and DTIs is not linear. The denial rates for DTIs above certain key

thresholds increase sharply with higher DTIs, but for the DTIs below the thresholds, the denial

rate may actually decrease with increased DTI. This is likely due to other confounding factors

that are correlated with DTI and not captured in this single bivariate graph. The goal in

presenting such observations is not to draw conclusions but rather to illustrate to users the

complexity of the issues when seeking to explain observed credit decisions.

6.7 Manufactured Home Secured Property

Type

The 2015 HMDA Rule added two new data points that are specific to manufactured homes. The

first is Manufactured Home Secured Property Type. Under the 2015 HMDA Rule, reporters of

manufactured home applications and loans use this data point to indicate whether the covered

loan or application is, or would have been, secured by a manufactured home and land, or by a

manufactured home only. Manufactured Home Secured Property Type is one of the

Discretionary Data Points as discussed in the introduction section of this article. This is one of

the data points that institutions that qualify for a partial exemption under the EGRRCPA are not

required to report. The allowable values of Manufactured Home Secured Property Type are code

1—Manufactured home and land, code 2—Manufactured home and not land, code 3—Not

applicable, and code 1111—Exempt.

Manufactured home loans secured by only manufactured homes and not secured by land (i.e.

those reported with code 2 for secured property type) are also commonly known as chattel loans.

Chattel loans are often different from mortgages for manufactured homes (i.e. loans secured by

manufactured homes and land) in many ways. Table 6.7.1 presents selected characteristics of

manufactured home loans by reported Manufactured Home Secured Property Type in 2019.

Overall, there are about 178,200 originated manufactured home loans reported in the 2019

HMDA data, only slightly higher than the 170,700 manufactured home loans reported in the

2018 data, about 106,100 of which are secured by both manufactured homes and land while

61

53,900 are chattel loans secured only by homes. The median interest rate for chattel loans is

8.49 percent, which is significantly higher than the median interest rate for non-chattel loans at

4.75 percent. In contrast, the median interest rate for non-chattel loans in 2018 was 5.125 and

that for chattel loans in 2018 was 8.29. The median income of chattel loan borrowers and the

median income of non-chattel borrowers are roughly the same; the two are only $1,000 apart.

The median credit score of chattel loan borrowers is 680 in 2019, 20 points lower than that of

non-chattel loan borrowers and one point higher than 2018. The median CLTV for chattel loans

is 83.1 percent, 4.9 percentage points lower than that of non-chattel loans at 89.0 percent. The

median DTI of chattel loan borrowers is 35.6 percent, slightly lower than the median DTI of

non-chattel loan borrowers at 37.7 percent. Almost all chattel loans are for home purchase, at 96

percent; in comparison, the share of home-purchase loans among non-chattel loans is 66

percent. In addition, 94 percent of chattel loans and 91 percent of non-chattel loans are fixed-

rate loans.

Table 6.7.2 shows that most chattel loans are conventional loans. About 53,200 chattel loans out

of 53,900 total are non-government closed-end loans. There are only a small number of chattel

loans issued through government programs.

Table 6.7.3 breaks down the secured property type of originated manufactured home loans by

race/ethnicity, age, neighborhood income, and geography. Including the loans with reported

secured property types of “Exempt” or “NA”, Table 6.7.3 shows that among manufactured home

borrowers Blacks are the most likely to take out chattel loans of all race/ethnicity groups. While

55.8 percent of Black manufactured home borrowers have a reported secured property type of

“manufactured home and not land,” 44.6 percent of Asian borrowers and 38.8 percent of

Hispanic White manufactured home borrowers take out chattel loans. Additionally, 23.6 percent

of manufactured home loans for non-Hispanic White borrowers are chattel loans, the smallest

share across all racial/ethnic groups.

Younger manufactured home borrowers are more likely to take out chattel loans than older

borrowers. The share falls from 38.5 percent for manufactured home borrowers younger than 25

to 25.9 percent for those 75 or older.

The share of chattel loans is higher in low/moderate-income census tracts, at 38.3 percent, than

the shares of chattel loans in middle- and high-income tracts, at 27.0 percent and 30.1 percent,

respectively. The majority of manufactured home loans are for homes in middle-income tracts.

In addition, 32.9 percent of manufactured home loans in metropolitan statistical areas are

chattel loans. In comparison, the shares of chattel loans among manufactured home loans are

23.8 percent in micropolitan areas and 27.9 percent in rural areas.

62

6.8 Manufactured Home Land Property

Interest

Manufactured Home Land Property Interest is another data point added in the 2015 HMDA

Rule that is applicable only to covered manufactured home loans or applications. Under the

2015 HMDA Rule, if the dwelling related to the property is a manufactured home and not a

multifamily dwelling, the reporter must report whether the applicant or borrower: (i) owns the

land on which the manufactured home is or will be located or, in the case of an application, did

or would have owned the land on which it would have been located, through a direct or indirect

ownership interest; or (ii) leases or, in the case of an application, would have leased the land

through a paid or unpaid leasehold. Manufactured Home Land Property Interest is one of the

Discretionary Data Points as discussed in the introduction section of this article. The

Manufactured Home Land Property Interest is one of the data points that institutions that

qualify for a partial exemption under the EGRRCPA are not required to report.

The allowable values of Manufactured Home Land Property Interest are: code 1—Direct

ownership, code 2—Indirect ownership, code 3—Paid leasehold, code 4—Unpaid leasehold, code

5—Not applicable, and code 1111—Exempt.

Table 6.8.1 presents some selected characteristic of manufactured home loans with different

land property interests.

Overall, about 121,600 manufactured home borrowers are reported as having direct ownership

of their land, 900 borrowers have indirect land ownership, 25,400 manufactured home loans

are on land with paid leaseholds, and another 12,000 are on land with unpaid leaseholds. The

median interest rate on loans is highest for properties with unpaid leaseholds at 8.99 percent,

followed by those with paid leaseholds at 8.60 percent. The median interest rate is lowest for

loans with direct ownership at 4.875 percent.

The median income of borrowers with unpaid leaseholds is $48,000, lower than the median

income of other borrowers. Borrowers with unpaid leaseholds also have a lower median credit

score (659) than borrowers with paid leaseholds (683) and those with direct ownership (699).

Borrowers who have unpaid leaseholds have higher median CLTVs (94.3 percent) than

borrowers with other types of property interests. A much higher share of loans to borrowers

with leaseholds are for the purpose of home purchase, at 99 percent for unpaid leasehold

borrowers and 96 percent for paid leasehold borrowers. This is higher than the home-purchase

shares of borrowers with direct ownership (70 percent) and indirect ownership (90 percent).

Table 6.8.2 shows that almost all loans with paid or unpaid leaseholds are conventional loans.

63

Table 6.8.3 breaks down the land property interest of originated manufactured home loans by

race/ethnicity, age, neighborhood income, and geography. Including the loans that reported the

land property interest as “Exempt” or “NA,” Table 6.8.3 shows that non-Hispanic White

borrowers are the most likely to have direct land ownership relative to other racial/ethnic

groups at 71.1 percent. In comparison, 59 percent of “Other” borrowers, 64.9 percent of

Hispanic White borrowers, 57.6 percent of Black borrowers, and 53 percent of Asian borrowers

have direct land ownership.

Young manufactured home borrowers are somewhat less likely to directly own the land than

older borrowers. While 59.9 percent of manufactured home borrowers younger than 25 have

direct ownership, that share was 67.2 percent for borrowers aged 25 to 34, 68.6 percent for

borrowers between the ages of 35 and 44, and around 70 percent for borrowers of older age

groups.

The share of manufactured home loan borrowers with direct ownership is lower in

low/moderate-income census tracts at 63.1 percent than in middle-income tracts (70.6 percent)

and high-income tracts (69.2 percent).

About 65.8 percent of manufactured home loans in metropolitan statistical areas feature direct

ownership. In comparison, the shares of loans with direct ownership are 76.3 percent in

micropolitan areas and 68.4 percent in rural areas.

Table 6.8.4 shows that, among the originated manufactured home loans secured by home and

land, 99.5 percent are reported to feature direct land ownership. Among the loans that are

secured by manufactured homes and not land, 29.7 percent feature direct ownership, 47.0

percent use paid leaseholds, 22.3 percent use unpaid leaseholds, and 1.0 percent feature indirect

ownership. Looked at from a related but different angle, among manufactured home loans with

direct ownership, about 13.2 percent are reported to be secured by a manufactured home and

not land.

6.9 Number of Affordable Units for

Multifamily Loan

The 2015 HMDA Rule added a new data point for loans and applications secured by multifamily

units. For multifamily loans, reporters are required to report the number of individual dwelling

units in multifamily dwelling properties securing the covered loans or, in the case of

applications, proposed to secure the covered loans that are income-restricted pursuant to

federal, state, or local affordable housing programs. These are referred to as “affordable units.”

Number of Multifamily Affordable Units is one of the Discretionary Data Points as discussed in

64

the introduction section of this article. The number of affordable units for multifamily loans is

one of the data points that institutions that qualify for a partial exemption under the EGRRCPA

are not required to report. A reporter would enter “0” for a covered loan or application related to

a multifamily dwelling that does not contain any income-restricted individual dwelling units,

“NA” if the requirement to report multifamily affordable units does not apply, or “Exempt” if the

reporter is exempt from reporting this information. Affordable units are disclosed in the public

loan-level 2019 HMDA data as a percentage, rounded to the nearest whole number, of the value

reported for the total number of individual dwelling units related to the property securing the

covered loan.

Table 6.9.1a shows, among all site-built multifamily originated loans, the number of loans, the

number of loans reported with one or more affordable units, and their relative shares by the

number of total unit bins disclosed in the public loan-level 2019 HMDA data. In total, out of

about 53,400 multifamily originated loans (up from 50,600 in 2018), about 5.5 percent or close

to 2,900 are for properties with at least one affordable unit in 2019, which was down from 7.1

percent or close to 3,600 in 2018. There are about 34,800 multifamily loans secured by

properties with between five and 24 units in 2019 (up from 32,600 in 2018), and about 860

such loans (or 2.5 percent) are reported to have at least one income-restricted unit, down from

about 1,300 (or 3.9 percent) in 2018. The share of multifamily loans with income-restricted

units is highest among multifamily loans with between 100 and 149 total units, at 16.0 percent.

Table 6.9.1b restricts the sample to the multifamily loans with income-restricted units, and

shows the distribution of the ratio between the number of income-restricted units and the

number of the total units securing each loan by the number of total-units bins disclosed in the

public loan-level 2019 HMDA data. Among the multifamily loans reported with income-

restricted units, more than half of them are exclusively income-restricted, with the number of

income-restricted units equal to or very close to the total number of units.

65

7. Pricing Outcomes and

Components

The DFA and the 2015 HMDA Rule added several data points regarding pricing of loans and

applications, and expanded the scope of the rate spread data point. This section introduces

readers to these new or expanded data points related to mortgage pricing and costs of the loan.

The mortgage pricing and the costs of a loan include many components, some of which could be

substitutes for one another (in other words, fungible) or may involve intertemporal tradeoffs

between the upfront costs of obtaining a loan and the longer-term costs during the life of a loan.

It is beyond the scope of this article to address the complex interrelationship of these pricing

components. Instead, this section provides some basic summary statistics based on the 2019

HMDA data, while introducing readers to these pricing data points.

7.1 Interest Rate

The 2015 HMDA Rule added a new requirement that institutions report the interest rate

applicable to the approved application, or to the covered loan at closing or account opening.

Interest Rate is one of the Discretionary Data Points as discussed in the introduction section of

this article. The interest rate is reported as a percentage, to at least three decimal places. This is

one of the data points that institutions that qualify for a partial exemption under the EGRRCPA

are not required to report. A reporter would report “NA” if the requirement to report interest

rate does not apply or “Exempt” if the reporter is exempt from reporting this information under

the EGRRCPA.

Table 7.1.1 reports selected summary statistics on the interest rates of originated loans by

enhanced loan type in 2019. The median interest rate for conventional conforming loans is 4.125

percent, 62.5 basis points lower than 2018, with its 5th percentile at 3.25 percent (meaning that

five percent of borrowers obtained interest rates at or below 3.25 percent and the rest obtained

higher interest rates), and 95th percentile at 5.80 percent. The median interest rate of jumbo

loans is 37.5 basis points lower than that of conventional conforming loans, at 3.75 percent in

2019, down from 4.25 percent in 2018. The 5th percentile of jumbo loans’ interest rates is 2.75

percent and the 95th percentile is 5.375 percent. This report notes that such a comparison has

not adjusted for the credit characteristics and loan characteristics of the loans as discussed in

other sections of this article.

66

The median interest rate of FHA loans is 4.125 percent (down from 4.75 percent in 2018), the

same as the median interest rate of conventional conforming loans. But the 5th percentile of FHA

loans’ interest rates is 12.5 basis points higher than the interest rate of conventional conforming

loans at the equivalent percentiles. The FHA loan interest rates’ 25th percentile (at 3.75 percent)

and 75th percentile (at 4.625 percent) are the same as the equivalent percentiles for conventional

conforming loans, and its 95th percentile (at 5.375 percent) is lower than that of the conventional

conforming loans. Together, the dispersion of interest rates on FHA loans is smaller than that of

conventional conforming loans. The median interest rate of VA loans is 3.75 percent (down from

4.5 percent in 2018), lower than that of all other enhanced loan types, except for jumbo

conventional loans. The 5th, 25th, 75th, and 95th percentiles of the VA loans’ interest rates are also

lower than the equivalent percentiles of FHA and RHS/FSA loans.

The median interest rate on HELOCs is higher than that of closed-end mortgages, at 5.34

percent (up from 5.0 percent in 2018). However, their 5th percentile is 2.74 percent, lower than

that of any closed-end mortgage loan type, while its 95th percentile is 8.62 percent, significantly

higher than the 95th percentile of any closed-end mortgage loan type. In other words, the median

interest rate of HELOCs is about 121.5 basis points higher than the median interest rate of

conventional conforming loans, but the HELOC interest rate displays a substantial degree of

variation, with a relatively high interest rate tail to the right. The median interest rate of reverse

mortgages is 4.482 percent down from the median interest rate of 4.827 in 2018), its 5th

percentile is 3.423 percent, and its 95th percentile is 6.750 percent.

[Table 7.1.1: Selected Summary Statistics of Interest Rate by Enhanced Loan Type]

Table 7.1.2 presents the median interest rates for closed-end enhanced loan types by loan

purpose, occupancy type, and lien status. The median interest rates of cash-out refinance loans

are higher than that of non-cash-out refinance loans for each enhanced loan type. The non-cash-

out refinance loans have lower median interest rates than the median interest rates of home-

purchase loans within each respective enhanced loan type. The median interest rates on home

improvement loans are higher than the median interest rates of home-purchase and refinance

loans for both conventional conforming and jumbo loans.

The median interest rate for conventional conforming loans secured by a second residence is the

same as that of principal residence conventional conforming loans at 4.0 percent, while the

median interest rate for jumbo loans secured by a second residence is 12.5 basis points lower

than that of jumbo loans secured by a principal residence which is at 3.75 percent. Keep in mind

such comparisons do not control for other underlying credit characteristics of the borrowers.

The median interest rate for conventional conforming loans secured by an investment property

is 5.0 percent, and the median interest rate for jumbo loans secured by an investment property

is very similar, at 4.99 percent.

67

The median interest rate for conventional conforming loans secured by subordinate liens is 5.34

percent, down slightly from 5.50 percent in 2018, but the drop is far smaller than the drop in

median interest rate for conventional conforming loans secured by first liens which stands at 4

percent in 2019 compared to 4.75 percent in 2018. The median interest rate for jumbo loans

secured by subordinate liens is 4.50 percent (down from 5.125 percent in 2018), 75 basis points

higher than the median interest rate of jumbo loans secured by a first lien.

Table 7.1.3 presents the median interest rate within enhanced loan types by race/ethnicity, age,

neighborhood income, and geography in 2019.

The median interest rate for Black borrowers with conventional conforming loans is 4.375

percent, the median interest rate for Hispanic White borrowers with conventional conforming

loans is 4.25 percent. Both are higher than the median interest rate of non-Hispanic White

borrowers at 4.125 percent. The median interest rate of Asian conventional conforming loan

borrowers is 3.99 percent, the lowest among all racial/ethnic groups within conventional

conforming loan type.

The median interest rates for Black borrowers and Hispanic White borrowers with jumbo loans

are both 4.0 percent. In comparison, the median interest rate of non-Hispanic White borrowers

for jumbo loans is 3.75 percent. Just like in the conforming loan market, Asian borrowers as a

group have the lowest median interest rate for jumbo loans, at 3.5 percent, among all

racial/ethnic groups.

The median interest rate for Black borrowers and Hispanic White borrowers with FHA loans are

both 4.25 percent, which is 12.5 basis points higher than that of non-Hispanic White borrowers.

The median interest rate for Asian FHA borrowers is 4.0 percent, again the lowest among all

racial/ethnic groups in that segment of the market.

The median interest rate for Black VA loan borrowers is 3.875 percent. The median interest rate

for Asian VA loan borrowers is 3.625 percent. The median interest rate for all other non-missing

race/ethnicity groups for VA loans is 3.75 percent.

The median interest rate for both Asian and Hispanic White borrowers taking out RHS/FSA

loans is 4 percent. The median interest rate for all other non-missing race/ethnicity groups for

RHS/FSA loans is 4.125 percent.

The median interest rates for HELOCs is 5.75 percent for Black borrowers and 5.625 percent for

Hispanic White borrowers. In comparison, non-Hispanic White borrowers of HELOCs have a

median interest rate of 5.33 percent, and Asian HELOC borrowers have a median interest rate of

5.09 percent.

68

The variation of median interest rates for closed-end mortgages over age is generally small, but

borrowers younger than 25 pay higher median interest rates than all other age groups for

conventional conforming, jumbo, FHA, and VA loans, respectively. The median interest rate for

HELOCs generally decreases with age, with the median interest rates for HELOC borrowers

younger than 25 and between 25 and 34 years old at 5.75 percent, and the median interest rates

for HELOC borrowers older than 75 and between 65 and 74 years old at 5.24 percent.

The median interest rate of conventional conforming loans for properties located in

low/moderate-income tracts is 4.31 percent, 18.5 basis points higher than the median interest

rate for conforming loans for properties in middle-income census tracts, and 31 basis points

higher than the median interest rate of conventional conforming loans in high-income tracts.

The median interest rates of jumbo loans in middle-income census tracts and in high-income

tracts are both 3.75 percent, 12.5 basis points lower than the median interest rate for jumbo

loans in low/moderate-income tracts. The median interest rate for FHA loans in low/moderate-

income census tracts (4.25 percent) is 12.5 basis point higher than the median interest rate for

FHA loans in both middle-income tracts and high-income tracts, with the same median interest

rate (4.125 percent). The median interest rate for the VA loans in low/moderate-income tracts is

3.875 percent, 12.5 basis point higher than in middle-income tracts (3.75 percent) and 17.5 basis

points higher than in high-income tracts (3.70 percent). For the RHS/FSA loans, the median

interest rates are both 4.125 percent for properties in low/moderate-income tracts and in

middle-income tracts, and 4.0 percent for high-income tracts.

The median interest rate is about 11 basis points higher for HELOCs in low/moderate-income

census tracts (at 5.5 percent) than for HELOCs in middle-income tracts (5.39 percent) and 25

basis points higher than for HELOCs in high-income tracts (5.25 percent).

At 4.125 percent, the median interest rate for conventional conforming loan borrowers living in

metropolitan statistical areas is 12.5 basis points lower than that for borrowers in micropolitan

statistical areas and rural areas (4.25 percent). The median interest rates for jumbo loans are

3.75 percent in metropolitan statistical areas, 3.875 percent in micropolitan statistical areas, and

4.25 percent in rural areas. The median interest rates for FHA loans are the same for

micropolitan statistical areas and rural areas at 4.25 percent, and 12.5 basis points higher than

the median interest rate for FHA loans in metropolitan statistical areas (4.125 percent). The

same pattern exists for VA loans. The median interest rates for RHS/FSA loans are the same in

metropolitan statistical areas, micropolitan statistical areas, and rural areas, at 4.125 percent.

There are only small differences in the median interest rate paid by HELOC borrowers in the

three geographic categories. The median interest rate of reverse mortgages for borrowers living

in rural areas (4.716 percent) is slightly higher than that of borrowers in micropolitan areas

(4.549 percent) by 16.7 basis points, who in turn have median interest rates slightly higher than

that of borrowers in metropolitan areas (4.465 percent) by 8.4 basis points.

69

It is worth emphasizing that the median interest rates discussed above do not take into

consideration the differences in the underlying credit characteristics of the borrowers or the

loans, such as credit score, CLTV, choice of loan term, whether the loan has a fixed rate or

adjustable rate, non-amortizing features, lien status, occupancy status, and whether the

borrowers have paid discount points or received lender credits, etc.

As previously noted, the interest rate reported is the rate at closing or account opening, which

means that for an adjustable-rate loan, the reported rate is the initial rate. Table 7.1.4 shows the

median interest rates of different enhanced loan types, separated by whether the loans are fixed-

rate or ARM loans. Notably in 2019, the median interest rates for conventional conforming

loans were 4.125 percent for both fixed-rate mortgages and adjustable-rate loans. The median

interest rate for jumbo loan fixed-rate mortgages is 3.875 percent and the median interest rate

for adjustable-rate jumbo loans is 3.375 percent. The median interest rate for fixed-rate FHA

mortgages is 4.125 percent and the median interest rate for adjustable-rate FHA loans is 4.0

percent. Similarly, the median interest rates for adjustable-rate loans are lower than the median

interest rates for fixed-rate loans for HELOCs and reverse mortgages, respectively.

Interest rates typically vary with the term of the loan as well. Table 7.1.5 shows the median

interest rates of different term lengths for fixed-rate conventional mortgages, including

conventional conforming mortgages and jumbo loans, respectively.

Within conventional conforming fixed-rate mortgages, the median interest rate for 30-year

loans is 4.125 percent. As the term shortens from 30 to 20 years, the median interest rate drops

to 4.0 percent. Fifteen-year fixed-rate mortgages have the lowest median interest rate among all

common loan terms for conventional conforming fixed-rate mortgages listed in Table 7.1.5, at

3.625 percent. On the other hand, five-year fixed-rate mortgages, the shortest-term loans among

the common terms listed in the table (excluding “other”), have the highest median interest rate

at 4.95 percent; and the median interest rate of 10-year fixed-rate mortgages is 4.5 percent.

Of the conventional jumbo fixed-rate mortgages, the median interest rate for 30-year loans is

3.875 percent and the median interest rate for 20-year loans is 3.75 percent. As with

conventional conforming loans, 15-year fixed-rate jumbo loans have the lowest median interest

rate among all common loan terms, at 3.375 percent. On the other hand, 5-year fixed-rate jumbo

mortgages have a median interest rate of 5.0 percent, and the median interest rate for 10-year

fixed-rate jumbo mortgages is 3.875 percent. (Note that 5-year and 10-year jumbo mortgages

are much less common than the jumbo loans of 30-year, 20-year, and 15-year terms.)

Finally, among the adjustable-rate mortgages, the interest rates also vary with the length of the

introductory rate period. Table 7.1.6 presents the median interest rate of adjustable-rate

mortgages of different introductory rate periods for conventional conforming loans and jumbo

loans. To control for the effect of different loan terms, Table 7.1.6 limits the sample to only

70

adjustable-rate mortgages with a 30-year term, which is the most common term for adjustable-

rate mortgages.

For adjustable-rate mortgages with an introductory rate period less than or equal to ten years, as

the introductory rate period increases the median interest rate generally decreases. The median

interest rates for conventional conforming ARMs with an introductory rate period less than one

year is 5.5 percent; for an introductory rate period of one year it is 4.0 percent; for an

introductory rate period of three years it is 4.375 percent; for an introductory rate period of five

years it is 4.375 percent as well; and for an introductory rate period of either seven or ten years

it is 3.75 percent. Similarly, the median interest rate for jumbo ARMs with an introductory rate

period of one year is 4.0 percent; for an introductory rate period of three years it is 3.70 percent;

for an introductory rate period of five years it is 3.625 percent; and for an introductory rate

period of either seven years or ten years it is 3.375 percent. On the other hand, the median

interest rate of 15-year conventional conforming non-fixed-rate mortgages is 3.95 percent, and

the median interest rate of 15-year jumbo non-fixed-rate mortgages is 3.875 percent

52

.

7.2 Rate Spread

Rate Spread, defined as the difference between the covered loan’ annual percentage rate (APR)

and the average prime offer rate (APOR) for a comparable type mortgage as of the date the

interest rate is set, was required to be reported for higher-priced closed-end mortgages prior to

2018.

53

Loans were classified as higher-priced if the APR exceeded the APOR for loans of a

similar type by at least 1.5 percentage points for first-lien loans or 3.5 percentage points for

junior-lien loans.

54

Pursuant to the DFA as implemented by the 2015 HMDA Rule, the required

reporting of rate spread is no longer limited to the higher-priced closed-end mortgages. Rate

spread must now be reported for all covered loans and applications that are approved but not

accepted and that are subject to Regulation Z, excluding assumptions, purchased covered loans,

52

The majority of closed-end mortgage loans reported under HMDA with a 15-year introductory rate period appear to

be what the industry refers to as step-rate mortgages, for which the interest rate is set for the first 15 years, and then

reset to the ongoing rate at that time for another 15 years until the end of the term.

53

“Average prime offer rate” means an annual percentage rate that is derived from average interest rates and other

loan pricing terms currently offered by a set of creditors to consumers for mortgage loans that have low-risk pricing

characteristics. The Bureau publishes tables of average prime offer rates by transaction type at least weekly and

publishes the methodology it uses to derive these rates. (https://ffiec.cfpb.gov/tools/rate-spread).

54

Prior to October 2009, loans were classified as higher-priced if the spread between the APR and the rate on a

Treasury bond of comparable term exceeded three percentage points for first-lien loans or five percentage points for

junior-lien loans, and the rate spread reported under HMDA used the comparison of APR to the rate on a Treasury

bond instead of the spread over the APOR.

71

and reverse mortgages.

55

The inclusion of mandatory reporting of open-end lines of credit by the

2015 HMDA Rule also added HELOCs into the rate spread reporting requirements. Rate Spread

for all loans is one of the Mandated Data Points as discussed in the introduction section of this

article. The rate spread is one of the data points that institutions that qualify for the EGRRCPA

partial exemption are not required to report. Rate spread is reported as a percentage to at least

three decimal places. It can be either positive or negative, depending upon whether it exceeds or

falls below the comparable APOR. Reporters would enter “NA” if the requirement to report rate

spread does not apply, or “Exempt” if the reporter is exempt from reporting the information

under the EGRCCPA.

The accompanying article to this one, titled “2019 Mortgage Market Activity and Trends”, has an

extensive discussion, using the rate spread data point, on higher-priced closed-end mortgages.

56

T0 avoid overlap, this section presents only some selected summary statistics of the distribution

of rate spread by enhanced loan type for originated loans, excluding reverse mortgages.

Table 7.2.1 presents the distribution of the rate spread by enhanced loan type. The median rate

spread for conventional conforming loans is 0.357 percent; for jumbo loans it is -0.111 percent;

for FHA loans it is 1.207 percent; for VA loans it is -0.019 percent; for RHS/FSA loans it is 0.692

percent; and for HELOCs it is 0.8 percent. It is important to note that APOR represents the

average interest rates and fees offered to prime borrowers for a first-lien closed-end

conventional conforming loan with an 80 percent LTV and the calculation of the rate spread is

essentially comparing the APR of an originated loan or HELOC to that average. Given the

different compositions of borrowers’ credit characteristics and different loan characteristics

across various enhanced loan types, caution should be used in interpreting the differences in

rate spread across different products.

Table 7.2.2 presents the median rate spread within each enhanced loan type by loan purpose,

occupancy type, and lien status.

57

The median rate spread of cash-out refinance loans is higher

than that of non-cash-out refinance loans except for FHA loans. The median rate spread of

home-purchase loans is higher than that of non-cash-out refinance loans for each enhanced loan

type. The median rate spread of loans secured by second residences is lower than for loans

secured by principal residences for each enhanced loan type, and the median rate spread for

loans secured by an investment property is higher than the median rate spread for loans secured

by a principal residence except for VA loans. The median rate spread for loans secured by a

55

See Regulation Z, 12 CFR part 1026.

56

See “2019 Mortgage Market Activity and Trends”, available at https://www.consumerfinance.gov/data-

research/research-reports/cfpb-data-point-2019-mortgage-market-activity-and-trends/

57

This discussion excludes all cells that are omitted from Table 7.2.2 each of which has a frequency count less than

500.

72

subordinate lien is higher than that of loans secured by first lien for all enhanced loan types

shown in the table.

Table 7.2.3 presents median rate spread within each enhanced loan product by race/ethnicity,

age, neighborhood income, and geography. Again, the median rate spreads displayed have not

controlled for the differences in underlying borrower credit characteristics, loan features, and

borrowers’ loan choices.

The median rate spread for the loans of Black borrowers is higher than for that of all other

racial/ethnic groups for each enhanced loan type. The median rate spread for the loans of Asian

borrowers is the lowest among all racial/ethnic groups for each enhanced loan type. The median

rate spread for the loans of Hispanic White borrowers is higher than that of loans for non-

Hispanic White borrowers for all enhanced loan type except for RHS/FSA loans.

The median rate spread for HELOCs decreases monotonically with age (except for HELOC

borrowers 24 years in age or younger). The median rate spreads among the youngest borrower

age groups generally are higher than the median rate spread for older groups within each

enhanced closed-end mortgage type, but the detailed patterns of rate spreads over age vary

across different enhanced loan types.

The median rate spread for loans in low/moderate-income tracts is higher than that of middle-

income tracts, within each enhanced loan type (except for jumbo loans), which in turn is higher

than the median rate spread of the loans in high-income tracts, within each enhanced loan type.

In addition, the median rate spread for loans in rural areas is higher than that in micropolitan

statistical areas, within each enhanced loan type, which in turn is higher than the median rate

spread in metropolitan areas, within each enhanced loan type except for HELOCs.

7.3 Total Loan Costs or Total Points and

Fees

The DFA, as implemented by the 2015 HMDA Rule, added Total Points and Fees as one of the

data points that institutions must report. Total Points and Fees is one of the Mandated Data

Points as discussed in the introduction section of this article. These are captured in two data

fields in 2019 HMDA data: Total Loan Costs, and Total Points and Fees, each applied to

different transactions as explained below. Total Loan Costs or Total Points and Fees applies only

73

to originated loans that are subject to specified requirements in Regulation Z.

58

Total Loan Costs

applies to originated loans that are subject to the TILA-RESPA Integrated disclosure

requirements in Regulation Z. Total Points and Fees applies to originated loans that are not

subject to those requirements but are covered by the Ability-to-Pay requirements in Regulation

Z . Institutions that qualify for the partial exemption under the EGRRCPA are not required to

report Total Loan Costs or Total Points and Fees.

Under Regulation C, other than for loans that are eligible for partial exemptions under the

EGRRCPA, in general, if a loan is subject to the TILA-RESPA Integrated Disclosure Rule

(“TRID”)

59

requirements, a reporter must report total loan costs as disclosed on the TRID

Closing Disclosure. TRID applies to most closed-end consumer credit transactions secured by

real property or co-ops, but does not apply to HELOCs, reverse mortgages, or mortgages secured

by a mobile home that is not attached to real property. In other words, open-end lines of credit,

reverse mortgages, and closed-end loans made primarily for a business purpose are not subject

to TRID, and hence financial institutions do not report Total Loan Costs for these transactions.

Loans secured by manufactured homes but not secured by the land do not report Total Loan

Costs either, since they do not require a TRID Closing Disclosure; they would report “Total

Points and Fees” instead, which is defined under the QM rule. Open-end lines of credit and

reverse mortgages are not required to report “Total Points and Fees” either.

Total Loan Costs are entered in dollars, or as “NA” for transactions for which this requirement

does not apply, or “Exempt” if the reporter is exempt from reporting this information under the

EGRRCPA. It is important to note that the total loan costs reported under HMDA are “borrower

paid.”

60

The total closing costs may be partially paid by the seller (in the home-purchase

transaction) or by others, but those should not be captured by the Total Loan Costs data point

reported under HMDA. The total loan costs are the sum of origination charges that the lender

charges, charges for the services that borrowers cannot shop for (e.g., appraisal fees or credit

report fees), and charges for services borrowers can shop for such as settlement agent or title

insurance fees. In other words, under HMDA reporting requirements, it includes the charges by

the lenders as well as the charges by third party service providers in connection with obtaining

the loan to the extent those are paid by the consumer rather than by a seller or other third party.

It is important to note that loan costs may be tied to the size of the loan and can be affected by

factors such as the size of the down payment relative to the loan (as that will drive the need for

58

See 12 CFR 1026.19(f).

59

See Regulation Z, 12 CFR part 1026.19(f).

60

On the TRID Closing Disclosure, this corresponds to the number on the summary line of Block D (titled “TOTAL

LOAN COSTS (Borrower-Paid)”) of the “Closing Cost Detail” Section on the “Borrower-Paid” column.

74

mortgage insurance) as well as by choices made by consumers (such as the purchase of owners

title insurance). The summary statistics reported in this section do not control for any such

factors and these factors may explain some of the differences observed across enhanced loan

types, loan purpose, demographic groups, etc.

Table 7.3.1 presents some basic summary statistics on Total Loan Costs by enhanced loan type

for those loans subject to reporting under HMDA. The table also excludes manufactured home

loans. The same exclusion rules also apply to Tables 7.3.2 and 7.3.3. The average total loan costs

for all loans is $4,809. The average total loan costs reported under HMDA for conventional

conforming loans is $3,888 and the median is $3,404; the average total loan costs for jumbo

loans is $6,210 and the median is $4,857; the average total loan costs for FHA loans is $7,810

and the median is $7,129; for VA loans, the average is $5,777 and the median is $4,274; and the

average total loan costs for RHS/FSA loans is $4,753 with a median of $4,547. With all

enhanced loan types combined, the median total loan costs for all site-built single-family closed-

end consumer purpose loans secured by real property reported under HMDA is $3,925.

Overall, not adjusting for loan amount and borrower/loan characteristics, FHA loan borrowers

on average and at the median pay higher total loan costs than borrowers of other enhanced loan

types, in absolute dollar terms. The jumbo loan borrowers are the second highest in average and

median total loan costs paid. The average and median total loan costs of conventional loan

borrowers are the lowest among all enhanced loan types. The 5th, 25th, 75th, and 95th percentiles

of the reported total loan costs of each enhanced loan type are also reported in Table 7.3.1.

Table 7.3.2 reports the median total loan costs of various enhanced loan types by loan purpose,

occupancy type, and lien status. The median total loan costs for cash-out refinance loans are

higher than those of non-cash-out refinances among all closed-end enhanced loan types. The

median total loan costs for home-purchase loans are higher than that of refinance loans

(including cash-out and non-cash-out refinance) for conventional conforming and jumbo loans.

On the other hand, the median total loan costs for home purchase loans are higher than that of

non-cash-out refinance loans but lower than that of cash-out refinance loans for FHA and VA

loans, respectively. In particular, the differences in total loan costs between home-purchase,

non-cash-out refinance and cash-out refinance loan for VA loans are substantial, with the

median total loan costs at $5, 330 for VA home-purchase loans, at $2,651 for VA non-cash-out

loans, and at $6,775 for VA cash-out refinance loans. The median total loan costs for home

improvement loans and loans that reported “other purpose” are much lower than loans of all

other purposes among conventional conforming loans.

The median total loan costs for loans secured by investment properties are higher than those of

loans secured by second residences among conventional conforming and jumbo loans

75

respectively, which in turn have higher median total loan costs than loans secured by principal

residences within each of these two enhanced loan types.

The median total costs for loans secured by a first lien are all much higher than the median total

costs for loans secured by subordinate liens across all enhanced loan types, most likely because

the majority of the loan costs of obtaining a subordinate lien mortgages have already occurred in

the first lien mortgages.

Table 7.3.3 presents the median total loan costs for different types of loans by race/ethnicity,

age, neighborhood income, and geography. There is no consistent pattern for the median total

loan costs in terms of absolute dollar values across all racial/ethnicity groups. For instance, the

median total loan costs for Asian borrowers are the highest among FHA, and VA loans,

respectively. But Asian borrowers’ median total loan costs rank lower than that of Hispanic

White borrowers for conventional conforming mortgages, and Asian borrowers’ median total

loan costs is the lowest among all race ethnic groups for jumbo loans. There is also no apparent

pattern for median total loan costs related to age and neighborhood income. In terms of

geography, the median total loan costs are higher for loans in metropolitan statistical areas than

the median total loan costs in micropolitan areas across all enhanced loan types, which in turn

are higher than the median total loan costs for loans in rural areas except for jumbo loans.

All tables discussed in this section so far are limited to site-built single-family homes. For

completeness, Table 7.3.4 presents the summary statistics on the total loan costs for loans

secured by both manufactured homes and the land and the total points and fees for loans that

are secured by only the manufactured home and not the land. The median total loan costs on

manufactured home loans secured by the manufactured home and land is $4,215 in 2019 up

from $3,933 in 2018. The median total points and fees on manufactured home loans secured by

the manufactured home but not land is $1,736, up from $1,525 in 2018. This report notes that

the total loan costs and the total points and fees are not directly comparable because they are

calculated differently based on different regulations.

7.4 Origination Charges

Origination Charges is another data point that the 2015 HMDA Rule requires institutions to

report for covered loans. Origination Charges is one of the Discretionary Data Points as

discussed in the introduction section of this article. In practical terms, under the Rule, if a loan

is subject to the requirement to provide a TRID Closing Disclosure, a reporter is required to

76

report the borrower-paid

61

origination charges, as disclosed on the TRID Closing Disclosure. As

with Total Loan Costs, this data point (Origination Charges) only applies to closed-end

consumer credit transactions secured by real property or co-ops. In other words, open-end lines

of credit, reverse mortgages, and loans or lines of credit made primarily for a business purpose

are not subject to TRID and hence do not report Origination Charges. Loans secured by

manufactured homes and not the land do not report Origination Charges either, since they do

not require a TRID Closing Disclosure. Institutions that qualify for a partial exemption under

the EGRRCPA are not required to report this data point.

Origination Charges are entered in dollars, or as “NA” for transactions for which this

requirement does not apply, or as “Exempt” if the reporter is exempt from reporting this

information under the EGRRCPA.

Table 7.4.1 presents some basic summary statistics on the origination charges by enhanced loan

type. As with total loan costs, these statistics do not control for various factors that may drive

variations in origination charges including, for example, loan size or choices made by consumers

in trading off interest rates and fees. Overall, the average origination charges for all single-

family consumer-purpose closed-end mortgages secured by real property reported under HMDA

is $1,852 and the median is $1,225. The average origination charges reported under HMDA for

conventional conforming loans is $1,868, the median is $1,250; the average for jumbo loans is

$2,503, the median is $1,150; the average for FHA loans is $1,823, the median is $1,303; the

average for VA loans is $1,550, the median is $794; the average for RHS/FSA loans is $1,424,

and its median is $1,199. The 5th, 25th, 75th, and 95th percentiles of the origination charges of

each enhanced loan type are also reported in Table 7.4.1. Compared to 2018, the mean and

median origination charges in 2019 dropped for jumbo loans and VA loans and rose in all other

enhanced loan types.

Table 7.4.2 reports the median origination charges by loan purpose, occupancy type and lien

status, separated by enhanced loan type. In 2019, the median origination charges for

conventional conforming loans are greater than the median origination charges for jumbo loans

with all three major loan purposes, i.e. the home-purpose, non-cash-out refinance and cash-out

refinance categories. Conventional conforming loans for home improvement or reported “other

purpose” have zero or near zero origination charges at the median. The median origination

charges on cash-out refinance loans are higher than those of non-cash-out refinance among all

closed-end loan types shown in Table 7.4.2. The median origination charges on refinance loans

are substantially higher than those of home-purchase loans among VA loans, with the median

61

As with total loan costs, the origination charges reported under HMDA are “borrower-paid.” To the extent that

some part of the origination charges may be paid by the seller (in the home purchase transaction) or paid by others,

those should not be captured by the origination charges data point reported under HMDA.

77

origination charges for VA non-cash-out refinance loans at $712, which represents a significant

drop from $1,976 in 2018, median origination charges for VA cash-out refinance loans at

$2,487, and median origination charges for VA home-purchase loans at $80, which was $268 in

2018.

Among conventional conforming loans, the median origination charges for loans secured by a

second residence are similar to those of loans secured by a principal residence, and the median

origination charges for loans secured by investment properties are higher than the median

origination charges of the other two occupancy categories. Among jumbo loans, the median

origination charges for loans secured by an investment property are also higher than the median

origination charges of the other two occupancy types.

The median origination charges for loans secured by subordinate liens are zero for conventional

conforming loans, likely because most of them are piggy-back loans whose origination charges

are covered by the first mortgages originated at the same time.

62

The median origination charges

for jumbo loans secured by subordinate liens are $495, also substantially lower than the median

origination charges for jumbo loans secured by first liens (by $660).

Table 7.4.3 presents the median origination charges for different types of loans by

race/ethnicity, age, neighborhood, income, and geography. The median origination charges that

Hispanic White conventional conforming loan borrowers paid are $1,445. In comparison,

median Black conventional conforming loan borrowers paid $1,295 in origination charges,

median Asian conventional conforming loan borrowers paid $1,290, and median non-Hispanic

White conventional conforming loan borrowers paid $1,195. The median origination charges

paid by both Black and Hispanic White jumbo loan borrowers are $1,290, while the median

origination charges paid by Asian jumbo loan borrowers are $1,175 and the median origination

charges paid by non-Hispanic White jumbo loan borrowers are $1,140. For FHA loans, the

median origination charges paid by Hispanic White borrowers are $1,485, the median

origination charges paid by Asian borrowers are $1,295, the median origination charges for

Black borrowers are $1,305, and the median origination charges for non-Hispanic White

borrowers are $1,290. The median origination charges for non-Hispanic White VA loans are

slightly higher than that of Black and Hispanic White groups. The median origination charges

for Non-Hispanic White borrowers for RHS/FSA loans are similar to that of Asian borrowers

and are lower than the median origination charges of all other groups.

62

Piggy-back loans are a second mortgage that is made at the same time as the main mortgage to allow borrowers

with low down payment savings to borrow additional money in order to qualify for a main mortgage without paying

for private mortgage insurance.

78

There is no apparent pattern, and there is only limited variation for median origination charges

related to age among conventional conforming, jumbo, and RHS/FSA loans. The median

origination charges however, somewhat increase with age for FHA loans, and significantly

increase with age for VA loans.

The median origination charges are lower for borrowers in high-income tracts than those in

middle-income tracts, which are lower than those in low/moderate-income tracts, across all

enhanced loan types, but only by a relatively small amount. The median origination charges are

lower in rural areas than metropolitan statistical areas among conventional conforming, jumbo

loans, but are slightly higher among FHA and VA loans.

7.5 Discount Points and Lender Credits

Discount Points and Lender Credits are two data points that the 2015 HMDA Rule requires

institutions to report for applicable originated loans. Discount Points and Lender Credits are

among the Discretionary Data Points as discussed in the introduction section of this article.

Discount Points is defined as the points paid to the creditor to reduce the interest rate,

expressed in dollars. Similar to Total Loan Costs and Origination Charges, Discount Points is

applicable only to the originated loans subject to the TRID Closing Disclosure requirements. In

other words, open-end lines of credit, reverse mortgages, loans made primarily for a business

purpose, and loans secured by manufactured homes but not the land do not require reporting of

Discount Points, since they are not subject to TRID Closing Disclosure requirements. Discount

Points is one of the data points that institutions that qualify for a partial exemption under the

EGRRCPA are not required to report.

Discount Points is reported in dollars based on the amount disclosed in the Closing Disclosure,

or “NA” if the requirement to report discount points does not apply, or “Exempt” if the reporter

is exempt from reporting this information under the EGRRCPA. Different from the Total Loan

Costs and the Origination Charges that are defined as “borrower-paid” under the 2015 HMDA

Rule, Discount Points required to be reported under the HMDA Rule are not limited to

“borrower-paid,” but also include any discount points that are paid by the seller or other

parties.

63

The discount points required to be reported by HMDA under Regulation C are equivalent to the sum of all columns

for line 01 (percent of Loan Amount (Points)) of Block A of the “Closing Cost Details” Section of the TRID Closing

Disclosure.

79

TRID rules restrict the Discount Points disclosed in the Closing Disclosure to a positive number.

In some transactions borrowers receive a rebate, sometimes known as “negative discount

points”, typically to cover some of the upfront costs of obtaining a loan and/or home, and in

exchange the borrower is charged a higher interest rate. Such a rebate (negative discount points)

is not captured separately on the Closing Disclosure and thus is not captured in the HMDA

discount points field. Instead, rebates that are directly tied to the interest rate that the borrower

received are included as a part of Lender Credits on the Closing Disclosure and in HMDA.

The Lending Credits data point, newly required under HMDA for applicable originated loans

starting in 2018 HMDA data, is defined as the amount of lender credits, as disclosed on the

TRID Closing Disclosure.

64

It is among the data points that institutions that qualify for the

EGRRCPA partial exemption are not required to report. Lender Credits is reported in dollars, or

“NA” if the requirement to report lending credit does not apply, or “Exempt” if the reporting

institution is exempt from reporting this information under the EGRRCPA. Similar to Total

Loan Costs, Origination Charges, and Discount Points, Lending Credits is not applicable to

open-end lines of credit, reverse mortgages, loans credit made primarily for a business purpose,

and loans secured by manufactured homes but not the land, since they do not require a TRID

Closing Disclosure.

Discount Points and the rebate (negative discount points) included in the lender credits are one

of the important factors related to the final interest rate that the borrowers received. However,

an analysis of how discount points paid and rebates received affects the interest rate is beyond

the scope of this article. Interest rates are also affected by many other factors, such as credit

score, LTV, CLTV, loan type, loan term, loan products, loan amount, occupancy type, lien status,

etc., and the complex behaviors of borrowers and lenders. Instead, this section presents some

basic summary statistics about the Discounts Points and Lender Credits data points reported in

the 2019 HMDA data.

The Discount Points reported under HMDA are in dollars. In practice, when lenders price the

loans and charge discount points on a transaction in exchange for a lower interest rate, discount

points are most commonly calculated in points (i.e. as a percentage of the loan amount, typically

stated as a number by multiply the percentage by 100). Taking that approach, Tables 7.5.1 and

7.5.2 divide the reported discount points by the reported loan amount and multiply by 100 to

convert the dollar amounts reported into points. Loans with missing data on Discount Points are

treated as having zero points.

64

On the TRID Closing Disclosure, the lending credits required to be reported under HMDA are on the “Lending

Credits” line of Block J (TOTAL CLOSING COST) of the “Closing Cost Details” Section.

80

As demonstrated in Table 7.5.1, of all site-built single-family closed-end forward mortgages not

primarily for business or commercial purposes, close to two thirds, or 65.7 percent had zero

discount points. About 15.2 percent of loans have discount points between zero and half a point;

8.4 percent have discount points above half a point but below one point. Overall, about 10.8

percent have reported discount points at one point or higher. Generally, as the discount points

increase, the share of loans having discount points within each consecutive discount points

range decreases, i.e. fewer borrowers are paying them. Among different enhanced loan types,

81.3 percent of jumbo loans paid no discount points at all. In comparison, 65.7 percent of

conventional conforming loans, 63.9 percent of FHA loans, 61.6 percent of VA loans and 64.9

percent of RHS/FSA loans paid no discount points. If the range is broadened to include

borrowers who paid less than one point, 97.4 percent of jumbo loan borrowers either paid no

discount points or paid less than one point. That percentage is 89.2 percent for conventional

conforming loans, 88.2 percent for FHA loans, 87.3 percent for VA loans, and 90.7 percent for

RHS/FSA loans.

Table 7.5.2 breaks down, by race/ethnicity, age, neighborhood income, and geography, the

percentages of loans that had reported discount points in incremental ranges relative to the loan

amount. As it shows, 70.3 percent of Asian borrowers and 67.7 percent of non-Hispanic White

borrowers paid no discount points at all. In comparison, 62.4 percent of Black borrowers, 62.5

percent of Hispanic White borrowers, and 60 percent of “Other” borrowers paid no discount

points. The same pattern generally exists in the higher discount point ranges. If the borrowers

that paid no discount points are combined with those that paid less than one discount point,

92.9 percent of Asian borrowers and 90.8 percent of non-Hispanic White borrowers either paid

no discount points or paid less than one point, compared to 86.1 percent of Black borrowers,

88.9 percent of Hispanic White borrowers, and 86.2 percent of “other” borrowers that paid zero

or less than one discount point.

The percentage of borrowers that paid no discount points decreases monotonically with age.

About 72 percent of borrowers younger than 25 paid no discount points. This percentage

decreases to 69.1 percent for borrowers between 25 and 34 years old, 67 percent for borrowers

between 35 and 44, and all the way to 59.5 percent for borrowers older than 74. The same age

pattern exists if borrowers who paid less than one discount point are included. In the same vein,

the older the borrowers become, the more likely that they would pay discount points in the

higher range, relatively.

The borrowers in high-income census tracts (68.0 percent) are more likely than the borrowers

in middle-income tracts (64.6 percent) not to pay any discount points, who in turn are more

likely than the borrowers in low/moderate-income tracts (63.3 percent) not to pay any discount

points. If expanded to include borrowers who paid less than one discount point, the same

pattern exists. The borrowers in high-income census tracts (91.2 percent) are more likely than

81

the borrowers in middle-income tracts (88.4 percent) not to pay any discount points or to pay

discount points less than one point, who in turn are more likely than the borrowers in

low/moderate-income tracts (87.0 percent) not to pay any discount points or to pay less than

one discount point. In addition, 89.5 percentage of the borrowers in metropolitan statistical

areas either paid no discount points or paid discount points less than one point. This is higher

than the percentage of the borrowers in micropolitan areas who paid no discount points or paid

discount points less than one point (87.9 percent), and the borrowers in rural areas who either

paid no discount points or paid discount points less than one point (87.4 percent).

The Lender Credits reported under HMDA are in dollars. Similar to the treatment of discount

points, to put the amounts of lender credits in relative terms, for Tables 7.5.3 to 7.5.4, the dollar

amount of the Lender Credits as reported in the data are converted to a percentage of the dollar

amount of the loan and multiplied by 100 to be expressed as points. Loans for which the

reported Lender Credit is filed as blank are treated as if the Lender Credit is zero.

As demonstrated in Table 7.5.3, of all site-built single-family closed-end forward mortgages not

primarily for business or commercial purposes, about 61.7 percent received no lender credits.

About 27.9 percent of loans received lender credits between zero and half a point; 5.6 percent

received lender credits above half a point but below one point. About 2.4 percent received lender

credits greater than or equal to one point but less than 1.5 points. Generally, the percentage of

loans within each consecutive lender credits range, when expressed as points relative to the loan

amount, decreases as the lender credits increase, i.e. fewer borrowers received them. Among

different enhanced loan types, 64.3 percent of conventional conforming loans received no lender

credits. In comparison, 48.6 percent of jumbo loans, 56 percent of FHA loans, 59.5 percent of

VA loans, and 60.7 percent of RHS/FSA loans received no lender credits. If the range is

broadened to include borrowers who received less than one point in lender credits, 96.0 percent

of conventional conforming loan borrowers either received no lender credits or received less

than one point in lender credit relative to the loan amount. That percentage is 98.4 percent for

jumbo loans, 90.0 percent for FHA loans, 96.7 percent for VA loans, and 95.8 percent for

RHS/FSA loans.

Table 7.5.4 breaks down, by race/ethnicity, age, neighborhood income, and geography, the

percentages of loans that received lender credits in incremental ranges relative to the loan

amount. As shown, 52.7 percent of Asian borrowers received no lender credit, the lowest among

all racial/ethnic groups. In comparison, 60.1 percent of Black borrowers, 62.6 percent of

Hispanic White borrowers, 62.2 percent of non-Hispanic White borrowers, and 61.9 percent of

“Other” borrowers received no lender credits.

The percentage of borrowers that received no lender credits exhibits no clear pattern related to

age. Borrowers in low/moderate-income tracts (62.5 percent) and borrowers in middle-income

82

tracts (63 percent) are more likely than the borrowers in high-income tracts (59.9 percent) to

receive no lender credits.

A higher percentage of the borrowers in rural areas (66.4 percent) received no lender credits,

than the borrowers in micropolitan areas (65.7 percent), who in turn is higher than the share of

borrowers in metropolitan areas (61.3 percent) that received no lender credits.

It is important to note that the summary statistics on the incidence and magnitude of discount

points and lender credits presented in this section have not controlled for the borrowers’ credit

characteristics and characteristics of the loans, which, if included (though beyond the scope of

this article), may help explain some of the differences observed across different categories of

loans, borrowers, neighborhood income, and geography etc. as shown above.

Lastly, Lender Credits, as disclosed in the Closing Disclosure and reported under HMDA, may

include lender credits given to borrowers for reasons other than choosing a higher interest rate

in exchange for reduced upfront costs.

65

In other words, the lender credits reported under

HMDA may not perfectly mirror the definition of the Discount Points reported under HMDA

and thus should not be viewed as the equivalence of the negative direction, i.e. being negative

discount points. To illustrate this issue, Table 7.5.5 shows for loans with reported discount

points within various ranges, the counts and percentage of the loans that also reported a lender

credit within certain ranges. For instance, among loans that reported zero discount points, about

55.4 percent had no lender credit, 31.0 percent had lender credits in the zero to 0.5 point range,

7.5 percent had lender credits in the 0.5 to one point range, and 3.2 percent had lender credits in

the one to 1.5 points range. At least some portions of those lender credits for the loans with zero

discount points could be “negative discount points” directly tied to the interest rates. However,

many loans that reported charged discount points reported receiving a lender credit as well. For

instance, for loans that reported discount points between one and 1.5 points, only 74.4 percent

reported no lender credits at all, the rest, or 25.6 percent, reported a lender credit that is

positive. In such cases, the positive lender credits reported are most likely not negative discount

points, but rather lender credits for other reasons.

65

For instance, the Lender Credits may include lender credits given to the borrowers to correct processing errors,

lender credits due to the banking relationship, lending credits for Community Reinvested Act (CRA) related loans,

lender credits due to promotional campaigns, etc.

83

8. Miscellaneous

HMDA data beginning in 2018 also incorporate a few miscellaneous changes to data points.

Compared to HMDA data prior to 2018, Legal Entity Identifier (LEI) replaces the respondent ID

coupled with the agency code that previously served as the main lender ID in HMDA data. Each

entity reporting under HMDA is required to obtain a LEI issued by either a utility endorsed by

the LEI Regulatory Oversight Committee; or a utility endorsed or otherwise governed by the

Global LEI Foundation (GLEIF) (or any successor of the GLEIF) after the GLEIF assumes

operational governance of the global LEI system. The users of the publicly released HMDA data

can use the HMDA panel to help link the current reporters to their previous agency code and

respondent ID if the reporters reported HMDA data in the past.

Universal Loan ID (ULI) is a unique ID assigned to each covered loan or application according

to requirements set by the 2015 HMDA Rule. ULI is one of the Mandated Data Points as

discussed in the introduction section of this article. ULI is one of the data points that

institutions that qualify for a partial exemption under the EGRRCPA are not required to report.

For reporters exempt under the EGRRCPA that choose not to report a ULI, the Bureau’s 2018

HMDA Rule sets out different requirements to report a non-universal loan identifier (NULI) for

the covered loans or applications. ULIs and NULIs are excluded from the public loan-level

HMDA data.

Reporters are required to collect and report into the Nationwide Mortgage Licensing System and

Registry (NMLSR) the mortgage loan originator unique identifier (NMLSR ID) for the mortgage

loan originator for applicable transactions. NMLSR ID is one of the Mandated Data Points as

discussed in the introduction section of this article. The NMLSR ID is one of the data points that

certain institutions are exempt from reporting for eligible transactions under the EGRRCPA.

The NMLSR ID is excluded from the public loan-level HMDA data.

Additionally, under the DFA as implemented by the 2015 HMDA Rule, reporters are required to

report the address of the property securing the covered loan or, in the case of an application,

proposed to secure the covered loan. Property Address is one of the Mandated Data Points as

discussed in the introduction section of this article. This fulfills the DFA’s mandate to collect

parcel IDs for the properties reported under HMDA. Property address is one of the data points

certain institutions are exempt from reporting for eligible transactions under the EGRRCPA.

The property addresses are not included in the public release HMDA data.

Finally, the 2015 HMDA Rule requires reporters to report, except for purchased covered loans,

the name of the automated underwriting system used by the financial institution to evaluate the

application and the result generated by that automated underwriting system. Automated

84

Underwriting System is one of the Discretionary Data Points as discussed in the introduction

section of this article. An “automated underwriting system,” defined under Regulation C means

an electronic tool developed by a securitizer, Federal government insurer, or Federal

government guarantor of closed-end mortgage loans or open-end lines of credit that provide a

result regarding the credit risk of the applicant and whether the covered loan is eligible to be

originated, purchased, insured, or guaranteed by that securitizer, Federal government insurer,

or Federal government guarantor.

66

The information regarding the automated underwriting system is among the data points that

certain institutions are exempt from reporting for eligible transactions under the EGRRCPA.

The automated underwriting system result and free form text fields used to report the name of

the automated underwriting system are excluded from the public loan-level HMDA data.

66

For purposes of Regulation C, a person is a securitizer, Federal government insurer, or Federal government

guarantor of closed-end mortgage loans or open-end lines of credit, respectively, if it has ever securitized, provided

Federal government insurance, or provided a Federal government guarantee for a closed-end mortgage loan or open-

end line of credit.

85

APPENDIX A: TABLES

TABLE 2.1.1: TOP 25 REPORTERS BY TOTAL OPEN-END ORIGINATIONS

Institution

type

Applications

(thousands)

Originations

(thousands)

Purchases

(thousands)

Assets

($ Millions)

Market

Share (%)

Bank of America, National

Association

Large bank

245

79

2

1,751,524

7.4

Wells Fargo Bank, National

Association

Large bank

185

51

7

1,747,398

4.8

U.S. Bank National Association

Large bank

95

43

0

456,026

4.1

Citizens Bank, National

Association

Large bank

81

40

0

121,996

3.7

PNC Bank, National Association

Large bank

61

36

0

370,002

3.4

The Huntington National Bank

Large bank

62

36

0

104,052

3.4

JPMorgan Chase Bank, National

Association

Large bank

126

34

0

2,140,778

3.2

Truist Bank

Large bank

50

29

0

216,077

2.7

SUNTRUST BANKS, INC.

Large bank

64

24

0

201,638

2.3

Fifth Third Bank, National

Association

Large bank

40

22

0

140,078

2.1

Third Federal Savings and Loan

Association of Cleveland

Large bank

28

17

0

13,855

1.6

TCF National Bank

Large bank

20

16

0

23,021

1.5

Regions Bank

Large bank

36

16

0

123,325

1.5

KeyBank National Association

Large bank

26

15

0

135,758

1.4

86

TD Bank, National Association

Large bank

37

14

0

302,669

1.3

Boeing Employees Credit Union

Credit union

19

12

0

302,669

1.1

AMERICAN ADVISORS GROUP

Ind. mort. co.

18

11

2

44,465

1.0

Manufacturers and Traders Trust

Company

Large bank

19

11

0

118,072

1.0

Navy Federal Credit Union

Credit union

22

10

0

90,566

1.0

Zions Bancorporation, National

Association

Large bank

18

10

0

66,081

1.0

BMO Harris Bank National

Association

Large bank

22

10

0

109,373

0.9

First-Citizens Bank & Trust

Company

Large bank

13

9

0

34,347

0.9

Citibank, National Association

Large bank

24

9

0

1,384,707

0.9

Bank of the West

Large bank

17

9

0

89,766

0.9

STATE EMPLOYEES'

Credit union

12

8

0

37,319

0.8

Top 25 institutions

...

1,342

573

10

10,125,559

53.6

All institutions

...

2,061

1,069

31

15,152,609

100.0

NOTE: Open-end records only. Ranked by open-end origination volume.

87

TABLE 2.1.2: OPEN-END REPORTERS BY ORIGINATION SIZE CATEGORY

Reporters

Originations

(thousands)

Applications (thousands)

Origination size category

1-99

413

8

12

100-199

54

8

13

200-499

118

42

64

500-999

127

93

143

1000-4999

136

275

441

>=5000

37

643

1,470

Total

885

1,069

2,142

NOTE: Open-end records only.

88

TABLE 2.2.1: TOP 10 REPORTERS BY TOTAL REVERSE MORTGAGE ORIGINATIONS

Institution

type

Applications

Originations

Purchases

Assets

($ Millions)

Market

Share (%)

AMERICAN ADVISORS GROUP

Ind. mort. co.

19,104

11,418

2,086

44,465

32.8

FINANCE OF AMERICA

REVERSE LLC

Ind. mort. co.

8,060

6,485

1,279

54,280

18.6

REVERSE MORTGAGE

FUNDING LLC

Ind. mort. co.

4,759

3,196

922

40,859

9.2

ONE REVERSE MORTGAGE,

LLC

Ind. mort. co.

5,780

2,931

0

86

8.4

LIBERTY HOME EQUITY

SOLUTIONS, INC.

Ind. mort. co.

4,364

2,446

1,422

23,298

7.0

SYNERGY ONE LENDING, INC.

Ind. mort. co.

3,126

2,303

19

2

6.6

Longbridge Financial, LLC

Ind. mort. co.

1,538

897

3,853

520

2.6

Open Mortgage, LLC

Ind. mort. co.

1,213

846

0

269

2.4

HighTechLending Inc

Ind. mort. co.

955

649

0

135

1.9

CHERRY CREEK MORTGAGE

CO., INC.

Ind. mort. co.

573

491

0

1,402

1.4

Top 10 institutions

...

49,472

31,662

9,581

165,316

90.9

All institutions

...

50,243

34,833

20,425

15,152,609

100.0

NOTE: Reverse mortgage records only. Ranked by reverse mortgage origination volume.

89

TABLE 2.2.2: REVERSE MORTGAGE REPORTERS BY ORIGINATION SIZE CATEGORY

Reporters

Originations

Applications

Origination size category

1-99

90

918

1,362

100-199

9

1,239

2,088

200-499

4

1,505

2,015

500-999

3

2,392

3,706

1000-4999

4

10,876

18,029

>=5000

2

17,903

27,164

Total

112

34,833

54,364

NOTE: Reverse mortgage records only.

90

TABLE 2.3.1: REVERSE MORTGAGE BY OPEN-END FLAG (COUNTS IN THOUSANDS)

Open-end line of credit

Yes

No

Exempt

Total

Count

%

Count

%

Count

%

Count

%

Reverse mortgage

Originations

Yes

56.4

74.7

19.1

25.3

0.0

75.5

100.0

No

2,116.8

12.5

14,883.3

87.5

0.5

0.0

17,000.6

100.0

Exempt

0.2

0.0

3.0

0.6

466.1

99.3

469.3

100.0

Total

2,173.4

12.4

14,905.5

85.0

466.6

2.7

17,545.5

100.0

All LARs

Yes

27.1

77.7

7.8

22.3

0.0

34.8

100.0

No

1,041.9

11.6

7,911.3

88.4

0.3

0.0

8,953.5

100.0

Exempt

0.1

0.0

2.1

0.6

334.7

99.4

336.9

100.0

Total

1,069.1

11.5

7,921.1

84.9

335.1

3.6

9,325.2

100.0

NOTE: All originations, and all LARs.

91

TABLE 2.3.2: CLOSED-END, HELOC AND REVERSE MORTGAGE BY ACTION TYPE (COUNTS IN THOUSANDS)

Closed-end

HELOC

Reverse

mortgage

Total

Action type

Originated

Count

7,911.3

1,041.9

34.8

8,988.0

%

53.2

49.2

46.1

52.6

Approved, not Accepted

Count

328.3

56.4

1.2

386.0

%

2.2

2.7

1.6

2.3

Denied

Count

1,699.3

784.2

7.6

2,491.1

%

11.4

37.0

10.0

14.6

Withdrawn

Count

1,960.5

139.3

6.6

2,106.4

%

13.2

6.6

8.7

12.3

Closed for Incompleteness

Count

629.5

77.7

4.8

712.1

%

4.2

3.7

6.4

4.2

Purchased

Count

2,210.1

17.3

20.4

2,247.8

%

14.8

0.8

27.1

13.2

92

Closed-end

HELOC

Reverse

mortgage

Total

Preapproval Request Denied

Count

74.1

0

74.1

%

0.5

0.0

0.4

Preapproval Approved not

Accepted

Count

70.2

0

70.2

%

0.5

0.0

0.4

Total

Count

14,883.3

2,116.8

75.5

17,075.7

%

100.0

NOTE: Closed-end records are defined as records with open-end lines of credit flag = 2 and reverse mortgage flag = 2. HELOCs are defined as records with

open-end lines of credit flag = 1 and reverse mortgage flag = 2. Reverse mortgages are defined as records with reverse mortgage flag = 1. Records with open-

end lines of credit flag = 1111 or reverse mortgage flag = 1111 due to the partial exemption under the EGRRCPA are excluded. The same definitions also apply

to all subsequent tables where these terms are used.

93

TABLE 2.3.3: CLOSED-END, HELOC AND REVERSE MORTGAGE ORIGINATIONS BY RACE/ETHNICITY, NEIGHBORHOOD INCOME, AND

GEOGRAPHY (COUNTS IN THOUSANDS)

Transaction type

Closed-end

HELOC

Reverse mortgage

Total

Count

%

Count

%

Count

%

Count

%

Borrower race and ethnicity

Asian

454.2

5.7

50.2

4.8

0.6

1.7

505.0

5.6

Black

475.8

6.0

33.5

3.2

2.5

7.2

511.8

5.7

Hispanic white

613.4

7.8

36.8

3.5

1.6

4.4

651.7

7.3

Joint

269.8

3.4

33.1

3.2

0.6

1.7

303.4

3.4

Non-Hispanic white

4,700.2

59.4

751.7

72.1

26.0

74.7

5,478.0

60.9

Other

62.4

0.8

8.9

0.9

0.2

0.6

71.5

0.8

Missing

1,335.5

16.9

127.7

12.3

3.3

9.6

1,466.6

16.3

Total

7,911.3

100.0

1,041.9

100.0

34.8

100.0

8,988.0

100.0

Neighborhood income

Low or moderate

1,315.8

16.7

121.9

11.7

6.5

18.7

1,444.1

16.1

Middle

3,459.4

44.0

436.1

41.9

15.4

44.3

3,910.9

43.7

High

3,096.0

39.3

482.2

46.4

12.9

37.0

3,591.1

40.1

Total

7,871.1

100.0

1,040.2

100.0

34.8

100.0

8,946.1

100.0

Geography

Metropolitan Area

7,130.9

90.1

947.0

90.9

31.8

91.4

8,109.7

90.2

94

Transaction type

Closed-end

HELOC

Reverse mortgage

Total

Count

%

Count

%

Count

%

Count

%

Micropolitan Area

481.8

6.1

62.4

6.0

1.9

5.4

546.1

6.1

Rural

298.7

3.8

32.5

3.1

1.1

3.2

332.3

3.7

Total

7,911.3

100.0

1,041.9

100.0

34.8

100.0

8,988.0

100.0

NOTE: Originations only. Originations with the values for "open-end lines of credit flag" or "reverse mortgage flag" equal to 1111 (Exempt under the EGRRCPA)

are excluded from the analysis. The total counts may vary across groups due to missing values in this table and other tables.

The following categorization rules apply to this table and all subsequent tables where race/ethnicity, neighborhood income, and geography comparisons are

displayed:

1) For race and ethnicity categorization, applications/loans are placed in one category for race and ethnicity. The application is designated as "joint" if one

applicant was reported as White and the other was reported as one or more minority races, or if the application is designated as White with one Hispanic

applicant and one non-Hispanic applicant. If there are two applicants and each reports a different minority race, the application is designated as two or more

minority races. If an applicant reports multiple races and one is White, that applicant is categorized under the minority race. Otherwise, the applicant is

categorized under the first race reported. "Missing" refers to applications in which the race of the applicant(s) has not been reported or is not applicable or the

application is categorized as White but ethnicity has not been reported. "Other" consists of applications by American Indians or Alaska Natives, Native Hawaiians

or other Pacific Islanders, and borrowers reporting two or more minority races.

2) The categories for the neighborhood-income group are based on the ratio of census-tract median family income to area median family income from the 2011-

15 American Community Survey data. Low- or moderate-income (or LMI) census tracts have census-tract median family income that is less than 80 percent of

estimated current area median family income (AMFI), middle-income census tracts have census-tract median family income that is at least 80 percent and less

than 120 percent of AMFI, and high-income census tracts have census-tract median family income that is at least 120 percent of AMFI.

95

3) For geography categorization, metropolitan areas refer to metropolitan statistical areas (MSA), micropolitan areas refer to micropolitan statistical areas, and

rural areas refer to areas that are neither in a metropolitan statistical areas nor in a micropolitan statistical area. The geography is based on the reported county

and state mapped to the list of metropolitan and micropolitan statistical areas published by the OMB on September 14, 2018.

Some records have county or state information reported as not applicable. Such records cannot be matched to the metropolitan and micropolitan statistical areas

list, and thus are excluded from this and all other tables in which metropolitan, micropolitan, and rural area comparisons are displayed. Note that such cases may

be for an application where the property location information was not known at the time when application was denied, withdrawn, or closed for incompleteness,

or if the property is in an MSA or Metropolitan Division (MD) where the reporting financial institution did not have a home or branch office and the financial

institution was not subject to Community Reinvestment Act of 1977. Specifically, according to Regulation C, a financial institution is required to report the state,

county of the property securing the covered loan or, in the case of an application, proposed to secure the covered loan if the property is located in an MSA or MD

in which the financial institution has a home or branch office or if the institution is subject to § 1003.4(e) of Regulation C. Furthermore § 1003.4(e) of Regulation

C states that banks and savings associations that are required to report data on small business, small farm, and community development lending under

regulations that implement the Community Reinvestment Act of 1977 shall also collect state, county information for property located outside of MSAs and MDs in

which the institution has a home or branch office, or outside of any MSA. Financial institutions can also voluntarily report county and state information even if

they are not required to. Given such requirements, it is likely that some records with state and county information reported as NA are in micropolitan statistical or

rural areas, but their metropolitan/micropolitan/rural status cannot be affirmatively determined and hence are omitted from the analyses.

96

TABLE 2.3.4: TRANSACTION TYPE BY LOAN TYPE (COUNTS IN THOUSANDS)

Transaction type

Closed-end

HELOC

Reverse mortgage

Total

Count

%

Count

%

Count

%

Count

%

Loan type

Originations

Conventional

5,843.7

73.9

1,041.8

100.0

3.4

9.7

6,888.9

76.6

Non-conventional

FHA

1,139.1

14.4

0.1

0.0

31.5

90.3

1,170.6

13.0

VA

829.0

10.5

0.0

829.0

9.2

RHS/FSA

99.6

1.3

0.0

0

0.0

99.6

1.1

Total

7,911.3

100.0

1,041.9

100.0

34.8

100.0

8,988.0

100.0

Purchases

Conventional

1,252.5

56.7

17.3

100.0

0.7

3.3

1,270.5

56.5

Non-conventional

FHA

589.3

26.7

0

0.0

19.7

96.7

609.1

27.1

VA

304.4

13.8

0.0

304.4

13.5

RHS/FSA

63.8

2.9

0

0.0

0

0.0

63.8

2.8

Total

2,210.1

100.0

17.3

100.0

20.4

100.0

2,247.8

100.0

NOTE: Originations and purchases with open-end lines of credit flag = 1111 or reverse mortgage flag = 1111 due to the partial exemptions under the EGRRCPA

are excluded.

97

TABLE 2.3.5: BASIC CHARACTERISTICS BY ENHANCED LOAN TYPE

Originations

(thousands)

Mean

Income ($

thousands)

Median

Income ($

thousands)

Mean Loan

Amount ($

thousands)

Median Loan

Amount ($

thousands)

Home Purchase

(%)

Refinance

(%)

First Lien

(%)

Enhanced loan type

Conventional

Conforming

5,320.5

134

95

246.3

223.3

52.3

44.9

94.5

Jumbo

373.9

494

300

1,010.5

815.0

54.8

44.5

99.0

Non-conventional

FHA

1,106.8

73

66

225.8

208.9

68.0

31.2

99.7

VA

815.6

83

75

291.7

263.4

48.5

51.2

100.0

RHS/FSA

98.9

54

52

154.8

146.5

97.6

2.4

99.9

HELOC

1,036.3

149

107

113.1

75.0

7.2

41.7

28.3

Reverse Mortgage

34.1

32

28

226.2

163.2

6.7

90.1

100.0

Total

8,786.3

138

92

263.7

215.0

50.6

43.1

88.1

NOTE: Site-built, single-family originations only. The median loan amounts in the table are calculated from non-public raw data reported by financial institutions.

The outliers are excluded from the analysis sample to produce consistent estimates.

98

TABLE 3.1.1: BORROWER AGE BY ENHANCED LOAN TYPE: ORIGINATIONS ONLY

Enhanced loan type

Conventional

Non-conventional

HELOC

Reverse

mortgage

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Age group share (%)

<=24

2.7

0.1

6.1

2.8

19.1

0.2

0.0

2.9

25-34

21.8

13.4

29.5

21.8

42.9

6.8

0.0

20.8

35-44

25.2

35.5

27.9

22.4

19.8

19.1

0.0

24.9

45-54

21.6

27.2

20.6

19.6

10.4

25.6

0.0

21.8

55-64

17.2

15.8

11.1

14.5

5.6

25.5

9.7

16.9

65-74

8.9

6.3

3.9

14.3

1.9

16.5

46.0

9.6

>=75

2.6

1.6

0.9

4.7

0.4

6.3

44.4

3.1

Total

100.0

Summary

Mean age

46.1

46.6

41.1

47.8

34.3

54.0

74.1

46.5

Median age

45.0

39.0

46.0

31.0

54.0

73.0

45.0

Count (thousands)

5,429.1

414.5

1,139.1

829.0

99.6

1,041.9

34.8

8,988.0

NOTE: Originations. Age is for applicants only, not taking into account of the co-applicant's age, in this and all other tables in this article. The mean and median

ages in the table are calculated from non-public raw data reported by financial institutions. The outliers are excluded from the analysis sample to produce

consistent estimates.

99

TABLE 3.1.2: BORROWER AGE BY RACE/ETHNICITY: ORIGINATIONS, EXCLUDING REVERSE MORTGAGES

Race and ethnicity

Asian

Black

Hispanic

white

Joint

Non-Hispanic

white

Other

Missing

Total

Age group share (%)

<=24

1.6

1.9

4.6

2.8

3.2

3.7

2.0

2.9

25-34

23.8

18.4

25.7

25.0

20.3

21.4

19.2

20.8

35-44

34.1

25.0

28.4

28.3

23.0

26.0

26.5

24.9

45-54

23.8

23.5

22.5

21.2

21.3

22.1

23.0

21.9

55-64

11.4

18.5

12.6

14.1

18.0

15.8

17.4

17.0

65-74

4.2

9.9

4.9

6.9

10.8

8.4

9.2

9.5

>=75

1.1

2.9

1.3

1.8

3.5

2.5

2.7

3.0

Total

100.0

Summary

Mean age

43.1

47.5

42.8

44.2

47.2

45.5

46.8

46.5

Median age

42.0

46.0

41.0

42.0

46.0

44.0

45.0

Count (thousands)

511.8

517.7

661.2

309.5

5,686.6

72.7

1,531.0

9,290.4

NOTE: Forward originations only, excluding reverse mortgages. The mean and median ages in the table are calculated from non-public raw data reported by

financial institutions. The outliers are excluded from the analysis sample to produce consistent estimates.

100

TABLE 3.1.3: BORROWER AGE BY RACE/ETHNICITY: HOME PURCHASE ORIGINATIONS, EXCLUDING HELOCS AND REVERSE MORTGAGES

Race and ethnicity

Asian

Black

Hispanic

white

Joint

Non-Hispanic

white

Other

Missing

Total

Age group share (%)

<=24

2.7

3.2

7.1

5.1

6.2

7.1

4.2

5.5

25-34

31.6

26.7

33.2

36.4

31.0

32.8

29.3

30.9

35-44

35.2

29.5

28.7

28.6

23.8

27.8

28.3

26.1

45-54

19.7

21.5

18.7

15.7

17.0

17.5

18.7

17.8

55-64

7.9

13.1

8.9

9.2

13.1

9.8

12.5

12.2

65-74

2.3

5.1

2.9

4.0

7.1

4.1

5.7

5.9

>=75

0.5

1.0

0.6

0.9

1.8

0.9

1.3

1.5

Total

100.0

Summary

Mean age

40.2

42.9

39.5

39.8

42.3

40.1

42.3

41.9

Median age

38.0

41.0

38.0

37.0

39.0

38.0

40.0

39.0

Count (thousands)

255.5

288.1

391.5

151.9

2,523.0

34.5

650.0

4,294.6

NOTE: Closed-end home-purchase originations. The mean and median ages in the table are calculated from non-public raw data reported by financial institutions.

The outliers are excluded from the analysis sample to produce consistent estimates.

101

TABLE 3.1.4: DENIAL RATES OF APPLICANTS AGE 62 OR OLDER BY ENHANCED LOAN TYPE (PERCENT)

Applicant age 62 or older

Yes

No

Total

Enhanced loan type

Conventional

Conforming

17.3

12.0

12.9

Jumbo

22.0

13.8

14.7

Non-conventional

FHA

35.9

17.7

19.3

VA

22.7

14.0

16.1

RHS/FSA

15.9

12.5

12.6

HELOC

34.6

46.8

40.6

Reverse Mortgage

17.0

26.1

17.1

Total

22.6

14.5

16.0

NOTE: Site-built single-family, first-lien, owner-occupied only. The denial rates are calculated based on applications that were denied, divided by (applications

that were denied + applications that were approved but not accepted + loans originated). The denial rate calculations do not include applications that were

withdrawn or files that were closed for incompleteness.

102

TABLE 3.2.1: DISTRIBUTION OF THE FIRST REPORTED RACE OF APPLICANTS (COUNTS IN THOUSANDS)

Count

%

First reported race

American Indian or Alaska Native

126.2

0.7

Asian

825.8

4.7

Black or African American

1,172.6

6.7

Native Hawaiian or Other Pacific

Islander

38.2

0.2

White

11,271.6

64.2

Asian Indian

55.8

0.3

Chinese

17.2

0.1

Filipino

17.5

0.1

Japanese

2.9

0.0

Korean

7.1

0.0

Vietnamese

8.4

0.0

Other Asian

17.7

0.1

Native Hawaiian

0.8

0.0

Guamanian or Chamorro

0.8

0.0

Samoan

0.5

0.0

Other Pacific Islander

13.4

0.1

Not available or Missing

3,968.9

22.6

Total

17,545.5

100.0

103

NOTE: Race of applicants only. The estimates are calculated from non-public raw data reported by financial institutions. Code 3 (Information not provided by

applicant in mail, internet, or telephone application), code 4 (Not applicable), and missing observations are combined into one category under "Not available or

missing."

104

TABLE 3.2.2: NUMBER OF RACES SELECTED BY THE FIRST REPORTED RACE OF APPLICANTS (PERCENT)

Number of races

0

1

2

3

4

5

Total

First reported race

American Indian or Alaska Native

0.0

59.7

35.4

3.8

0.6

0.4

100.0

Asian

0.0

94.3

5.4

0.3

0.0

100.0

Black or African American

0.0

97.5

2.4

0.1

0.0

100.0

Native Hawaiian or Other Pacific

Islander

0.0

87.6

12.0

0.4

0.0

100.0

White

0.0

99.8

0.2

0.0

100.0

Not Available or Missing

100.0

0.0

100.0

Total

22.6

76.4

0.9

0.1

0.0

100.0

NOTE: Race of applicants only. The estimates are calculated from non-public raw data reported by financial institutions. Code 3 (Information not provided by

applicant in mail, internet, or telephone application), code 4 (Not applicable), and missing observations are combined into one category under "Not available or

missing." The disaggregated categories of Asian and Native Hawaiian or Other Pacific Islander are aggregated for this analysis.

105

TABLE 3.2.3: SECOND REPORTED RACE CONDITIONAL ON THE FIRST REPORTED RACE OF APPLICANTS

Second reported race

American Indian

or Alaska Native

Asian

Black or African

American

Native Hawaiian

or Other Pacific

Islander

White

Asian Indian

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

First reported race

American Indian or Alaska

Native

0

0.0

2.4

1.9

9.1

7.2

0.4

0.3

37.7

29.9

0.2

0.1

Asian

0.0

0

0.0

3.1

0.4

4.6

0.6

41.2

5.0

147.8

17.9

Black or African American

0.1

0.0

0

0.0

1.0

0.1

24.9

2.1

0.2

0.0

Native Hawaiian or Other Pacific

Islander

0.0

0

0.0

5.6

14.6

0.1

White

0.5

0.0

0.2

0.0

0.1

0.0

2.1

0.0

Asian Indian

0.0

1.7

3.0

0.0

0.1

0.2

0

0.0

Chinese

0

0.0

0.6

3.7

0.0

0.1

0.4

0.0

Filipino

0

0.0

0.2

1.3

0.0

0.1

0.0

0.1

0.2

1.3

0.0

Japanese

0

0.0

0.1

2.0

0.0

0.6

0.0

0.1

3.8

0

0.0

Korean

0

0.0

0.3

4.3

0.0

0.2

0.0

0.1

1.1

0.0

Vietnamese

0

0.0

0.3

3.1

0.0

0.5

0.0

Other Asian

0.0

0.3

1.9

0.1

0.3

0.0

0.1

0.3

1.7

0.0

0.1

Native Hawaiian

0

0.0

0

0.0

0

0.0

0

0.0

2.8

0

0.0

Guamanian or Chamorro

0

0.0

0

0.0

0

0.0

0

0.0

1.7

0

0.0

106

Second reported race

American Indian

or Alaska Native

Asian

Black or African

American

Native Hawaiian

or Other Pacific

Islander

White

Asian Indian

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Samoan

0

0.0

0

0.0

0

0.0

0

0.0

2.5

0

0.0

Other Pacific Islander

0

0.0

0

0.0

0.1

0.7

0.0

0.1

Not available or missing

0

0.0

0

0.0

0

0.0

0

0.0

0

0.0

Total

0.6

0.0

6.1

0.0

12.5

0.1

6.2

0.0

110.5

0.6

150.4

0.9

107

TABLE 3.2.3: NUMBER OF RACES SELECTED BY THE FIRST REPORTED RACE OF APPLICANTS (COUNTS IN THOUSANDS) continued

Second reported race

Chinese

Filipino

Japanese

Korean

Vietnamese

Other Asian

Native Hawaiian

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

First reported race

American Indian or Alaska Native

0.1

0.0

0.4

0.3

0.0

Asian

120.0

14.5

69.1

8.4

15.6

1.9

37.5

4.5

52.1

6.3

49.8

6.0

0.0

Black or African American

0.2

0.0

0.4

0.0

0.1

0.0

0.2

0.0

0.1

0.0

0.6

0.1

0.0

Native Hawaiian or Other Pacific

Islander

0.1

0.3

0.4

1.0

0.0

0.1

0.0

0.1

0.3

4.6

12.0

White

1.7

0.0

4.7

0.0

2.1

0.0

1.5

0.0

0.9

0.0

5.7

0.1

0.4

0.0

Asian Indian

0.1

0.0

0.1

0.3

0.5

0.0

Chinese

0

0.0

0.2

0.9

0.1

0.5

0.0

0.2

0.9

0.1

0.4

0.0

0.3

Filipino

0.0

0.1

0

0.0

0.1

0.4

0.0

0.1

0.0

0.1

0.4

0.0

0.2

Japanese

0.0

0.1

0.0

0.1

0

0.0

1.0

0.0

0.1

0.0

0.8

0.0

1.5

Korean

0.0

0.1

0

0.0

0.1

0.0

0.4

0.0

0.1

Vietnamese

0.0

0

0.0

0.4

0.0

Other Asian

0.0

0.1

0.0

0

0.0

0

0.0

Native Hawaiian

0.0

0.5

0.0

0.8

0.0

0.6

0.0

0.2

0

0.0

0

0.0

0

0.0

Guamanian or Chamorro

0

0.0

0.1

0.0

0.1

0

0.0

0

0.0

0.1

0.0

0.1

108

Second reported race

Chinese

Filipino

Japanese

Korean

Vietnamese

Other Asian

Native Hawaiian

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Samoan

0

0.0

0

0.0

0

0.0

0

0.0

0

0.0

0

0.0

0

0.0

Other Pacific Islander

0.0

0.1

0.0

0.2

0

0.0

Not available or missing

0

0.0

0

0.0

0

0.0

0

0.0

0

0.0

0

0.0

Total

122.1

0.7

74.9

0.4

18.1

0.1

39.3

0.2

53.3

0.3

57.2

0.3

5.3

0.0

109

TABLE 3.2.3: NUMBER OF RACES SELECTED BY THE FIRST REPORTED RACE OF APPLICANTS (COUNTS IN THOUSANDS)

continued

Second reported race

Guamanian or

Chamorro

Samoan

Other Pacific

Islander

Not available or

missing

Total

Count

%

Count

%

Count

%

Count

%

Count

%

First reported race

American Indian or Alaska Native

0.0

0.3

75.4

59.8

126.2

100.0

Asian

0.0

0.4

0.0

284.4

34.4

825.8

100.0

Black or African American

0.0

0.8

0.1

1,143.8

97.6

1,172.6

100.0

Native Hawaiian or Other Pacific

Islander

1.6

4.2

1.6

4.3

5.7

14.9

18.4

48.2

38.2

100.0

White

0.4

0.0

0.3

0.0

2.9

0.0

11,248.2

99.8

11,271.6

100.0

Asian Indian

0

0.0

0.1

0.2

53.4

95.7

55.8

100.0

Chinese

0.0

0.1

15.9

92.3

17.2

100.0

Filipino

0.0

0.2

0.0

0.2

0.9

16.6

95.0

17.5

100.0

Japanese

0.0

0.3

0.0

0.2

0.0

0.3

2.6

89.1

2.9

100.0

Korean

0.0

0.1

0.0

0.1

0.0

0.1

6.6

93.5

7.1

100.0

Vietnamese

0

0.0

0

0.0

0.2

8.0

95.6

8.4

100.0

Other Asian

0.0

0.1

0.0

0.6

3.4

16.4

92.3

17.7

100.0

Native Hawaiian

0.0

1.1

0.0

1.1

0.0

3.5

0.8

89.4

0.8

100.0

110

Second reported race

Guamanian or

Chamorro

Samoan

Other Pacific

Islander

Not available or

missing

Total

Count

%

Count

%

Count

%

Count

%

Count

%

Guamanian or Chamorro

0

0.0

0.1

0.0

1.6

0.7

96.0

0.8

100.0

Samoan

0

0.0

0

0.0

3.9

0.5

93.5

0.5

100.0

Other Pacific Islander

0.0

0

0.0

13.3

98.8

13.4

100.0

Not available or missing

0

0.0

0

0.0

3,968.9

100.0

3,968.9

100.0

Total

2.1

0.0

2.0

0.0

11.1

0.1

16,873.8

96.2

17,545.5

100.0

NOTE: Race of applicants only. The estimates are calculated from non-public raw data reported by financial institutions. Code 3 (Information not provided by

applicant in mail, internet, or telephone application), code 4 (Not applicable), and missing observations are combined into one category under "Not available or

missing".

111

TABLE 3.2.4: DISTRIBUTION OF THE FIRST REPORTED ETHNICITY OF APPLICANTS (COUNTS IN THOUSANDS)

Count

%

First reported ethnicity

Hispanic or Latino

1,567.9

8.9

Not Hispanic or Latino

11,921.4

67.9

Mexican

44.9

0.3

Puerto Rican

15.1

0.1

Cuban

7.3

0.0

Other Hispanic or Latino

61.9

0.4

Not available or Missing

3,926.9

22.4

Total

17,545.5

100.0

NOTE: Ethnicity of applicants only. The estimates are calculated from non-public raw data reported by financial institutions. Code 3 (Information not provided by

applicant in mail, internet, or telephone application), code 4 (Not applicable), and missing observations are combined into one category under "Not available or

missing."

112

TABLE 3.2.5: SECOND REPORTED ETHNICITY CONDITIONAL ON THE FIRST REPORTED ETHNICITY OF APPLICANTS (COUNTS IN

THOUSANDS)

Second reported ethnicity

Hispanic or

Latino

Not Hispanic

or Latino

Mexican

Puerto Rican

Cuban

Other Hispanic

or Latino

Not available or

missing

Total

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

First reported ethnicity

Hispanic or Latino

0.0

11.0

0.7

432.3

27.6

97.5

6.2

43.0

2.7

139.3

8.9

844.9

53.9

1,567.9

100.0

Not Hispanic or Latino

0.1

0.0

1.4

0.0

1.1

0.0

0.7

0.0

22.7

0.2

11,895.4

99.8

11,921.4

100.0

Mexican

2.6

5.7

0.1

0.0

0.2

0.4

0.1

0.5

1.0

41.6

92.5

44.9

100.0

Puerto Rican

0.6

4.3

0.1

0.4

0.0

0.1

0.8

0.2

1.4

14.0

93.1

15.1

100.0

Cuban

0.3

3.7

0.0

0.5

0.0

0.1

0.0

0.1

1.9

6.8

93.8

7.3

100.0

Other Hispanic or Latino

1.1

1.7

1.5

2.4

0.0

59.3

95.8

61.9

100.0

Not available or missing

0.0

3,926.9

100.0

3,926.9

100.0

Total

4.7

0.0

12.6

0.1

433.7

2.5

98.8

0.6

43.8

0.2

162.8

0.9

16,789.0

95.7

17,545.5

100.0

NOTE: Ethnicity of applicants only. The estimates are calculated from non-public raw data reported by financial institutions. Code 3 (Information not provided by

applicant in mail, internet, or telephone application), code 4 (Not applicable), and missing observations are combined into one category under "Not available or

missing."

113

TABLE 3.2.6: NUMBER OF ETHNICITY FIELDS REPORTED BY APPLICANTS (COUNTS IN THOUSANDS)

Count

%

Number of ethnicities

1

12,861.7

94.4

2

745.6

5.5

3

10.9

0.1

4

0.3

0.0

5

0.1

0.0

Total

13,618.5

100.0

NOTE: Ethnicity of applicants only. The estimates are calculated from non-public raw data reported by financial institutions.

114

TABLE 3.3.1: RACE AND ETHNICITY DETERMINED BY VISUAL OBSERVATION OR SURNAME (COUNTS IN THOUSANDS)

Count

%

Applicant's race

Collected on the basis of visual observation or

surname

670.8

3.8

Not collected on the basis of visual observation

or surname

14,166.5

80.7

Not applicable or missing

2,708.1

15.4

Total

17,545.5

100.0

Co-applicant's race

Collected on the basis of visual observation or

surname

292.7

1.7

Not collected on the basis of visual observation

or surname

6,267.3

35.7

No co-applicant

9,158.8

52.2

Not applicable or missing

1,826.8

10.4

Total

17,545.5

100.0

Applicant's ethnicity

Collected on the basis of visual observation or

surname

664.2

3.8

Not collected on the basis of visual observation

or surname

14,145.3

80.6

Not applicable or missing

2,735.9

15.6

Total

17,545.5

100.0

Co-applicant's ethnicity

115

Count

%

Collected on the basis of visual observation or

surname

291.1

1.7

Not collected on the basis of visual observation

or surname

6,254.7

35.6

No co-applicant

9,104.2

51.9

Not applicable or missing

1,895.5

10.8

Total

17,545.5

100.0

NOTE: The estimates are calculated from non-public raw data reported by financial institutions. Code 3 (Information not provided by applicant in mail, internet,

or telephone application), code 4 (Not applicable), and missing observations are combined into one category under "Not available or missing."

116

TABLE 3.3.2: RACE OF APPLICANTS DETERMINED BY VISUAL OBSERVATION OR SURNAME BY THE FIRST REPORTED RACE (COUNTS IN

THOUSANDS)

Collected on the basis of visual observation or surname

Yes

No

Not applicable or

Missing

Total

Count

%

Count

%

Count

%

Count

%

First reported race

American Indian or Alaska Native

4.4

3.5

114.7

90.8

7.2

5.7

126.2

100.0

Asian

34.2

3.6

859.0

90.2

59.1

6.2

952.4

100.0

Black or African American

55.6

4.7

1,032.2

88.0

84.8

7.2

1,172.6

100.0

Native Hawaiian or Other Pacific

Islander

2.9

5.4

47.6

88.6

3.2

5.9

53.8

100.0

White

573.6

5.1

9,964.2

88.4

733.9

6.5

11,271.6

100.0

Not Available or Missing

0.0

2,148.9

54.1

1,820.0

45.9

3,968.9

100.0

Total

670.8

3.8

14,166.5

80.7

2,708.1

15.4

17,545.5

100.0

NOTE: Race of applicants only. The estimates are calculated from non-public raw data reported by financial institutions. Code 3 (Information not provided by

applicant in mail, internet, or telephone application), code 4 (Not applicable), and missing observations are combined into one category under "Not available or

missing."

117

TABLE 3.3.3: ETHNICITY OF APPLICANTS DETERMINED BY VISUAL OBSERVATION OR SURNAME BY THE FIRST REPORTED ETHNICITY

(COUNTS IN THOUSANDS)

Collected on the basis of visual observation or

surname

Total

Yes

No

Not applicable or

missing

Count

%

Count

%

Count

%

Count

%

Hispanic or Latino

76.4

4.5

1,505.9

88.7

114.8

6.8

1,697.1

100.0

Not Hispanic or Latino

587.9

4.9

10,544.2

88.4

789.3

6.6

11,921.4

100.0

Not available or missing

0.0

2,095.2

53.4

1,831.7

46.6

3,926.9

100.0

Total

664.2

3.8

14,145.3

80.6

2,735.9

15.6

17,545.5

100.0

NOTE: Ethnicity of applicants only. The estimates are calculated from non-public raw data reported by financial institutions. Code 3 (Information not provided by

applicant in mail, internet, or telephone application), code 4 (Not applicable), and missing observations are combined into one category under "Not available or

missing."

118

TABLE 3.3.4: DISTRIBUTION OF SEX OF APPLICANTS AND CO-APPLICANTS (COUNTS IN THOUSANDS)

Count

%

Applicant's sex

Male

9,831.1

56.0

Female

4,847.9

27.6

Both male & female

8.7

0.0

Not available or Missing

2,857.7

16.3

Total

17,545.5

100.0

Co-applicant's sex

Male

1,697.6

9.7

Female

4,733.7

27.0

No co-applicant

9,156.5

52.2

Both male & female

6.0

0.0

Not available or Missing

1,951.7

11.1

Total

17,545.5

100.0

NOTE: The estimates are calculated from non-public raw data reported by financial institutions. Code 3 (Information not provided by applicant in mail, internet,

or telephone application), code 4 (Not applicable), and missing observations are combined into one category under "Not available or missing."

119

TABLE 3.3.5: SEX OF APPLICANTS DETERMINED BY VISUAL OBSERVATION OR SURNAME (COUNTS IN THOUSANDS)

Collected on the basis of visual observation or

surname

Total

Yes

No

Not applicable or

missing

Count

%

Count

%

Count

%

Count

%

Male

460.8

4.7

8,705.1

88.5

665.3

6.8

9,831.1

100.0

Female

225.5

4.7

4,296.8

88.6

325.6

6.7

4,847.9

100.0

Both male & female

0.0

8.3

95.0

0.4

5.0

8.7

100.0

Not available or missing

0.0

1,120.1

39.2

1,737.6

60.8

2,857.7

100.0

Total

686.2

3.9

14,130.3

80.5

2,728.9

15.6

17,545.5

100.0

NOTE: Sex of applicants only. The estimates are calculated from non-public raw data reported by financial institutions. Code 3 (Information not provided by

applicant in mail, internet, or telephone application), code 4 (Not applicable), and missing observations are combined into one category under "Not available or

missing."

120

TABLE 4.1: ACTION TYPE BY PROPERTY TYPE (COUNTS IN THOUSANDS)

Property type

Site-built

Single-family

Site-built

Multifamily

Manufactured

home

Total

Action type

Originated

Count

9,093.6

53.4

178.2

9,325.2

%

53.8

79.7

30.6

53.1

Approved, not Accepted

Count

366.9

1.6

33.9

402.5

%

2.2

2.5

5.8

2.3

Denied

Count

2,320.9

5.3

212.8

2,539.0

%

13.7

7.9

36.6

14.5

Withdrawn

Count

2,096.2

5.5

42.7

2,144.4

%

12.4

8.2

7.3

12.2

Closed for Incompleteness

Count

635.5

0.4

82.1

718.1

%

3.8

0.7

14.1

4.1

Purchased

121

Property type

Site-built

Single-family

Site-built

Multifamily

Manufactured

home

Total

Count

2,235.4

0.8

29.6

2,265.7

%

13.2

1.1

5.1

12.9

Preapproval Request Denied

Count

75.5

0.0

1.5

77.0

%

0.4

0.0

0.3

0.4

Preapproval Approved not

Accepted

Count

72.7

0.0

0.9

73.6

%

0.4

0.0

0.2

0.4

Total

Count

16,896.7

67.1

581.7

17,545.5

%

100.0

122

TABLE 4.2: SITE-BUILT SINGLE FAMILY NUMBER OF UNITS BY ACTION TYPE (COUNTS IN THOUSANDS)

Number of units

1

2

3

4

Total

Action type

Originated

Count

8,885.3

148.1

34.2

25.9

9,093.6

%

97.7

1.6

0.4

0.3

100.0

Approved, not Accepted

Count

355.7

7.8

1.9

1.5

366.9

%

97.0

2.1

0.5

0.4

100.0

Denied

Count

2,244.3

54.5

13.3

8.9

2,320.9

%

96.7

2.3

0.6

0.4

100.0

Withdrawn

Count

2,041.4

38.1

9.3

7.3

2,096.2

%

97.4

1.8

0.4

0.3

100.0

Closed for Incompleteness

Count

618.2

12.3

3.1

1.8

635.5

%

97.3

1.9

0.5

0.3

100.0

Purchased

Count

2,189.4

33.8

7.3

5.0

2,235.4

123

Number of units

1

2

3

4

Total

%

97.9

1.5

0.3

0.2

100.0

Preapproval Request Denied

Count

74.2

1.0

0.2

0.1

75.5

%

98.3

1.4

0.2

100.0

Preapproval Approved not

Accepted

Count

71.3

1.0

0.2

72.7

%

98.1

1.4

0.2

0.3

100.0

Total

Count

16,479.9

296.7

69.4

50.7

16,896.7

%

97.5

1.8

0.4

0.3

100.0

NOTE: Site-built single-family.

124

TABLE 4.3: MANUFACTURED HOME NUMBER OF UNITS BY ACTION TYPE

Number of units

1

2

3

4

5-24

25-49

50-99

100-149

>= 150

Total

Action type

Originated

Count

175,009

1,125

239

118

697

309

324

144

230

178,195

%

98.2

0.6

0.1

0.4

0.2

0.1

100.0

Approved, not Accepted

Count

33,778

85

11

9

23

18

9

2

3

33,938

%

99.5

0.3

0.0

0.1

0.0

100.0

Denied

Count

212,343

218

30

11

69

37

28

9

8

212,753

%

99.8

0.1

0.0

100.0

Withdrawn

Count

42,504

90

19

4

37

18

14

9

14

42,709

%

99.5

0.2

0.0

0.1

0.0

100.0

Closed for Incompleteness

Count

82,099

36

5

3

2

0

1

0

82,147

%

99.9

0.0

100.0

Purchased

Count

29,536

8

3

0

1

0

6

29,557

125

Number of units

1

2

3

4

5-24

25-49

50-99

100-149

>= 150

Total

%

99.9

0.0

100.0

Preapproval Request Denied

Count

1,515

2

0

1

1,518

%

99.8

0.1

0.0

0.1

100.0

Preapproval Approved not

Accepted

Count

929

2

0

931

%

99.8

0.2

0.0

100.0

Total

Count

577,713

1,566

307

148

828

382

377

165

262

581,748

%

99.3

0.3

0.1

0.0

0.1

0.0

100.0

NOTE: Manufactured homes.

126

TABLE 4.4: SITE-BUILT MULTIFAMILY NUMBER OF UNITS BY ACTION TYPE

Number of units

5-24

25-49

50-99

100-149

>= 150

Total

Action type

Originated

Count

34,799

6,806

4,941

2,186

4,687

53,419

%

65.1

12.7

9.2

4.1

8.8

100.0

Approved, not Accepted

Count

1,218

202

113

49

67

1,649

%

73.9

12.2

6.9

3.0

4.1

100.0

Denied

Count

3,954

558

353

147

269

5,281

%

74.9

10.6

6.7

2.8

5.1

100.0

Withdrawn

Count

3,133

654

525

314

858

5,484

%

57.1

11.9

9.6

5.7

15.6

100.0

Closed for Incompleteness

Count

302

59

36

16

23

436

%

69.3

13.5

8.3

3.7

5.3

100.0

Purchased

Count

484

97

80

32

77

770

127

Number of units

5-24

25-49

50-99

100-149

>= 150

Total

%

62.9

12.6

10.4

4.2

10.0

100.0

Preapproval Request Denied

Count

8

2

0

10

%

80.0

20.0

0.0

100.0

Preapproval Approved not

Accepted

Count

4

1

0

6

%

66.7

16.7

0.0

100.0

Total

Count

43,902

8,379

6,049

2,744

5,981

67,055

%

65.5

12.5

9.0

4.1

8.9

100.0

NOTE: Site-built mutlifamily homes.

128

TABLE 5.1.1: BUSINESS OR COMMERCIAL PURPOSE FLAG BY ACTION TYPE (COUNTS IN THOUSANDS)

Primarily for business or commercial purpose

Yes

No

Exempt

Total

Count

%

Count

%

Count

%

Count

%

Action type

Originated

329.5

3.5

8,660.6

92.9

335.2

3.6

9,325.2

100.0

Approved, not Accepted

12.9

3.2

373.2

92.7

16.3

4.1

402.5

100.0

Denied

69.5

2.7

2,421.9

95.4

47.6

1.9

2,539.0

100.0

Withdrawn

63.7

3.0

2,043.0

95.3

37.7

1.8

2,144.4

100.0

Closed for Incompleteness

16.6

2.3

695.5

96.9

6.0

0.8

718.1

100.0

Purchased

30.0

1.3

2,217.1

97.9

18.6

0.8

2,265.7

100.0

Preapproval Request Denied

0.8

1.0

73.4

95.3

2.8

3.7

77.0

100.0

Preapproval Approved not

Accepted

1.4

1.8

68.9

93.5

3.4

4.6

73.6

100.0

Total

524.3

3.0

16,553.6

94.3

467.6

2.7

17,545.5

100.0

NOTE: All LARs.

129

TABLE 5.1.2: BUSINESS OR COMMERCIAL PURPOSE FLAG BY PROPERTY TYPE (COUNTS IN THOUSANDS)

Primarily for business or commercial purpose

Yes

No

Exempt

Total

Count

%

Count

%

Count

%

Count

%

Property type

Site-built Single-family

283.9

3.1

8,504.1

93.5

305.6

3.4

9,093.6

100.0

Site-built Multifamily

40.5

75.8

0.2

0.4

12.7

23.8

53.4

100.0

Manufactured Home

5.0

2.8

156.3

87.7

16.9

9.5

178.2

100.0

Total

329.5

3.5

8,660.6

92.9

335.2

3.6

9,325.2

100.0

NOTE: Originations.

130

TABLE 5.1.3: BUSINESS OR COMMERCIAL PURPOSE FLAG BY ENHANCED LOAN TYPE (COUNTS IN THOUSANDS)

Primarily for business or commercial purpose

Yes

No

Exempt

Total

Count

%

Count

%

Count

%

Count

%

Enhanced loan type

Conventional

Conforming

253.5

4.8

5,066.7

95.2

0.3

0.0

5,320.5

100.0

Jumbo

13.3

3.6

360.7

96.4

0.0

373.9

100.0

Non-conventional

FHA

1.6

0.1

1,105.3

99.9

0.0

1,106.8

100.0

VA

1.7

0.2

814.0

99.8

0.0

815.6

100.0

RHS/FSA

0.0

98.9

100.0

0.0

98.9

100.0

HELOC

13.1

1.3

1,022.8

98.7

0.4

0.0

1,036.3

100.0

Reverse Mortgage

0.0

34.1

100.0

0.0

34.1

100.0

Total

283.2

3.2

8,502.4

96.8

0.7

0.0

8,786.3

100.0

NOTE: Site-built single-family originations.

131

TABLE 5.1.4: BUSINESS OR COMMERCIAL PURPOSE FLAG BY LOAN PURPOSE (COUNTS IN THOUSANDS)

Primarily for business or commercial purposes

Yes

No

Exempt

Total

Count

Row %

Col %

Count

Row %

Col %

Count

Row %

Col %

Count

Row %

Col %

Loan purpose

Home purchase

159.6

3.7

56.2

4,053.3

93.0

47.7

147.3

3.4

48.2

4,360.3

100.0

47.9

Home improvement

13.5

2.5

4.8

499.7

93.5

5.9

21.2

4.0

6.9

534.5

100.0

5.9

Other

2.5

0.5

0.9

466.4

96.2

5.5

15.7

3.2

5.1

484.6

100.0

5.3

NA

1.4

20.8

0.5

5.3

76.5

0.1

0.2

2.6

0.1

6.9

100.0

0.1

Non-cash-out refi

61.6

2.8

21.7

2,090.0

93.7

24.6

79.8

3.6

26.1

2,231.3

100.0

24.5

Cash-out refi

45.2

3.1

15.9

1,389.4

94.1

16.3

41.4

2.8

13.5

1,476.0

100.0

16.2

Total

283.9

3.1

100.0

8,504.1

93.5

100.0

305.6

3.4

100.0

9,093.6

100.0

NOTE: Site-built single-family originations.

132

TABLE 5.1.5: BUSINESS OR COMMERCIAL PURPOSE FLAG BY OCCUPANCY STATUS (COUNTS IN THOUSANDS)

Primarily for business or commercial purposes

Yes

No

Exempt

Total

Count

Row %

Col %

Count

Row %

Col %

Count

Row %

Col %

Count

Row %

Col %

Occupancy status

Principal Residence

11.8

0.1

4.1

8,055.0

97.2

94.7

222.8

2.7

72.9

8,289.5

100.0

91.2

Second Residence

1.1

0.4

246.4

95.5

2.9

10.4

4.0

3.4

258.0

100.0

2.8

Investment Property

271.1

49.6

95.5

202.7

37.1

2.4

72.4

13.3

23.7

546.2

100.0

6.0

Total

283.9

3.1

100.0

8,504.1

93.5

100.0

305.6

3.4

100.0

9,093.6

100.0

NOTE: Site-built single-family originations.

133

TABLE 5.1.6: BUSINESS OR COMMERCIAL PURPOSE FLAG BY RACE AND ETHNICITY (COUNTS IN THOUSANDS)

Primarily for business or commercial purposes

Yes

No

Exempt

Total

Count

Row %

Col %

Count

Row %

Col %

Count

Row %

Col %

Count

Row %

Col %

Race and ethnicity

Asian

27.8

5.4

9.8

475.7

93.2

5.6

7.1

1.4

2.3

510.6

100.0

5.6

Black

10.2

2.0

3.6

493.1

96.5

5.8

7.8

1.5

2.6

511.1

100.0

5.6

Hispanic white

13.6

2.1

4.8

626.0

96.3

7.4

10.3

1.6

3.4

650.0

100.0

7.1

Joint

5.3

1.7

1.8

293.8

96.2

3.5

6.3

2.1

305.4

100.0

3.4

Non-Hispanic white

107.4

1.9

37.8

5,261.1

94.2

61.9

218.1

3.9

71.4

5,586.5

100.0

61.4

Other

1.4

2.0

0.5

67.9

96.2

0.8

1.3

1.8

0.4

70.5

100.0

0.8

Missing

118.3

8.1

41.7

1,286.5

88.1

15.1

54.7

3.7

17.9

1,459.5

100.0

16.0

Total

283.9

3.1

100.0

8,504.1

93.5

100.0

305.6

3.4

100.0

9,093.6

100.0

NOTE: Site-built single-family originations.

134

TABLE 5.2.1: ACTION TYPE BY LOAN PURPOSE: ALL LARS, SINGLE-FAMILY (COUNTS IN THOUSANDS)

Loan purpose

Home

purchase

Home

improve-

ment

Non-cash-out

refi

Cash-out refi

Other

NA

Total

Action type

Originated

Count

4,360.3

534.5

2,231.3

1,476.0

484.6

6.9

9,093.6

Col %

58.5

47.0

54.8

50.0

44.6

3.6

53.8

Row %

47.9

5.9

24.5

16.2

5.3

0.1

100.0

Approved, not Accepted

Count

144.6

31.3

101.7

60.4

28.4

0.4

366.9

Col %

1.9

2.8

2.5

2.0

2.6

0.2

2.2

Row %

39.4

8.5

27.7

16.5

7.7

0.1

100.0

Denied

Count

462.6

431.9

515.0

481.0

427.8

2.8

2,320.9

Col %

6.2

38.0

12.6

16.3

39.4

1.5

13.7

Row %

19.9

18.6

22.2

20.7

18.4

0.1

100.0

Withdrawn

Count

835.8

91.6

592.9

479.6

94.0

2.3

2,096.2

Col %

11.2

8.1

14.6

16.2

8.7

1.2

12.4

Row %

39.9

4.4

28.3

22.9

4.5

0.1

100.0

Closed for Incompleteness

135

Loan purpose

Home

purchase

Home

improve-

ment

Non-cash-out

refi

Cash-out refi

Other

NA

Total

Count

133.9

40.0

244.3

173.0

40.4

3.9

635.5

Col %

1.8

3.5

6.0

5.9

3.7

2.1

3.8

Row %

21.1

6.3

38.4

27.2

6.4

0.6

100.0

Purchased

Count

1,373.4

7.5

388.3

281.8

10.6

173.8

2,235.4

Col %

18.4

0.7

9.5

1.0

91.4

13.2

Row %

61.4

0.3

17.4

12.6

0.5

7.8

100.0

Preapproval Request Denied

Count

75.5

0.0

75.5

Col %

1.0

0.0

0.4

Row %

100.0

0.0

100.0

Preapproval Approved not

Accepted

Count

72.7

0.0

72.7

Col %

1.0

0.0

0.4

Row %

100.0

0.0

100.0

Total

Count

7,458.7

1,136.7

4,073.6

2,951.7

1,085.7

190.1

16,896.7

Col %

100.0

Row %

44.1

6.7

24.1

17.5

6.4

1.1

100.0

136

NOTE: Site-built single-family homes.

137

TABLE 5.2.2: LOAN PURPOSE BY ENHANCED LOAN TYPE

Loan purpose

Home

purchase

Home

improvement

Non-cash-out

refi

Cash-out refi

Other

NA

Total

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Enhanced loan type

Conventional

Conforming

2,715.1

51.0

137.7

2.6

1,328.3

25.0

1,003.5

18.9

132.0

2.5

4.0

0.1

5,320.5

100.0

Jumbo

200.9

53.7

2.1

0.6

114.4

30.6

48.9

13.1

7.4

2.0

0.2

0.1

373.9

100.0

Non-conventional

FHA

750.2

67.8

7.0

0.6

188.3

17.0

155.8

14.1

4.2

0.4

1.3

0.1

1,106.8

100.0

VA

395.3

48.5

1.9

0.2

266.6

32.7

150.7

18.5

0.8

0.1

0.3

0.0

815.6

100.0

RHS/FSA

96.5

97.6

0.0

2.3

0.1

0.0

98.9

100.0

Total

4,158.1

53.9

148.8

1.9

1,899.9

24.6

1,358.9

17.6

144.4

1.9

5.8

0.1

7,715.9

100.0

NOTE: Site-built single-family closed-end originations.

138

TABLE 5.2.3: LOAN PURPOSE BY RACE/ETHNICITY, AGE, NEIGHBORHOOD INCOME, AND GEOGRAPHY: CLOSED-END ORIGINATIONS,

SINGLE-FAMILY (COUNTS IN THOUSANDS)

Loan purpose

Home

purchase

Home

improvement

Non-cash-out

refi

Cash-out refi

Other

NA

Total

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Borrower race and ethnicity

Asian

254.4

56.2

4.1

0.9

135.9

30.0

52.1

11.5

5.9

1.3

0.3

0.1

452.7

100.0

Black

280.7

60.1

7.0

1.5

100.5

21.5

72.6

15.5

6.2

1.3

0.4

0.1

467.3

100.0

Hispanic white

381.3

63.4

8.1

1.3

113.5

18.9

89.9

14.9

8.4

1.4

0.3

0.0

601.5

100.0

Joint

148.6

56.0

5.3

2.0

64.5

24.3

42.5

16.0

4.5

1.7

0.2

0.1

265.6

100.0

Non-Hispanic white

2,446.6

53.3

100.1

2.2

1,117.9

24.3

828.7

18.0

97.8

2.1

3.1

0.1

4,594.2

100.0

Other

32.7

54.4

1.1

1.9

13.5

22.5

11.6

19.2

1.2

2.0

0.0

0.1

60.2

100.0

Missing

613.8

48.2

22.9

1.8

354.1

27.8

261.6

20.5

1.6

1.5

0.1

1,274.5

100.0

Total

4,158.1

53.9

148.8

1.9

1,899.9

24.6

1,358.9

17.6

144.4

1.9

5.8

0.1

7,715.9

100.0

Age group

<=24

222.8

91.8

0.7

0.3

15.4

6.4

2.9

1.2

0.9

0.4

0.1

0.0

242.7

100.0

25-34

1,276.5

73.4

13.5

0.8

334.4

19.2

100.0

5.7

13.9

0.8

0.0

1,739.1

100.0

35-44

1,081.3

54.7

34.5

1.7

536.9

27.2

290.5

14.7

30.5

1.5

1.4

0.1

1,975.1

100.0

45-54

732.3

44.8

38.6

2.4

445.8

27.3

379.9

23.2

37.6

2.3

1.1

0.1

1,635.4

100.0

55-64

494.5

41.2

33.5

2.8

317.5

26.5

319.5

26.6

34.5

2.9

0.7

0.1

1,200.2

100.0

65-74

239.0

36.9

17.7

2.7

177.3

27.4

193.1

29.8

19.7

3.0

0.3

0.1

647.1

100.0

139

Loan purpose

Home

purchase

Home

improvement

Non-cash-out

refi

Cash-out refi

Other

NA

Total

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

>=75

58.5

31.2

5.1

2.7

53.4

28.4

64.1

34.2

6.5

3.4

0.1

0.0

187.8

100.0

Total

4,104.9

53.8

143.5

1.9

1,880.8

24.7

1,350.1

17.7

143.6

1.9

4.6

0.1

7,627.3

100.0

Neighborhood income

Low or moderate

730.6

57.7

24.3

1.9

262.3

20.7

224.7

17.8

22.5

1.8

1.1

0.1

1,265.4

100.0

Middle

1,829.8

54.6

67.5

2.0

784.5

23.4

600.4

17.9

65.7

2.0

2.1

0.1

3,350.1

100.0

High

1,578.4

51.5

54.6

1.8

845.3

27.6

528.7

17.3

55.1

1.8

2.1

0.1

3,064.2

100.0

Total

4,138.8

53.9

146.5

1.9

1,892.0

24.6

1,353.8

17.6

143.4

1.9

5.3

0.1

7,679.7

100.0

Geography

Metropolitan Area

3,741.3

53.5

130.3

1.9

1,751.1

25.0

1,240.2

17.7

126.7

1.8

4.9

0.1

6,994.4

100.0

Micropolitan Area

262.3

58.4

10.1

2.2

93.1

20.7

73.2

16.3

10.4

2.3

0.3

0.1

449.4

100.0

Rural

154.6

56.8

8.4

3.1

55.7

20.5

45.6

16.7

7.3

2.7

0.6

0.2

272.1

100.0

Total

4,158.1

53.9

148.8

1.9

1,899.9

24.6

1,358.9

17.6

144.4

1.9

5.8

0.1

7,715.9

100.0

NOTE: Site-built single-family closed-end originations.

140

TABLE 5.2.4: LOAN PURPOSE BY LIEN STATUS: CLOSED-END ORIGINATIONS, SINGLE-FAMILY (COUNTS IN THOUSANDS)

Lien status

First lien

Subordinate lien

Total

Count

Row %

Col %

Count

Row %

Col %

Count

Row %

Col %

Loan purpose

Home purchase

4,060.6

97.7

54.8

97.5

2.3

32.5

4,158.1

100.0

53.9

Home improvement

72.0

48.4

1.0

76.8

51.6

25.5

148.8

100.0

1.9

Non-cash-out refi

1,866.4

98.2

25.2

33.4

1.8

11.1

1,899.9

100.0

24.6

Cash-out refi

1,332.8

98.1

18.0

26.2

1.9

8.7

1,358.9

100.0

17.6

Other

78.0

54.0

1.1

66.3

46.0

22.1

144.4

100.0

1.9

NA

5.6

97.3

0.1

0.2

2.7

0.1

5.8

100.0

0.1

Total

7,415.4

96.1

100.0

300.4

3.9

100.0

7,715.9

100.0

NOTE: Site-built single-family closed-end originations.

141

TABLE 5.2.5: MEDIAN LOAN AMOUNT: LOAN PURPOSE BY ENHANCED LOAN TYPE, CLOSED-END ORIGINATIONS, SINGLE-FAMILY (DOLLARS

IN THOUSANDS)

Enhanced loan type

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Loan purpose

Home purchase

226.1

788.0

205.7

258.3

146.4

230.0

Home improvement

50.5

750.0

116.8

257.7

146.5

55.0

Non-cash-out refi

254.0

861.0

235.6

276.6

173.7

266.6

Cash-out refi

210.0

823.5

197.2

253.2

207.1

220.0

Other

60.0

923.8

92.1

260.7

25.0

65.0

NA

276.2

798.3

255.0

306.7

331.8

279.0

Total

223.3

815.0

208.9

263.4

146.5

232.1

NOTE: Site-built single-family closed-end originations. The median loan amounts in the table are calculated from non-public raw data reported by financial

institutions. The outliers are excluded from the analysis sample to produce consistent estimates.

142

TABLE 5.3.1: TOP 20 MOST COMMON LOAN TERMS OF CLOSED-END ORIGINATIONS

Count

%

Loan term

(months)

360

6,205,591

80.5

180

724,840

9.4

240

259,927

3.4

120

163,742

2.1

60

58,836

0.8

300

45,704

0.6

12

30,102

0.4

84

16,052

0.2

372

15,733

0.2

144

13,067

0.2

348

12,834

0.2

324

10,928

0.1

36

9,228

0.1

336

7,104

0.1

72

6,994

0.1

6

5,057

0.1

96

3,941

0.1

216

3,797

0.0

143

Count

%

9

3,684

0.0

369

3,561

0.0

Total

7,704,355

100.0

NOTE: Site-built single-family closed-end originations.

144

TABLE 5.3.2: TOP 20 MOST COMMON LOAN TERMS OF HELOC ORIGINATIONS

Count

%

Loan term

(months)

360

471,941

46.3

300

189,437

18.6

361

81,599

8.0

240

68,640

6.7

120

59,243

5.8

480

43,761

4.3

180

38,634

3.8

60

22,212

2.2

444

5,170

0.5

264

3,880

0.4

12

3,857

0.4

144

3,125

0.3

355

2,916

0.3

420

2,189

0.2

156

1,778

0.2

59

1,777

0.2

204

1,726

0.2

168

1,606

0.2

145

Count

%

35

1,321

0.1

84

1,243

0.1

Total

1,018,628

100.0

NOTE: Site-built single-family HELOC originations.

146

TABLE 5.3.3: COMMON LOAN TERMS BY LOAN PURPOSE, RACE/ETHNICITY, AGE, NEIGHBORHOOD INCOME, AND GEOGRAPHY: CLOSED-END

ORIGINATIONS (COUNTS IN THOUSANDS)

Loan term

5 years

10 years

15 years

20 years

30 years

Other

Total

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Loan purpose

Home purchase

17.1

0.4

34.7

0.8

179.5

4.3

46.3

1.1

3,785.6

91.1

90.2

2.2

4,153.4

100.0

Home improvement

13.0

8.8

28.9

19.6

32.0

21.6

16.4

11.1

37.1

25.1

20.3

13.7

147.7

100.0

Other

10.8

7.6

26.6

18.6

30.6

21.4

14.7

10.2

45.6

31.9

14.9

10.4

143.2

100.0

NA

0.1

1.3

0.1

1.3

0.4

6.8

0.2

3.3

3.8

67.2

1.1

20.1

5.7

100.0

Non-cash-out refi

12.9

0.7

40.7

2.1

271.6

14.3

105.8

5.6

1,336.8

70.5

129.3

6.8

1,897.2

100.0

Cash-out refi

4.9

0.4

32.7

2.4

210.8

15.5

76.5

5.6

996.6

73.4

35.7

2.6

1,357.2

100.0

Total

58.8

0.8

163.7

2.1

724.8

9.4

259.9

3.4

6,205.6

80.5

291.4

3.8

7,704.4

100.0

Borrower race and

ethnicity

Asian

1.3

0.3

5.2

1.1

60.3

13.3

12.9

2.9

364.9

80.7

7.8

1.7

452.4

100.0

Black

3.2

0.7

8.9

1.9

26.7

5.7

10.7

2.3

400.5

85.8

16.4

3.5

466.5

100.0

Hispanic white

2.5

0.4

9.7

1.6

38.3

6.4

16.5

2.7

519.9

86.6

13.7

2.3

600.6

100.0

Joint

1.3

0.5

5.3

2.0

23.2

8.7

8.1

3.1

219.7

82.8

7.7

2.9

265.3

100.0

Non-Hispanic white

33.4

0.7

110.2

2.4

452.2

9.9

165.3

3.6

3,668.0

80.0

158.4

3.5

4,587.6

100.0

Other

0.3

0.5

1.1

1.8

4.3

7.1

1.8

3.0

50.6

84.2

2.0

3.4

60.1

100.0

Missing

16.8

1.3

23.4

1.8

119.8

9.4

44.6

3.5

981.9

77.2

85.3

6.7

1,271.9

100.0

147

Loan term

5 years

10 years

15 years

20 years

30 years

Other

Total

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Total

58.8

0.8

163.7

2.1

724.8

9.4

259.9

3.4

6,205.6

80.5

291.4

3.8

7,704.4

100.0

Age group

<=24

0.8

0.3

3.4

1.4

6.6

2.7

2.2

0.9

226.0

93.2

3.4

1.4

242.5

100.0

25-34

5.3

0.3

17.1

1.0

78.0

4.5

30.9

1.8

1,567.8

90.2

38.4

2.2

1,737.4

100.0

35-44

8.8

0.4

29.2

1.5

162.5

8.2

69.1

3.5

1,640.8

83.2

62.5

3.2

1,972.9

100.0

45-54

11.4

0.7

40.9

2.5

206.9

12.7

74.0

4.5

1,239.7

75.9

60.3

3.7

1,633.1

100.0

55-64

12.0

1.0

41.4

3.5

171.4

14.3

50.2

4.2

873.9

72.9

49.2

4.1

1,198.1

100.0

65-74

6.2

1.0

20.1

3.1

75.7

11.7

22.5

3.5

495.4

76.7

26.2

4.1

646.1

100.0

>=75

1.9

1.0

5.7

3.0

17.8

9.5

6.0

3.2

148.6

79.2

7.5

4.0

187.5

100.0

Total

46.4

0.6

157.8

2.1

718.8

9.4

254.9

3.3

6,192.1

81.3

247.5

3.2

7,617.6

100.0

Neighborhood income

Low or moderate

13.2

1.0

28.7

2.3

101.0

8.0

37.8

3.0

1,032.7

81.8

49.5

3.9

1,263.0

100.0

Middle

28.0

0.8

79.9

2.4

309.2

9.2

116.8

3.5

2,680.7

80.1

130.7

3.9

3,345.3

100.0

High

17.1

0.6

54.3

1.8

312.5

10.2

104.5

3.4

2,472.0

80.8

100.6

3.3

3,061.0

100.0

Total

58.4

0.8

162.9

2.1

722.7

9.4

259.2

3.4

6,185.3

80.7

280.8

3.7

7,669.3

100.0

Geography

Metropolitan Area

49.4

0.7

141.2

2.0

643.7

9.2

230.2

3.3

5,675.5

81.2

245.5

3.5

6,985.6

100.0

Micropolitan Area

5.6

1.3

13.4

3.0

49.6

11.1

18.2

4.1

339.3

75.7

22.0

4.9

448.2

100.0

148

Loan term

5 years

10 years

15 years

20 years

30 years

Other

Total

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Rural

3.7

1.4

9.1

3.4

31.5

11.7

11.5

4.2

190.7

70.5

23.9

8.8

270.5

100.0

Total

58.8

0.8

163.7

2.1

724.8

9.4

259.9

3.4

6,205.6

80.5

291.4

3.8

7,704.4

100.0

NOTE: Site-built single-family closed-end originations.

149

TABLE 5.3.4: COMMON LOAN TERMS BY RACE/ETHNICITY, AGE, NEIGHBORHOOD INCOME, AND GEOGRAPHY: HELOC ORIGINATIONS

(COUNTS IN THOUSANDS)

Loan term

5 years

10 years

15 years

20 years

25 years

30 years

40 years

Other

Total

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Cou

nt

%

Count

%

Borrower race and

ethnicity

Asian

0.4

0.8

0.9

1.7

1.3

2.6

1.5

3.1

6.8

13.5

36.0

72.1

2.0

4.1

1.0

2.1

49.9

100.0

Black

0.5

1.7

1.5

4.6

1.9

5.8

2.0

6.1

5.6

17.4

16.4

51.4

3.0

9.6

1.1

3.4

31.9

100.0

Hispanic white

0.7

1.9

1.3

3.5

1.3

3.6

2.8

7.8

5.9

16.1

22.2

61.1

1.3

3.6

0.9

2.4

36.4

100.0

Joint

0.9

2.8

1.7

5.3

1.2

3.7

2.5

7.6

5.8

17.7

18.0

55.4

1.4

4.4

1.0

3.2

32.6

100.0

Non-Hispanic white

16.9

2.3

47.5

6.5

27.9

3.8

52.2

7.1

137.1

18.7

389.9

53.1

31.5

4.3

31.9

4.3

734.9

100.0

Other

0.2

2.0

0.2

2.5

0.2

2.5

0.4

4.7

1.3

15.1

5.7

65.5

0.4

4.5

0.3

3.1

8.8

100.0

Missing

2.6

2.1

6.2

5.0

4.8

3.9

7.3

5.9

27.1

21.8

65.3

52.6

4.0

3.2

6.9

5.5

124.2

100.0

Total

22.2

2.2

59.2

5.8

38.6

3.8

68.6

6.7

189.4

18.6

553.5

54.3

43.8

4.3

43.2

4.2

1,018.6

100.0

Age group

<=24

0.1

4.6

0.2

12.2

0.1

6.2

0.1

7.6

0.3

16.7

0.8

43.3

0.1

3.9

0.1

5.5

1.8

100.0

25-34

2.5

3.7

5.2

7.6

3.2

4.6

6.0

8.8

12.4

18.1

33.0

48.1

3.2

4.6

3.0

4.4

68.5

100.0

35-44

5.6

2.9

12.4

6.4

8.2

4.2

15.5

8.0

33.8

17.4

102.5

52.7

8.6

4.4

7.9

4.1

194.7

100.0

45-54

5.4

2.1

14.8

5.7

10.2

3.9

18.6

7.1

46.7

18.0

142.6

54.9

11.3

4.4

10.1

3.9

259.8

100.0

55-64

4.8

1.9

14.5

5.6

9.7

3.7

16.6

6.4

49.6

19.2

142.3

55.0

10.7

4.1

10.5

4.1

258.7

100.0

65-74

2.7

1.6

8.7

5.2

5.3

3.2

8.7

5.2

32.9

19.7

94.9

56.7

7.1

4.3

7.0

4.2

167.4

100.0

150

Loan term

5 years

10 years

15 years

20 years

25 years

30 years

40 years

Other

Total

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Cou

nt

%

Count

%

>=75

0.9

1.4

3.1

4.9

1.9

2.9

3.0

4.6

13.1

20.2

37.1

57.5

2.8

4.3

2.7

4.2

64.5

100.0

Total

22.0

2.2

58.9

5.8

38.6

3.8

68.6

6.8

188.8

18.6

553.4

54.5

43.8

4.3

41.3

4.1

1,015.4

100.0

Neighborhood income

Low or moderate

2.4

2.0

6.5

5.5

4.9

4.1

8.8

7.4

24.3

20.6

61.0

51.7

5.1

4.3

5.0

4.2

118.0

100.0

Middle

10.4

2.4

27.7

6.5

17.2

4.1

30.5

7.2

84.7

20.0

217.1

51.3

18.1

4.3

17.8

4.2

423.5

100.0

High

9.4

2.0

24.9

5.2

16.4

3.4

29.2

6.1

79.9

16.8

275.0

57.8

20.5

4.3

20.3

4.3

475.5

100.0

Total

22.2

2.2

59.2

5.8

38.4

3.8

68.5

6.7

188.9

18.6

553.1

54.4

43.7

4.3

43.0

4.2

1,017.0

100.0

Geography

Metropolitan Area

19.3

2.1

47.8

5.1

33.1

3.6

64.0

6.9

171.0

18.4

513.6

55.3

41.5

4.5

37.8

4.1

928.1

100.0

Micropolitan Area

1.9

3.2

7.7

13.0

3.4

5.7

2.9

4.9

11.7

19.6

27.2

45.7

1.7

2.9

3.0

5.0

59.5

100.0

Rural

1.0

3.2

3.7

12.0

2.1

6.8

1.7

5.3

6.8

21.9

12.7

41.0

0.6

1.9

2.4

7.8

31.0

100.0

Total

22.2

2.2

59.2

5.8

38.6

3.8

68.6

6.7

189.4

18.6

553.5

54.3

43.8

4.3

43.2

4.2

1,018.6

100.0

NOTE: Site-built single-family HELOC originations.

151

TABLE 5.3.5: COMMON LOAN TERMS BY ENHANCED LOAN TYPE: CLOSED-END AND HELOC ORIGINATIONS (COUNTS IN THOUSANDS)

Loan term

5 years

10 years

15 years

20 years

30 years

Other

Total

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Enhanced loan type

Conventional

Conforming

57.8

1.1

161.3

3.0

680.6

12.8

244.5

4.6

3,956.0

74.5

211.1

4.0

5,311.2

100.0

Jumbo

0.8

0.2

1.5

0.4

16.2

4.3

3.5

0.9

334.4

89.5

17.3

4.6

373.7

100.0

Non-conventional

FHA

0.1

0.0

0.7

0.1

14.6

1.3

6.5

0.6

1,055.8

95.5

28.2

2.6

1,105.9

100.0

VA

0.1

0.0

0.3

0.0

13.4

1.6

5.4

0.7

761.2

93.4

34.5

4.2

814.9

100.0

RHS/FSA

0.0

98.2

99.6

0.4

98.6

100.0

HELOC

22.2

2.2

59.2

5.8

38.6

3.8

68.6

6.7

471.9

46.3

358.0

35.1

1,018.6

100.0

Total

81.0

0.9

223.0

2.6

763.5

8.8

328.6

3.8

6,677.5

76.6

649.4

7.4

8,723.0

100.0

NOTE: Site-built single-family closed-end and HELOC originations.

152

TABLE 5.4.1: FIXED RATE VS ARM BY ENHANCED LOAN TYPE: ORIGINATION (COUNTS IN THOUSANDS)

ARM or Fixed Rate

ARM

Fixed rate

Total

Count

%

Count

%

Count

%

Enhanced loan type

Conventional

Conforming

294.5

5.5

5,020.6

94.5

5,315.1

100.0

Jumbo

155.9

41.7

217.9

58.3

373.8

100.0

Non-conventional

FHA

2.6

0.2

1,104.2

99.8

1,106.8

100.0

VA

1.2

0.2

814.4

99.8

815.6

100.0

RHS/FSA

0.0

98.9

100.0

98.9

100.0

HELOC

835.2

80.6

200.9

19.4

1,036.2

100.0

Reverse Mortgage

17.8

52.0

16.4

48.0

34.1

100.0

Total

1,307.2

14.9

7,473.3

85.1

8,780.5

100.0

NOTE: Site-built single-family originations.

153

TABLE 5.4.2: FIXED RATE VS. ARM BY RACE/ETHNICITY, AGE, NEIGHBORHOOD INCOME, AND GEOGRAPHY: CLOSED-END ORIGINATIONS

(COUNTS IN THOUSANDS)

ARM or Fixed Rate

ARM

Fixed rate

Total

Count

%

Count

%

Count

%

Borrower race and ethnicity

Asian

63.0

13.9

389.1

86.1

452.1

100.0

Black

12.1

2.6

455.2

97.4

467.2

100.0

Hispanic white

14.9

2.5

586.3

97.5

601.2

100.0

Joint

12.9

4.9

252.5

95.1

265.4

100.0

Non-Hispanic white

259.8

5.7

4,332.3

94.3

4,592.1

100.0

Other

1.7

2.8

58.4

97.2

60.1

100.0

Missing

89.9

7.1

1,182.3

92.9

1,272.2

100.0

Total

454.2

5.9

7,256.0

94.1

7,710.2

100.0

Age group

<=24

7.3

3.0

235.4

97.0

242.7

100.0

25-34

74.4

4.3

1,663.9

95.7

1,738.3

100.0

35-44

111.9

5.7

1,862.0

94.3

1,974.0

100.0

45-54

104.6

6.4

1,529.6

93.6

1,634.2

100.0

55-64

79.0

6.6

1,120.0

93.4

1,199.0

100.0

65-74

38.7

6.0

607.7

94.0

646.4

100.0

154

ARM or Fixed Rate

ARM

Fixed rate

Total

Count

%

Count

%

Count

%

>=75

12.7

6.8

174.9

93.2

187.6

100.0

Total

428.6

5.6

7,193.5

94.4

7,622.1

100.0

Neighborhood income

Low or moderate

53.1

4.2

1,211.2

95.8

1,264.3

100.0

Middle

149.6

4.5

3,198.4

95.5

3,348.0

100.0

High

248.6

8.1

2,813.3

91.9

3,061.9

100.0

Total

451.3

5.9

7,222.9

94.1

7,674.2

100.0

Geography

Metropolitan Area

408.0

5.8

6,581.2

94.2

6,989.1

100.0

Micropolitan Area

28.3

6.3

421.0

93.7

449.3

100.0

Rural

17.9

6.6

253.9

93.4

271.8

100.0

Total

454.2

5.9

7,256.0

94.1

7,710.2

100.0

NOTE: Site-built single-family closed-end originations.

155

TABLE 5.4.3: DENIAL RATE: FIXED RATE VS. ARM BY ENHANCED LOAN TYPE (PERCENT)

ARM or Fixed Rate

ARM

Fixed rate

Total

Enhanced loan type

Conventional

Conforming

12.8

15.1

14.9

Jumbo

12.8

16.3

14.9

Non-conventional

FHA

31.5

19.4

19.5

VA

21.1

16.1

HELOC

41.1

43.7

41.6

Reverse Mortgage

19.0

14.9

17.1

Total

33.4

17.0

20.0

NOTE: Site-built single-family homes, excluding RHS/FSA applications. The denial rates are calculated based on applications that were denied, divided by

(applications that were denied + applications that were approved but not accepted + loans originated). The denial rate calculations do not include applications

that were withdrawn or files that were closed for incompleteness.

156

TABLE 5.4.4: TOP 20 MOST COMMON INTRODUCTORY RATE PERIOD FOR CLOSED-END ORIGINATIONS, RANKED

Count

%

Introductory rate period

(months)

84

138,295

30.45

60

134,052

29.51

120

111,364

24.52

36

18,287

4.03

180

12,917

2.84

1

8,540

1.88

12

5,293

1.17

62

4,542

1.00

96

2,125

0.47

72

1,990

0.44

61

1,945

0.43

121

1,567

0.34

132

1,501

0.33

63

1,100

0.24

6

971

0.21

9

879

0.19

66

792

0.17

24

536

0.12

157

Count

%

85

519

0.11

48

469

0.10

Total

454,219

100.00

NOTE: Site-built single-family closed-end originations.

158

TABLE 5.4.5: COMMON INTRODUCTORY RATE PERIOD BY LOAN TERM: CLOSED-END ORIGINATIONS

Loan term

5 years

10 years

15 years

20 years

30 years

Other

Total

Introductory rate period

1 year

123

300

495

239

2,944

1,171

5,272

3 years

109

726

1,805

1,884

12,899

1,206

18,629

5 years

393

5,461

8,421

8,993

111,902

6,822

141,992

7 years

0

350

1,373

764

134,198

2,211

138,896

10 years

1

186

959

958

109,106

1,912

113,122

15 years

2

0

503

101

12,178

129

12,913

< 1 year

551

851

759

194

1,828

6,449

10,632

Other

88

266

363

464

3,431

7,861

12,473

Total

1,267

8,140

14,678

13,597

388,486

27,761

453,929

NOTE: Site-built single-family closed-end originations.

159

TABLE 5.4.6: TOP 20 MOST COMMON INTRODUCTORY RATE PERIOD FOR HELOC ORIGINATIONS, RANKED

Count

%

Introductory rate period

(months)

1

573,604

68.7

12

127,235

15.2

6

73,558

8.8

3

20,870

2.5

60

8,140

1.0

36

6,082

0.7

2

5,170

0.6

9

4,916

0.6

24

4,594

0.6

61

2,353

0.3

48

2,136

0.3

25

1,374

0.2

120

1,077

0.1

11

934

0.1

13

830

0.1

4

742

0.1

84

459

0.1

5

321

0.0

160

Count

%

7

238

0.0

8

224

0.0

Total

835,244

100.0

NOTE: Site-built single-family HELOC originations.

161

TABLE 5.5.1: NON-AMORTIZING FEATURES BY TRANSACTION TYPE: ORIGINATIONS (COUNTS IN THOUSANDS)

Transaction type

Closed-end

HELOC

Total

Count

%

Count

%

Count

%

Balloon payment

Yes

121.8

1.6

103.3

10.0

225.1

2.6

No

7,591.0

98.4

932.9

90.0

8,523.9

97.4

Exempt

3.1

0.0

0.1

0.0

3.1

0.0

Total

7,715.9

100.0

1,036.3

100.0

8,752.1

100.0

Interest-only payments

Yes

158.0

2.1

547.4

52.8

705.5

8.1

No

7,554.8

97.9

488.7

47.2

8,043.5

91.9

Exempt

3.1

0.0

0.1

0.0

3.1

0.0

Total

7,715.9

100.0

1,036.3

100.0

8,752.1

100.0

Negative amortization

Yes

1.1

0.0

2.3

0.2

3.5

0.0

No

7,711.7

100.0

1,033.8

99.8

8,745.5

99.9

Exempt

3.1

0.0

0.1

0.0

3.1

0.0

Total

7,715.9

100.0

1,036.3

100.0

8,752.1

100.0

Other non-amortizing features

Yes

10.9

0.1

43.3

4.2

54.3

0.6

162

Transaction type

Closed-end

HELOC

Total

Count

%

Count

%

Count

%

No

7,701.9

99.8

992.9

95.8

8,694.7

99.3

Exempt

3.1

0.0

0.1

0.0

3.1

0.0

Total

7,715.9

100.0

1,036.3

100.0

8,752.1

100.0

NOTE: Site-built single-family closed-end and HELOC originations.

163

TABLE 5.5.2: NON-AMORTIZING FEATURES BY ENHANCED LOAN TYPE: CLOSED-END ORIGINATIONS (COUNTS IN THOUSANDS)

Enhanced Loan Type

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Balloon payment

Yes

112.3

2.1

9.4

2.5

0.1

0.0

121.8

1.6

No

5,206.2

97.9

364.5

97.5

1,106.2

99.9

815.2

100.0

98.9

100.0

7,591.0

98.4

Exempt

2.0

0.0

0.1

0.0

0.6

0.1

0.4

0.1

0.0

3.1

0.0

Total

5,320.5

100.0

373.9

100.0

1,106.8

100.0

815.6

100.0

98.9

100.0

7,715.9

100.0

Interest-only

payments

Yes

102.3

1.9

55.0

14.7

0.5

0.0

0.3

0.0

158.0

2.1

No

5,216.3

98.0

318.9

85.3

1,105.8

99.9

815.0

99.9

98.9

100.0

7,554.8

97.9

Exempt

2.0

0.0

0.1

0.0

0.6

0.1

0.4

0.1

0.0

3.1

0.0

Total

5,320.5

100.0

373.9

100.0

1,106.8

100.0

815.6

100.0

98.9

100.0

7,715.9

100.0

Negative amortization

Yes

1.1

0.0

1.1

0.0

No

5,317.4

99.9

373.9

100.0

1,106.3

100.0

815.2

100.0

98.9

100.0

7,711.7

100.0

Exempt

2.0

0.0

0.1

0.0

0.6

0.1

0.4

0.1

0.0

3.1

0.0

Total

5,320.5

100.0

373.9

100.0

1,106.8

100.0

815.6

100.0

98.9

100.0

7,715.9

100.0

164

Enhanced Loan Type

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Other non-amortizing

features

Yes

8.7

0.2

0.4

0.1

0.9

0.1

0.8

0.1

10.9

0.1

No

5,309.8

99.8

373.5

99.9

1,105.4

99.9

814.4

99.9

98.8

99.9

7,701.9

99.8

Exempt

2.0

0.0

0.1

0.0

0.6

0.1

0.4

0.1

0.0

3.1

0.0

Total

5,320.5

100.0

373.9

100.0

1,106.8

100.0

815.6

100.0

98.9

100.0

7,715.9

100.0

NOTE: Site-built single-family closed-end originations.

165

TABLE 5.5.3: SELECTED CHARACTERISTICS OF CLOSED-END ORIGINATIONS BY NON-AMORTIZING FEATURES

Total

Originations

(thousands)

Share

Purchase

(%)

Share

Refi (%)

Share

Cashout

Refi (%)

Median

Interest

Rate (%)

Median

Loan

Amount

Median

Income

(thousands)

Median

Credit

Score

Median

CLTV

Median

DTI

Balloon payment

Yes

121.8

65.1

28.2

7.9

5.75

125,000

94

725

75.6

36.8

No

7,591.0

54.8

43.3

18.1

4.00

233,700

90

742

80.0

37.5

Exempt

3.1

15.3

82.4

52.4

4.12

284,075

95

693

75.0

40.1

Total

7,715.9

54.9

43.0

17.9

4.00

232,100

90

742

80.0

37.5

Interest-only payments

Yes

158.0

65.1

29.3

9.7

4.50

369,000

194

768

73.8

35.8

No

7,554.8

54.7

43.3

18.1

4.00

230,850

90

741

80.0

37.5

Exempt

3.1

15.3

82.4

52.4

4.12

284,075

95

693

75.0

40.1

Total

7,715.9

54.9

43.0

17.9

4.00

232,100

90

742

80.0

37.5

Negative amortization

Yes

1.1

44.9

35.0

4.9

5.50

83,825

75

739

70.0

33.9

No

7,711.7

54.9

43.0

17.9

4.00

232,110

90

742

80.0

37.5

Exempt

3.1

15.3

82.4

52.4

4.12

284,075

95

693

75.0

40.1

Total

7,715.9

54.9

43.0

17.9

4.00

232,100

90

742

80.0

37.5

Other non-amortizing

features

Yes

10.9

68.4

28.6

11.9

4.00

182,000

85

723

85.0

36.5

No

7,701.9

54.9

43.0

17.9

4.00

232,200

90

742

80.0

37.5

166

Total

Originations

(thousands)

Share

Purchase

(%)

Share

Refi (%)

Share

Cashout

Refi (%)

Median

Interest

Rate (%)

Median

Loan

Amount

Median

Income

(thousands)

Median

Credit

Score

Median

CLTV

Median

DTI

Exempt

3.1

15.3

82.4

52.4

4.12

284,075

95

693

75.0

40.1

Total

7,715.9

54.9

43.0

17.9

4.00

232,100

90

742

80.0

37.5

NOTE: Site-built single-family closed-end originations. Median loan amount, credit score, and DTI in the table are calculated from non-public raw data reported

by financial institutions. The outliers are excluded from the analysis sample to produce consistent estimates. Cells with frequency counts (of valid non-missing

values) less than 500 are omitted from the table.

167

TABLE 5.5.4: BALLOON AND INTEREST-ONLY FEATURES BY RACE ETHNICITY: CLOSED-END ORIGINATIONS (COUNTS IN THOUSANDS)

Race and Ethnicity

Asian

Black

Hispanic

white

Joint

Non-Hispanic

white

Other

Missing

Total

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Balloon payment

Yes

2.7

0.6

3.8

0.8

5.3

0.9

1.8

0.7

42.4

0.9

0.5

0.9

65.5

5.1

121.8

1.6

No

449.9

99.4

463.4

99.2

595.9

99.1

263.7

99.3

4,550.0

99.0

59.6

99.1

1,208.5

94.8

7,591.0

98.4

Exempt

0.1

0.0

0.2

0.0

0.3

0.1

1.8

0.0

0.1

0.5

0.0

3.1

0.0

Total

452.7

100.0

467.3

100.0

601.5

100.0

265.6

100.0

4,594.2

100.0

60.2

100.0

1,274.5

100.0

7,715.9

100.0

Interest-only

payments

Yes

4.6

1.0

3.1

0.7

3.9

0.7

3.3

1.2

81.1

1.8

0.4

0.7

61.6

4.8

158.0

2.1

No

448.0

99.0

464.0

99.3

597.2

99.3

262.2

98.7

4,511.3

98.2

59.8

99.3

1,212.3

95.1

7,554.8

97.9

Exempt

0.1

0.0

0.2

0.0

0.3

0.1

1.8

0.0

0.1

0.5

0.0

3.1

0.0

Total

452.7

100.0

467.3

100.0

601.5

100.0

265.6

100.0

4,594.2

100.0

60.2

100.0

1,274.5

100.0

7,715.9

100.0

NOTE: Site-built single-family closed-end originations.

168

TABLE 5.5.5: BALLOON AND INTEREST-ONLY FEATURES BY AGE: CLOSED-END ORIGINATIONS (COUNTS IN THOUSANDS)

Age group

<= 24

25-34

35-44

45-54

55-64

65-74

>= 75

Total

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Balloon payment

Yes

2.3

0.9

12.3

0.7

16.0

0.8

14.5

0.9

11.6

1.0

5.5

0.9

1.9

1.0

64.0

0.8

No

240.4

99.1

1,726.5

99.3

1,958.5

99.2

1,620.1

99.1

1,187.9

99.0

641.2

99.1

185.8

98.9

7,560.4

99.1

Exempt

0.0

0.3

0.0

0.7

0.0

0.8

0.1

0.7

0.1

0.3

0.1

2.8

0.0

Total

242.7

100.0

1,739.1

100.0

1,975.1

100.0

1,635.4

100.0

1,200.2

100.0

647.1

100.0

187.8

100.0

7,627.3

100.0

Interest-only

payments

Yes

1.3

0.5

15.9

0.9

29.3

1.5

28.9

1.8

24.8

2.1

12.8

2.0

4.1

2.2

117.0

1.5

No

241.4

99.5

1,722.9

99.1

1,945.2

98.5

1,605.7

98.2

1,174.7

97.9

633.9

98.0

183.6

97.8

7,507.4

98.4

Exempt

0.0

0.3

0.0

0.7

0.0

0.8

0.1

0.7

0.1

0.3

0.1

2.8

0.0

Total

242.7

100.0

1,739.1

100.0

1,975.1

100.0

1,635.4

100.0

1,200.2

100.0

647.1

100.0

187.8

100.0

7,627.3

100.0

NOTE: Site-built single-family closed-end originations.

169

TABLE 5.5.6: BALLOON AND INTEREST-ONLY FEATURES BY GEOGRAPHY: CLOSED-END ORIGINATIONS (COUNTS IN THOUSANDS)

Geography

Metropolitan area

Micropolitan area

Rural

Total

Count

%

Count

%

Count

%

Count

%

Balloon payment

Yes

96.1

1.4

9.1

2.0

16.6

6.1

121.8

1.6

No

6,895.4

98.6

440.2

98.0

255.4

93.9

7,591.0

98.4

Exempt

2.9

0.0

0.1

0.0

3.1

0.0

Total

6,994.4

100.0

449.4

100.0

272.1

100.0

7,715.9

100.0

Interest-only payments

Yes

133.3

1.9

10.0

2.2

14.7

5.4

158.0

2.1

No

6,858.1

98.1

439.2

97.8

257.4

94.6

7,554.8

97.9

Exempt

2.9

0.0

0.1

0.0

3.1

0.0

Total

6,994.4

100.0

449.4

100.0

272.1

100.0

7,715.9

100.0

NOTE: Site-built single-family closed-end originations.

170

TABLE 5.6.1: HAVING PREPAYMENT PENALTY TERM (YES/NO) BY ENHANCED LOAN TYPE: ORIGINATIONS (COUNTS IN THOUSANDS)

Has prepayment penalty term

No

Yes

Total

Count

%

Count

%

Count

%

Enhanced loan type

Conventional

Conforming

5,043.2

99.5

23.2

0.5

5,066.4

100.0

Jumbo

359.6

99.7

1.1

0.3

360.6

100.0

Non-conventional

FHA

1,105.2

100.0

0.0

1,105.3

100.0

VA

813.9

100.0

0.0

814.0

100.0

RHS/FSA

98.9

100.0

0.0

98.9

100.0

HELOC

778.5

76.1

244.6

23.9

1,023.1

100.0

Total

8,199.3

96.8

269.0

3.2

8,468.3

100.0

NOTE: Site-built single-family closed-end or HELOC originations.

171

TABLE 5.6.2: CLOSED-END ORIGINATIONS WITH OR WITHOUT PREPAYMENT PENALTY TERM BY RACE/ETHNICITY, AGE, AND GEOGRAPHY

(COUNTS IN THOUSANDS)

Has prepayment penalty term

No

Yes

Total

Count

%

Count

%

Count

%

Borrower race and ethnicity

Asian

383.3

99.7

1.3

0.3

384.5

100.0

Black

213.4

99.5

1.0

0.5

214.4

100.0

Hispanic white

362.0

99.4

2.3

0.6

364.3

100.0

Joint

178.1

99.5

0.8

0.5

178.9

100.0

Non-Hispanic white

3,397.5

99.5

16.0

0.5

3,413.5

100.0

Other

34.3

99.5

0.2

0.5

34.5

100.0

Missing

834.1

99.7

2.7

0.3

836.9

100.0

Total

5,402.8

99.6

24.3

0.4

5,427.1

100.0

Age group

<=24

133.9

99.9

0.1

134.0

100.0

25-34

1,167.1

99.8

2.4

0.2

1,169.5

100.0

35-44

1,405.5

99.6

5.2

0.4

1,410.8

100.0

45-54

1,178.4

99.5

6.3

0.5

1,184.7

100.0

55-64

911.7

99.4

5.9

0.6

917.6

100.0

65-74

467.6

99.3

3.2

0.7

470.8

100.0

172

Has prepayment penalty term

No

Yes

Total

Count

%

Count

%

Count

%

>=75

133.7

99.1

1.2

0.9

134.9

100.0

Total

5,397.9

99.6

24.2

0.4

5,422.1

100.0

Geography

Metropolitan Area

4,943.2

99.6

21.4

0.4

4,964.6

100.0

Micropolitan Area

285.9

99.4

1.7

0.6

287.6

100.0

Rural

173.6

99.3

1.3

0.7

174.9

100.0

Total

5,402.8

99.6

24.3

0.4

5,427.1

100.0

NOTE: Site-built single-family closed-end originations.

173

TABLE 5.6.3: HELOC ORIGINATIONS WITH OR WITHOUT PREPAYMENT PENALTY TERM BY NON-AMORTIZING FEATURES, RACE/ETHNICITY,

AGE, AND GEOGRAPHY (COUNTS IN THOUSANDS)

Has prepayment penalty term

No

Yes

Total

Count

%

Count

%

Count

%

ARM or Fixed

ARM

604.3

73.2

220.9

26.8

825.2

100.0

Fixed Rate

174.2

88.0

23.7

12.0

197.9

100.0

Total

778.5

76.1

244.6

23.9

1,023.1

100.0

Balloon payment

Yes

83.0

83.1

16.9

99.8

100.0

No

695.5

75.3

227.7

24.7

923.2

100.0

Total

778.5

76.1

244.6

23.9

1,023.1

100.0

Interest-only payments

Yes

418.4

77.6

121.0

22.4

539.4

100.0

No

360.1

74.4

123.6

25.6

483.7

100.0

Total

778.5

76.1

244.6

23.9

1,023.1

100.0

Negative amortization

Yes

1.3

56.3

1.0

43.7

2.3

100.0

No

777.2

76.1

243.6

23.9

1,020.8

100.0

Total

778.5

76.1

244.6

23.9

1,023.1

100.0

174

Has prepayment penalty term

No

Yes

Total

Count

%

Count

%

Count

%

Other non-amortizing features

0.0

Yes

42.3

100.0

0.0

42.3

100.0

No

736.2

75.1

244.6

24.9

980.8

100.0

Total

778.5

76.1

244.6

23.9

1,023.1

100.0

Borrower race and ethnicity

Asian

33.3

67.8

15.8

32.2

49.1

100.0

Black

26.1

79.4

6.8

20.6

32.9

100.0

Hispanic white

27.8

76.8

8.4

23.2

36.3

100.0

Joint

25.3

77.7

7.2

22.3

32.6

100.0

Non-Hispanic white

566.7

76.4

174.8

23.6

741.5

100.0

Other

6.4

73.6

2.3

26.4

8.8

100.0

Missing

92.8

76.0

29.2

24.0

122.0

100.0

Total

778.5

76.1

244.6

23.9

1,023.1

100.0

Age group

<=24

1.6

85.8

0.3

14.2

1.9

100.0

25-34

57.6

83.5

11.3

16.5

68.9

100.0

35-44

157.2

80.4

38.4

19.6

195.6

100.0

45-54

200.0

76.5

61.5

23.5

261.5

100.0

175

Has prepayment penalty term

No

Yes

Total

Count

%

Count

%

Count

%

55-64

193.1

74.1

67.5

25.9

260.5

100.0

65-74

122.3

72.4

46.5

27.6

168.8

100.0

>=75

45.9

70.8

19.0

29.2

64.9

100.0

Total

777.7

76.1

244.4

23.9

1,022.1

100.0

Geography

Metropolitan Area

708.9

76.1

222.8

23.9

931.7

100.0

Micropolitan Area

45.9

76.2

14.3

23.8

60.2

100.0

Rural

23.7

76.1

7.5

23.9

31.2

100.0

Total

778.5

76.1

244.6

23.9

1,023.1

100.0

NOTE: Site-built single-family HELOC originations.

176

TABLE 5.6.4: MOST COMMON PREPAYMENT PENALTY TERM FOR CLOSED-END MORTGAGES AND HELOCS

Count

%

Prepayment penalty term

(months)

Closed-end

36

18,197

74.7

24

3,494

14.3

12

1,591

6.5

Total top 3

23,282

95.6

Total

24,358

100.0

HELOC

36

191,078

78.1

24

48,928

20.0

12

4,545

1.9

Total top 3

244,551

100.0

Total

244,596

100.0

NOTE: Site-built single-family closed-end and HELOC originations.

177

TABLE 5.7.1: CHANNEL BY ENHANCED LOAN TYPE: ORIGINATIONS (COUNTS IN THOUSANDS)

Channel

Directly

submitted,

initially payable

Directly

submitted, not

initially payable

Not directly

submitted,

initially payable

Not directly

submitted, not

initially payable

Directly

submitted

exempt, initially

payable exempt

Total

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Enhanced loan type

Conventional

Conforming

4,477.8

84.2

76.9

1.4

559.2

10.5

205.7

3.9

1.0

0.0

5,320.5

100.0

Jumbo

316.6

84.7

1.8

0.5

26.3

7.0

29.2

7.8

0.0

373.9

100.0

Non-conventional

FHA

873.3

78.9

22.0

2.0

176.4

15.9

35.1

3.2

0.0

1,106.8

100.0

VA

669.0

82.0

14.5

1.8

94.9

11.6

37.2

4.6

0.0

815.6

100.0

RHS/FSA

76.2

77.0

2.6

13.0

13.2

7.1

7.2

0.0

98.9

100.0

HELOC

1,013.5

97.8

0.3

0.0

22.1

2.1

0.4

0.0

0.1

0.0

1,036.3

100.0

Reverse Mortgage

21.3

62.3

0.1

0.2

9.6

28.1

3.2

9.4

0.0

34.1

100.0

Total

7,447.6

84.8

118.1

1.3

901.5

10.3

317.8

3.6

1.1

0.0

8,786.3

100.0

NOTE: Site-built single-family originations.

178

TABLE 5.7.2: CHANNEL BY RACE/ETHNICITY, AGE, NEIGHBORHOOD INCOME, AND GEOGRAPHY (COUNTS IN THOUSANDS)

Channel

Directly

submitted,

initially payable

Directly

submitted, not

initially payable

Not directly

submitted,

initially payable

Not directly

submitted, not

initially payable

Directly

submitted

exempt, initially

payable exempt

Total

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Borrower race and ethnicity

Asian

327.8

72.4

7.2

1.6

88.4

19.5

29.0

6.4

0.2

0.1

452.7

100.0

Black

390.2

83.5

5.9

1.3

55.2

11.8

15.9

3.4

0.0

467.3

100.0

Hispanic white

472.1

78.5

10.1

1.7

96.1

16.0

23.1

3.8

0.1

0.0

601.5

100.0

Joint

223.9

84.3

3.5

1.3

28.1

10.6

10.1

3.8

0.1

0.0

265.6

100.0

Non-Hispanic white

3,882.3

84.5

71.2

1.5

446.5

9.7

193.8

4.2

0.3

0.0

4,594.2

100.0

Other

50.9

84.6

0.8

1.4

6.6

10.9

1.9

3.1

0.0

60.2

100.0

Missing

1,065.7

83.6

19.1

1.5

148.9

11.7

40.5

3.2

0.3

0.0

1,274.5

100.0

Total

6,412.9

83.1

117.8

1.5

869.9

11.3

314.3

4.1

1.0

0.0

7,715.9

100.0

Age group

<=24

201.6

83.1

5.1

2.1

26.5

10.9

9.4

3.9

0.0

242.7

100.0

25-34

1,426.5

82.0

32.3

1.9

208.4

12.0

71.8

4.1

0.1

0.0

1,739.1

100.0

35-44

1,608.4

81.4

32.5

1.6

247.1

12.5

87.1

4.4

0.2

0.0

1,975.1

100.0

45-54

1,354.2

82.8

23.4

1.4

189.4

11.6

68.2

4.2

0.2

0.0

1,635.4

100.0

55-64

1,018.8

84.9

15.0

1.3

120.4

10.0

45.8

3.8

0.2

0.0

1,200.2

100.0

65-74

558.6

86.3

7.1

1.1

58.0

9.0

23.3

3.6

0.1

0.0

647.1

100.0

179

Channel

Directly

submitted,

initially payable

Directly

submitted, not

initially payable

Not directly

submitted,

initially payable

Not directly

submitted, not

initially payable

Directly

submitted

exempt, initially

payable exempt

Total

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

>=75

163.0

86.8

1.9

1.0

16.3

8.7

6.7

3.6

0.0

187.8

100.0

Total

6,331.1

83.0

117.2

1.5

866.0

11.4

312.2

4.1

0.8

0.0

7,627.3

100.0

Geography

Metropolitan Area

5,788.2

82.8

109.2

1.6

816.4

11.7

279.7

4.0

1.0

0.0

6,994.4

100.0

Micropolitan Area

390.8

87.0

5.2

1.1

34.2

7.6

19.2

4.3

0.0

449.4

100.0

Rural

233.9

86.0

3.4

1.2

19.3

7.1

15.5

5.7

0.0

272.1

100.0

Total

6,412.9

83.1

117.8

1.5

869.9

11.3

314.3

4.1

1.0

0.0

7,715.9

100.0

NOTE: Site-built single-family closed-end originations.

180

TABLE 5.7.3: DENIAL RATES BY CHANNEL & ENHANCED LOAN TYPE (PERCENT)

Channel

Directly

submitted, initially

payable

Directly

submitted, not

initially payable

Not directly

submitted,

initially payable

Not directly

submitted, not

initially payable

Total

Enhanced loan type

Conventional

Conforming

13.5

8.6

6.3

5.8

12.4

Jumbo

11.9

8.3

17.0

19.0

12.9

Non-conventional

FHA

14.3

40.6

11.3

11.1

14.6

VA

13.4

36.9

5.8

6.1

12.9

RHS/FSA

8.5

15.8

14.3

8.6

9.5

HELOC

40.5

28.4

10.4

13.4

40.0

Reverse Mortgage

11.6

9.3

11.0

5.8

10.9

Total

18.5

21.8

7.9

17.2

NOTE: Site-built single-family homes. The denial rates are calculated based on applications that were denied, divided by (applications that were denied +

applications that were approved but not accepted + loans originated). The denial rate calculations do not include applications that were withdrawn or files that

were closed for incompleteness.

181

TABLE 6.1.1: OCCUPANCY TYPE BY ENHANCED LOAN TYPE: ORIGINATIONS (COUNTS IN THOUSANDS)

Occupancy type

Principal

residence

Secondary

residence

Investment

property

Total

Count

%

Count

%

Count

%

Count

%

Enhanced loan type

Conventional

Conforming

4,685.4

88.1

202.2

3.8

432.9

8.1

5,320.5

100.0

Jumbo

327.2

87.5

30.1

8.0

16.7

4.5

373.9

100.0

Non-conventional

FHA

1,105.3

99.9

0.2

0.0

1.3

0.1

1,106.8

100.0

VA

812.1

99.6

0.4

0.1

3.1

0.4

815.6

100.0

RHS/FSA

98.9

100.0

0.0

98.9

100.0

HELOC

1,002.7

96.8

14.6

1.4

18.9

1.8

1,036.3

100.0

Reverse Mortgage

34.1

100.0

0.0

34.1

100.0

Total

8,065.7

91.8

247.5

2.8

473.1

5.4

8,786.3

100.0

NOTE: Site-built single-family originations.

182

TABLE 6.1.2: SELECTED CHARACTERISTIC BY OCCUPANCY TYPE

Total

Originations

(thousands)

Share

Purchase

(%)

Share

Refi (%)

Share

Cashout

Refi (%)

Median

Interest

Rate (%)

Median

property

value

Median

Loan

Amount

Median

Income

(thousands)

Median

Credit

Score

Median

CLTV

Median

DTI

Occupancy type

Conventional conforming

Principal residence

4,685.4

50.8

46.4

19.8

4.00

320,000

228,400

92

754

80.0

36.0

Secondary residence

202.2

76.8

22.3

7.2

4.00

312,000

225,000

158

776

80.0

35.8

Investment property

432.9

56.9

40.0

19.7

5.00

240,000

161,250

130

763

75.0

37.2

Total

5,320.5

52.3

44.9

19.3

4.12

313,300

223,250

95

756

79.4

36.1

Jumbo

Principal residence

327.2

54.1

45.4

13.6

3.75

1,150,000

805,578

288

773

77.9

34.6

Secondary residence

30.1

65.6

33.8

8.4

3.62

1,250,000

838,000

502

778

74.0

33.4

Investment property

16.7

50.9

45.9

16.4

4.99

1,500,000

960,000

411

766

66.0

35.2

Total

373.9

54.8

44.5

13.3

3.75

1,170,000

815,000

300

774

76.5

34.5

NOTE: Site-built single-family closed-end conventional conforming and jumbo originations. Median property values, loan amounts, credit scores, and DTIs in the

table are calculated from non-public raw data reported by financial institutions. The outliers are excluded from the analysis sample to produce consistent

estimates.

183

TABLE 6.1.3: OCCUPANCY TYPE BY RACE/ETHNICITY, AGE, NEIGHBORHOOD INCOME, AND GEOGRAPHY: CLOSED-END CONVENTIONAL

ORIGINATIONS (COUNTS IN THOUSANDS)

Occupancy type

Principal

residence

Secondary

residence

Investment

property

Total

Count

%

Count

%

Count

%

Count

%

Borrower race and ethnicity

Asian

346.5

84.3

11.7

2.8

52.9

12.9

411.1

100.0

Black

200.6

89.7

5.3

2.4

17.7

7.9

223.5

100.0

Hispanic white

340.0

90.1

8.8

2.3

28.5

7.5

377.2

100.0

Joint

167.1

90.9

6.8

3.7

9.9

5.4

183.8

100.0

Non-Hispanic white

3,159.3

89.9

158.9

4.5

195.4

5.6

3,513.6

100.0

Other

32.4

90.7

1.0

2.9

2.3

6.4

35.7

100.0

Missing

766.7

80.7

39.8

4.2

143.0

15.1

949.5

100.0

Total

5,012.5

88.0

232.3

4.1

449.7

7.9

5,694.5

100.0

Age group

<=24

132.1

97.5

0.8

0.6

2.6

1.9

135.5

100.0

25-34

1,135.4

95.2

12.7

1.1

44.0

3.7

1,192.1

100.0

35-44

1,323.2

90.6

37.5

2.6

99.4

6.8

1,460.1

100.0

45-54

1,070.5

86.7

63.1

5.1

101.5

8.2

1,235.1

100.0

55-64

805.5

84.3

75.3

7.9

74.6

7.8

955.4

100.0

65-74

418.9

85.8

35.0

7.2

34.1

7.0

488.0

100.0

184

Occupancy type

Principal

residence

Secondary

residence

Investment

property

Total

Count

%

Count

%

Count

%

Count

%

>=75

122.9

87.7

6.7

4.8

10.6

7.6

140.2

100.0

Total

5,008.5

89.3

231.1

4.1

366.8

6.5

5,606.5

100.0

Neighborhood income

Low or moderate

701.5

80.7

28.6

3.3

139.7

16.1

869.8

100.0

Middle

2,056.3

87.9

110.2

4.7

172.7

7.4

2,339.2

100.0

High

2,237.1

91.2

92.1

3.8

123.7

5.0

2,452.9

100.0

Total

4,995.0

88.2

230.9

4.1

436.1

7.7

5,661.9

100.0

Geography

Metropolitan Area

4,620.8

88.9

171.5

3.3

403.0

7.8

5,195.3

100.0

Micropolitan Area

249.5

81.9

31.2

10.3

23.7

7.8

304.5

100.0

Rural

142.2

73.1

29.6

15.2

22.9

11.7

194.7

100.0

Total

5,012.5

88.0

232.3

4.1

449.7

7.9

5,694.5

100.0

NOTE: Site-built single-family closed-end conventional originations.

185

TABLE 6.1.4: OCCUPANCY TYPE BY ACTION TYPE: CONVENTIONAL CONFORMING AND JUMBO LARS (COUNTS IN THOUSANDS)

Action type

Originated

Approved,

not

accepted

Denied

Withdrawn

Closed for

incomplete

ness

Purchased

Preapproval

request

denied

Preapproved,

not accepted

Total

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Occupancy type

Conventional

Principal residence

4,685.4

57.0

171.3

2.1

848.5

10.3

1,069.8

13.0

286.3

3.5

1,064.4

13.0

44.3

0.5

46.3

0.6

8,216.4

100.0

Secondary residence

202.2

59.4

6.8

2.0

28.5

8.4

44.0

12.9

9.1

2.7

47.0

13.8

1.0

0.3

2.1

0.6

340.7

100.0

Investment property

432.9

56.1

21.6

2.8

91.6

11.9

115.4

15.0

28.5

3.7

77.9

10.1

1.5

0.2

2.6

0.3

772.1

100.0

Total

5,320.5

57.0

199.7

2.1

968.6

10.4

1,229.3

13.2

323.9

3.5

1,189.2

12.7

46.8

0.5

50.9

0.5

9,329.1

100.0

Jumbo

Principal residence

327.2

57.5

14.4

2.5

59.5

10.5

85.8

15.1

21.1

3.7

49.3

8.7

3.1

0.6

8.6

1.5

569.0

100.0

Secondary residence

30.1

63.6

1.0

2.2

4.5

9.4

6.4

13.6

1.6

3.5

3.1

6.6

0.1

0.3

0.4

0.9

47.3

100.0

Investment property

16.7

54.8

1.4

4.5

14.7

5.4

17.6

1.2

4.1

1.1

3.6

0.1

0.2

0.6

30.5

100.0

Total

373.9

57.8

16.8

2.6

68.5

10.6

97.6

15.1

24.0

3.7

53.5

8.3

3.3

0.5

9.2

1.4

646.8

100.0

NOTE: Site-built single-family closed-end conventional conforming and jumbo LARs.

186

TABLE 6.2.1: PROPERTY VALUE BY ENHANCED PRODUCT TYPE: ORIGINATIONS (DOLLARS IN THOUSANDS)

Mean Property Value

Median Property Value

Enhanced loan type

Conventional

Conforming

390.0

313.3

Jumbo

1,547.1

1,170.0

Non-conventional

FHA

244.4

223.5

VA

316.5

280.0

RHS/FSA

157.0

148.0

HELOC

475.8

344.0

Reverse Mortgage

518.3

351.0

Total

423.8

303.0

NOTE: Site-built single-family originations. The mean and median property values in the table are calculated from non-public raw data reported by financial

institutions. The outliers are excluded from the analysis sample to produce consistent estimates.

187

TABLE 6.2.2: MEDIAN PROPERTY VALUE BY LOAN PURPOSE, OCCUPANCY TYPE, AND LIEN STATUS: CLOSED-END ORIGINATIONS (DOLLARS

IN THOUSANDS)

Enhanced loan type

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Loan purpose

Home purchase

284.7

1,025.0

212.1

262.8

147.5

269.0

Home improvement

270.0

1,250.0

215.0

290.0

270.0

Other

270.0

1,620.0

155.0

297.5

280.0

NA

380.0

275.0

355.0

Non-cash-out refi

375.0

1,330.0

256.0

310.0

176.0

365.0

Cash-out refi

335.0

1,317.0

245.0

289.0

325.0

Total

313.3

1,170.0

223.5

280.0

148.0

300.0

Occupancy type

Principal Residence

320.0

1,150.0

223.5

280.0

148.0

300.0

Second Residence

312.0

1,250.0

349.0

Investment Property

240.0

1,500.0

195.0

250.0

248.0

Total

313.3

1,170.0

223.5

280.0

148.0

300.0

Lien status

First Lien

316.0

1,170.0

223.5

280.0

148.0

300.0

Subordinate Lien

260.0

961.8

230.0

262.0

188

Enhanced loan type

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Total

313.3

1,170.0

223.5

280.0

148.0

300.0

NOTE: Site-built single-family closed-end originations. The median property values in the table are calculated from non-public raw data reported by financial

institutions. The outliers are excluded from the analysis sample to produce consistent estimates. Cells with frequency counts (of valid non-missing values) less

than 500 are omitted from the table.

189

TABLE 6.4.1A: BORROWER CREDIT SCORING MODEL BY ENHANCED LOAN TYPE: ORIGINATIONS (COUNTS IN THOUSANDS)

Enhanced loan type

Conventional

Non-conventional

HELOC

Reverse

mortgage

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Credit scoring model

Equifax Beacon 5.0

1,653.3

31.1

125.1

33.4

316.3

28.6

231.8

28.4

29.5

29.9

204.3

19.7

0.9

2.5

2,561.2

29.2

Experian Fair Isaac Risk

Model v2

1,325.8

24.9

82.2

22.0

287.9

26.0

211.0

25.9

25.6

25.9

186.5

18.0

0.5

1.5

2,119.7

24.1

TransUnion FICO Risk

Score

Classic 04

1,487.1

27.9

104.0

27.8

312.5

28.2

224.6

27.5

30.1

30.4

105.5

10.2

0.6

1.7

2,264.3

25.8

Classic 98

11.6

0.2

0.1

2.3

0.2

1.1

0.1

0.3

8.0

0.8

0.0

23.4

0.3

VantageScore

2.0

5.5

0.1

0.2

0.1

1.5

0.1

1.1

0.1

0.0

8.7

0.1

3.0

2.9

0.1

0.0

0.2

0.0

12.3

1.2

0.0

15.5

0.2

More than 1 credit scoring

model

223.4

4.2

8.9

2.4

73.3

6.6

32.0

3.9

6.5

50.5

4.9

0.5

1.6

395.1

4.5

Other credit scoring model

150.5

2.8

4.5

1.2

4.8

0.4

4.6

0.6

388.4

37.5

0.0

553.4

6.3

Not applicable

459.1

8.6

48.7

13.0

108.0

9.8

109.4

13.4

6.2

6.3

80.5

7.8

31.6

92.7

843.7

9.6

Exempt

1.1

0.0

0.1

0.0

1.2

0.0

Total

5,320.5

100.0

373.9

100.0

1,106.8

100.0

815.6

100.0

98.9

100.0

1,036.3

100.0

34.1

100.0

8,786.3

100.0

190

NOTE: Site-built single-family originations. Borrowers only, not including co-borrowers.

191

TABLE 6.4.1B: CO-BORROWER CREDIT SCORING MODEL BY ENHANCED LOAN TYPE: ORIGINATIONS (COUNTS IN THOUSANDS)

Enhanced loan type

Conventional

Non-conventional

HELOC

Reverse

mortgage

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Credit scoring model

Equifax Beacon 5.0

375.4

7.1

32.3

8.6

59.8

5.4

55.0

6.7

4.9

62.3

6.0

0.3

0.8

589.9

6.7

Experian Fair Isaac Risk

Model v2

301.7

5.7

21.8

5.8

51.8

4.7

44.0

5.4

4.1

69.7

6.7

0.1

0.4

493.2

5.6

TransUnion FICO Risk

Score

Classic 04

338.2

6.4

27.3

7.3

57.3

5.2

52.1

6.4

4.9

39.5

3.8

0.2

0.4

519.5

5.9

Classic 98

4.8

0.1

0.0

0.5

0.0

0.1

0.0

4.8

0.5

0.0

10.4

0.1

VantageScore

2.0

2.5

0.0

0.1

0.0

0.6

0.1

0.5

0.1

0.0

3.9

0.0

3.0

1.2

0.0

0.1

0.0

6.5

0.6

0.0

7.9

0.1

More than 1 credit scoring

model

80.2

1.5

3.7

1.0

24.0

2.2

12.8

1.6

24.5

2.4

0.3

0.8

147.0

1.7

Other credit scoring model

60.4

1.1

2.0

0.5

1.3

0.1

1.3

0.2

0.1

130.2

12.6

0.0

195.2

2.2

Not applicable

1,324.1

24.9

141.7

37.9

245.8

22.2

237.4

29.1

13.8

13.9

233.2

22.5

21.4

62.8

2,217.4

25.2

No co-applicant

2,830.7

53.2

144.9

38.8

665.7

60.1

412.5

50.6

69.5

70.3

465.5

44.9

11.9

34.8

4,600.7

52.4

Exempt

1.1

0.0

0.1

0.0

1.2

0.0

192

Enhanced loan type

Conventional

Non-conventional

HELOC

Reverse

mortgage

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Total

5,320.5

100.0

373.9

100.0

1,106.8

100.0

815.6

100.0

98.9

100.0

1,036.3

100.0

34.1

100.0

8,786.3

100.0

NOTE: Site-built single-family originations.

193

TABLE 6.4.2: CREDIT SCORE DISTRIBUTION BY ENHANCED LOAN TYPE: ORIGINATIONS

Credit Score

Mean

Median

P5

P25

P75

P95

SD

Enhanced loan type

Conventional

Conforming

747

756

659

715

786

809

48

Jumbo

765

774

700

746

791

808

35

Non-conventional

FHA

668

663

599

637

695

757

47

VA

709

613

662

760

801

61

RHS/FSA

696

691

626

659

730

782

48

HELOC

764

773

671

730

804

836

53

Reverse Mortgage

729

749

574

686

792

813

76

Total

736

746

633

696

783

810

57

NOTE: Site-built single-family originations. The median credit scores in the table are calculated from non-public raw data reported by financial institutions. The

outliers are excluded from the analysis sample to produce consistent estimates.

194

TABLE 6.4.3: MEDIAN CREDIT SCORE OF EACH CLOSED-END ENHANCED LOAN TYPE BY LOAN PURPOSE, OCCUPANCY TYPE, AND LIEN

STATUS: ORIGINATIONS

Total

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Loan purpose

Home purchase

757

775

663

712

691

739

Home improvement

751

769

673

704

748

Other

738

771

656

710

739

NA

760

674

739

Non-cash-out refi

762

774

668

709

701

754

Cash-out refi

744

764

659

703

732

Total

756

774

663

709

691

742

Occupancy type

Principal Residence

754

773

663

709

691

739

Second Residence

776

778

776

Investment Property

763

766

700

733

763

Total

756

774

663

709

691

742

Lien status

First Lien

757

774

663

709

691

743

Subordinate Lien

724

760

706

724

195

Total

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Total

756

774

663

709

691

742

NOTE: Site-built single-family closed-end originations. The median credit scores in the table are calculated from non-public raw data reported by financial

institutions. The outliers are excluded from the analysis sample to produce consistent estimates. Cells with frequency counts (of valid non-missing values) less

than 500 are omitted from the table.

196

TABLE 6.4.4: MEDIAN CREDIT SCORE BY ENHANCED LOAN TYPE, RACE/ETHNICITY, AGE, NEIGHBORHOOD INCOME, AND GEOGRAPHY:

ORIGINATIONS

Total

Conventional

Non-conventional

HELOC

Reverse

Mortgage

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Borrower race and ethnicity

Asian

763

775

672

732

707

780

763

Black

724

749

656

676

675

736

694

Hispanic white

733

758

665

699

695

753

714

Joint

749

774

656

707

678

766

739

Non-Hispanic white

759

774

664

715

692

776

759

752

Other

740

765

663

697

690

762

723

Missing

756

774

663

712

692

772

746

Total

756

774

663

709

691

773

749

746

Age group

<=24

727

673

700

693

722

711

25-34

753

771

663

712

694

754

739

35-44

757

774

660

709

685

763

746

45-54

751

772

661

700

685

766

742

55-64

759

775

664

703

692

778

754

65-74

771

781

671

717

706

789

739

767

197

Total

Conventional

Non-conventional

HELOC

Reverse

Mortgage

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

>=75

777

784

675

730

794

763

775

Total

756

774

663

709

691

773

749

746

Neighborhood income

Low or moderate

745

772

663

696

686

761

728

Middle

753

772

662

705

691

771

752

739

High

762

774

664

719

697

778

756

758

Total

756

774

663

709

691

773

749

746

Geography

Metropolitan Area

756

774

663

709

693

774

749

747

Micropolitan Area

752

775

660

705

688

771

737

Rural

750

774

659

700

688

768

735

Total

756

774

663

709

691

773

749

746

NOTE: Site-built single-family originations. The median credit scores in the table are calculated from non-public raw data reported by financial institutions. The

outliers are excluded from the analysis sample to produce consistent estimates. Cells with frequency counts (of valid non-missing values) less than 500 are

omitted from the table.

198

TABLE 6.5.1: CLTV DISTRIBUTION BY ENHANCED LOAN TYPE: ORIGINATIONS (IN PERCENT)

CLTV

Mean

Median

P5

P25

P75

P95

SD

Enhanced loan type

Conventional

Conforming

74.4

79.4

36.4

65.0

89.7

97.0

18.9

Jumbo

72.6

76.5

43.1

65.2

80.0

90.0

14.9

Non-conventional

FHA

92.4

96.5

73.9

90.0

96.5

100.4

10.3

VA

93.8

100.0

69.6

90.0

100.0

101.4

11.6

RHS/FSA

98.3

100.0

89.4

98.0

101.0

5.5

HELOC

64.5

71.1

19.0

50.0

80.0

90.0

22.4

Reverse Mortgage

49.2

49.6

11.3

34.0

56.2

100.0

23.6

Total

77.0

80.0

35.8

67.9

95.0

100.0

20.0

NOTE: Site-built single-family originations. The outliers are excluded from the analysis sample to produce consistent estimates.

199

TABLE 6.5.2A: MEDIAN CLTV FOR CLOSE-END HOME-PURCHASE LOANS BY RACE/ETHNICITY, AGE, NEIGHBORHOOD INCOME, AND

GEOGRAPHY (IN PERCENT)

Enhanced loan type

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Borrower race and ethnicity

Asian

80.0

96.5

100.0

100.4

80.0

Black

95.0

81.7

96.5

100.0

100.1

96.5

Hispanic white

91.0

80.0

96.5

100.0

100.1

96.5

Joint

85.0

80.0

96.5

100.0

100.2

93.4

Non-Hispanic white

80.0

96.5

100.0

90.0

Other

90.0

80.0

96.5

100.0

96.5

Missing

80.0

96.5

100.0

90.0

Total

82.0

80.0

96.5

100.0

90.0

Age group

<=24

95.0

96.5

100.0

96.5

25-34

90.0

80.0

96.5

100.0

95.0

35-44

85.0

80.0

96.5

100.0

90.0

45-54

80.0

96.5

100.0

90.0

55-64

80.0

96.5

100.0

80.0

65-74

79.7

76.2

96.5

100.0

80.0

200

Enhanced loan type

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

>=75

77.7

75.0

96.5

100.0

80.0

Total

83.7

80.0

96.5

100.0

90.1

Neighborhood income

Low or moderate

90.0

80.0

96.5

100.0

95.0

Middle

85.0

80.0

96.5

100.0

95.0

High

80.0

96.5

100.0

85.0

Total

82.5

80.0

96.5

100.0

90.0

Geography

Metropolitan Area

82.6

80.0

96.5

100.0

100.1

90.0

Micropolitan Area

82.3

80.0

96.5

100.0

95.0

Rural

80.0

75.0

96.5

100.0

99.8

90.0

Total

82.0

80.0

96.5

100.0

90.0

NOTE: Site-built single-family closed-end home-purchase originations. The outliers are excluded from the analysis sample to produce consistent estimates. One

cell with frequency counts (of valid non-missing values) less than 500 are omitted from the table.

201

TABLE 6.5.2B: MEDIAN CLTV FOR CLOSE-END REFINANCE LOANS BY RACE/ETHNICITY, AGE, NEIGHBORHOOD INCOME, AND GEOGRAPHY

(IN PERCENT)

Enhanced loan type

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Borrower race and ethnicity

Asian

69.0

70.0

85.0

93.0

70.0

Black

72.0

75.0

85.0

96.6

79.6

Hispanic white

70.0

70.5

85.0

95.3

74.2

Joint

71.2

70.0

85.0

93.9

75.0

Non-Hispanic white

70.7

70.0

85.0

94.2

99.0

74.0

Other

70.0

85.0

94.4

75.0

Missing

70.0

69.0

85.0

93.6

74.0

Total

70.1

70.0

85.0

94.4

98.9

74.0

Age group

<=24

79.2

96.0

99.2

83.5

25-34

77.5

76.1

90.8

97.9

99.6

79.7

35-44

74.5

73.7

85.0

96.0

75.9

45-54

70.0

68.8

85.0

94.7

73.3

55-64

65.2

64.4

83.9

92.8

69.0

65-74

60.0

82.1

90.0

66.8

202

Enhanced loan type

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

>=75

57.0

56.4

80.2

90.0

63.7

Total

70.2

70.0

85.0

94.4

98.9

74.1

Neighborhood income

Low or moderate

70.0

72.0

85.0

94.6

74.4

Middle

71.0

71.5

85.0

94.7

99.0

75.0

High

70.0

69.5

85.0

94.0

72.9

Total

70.2

70.0

85.0

94.4

98.9

74.1

Geography

Metropolitan Area

70.1

70.0

85.0

94.4

99.0

74.0

Micropolitan Area

70.5

67.5

85.0

94.6

74.7

Rural

70.0

67.3

85.0

94.1

73.4

Total

70.1

70.0

85.0

94.4

98.9

74.0

NOTE: Site-built single-family closed-end refinance originations. The outliers are excluded from the analysis sample to produce consistent estimates. Cells with

frequency counts (of valid non-missing values) less than 500 are omitted from the table.

203

TABLE 6.6.1: DTI DISTRIBUTION BY ENHANCED LOAN TYPE (IN PERCENT): CLOSED-END AND HELOC ORIGINATIONS

DTI

Mean

Median

P5

P25

P75

P95

SD

Enhanced loan type

Conventional

Conforming

35

36

17

28

43

49

14

Jumbo

33

34

16

27

40

47

29

Non-conventional

FHA

43

44

24

37

50

56

11

VA

40

41

19

33

48

58

13

RHS/FSA

35

36

22

30

40

44

7

HELOC

40

28

0

64

131

45

Total

36

37

18

29

44

51

15

NOTE: Site-built single-family closed-end and HELOC originations. The DTIs used in the calculations are from non-public raw data reported by financial

institutions. The outliers are excluded from the analysis sample to produce consistent estimates.

204

TABLE 6.6.2: MEDIAN DTI OF ORIGINATED CLOSED-END MORTGAGES BY ENHANCED LOAN TYPE, LOAN PURPOSE, OCCUPANCY TYPE AND

LIEN STATUS (IN PERCENT)

Enhanced loan type

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Loan purpose

Home purchase

37

35

44

42

36

38

Home improvement

34

35

40

34

Other

36

39

42

36

NA

35

44

36

Non-cash-out refi

34

33

40

0

30

34

Cash-out refi

37

36

44

41

38

Total

36

34

44

41

36

37

Occupancy type

Principal Residence

36

35

44

41

36

38

Second Residence

36

33

36

Investment Property

37

35

37

Total

36

34

44

41

36

37

Lien status

First Lien

36

34

44

41

36

37

Subordinate Lien

37

34

38

37

205

Enhanced loan type

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Total

36

34

44

41

36

37

NOTE: Site-built single-family closed-end originations. The median DTIs in the table are calculated from non-public raw data reported by financial institutions.

The outliers are excluded from the analysis sample to produce consistent estimates. Cells with frequency counts (of valid non-missing values) less than 500 are

omitted from the table.

206

TABLE 6.6.3: MEDIAN DTI OF ORIGINATED CLOSED-END MORTGAGES BY ENHANCED LOAN TYPE, RACE/ETHNICITY, AGE, NEIGHBORHOOD

INCOME, AND GEOGRAPHY (IN PERCENT)

Enhanced loan type

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Borrower race and ethnicity

Asian

38

36

46

44

37

38

Black

39

36

45

43

37

41

Hispanic white

40

36

46

43

36

41

Joint

35

34

43

41

36

37

Non-Hispanic white

35

34

43

40

36

Other

38

36

43

36

40

Missing

36

34

44

42

37

38

Total

36

34

44

41

36

37

Age group

<=24

37

43

41

35

39

25-34

36

34

44

42

36

38

35-44

36

34

44

41

36

37

45-54

36

34

44

40

36

37

55-64

36

35

44

40

35

37

65-74

38

45

42

36

39

207

Enhanced loan type

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

>=75

39

45

43

40

Total

36

34

44

41

36

37

Neighborhood income

Low or moderate

37

35

44

41

36

39

Middle

36

35

44

41

36

38

High

36

34

44

41

36

37

Total

36

34

44

41

36

37

Geography

Metropolitan Area

36

34

44

41

36

38

Micropolitan Area

35

34

42

40

35

36

Rural

34

33

41

40

35

36

Total

36

34

44

41

36

37

NOTE: Site-built single-family closed-end originations. The median DTIs in the table are calculated from non-public raw data reported by financial institutions.

The outliers are excluded from the analysis sample to produce consistent estimates. Cells with frequency counts (of valid non-missing values) less than 500 are

omitted from the table.

208

TABLE 6.7.1: SELECTED CHARACTERISTICS OF MANUFACTURED HOME ORIGINATIONS BY SECURED PROPERTY TYPE (COUNT AND INCOME

IN THOUSANDS)

Count

Median

Interest

Rate (%)

Median

Income

($thousa

nds)

Median

Credit

Score

Median

CLTV

Median

DTI

Share

Purchase

(%)

Fixed rate

share (%)

Secured property type

Manufactured home and land

106.1

4.750

54

700

89.0

37.7

66

91

Manufactured home and not land

53.9

8.490

53

680

83.1

35.6

96

94

NA

1.3

4.683

60

710

74.9

39.7

64

80

Exempt

16.9

6.600

59

757

61.0

30.9

56

83

Total

178.2

5.500

54

695

86.8

37.0

74

92

NOTE: Manufactured home originations. The median credit scores and DTIs in the table are calculated from non-public raw data reported by financial

institutions. The outliers are excluded from the analysis sample to produce consistent estimates. Cells with frequency counts (of valid non-missing values) less

than 500 are omitted from the table.

209

TABLE 6.7.2: MANUFACTURED HOME ORIGINATION SECURED PROPERTY TYPE BY TRANSACTION TYPE AND LOAN TYPE (THOUSANDS)

Enhanced loan type

Excluding reverse

Reverse

Mortgage

Total

Closed-end

Open-end

Conventional

FHA

VA

RHS/FSA

Secured property type

Manufactured home and land

55.6

31.2

13.1

0.6

4.8

0.7

106.0

Manufactured home and not land

53.2

0.5

0.2

0.0

53.9

NA

1.0

0.2

0.1

0.0

1.3

Total

109.7

31.9

13.3

0.6

4.9

0.7

161.2

NOTE: Manufactured home originations.

210

TABLE 6.7.3: MANUFACTURED HOME ORIGINATION SECURED PROPERTY TYPE BY RACE/ETHNICITY, AGE, NEIGHBORHOOD INCOME, AND

GEOGRAPHY

Secured property type

Manufactured home

and land

Manufactured home

and not land

NA

Exempt

Total

Count

%

Count

%

Count

%

Count

%

Count

%

Borrower race and ethnicity

Asian

0.5

48.6

0.5

44.6

0.0

0.8

0.1

6.0

1.1

100.0

Black

3.4

38.8

4.9

55.8

0.0

0.2

0.4

5.1

8.8

100.0

Hispanic white

7.0

56.2

4.8

38.8

0.1

0.8

0.5

4.2

12.4

100.0

Joint

2.7

59.5

1.5

33.4

0.0

0.5

0.3

6.5

4.6

100.0

Non-Hispanic white

78.9

64.9

28.7

23.6

0.5

0.4

13.6

11.2

121.6

100.0

Other

1.1

49.4

1.0

44.8

0.0

0.3

0.1

5.5

2.3

100.0

Missing

12.5

45.5

12.5

45.4

0.7

2.6

1.8

6.5

27.5

100.0

Total

106.1

59.6

53.9

30.2

1.3

0.7

16.9

9.5

178.2

100.0

Age group

<=24

6.8

54.0

4.8

38.5

0.0

0.3

0.9

7.2

12.6

100.0

25-34

21.7

60.2

11.8

32.8

0.1

0.3

2.4

6.6

36.1

100.0

35-44

18.6

60.1

9.7

31.3

0.1

0.4

2.5

8.2

30.9

100.0

45-54

19.9

61.4

9.0

27.8

0.1

0.4

3.4

10.4

32.3

100.0

55-64

20.1

60.8

9.3

28.0

0.1

0.4

3.6

10.8

33.1

100.0

65-74

13.6

62.5

5.9

27.1

0.1

0.4

2.2

10.0

21.8

100.0

211

Secured property type

Manufactured home

and land

Manufactured home

and not land

NA

Exempt

Total

Count

%

Count

%

Count

%

Count

%

Count

%

>=75

4.6

63.9

1.9

25.9

0.0

0.3

0.7

9.9

7.2

100.0

Total

105.3

60.5

52.4

30.1

0.7

0.4

15.7

9.0

174.0

100.0

Neighborhood income

Low or moderate

19.2

53.4

13.8

38.3

0.3

1.0

2.6

7.3

36.0

100.0

Middle

70.0

62.3

30.3

27.0

0.7

0.6

11.3

10.1

112.3

100.0

High

16.4

60.9

8.1

30.1

0.3

0.9

2.2

8.1

27.0

100.0

Total

105.6

60.3

52.2

29.8

1.3

0.8

16.1

9.2

175.3

100.0

Geography

Metropolitan Area

65.6

58.2

37.1

32.9

1.0

0.9

9.0

8.0

112.8

100.0

Micropolitan Area

23.2

66.7

8.3

23.8

0.2

0.5

3.1

9.0

34.8

100.0

Rural

17.3

56.5

8.6

27.9

0.1

0.3

4.7

15.3

30.6

100.0

Total

106.1

59.6

53.9

30.2

1.3

0.7

16.9

9.5

178.2

100.0

NOTE: Manufactured home originations.

212

TABLE 6.8.1: SELECTED CHARACTERISTICS OF MANUFACTURED HOME ORIGINATIONS BY LAND PROPERTY INTEREST (COUNT AND

INCOME IN THOUSANDS)

Count

Median

Interest

Rate (%)

Median

Income

($thousa

nds)

Median

Credit

Score

Median

CLTV

Median

DTI

Share

Purchase

(%)

Fixed rate

share (%)

Land property

interest

Direct ownership

121.6

4.875

55

699

89.0

37.5

70

91

Indirect ownership

0.9

6.500

61

739

75.0

34.6

90

96

Paid leasehold

25.4

8.600

53

683

80.0

36.0

96

93

Unpaid leasehold

12.0

8.990

48

659

94.3

35.1

99

96

NA

1.4

4.740

59

710

75.0

39.4

64

80

Exempt

16.9

6.600

59

61.0

56

84

Total

178.2

5.500

54

695

86.8

37.0

74

92

NOTE: Manufactured home originations. The median credit scores and DTIs in the table are calculated from non-public raw data reported by financial

institutions. The outliers are excluded from the analysis sample to produce consistent estimates. Cells with frequency counts (of valid non-missing values) less

than 500 are omitted from the table.

213

TABLE 6.8.2: MANUFACTURED HOME ORIGINATION LAND PROPERTY INTEREST BY ENHANCED LOAN TYPE (THOUSANDS)

Enhanced loan type

Excluding reverse

Reverse

Mortgage

Total

Closed-end

Open-end

Conventional

FHA

VA

RHS/FSA

Land property

interest

Direct ownership

70.6

31.6

13.2

0.6

4.9

0.7

121.5

Indirect ownership

0.9

0.0

0.1

0.0

0.9

Paid leasehold

25.3

0.1

0.0

25.4

Unpaid leasehold

12.0

0.1

0.0

12.0

NA

1.0

0.2

0.1

0.0

1.4

Exempt

0.0

Total

109.7

31.9

13.3

0.6

4.9

0.7

161.2

NOTE: Manufactured home originations.

214

TABLE 6.8.3: MANUFACTURED HOME ORIGINATION LAND PROPERTY INTEREST BY RACE/ETHNICITY, AGE, NEIGHBORHOOD INCOME, AND

GEOGRAPHY (COUNTS IN THOUSANDS)

Land property interest

Direct

ownership

Indirect

ownership

Paid leasehold

Unpaid

leasehold

NA

Exempt

Total

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Borrower race and ethnicity

Asian

0.6

53.0

0.0

0.5

0.4

35.9

0.0

3.8

0.0

0.8

0.1

6.0

1.1

100.0

Black

5.0

57.6

0.0

0.2

1.5

17.1

1.7

19.7

0.0

0.2

0.4

5.1

8.8

100.0

Hispanic white

8.0

64.9

0.0

0.4

2.7

21.8

1.0

7.9

0.1

0.8

0.5

4.2

12.4

100.0

Joint

3.0

66.7

0.0

0.7

0.9

19.1

0.3

6.4

0.0

0.6

0.3

6.5

4.6

100.0

Non-Hispanic white

86.4

71.1

0.7

0.6

13.8

11.4

6.5

5.4

0.5

0.4

13.6

11.2

121.6

100.0

Other

1.4

59.0

0.0

0.4

18.9

0.4

15.8

0.0

0.3

0.1

5.5

2.3

100.0

Missing

17.1

62.4

0.1

0.4

5.6

20.5

2.1

7.6

0.7

2.6

1.8

6.5

27.5

100.0

Total

121.6

68.3

0.9

0.5

25.4

14.2

12.0

6.8

1.4

0.8

16.9

9.5

178.2

100.0

Age group

<=24

7.5

59.9

0.1

0.4

1.8

14.5

2.2

17.6

0.0

0.4

0.9

7.2

12.6

100.0

25-34

24.2

67.2

0.1

0.4

5.3

14.7

3.9

10.8

0.1

0.4

2.4

6.6

36.1

100.0

35-44

21.2

68.6

0.1

0.4

4.6

14.8

2.3

7.5

0.1

0.5

2.5

8.2

30.9

100.0

45-54

22.8

70.6

0.1

0.4

4.3

13.2

1.6

5.0

0.1

0.4

3.4

10.4

32.3

100.0

55-64

23.2

70.0

0.2

0.7

4.9

14.7

1.1

3.3

0.1

0.4

3.6

10.8

33.1

100.0

65-74

15.3

70.2

0.2

0.9

3.4

15.4

0.6

3.0

0.1

0.5

2.2

10.0

21.8

100.0

215

Land property interest

Direct

ownership

Indirect

ownership

Paid leasehold

Unpaid

leasehold

NA

Exempt

Total

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

>=75

5.0

69.9

0.1

0.8

1.1

15.6

0.2

3.4

0.0

0.4

0.7

9.9

7.2

100.0

Total

119.3

68.6

0.9

0.5

25.3

14.6

12.0

6.9

0.7

0.4

15.7

9.0

174.0

100.0

Neighborhood income

Low or moderate

22.7

63.1

0.2

0.6

7.6

21.2

2.5

6.8

0.3

1.0

2.6

7.3

36.0

100.0

Middle

79.3

70.6

0.6

0.5

13.2

11.7

7.2

6.4

0.8

0.7

11.3

10.1

112.3

100.0

High

18.7

69.2

0.1

0.5

3.9

14.3

1.8

6.8

0.3

1.0

2.2

8.1

27.0

100.0

Total

120.7

68.8

0.9

0.5

24.7

14.1

11.5

6.6

1.4

0.8

16.1

9.2

175.3

100.0

Geography

Metropolitan Area

74.2

65.8

0.7

0.6

21.9

19.4

6.0

5.3

1.1

0.9

9.1

8.0

112.8

100.0

Micropolitan Area

26.5

76.3

0.1

0.4

2.0

5.9

2.7

7.9

0.2

0.6

3.1

9.0

34.8

100.0

Rural

21.0

68.4

0.1

0.4

1.5

4.7

3.3

10.9

0.1

0.4

4.7

15.3

30.6

100.0

Total

121.6

68.3

0.9

0.5

25.4

14.2

12.0

6.8

1.4

0.8

16.9

9.5

178.2

100.0

NOTE: Manufactured home originations.

216

TABLE 6.8.4: MANUFACTURED HOME ORIGINATION LAND PROPERTY INTEREST BY SECURED PROPERTY TYPE

Land property interest

Direct

ownership

Indirect

ownership

Paid leasehold

Unpaid

leasehold

NA

Exempt

Total

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Secured property type

Manufactured home and land

105.6

99.5

0.4

0.3

0.1

0.0

0.1

0.0

106.1

100.0

Manufactured home and not land

16.0

29.7

0.6

1.0

25.3

47.0

12.0

22.3

0.0

53.9

100.0

NA

0.0

2.6

0.0

1.3

97.4

0.0

1.3

100.0

Exempt

0.0

16.9

100.0

16.9

100.0

Total

121.6

68.3

0.9

0.5

25.4

14.2

12.0

6.8

1.4

0.8

16.9

9.5

178.2

100.0

NOTE: Manufactured home originations.

217

TABLE 6.9.1A: MULTIFAMILY HOME ORIGINATION NUMBER OF AFFORDABLE UNITS BY DISCLOSED TOTAL UNITS

# of loans

# of loans with

affordable units

Share (%)

Disclosed units

5-24

34,799

860

2.5

25-49

6,806

484

7.1

50-99

4,941

620

12.5

100-149

2,186

349

16.0

150+

4,687

628

13.4

Total

53,419

2,941

5.5

NOTE: Site-built multifamily originations.

218

TABLE 6.9.1B: DISTRIBUTION OF AFFORDABLE UNITS / NUMBER OF TOTAL UNITS

% of Affordable Units / Number of Total Units

Mean

Median

P5

P25

P75

P95

Disclosed units

5-24

68

100

8

27

100

25-49

66

97

5

20

100

50-99

77

100

10

54

100

100-149

79

100

9

69

100

150+

67

94

4

22

100

Total

71

99

6

30

100

NOTE: Site-built multifamily originations with at least one affordable unit. The percentages of affordable units / number of total units in the table are calculated

from non-public raw data reported by financial institutions and may differ slightly from the public data due to rounding.

219

TABLE 7.1.1: DISTRIBUTION OF INTEREST RATES BY ENHANCED LOAN TYPE (PERCENT)

Interest Rate

Mean

Median

P5

P25

P75

P95

SD

Enhanced loan type

Conventional

Conforming

4.275

4.125

3.250

3.750

4.625

5.800

1.033

Jumbo

3.869

3.750

2.750

3.375

4.125

5.375

0.934

Non-conventional

FHA

4.231

4.125

3.375

3.750

4.625

5.375

0.656

VA

3.858

3.750

3.125

3.500

4.125

4.875

0.599

RHS/FSA

4.120

4.125

3.375

3.750

4.500

5.000

0.559

HELOC

5.382

5.340

2.740

4.480

6.240

8.620

1.715

Reverse Mortgage

4.645

4.482

3.423

3.904

5.261

6.750

0.993

Total

4.343

4.125

3.125

3.750

4.750

6.250

1.135

NOTE: Site-built single-family originations. The outliers are excluded from the analysis sample to produce consistent estimates.

220

TABLE 7.1.2: MEDIAN INTEREST RATE BY ENHANCED LOAN TYPE, LOAN PURPOSE, OCCUPANCY TYPE, AND LIEN STATUS (PERCENT)

Enhanced loan type

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Loan purpose

Home purchase

4.180

3.875

4.250

3.990

4.125

Home improvement

5.000

4.125

4.625

3.875

5.000

Other

4.990

3.425

4.375

3.750

4.875

NA

4.250

3.750

4.000

Non-cash-out refi

3.875

3.625

3.875

3.500

3.625

3.875

Cash-out refi

4.250

3.750

4.125

3.875

4.125

Total

4.125

3.750

4.125

3.750

4.125

4.000

Occupancy type

Principal Residence

4.000

3.750

4.125

3.750

4.125

4.000

Second Residence

4.000

3.625

4.000

Investment Property

5.000

4.990

4.125

3.500

5.000

Total

4.125

3.750

4.125

3.750

4.125

4.000

Lien status

First Lien

4.000

3.750

4.125

3.750

4.125

4.000

Subordinate Lien

5.340

4.500

5.950

5.340

Total

4.125

3.750

4.125

3.750

4.125

4.000

221

NOTE: Site-built single-family closed-end originations. The outliers are excluded from the analysis sample to produce consistent estimates. Cells with frequency

counts (of valid non-missing values) less than 500 are omitted from the table.

222

TABLE 7.1.3: MEDIAN INTEREST RATE BY ENHANCED LOAN TYPE, RACE/ETHNICITY, AGE, NEIGHBORHOOD INCOME, AND GEOGRAPHY

(PERCENT)

Enhanced loan type

Conventional

Non-conventional

HELOC

Reverse

Mortgage

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Borrower race and ethnicity

Asian

3.990

3.500

4.000

3.625

4.000

5.090

4.739

3.990

Black

4.375

4.000

4.250

3.875

4.125

5.750

4.521

4.250

Hispanic white

4.250

4.000

4.250

3.750

4.000

5.625

4.383

4.250

Joint

4.000

3.750

4.125

3.750

4.125

5.490

4.293

4.000

Non-Hispanic white

4.125

3.750

4.125

3.750

4.125

5.330

4.462

4.125

Other

4.250

3.875

4.125

3.750

4.125

5.500

4.125

Missing

4.125

3.750

4.125

3.625

4.000

5.260

4.661

4.100

Total

4.125

3.750

4.125

3.750

4.125

5.340

4.482

4.125

Age group

<=24

4.250

3.875

4.250

3.875

4.125

5.750

4.250

25-34

4.000

3.750

4.125

3.750

4.125

5.750

4.025

35-44

4.000

3.750

4.125

3.750

4.125

5.500

4.000

45-54

4.125

3.750

4.125

3.750

4.125

5.440

4.125

55-64

4.125

3.750

4.125

3.750

4.125

5.250

4.346

4.125

65-74

4.125

3.690

4.125

3.750

4.125

5.240

4.375

4.125

223

Enhanced loan type

Conventional

Non-conventional

HELOC

Reverse

Mortgage

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

>=75

4.125

3.625

4.125

3.750

5.240

4.644

4.250

Total

4.125

3.750

4.125

3.750

4.125

5.340

4.482

4.125

Neighborhood income

Low or moderate

4.310

3.875

4.250

3.875

4.125

5.500

4.491

4.250

Middle

4.125

3.750

4.125

3.750

4.125

5.390

4.452

4.125

High

4.000

3.750

4.125

3.700

4.000

5.250

4.519

4.000

Total

4.125

3.750

4.125

3.750

4.125

5.340

4.482

4.125

Geography

Metropolitan Area

4.125

3.750

4.125

3.750

4.125

5.340

4.465

4.125

Micropolitan Area

4.250

3.875

4.250

3.875

4.125

5.375

4.549

4.250

Rural

4.250

3.875

4.125

5.280

4.716

4.250

Total

4.125

3.750

4.125

3.750

4.125

5.340

4.482

4.125

NOTE: Closed-end single-family originations. The outliers are excluded from the analysis sample to produce consistent estimates. Cells with frequency counts

(of valid non-missing values) less than 500 are omitted from the table.

224

TABLE 7.1.4: MEDIAN INTEREST RATE: FIXED RATE VS. ARM (PERCENT)

Enhanced loan type

Conventional

Non-conventional

HELOC

Reverse

Mortgage

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

ARM or Fixed Rate

ARM

4.125

3.375

4.000

3.875

5.250

4.195

4.750

Fixed rate

4.125

3.875

4.125

3.750

4.125

5.500

4.972

4.125

Total

4.125

3.750

4.125

3.750

4.125

5.340

4.484

4.125

NOTE: Site-built single-family originations. The outliers are excluded from the analysis sample to produce consistent estimates. One cell with frequency count (of

valid non-missing values) less than 500 is omitted from the table.

225

TABLE 7.1.5: MEDIAN INTEREST RATE BY COMMON LOAN TERM: CONVENTIONAL FIXED RATE MORTGAGES (PERCENT)

Enhanced loan type

Conventional

Total

Conforming

Jumbo

Loan term

5 years

4.950

5.000

4.950

10 years

4.500

3.875

4.490

15 years

3.625

3.375

3.625

20 years

4.000

3.750

4.000

30 years

4.125

3.875

4.125

Other

4.550

4.875

4.590

Total

4.125

3.875

4.125

NOTE: Site-built single-family closed-end conventional fixed rate originations. The outliers are excluded from the analysis sample to produce consistent

estimates.

226

TABLE 7.1.6: MEDIAN INTEREST RATE BY COMMON INTRODUCTORY RATE PERIOD FOR CONVENTIONAL ARMS, 30-YR TERM (PERCENT)

Enhanced loan type

Conventional

Total

Conforming

Jumbo

Introductory rate period

1 year

4.000

3 years

4.375

3.700

4.250

5 years

4.375

3.625

4.125

7 years

3.750

3.375

3.600

10 years

3.750

3.375

3.500

15 years

3.950

3.875

< 1 year

5.500

3.875

4.875

Other

4.250

4.125

4.250

Total

3.875

3.375

3.750

NOTE: Site-built single-family closed-end conventional ARM originations with a 30-year term. The outliers are excluded from the analysis sample to produce

consistent estimates.

227

TABLE 7.2.1: DISTRIBUTION OF RATE SPREAD BY ENHANCED LOAN TYPE (PERCENT)

Rate Spread

Mean

Median

P5

P25

P75

P95

SD

Enhanced loan type

Conventional

Conforming

0.511

0.357

-0.268

0.087

0.747

1.837

20.212

Jumbo

0.005

-0.111

-0.695

-0.358

0.164

1.037

13.143

Non-conventional

FHA

1.293

1.207

0.447

0.866

1.603

2.380

11.926

VA

0.049

-0.019

-0.607

-0.282

0.304

0.961

0.513

RHS/FSA

0.695

0.692

0.039

0.401

0.984

1.347

0.520

HELOC

0.768

0.800

-2.100

0.000

1.630

3.650

5.094

Total

0.580

0.422

-0.476

0.055

0.983

2.178

16.511

NOTE: Site-built single-family forward originations.

228

TABLE 7.2.2: MEDIAN RATE SPREAD BY ENHANCED LOAN TYPE, LOAN PURPOSE, OCCUPANCY TYPE, AND LIEN STATUS (PERCENT)

Enhanced loan type

Conventional

Non-conventional

HELOC

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Loan purpose

Home purchase

0.351

-0.091

1.315

0.066

0.701

0.990

0.441

Home improvement

1.300

0.101

1.657

0.149

0.970

1.060

Other

1.280

-0.268

1.404

-0.102

0.840

0.930

NA

0.688

0.906

0.841

Non-cash-out refi

0.226

-0.164

1.001

-0.206

0.404

0.470

0.209

Cash-out refi

0.484

-0.048

1.001

0.156

0.750

0.501

Total

0.357

-0.111

1.207

-0.019

0.692

0.800

0.422

Occupancy type

Principal Residence

0.348

-0.108

1.207

-0.019

0.692

0.800

0.419

Second Residence

0.296

-0.175

0.580

0.256

Investment Property

1.125

0.085

1.358

-0.155

1.400

1.115

Total

0.357

-0.111

1.207

-0.019

0.692

0.800

0.422

Lien status

First Lien

0.340

-0.114

1.206

-0.019

0.692

0.460

0.381

Subordinate Lien

1.675

0.516

4.690

0.970

1.130

Total

0.357

-0.111

1.207

-0.019

0.692

0.800

0.422

229

NOTE: Site-built single-family forward originations. Cells with frequency counts (of valid non-missing values) less than 500 are omitted from the table.

230

TABLE 7.2.3: MEDIAN RATE SPREAD BY ENHANCED LOAN TYPE, RACE/ETHNICITY, AGE, NEIGHBORHOOD INCOME, AND GEOGRAPHY

(PERCENT)

Total

Conventional

Non-conventional

HELOC

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Borrower race and ethnicity

Asian

0.173

-0.277

1.050

-0.155

0.572

0.650

0.168

Black

0.576

0.094

1.307

0.046

0.751

1.240

0.749

Hispanic white

0.541

0.046

1.297

-0.005

0.685

1.200

0.747

Joint

0.319

-0.109

1.193

-0.059

0.658

0.890

0.350

Non-Hispanic white

0.352

-0.073

1.187

-0.009

0.695

0.770

0.405

Other

0.447

0.008

1.167

-0.020

0.686

0.980

0.520

Missing

0.343

-0.144

1.143

-0.070

0.655

0.800

0.380

Total

0.357

-0.111

1.207

-0.019

0.692

0.800

0.422

Age group

<=24

0.556

1.317

0.096

0.703

1.190

0.723

25-34

0.333

-0.153

1.243

-0.016

0.672

1.220

0.429

35-44

0.322

-0.111

1.201

-0.050

0.689

1.030

0.394

45-54

0.382

-0.091

1.187

-0.020

0.729

0.850

0.446

55-64

0.376

-0.100

1.157

-0.004

0.746

0.680

0.417

65-74

0.362

-0.126

1.112

-0.013

0.729

0.610

0.360

231

Total

Conventional

Non-conventional

HELOC

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

>=75

0.393

-0.150

1.076

-0.015

0.610

0.379

Total

0.357

-0.111

1.207

-0.019

0.692

0.800

0.422

Neighborhood income

Low or moderate

0.498

-0.103

1.297

0.054

0.767

0.990

0.648

Middle

0.403

-0.070

1.216

0.008

0.697

0.820

0.490

High

0.273

-0.120

1.119

-0.080

0.612

0.740

0.289

Total

0.357

-0.111

1.208

-0.019

0.693

0.800

0.422

Geography

Metropolitan Area

0.347

-0.115

1.196

-0.029

0.656

0.810

0.409

Micropolitan Area

0.461

-0.025

1.315

0.063

0.753

0.720

0.541

Rural

0.491

0.009

1.341

0.085

0.767

0.710

0.571

Total

0.357

-0.111

1.207

-0.019

0.692

0.800

0.422

NOTE: Site-built single-family forward originations. Cells with frequency counts (of valid non-missing values) less than 500 are omitted from the table.

232

TABLE 7.3.1: DISTRIBUTION OF TOTAL LOAN COSTS BY ENHANCED LOAN TYPE ($)

Total loan costs

Mean

Median

P5

P25

P75

P95

SD

Enhanced loan type

Conventional

Conforming

3,888

3,404

120

2,364

4,872

8,680

7,321

Jumbo

6,210

4,857

635

3,327

7,300

15,480

8,782

Non-conventional

FHA

7,810

7,129

3,180

5,373

9,305

13,974

15,458

VA

5,777

4,274

0

2,224

7,866

15,549

12,116

RHS/FSA

4,753

4,547

1,586

3,497

5,667

8,028

6,920

Total

4,809

3,925

250

2,598

6,125

11,325

9,738

NOTE: Site-built single-family closed-end originations.

233

TABLE 7.3.2: MEDIAN TOTAL LOAN COSTS BY ENHANCED LOAN TYPE ($)

Enhanced loan type

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Loan purpose

Home purchase

3,689

5,786

7,335

5,330

4,551

4,382

Home improvement

280

4,989

6,371

6,794

375

Other

499

4,444

4,992

5,134

685

NA

2,354

6,427

3,839

Non-cash-out refi

2,950

3,717

5,992

2,651

4,375

3,150

Cash-out refi

3,626

4,189

7,560

6,775

4,222

Total

3,404

4,857

7,129

4,274

4,547

3,925

Occupancy type

Principal Residence

3,369

4,796

7,130

4,283

4,547

3,931

Second Residence

3,546

5,339

3,679

Investment Property

4,121

6,521

5,145

2,049

4,135

Total

3,404

4,857

7,129

4,274

4,547

3,925

Lien status

First Lien

3,498

4,875

7,136

4,274

4,547

4,022

Subordinate Lien

110

2,285

1,453

120

Total

3,404

4,857

7,129

4,274

4,547

3,925

234

NOTE: Site-built single-family closed-end originations. Cells with frequency counts (of valid non-missing values) less than 500 are omitted from the table.

235

TABLE 7.3.3: MEDIAN TOTAL LOAN COSTS BY ENHANCED LOAN TYPE, RACE/ETHNICITY, AGE, NEIGHBORHOOD INCOME, AND GEOGRAPHY

($)

Total

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Borrower race and ethnicity

Asian

3,653

4,184

8,707

5,100

5,286

3,897

Black

3,571

5,809

7,202

3,750

4,828

4,759

Hispanic white

3,885

6,164

7,991

4,450

5,112

4,962

Joint

3,578

4,992

7,966

4,727

5,152

4,232

Non-Hispanic white

3,251

4,851

6,736

4,267

4,442

3,665

Other

3,657

5,415

7,168

4,292

4,749

4,469

Missing

3,747

5,114

7,152

4,395

4,809

4,281

Total

3,404

4,857

7,129

4,274

4,547

3,925

Age group

<=24

3,102

6,163

4,826

4,250

3,985

25-34

3,463

4,844

7,008

4,562

4,604

4,095

35-44

3,506

4,862

7,447

4,287

4,760

4,110

45-54

3,411

4,849

7,385

4,172

4,618

3,940

55-64

3,336

4,849

6,998

4,076

4,443

3,689

65-74

3,270

4,790

6,620

3,936

4,312

3,532

236

Total

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

>=75

3,186

4,709

6,580

4,505

3,487

Total

3,404

4,849

7,128

4,273

4,547

3,925

Neighborhood income

Low or moderate

3,400

4,983

6,788

3,982

4,503

4,042

Middle

3,331

4,789

6,976

4,169

4,500

3,880

High

3,483

4,865

7,732

4,548

4,766

3,928

Total

3,403

4,855

7,128

4,272

4,548

3,925

Geography

Metropolitan Area

3,448

4,861

7,291

4,350

4,823

3,979

Micropolitan Area

3,027

4,755

5,786

3,776

4,224

3,510

Rural

2,919

4,851

5,577

3,462

4,045

3,352

Total

3,404

4,857

7,129

4,274

4,547

3,925

NOTE: Site-built single-family closed-end originations. Cells with frequency counts (of valid non-missing values) less than 500 are omitted from the table.

237

TABLE 7.3.4: TOTAL LOAN COSTS AND POINTS AND FEES OF MANUFACTURED HOME LOANS ($)

Mean

Median

P5

P25

P75

P95

Manufactured home non-chattel

loans

Total loans costs

4,794

4,215

325

2,590

6,227

10,350

Manufactured home chattel loans

Total points and fees

1,736

1,548

70

749

2,622

3,601

NOTE: Manufactured home originations.

238

TABLE 7.4.1: DISTRIBUTION OF ORIGINATION CHARGES BY ENHANCED LOAN TYPE ($)

Origination charges

Mean

Median

P5

P25

P75

P95

SD

Enhanced loan type

Conventional

Conforming

1,868

1,250

0

750

2,310

5,882

2,773

Jumbo

2,503

1,150

0

745

1,910

10,447

4,595

Non-conventional

FHA

1,823

1,303

0

500

2,459

5,438

3,266

VA

1,550

794

0

2,164

6,026

3,391

RHS/FSA

1,424

1,199

0

765

1,861

3,600

2,322

Total

1,852

1,225

0

600

2,300

5,954

3,036

NOTE: Site-built single-family closed-end originations

239

TABLE 7.4.2: MEDIAN ORIGINATION CHARGES BY ENHANCED LOAN TYPE, LOAN PURPOSE, OCCUPANCY TYPE, AND LIEN STATUS ($)

Enhanced loan type

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Loan purpose

Home purchase

1,216

1,175

1,342

80

1,198

Home improvement

0

1,250

1,722

1,800

0

Other

10

815

1,693

2,501

100

NA

308

71

325

Non-cash-out refi

1,250

1,082

258

712

1,345

1,148

Cash-out refi

1,631

1,175

2,088

2,487

1,731

Total

1,250

1,150

1,303

794

1,199

1,225

Occupancy type

Principal Residence

1,245

1,155

1,303

795

1,199

1,208

Second Residence

1,215

995

1,195

Investment Property

1,827

1,430

1,309

295

1,800

Total

1,250

1,150

1,303

794

1,199

1,225

Lien status

First Lien

1,290

1,155

1,310

794

1,199

1,265

Subordinate Lien

0

495

1,005

0

Total

1,250

1,150

1,303

794

1,199

1,225

240

NOTE: Site-built single-family closed-end originations. Cells with frequency counts (of valid non-missing values) less than 500 are omitted from the table.

241

TABLE 7.4.3: MEDIAN ORIGINATION CHARGES BY ENHANCED LOAN TYPE, RACE/ETHNICITY, AGE, NEIGHBORHOOD INCOME AND

GEOGRAPHY ($)

Enhanced loan type

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Borrower race and ethnicity

Asian

1,290

1,175

1,295

645

1,195

1,250

Black

1,295

1,290

1,305

709

1,223

1,246

Hispanic white

1,445

1,290

1,485

729

1,420

1,419

Joint

1,276

1,155

1,305

692

1,295

1,210

Non-Hispanic white

1,195

1,140

1,290

750

1,194

1,185

Other

1,360

1,295

1,415

625

1,280

1,295

Missing

1,400

1,100

1,295

995

1,201

1,325

Total

1,250

1,150

1,303

794

1,199

1,225

Age group

<=24

1,095

1,276

0

1,161

1,095

25-34

1,205

1,155

1,285

188

1,197

1,190

35-44

1,285

1,175

1,295

595

1,230

1,245

45-54

1,290

1,155

1,350

895

1,240

1,252

55-64

1,274

1,090

1,435

1,000

1,240

1,255

65-74

1,250

1,000

1,580

1,251

1,220

1,257

242

Enhanced loan type

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

>=75

1,235

995

1,696

1,345

1,250

Total

1,250

1,151

1,303

794

1,199

1,225

Neighborhood income

Low or moderate

1,290

1,190

1,393

820

1,245

1,290

Middle

1,250

1,175

1,300

794

1,198

1,225

High

1,245

1,130

1,285

770

1,174

1,199

Total

1,250

1,150

1,302

792

1,198

1,225

Geography

Metropolitan Area

1,270

1,155

1,300

792

1,198

1,240

Micropolitan Area

1,140

1,060

1,306

773

1,200

1,145

Rural

1,090

1,015

1,353

829

1,195

1,115

Total

1,250

1,150

1,303

794

1,199

1,225

NOTE: Site-built single-family closed-end originations. Cells with frequency counts (of valid non-missing values) less than 500 are omitted from the table.

243

TABLE 7.5.1: RANGE OF DISCOUNT POINTS (EXPRESSED IN POINTS) BY ENHANCED LOAN TYPE (COUNTS IN THOUSANDS)

Enhanced Loan Type

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Discount points

0

3,330.5

65.7

293.1

81.3

706.0

63.9

501.6

61.6

64.1

64.9

4,895.3

65.7

(0 - 0.5)

781.7

15.4

43.7

12.1

158.2

14.3

130.9

16.1

15.6

15.8

1,130.1

15.2

[0.5 - 1)

410.4

8.1

14.6

4.0

110.1

10.0

77.9

9.6

9.9

10.0

622.9

8.4

[1 - 1.5)

207.2

4.1

5.3

1.5

52.3

4.7

38.0

4.7

4.1

306.9

4.1

[1.5 - 2)

152.4

3.0

2.3

0.6

41.7

3.8

44.3

5.4

2.9

3.0

243.6

3.3

[2 - 2.5)

91.1

1.8

0.9

0.2

19.2

1.7

11.3

1.4

1.0

123.5

1.7

[2.5 - 3)

57.0

1.1

0.5

0.1

12.2

1.1

6.4

0.8

0.7

0.8

76.7

1.0

[3 - 3.5)

21.5

0.4

0.2

0.1

3.5

0.3

2.1

0.3

0.2

27.5

0.4

[3.5 - 4)

9.3

0.2

0.1

0.0

1.5

0.1

1.0

0.1

12.0

0.2

>= 4

5.9

0.1

0.0

0.5

0.0

0.5

0.1

7.1

0.1

Total

5,067.0

100.0

360.7

100.0

1,105.3

100.0

814.0

100.0

98.9

100.0

7,445.8

100.0

NOTE: Site-built single-family closed-end originations, not primarily for a business or commercial purpose. Discount points are expressed in points relative to the

loan amount. The loan amounts used in the discount point calculations are from non-public raw data reported by financial institutions.

244

TABLE 7.5.2: RANGE OF DISCOUNT POINTS (EXPRESSED IN POINTS) BY RACE/ETHNICITY, AGE, NEIGHBORHOOD INCOME, AND

GEOGRAPHY (PERCENT)

Discount points

0

(0 - 0.5)

[0.5 - 1)

[1 - 1.5)

[1.5 - 2)

[2 - 2.5)

[2.5 - 3)

[3 - 3.5)

[3.5 - 4)

>= 4

Total

Borrower race and

ethnicity

Asian

70.3

15.0

7.6

3.3

2.0

0.9

0.6

0.2

0.1

100.0

Black

62.4

14.2

9.5

5.0

4.5

2.1

1.3

0.5

0.2

0.3

100.0

Hispanic white

62.5

16.5

9.9

4.7

3.3

1.5

0.9

0.3

0.2

100.0

Joint

66.3

15.9

8.5

4.0

2.9

1.2

0.7

0.2

0.1

100.0

Non-Hispanic white

67.7

15.2

7.9

3.7

2.8

1.4

0.9

0.3

0.1

100.0

Other

60.0

16.2

10.0

5.2

4.3

2.0

1.4

0.5

0.2

0.1

100.0

Missing

59.7

14.6

9.2

5.5

5.1

3.0

1.9

0.7

0.3

0.1

100.0

Total

65.7

15.2

8.4

4.1

3.3

1.7

1.0

0.4

0.2

0.1

100.0

Age group

<=24

72.0

14.7

7.6

2.9

1.7

0.6

0.4

0.1

100.0

25-34

69.1

15.7

8.0

3.3

2.1

0.9

0.5

0.2

0.1

100.0

35-44

67.0

15.4

8.3

3.9

2.8

1.3

0.8

0.3

0.1

100.0

45-54

64.7

15.1

8.7

4.4

3.5

1.8

1.1

0.4

0.2

0.1

100.0

55-64

62.8

14.7

8.6

4.8

4.3

2.4

1.5

0.6

0.2

0.1

100.0

65-74

60.2

14.6

8.6

5.1

5.4

3.0

1.9

0.7

0.3

0.2

100.0

>=75

59.5

14.1

8.6

5.2

5.8

3.3

2.1

0.8

0.4

0.2

100.0

245

Discount points

0

(0 - 0.5)

[0.5 - 1)

[1 - 1.5)

[1.5 - 2)

[2 - 2.5)

[2.5 - 3)

[3 - 3.5)

[3.5 - 4)

>= 4

Total

65.7

15.2

8.4

4.1

3.3

1.7

1.0

0.4

0.2

0.1

100.0

Neighborhood income

Low or moderate

63.3

14.7

9.0

4.7

3.9

2.1

1.3

0.5

0.2

100.0

Middle

64.6

15.1

8.7

4.3

3.6

1.8

1.2

0.4

0.2

0.1

100.0

High

68.0

15.4

7.8

3.7

2.7

1.3

0.8

0.3

0.1

100.0

Total

65.7

15.2

8.4

4.1

3.3

1.7

1.0

0.4

0.2

0.1

100.0

Geography

Metropolitan Area

65.8

15.3

8.4

4.1

3.2

1.6

1.0

0.4

0.2

0.1

100.0

Micropolitan Area

65.3

14.3

8.3

4.3

3.8

2.0

1.3

0.5

0.2

0.1

100.0

Rural

66.0

13.4

8.0

4.2

3.8

2.2

1.4

0.5

0.2

0.1

100.0

Total

65.7

15.2

8.4

4.1

3.3

1.7

1.0

0.4

0.2

0.1

100.0

NOTE: Site-built single-family closed-end originations, not primarily for a business or commercial purpose. Discount points are expressed in points relative to the

loan amount. The loan amounts used in the discount point calculations are from non-public raw data reported by financial institutions.

246

TABLE 7.5.3: RANGE OF LENDER CREDITS (EXPRESSED IN POINTS) BY ENHANCED LOAN TYPE (COUNTS IN THOUSANDS)

Enhanced Loan Type

Conventional

Non-conventional

Total

Conforming

Jumbo

FHA

VA

RHS/FSA

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Lender credits

0

3,256.4

64.3

175.2

48.6

618.9

56.0

484.1

59.5

60.1

60.7

4,594.6

61.7

(0 - 0.5)

1,338.7

26.4

158.6

44.0

291.7

26.4

261.6

32.1

29.9

30.2

2,080.5

27.9

[0.5 - 1)

268.0

5.3

21.0

5.8

84.5

7.6

41.3

5.1

4.8

4.9

419.6

5.6

[1 - 1.5)

106.3

2.1

4.1

1.1

50.3

4.5

16.4

2.0

1.9

179.0

2.4

[1.5 - 2)

50.6

1.0

1.3

0.4

27.7

2.5

6.2

0.8

1.0

86.8

1.2

[2 - 2.5)

23.9

0.5

0.3

0.1

14.7

1.3

2.4

0.3

0.6

41.9

0.6

[2.5 - 3)

10.0

0.2

0.1

0.0

8.8

0.8

1.2

0.1

0.4

20.5

0.3

[3 - 3.5)

4.9

0.1

0.0

4.3

0.4

0.1

9.8

0.1

[3.5 - 4)

2.7

0.1

0.0

2.1

0.2

0.0

5.0

0.1

>= 4

5.5

0.1

0.0

2.3

0.2

0.0

8.0

0.1

Total

5,067.0

100.0

360.7

100.0

1,105.3

100.0

814.0

100.0

98.9

100.0

7,445.8

100.0

NOTE: Site-built single-family closed-end originations, not primarily for a business or commercial purpose. Lender credits are expressed in points relative to the

loan amount. The loan amounts used in the lender credits calculations are from non-public raw data reported by financial institutions.

247

TABLE 7.5.4: RANGE OF LENDER CREDITS (EXPRESSED IN POINTS) BY RACE/ETHNICITY, AGE, NEIGHBORHOOD INCOME, AND GEOGRAPHY

(PERCENT)

Lender credits

0

(0 - 0.5)

[0.5 - 1)

[1 - 1.5)

[1.5 - 2)

[2 - 2.5)

[2.5 - 3)

[3 - 3.5)

[3.5 - 4)

>= 4

Total

Borrower race and

ethnicity

Asian

52.7

31.8

9.6

3.4

1.4

0.6

0.3

0.1

100.0

Black

60.1

27.2

6.2

3.0

1.6

0.8

0.5

0.3

0.1

0.3

100.0

Hispanic white

62.6

26.3

5.3

2.7

1.5

0.8

0.4

0.2

0.1

0.2

100.0

Joint

60.9

29.7

5.5

2.2

1.0

0.4

0.2

0.1

0.0

100.0

Non-Hispanic white

62.2

28.2

5.2

2.2

1.1

0.5

0.2

0.1

100.0

Other

61.9

27.4

5.8

2.5

1.2

0.5

0.3

0.2

0.1

0.2

100.0

Missing

63.4

26.2

5.7

2.5

1.2

0.5

0.3

0.1

100.0

Total

61.7

27.9

5.6

2.4

1.2

0.6

0.3

0.1

100.0

Age group

<=24

62.5

27.3

5.2

2.2

1.2

0.7

0.4

0.2

0.1

0.2

100.0

25-34

59.9

28.8

6.1

2.6

1.3

0.7

0.3

0.1

100.0

35-44

59.9

29.0

6.1

2.6

1.2

0.6

0.3

0.1

100.0

45-54

62.0

27.8

5.6

2.4

1.1

0.5

0.3

0.1

100.0

55-64

64.2

26.6

5.1

2.1

1.0

0.5

0.2

0.1

100.0

65-74

65.3

26.3

4.6

1.9

0.9

0.5

0.2

0.1

100.0

>=75

65.5

26.2

4.5

1.9

0.9

0.4

0.2

0.1

100.0

248

Lender credits

0

(0 - 0.5)

[0.5 - 1)

[1 - 1.5)

[1.5 - 2)

[2 - 2.5)

[2.5 - 3)

[3 - 3.5)

[3.5 - 4)

>= 4

Total

61.7

27.9

5.6

2.4

1.2

0.6

0.3

0.1

100.0

Neighborhood income

Low or moderate

62.5

25.9

5.6

2.7

1.4

0.8

0.4

0.3

0.1

0.3

100.0

Middle

63.0

27.0

5.4

2.3

1.1

0.6

0.3

0.1

100.0

High

59.9

29.8

5.9

2.4

1.1

0.5

0.2

0.1

0.0

100.0

Total

61.7

28.0

5.6

2.4

1.2

0.6

0.3

0.1

100.0

Geography

Metropolitan Area

61.3

28.1

5.8

2.5

1.2

0.6

0.3

0.1

100.0

Micropolitan Area

65.7

26.7

4.2

1.7

0.8

0.4

0.2

0.1

100.0

Rural

66.4

26.0

4.2

1.6

0.8

0.4

0.3

0.1

100.0

Total

61.7

27.9

5.6

2.4

1.2

0.6

0.3

0.1

100.0

NOTE: Site-built single-family closed-end originations, not primarily for a business or commercial purpose. Lender credits are expressed in points relative to the

loan amount. The loan amounts used in the lender credits calculations are from non-public raw data reported by financial institutions.

249

TABLE 7.5.5: RANGE OF DISCOUNT POINTS (EXPRESSED IN POINTS) BY RANGE OF LENDER CREDIT (EXPRESSED IN POINTS) (COUNTS IN

THOUSANDS)

Lender credits

0

(0 - 0.5)

[0.5 - 1)

[1- 1.5)

[1.5 - 2)

[2 - 2.5)

Count

%

Count

%

Count

%

Count

%

Count

%

Count

%

Discount points

0

2,713.3

55.4

1,516.1

31.0

364.9

7.5

157.6

3.2

77.0

1.6

35.9

0.7

(0 - 0.5)

808.9

71.6

283.7

25.1

21.7

1.9

7.5

0.7

3.3

0.3

2.0

0.2

[0.5 - 1)

456.5

73.3

137.8

22.1

17.4

2.8

5.2

0.8

2.3

0.4

1.4

0.2

[1 - 1.5)

228.3

74.4

62.1

20.2

7.3

2.4

4.6

1.5

1.6

0.5

1.1

0.3

[1.5 - 2)

190.1

78.0

41.8

17.2

4.6

1.9

2.4

1.0

1.7

0.7

0.8

0.3

[2 - 2.5)

99.5

80.6

18.9

15.3

1.8

1.4

1.0

0.8

0.5

0.4

[2.5 - 3)

61.6

80.3

12.3

16.1

1.1

1.4

0.5

0.6

0.3

0.4

0.2

[3 - 3.5)

22.0

80.1

4.3

15.5

0.5

1.8

0.2

0.6

0.1

0.3

0.1

0.3

[3.5 - 4)

9.2

77.0

2.2

18.1

0.3

2.1

0.1

0.7

0.0

0.3

0.0

0.2

>= 4

5.1

71.4

1.4

19.9

0.2

2.9

0.1

0.9

0.0

0.6

0.0

0.3

Total

4,594.6

61.7

2,080.5

27.9

419.6

5.6

179.0

2.4

86.8

1.2

41.9

0.6

NOTE: Site-built single-family closed-end originations, not primarily for a business or commercial purpose. Discount points and lender credits are expressed in

points relative to the loan amount. The loan amounts used in the discount point and lender credit calculations are from non-public raw data reported by financial

institutions.

250

TABLE 7.5.5: RANGE OF DISCOUNT POINTS (EXPRESSED IN POINTS) BY RANGE OF LENDER CREDIT (EXPRESSED IN POINTS) (COUNTS IN

THOUSANDS)

continued

Lender credits

[2.5 - 3)

[3 - 3.5)

[3.5 - 4)

>= 4

Total

Count

%

Count

%

Count

%

Count

%

Count

%

Discount points

0

15.5

0.3

7.0

0.1

3.2

0.1

4.8

0.1

4,895.3

100.0

(0 - 0.5)

1.6

0.1

0.6

0.1

0.4

0.0

0.5

0.0

1,130.1

100.0

[0.5 - 1)

1.1

0.2

0.5

0.1

0.2

0.0

0.4

0.1

622.9

100.0

[1 - 1.5)

0.9

0.3

0.6

0.2

0.3

0.1

0.4

0.1

306.9

100.0

[1.5 - 2)

0.8

0.3

0.6

0.2

0.4

0.2

0.5

0.2

243.6

100.0

[2 - 2.5)

0.3

0.2

0.3

0.2

0.3

0.2

0.5

0.4

123.5

100.0

[2.5 - 3)

0.2

0.4

0.5

76.7

100.0

[3 - 3.5)

0.0

0.1

0.2

0.1

0.2

0.9

27.5

100.0

[3.5 - 4)

0.0

0.2

0.0

0.2

0.0

0.2

0.1

1.2

12.0

100.0

>= 4

0.0

0.3

0.0

0.4

0.0

0.2

3.2

7.1

100.0

Total

20.5

0.3

9.8

0.1

5.0

0.1

8.0

0.1

7,445.8

100.0

NOTE: Site-built single-family closed-end originations, not primarily for a business or commercial purpose. Discount points and lender credits are expressed in

points relative to the loan amount. The loan amounts used in the discount point and lender credit calculations are from non-public raw data reported by financial

institutions.

251

252

APPENDIX B: FIGURES

FIGURE 3.1.1 : DENIAL RATE BY APPLICANT AGE

NOTE: Site-built single-family, principal residence, first lien applications (excluding applications that were withdrawn or incomplete).

253

FIGURE 6.4.1 HISTOGRAM OF CREDIT SCORES BY ENHANCED LOAN TYPE: ORIGINATED LOANS ONLY

NOTE: Site-built single-family originations. The vertical reference line represents a credit score of 620.

254

FIGURE 6.4.2 HISTOGRAM OF CREDIT SCORES BY ENHANCED LOAN TYPE: APPLICATIONS

NOTE: Site-built single-family applications. The vertical reference line represents a credit score of 620.

255

FIGURE 6.4.3.1 HISTOGRAM OF CREDIT SCORES BY RACE AND ETHNICITY: CONVENTIONAL CONFORMING APPLICATIONS

NOTE: Site-built single-family, closed-end conventional conforming applications. The vertical reference line represents a credit score of 620.

256

FIGURE 6.4.3.2 HISTOGRAM OF CREDIT SCORES BY RACE AND ETHNICITY: JUMBO APPLICATIONS

NOTE: Site-built single-family, closed-end conventional jumbo applications. The vertical reference line represents a credit score of 620.

257

FIGURE 6.4.3.3 HISTOGRAM OF CREDIT SCORES BY RACE AND ETHNICITY: FHA APPLICATIONS

NOTE: Site-built single-family, closed-end FHA applications. The vertical reference line represents a credit score of 620.

258

FIGURE 6.4.3.4 HISTOGRAM OF CREDIT SCORES BY RACE AND ETHNICITY: VA APPLICATIONS

NOTE: Site-built single-family, closed-end VA applications. The vertical reference line represents a credit score of 620.

259

FIGURE 6.4.3.5 HISTOGRAM OF CREDIT SCORES BY RACE AND ETHNICITY: RHS/FSA APPLICATIONS

NOTE: Site-built single-family, closed-end RHS/FSA applications. The vertical reference line represents a credit score of 620.

260

FIGURE 6.4.3.6 HISTOGRAM OF CREDIT SCORES BY RACE AND ETHNICITY: HELOC APPLICATIONS

NOTE: Site-built single-family, HELOC applications. The vertical reference line represents a credit score of 620.

261

FIGURE 6.4.4 DENIAL RATE BY CREDIT SCORE

NOTE: Site-built single-family, principal residence, first-lien applications (excluding applications that were withdrawn or incomplete).

262

FIGURE 6.4.5 DENIAL RATE BY CREDIT SCORE: CONVENTIONAL CONFORMING HOME-PURCHASE, 30-YEAR FIXED RATE APPLICATIONS

NOTE: Site-built single-family, closed-end, principal residence, first-lien, 30-year term, fixed-rate, conventional conforming applications (excluding applications

that were withdrawn or incomplete), with CLTV<=120.

263

FIGURE 6.4.6 CLTV BY CREDIT SCORE: CONVENTIONAL CONFORMING HOME-PURCHASE, 30-YEAR FIXED RATE APPLICATIONS

NOTE: Site-built single-family, closed-end, principal residence, first-lien, 30-year term, fixed-rate, conventional conforming applications (excluding applications

that were withdrawn or incomplete), with CLTV<=120.

264

FIGURE 6.5.1A HISTOGRAM OF CLTV: CONVENTIONAL CONFORMING HOME-PURCHASE LOANS

NOTE: Site-built single-family closed-end conventional conforming, home-purchase originations. The vertical reference line represents CLTV equal to 80%.

265

FIGURE 6.5.1B HISTOGRAM OF CLTV: CONVENTIONAL CONFORMING REFINANCE LOANS

NOTE: Site-built single-family closed-end conventional conforming, refinance originations. The vertical reference line represents CLTV equal to 80%.

266

FIGURE 6.5.2A HISTOGRAM OF CLTV: JUMBO HOME-PURCHASE LOANS

NOTE: Site-built single-family closed-end jumbo, home-purchase originations. The vertical reference line represents CLTV equal to 80%.

267

FIGURE 6.5.2B HISTOGRAM OF CLTV: JUMBO REFINANCE LOANS

NOTE: Site-built single-family closed-end jumbo, refinance originations. The vertical reference line represents CLTV equal to 80%.

268

FIGURE 6.5.3A HISTOGRAM OF CLTV: FHA HOME-PURCHASE LOANS

NOTE: Site-built single-family closed-end FHA, home-purchase originations. The vertical reference line represents CLTV equal to 96.5%.

269

FIGURE 6.5.3B HISTOGRAM OF CLTV: FHA REFINANCE LOANS

NOTE: Site-built single-family closed-end FHA, refinance originations. The vertical reference line represents CLTV equal to 96.5%.

270

FIGURE 6.5.4A HISTOGRAM OF CLTV: VA HOME-PURCHASE LOANS

NOTE: Site-built single-family closed-end VA, home-purchase originations. The vertical reference line represents CLTV equal to 100%.

271

FIGURE 6.5.4B HISTOGRAM OF CLTV: VA REFINANCE LOANS

NOTE: Site-built single-family closed-end VA, refinance originations. The vertical reference line represents CLTV equal to 100%.

272

FIGURE 6.5.5A HISTOGRAM OF CLTV: RHS/FSA HOME-PURCHASE LOANS

NOTE: Site-built single-family closed-end RHS/FSA, home-purchase originations. The vertical reference line represents CLTV equal to 100%.

273

FIGURE 6.5.5B HISTOGRAM OF CLTV: RHS/FSA REFINANCE LOANS

NOTE: Site-built single-family closed-end RHS/FSA, refinance originations. The vertical reference line represents CLTV equal to 100%.

274

FIGURE 6.5.6 HISTOGRAM OF CLTV: HELOC

NOTE: Site-built single-family HELOC originations. The vertical reference line represents CLTV equal to 80%.

275

FIGURE 6.6.1 HISTOGRAM OF DTI: CONVENTIONAL CONFORMING LOANS

NOTE: Site-built single-family closed-end conventional conforming originations. The three vertical reference lines represent DTI equal to 43%, 45%, and 50%,

respectively.

276

FIGURE 6.6.2 HISTOGRAM OF DTI: JUMBO LOANS

NOTE: Site-built single-family closed-end jumbo originations. The three vertical reference lines represent DTI equal to 43%, 45%, and 50%, respectively.

277

FIGURE 6.6.3 HISTOGRAM OF DTI: FHA LOANS

NOTE: Site-built single-family closed-end FHA originations. The four vertical reference lines represent DTI equal to 43%, 45%, 50%, and 57%, respectively.

278

FIGURE 6.6.4 HISTOGRAM OF DTI: VA LOANS

NOTE: Site-built single-family closed-end VA originations. The two vertical reference lines represent DTI equal to 45% and 50% respectively.

279

FIGURE 6.6.5 HISTOGRAM OF DTI: RHS/FSA LOANS

NOTE: Site-built single-family closed-end RHS/FSA originations. The two vertical reference lines represent DTI equal to 40% and 50% respectively.

280

FIGURE 6.6.6 HISTOGRAM OF DTI: HELOC

NOTE: Site-built single-family HELOC originations. The four vertical reference lines represent DTI equal to 43%, 45%, 50%, and 55%, respectively.

281

FIGURE 6.6.7 DENIAL RATE BY DTI

NOTE: Site-built single-family, principal residence, first-lien applications (excluding applications that were withdrawn or incomplete). The sample is limited to

DTI>=0 and DTI<=100%.