Saturday, April 5, 2014

Why the consensus house price forecasts are wrong.

Mark Fleming, Corelogic’s chief economist, views are typical of the consensus.  In his February post, Popping the Housing Bubbly Theory,  he concludes;  “analyzing home price levels relative to fundamental prices leads us to conclude that there is no need to fear a bubble for at least a few years to come, if at all.” He believes that house prices will continue to rise through 2015.   The housing consensus has been wrong for the last three years, so will it be right now?

This data for the Case Shiller index illustrated below demonstrates the rapid increase in prices which began in March 2012, peaked in March 2013, only to fall 55% in the last 10 Months.
Prices in the Case Shiller California cities (Los Angeles, San Francisco, and San Diego) average 273% above the start of the cycle in June of 1997, while inflation is 146% so in real terms prices are 185% above the start of the cycle. In normal times the prices would fall, but these are not normal times.  Government programs and low interest rates lead them to grow faster.   In this particular case prices declined in 2011 in advance of expected return to normal FHA loan Limits. When the higher loan limits were restored prices jumped. The increases fell off in June 2012 but by then expectations of higher prices, investor purchases, and the fear of increasing mortgage rates sent prices soaring.  The high priced cities in California experienced the sharpest growth.  They increased 36% in two years.

 This increase cannot be explained by macroeconomic indicators.  It is more likely an example of Fed Governor Jeremy Stein’s concern about reaching for yield causing distortions in the housing market.  Pushing prices up that will come down painfully in the future. California was the area most affected by investor purchases because the high prices make this an attractive strategy. 

The Case Shiller index is priced far above the national market. It is an index of volatile markets, but it is an excellent index of price behavior.   In the current reporting period San Francisco and San Diego are growing at over 20% per year. Since California cities make up 27% of the Case Shiller and 12% on national housing value they have a big impact.  The FHFA expanded index of national prices, which is similar to the Case Shiller in methodology and result, grew at 5% per year in the latest quarter. Outside the 20 cities price increases were only 1 % per year. 

 Regional housing cycles in the United States have ended up where they were adjusted for inflation. They are slow moving and go on for years regardless of the business cycles. This is illustrated in the following chart.
The growth in prices to the peak correlates with the drop to the second quarter 2009 at 92%.  After this period, government stimulus programs raised the Case Shiller cites but the areas outside these cities continued to decline. The real price of the Case Shiller cites are 45% above the start of the bubble while the prices outside these cities are 5% below.

Of course some markets outside the 20 cities are high priced and volatile and some of the 20 cities are stable. Real Estate is driven by location, location, location.  But this data indicates that housing can be considered as a two track market. In markets which are restrictive in permitting new housing, programs to help lower income buyer will drive prices in the entire market. More purchases at the low end create move up buyers and prices increase in all price ranges. 
The downside of monthly data is the high standard deviation. The high priced cites have the highest standard deviation.  The time lag is substantial. The data for March reporting covered January closings, most of which were negotiated in November. Both the Case Shiller and FHFA report data over a thr
ee months period as the data becomes available.  In April reporting, more January closing data will be added to the first month’s batch. The final batch will arrive in May reporting. The second batch of data produces the greatest change so the first month of data does get the largest revision.

One problem casted by the time lag is the effect of the cold winter. This is likely to have depressed purchase activity and prices starting with the December to March negotiations which will be seen in April to August reporting. But that effect will be difficult to identify. 
Fleming talks about prices being undervalued because of rising income and greater housing affordability. This theory will be tested very soon.  An obvious conclusion is that prices will decline in the Case Shiller cities in the future contrary to Fleming’s view.