For anyone who wants to know, finding out and arguing all of the seasonal adjustments being used...
start with ARIMA and there is even free software from here.
But it's our friend Gaussian distributions and so on, although such a large seasonal layoff such as the auto plants assumed to be shutting down for retooling....
I don't think one necessarily has to run this but at least understand seasonality
and realize this is done over time, a long time by running auto correlations with the past...
so a large anomolie like the entire auto industry shutting down months early....
would throw the numbers for seasonal...while not being the total seasonal adjustments.
Anywho if one really wants to Q.E.D. in this argument, you can literally dig deep into the seasonal methods for employment in the U.S.
After I wrote this, I look over to read Naked Capitalism also pointing to oil, commodities as deserving a larger consideration in a potential W.
The NC piece is worth the read for they note the headline sounds soft enough, but in truth the FT Roubini piece has a series of points that are not so uuuuy sounding.
(read the full FT Roubini piece too, hell, read the web, read books, read everything including the instructions on the vacuum cleaner, what the hell).
time out guys, OK? I enjoy both arguments amd both perspectives, and I don't want to have keep chasing you both around the net just to be able to get your perspectives.
I've been watching this mess and reading a number of sites, Including Bonddad, and I think he has an emotional investment in the success of this administration's approach. I don't get the impression he sees what the situation on the ground is--the map is not the terrain. why would we want to return to the status qou ante? until we clear the ledgers of all the red ink, we are fu**ed long term. Short term bounce,sure, but it's like sanding an icy road- we may not fall flat on our ass, but we're sure as hell not going sprint back to prosperity
you mixed initial with total unemployment predictions which middle was implying with the seasonal auto worker adjustment that was already previously done earlier in the year. (and I still don't know those raw numbers btw)
Below are weekly initial claims, not seasonally adjusted, seasonally adjusted and moving average, for May 1, 2009 to August 15, 2009.
So, I think I have narrowed down this argument. middle is referring to just the auto layoffs and not to all of the seasonal adjustments done in the entire BLS time series for unemployment and various labor markets due to cyclical seasonal conditions.
So, the next question is....what specifically are the seasonal adjustments, across the entire labor market and what percentage of that is for auto layoffs they normally assume take place during the summer for retooling?
and we have a host of posts, from all people in this "heated discussion" I might add, who are all demanding some sort of modern day public works program to get U.S. citizens/LPR income.
Yet on the MSM, all they can talk about is the consumer! Jack asses, that consumer is your neighbor who was just foreclosed on!
of double dip recession is growing. Man, we are sooo screwed but we are arguing over unemployment data points.
There are other macroeconomic factors - how about the fact that consumer spending makes up 70% of gdp but we have households still loaded w/debt. We have a financial oligarchy that is still running unrestrained. We have wage stagnation and growing income inequality. We are about to experience 'purgatory' or 'new normal' and there are ABSOLUTELY NO policy proposals addressing that.
if you think I'm wrong, prove me wrong with a calculator.
What in either of your two posts do you believe I am misprepresenting? Show me the numbers.
Right now here's what I see: you believed that non-seasonally adjusted initial jobless claims showed that the "real" jobless claims number had bottomed in June and was moving up, and showed the "real" trend. Now you are saying that it's the seasonally adjusted number that are the "real" numbers, and bottomed in July. Which is it?
Or, put another way, what number, seasonally adjusted or unadjusted, at what time would prove you wrong or prove you right?
now come on folks, what is all of this about anyway?
I don't really understand what is going on here as of late...
why the hostility, the attacks...
you guys, it's friggin' numbers!
I don't want to ban anyone, I want everyone to return to their logical brain and start paying attention to the points.
I just did and ya know what am I missin' here?
Also, may I point out that within Academia circles, one will deal with many a critic, often public and often to incense. Don't let things get under your skin in so many words or baby with bath water.
1. Every single one of you is wrong in using the Black Swan analogy for simply trying to play pin the tale on the end of the economic recession. Two very different concepts here.
2. Am I missing something? I mean even the main stream economists are predicting unemployment will rise? Are they pulling that out of their ass? (like many things they do pull out of there)?
3. How come I'm staring at these statistics and also seeing recently initial claims weeklies and seeing a rise?
$. What are you guyings going to town over the unemployment numbers? In all seriousness, it's like a public brawl of some burly dudes all the while the house they are fighting in front of is burning down! Everybody get over this data point argument and start looking at the same we got > 10% plus unemployment in 26 states, 16% + U6 and flip out! That's incredible! Let's all talk about that!
Now, look, I believe bonddad's analysis is dead wrong in many ways and that's why EP exists, so we can argue on these things and look at different aspects.
I wrote up many posts on the various EIs and said, ya know what, this is the end of the cliff but sure doesn't look like any recovery in that you get a "V" with GDP correlated to decreased unemployment and so on that is trying to be implied.
But what I don't get is jumping on this whole black swan thing...anybody really read the book?
Or do they all just think Taleb is an asshole (I get a strong impression that's true), hence trying to dismiss his very good insights into structured finance, derivatives and mathematical modeling?
By the way, what is this continual use of the Black Swan? The book is about derivatives really, structured finance, use of Gaussian distributions in mathematical models that are not valid...as well as those events in the "tails" of Gaussian distribution too often dismissed as improbable.
I would say the money market situation that Bernanke guaranteed was indeed a black swan and also what the book is talking about...
on the other hand, why is this term being used in reference to trying to play the game pin the tail on the recession end point?
firstly I will be moving this entire thread to middle's post...
but secondly, I think one needs to explicitly talk about initial claims vs. ongoing unemployment.
Unemployment is rising. We have some new data that doesn't look good, we'll see when the monthlies come out but the actual rate is increasing, plus people are falling off of the roles...
So, can we please differentiate between initial claims vs. the overall U3, U6 seasonally adjusted or not?
this set of numbers is produced through the manipulation of the actual numbers to conform to a statistical model that attempts to remove seasonal variance to make numbers comparable across seasons. Strip that statistical model away and the numbers tell a far different story.
The advance number of actual initial claims under state programs, unadjusted, totaled 667,534 in the week ending July 11, an increase of 86,389 from the previous week. There were 483,981 initial claims in the comparable week in 2008.
Say again what. The numbers fell, but they actually rose? So which one is right. The hard answer is that they both are supposed to be, but they clearly aren't. Why?
Because the expectations ingrained in the model are wrong. It's a baby black swan....
....
If we mock up the seasonally adjusted and real numbers since the start of the year we get this:
[graph]
Again, there are two different stories, if you want a happy ending you use the blue, seasonally adjusted line. If you want to be more honest about what's happening you use the red [unadjusted] line, and you see that unemployment is actually on the rise after having settled down.
==============
You clearly and unequivocally said that the seasonally adjusted numbers were hiding the "real" upward trend that was revealed in the unadjusted numbers.
So, how has that worked out? Here are the raw weekly, unadjusted jobless claims numbers that you said in early July were the "truth" and were going to head back up:
Notice that in the latest week we've had the lowest jobless claims of the entire year 2009. So the thesis of your diary from last month was wrong.
And with your diary today, your logic admits it:
The same flaw in the seasonal adjusting formula that created the appearance of falling initial claims while they were actually rising is now striking back. With the exception of some very small job increases coming from the CARS program that will most likely materialize in September, there is no huge increase in auto employment. Because the layoff didn't happen at the normal time. And there is going to be a layoff at GM either at the end of September or early October.
The bottom line is that things weren't getting better in July, and they are getting worse right now.
In other words, all of a sudden it is the "seasonally adjusted" numbers which are the "real" numbers, and the unadjusted numbers which are obscuring the truth.
You can't have it both ways. Either you stick with the unadjusted, which went up last month and are dramatically down now, or you go with the seasonally adjusted numbers.
On that score, let me reiterate your conclusion from your diary last month:
unemployment is actually on the rise after having settled down
We already know that unadjusted claims are down dramatically since, but was June the month of "settling down" after which jobless claims were going higher? Here's the seasonally adjusted data starting from the June "settling down" data point:
The numbers simply show that your thesis was incorrect. At no point since June has the average approached the 600,000+ level at which it had, in your terms, "settled down", and from which it was "actually on the rise."
If the 4 week moving average, seasonally adjusted, moves up over 618,000 you've got a point. Right now what I see is an average which is consistently lower than June's and is moving sideways in the 560-570,000 range.
I'm sorry, but as of now your diary's thesis has to be regarded as a dead black swan.
although what a great phrase. I also like Reich's "out of the woods means you're out to sea".
I don't think it's a "personal attack" but I do believe some people became too emotionally vested in to the predictions.
If it makes you feel better I've seen similar type of things with Jerome A Paris and even midtowng.
Even better on the financial blogs is a major war on the effectiveness of the Stimulus with me, sitting in the peanut gallery wondering if a little sexism is coming into play (although I think the entire multiplier, grand scale "throw money at it" is wrong, and should be very targeted government expenditures, mostly public works or production investments of some sort)
I saw that whole thing and thought, ya know, shall we continue to work on our statistics and understanding everyone?
But on the other hand, when things are mathematically, logic, analytical driven, I have always been astounded at the politics.
Seriously, in Engineering, I'm like, "hey awesome idea, great design, obvious improvements" yet if you are in a bad team, bad whatever, someone will pound on the person with the great idea simply due to ego issues.
On the other hand, I've seen people fight over which way the toilet paper is supposed to go on the roll. So....go figure.
I'm thrilled for I cannot tell you how many times on EP I had said we need a new term, a new official term for the oxymoron, jobless recovery so honestly I hope the phrase is truly coined! ;)
This post was also reposted by Calculated Risk, which is fine by me. Hello Calculated Risk readers!
For anyone who wants to know, finding out and arguing all of the seasonal adjustments being used...
start with ARIMA and there is even free software from here.
But it's our friend Gaussian distributions and so on, although such a large seasonal layoff such as the auto plants assumed to be shutting down for retooling....
I don't think one necessarily has to run this but at least understand seasonality
and realize this is done over time, a long time by running auto correlations with the past...
so a large anomolie like the entire auto industry shutting down months early....
would throw the numbers for seasonal...while not being the total seasonal adjustments.
Anywho if one really wants to Q.E.D. in this argument, you can literally dig deep into the seasonal methods for employment in the U.S.
After I wrote this, I look over to read Naked Capitalism also pointing to oil, commodities as deserving a larger consideration in a potential W.
The NC piece is worth the read for they note the headline sounds soft enough, but in truth the FT Roubini piece has a series of points that are not so uuuuy sounding.
(read the full FT Roubini piece too, hell, read the web, read books, read everything including the instructions on the vacuum cleaner, what the hell).
time out guys, OK? I enjoy both arguments amd both perspectives, and I don't want to have keep chasing you both around the net just to be able to get your perspectives.
I've been watching this mess and reading a number of sites, Including Bonddad, and I think he has an emotional investment in the success of this administration's approach. I don't get the impression he sees what the situation on the ground is--the map is not the terrain. why would we want to return to the status qou ante? until we clear the ledgers of all the red ink, we are fu**ed long term. Short term bounce,sure, but it's like sanding an icy road- we may not fall flat on our ass, but we're sure as hell not going sprint back to prosperity
you mixed initial with total unemployment predictions which middle was implying with the seasonal auto worker adjustment that was already previously done earlier in the year. (and I still don't know those raw numbers btw)
Below are weekly initial claims, not seasonally adjusted, seasonally adjusted and moving average, for May 1, 2009 to August 15, 2009.
So, I think I have narrowed down this argument. middle is referring to just the auto layoffs and not to all of the seasonal adjustments done in the entire BLS time series for unemployment and various labor markets due to cyclical seasonal conditions.
So, the next question is....what specifically are the seasonal adjustments, across the entire labor market and what percentage of that is for auto layoffs they normally assume take place during the summer for retooling?
and we have a host of posts, from all people in this "heated discussion" I might add, who are all demanding some sort of modern day public works program to get U.S. citizens/LPR income.
Yet on the MSM, all they can talk about is the consumer! Jack asses, that consumer is your neighbor who was just foreclosed on!
of double dip recession is growing. Man, we are sooo screwed but we are arguing over unemployment data points.
There are other macroeconomic factors - how about the fact that consumer spending makes up 70% of gdp but we have households still loaded w/debt. We have a financial oligarchy that is still running unrestrained. We have wage stagnation and growing income inequality. We are about to experience 'purgatory' or 'new normal' and there are ABSOLUTELY NO policy proposals addressing that.
Carry on.
RebelCapitalist.com - Financial Information for the Rest of Us.
MfM's diaries are about initial jobless claims. The unemployment rate is another beast entirely.
if you think I'm wrong, prove me wrong with a calculator.
What in either of your two posts do you believe I am misprepresenting? Show me the numbers.
Right now here's what I see: you believed that non-seasonally adjusted initial jobless claims showed that the "real" jobless claims number had bottomed in June and was moving up, and showed the "real" trend. Now you are saying that it's the seasonally adjusted number that are the "real" numbers, and bottomed in July. Which is it?
Or, put another way, what number, seasonally adjusted or unadjusted, at what time would prove you wrong or prove you right?
now come on folks, what is all of this about anyway?
I don't really understand what is going on here as of late...
why the hostility, the attacks...
you guys, it's friggin' numbers!
I don't want to ban anyone, I want everyone to return to their logical brain and start paying attention to the points.
I just did and ya know what am I missin' here?
Also, may I point out that within Academia circles, one will deal with many a critic, often public and often to incense. Don't let things get under your skin in so many words or baby with bath water.
1. Every single one of you is wrong in using the Black Swan analogy for simply trying to play pin the tale on the end of the economic recession. Two very different concepts here.
2. Am I missing something? I mean even the main stream economists are predicting unemployment will rise? Are they pulling that out of their ass? (like many things they do pull out of there)?
3. How come I'm staring at these statistics and also seeing recently initial claims weeklies and seeing a rise?
$. What are you guyings going to town over the unemployment numbers? In all seriousness, it's like a public brawl of some burly dudes all the while the house they are fighting in front of is burning down! Everybody get over this data point argument and start looking at the same we got > 10% plus unemployment in 26 states, 16% + U6 and flip out! That's incredible! Let's all talk about that!
Now, look, I believe bonddad's analysis is dead wrong in many ways and that's why EP exists, so we can argue on these things and look at different aspects.
I wrote up many posts on the various EIs and said, ya know what, this is the end of the cliff but sure doesn't look like any recovery in that you get a "V" with GDP correlated to decreased unemployment and so on that is trying to be implied.
But what I don't get is jumping on this whole black swan thing...anybody really read the book?
Or do they all just think Taleb is an asshole (I get a strong impression that's true), hence trying to dismiss his very good insights into structured finance, derivatives and mathematical modeling?
either you ban New Deal Democrat, or I leave Economic Populist forever.
He was given the opportunity to admit that he had made a mistake in being fooled by the data, but he's chosen to try to spin what he's saying.
He's either lacking in intelligence or integrity, and I'm not going to stick around here if he's going to keep pulling this shit.
By the way, what is this continual use of the Black Swan? The book is about derivatives really, structured finance, use of Gaussian distributions in mathematical models that are not valid...as well as those events in the "tails" of Gaussian distribution too often dismissed as improbable.
I would say the money market situation that Bernanke guaranteed was indeed a black swan and also what the book is talking about...
on the other hand, why is this term being used in reference to trying to play the game pin the tail on the recession end point?
firstly I will be moving this entire thread to middle's post...
but secondly, I think one needs to explicitly talk about initial claims vs. ongoing unemployment.
Unemployment is rising. We have some new data that doesn't look good, we'll see when the monthlies come out but the actual rate is increasing, plus people are falling off of the roles...
So, can we please differentiate between initial claims vs. the overall U3, U6 seasonally adjusted or not?
Here's you in your diary from early July:
==============
You clearly and unequivocally said that the seasonally adjusted numbers were hiding the "real" upward trend that was revealed in the unadjusted numbers.
So, how has that worked out? Here are the raw weekly, unadjusted jobless claims numbers that you said in early July were the "truth" and were going to head back up:
2009-06-20 568552
2009-06-27 559857
2009-07-04 581145
2009-07-11 671242
2009-07-18 585575
2009-07-25 511628
2009-08-01 466695
2009-08-08 482622
2009-08-15 454216
Notice that in the latest week we've had the lowest jobless claims of the entire year 2009. So the thesis of your diary from last month was wrong.
And with your diary today, your logic admits it:
In other words, all of a sudden it is the "seasonally adjusted" numbers which are the "real" numbers, and the unadjusted numbers which are obscuring the truth.
You can't have it both ways. Either you stick with the unadjusted, which went up last month and are dramatically down now, or you go with the seasonally adjusted numbers.
On that score, let me reiterate your conclusion from your diary last month:
We already know that unadjusted claims are down dramatically since, but was June the month of "settling down" after which jobless claims were going higher? Here's the seasonally adjusted data starting from the June "settling down" data point:
2009-06-06 622750
2009-06-13 616750
2009-06-20 618000
2009-06-27 616000
2009-07-04 607000
2009-07-11 585000
2009-07-18 567250
2009-07-25 560250
2009-08-01 556500
2009-08-08 565750
2009-08-15 570000
The numbers simply show that your thesis was incorrect. At no point since June has the average approached the 600,000+ level at which it had, in your terms, "settled down", and from which it was "actually on the rise."
If the 4 week moving average, seasonally adjusted, moves up over 618,000 you've got a point. Right now what I see is an average which is consistently lower than June's and is moving sideways in the 560-570,000 range.
I'm sorry, but as of now your diary's thesis has to be regarded as a dead black swan.
although what a great phrase. I also like Reich's "out of the woods means you're out to sea".
I don't think it's a "personal attack" but I do believe some people became too emotionally vested in to the predictions.
If it makes you feel better I've seen similar type of things with Jerome A Paris and even midtowng.
Even better on the financial blogs is a major war on the effectiveness of the Stimulus with me, sitting in the peanut gallery wondering if a little sexism is coming into play (although I think the entire multiplier, grand scale "throw money at it" is wrong, and should be very targeted government expenditures, mostly public works or production investments of some sort)
I saw that whole thing and thought, ya know, shall we continue to work on our statistics and understanding everyone?
But on the other hand, when things are mathematically, logic, analytical driven, I have always been astounded at the politics.
Seriously, in Engineering, I'm like, "hey awesome idea, great design, obvious improvements" yet if you are in a bad team, bad whatever, someone will pound on the person with the great idea simply due to ego issues.
On the other hand, I've seen people fight over which way the toilet paper is supposed to go on the roll. So....go figure.
I bet this phrase catches on.
Krugman must check the posts that link to him.
Friggin awesome.
I'm thrilled for I cannot tell you how many times on EP I had said we need a new term, a new official term for the oxymoron, jobless recovery so honestly I hope the phrase is truly coined! ;)
This post was also reposted by Calculated Risk, which is fine by me. Hello Calculated Risk readers!
that's where I got it from
Pages