Yesterday, the media spin went into full overdrive as the most recent housing data was released on starts and permits. The following is Econoday's summary:
"Don't let the headline fool you, the housing starts & permits report points to solid strength for the housing sector. Starts came in at a 1.036 million rate in May which is down 11.1 percent from the April rate but the April rate, which was already one for the record books, is now revised higher to 1.165 million for, and this is no misprint, a 22.1 percent gain from March. Sealing matters is another gigantic surge in permits, up 11.8 percent to 1.275 million following a 9.8 percent gain in April. Forecasters will be revising their second-quarter GDP estimates higher following today's report, not to mention their estimates for Thursday's index of leading economic indicators where permits are one of the components.
Permits are the leading indicator in the report, and the latest rate is the best since way back in August 2007. The gain is centered in the Northeast followed by the Midwest. Turning to starts, the monthly step back is split between all regions with the Northeast, in contrast to permits, showing the largest percentage decrease."
As usual, there is much more to the story than just the headline data point if we are to understand what is happening within the "real" economy. The following series of charts will hopefully dispel some of these misconceptions and faulty analysis.
Seasonal Adjustments Gone Haywire
The first issue is how the data is reported. Taking a number, annualizing it, and then seasonally adjusting it, distorts the actual data to a great degree from time to time. This is one of the reasons why the BEA is now trying to readjust the calculation of GDP due to "residual bias" in the seasonally adjusted data. The chart below shows the 12-month average of the NON-seasonally adjusted permits as compared to the NON-ANNUALIZED seasonally adjusted data.
(Using a 12-month average smooths the very volatile non-seasonally adjusted data. It also elimates the need for the quirky seasonal adjustment/annualization process that distorts the data by trying to smooth it. Sometimes simple is better.)
As you will notice, the seasonally adjusted data has a tendency to be very "spiky." Those data spikes are revised away in forthcoming reports which is why focusing on a single monthly data point, and the extrapolating a view, can lead to disappointing outcomes in the future.
In May, the non-seasonally adjusted average of building permits was 90,740 permits up from 89,120 permits in April. This was a 1.8% increase in permits for the month which is a far cry from the 11% increase reported by the Census Bureau.
No, It Isn't Housing Story
There are two other problems with extrapolating a view on the housing market from building permit data. The first is that just because a permit is pulled does NOT mean that a house will actually be built. Projects are regularly cancelled for a variety of reasons. However, more importantly, is that the building permit data is muddled by the issue that the permits are not just for single family homes but also multi-family apartments. This is crucially important to understand as the U.S. is now a "renter" nation as I discussed in my recent report on housing.
"There has been very little recovery in housing. With six years of economic recovery now in the rear view mirror, it is clear that the average American is not recovering as evidenced by the lowest level of home ownership since 1980."
"At 63.7%, the current level of home ownership shows the lack of 'true' household formation which is a detraction from longer-term economic growth. The weak rate of growth and widening wealth gap certainly explain the fall in homeownership rates and the lowest level of birthrates on record."
However, the recent reports of sales, starts, permits, and completions have all certainly improved in recent months. Those transactions must be showing up somewhere, right?
The chart below shows the number of homes that are renter occupied versus the seasonally adjusted homeownership rate."
The issue of a "renter nation" was confirmed by yesterday's building permit data which showed that permits for "single family" homes plunged last month. With permits near the lows of the "financial crisis," it begs the question as to where the surge in permits came from.
Oh, there it is. Permits for multi-family structures of 5-units or more surged to a near record high.
The boost to builder optimism has not been due to a surge in demand for single family homes but rather to a massive increase in the building of multi-family apartments.
It has long been hoped that the recovering economy would lead to a demand for single family homes particularly as "millennials" come of age and start to raise families of their own. Home ownership would then lead to demand for everything from furniture to lawn equipment and all the services required to maintain a housing structure in turn boosting economic growth. However, this has yet to be the case.
The problem with the massive build up of multi-unit family structures is that eventually there becomes a "glut" in the number of units available to be rented. While demand is hot, builders tend to overbuild units to excess. The eventual reversions have historically been quite nasty.
Recovery Still Very Limited
Despite the ebullient headlines yesterday, the housing recovery is still very muted especially when considering the enormous lengths the Government, Fannie Mae/Mac and Federal Reserve have gone to in order to ignite that engine of growth. As shown in the chart below, the total housing recovery is still well below historical norms.
Furthermore, as Anthony Sanders pointed out recently, the increase in activity has come at the expense of loosening lending standards back to levels seen prior to the financial crisis.
"Fannie Mae and Freddie Mac are the primary purchasers of single-family mortgages since the housing bubble burst. The FHA is an insurer, not a loan purchaser.
Here is some empirical evidence from Fannie Mae mortgage-backed securities (MBS). Here is the average loan-to-value (LTV) ratio for Fannie Mae 4% coupon MBS."
"At least for 4% coupon Fannie MBS, the average LTV is higher now than during the housing bubble! So much for the story that Fannie and Freddie are "too tight" with mortgage credit.
Now, in terms of credit (FICO) score, Fannie 4% coupon MBS is higher today than during the housing bubble. But it continues to decline."
Importantly, the data clearly shows that to stimulate housing activity it requires ultra-low interest rates and looser credit standards. Given that individuals "buy payments," rather than houses, increases in interest rates have an almost immediate negative impact on housing activity. This is due to individuals being priced out of home buying decisions AND the psychology of "waiting" for rates to fall once again.
With the Federal Reserve now poised to raise rates, and further tighter monetary accommodation, the impact to the housing activity could be more severe than most have anticipated as speculative "buy-to-rent" activity is suppressed.
One thing is for certain; the housing data needs to be viewed very skeptically. The headline data points are not really telling much of the actual underlying economic story.