HMDA Data

1
HMDA Data

• Robert B. Avery
• Association of Public Data Users Annual
  Conference 2008
• September 24, 2008
• This presentation reflects the joint work of
  Robert B. Avery, Kenneth P. Brevoort and Glenn B.
  Canner. The views expressed here are those of
  the author and do not necessarily represent those
  of the Board of Governors of the Federal Reserve
  System or members of the staff.

2
HMDA Background

• The Home Mortgage Disclosure Act (HMDA) of 1975
  enacted to address concerns about redlining
  practices
• Requires most lenders in metropolitan areas to
  disclose information about their mortgage lending
  activity each year by census tract and type of
  loan.
• HMDA amended by Congress in 1989 to add
  application-by-application disclosure of
• Race, ethnicity and sex of applicants
• Borrower income
• Disposition of applications (approval, denial,
  etc.) and reasons for denial
• Loan size, type of loan, and census tract
• Information on the sale of the loan in the
  secondary market

3
HMDA Price Disclosures

• Federal Reserve amended Regulation C in 2002 to
  require disclosure of information about loans
  with prices above designated thresholds such
  loans are referred to as higher-priced loans
  but are not equivalent necessarily to subprime.
• What is disclosed?
• Rate spread in percentage points
• Spread is difference between APR on loan and
  yield on Treasury security of comparable
  maturity
• Spread reported for first lien loans if above 3
  percentage points
• Spread reported for subordinate lien loans if
  above 5 percentage points
• Lien status (first, subordinate, no lien)

4
The 2007 HMDA Data

• For 2007 8,610 lenders reported on roughly 21
  million loan applications down about 25 percent
  from 2006
• Reporters include 6,858 banking institutions and
  1,752 mortgage companies -- including 1,124
  independent mortgage companies
• Number of reporters fell about 3 percent
  virtually all the decline due to independent
  mortgage companies that ceased operations

5
FHA and Nonowner-Occupant Lending

• FHA-lending
• FHA share of market fell from 16 percent as
  recently as 2000 to only 3 percent in 2006. FHA
  share rebounded to 5 percent in 2007 and
  indications are it is growing rapidly in 2008
• Growth of nonowner-occupied lending
• In mid-1990s share was 5 percent, reached peak in
  2006 at 17 percent share fell to 15 percent in
  2007
• Important in current market because investors are
  quick to walk away from property when values fall

6
Piggyback Lending

• In recent years, Piggyback loans were a popular
  financing arrangement for homebuyers seeking to
  avoid PMI and minimize downpayment
• Volume of piggys fell 60 percent from 2006 to
  2007
• From 2006 to 2007 the use of piggys changed more
  were used to keep the first lien loan within the
  GSE loan limits and fewer used as a substitute
  for PMI or as part of subprime 80-10-10 loans

7
Incidence of Higher-Priced Lending

• Incidence of higher-priced lending has changed
  over time
• 
  Year Percent
• 2004
  16
• 2005
  26
• 2006
  29
• 2007
  18
• Change in incidence may be caused by
• Artifact of how Regulation C determines which
  loans are higher-priced in combination with the
  flattening of the yield curve
• Changing risk profiles of borrowers
• Changing business practices or risk tolerances of
  lenders
• Nonreporting by lenders ceasing operations to the
  extent they were more focused on the
  higher-priced segment

8
Who Extends Higher-Priced Loans

• Share of higher-priced lending accounted for by
  depositories rose sharply in 2007 as independent
  mortgage company share fell

9
Effects of Closed Lenders
Loan Counts by Month, 2006 and 2007
10
Percentage Change in Home Purchase Lending by
Race or Ethnicity, 2006. H1 to 2007.H2

• Lending to Blacks and Hispanic whites fell more
  than to non-Hispanic whites
• Racial pattern holds across all income groups
• Falloff in lending to Blacks driven by decline in
  higher-priced lending

11
Percentage Changes in Lending by Type of Lender

• Independent mortgage companies experienced
  greater declines than depository institutions
  regardless of location for all types of lending

12
House Price Changes and Changes in Lending

• Falloff in lending activity is related to
  patterns of house price changes in preceding
  years
• MSAs that had larger declines in prices from
  2006.Q4 to 2008.Q1 experienced larger declines in
  lending
• Falloff in activity was greater for MSAs that had
  experienced sharp increases in home prices from
  2003.Q1 to 2006.Q4
• Lending declined 53 percent in MSAs with sharp
  declines preceded by sharp increases, compared to
  a 5 percent decline in MSAs with sharp declines
  preceded by smaller appreciation

13
Change in House Price Index from December 2006 to
March 2008
14
Percent Change in Lending from the First Half of
2006 to the Second Half of 2007
15
Incidence of Higher-priced Lending by Race or
Ethnicity, 2nd Half of 2006, 2007

• Wide differences in the incidence of higher
  priced lending across groups blacks and Hispanic
  whites have elevated rates

16
Denial rates by Race and Ethnicity, 2nd Half of
2006, 2007

• Wide differences in denial rates blacks,
  Hispanics have elevated rates of loan denial

17
Modifying HMDA to be Consistent with New HOEPA
Rules

• Definition of higher-priced lending in HOEPA
• APR spread for each loan calculated by comparing
  APR on loan with mortgage rates for lowest risk
  prime borrowers
• Prime rates from weekly Freddie Mac PMMS
• Loan products not included in PMMS are based on a
  an interpolation of the spread differences
  between PMMS rate on similar products and
  Treasury yield and adding the interpolated spread
  to the comparable Treasury yield
• Starting date for the new rule