May 7, 2014

Analyzing the U.S. new homes market is an exercise in frustration

The figures on the U.S. new homes market, while greatly improved from several years ago, are still serving up disappointments.

Housing starts, after exceeding one million units (seasonally adjusted and annualized/SAAR) in three months last year (March, November and December), have tailed off to 923,000 units on average in the first quarter of this year.

In the latest joint >press release from the Census Bureau and Department of Housing and Urban Development (HUD), the number of U.S. new single-family homes sold in March 2014 was -14.5% versus February 2014 and -13.3% compared with March of 2013.

Harsh winter weather in the early months of this year undoubtedly held up residential real estate activity. Nevertheless, the usual spring bounce in the sector has been slow to materialize.

What I’m about to say won’t come as a surprise to anyone. It’s pretty clear the U.S. new housing market is seriously out of equilibrium.

But it’s the extent of one glaring discrepancy that is becoming quite noteworthy. This requires some explanation and background.

Annual average U.S. housing starts have demonstrated remarkable consistency in each of the past five decades.

In the 1960s, the “mean” was 1.406 million units; in the 1970s, 1.768 million; 1980s, 1.492; 1990s, 1.371; and in the 00s, 1.536.

The 00s kicking of this century recorded the greatest volatility, with a dizzying descent from 2.068 million units in 2005 to only 554,000 units in 2009.

The annual average over all five decades (i.e., the full 50 years) was 1.515 million units.

It appears safe to say that 1.5 million units is the “norm” for U.S. housing starts. That figure may even be conservative.

Housing is a “stock”. As the stock rises over time, a fixed rate of demolitions requires a higher replacement number.

Also, while America’s population growth rate has slowed, — it was +0.7% last year — there were still 20-plus million more people living in the country last year who needed additional housing.

The number of new housing units required is dependent on the actual number of new citizens. A gradual change in the population growth rate is more of a long-term influence.

Based on these “stock” effects, an annual level of housing starts between 1.5 million and 1.7 million units seems to be what’s warranted in the mid-2010s.

But maybe Americans won’t be moving around the country to the same degree as in the past. Or maybe the illegal immigrant count has fallen in a manner that’s not widely known.

Therefore, let’s stick with our “trusty” and lower-end 1.5 million units in the arithmetic which follows.

The last time U.S. new home starts exceeded that benchmark was in 2006 (1.801 million units).

Since then, the shortfalls (i.e., versus 1.5 million) have been as follows: 145,000 units in 2007; 595,000 units in 2008; 946,000 in 2009; 913,000 in 2010; 891,000 in 2011; 719,000 in 2012; and 575,000 in 2013.

The total deficit in those seven years combined has been 4.8 million units.

In other words, since 2006, U.S. housing starts have fallen behind their traditional or equilibrium annual level by more than three whole years of groundbreakings.

Can it be that the latest seven-year weakness has been a counter-weight to the exceptional strength that preceded it between 2000 and 2006?

It’s true that in every one of those earlier seven years, annual starts were higher than 1.5 million units, with the largest excess occurring in 2005, at +568,000 units.

The combined surplus, 2000 to 2006, was 2.0 million units.

Subtract 2.0 from 4.8 and that still leaves a deficit of 2.8 million units, or nearly two years’ worth of “normal” starts.

Among the many factors that drive new residential construction, jobs and incomes are near the top of the list. (Mortgage rates are also crucial, but they remain near record lows.)

U.S. total employment has been slow to improve. It has finally returned close to its pre-recession peak.

If U.S. home starts were to return close to “normal”, the resulting rise in construction employment would lift the number of total jobs in the economy well beyond its former peak.

This would put to rest many of the concerns about the labour market.

The day will come — in a year or two (hopefully not any longer) — when we’ll see skyrocketing U.S. home starts.

At that time, analysts will “wisely” (and here’s where you can roll your eyes) point to pent-up demand as the obvious source of the strength.

But for now, we economists are locked in frustration knowing that the forward momentum should be manifest, but that it’s having trouble becoming untethered.

For more articles by Alex Carrick on the Canadian and U.S. economies, please see his market insights. Mr. Carrick also has an economics blog.

U.S. new home inventory - February 2014

Based on seasonally adjusted data (single-family housing).

Data source: U.S. Census Bureau and U.S. Department of Housing and

Urban Development/Chart: Reed Construction Data - CanaData.

Print | Comment