Preliminary estimates indicate
that US homebuilders began construction on a seasonally adjusted 1.036 million
homes at an annualized rate in March, the most since June 2008. While pundits
and prognosticators are likely to trumpet such numbers as an indication of a
strong housing recovery, as has so often been the case during this recovery,
the underlying data is weaker than the headline figure suggests. First, the
strength in the headline figure is entirely the result of strength in the
multi-family sector of the market, which includes townhouses, condos, and
apartment buildings. Single-family housing
starts actually declined in March, and, aside from February, have been
essentially flat over the last four months. As housing starts began to gain
strength over the last two years, multi-family starts as a percentage of total
starts have grown. In monthly data from 1959 through 2012, multi-family starts
average 24% of all starts. However, in 2012 alone, they averaged 30% of starts.
They represented 38% of all starts last month. Multi-family starts are far more
volatile than single-family starts. As such one must be careful in ascribing strength
to a market in which the volatile multi-family sector is a growing portion.
A deeper underlying concern is
the slowing growth in housing permits, an indicator of future starts. While
monthly permits grew rapidly during the first nine months of 2012, the growth
rate in permits issued has been essentially flat for the last six months
outside of a jump in February. This suggests that starts will likely grow at a
slower rate in the coming months.
Finally, it is important to put
the improvement we have seen in housing starts over the last two years into a
larger, long-term perspective. Starts have improved dramatically, more than
doubling from the historic lows reached in 2009. This is unambiguously positive
on an absolute basis; i.e. more starts mean more economic activity than less
starts. Yet, note that prior to the
crisis, monthly starts only fell below 1,000,000 on a seasonally-adjusted
annualized basis in five months, only one of which was lower than 921,000. In
short, despite the recovery, the current level of starts would rank among the
lowest from 1959 through mid-2008. As such, the impact on US GDP growth has been
modest. If the rate of growth decreases, as permit data suggests, the impact
will be more modest yet.
The US housing market has shown
signs of recovery over the last two years and this improvement is welcome.
However, housing starts remain among the lowest levels in the nearly 50 years
of data prior to the credit crisis. Additionally, over the last two years, the
highly volatile multi-family sector has grown significantly as an overall
percentage of housing starts. Finally, over the last six months, the growth in
permits has slowed significantly, suggesting that the growth rate in starts
will slow significantly in the coming months. Residential housing is a far
smaller part of the US economic pie than it had been in 2006 and there are
signs that the relatively rapid growth in housing starts we experienced over
the last two years is slowing. All of this suggests a healthy dose of
skepticism over the perceived positive impacts of an improved housing market on
US economic activity.