House Prices and Fundamentals Under the Microscope

Steve Oliner* American Enterprise Institute and UCLA Ziman Center

Presented at Meeting of Collateral Risk Network Washington, DC October 30, 2013

*This presentation describes work done jointly with Ed Pinto at AEI and Morris Davis at the University of Wisconsin.
1

Part 1: Setting the Stage

House Prices: What Do We Know?

Housing busts can be very damaging to the economy and financial system. Best to prevent them if possible. House prices can stray from fundamentals (construction cost, rent, income). Eventually correct back. Risk varies over the cycle. Increases during the boom. Land is the risky part of the housing bundle.

Residential Land and House Prices

In recent cycle, land prices moved up and down much more than house prices.

300

Index, 2002:H2 = 100 Half-yearly Residential land price index, 23 cities

300

250

250

200

200

150

150

Case-Shiller 20-city house price index

100 100

50

50

0 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

Source: Nichols, Oliner, and Mulhall (2013).

4

Open Questions
How do land and house prices move within a metro area? Is there a common spatial pattern across cities? How does risk correlate with the size of the gap between house prices and fundamentals? Are there threshold effects? What are the most effective ways for regulators to lean against house price booms? Can simple rules work? Micro data are important for addressing all of these questions.
5

This Study
Uses property-level data to pop the hood on housing valuation within metro areas.
Focus on detached single-family homes.

Initial analysis for Montgomery County, MD.

Borders Washington DC. Large, diverse county. About 1 million residents.

In the process of extending the analysis to 25 other counties in 10 metro areas.

Part 2: Data and Methodology

Key Data Sources for Montgomery County

Property-level data
House characteristics, location, and AVM value: FNC, Inc. Reconstruction cost as new: Marshall & Swift/Boeckh (MSB)

Zip-level data
House price indexes: FNC and Zillow Construction cost indexes: MSB

Property-Level Snapshot in 2013:Q3

Dataset contains 167,000 detached single-family homes, which is about 90% of all such homes in Montgomery County. Have AVM and reconstruction cost as new for each home. Adjust reconstruction cost for depreciation
Estimate depreciation from Montgomery County dataset Find depreciation of about 1.2% per year to age 20, slower depreciation after that.

Land value = house AVM value adj. reconstruction cost. Land is a shorthand for amenities of a given location.

Creating Time Series

Roll the property-level AVMs back from 2013:Q3 to 2000:Q1 using FNC zip-level house price indexes. Roll the property-level adjusted reconstruction cost back from 2013:Q3 to 2000:Q1 using MSB zip-level cost indexes.
For Montgomery County, almost all zips have same cost index

Back out implied land value for each property in each quarter.

10

Part 3: Results

11

House Price Decline in Montgomery County, 2006-2012

Wide variation across zips. Smallest declines in close-in areas, largest declines in outlying areas.
20872, 21771

Percent

Less than 10* 10-20 20-30 30-40

20871 20882, 21797

More than 40
20777, 20833, 20860, 20861, 20862

20876 20838, 20839, 20841, 20842 20879 20886

20879

20879 20832 20855 20868, 20905 20853 20850 20851 20904 20902 20895, 20896 20814 20815 20812, 20816 20910 20912 20901 20903 20906

20874 20837 20878

20877, 20880

20707, 20866

20852 20854

20817, 20818

20903

* Includes zip 20815, where index rose 7 percent. Source: Authors' calculations using data from FNC, Inc.

12

House Prices in 20815 and 20903

Very different cycles in two zip codes less than 10 miles apart.
Index, year 2000=100
300 Annual average 260 20903 -- Silver Spring 220 20815 -- Chevy Chase 180 180 220 260 300

140

140

100

100

60 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

60

Source: FNC, Inc.

13

Average Land Price in Montgomery County, 2013:Q3

Highest in close-in zips, with downward gradient moving away from DC.
20872, 21771

Dollar per square foot

Less than 15 15-25
20871 20882, 21797

25-40 40-70 More than 70

20777, 20833, 20860, 20861, 20862 20879 20832 20855 20868, 20905 20853 20850 20851 20904 20902 20895, 20896 20814 20815 20812, 20816 20910 20912 20901 20903 20906

20876 20838, 20839, 20841, 20842 20879 20886

20879

20874 20837 20878

20877, 20880

20707, 20866

20852 20854

20817, 20818

20903

Source: Authors' calculations using data from FNC, Inc. and Marshall & Swift/Boeckh.

14

Average Lot Size in Montgomery County, 2013:Q3

Largest in outlying zips, dropping to less than 1/4 acre in close-in zips.
20872, 21771

Square feet

Less than 11,000 11,000-15,000 15,000-25,000

20871 20882, 21797

25,000-40,000

More than 40,000

20777, 20833, 20860, 20861, 20862

20876 20838, 20839, 20841, 20842 20879 20886

20879

20879 20832 20855 20868, 20905 20853 20850 20851 20904 20902 20895, 20896 20814 20815 20812, 20816 20910 20912 20901 20903 20906

20874 20837 20878

20877, 20880

20707, 20866

20852 20854

20817, 20818

20903

Source: Authors' calculations using data from FNC, Inc. and Marshall & Swift/Boeckh.

15

Average Lot Value in Montgomery County, 2013:Q3

By far the highest in close-in zips even though the lots are small.
20872, 21771

Dollars

Less than 200,000 200,000-250,000 250,000-400,000 400,000-600,000

20871 20882, 21797

More than 600,000

20777, 20833, 20860, 20861, 20862

20876 20838, 20839, 20841, 20842 20879 20886

20879

20879 20832 20855 20868, 20905 20853 20850 20851 20904 20902 20895, 20896 20814 20815 20812, 20816 20910 20912 20901 20903 20906

20874 20837 20878

20877, 20880

20707, 20866

20852 20854

20817, 20818

20903

Source: Authors' calculations using data from FNC, Inc. and Marshall & Swift/Boeckh.

16

Average Land Share of Property Value in Montgomery County, 2013:Q3

Excceds 65% in some close-in zips, dropping to less than 35% in some outlying zips.
20872, 21771

Percent

Less than 35 35-45 45-55 55-65

20871 20882, 21797

More than 65
20777, 20833, 20860, 20861, 20862

20876 20838, 20839, 20841, 20842 20879 20886

20879

20879 20832 20855 20868, 20905 20853 20850 20851 20904 20902 20895, 20896 20814 20815 20812, 20816 20910 20912 20901 20903 20906

20874 20837 20878

20877, 20880

20707, 20866

20852 20854

20817, 20818

20903

Source: Authors' calculations using data from FNC, Inc. and Marshall & Swift/Boeckh.

17

Land Prices over Time

For zips grouped by land price in 2013:Q3. Very large swing for low-price zips, much less for high-price zips.
120

Index, 2006=100 Annual

120

More than $70/sf

100

100

$40-$70/sf
80 80

$25-$40/sf
60 60

$15-$25/sf
40 40

Less than $15/sf

20 20

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

Note: Observations for 2013 represent average of Q1, Q2, and Q3. Source: Authors calculations using data from FNC, Inc. and Marshall & Swift/Boeckh.

18

Prices and Construction Costs: 2000-2006

Land prices rose more than house prices everywhere, especially in zips with low land prices. Construction costs accounted for little of the rise in house prices.
600

Percent change

500 Land Price 400 House price Construction cost

300

200

100

0 Less than $15/sf $15-$25/sf $25-$40/sf Zips grouped by land price in 2013:Q3 $40-$70/sf More than $70/sf

Source: Authors calculations based on data from FNC, Inc. and Marshall & Swift/Boeckh.

19

Prices and Construction Costs: 2006-2012

Land prices fell more than house prices everywhere, especially in zips with low land prices. Construction costs continued to rise.
Zips grouped by land price in 2013:Q3 Less than $15/sf 20 10 0 -10 -20 -30 -40 Land Price -50 -60 -70 House price Construction cost $15-$25/sf $25-$40/sf $40-$70/sf More than $70/sf

Percent change

Source: Authors calculations based on data from FNC, Inc. and Marshall & Swift/Boeckh.

20

Land Shares over Time

For zips grouped by land price in 2013:Q3. Huge swing in land share for lowprice zips, while share is much more stable for high-price zips.
80%

Percent Annual

More than $70/sf $40-$70/sf

70%

60%

$25-$40/sf
50%

$15-$25/sf
40%

30%

Less than $15/sf

20% 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

Note: Observations for 2013 represent average of Q1, Q2, and Q3. Source: Authors' calculations using data from FNC, Inc. and Marshall & Swift/Boeckh.

21

2000-12 Cycle: House Prices Fell More in Zips with Previous Large Rise in Land Share
10

0 10 House price change, 2006 to 2012, percent 15 20 25 30 35 40 45 50

-10

-20

-30

-40

-50

Maximum rise in land share from 2000 to any year through 2006, percentage points

Source: Authors' calculations using data from FNC, Inc. and Marshall & Swift/Boeckh.

22

2000-12 Cycle: House Prices Fell More in Zips with Previous Large Rise in Price-Rent Ratio
10

House price change, 2006 to 2012, percent

60

70

80

90

100

110

120

130

140

-10

-20

-30

-40

-50

Maximum rise in price-rent ratio from 2000 to any year through 2006, index points

Source: Authors' calculations using data from FNC, Inc., Marshall & Swift/Boeckh, and REIS.

23

2000-12 Cycle: House Prices Fell More in Zips with Previous Large Rise in Price-Income Ratio
10

House price change, 2006 to 2012, percent

0 40 60 80 100 120 140 160

-10

-20

-30

-40

-50

Maximum rise in price-income ratio from 2000 to any year through 2006, index points

Source: Authors' calculations using data from FNC, Inc., Marshall & Swift/Boeckh, and ScoreLogix.

24

Metro Areas in Full Study

East Coast Boston Washington DC Miami Far West Seattle Los Angeles Phoenix Interior Chicago Detroit Memphis Oklahoma City

MSAs represent range of markets across the U.S. Counties selected to capture range of experience within the MSA. 26 counties in all. Total population of selected counties: about 40 million
25

Summary
Housing submarkets are important in Montgomery County. House price swing was largest in outlying areas; it was much smaller in affluent, close-in areas. Land prices were more volatile than house prices, especially in the outlying areas. Rapidly rising land share was associated with sharp house price correction later on. Results suggest that clearest signal about over-valuation could come from developments in the outlying areas.
26