patrick.net

In June 2009, as the economic recovery was just getting underway, we reaffirmed this view (US Daily Financial Comment, “L is for Housing”, June 25, 2009).  In particular, we highlighted expectations that: (1) homeownership would continue to decline; (2) demand for owner-occupied housing would therefore remain weak; (3) as a result, the overhang of unoccupied homes would take a long time to work off; (4) while this was occurring, sales of existing homes would dominate total sales; and (5) sales of new homes and starts of single-family units would remain depressed.

However, as starts and sales began to exceed our expectations, we marked up our estimates of frictional single-family starts modestly in September—from an original guesstimate of a 500,000 annual rate to about 600,000 by late 2010—on the view that disparities between the types and locations of vacant, unsold homes and the types and locations of homes in demand would justify modest increases from the depths to which home sales and starts had fallen in early 2009.  (See US Economics Analyst, “Homebuilding: More Upside, But Much Less than Usual.” September 11, 2009.  In theory, if houses were completely commoditized— i.e., movable, capable of being resized, etc.—then starts would fall to zero until the excess supply was worked off. However, because housing is inherently heterogeneous, new starts will occur in regions and for types of housing where demand exceeds supply.  In other words, disparities between the types and locations of housing demanded and supplied will drive production in an environment in which excess supply prevails.)  Even so, our expectations for homebuilding remain well below the standards of most economic recoveries, when activity in this sector routinely contributed about 1 percentage point to the first year or two of real GDP growth.  For example, we currently expect residential investment to contribute an average of 0.4 points to annualized real GDP growth from mid-2009 through the end of 2011.  This includes the effect of a recent surge in sales of existing homes on brokerage commissions, which count as part of residential investment.

In recent months the wisdom of this upgrade has come into question as housing starts have failed to increase and new home sales have weakened anew.  As shown in Exhibit 1 below, the latest decline in new home sales brought the three-month average of sales slightly below the low end of prior experience for housing cycles.  (In the exhibit, the past experience includes cycles in home sales with troughs in October 1966, April 1970, February 1975, April 1982, and February 1991.  The 1966-67 housing cycle did not culminate in an economy-wide recession.  We have excluded the cycle associated with the 1980-81 recession/recovery because it was so brief, while the 2001-02 recession/recovery did not have a meaningful cycle in housing activity.)

Meanwhile, the National Association of Home Builders’ (NAHB’s) housing market index – a gauge of builders’ assessments of the demand for new homes—has also suffered renewed setbacks – as have mortgage loan applications for home purchase.  These developments, coupled with a recent sharp drop in pending home sales in November, reinforce our sense that the homebuyer tax credit – which was originally due to expire on November 30—helped boost home sales (both new and existing) and may have had a stronger impact we originally thought.  Although the tax credit was eventually extended (to the end of April, 2010) and expanded (to provide some help to existing homeowners), the effect of these measures has yet to be seen.  We suspect they will not be as powerful as the original tax credit; otherwise builders would probably be reporting better prospects.  If this suspicion is right, then single-family starts and residential investment are apt to disappoint to the downside of our forecasts.

Ed McKelvey

Goldman Sachs Financial Conditions Index^SM*(October 20, 2003=100)
*Revised as described in our April 8, 2005, US Economics Analyst.