House Price Forecasts and Their Complex Relationship with Interest Rates February 15, 2006

1
House Price Forecasts and Their Complex
Relationship with Interest Rates February 15,
2006
Fidelity Hansen Qualitys Analytics Group
Mike Sklarz (Head) Jim Follain (SVP for Mortgage
Valuation) Carl Bonham and Norm Miller
(Consultants)
2
Motivation and Purpose of Presentation

• Motivation We sense great interest in
  disciplined, third-party, and transparent
  processes to forecast house prices at local
  levels
• Investors seek insight about their exposure
• Experts acknowledge the critical role of local
  factors
  
• IT and data innovations make more readily
  available
  
• Purpose Discuss Fidelitys approach
• Provide examples of both MSA and zip code
  forecasts
  
• Highlight the complex relationship between
  interest rates and house price forecasts
  
• Changes in core rates impact house price
  forecasts
  
• Nature of house price forecasts affect risk-based
  pricing of credit risk
  

3
Key Themes

• Positive house price appreciation is still the
  best estimate for many markets
  
• Flat nominal growth is the typical aftermath of
  an unusually large run up in house prices
  
• Stress scenarios with double-digit price declines
  over three years are common
  
• The responses of house prices to interest rate
  shocks vary widely among MSAs
  
• Risk-based pricing of MSA credit risk is a viable
  alternative to the nuclear option and its many
  manifestations

4
Fidelitys Approach to Forecasting

• MSA prices driven by mortgage rates and local
  fundamentals
  
• Local employment
• Affordable house price median income
• Zip code forecasts incorporate local sales
  activity and persistent relationships with
  movements in the MSA average
  
• Functional forms driven by local conditions and
  history
  
• Scenarios for the fundamental drivers are
  transparent and based upon local experience
  
• The stress test for LA reflects its experience
• Subjective adjustments are incorporated
• Forecasts available for 375 CBSAs and thousands
  of zip codes

5
Key results for Average of CBSAs

• We construct a weighted average of the forecasts
  for 375 MSAs using recent sales as weights
  
• Most scenarios show declines in house price
  appreciation
  
• Expected annual growth rate on base path is 1.6
  percent per year over the next four years for
  weighted average of CBSAs
  
• The most stressful scenario reaches a trough at
  the end of 2009 and averages a -1.50 percent per
  year decline

6
Fidelity Nationals Analytic Group Index
7

• Most scenarios call for a substantial slowdown
  in house price growth
  
• About one-third show modest declines

8

• The base case scenario calls for annual growth
  of about 1.5 percent
  
• The most stressful path shows annual declines of
  about -1.5 percent thru 2009

9

The Case of Los Angeles Flat growth in base ca
se but considerable downside risk
10
Atlanta is expected to show continued growth and
a less severe stress scenario
11
Flat growth is the expectation for a wide
diversity of places
12
Four MSAs Face Serious Stress Scenarios
13
90680 is NW of Garden Grove and 92620 is further
south and east
14
We focus upon the differential impact of two
interest rate scenarios on house prices
15
San Francisco Substantial Impact of an Interest
Rate Increase on House Prices
16
Providence Substantial impact of an interest
rate hike on
base case house price forecast
17
Examples of large and small impacts of an
interest rate hike

• Smaller impacts
• Flagstaff
• Ft. Lauderdale
• Washington DC
• Detroit
• New Haven
• San Diego
• Dallas

• Larger impacts
• San Jose
• Honolulu
• New York City
• Miami
• West Palm Beach
• Las Vegas
• St. Louis

18
A Risk-Based Pricing Approach to House Price
Uncertainty

• Credit Risk Spread Measures the additional
  interest rate differential a lender should charge
  to earn the same rate of return losses on a
  mortgage due to default
• Incorporates expected losses and a capital
  charge
  
• Uses MSA specific house price forecasts by
  Fidelity Hansen Qualitys Analytics Group
  
• Computes the credit risk spread appropriate to
  each MSA for a prime 95 percent LTV mortgage to a
  borrower with a 680 FICO score
  
• Article published in the Mortgage Banking
  Magazine, October 2005 by Jim Follain and Mike
  Sklarz.

19
Credit Spreads by CBSA
20
Next Steps

• Market House Price Forecasting Package
• Includes data and forecasts
• Contact jfollain_at_hanqual.com or
  mike.sklarz_at_fnf.com
  
• Continue development
• Focus attention on larger CBSAs
• Investigate the impact of affordable mortgage
  products
  
• Incorporate the impact of external demand
• Deepen investigation of zip code information
• Create more stable indexes
• Explain variations in BETAS among zip codes