New Trends in U.S.

Mortgage
Origination and Possible Concerns
National Association of Realtors Conference
June 9, 2017
Nancy Wallace
Haas School of Business
 Startling Fact
Majority of U.S. mortgages are now originated by
non-depositories (not banks or credit unions).

Independent mortgage companies market share has

risen:
14% 38% (07-15).
Independent mortgage companies relying on
algorithmic underwriting (fintech) market share has
risen:
4% 13% (07-15).
Source: HMDA, SEC filings
 Mortgage Origination Loan Amounts by
Originator Type
Independent Mort. Co. (Fintech) Independent Mort. Co. Depositories
900
800
700
600
500
400
300
200
100
0
2007 2010 2015
Source: HMDA
Importance of Non-Depository Origination
for GSE and GNMA Securitization
FNMA FHMC GNMA
80%

70%
% of Securitization Volume

60%

50%

40%

30%

20%

10%

0%
Depositories 2015 Depositories Non Depositories Non Depositories
2016(Q1) 2015 2016(Q1)

Source: Inside Mortgage Finance

 GSE, FHA/VA Originations as Percent of
Total U.S. Origination
VA Loans FHA Endorsements
GSE New MBS GSE, FHA/VA to Total
2,000 120%
1,800
1,600 100%
1,400 80%
$ Billions

1,200

Percent
1,000 60%
800
600 40%
400 20%
200
0 0%
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
Source: Inside Mortgage Finance
U.S. MORTGAGE
ORIGINATION FUNDING
STRUCTURES
 Flow of Mortgage Origination is Not Highly
Correlated with Deposit Flow
Home Mortgage: Flow (F217) Time and savings deposits: Flow (F205)
1200

1000

800

600
$ Billions

400

200

0
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
-200

-400
Source: Flow of Funds
Source: Federal Reserve Flow of Funds
 8-K Recorded MRA filings and Non-
Depository Origination Share
100 45%
Passage of BAPCPA 2005

Percentage of Originations by Non-Depositories

90
8K Filings for Master Repurchase Agreements

40%

80
35%
70
30%
60
25%
50
20%
40
15%
30
10%
20

10 5%

0 0%
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Source: HMDA, SEC Edgar

 Pre-Crisis Outcomes for Top Forty U.S.
Mortgage Originators in 2006
Firm Type Originations % of Firm % of Firm
as % of Originations Failures1.
Total with MRAs
Commercial banks 38.0% 0.0% 0.0%
Federal Savings 1.9% 60.6% 66.7%
Banks
Savings and Loans 29.0% 64.8% 100.0%
Affiliated Mortgage 12.7% 100.0% 89.0%
Companies
Independent 3.5% 84.0% 66.7%
Mortgage Companies
Real Estate 10.7% 100.0% 100.0%
Investment Trusts
1. Supervisory closure, Chapter 11 closure, distressed closure.
Source: Stanton, Walden, and Wallace, 2014.
 Systemic Risks
Repo buyer runs rapid loss of mortgage
origination capacity, mortgage fire sales.
Unfunded rep and warranty guarantees risks
to GNMA/GSEs.
Un-priced liquidity provided by GNMA/GSEs
repo is a bet on securitization speeds.
POST CRISIS TRENDS IN
MORTGAGE UNDERWRITING:
EXTREMELY TIGHT CREDIT
BOX
 70
72
74
76
78
80
82
84
86
2005Q1
2005Q3
2006Q1
2006Q3
2007Q1

Source: Inside Mortgage Finance

2007Q3
2008Q1
2008Q3
2009Q1
Fannie Purchase

2009Q3
2010Q1
2010Q3
2011Q1
2011Q3
2012Q1
2012Q3
Freddie Purchase

2013Q1
2013Q3
2014Q1
2014Q3
GSE Underwriting Characteristics: LTV

2015Q1
 710
720
730
740
750
760
770
2005Q1
2005Q3
2006Q1
2006Q3
2007Q1
2007Q3

Source: Inside Mortgage Finance

2008Q1
2008Q3
2009Q1
2009Q3
FNMA Purchase

2010Q1
2010Q3
2011Q1
2011Q3
2012Q1
2012Q3
2013Q1
FHMC Purchase

2013Q3
2014Q1
2014Q3
2015Q1
2016Q1
GSE Underwriting Characteristics: FICO

2017Q1
 GSE Underwriting Characteristics: DTI

Source: Inside Mortgage Finance

 GNMA and GSE Average Loan Amount
GNMA GSE
350000

300000

250000

200000

150000

100000

50000

0
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
Source: Dataquick, McDash LLC
 GSE and GNMA Origination Coupons
GNMA GSE
0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
Source: Dataquick, McDash
 GNMA and GSE Credit Scores
GNMA GSE
780
760
740
720
700
680
660
640
620
600
580
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

Source: Dataquick, McDash

 Summary on Underwriting Standards
Mortgage lending standards are tight by historical
standards.
FICO scores of 750 (based on 2009 census)
correspond to an income of greater than $120,000.
Median annual income of U.S. homeowners is about
$60,000.
Recent study (Elwell, 2014) finds in 2012 that a
reasonable definition of the middle class is as annual
income greater than $39,736 and less than $104,087.
Credit overlays (Loan Level Price Adjustments
LLPAs) of the GSEs are too stringent.
EFFECT OF ALGORITHMIC
UNDERWRITING ON THE
AVAILABILITY OF
MORTGAGE CREDIT
 Lending Club Disclosed Risk
The Lending Club platform is a novel approach to
borrowing that may fail to comply with borrower
protection laws such as state usury laws, other
interest rate limitation or federal and state
consumer protection laws such as the Truth in
Lending Act, the Equal Credit Opportunity Act, the
Fair Credit Reporting Act and the Fair Debt
Collection Practices Act and their state
counterparts

Prospectus Supplement (Disclosure Report No. 2) (to Prospectus dated

October 13, 2008) Lending Club
 Racial Distribution of Originated Loans by
Lender Type and Year
1

0.8

0.6

0.4

0.2

0
2007 2010 2015 2007 2010 2015 2007 2010 2015
Independent Mort. Co. Independent Mort. Co. (Non- Depositories
(Fintech) Fintech)
Black/Hispanic White/Asian Unknown

Source: HMDA
 Mortgage Application Decision for Below Median
Income (Within County) Applicants
1

0.8

0.6

0.4

0.2

0
2007 2010 2015 2007 2010 2015 2007 2010 2015
Fintech Non-Fintech Traditional
Approved (Not Accepted) Denied Originated Withdrawn/Other

Source: HMDA
 Mortgage Application Decision for Below Median
Income White/Asian (Within County) Applicants
1

0.8

0.6

0.4

0.2

0
2007 2010 2015 2007 2010 2015 2007 2010 2015
Fintech Non-Fintech Traditional
Approved (Not Accepted) Denied Originated Withdrawn/Other

Source: HMDA
 Application Decision for Below Median Income
Black/Hispanic (Within County) Applicants
1

0.8

0.6

0.4

0.2

0
2007 2010 2015 2007 2010 2015 2007 2010 2015
Fintech Non-Fintech Traditional
Approved (Not Accepted) Denied Originated Withdrawn/Other

Source: HMDA
 Innovations in Algorithmic Underwriting
Consumer protection must be at the forefront of
innovations in underwriting technology.
Streamlining consumer facing underwriting
technology using algorithms, novel credit
scrapping technologies, and big data analytics:
Promise of significant benefits for consumers,
Potential for significant risks opaque nature of big
data, nature of algorithmic objective functions,
potential for cyber redlining.
 Conclusion
Pervasive reliance on repo funding for mortgage origination
poses potentially significant risks.
Loss of access to credit markets in rising interest rate environment --
runs.
Counter party risk to GSEs and GNMA from poorly capitalized
originators.
Vulnerability of servicing positions.
Credit box remains tight despite growth in 97% programs at
GSEs.
Improvements due to GSE 2013 sunset restrictions on putback liability.
Remains risk of False Claims Act (treble damages) liabilities.
Especially a problem for the middle class, minorities,
Need for careful evaluation of algorithmic underwriting and
the rise of the 8 minute mortgage.