Chart Of The Day: Housing Starts Adjusted vs Unadjusted

Tyler Durden's picture

No commentary necessary.

And focusing only on the November to December "comparable" Seasonal adjustments, here is the move from November to December in 2010, 2011 and 2012. Spot the outlier:

Source: Housing Starts NSA; Housing Starts SA, December New Residential Construction

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
zilverreiger's picture

So is that good news or bad news? What is a housing start anyway

Cult_of_Reason's picture

What's up with the huge parabolic DAX spike (from negative to +0.8% within ~3 minutes) and EUR/USD drop?

Stop hunt? Anybody?

Racer's picture

Probably because of someone protection their position for witching tomorrow

Cult_of_Reason's picture

DAX and Copper went up vertical on no news ~3 minutes before official 8:30 am US data, EUR/USD went down at the same time.

redpill's picture

Seasonally adjusted housing starts numbers don't mean much in the wintertime because volume is so low. But, the closing months of 2012 were certainly much better than the same months in 2011 an 2010, which can be corroborated from other sources not tainted by x-12 arima.

Falconsixone's picture

That's just someone playing on their computer. Pay no attention.


or, it's time to crash the dollar (because it's worthless).

fonzannoon's picture

Bro....I'll buy you some champagne Bro...

French Frog's picture


This is just to say that the article at the top left of the main page does not load:

Anyone else having the same issue?

CPL's picture

No problems here.


Go to the page.

Hold down the shift key

and click on the reload button in your brower.


or Clear cache and clear cookies.

Falconsixone's picture

I found the shift key but what's the reload botton on my brower?

qqqqtrader's picture

ctrl + F4?

that should do ya

Raymond K Hessel's picture

Left Right Left Left Up Down Square Square Circle Triangle Down

JeffB's picture

<ctrl> + <R> on Firefox & IE(?)

or right click the tab & select "reload tab".

Cognitive Dissonance's picture

On Internet Explorer it is the round circle with the arrow embedded in it next to the "X" in the box where you type the web site address you wish to visit.

<Was that a test and did I pass? I'm over 50 so I get extra points if I did.>

CPL's picture

The weekly doesn't match with the reality.


Both lines should follow each other like a flock of birds.

EscapeKey's picture

Well, if it's just a linear filter applied, then the only data fields which really matter is the entry which has now slipped off the filter (ie, 12 months ago), and the new entry (ie, this months entry). However, that would suggest truly atrocious data have recently been eliminated, but it would also suggest this recent development is about to come to a sudden end.

A 24-month (or longer, ideally) dataset would be quite interesting, as one should be able to deduce the weights used to calculate the adjusted.

NoDebt's picture

I wish they would start those lines at the same point so I don't do what I just did.  Which was wonder for about a minute "so..... we were building more houses than the seasonally adjusted number in 2011 but we're building fewer houses than the seasonally adjusted number in 2012??"

No, dum-dum (me) that's NOT what the graph says.  It says the seasonally adjusted numbers have been showing a consistent and rising over-estimation of reality.  Oh, OK, that makes more sense, I says to myself.

EscapeKey's picture

First off - I agree. It's complete bollocks.

But ATTEMPTING to justify this graph, imagine if the numbers from 12 months were truly awful. This would mean that if they were now replaced by better numbers, we'd see a strong growth in the adjusted number. Obviously, the weights can't be linear, or you'd just see the average of the unadjusted be the current adjusted, so there's some sort of magic in there. In fact, it's probably magic which is adjusted on a per-month basis (for some dubious reason cocksucking economics students wholeheartedly agree with), and for some even more dubious reason (OMG NATIONAL SECURITY!!eleventyone11) can't be released.

But given the unadjusted set of data, one MIGHT be able to at least come up with some sort of weight (and delta) estimate. Which no doubt would require intense doctoring and curve fitting.

dow2000's picture

they must be getting desperate...

Yen Cross's picture

Renters A-Plenty

LawsofPhysics's picture

Yes, unfortunately, unemployed renters with no wages are a problem (unless they agree to ride a sationary bike and provide power for the rest of the building).

Squid Vicious's picture

In my neck of the woods they are building houses that go unsold for many months ... not sure what that's all about guess it's keeping people busy?

LongSoupLine's picture

I'll comment...

Fuck off NAR, you fucking data stuffing assholes.

Ruger556's picture

in this case, down means up

pods's picture

I think the official NAR term for this phenomenon is "Bi-winning."


Dr. Engali's picture

My house wouldn't start this morning so I adjusted the thermostat.

Yes_Questions's picture



That's fucking funny right there!


teolawki's picture

Hey! Hands off! That's Benny's job!

Headbanger's picture

It's bad news. 

Gold took a hit on it


Me thinks the good SA unemployment claims + housing starts

numbers are just the beginning of the excuses needed for the

Fed to start backing off on their balance sheet which means

margin rates start going up and there goes the punch bowl!

fonzannoon's picture

We are going to spend all year watching the market climb and gold drop on the speculation that QE is no longer needed. We will pay little attention to countries bringing home their gold. Shit they are only bringing it home to make it easier to sell in their "gold for cash" shops that have sprung up everywhere.

I wish we could fast forward to the month after they stop QE.

DeadFred's picture

That way the Fed can buy enough gold to send to Germany.

GetZeeGold's picture



Anytime you manufacture the numbers on jobs you have to take gold's all in the manual somewhere.

Carl LaFong's picture

Seriously? The minute the FED backs off printing money and buying treasuries and interest rates start to rise, the economy and the stock market will crash. You clearly have no idea what's going on. Get a grip.

SheepDog-One's picture

There's got to be VERY little wiggle room for the FED criminals at this point.

azzhatter's picture

Benny's favorite song- Blowing Bubbles

hooligan2009's picture

maybe he gets a "Thriller" nightmareliner dream every night from a monkey.

andrewp111's picture

Notice that the 2 graphs are on diferent scales. The left hand Y scale (blue curve) has 5x the range than the right hand scale. So yes, the slope on teh red curve looks much larger, but it is not. The red curve has a rise of 80 over the whole x range, while the blue curve has a rise of 50.  Sort of what you would expect from seasonal adjustments - which by definition must cancel to zero over a 1 year cycle.

cahadjis's picture

No, thats the percentage scale, look again?

rguptatx's picture

I agree - not sure why the RHS is even there. Both series of data relate to LHS only - hence are internally consistent. 

Tyler Durden's picture

So the same seasonal adjustment that saw the worst NSA number since March result in a SA surge by 100K, was applied in December 2011 when a comparable decline in NSA starts led to a 9k drop in SA numbers a year ago?

Maybe not.

In case there is confusion, here it is visually. Curious how percentages explain the amusing 2012 seasonally adjusted outlier.

Yen Cross's picture

I'm pretty much flat Tyler. I smell a storm a brewing/

catacl1sm's picture

Please do not reply to trolls with facts, it makes their cock limp.

IamtheREALmario's picture

Does anyone else think that it is odd that the seasonal adjustments are very different year-to-year?

Racer's picture

Of course not, it is all smoke and mirrors so par for the CON