Here’s what we know:
1) Over the past 6 years, we’ve built far fewer single family houses than we have since at least the 1940’s. We’re undersupplied.
2) The Millennial Generation is slightly bigger than the Baby Boomer generation, and the peak population cohorts of the Millennials are currently in their early 20s. The closer this cohort gets to its early 30s the more single family housing they’ll demand. The demographic demand story for housing over the next decade is the strongest it’s been in 20+ years.
3) The housing stock we have is old. The decade that produced the largest number of homes still standing is the 1970’s. In addition to adding new supply we’ll have to tear down and replace old supply (chart below h/t George Pearkes).
When you combine the undersupply from the past decade, the coming demand from the next decade, and the age/condition of the existing stock, it should be a no-brainer to expect housing construction activity to be picking up significantly right now, and it’s not. So what’s going on?
The problem, as Evan Soltas pointed out so clearly yesterday, is that housing is the weakest link in the economy right now, not employment. Households will only move out of their apartments and parents’ houses into homes of their own if they have [good] jobs. Builders will only build (and hire construction workers to build) when there are buyers with jobs looking to buy. But what do you do when the missing jobs are in the housing sector? It becomes a chicken and egg problem.
To get an idea of what I’m talking about, here’s a look at residential fixed investment as a % of GDP compared to the unemployment rate for when residential fixed investment as a % of GDP is less than 4%. You can see that 4% is historically very low:
As you’d expect, there’s generally a strong relationship between the unemployment rate and the employment-sensitive housing sector. I shaded past cycles different colors to highlight contrasts, and only included the past 6 quarters for this cycle.
I bolded 2 quarters in 1982 because those are the only two quarters prior to this cycle where RFI/GDP was as low as it is now. In those 2 quarters the unemployment rate was 9.6% and 10.1%. In the first half of 2014, with similar levels of residential fixed investment activity, the unemployment rate was 6.7% and 6.1%. Similar level of housing activity, much lower unemployment today.
A look at the 1974-95, 1981-83, and 1991-92 periods show housing activity levels higher than we’re seeing today, but also higher unemployment rates. Ex-housing, employment really is quite robust in 2014.
What about those first periods from 1966-1970? They, I believe, hold the key to our future. In 1966-67 the unemployment rate was very low, sub-4%, while housing activity was also low. That was also the period when the older Baby Boomers were in their early 20s. Demographers could’ve told you that we’d need to build a lot of houses in the 1970’s (and as mentioned above, they did). But where were they going to find the workers to do so when the unemployment rate was sub-4%?
All of a sudden the inflation in the 1970’s makes more sense — employers were in a vicious competition for workers to do all the work needed that decade.
At the moment, the job market isn’t strong enough to get builders off the sidelines and building. The rental market is tight, as is the homebuying market, especially at the low end. By the time the job market and wage growth are strong enough to induce builders to get back to mid-cycle levels of production — forget make-up production or an overshoot — it’ll be clear just how inadequate the labor supply is to meet this demand, and we’ll get wage growth that we haven’t seen in 20+ years.