Sunday, April 22, 2012

Housing recovery is national, excluding Case Shiller

The Case Shiller 20 city index is reasonably accurate, but unrepresentative.  House prices in these cities are twice the national level.  While the Case Shiller index declined 34% since the peak in 2006, prices outside these regions declined only 13%.  This is summarized below.

Price Indexes: Case Shiller, FHFA and Heartland

The Heartland index (GREEN) is the sum of FHFA (Federal Housing Finance Administration) data for regions outside the Case Shiller 20 SMAs.  It contains volatile urban and upscale SMAs, but it consists largely of lower density areas.  The 20 Case Shiller SMAs (BLUE) and FHFA data for these SMAs track with reasonably consistency.  Case Shiller’s national index contains only urban areas so it is meaningless.  The FHFA index extended with FHA and private sales is used as the national index (RED).

Case Shiller peaked at 258% of the price at the beginning of the cycle in 1997, while the Heartland peaked at 136%.  When measured by Case Shiller, the house purchase tax credit and higher loan limits raised prices, but in the invisible Heartland markets they had little effect. 

How far do prices have to for housing to recover?

Housing Price Forecast

The following table illustrates the status of the 20 Case Shiller cities.  For example, Los Angeles peaked in 2006 at 358% of the starting price.  By December 2011 it was 212% - a drop of 41%.  In the last three months, prices fell at an annualized rate of 10%.  By contrast the recovered cities now average 116% of the price at the start of the bubble - about the same as the Heartland index.

After a severe downturn prices move up and down with national and regional economic pressures.  For example, Detroit’s brush with bankruptcy led to a decline in the fourth quarter, causing total recovered markets to decline, but Detroit was still higher than in the first quarter.  Case Shiller declined at an annualized rate of 8% in the fourth quarter - triple the second quarter rate.  Case Shiller data timing problems in markets such as Atlanta overstated the decline, but three months data illustrates trends not seen in annual comparisons.

Why Case Shiller Cities will fall so much


Prices at the start of the cycle reflect premium prices in locations such as Los Angeles.  These Urban areas limit expansion of the housing supply when demand goes up.  They delay new construction and increase the cost of permitting.  All of this increases prices creating the conviction that prices will increase forever.  

This atmosphere infects lower priced markets but prices do not rise rapidly because new housing is built to meet demand.  For example Cleveland, Charlotte and Dallas reached peak prices of 137% to 134% of the beginning of the cycle just like the Heartland regions.  Cleveland which is heavily automotive declined by 19%, Charlotte in a less depressed market dropped 11% and Dallas 7%. 

Credit expanded rapidly during the housing bubble. This easy money had a limited impact in most of the country but it ignited a price explosion in markets where housing could not be built rapidly built to meet demand.  Investors soon entered the market to flip houses.  The credit expansion was an accelerant, the same as throwing gasoline on the fire in high priced markets.  Government programs to address the housing crisis increased prices in the cities where the prices took off with the credit expansion, but this was a temporary boost. 

There are no economic fundamentals that explain for why prices in the Case Shiller cities rose 258% in nine years.  Most forecasters say that prices will soon bottom out, but they cannot point to economic fundamentals to explain why prices of cities like Los Angeles and New York should remain well above the level at the start of the bubble, when prices is the bulk of the country did not keep up with inflation.

The U.S. has a history of regional housing cycles.  Higher price increases in the expansion lead to sharper declines in the contraction.  This works because price declines are self-reinforcing.  They increase foreclosures and put off buyers.  The cycles are much longer than people outside the industry expect.  I worked for both American Standard and Kohler in planning and acquisitions.  Both of these organizations knew that local contractions went on for years even when national prices were increasing and set local budgets accordingly. 

Housing cycles do not coincide with business cycles, so an economic recovery will help, but not end a housing contraction.   The price expansion lasted 9 years.  Contractions usually last 80% as long as the expansion.  (A full discussion of North American cycles can be found in Cunningham and Kolet, Bank of Canada working paper 2007-2.)  Because the house purchase tax credit and FHA financing interrupted the contraction, it will last longer.  They will not end in some markets until 2016. 

Why Housing cycles are usually a round trip

Some cities to rise in attractiveness and others fade over decades.  But historically the housing cycle usually ends up the same place as it started adjusted for the normal annual growth in value.  In a high priced city such as Los Angeles, homes are more expensive because land is the bulk of the price.  A home buyer in Los Angeles is speculating in land.  A buyer in Dallas is actually buying a house.

At the start of the cycle in 1997, Los Angeles prices were 164% of Dallas prices, but they were 440% above at the peak.  Now they are 280%.  If prices do not go back to the previous relationship, local businesses will leave because costs are uncompetitive.  For example, Boeing consolidated some of its Los Angeles operations in Oklahoma City.  More and more businesses and jobs are moving out.  Businesses may stay but functions are outsourced.  It does not help that higher property tax revenue in the boom leads to the granting of permanent overly generous local government benefits.  The higher housing costs make local business and consumer services more expensive.  Dentists, restaurants plumbers all become more costly.  Employees are happy to leave, as are businesses that are not tied to local customers.  

With inflation at 141% of 1997, the Dallas price of 125% is 16 percentage points below inflation.  For Los Angeles prices to return to 64% above Dallas prices would have to fall to $300,000 from $875,000 in 2006. It now averages $517,000.  Given the damage to the Los Angeles economy, prices may actually have to fall below $300,000.  So if the programs to stop declining house prices had succeeded, Los Angeles long term job losses outside housing would have been disastrous.

When purchasing real estate or mortgage backed securities it is critical to understand were prices are going to end up.  Forecasts biased on a knowledge of housing cycles are drastically more accurate than other approaches  

Larger FHA Deficit

The attraction of the housing fix is that GDP will surge and jobs will come back so why not ignore the off budget costs and try?  The problem is not that Los Angeles prices are falling.  The problem is that they were not allowed to fall to a sustainable level.  Buyers who took advantage of the $8000 tax credit and 3.5% down FHA financing to buy a house in Los Angeles are already underwater.  Joseph Gyourko estimates that the FHA’s one trillion fund which HUD projects with a positive balance is $50 to $100 billion negative.  Case Shiller declines projected in the analysis will increase this loss significantly.  

The result of this expensive experiment proves that propping up unsustainable house prices in volatile cities is impossible.  The least painful solution is for the prices to come down to a sustainable level to minimize the loss of jobs in overpriced cities.