A post last May suggested that house prices were bottoming, in contrast to consensus expectations at the time of a further 10-15% decline. This proved insufficiently upbeat. The Halifax index had already troughed in April and rose by 9%, or 14% annualised, over the subsequent eight months to December. The Nationwide index was also 9% above its low by December.
 
The first chart compares the quarterly evolution of real house prices since the peak in the third quarter of 2007 with their development after three prior major tops, in 1973, 1979 and 1989. (The comparison is based on the Nationwide measure, which has a longer history than the Halifax, deflated by the retail prices index.) Up to the first quarter of last year, the fall in real prices was closely tracking the major declines of 1973-77 and 1989-96. This suggested a further loss of 15-22% – roughly consistent, given inflation, with the consensus forecast for nominal prices.
 
After the first quarter, however, real prices moved sharply higher to converge with the less severe 1979-82 decline.
 
What explains the sudden shift from early 2009? The cut in Bank rate from 5.0% in September 2008 to 0.5% in March 2009 led to a steep drop in mortgage interest bills, reducing delinquencies and forced selling. Household interest payments fell from 10.5% of disposable income in the third quarter of 2008 to 6.9% a year later. This is close to the historical low (since 1987) of 6.7% in the first quarter of 2002.
 
Perhaps the decision to embark on quantitative easing in February 2009 also contributed to the shift. This may have boosted longer-term inflationary expectations, thereby increasing the attraction of housing as a store of real value relative to other assets, particularly cash savings.
 
If real house prices were to continue to follow their path after 1979, they would rise by 4% in the year to the fourth quarter of 2010 and 7% in the subsequent year. Assuming retail price inflation of 3%, the cumulative nominal gain over the two years would be 18% – sufficient for the Nationwide measure to surpass its 2007 peak by late 2011.
 
Such a scenario would imply housing remaining expensive by historical standards. As argued in the previous post, the "national rental yield" is a better measure of valuation than the house price to earnings ratio. This yield ended 2009 at about 3.4% versus a long-term average of 3.6%, implying overvaluation of about 5% – see second chart. (The yield rose slightly during 2009, despite the recovery in house prices, because of rapid growth in the national accounts rents measure.)
 
Historically-high valuations, however, may prove sustainable as long as real interest rates remain low or – in the case of cash rates – negative. The gap between the rental yield and the real yield on long-term index-linked gilts is at a record high – second chart.
 
The housing market recovery would be at risk if the MPC were to restore a positive real level of official interest rates. The consensus view, however, is that Bank rate will remain below both current inflation and the 2% target throughout 2010.