This note was originally published June 17, 2013 at 12:42 in Financials
Typically we pay little attention to the builder confidence survey, but given the recent underperformance in the builder complex we thought it was worth some further consideration. Looking at the history of the series, which unfortunately only dates back to 1985, there have been four prior episodes of builder confidence crossing above 50 and then moving on steadily to 65-70. In the first chart below we illustrate this tendency. The four prior peaks were 65 in June, 1986, 71 in November, 1993, 78 in December, 1998 and 72 in June, 2005. We find it interesting that builder confidence tends to autocorrelate (i.e. move in the same, self-reinforcing direction for long periods of time), just as home prices and housing demand do.
If we assume that builder confidence continues to trend higher, toward its prior peaks in the 65-78 range, what are the implications for construction activity? The chart below shows the historical relationship between builder confidence and construction activity. There are several takeaways from this chart. The x-axis shows the HMI (builder confidence) reading, while the y-axis is the concurrent level of single family housing starts in thousands of units annualized. Over the history of the series (1985-Present), we've seen a 16k change in SF starts for every one point change in the HMI index. Interestingly, based on today's builder confidence reading of 52, the fair value for SF starts is 1.137 million vs. the current reading of 610k, or 87% above where it is today. If we assume that confidence rises to 65, then SF starts should reach 1.35 million and if we assume 78, we get 1.56 million.
Bear in mind, this is single family starts. Historically, as the chart below shows, single family starts have averaged 72% of total starts from 1960-present. The most recent reading was identical at 72%, but if we take the average of the last six months, single family starts have been averaging 67% of total starts. If we divide the above figures by the long-term average of 72%, it translates to high side estimates for total starts of 1.58 million (today's HMI of 52), 1.875 million (HMI 65), and 2.17 million (HMI 78). These rates of construction are generally consistent with the levels we would expect to see given the current rates of new household formation we're seeing coupled with the historical relationship between household formation and housing construction.
It's also worth noting that the recovery in builder confidence remains more parabolic than linear, i.e. supportive of our conclusion that the housing recovery continues to show signs of acceleration. It doesn't hurt that mortgage rates have also backed off their recent highs, ticking down to 3.94% from their recent highs of 4.16%. As the chart below shows, the second order measure of trajectory remains much stronger (0.92) than the first order (0.76).
Housing's momentum has been one of the three central tenets of our bullish stance on Financials, alongside the labor market recovery and the Fed's asset purchase program. Today's signs of ongoing recovery, and even acceleration, are a welcome sign suggesting that our bullish intermediate to long term view on the Financials remains on track. However, in the short-run, we remain cautious for several reasons including the predictable deceleration in seasonally-adjusted labor data, elevated uncertainty around the Fed's willingness to continue its asset purchase program, and steadily deteriorating conditions in the credit markets.
The two charts below look at the long-term as well as the short-term relationship between builder confidence and builder activity levels.
Joshua Steiner, CFA
Jonathan Casteleyn, CFA, CMT