Skip to main content

Just Released: Analysis of Housing Finance Issues Using the American Housing Survey

HUD.GOV HUDUser.gov
eList
AHS

The announcement below was posted on the HUD USER announcement mailing list. I'm crossposting it here because it concerns the AHS. My apologies to those of you who subscribe to both mailing lists.

---------------------------------------------------------------------------------------------------------------

From: HUD USER News

The biannual American Housing Survey (AHS) national
sample provides housing researchers with robust and
reliable data on household size, location, tenure,
housing costs, and other variables. But while the AHS is
one of the most useful housing datasets available, data
collected about mortgages are generally underused in
research because the reliability of these variables has
not been verified.

A recent study commissioned by HUD's Office of Policy
Development and Research (PD&R) titled "Analysis of
Housing Finance Issues Using the American Housing Survey"
analyzes the validity of these mortgage-related
variables. The authors analyze the extent to which sample
size, survey design, and interview response affect the
accuracy and consistency of the mortgage data contained
in the AHS.

The study used a replication and a longitudinal analysis
to test the reliability of these and other mortgage-
related variables. The replication analysis was performed
using benchmark data sets from a number of sources,
including Home Mortgage Disclosure Act (HMDA) data and
Residential Finance Survey (RFS) data, just to name a
few. The longitudinal analysis examined the internal
reliability of AHS variables over time. In both
analyses, the authors emphasize the importance of
recognizing the sampling variability of estimates
obtained from the AHS.

The analysis showed that AHS market measures are
generally accurate for volume estimates or mortgage
origination volume. The AHS also tended to report
underestimated refinance loan origination amounts and
understated amounts for ARM mortgage payment product
type.

Internally, several inconsistencies were found between
data reported over the years. The authors expressed
concern that the accuracy of owners' responses declines
over time. The authors found that, for most variables,
between 80 and 90 percent of respondents reported most
mortgage information consistently over time, but
information on origination amount, monthly payment
amount, and interest rate information was far less
reliable, with only 40 percent of respondents reporting
interest rate data consistently.

The report provides a list of AHS mortgage variables that
are reliable. These include:

o Mortgage Originations for Home-Purchase Loans;
o Volume and Share by Market Segment;
o Original Mortgage Amount;
o Interest Rate;
o First-time homebuyer; and
o Borrower Race/Ethnicity.

The following variables were found to be unreliable or
uncertain:

o Refinances;
o Jumbo Loans;
o Payment Product Types;
o Loan-to-Value Ratios; and
o Borrower Income.

The study concludes with recommendations for
modifications to the survey, such as new questions about
the number of recent refinances and the mortgage payment
status to capture delinquency or default.

"Analysis of Housing Finance Issues Using the American
Housing Survey" is available for download from HUD USER
at
https://www.huduser.gov/portal/datasets/ahs.html
or in printed form for a nominal charge by calling HUD
USER at 1-800-245-2691.
--------------------------------------
Please contact HUD USER at:
HUD USER
P.O. Box 23268
Washington, DC 20026-3268
1-800-245-2691
1-800-927-7589 (TDD)
202-708-9981 (fax)
--------------------------------------

Dav Vandenbroucke
Economist
U.S. Dept. HUD
david_a._vandenbroucke@hud.gov
202-708-1060 ext. 5890