“Believe none of what you hear, and only half of what you see.”
-Benjamin Franklin (maybe, if you believe that)
So, the rash of recent housing data has (again) successfully managed to steep investors in an paralytic bath of informational ambiguity.
- We had the strongest sales figures since 2007 in 1Q17 but then demand seemingly rolled over according to the latest April volume data.
- Household formation is either accelerating (HVS Survey) or cratering (CPS Survey) depending on which survey you believe.
- Supply is at all-time tights which will continue to constrain the upside in demand growth but will also support continued acceleration in price growth which, historically, is positive for housing related equities.
- Prices continue to rise at a positive spread to incomes and there are overheated regional markets but affordability is still very good nationally.
- Mortgage Interest rates are off the highs but also off the lows and the recurrent ebb and flow of interest rate angst can dominate short-term price action in the equities … or not, depending on which way the winds of investor whimsicality happen to be blowing at the time.
- First-time buyer demand continues to improve but immigration trends may be worsening.
- Lot supply & regulatory pressures continue to persist but those are known knowns, right?
- Labor cost pressure may be easing on the margin, but what about rising material costs and the flow through impact of that 20% tariff on Canadian soft lumber imports?
But you already knew all that.
I mean who doesn’t have an extra 80 hours a month to devote to properly distilling and contextualizing the dizzying deluge of high-frequency housing data.
It’s not like reconciling 30 monthly data points of varying import, all of which are out-of-phase in terms of the month they are describing, is some Gandalf’ian analytical-logistical exercise.
And housing certainly isn’t locked in a delicate supply-demand-price dance that could tip out of balance at any point and shift the rate-of-change scales in a way that carries any real investment implications.
Besides, not giving your kid a bath is the new hipster Millennial social meme and doing laundry is for losers anyway.
And everyone knows “needing sleep” is just part of the deep state conspiracy. It’s all mental, just walk it off.
Anyway, back to the Domestic Macro Grind…
Across the Housing space, the last couple days saw these two beauties on the demand front:
- New Home Sales (NHS) in April fell -11.4% sequentially and decelerated to just +0.5% YoY, the slowest pace of growth in 15-months.
- Existing Home Sales (EHS) fell -2.3% sequentially and decelerated -400bps to +1.6% YoY in April.
Did housing demand suddenly atrophy in April?
We think not. More probably, it’s a statistical distortion stemming from the shifting Easter Holiday.
We discussed the Easter Distortion ahead of the data in our 2Q Housing Themes call on Monday but it’s worth recapitulating for broader consumption.
The gist of it is fairly straight forward:
Statistical seasonal adjustments – which are supposed to adjust for seasonality in demand, the number of working days in the month and holiday and other calendar effects – incompletely correct for the shifting Easter holiday.
This dynamic, empirically, manifests as optically disappointing growth in the month containing Easter and rebound strength in the subsequent month as the effect resolves.
Take last year, for example, as we saw a similar dynamic unfold.
When Easter shifted from April to March, April NHS printed its largest gain in 2 decades … the punditry heavens parted, the industry celebrated the long-awaited housing breakout and Panglossian housing bulls frolicked blissfully with virgin unicorns in the re-fertile fields of realized market renormalization.
Then …. May reversed April’s gain and both March and April were revised lower.
Some quick technical background may help the contextualization:
The preliminary Monthly New Home Sales data are derived from the Census Bureau’s Survey of Construction (SOF) and are imputed based on the issuance of permits. In other words, it’s a derived measure with significant standard error and not some direct count of contract signings.
Further, NHS carry the largest revision of any residential construction or volume related series - a reality which stems largely from the fact that the count is based on permit issuance and many homes have a sales contract prior to a permit being issued.
In last April’s false breakout, New Home Sales were originally estimated to be +619K. A year later, the “official/final” estimate now sits -9% lower at +566K.
Today’s Chart of the Day below visualizes the Distortion dynamic.
It’s a histogram showing the distribution of sequential percentage changes (M/M % Chg) in NHS from 2001-Present.
As can be seen, the black bars – which represent the full data set - are approximately normally distributed with a small positive central tendency, which is what you would expect to see.
We then plotted the April values for years in which Easter occurred in April. Those are shown in red. While the 'n' isn’t particularly large, the distribution is interesting as it’s clearly skewed left, meaning that those April's seem to be consistently worse than the broader population.
Meanwhile, the distribution for May plots with positive skew as the distortion reversal generally supports sequential strength.
Now, to be clear, we’re not expecting some moonshot acceleration chased by further step function increases in demand growth.
We are, however, expecting the tyranny of Easter's distortive effects to mostly reverse next month followed by steady modest-to-moderate improvement similar to that which has characterized the last couple of quarters as demand remains robust and supply shows further, trudging improvement.
Our immediate-term Global Macro Risk Ranges are now:
UST 10yr Yield 2.18-2.39%
Oil (WTI) 47.03-52.28
May the data set you free …. or get you more confused … depending on if it’s believable which it usually is, sometimes.
Christian B. Drake
U.S. Macro Analyst