Also troubling is that homes sold before starting are down over 22% from a year ago, homes sold while under construction down closer to 35%. Homes sold after completion are up about 3% year on year. Builders are assuming lots more risk in this pattern of home sales. Buyers are holding out for finished units, and why not?
In noticed something a little strange in the regional data. It looks like all the weakness in July was in the midwest which was down 21.3%. The south was also weak, down 8.0% but the NE was up 1.8%, and the west was up 11.7%. Anybody know any reason why this should be like that. It also sort of takes the wind out of the theory that the weakness would only occur in the overheated makets on the coasts. Also, anyybody see any data on revisions to previous months data. I know these figures havea fairly large standard error and are often revised, but didn't see the data.
CR, Have you considered including links to real world explanations of the indicators we're relying upon?
From Smartmoney.com Rex Nutting 8/20/06:
"New home sales are reported to the government when a contract is signed, not at the closing, so some of the sales seen in recent months have been phantoms that disappeared when the buyers got cold feet.
The incentives also create a problem with another closely watched statistic: median home prices. If the home builder offers to subsidize a cheaper mortgage, or give away a new car, or provide some other goodie, that would artificially inflate the reported sales price.
Median sales prices for new homes are up just 2.3% in the past year, but could be negative in nominal terms after adjusting for the value of the giveaways."
Dont forget the cancellations are not included in the numbers, ie. they dont revise for the cancellations. So sales are much worse than they appear - if that possible!
looks like 2002 sales volume is just round the corner unless by some miraculous event buying increases signficantly. ain't happening until prices come down 10%+. homebuilder ceo's and cheerleader lereah think that a 5-10% correction will open the floodgates to new buyers taking advantage of a 'one time' dip opportunity. yeah let me stick my hand out as i see the cleaver coming down.
I will not sully myself by addressing the jerk who got his number wrong and then tossed around the offensive (and somewhat obscure) term "progressive serial liar." Rather, I will just note that Bloomberg has the NSA sales pace in July of last year at 117k homes, just as you said. The census website is not responding, so I could not check the original source.
Here's an amusing thought looking at the graphs. Sales appear to be bound in the 400-800K/yr range from '71 to late '95 with recessions in those traditional cycles foreshadowed nicely. They appeared to pick up considerly from '96 to '98 and then flatten off in '98-'01 before taking a real leap. Now we had normal business cycles in the 1rst period, an investment-led boom in the 2nd and of course monetary policy induced housing jumps in the last. What happens if sales revert to the long-run mean ? And if so which mean? That is the 800-1000 mean or the 400-800 mean ?
Rather, I will just note that Bloomberg has the NSA sales pace in July of last year at 117k homes, just as you said.
Hey, jerk yourself. Can you read? Yes the july 2005 number is 117,000... NOW. As late as Sept 2005 the number was being reported as 118,000. Don't tell me you didn't sully yourself. You called me a jerk and then were such a self assured hypocritical ass that you tried to deny it in the next sentence. Nothing you said needed the gratuitous insult. Nothing I wrote deserved the insult. Think twice, post once when starting these things.
The issue of progressive serial lying is explained in the url attached to the original comment. The new homes sales rate has been consistently overreported when preliminary and almost always adjusted down twice in subsequent reports when the data no longer matters except as a way to minimize month over month declines.
There are easier ways to disagree than to open up with "jerk" and then setting yourself on a higher plain by incorrectly claiming to refuse to sully yourself.
You're right. You don't deserve insults. You deserve a spanking and stern words from your mommy. Since she let's you use the internet unsupervised, I did the best I could in her place.
You toss "liar" around - I don't care at whom - and then pretend to be the injured party for being called a jerk? Don't think so. I was being polite.
Hey everyone, Want something to scream about? The Commerce Dept. p.r. statement this morning states the number is 4.3% (+-11.5%) below the revised June rate of 1,120,000.
Just for a little perspective, the "liars" apparently in question revised year ago data by 0.8%. Let's see a show of hands - who knows their own weight, right this very minute, to within 0.8%? Savings account balance? Today's high temperature?
k harris, I think the Census Bureau does the best job they can - I meant to add I'm not comfortable with the term "liar" - but I do think they have a systemic estimation problem during this downturn.
Historically revisions have been either direction; during this downturn, revisions have almost been exclusively down. In fact, I'd be willing to bet that the final number for July will be below tpday's announced number. I doubt I'd find any takers!
k harris,
You don't like the word liar. Fine. You don't even like other people to use the word liar. Your opinion is noted. You've got the evidence in these half dozen related posts of a systematic misrepresentation of market changing data. You may not want to call a persistent and systematic and presistent and easily correctable numerical bias a lie but you've no standing to tell other people to not draw their own conclusions.
We have the same problem in the inflation numbers bias. Anyone who delves deeply into the issue concludes that the CPI is nothing more than a carefully crafted lie.. err... systematic, persistent, easily correctable numerical bias that materially misrepresents market moving data for the direct benefit of those producing the data. Feel fre to use that mouthful but don't pretend you don't recognise it for what it is. I did not use the term lightly and went to great pains to provide the supporting evidence just like my mommy taught me. My mommy also taught me that some people need to hide behind airs of superiority and bellicosity (former English teacher) to cover for their personal inadequacies. My mommy also taught me how to expose would be bullies as the small bundles of insecurity they truly are. Move along, nothing to see here.
Welcome to Limbaugh land, where those who agree with you constitute "anyone who delves deeply into the issue" and everybody else is a liar. You believing there is a "persistent and systematic and persistent and easily correctable numerical bias" doesn't make it true especially the part where it's persistent twice. This is with world of argument by assertion and character assassination, backed by what our psychologist friends call "confirmation bias." You have decided something is true, and only see evidence that confirms what you believe.
Yeah, "liar" isn't really the sort of thing you want to be calling people, especially when they aren't around to defend themselves, but that is far from the entire point. It isn't just small-minded, it's low-brow. Why listen to people who disagree with you when you've got it all figured out. Because frankly, I am not aware of a cogent piece of analysis anywhere in the history of intellectual endeavor that included the proposition "those who disagree are liars" leading on to "and so I'm right - QED."
It's somehow OK for you to mount an argument by character assassination, but heaven forbid that anybody call you on it. Them what disagrees with you are liars, but I have no standing to call you on it? Pathetic. Clean up your mouth.
While you're at it, call the guys at Census and ask 'em how they do their job. I've never had much trouble getting the guys who put the data together to answer questions. What you'll find is that they know tremendously more about the data than you do, including where the biases are in the inputs. Because that's usually where bias comes from. Statisticians don't insert the bias. They fight it. There are some forms of bias you can correct for, some you can't. It's against the rules to just say "oh, that number always gets revised down, so I'll just change it." That would be lying.
CR,
My apologies for being a lightning rod. I've correct that condition. My best wishes that the lightning goes away as well. If not, my sincerest apologies for my accidental dimunition of the your fine blog.
Eeeek...
Jan 2006 - $244,900
Feb 2006 - $250,800
Mar 2006 - $238,800
Apr 2006 - $257,000
May 2006 - $235,800
Jun 2006 - $233,800
Jul 2006 - $230,000
Cutting prices is supposed to make sales go up, not down.
Not good... not good.
Also troubling is that homes sold before starting are down over 22% from a year ago, homes sold while under construction down closer to 35%. Homes sold after completion are up about 3% year on year. Builders are assuming lots more risk in this pattern of home sales. Buyers are holding out for finished units, and why not?
In noticed something a little strange in the regional data. It looks like all the weakness in July was in the midwest which was down 21.3%. The south was also weak, down 8.0% but the NE was up 1.8%, and the west was up 11.7%. Anybody know any reason why this should be like that. It also sort of takes the wind out of the theory that the weakness would only occur in the overheated makets on the coasts. Also, anyybody see any data on revisions to previous months data. I know these figures havea fairly large standard error and are often revised, but didn't see the data.
CR, Have you considered including links to real world explanations of the indicators we're relying upon?
From Smartmoney.com Rex Nutting 8/20/06:
"New home sales are reported to the government when a contract is signed, not at the closing, so some of the sales seen in recent months have been phantoms that disappeared when the buyers got cold feet.
The incentives also create a problem with another closely watched statistic: median home prices. If the home builder offers to subsidize a cheaper mortgage, or give away a new car, or provide some other goodie, that would artificially inflate the reported sales price.
Median sales prices for new homes are up just 2.3% in the past year, but could be negative in nominal terms after adjusting for the value of the giveaways."
Dont forget the cancellations are not included in the numbers, ie. they dont revise for the cancellations. So sales are much worse than they appear - if that possible!
There were 117,000 New Homes sold in July 2005.
Funny, that number didn't show up in the Sept '05 numbers that showed 118,000. Progressive serial lying has become acceptable.
Consider my monthly post repeated here as well.
Calculated Risk: New Homes Sales Revisions
looks like 2002 sales volume is just round the corner unless by some miraculous event buying increases signficantly. ain't happening until prices come down 10%+. homebuilder ceo's and cheerleader lereah think that a 5-10% correction will open the floodgates to new buyers taking advantage of a 'one time' dip opportunity. yeah let me stick my hand out as i see the cleaver coming down.
CR,
I will not sully myself by addressing the jerk who got his number wrong and then tossed around the offensive (and somewhat obscure) term "progressive serial liar." Rather, I will just note that Bloomberg has the NSA sales pace in July of last year at 117k homes, just as you said. The census website is not responding, so I could not check the original source.
Here's an amusing thought looking at the graphs. Sales appear to be bound in the 400-800K/yr range from '71 to late '95 with recessions in those traditional cycles foreshadowed nicely. They appeared to pick up considerly from '96 to '98 and then flatten off in '98-'01 before taking a real leap. Now we had normal business cycles in the 1rst period, an investment-led boom in the 2nd and of course monetary policy induced housing jumps in the last. What happens if sales revert to the long-run mean ? And if so which mean? That is the 800-1000 mean or the 400-800 mean ?
Cuiouser and curioser, said Alice.
Rather, I will just note that Bloomberg has the NSA sales pace in July of last year at 117k homes, just as you said.
Hey, jerk yourself. Can you read? Yes the july 2005 number is 117,000... NOW. As late as Sept 2005 the number was being reported as 118,000. Don't tell me you didn't sully yourself. You called me a jerk and then were such a self assured hypocritical ass that you tried to deny it in the next sentence. Nothing you said needed the gratuitous insult. Nothing I wrote deserved the insult. Think twice, post once when starting these things.
The issue of progressive serial lying is explained in the url attached to the original comment. The new homes sales rate has been consistently overreported when preliminary and almost always adjusted down twice in subsequent reports when the data no longer matters except as a way to minimize month over month declines.
There are easier ways to disagree than to open up with "jerk" and then setting yourself on a higher plain by incorrectly claiming to refuse to sully yourself.
kharris,
I don't think that Robert meant offense to you. I interpreted his comment as a critique of the data that is produced, not of you.
Keith, I read it several times but for me was the description of my "progressive serial liar" phrase as offensive and obscure.
All, please be cool. The criticism was of the downward revisions from the Census Bureau.
Best to All.
Robert,
You're right. You don't deserve insults. You deserve a spanking and stern words from your mommy. Since she let's you use the internet unsupervised, I did the best I could in her place.
You toss "liar" around - I don't care at whom - and then pretend to be the injured party for being called a jerk? Don't think so. I was being polite.
Hey everyone, Want something to scream about? The Commerce Dept. p.r. statement this morning states the number is 4.3% (+-11.5%) below the revised June rate of 1,120,000.
Hello? Plus or minus 11.5% !!!!!!!!!
Just for a little perspective, the "liars" apparently in question revised year ago data by 0.8%. Let's see a show of hands - who knows their own weight, right this very minute, to within 0.8%? Savings account balance? Today's high temperature?
That's certainly not so difficult,
+-10%, is it?
k harris, I think the Census Bureau does the best job they can - I meant to add I'm not comfortable with the term "liar" - but I do think they have a systemic estimation problem during this downturn.
Historically revisions have been either direction; during this downturn, revisions have almost been exclusively down. In fact, I'd be willing to bet that the final number for July will be below tpday's announced number. I doubt I'd find any takers!
Let's all be cool. It's a nice day here.
Best Regards.
k harris,
You don't like the word liar. Fine. You don't even like other people to use the word liar. Your opinion is noted. You've got the evidence in these half dozen related posts of a systematic misrepresentation of market changing data. You may not want to call a persistent and systematic and presistent and easily correctable numerical bias a lie but you've no standing to tell other people to not draw their own conclusions.
We have the same problem in the inflation numbers bias. Anyone who delves deeply into the issue concludes that the CPI is nothing more than a carefully crafted lie.. err... systematic, persistent, easily correctable numerical bias that materially misrepresents market moving data for the direct benefit of those producing the data. Feel fre to use that mouthful but don't pretend you don't recognise it for what it is. I did not use the term lightly and went to great pains to provide the supporting evidence just like my mommy taught me. My mommy also taught me that some people need to hide behind airs of superiority and bellicosity (former English teacher) to cover for their personal inadequacies. My mommy also taught me how to expose would be bullies as the small bundles of insecurity they truly are. Move along, nothing to see here.
Robert,
Welcome to Limbaugh land, where those who agree with you constitute "anyone who delves deeply into the issue" and everybody else is a liar. You believing there is a "persistent and systematic and persistent and easily correctable numerical bias" doesn't make it true especially the part where it's persistent twice. This is with world of argument by assertion and character assassination, backed by what our psychologist friends call "confirmation bias." You have decided something is true, and only see evidence that confirms what you believe.
Yeah, "liar" isn't really the sort of thing you want to be calling people, especially when they aren't around to defend themselves, but that is far from the entire point. It isn't just small-minded, it's low-brow. Why listen to people who disagree with you when you've got it all figured out. Because frankly, I am not aware of a cogent piece of analysis anywhere in the history of intellectual endeavor that included the proposition "those who disagree are liars" leading on to "and so I'm right - QED."
It's somehow OK for you to mount an argument by character assassination, but heaven forbid that anybody call you on it. Them what disagrees with you are liars, but I have no standing to call you on it? Pathetic. Clean up your mouth.
While you're at it, call the guys at Census and ask 'em how they do their job. I've never had much trouble getting the guys who put the data together to answer questions. What you'll find is that they know tremendously more about the data than you do, including where the biases are in the inputs. Because that's usually where bias comes from. Statisticians don't insert the bias. They fight it. There are some forms of bias you can correct for, some you can't. It's against the rules to just say "oh, that number always gets revised down, so I'll just change it." That would be lying.
CR,
My apologies for being a lightning rod. I've correct that condition. My best wishes that the lightning goes away as well. If not, my sincerest apologies for my accidental dimunition of the your fine blog.