by Calculated Risk on 12/08/2006 03:20:00 PM

From Freddie Mac chief economist Frank Nothaft, et al: Anatomy of a Housing Recovery

"The housing market has been central to the economic outlook for the past several years. A buoyant residential sector provided critical support in the recovery from the 2001 recession. More recently, the housing slowdown since mid-2005 has been instrumental in helping realize the Fed's goal of moderating economic growth, in order to nip inflationary pressures in the bud. Over the next several quarters, housing will again be key to the outlook, as a prolonged downturn could threaten more severe consequences for the overall economy, while a housing recovery could engender a broader upturn. Given the importance of housing at the current juncture, two key questions arise: What will a housing recovery look like; and, When will it arrive?

Previous housing booms help provide a historical context. Residential investment as a share of GDP, a broad indicator of overall housing activity, rose to nearly 6 percent in both 1973 and 1978, compared to a long-run average of 4 ½ percent. The investment share subsequently slid to approximately 4 percent over the ensuing six to eight quarters, before stabilizing. After reaching a trough, real residential investment rose in the following four quarters. Price appreciation (as measured by the yearly change in repeat-sales indexes) tends to slow after a peak, but did not turn negative on a national basis after these booms. Rather, the boom periods are generally followed by extended sluggish price gains, though economically depressed local and regional markets may register outright declines."

Click on graph for larger image.

Residential investment (RI) has now fallen to 5.7% of GDP from the peak of 6.3% in the second half of 2005.

"During the recent boom, residential investment rose to a slightly higher share of GDP, reaching 6 ¼ percent in the second half of 2005. Investment has been declining rapidly since then, however, falling at an 18 percent annual rate (adjusted for inflation) in the third quarter, and is on track to a similar decline in the current period. This trend would reduce residential investment relative to GDP to 4 ½ percent in the second quarter of 2007, a change in line with declines in previous episodes, and over a similar six- to eight-quarter time frame."

In the previous cycles mentioned, starting in 1973 and 1978, residential investment as a percent of GDP eventually declined to 3.7% and 3.2% respectively. To reduce residential investment relative to GDP to 4%, the current trend would have to continue through all of 2007. The Freddie Mac analysis excludes the early '90s bust: residential investment peaked at only 5.0% of GDP and declined over 13 quarters to 3.3% of GDP in early 1991.

I'm not sure why Nothaft is suggesting residential investment will only decline to the long-run average of 4 ½ percent. Also, for the previous residential investment downturns, the period from peak to trough were: early '70s - 8 quarters, late 70's, early '80s - 16 quarters, and late 80's, early '90s - 18 quarters. So the "six to eight quarters" timeframe seems optimistic.

"By contrast, forecasts of a more serious and prolonged housing slowdown fall well outside the range of post-war experience. For example, consider an annualized 18 percent decline in residential construction every quarter through end-2008. Under such a scenario, the drop in residential investment relative to GDP would be 50 percent greater than what occurred during the 1973-74 recession, reaching roughly 3 percent, a result worse than during even the most severe post-war recessions."

The bottom of three previous troughs were: 3.7%, 3.3% and 3.2% relative to GDP. At the current rate of decline, the current downturn could last to about mid-2008 to reach about the same level. And the duration would only be about 11 quarters - shorter than two of the three previous downturns.

"... We expect housing markets to stabilize some time during the first half of 2007...."

At each step of this analysis, Dr. Nothaft has shaded to the most positive view. The three previous cycles lasted from seven to eighteen quarters, and that became "six to eight quarters". The three previous bottoms were all in the mid 3% range (3.2% to 3.7%), yet Dr. Nothaft forecast the bottom of the current decline at the long-run average of 4 ½ percent. And, as Nothaft noted, a decline at the current pace to the long-run average would persist through mid-2007, but then Nothaft forecast the market to "stabilize some time during the first half of 2007".

If we are using a "historical context", why not forecast a more normal cycle downturn of 11 quarters and a bottom of 3.4%? Of course the bottom wouldn't occur until Q2 2008 at the current pace.