A typical analysis from policy makers, like the Federal Reserve, points out that activity in the housing market has declined so far this year; but that assertion focuses primarily on housing starts. A more complete picture from the national data shows that the housing market bounced back strongly in the 2019 first quarter from a steep 2018 fourth quarter slide. The housing market was also able to hold on to those gains in the 2019 second quarter.
The chart below shows the trend in seasonally-adjusted annualized U.S. single-family starts and permits. Except for the January 2019 spike, it suggests that single-family housing activity has been “rolling over,” gradually trending downward since the peak in early 2018.
On an unadjusted basis, single-family housing starts have been in a clear downtrend for the past three consecutive quarters. Against prior year results, they were down 7.0% in 18Q4, 3.2% in 19Q1 and 6.2% in 19Q2. (The 19Q2 figures are preliminary and still subject to revision.). The permits data roughly mirror the trend in starts, with mid-single-digit declines in 19Q1 and 19Q2.
New home sales have been quite volatile over the past year, as indicated in the chart below, which shows unadjusted quarterly new home sales.
New home sales fell off a cliff in the final quarter of 2018, down 13.2%, but they rebounded strongly in the first quarter of 2019 and held their gains in the 2019 second quarter. The 3.2% gain in new home sales in 19Q1 is especially remarkable considering that the prior year quarter was viewed as exceptionally strong, due to the expected benefits of the recently passed Tax Cut and Jobs Act and with mortgage rates still at low levels.
Some economists have suggested that the 19Q1 rebound in new home sales was “pulling sales forward” due to the sharp decline in mortgage rates. In my view, however, it was pulling sales back, as potential buyers who had moved to the sidelines at the end of 2018 came back in to the market.
The declines in single family housing starts must be considered in light of this volatility in new home sales. With sales down 13.2% in 18Q4, builders might have considered cutting back starts even more aggressively. However, when mortgage rates began to fall sharply in November 2018, they became more optimistic about the outlook for sales for the upcoming spring selling season.
Still, the 18Q4 drop in new home sales left them with more inventory under construction as 2018 came to a close. Even with the expected rebound in new home sales from 18Q4 levels, they presumably judged that they would be able to meet the demand with fewer starts. With the rebound in sales and declines in starts, single-family homes under construction have been easing back to more sustainable levels.
Going forward, assuming no significant decline in general economic activity, housing starts should begin to show gains (vs. the prior year) in the 2019 second half. Those expected gains will likely come from a return of under construction inventories to more sustainable levels and against favorable prior year comparisons.
For the full year, I currently anticipate single family housing starts will come in at 1.26 million units, ahead of the consensus estimate of 1.22 million units, as reported in the Philly Fed’s Survey of Professional Forecasters (SPF) My forecast assumes that multi-family starts will be roughly flat at 375,000 units (after being down about 1% in the 2019 first half) and that single-family starts will be 882,000 units, up 0.7% for the year (after being down 4.9% in the 2019 first half). This forecast presumes GDP growth consistent with the latest view of the SPF forecasters, who anticipate real GDP growth of 2.1% in 19Q3 and 2.2% in 19Q4.
Given favorable prior year comparisons in the 2019 second half, I project that new home sales will total 654,000 units in 2019, up 6.0% for the full year. Year-to-date, through June, new home sales were up 1.7% vs. 2018. My projection assumes that annualized real GDP growth will average about 2.1% in the 2019 second half.
Underlying demand for housing has historically been driven most closely by employment growth and consumer confidence. So far this year, employment growth has remained solid, but there are clear signs that it is slowing. Monthly gains in private payrolls have averaged 165,000 per month in 2019, compared with the 2012-2018 average of 203,000 per month. With unemployment low, further improvement in housing demand can come from wage growth, which is now running above 3% annually.
Consumer confidence, as measured by the Conference Board, has been volatile in recent months; but it rebounded sharply in July to its highest level this year. Both the employment situation and consumer confidence are therefore supportive of continued improvement in housing.
On the other hand, there are also concerns of an emerging self-inflicted slowdown in economic activity, driven by the escalating trade war with China. So far this year, industrial activity has lost considerable momentum, as indicated by the fourth consecutive monthly decline in the PMI Manufacturing Index to 51.2 in July, which is only slightly above the neutral level of 50.0 that signals neither expansion nor decline. A year ago, the PMI was around 60.0.
Similarly, the Federal Reserve’s index of industrial production has been essentially flat since the beginning of the year. Year-over-year, the index was up 1.3% in June. However, year-over-year gains in industrial production were running as high as 6% a year ago. Still, the stabilization in the index over the past few months is a welcome sign.
For now, the ongoing strength in employment and housing have flowed from the general momentum of the economy, which has benefited from low interest rates and the TCJA. If major sectors of the economy, such as industrial, continue to slow (or possibly slip into decline) that will eventually lead to fewer job gains and slower demand for housing.
This yea’s rebound in housing is also evident in homebuilder new order trends, as shown in the chart below:
The average year-over-year increase in 19Q2 net new orders for the nine (of eleven) builders that I follow who have reported results so far was 11.2%, up sharply from a 0.7% decline in 19Q1 and a 9.7% plunge in 18Q4. (The nine builders include BZH, DHI, KBH, LEN, MDC, MHO, MH, NVR and PHM. TOL and HOV, the other two builders that I follow, will not report results for the 19Q2 calendar results until later this month or early in September.),
Clearly, the 2019 second quarter’s 11.2% gain exceeded the change in total U.S. new home sales, which were flat in the quarter. Accordingly, while builder net orders can continue to show solid gains (assuming an accommodative economic environment), it is highly unlikely that average order growth will continue at the elevated rate of 11.2% for long.
The stronger relative order performance by the homebuilders (vs. the national average) is due to an expansion in the number of active subdivisions, a mix shift toward more affordable homes and a modest pick-up in absorptions. That suggests that the large publicly-traded builders have wrested market share from smaller builders and also from the existing home market.
The turnaround from the 18Q4 freefall has helped the homebuilders outperform the broader market this year. The Lark Research Homebuilder Stock Price Index, which tracks the share price performance of the eleven builders, is up 26.6% through August 2, better than the gains of 17% in the S&P 500 and 13.7% in the Russell 2000.
It is also worth noting that the homebuilders were the second best performing Dow Jones U.S. Total Market Industry Group last week (ended August 2) and one of only 18 (out of a total of 143 industry groups) to post a gain for the week. My homebuilder stock index was up 0.6% for the week, compared with declines of 3.1% for the S&P 500 and 2.9% for the Russell 2000.
While there is good reason to think that the homebuilders can continue to outperform the market (assuming no meaningful slowdown in the economy), the strongest relative gains have probably already been made. This year’s rebound has brought average forward P/E multiples for the builders up a notch or two, but the sector still trades at only 10.4 times consensus 2019 estimates and 9.4 times projected 2020 consensus earnings. Further multiple expansion probably depends on an improved outlook for U.S. economic growth. Still, the group can deliver high single-digit to low double-digit price returns by meeting or exceeding Street earnings expectations with no multiple expansion.
At this point, I suspect that the greatest upside potential rests mostly with those builders who have to date underperformed their peers. Nine of the eleven builders, all except for HOV and TOL, have outperformed the broader market this year. (TOL still shows a gain of 9.4% year-to-date). The builders that have underperformed the peer group generally have the lowest valuations. They include LEN, PHM and TOL, among the larger cap names, and BZH and MHO, among the smaller cap names. (MHO is up 70.6% year-to-date, but still trades at only 8.4 times expected 2019 earnings and 7.8 times projected 2020 earnings – a big discount to the market and meaningfully cheaper than its peer group.)
To be sure, some of the underperformance for individual builder share prices has been due to disappointing performance. For example, Lennar’s orders were only flat in 19Q2, the worst showing among peers. PulteGroup’s unit closing declined 2.6%, also the worst showing among peers. Beazer’s share price, meanwhile, is still recovering from the impairment charges that it took in 19Q1 (its fiscal 2019 second quarter).
TOL reported the largest percentage decline in 19Q1 net orders (-9.7%) among its peers and is also the only publicly-traded builder without a product push toward the hot first-time buyer market. Still, I anticipate that it will report some improvement in new order trends on the heels of its National Sales Event, when it reports earnings later this month.
If the homebuilders are able to sustain their recent advance, I expect that many of the laggards will catch up, while some of the recent outperformers may see their stocks trade sideways for a while.
August 4, 2019
Stephen P. Percoco
839 Dewitt Street
Linden, New Jersey 07036
© Lark Research. All rights reserved. Reproduction without permission is prohibited.