Another reason is that completions are being accelerated to get land off the homebuilders' books. Homebuilders are having a liquidity crunch. Land is a very illiquid asset. And right now there are no other builders who want to buy lots. The best way to sell land is to build a house on it and sell the house. In the meantime, before the house sells you can get construction financing, generating some cash for current operations.
Is it possible the BLS is simply lying? I mean, it wouldn't be the first time the Bush administration has cooked the numbers for political reasons.
It's getting to the point that I am beginning to doubt ALL government statistics. Remember, a recent executive order allowed Bush to appoint political 'commissars' at all agencies, ala the Soviet Union. Why should we believe that unemployment fell AGAIN when all empirical evidence indicates in should have risen?
But even if you look at completions, wouldn't you still expect more job losses at this point?
I've got another theory that you sort of touched on. From what I've read, the high-end market is holding up much better than the low-end RE market. So the starts that have dropped off the most are the ones that require the least amount of workers to complete. Similarly, there is a greater than normal percentage of high-end homes in the total recent starts.
What do you think? Plausible? Not plausible?
Is there a way to check on the values of new home starts over the last year?
So what you're saying is you believe parts of the employment data are legit (the 374,000 increase in those out of the labor force), but they're lying about other parts (the construction job losses). Is that it?
So what you're saying is you believe parts of the employment data are legit (the 374,000 increase in those out of the labor force), but they're lying about other parts (the construction job losses). Is that it?
This is the problem that is presented when we have an administration that has lied from day one til today. From the war in Iraq to the FBI illegal activities that were denied by all until today. What to believe? When the stock market and low unemploymnet rate are the only thing this administration has to brag about, it is time to examine the teeth in those two horse's mouths.
Steve
Many of the internal numbers are probably correct. It is the overall interpretation of those numbers that is suspect.
Without having seen the original document, I would guess the BLS news release did not mention anything about the effect of the huge decrease in the labor force on the unemployment rate. This would be a significant omission.
Furthermore, how could the household survey possibly come up with a legitimate increase in employment in the RE sector, when virtually everything about it has been imploding recently?
There's just too much inconsistency for the overall conclusions to be believable. However, you are free to believe them, of course. I will act accordingly ... and have been selling stock into this dead cat bounce.
According to Perma Bear the answer to this question is simple: You believe the parts you like and disregard the parts you don't like as "fixed."
It's one thing to be suspicious of government numbers, but it's a whole different ballgame to be suspicious and supportive at the same time, depending on whether or not such data backs your view.
Let me ask you this. If you were to fudge the household survey so that it would yield a lower unemployment rate, would you do it by fixing the number of jobs that were created or the number of people who left the workforce?
Steve - according to my model, building is still very high. However, it's not because of SFR, but because of a higher volume of starts of large multi-family.
Now, once I reloaded everything and went into look at what it was actually doing, I was stricken with horror. It's calculating it five different ways, tossing up to two fliers, comparing to construction value numbers. Also, it's using some pay-for-play numbers from NAHB. The purpose of that module is actually to project risk forward over the life of a projected C&D loan from initiation through final sale rather than actual construction employment, although over a quarter there is necessarily a correlation.
The only thing I can say is that it has worked in the past. What it does show now is that construction is proceeding at a rapid pace, although I am guessing that a lot of these larger projects will drop out in various areas. One interesting thing it shows is that unskilled labor should be dropping out, whereas skilled labor should be in relatively high demand. In some localities construction is already very slow, but in others it is still very high.
Just one thing I wanted to point out. While I agree that official numbers are almost certainly undercounting layoffs because the illegals are probably the first to go...
Remember that the HIRING of the illegals also didn't show up in the data during the expansion. In other words, the distortions caused by the illegals should be equal on the way up and the way down.
Thus, the firing of the illegals shouldn't be a reason why layoffs don't seem to be tracking starts that well - yet.
Steve,
I don't think the Bush alters unemployment statistics. The BLS has the same methods (and sometimes management) they've had for decades.
Maybe the Bush administration would like to manipulate the data, but they are at least being watched by the other party that would enjoy seeing the numbers appear worse.
Remember that the HIRING of the illegals also didn't show up in the data during the expansion.
and he needs to know (and remember of course) that the recent benchmark revisions (~1M added) purport to take that illegal hiring (this B what it is folks: unprosecuted illegal hiring by specified and self-identified companies) into account.
Some accounting.
As I have said before, your expectation of a fast drop in construction jobs does not match what has happened in past housing downswings. In prior housing downturns the loss of 700k to 1 million jobs happened over 25 to 30 months and resulted in about 30k of job losses per month. The job losses only happened late into the housing downturns.
This data can be found at Economagic: Economic Time Series Page by looking at the BLS US NonFarm Employment Construction data. For example in the housing downturn of 1973-1974 that is on your graph, housing starts peaked in Jan of 1973. But construction jobs peaked in December 1973 at 4,239,000. The low point was reached in July of 1975 at 3,501,000 - a lost of 740,000 jobs in 19 months. That comes out to 39K jobs/month.
From your graph this was a rapid downswing comparable to the current situation and it is interesting that the losses did not start until 1 year into the downswing. If you look at the other housing downturn data, you will see the exact same pattern of slow and delayed job losses. This is nothing out of the norm..... only our expectations may be out of the norm.
Fred, thanks for your input - and maybe I'm wrong. But I'd caution in looking at total construction employment, as opposed to residential construction employment.
Historically nonresidential construction follows residential construction with a fairly predictable lag. See this post on investment lags.
So if you look at overall construction employment following a housing downturn, it appears to take longer than I'm suggesting. But I'm focused on residential construction employment - and there was no data before 1985 - so we really don't know what happened in prior periods.
OTOH, it could take longer for these jobs to be lost. We will see soon!
Calmo,
I appreciate the correction, but the point I am trying to make is that the accounting for illegals should be the same on the way up as the way down. It doesn't really matter whether the adjustments included illegals on the way up or not, so long as we are using the same methodology on the way down, does it?
This is rebuttal to the argument that the graph is distorting reality, but only on the way down.
jeez cr
cant ya be a bit more analytical here?
CR-
Another reason is that completions are being accelerated to get land off the homebuilders' books. Homebuilders are having a liquidity crunch. Land is a very illiquid asset. And right now there are no other builders who want to buy lots. The best way to sell land is to build a house on it and sell the house. In the meantime, before the house sells you can get construction financing, generating some cash for current operations.
Is it possible the BLS is simply lying? I mean, it wouldn't be the first time the Bush administration has cooked the numbers for political reasons.
It's getting to the point that I am beginning to doubt ALL government statistics. Remember, a recent executive order allowed Bush to appoint political 'commissars' at all agencies, ala the Soviet Union. Why should we believe that unemployment fell AGAIN when all empirical evidence indicates in should have risen?
Just my 2 cents.
Why should we believe that unemployment fell AGAIN when all empirical evidence indicates in should have risen?
Oh, so they're only cooking some of the books?
CR,
But even if you look at completions, wouldn't you still expect more job losses at this point?
I've got another theory that you sort of touched on. From what I've read, the high-end market is holding up much better than the low-end RE market. So the starts that have dropped off the most are the ones that require the least amount of workers to complete. Similarly, there is a greater than normal percentage of high-end homes in the total recent starts.
What do you think? Plausible? Not plausible?
Is there a way to check on the values of new home starts over the last year?
Mish has a striking analysis of the numbers which comes close to proving what i said above.
Mish's Global etc.
Its in his comment about the 374,000 who dropped out of the labor force and the birth/death adjustment, which is just ludicrous this month.
perma bear,
So what you're saying is you believe parts of the employment data are legit (the 374,000 increase in those out of the labor force), but they're lying about other parts (the construction job losses). Is that it?
perma bear,
So what you're saying is you believe parts of the employment data are legit (the 374,000 increase in those out of the labor force), but they're lying about other parts (the construction job losses). Is that it?
Thanks CR, Great Post!
Steve:
This is the problem that is presented when we have an administration that has lied from day one til today. From the war in Iraq to the FBI illegal activities that were denied by all until today. What to believe? When the stock market and low unemploymnet rate are the only thing this administration has to brag about, it is time to examine the teeth in those two horse's mouths.
Steve
Many of the internal numbers are probably correct. It is the overall interpretation of those numbers that is suspect.
Without having seen the original document, I would guess the BLS news release did not mention anything about the effect of the huge decrease in the labor force on the unemployment rate. This would be a significant omission.
Furthermore, how could the household survey possibly come up with a legitimate increase in employment in the RE sector, when virtually everything about it has been imploding recently?
There's just too much inconsistency for the overall conclusions to be believable. However, you are free to believe them, of course. I will act accordingly ... and have been selling stock into this dead cat bounce.
What to believe?
According to Perma Bear the answer to this question is simple: You believe the parts you like and disregard the parts you don't like as "fixed."
It's one thing to be suspicious of government numbers, but it's a whole different ballgame to be suspicious and supportive at the same time, depending on whether or not such data backs your view.
perma bear,
Let me ask you this. If you were to fudge the household survey so that it would yield a lower unemployment rate, would you do it by fixing the number of jobs that were created or the number of people who left the workforce?
Steve - according to my model, building is still very high. However, it's not because of SFR, but because of a higher volume of starts of large multi-family.
Now, once I reloaded everything and went into look at what it was actually doing, I was stricken with horror. It's calculating it five different ways, tossing up to two fliers, comparing to construction value numbers. Also, it's using some pay-for-play numbers from NAHB. The purpose of that module is actually to project risk forward over the life of a projected C&D loan from initiation through final sale rather than actual construction employment, although over a quarter there is necessarily a correlation.
The only thing I can say is that it has worked in the past. What it does show now is that construction is proceeding at a rapid pace, although I am guessing that a lot of these larger projects will drop out in various areas. One interesting thing it shows is that unskilled labor should be dropping out, whereas skilled labor should be in relatively high demand. In some localities construction is already very slow, but in others it is still very high.
Re Illegal migrant workers:
Just one thing I wanted to point out. While I agree that official numbers are almost certainly undercounting layoffs because the illegals are probably the first to go...
Remember that the HIRING of the illegals also didn't show up in the data during the expansion. In other words, the distortions caused by the illegals should be equal on the way up and the way down.
Thus, the firing of the illegals shouldn't be a reason why layoffs don't seem to be tracking starts that well - yet.
Steve,
I don't think the Bush alters unemployment statistics. The BLS has the same methods (and sometimes management) they've had for decades.
Maybe the Bush administration would like to manipulate the data, but they are at least being watched by the other party that would enjoy seeing the numbers appear worse.
NDD writes:
Remember that the HIRING of the illegals also didn't show up in the data during the expansion.
and he needs to know (and remember of course) that the recent benchmark revisions (~1M added) purport to take that illegal hiring (this B what it is folks: unprosecuted illegal hiring by specified and self-identified companies) into account.
Some accounting.
As I have said before, your expectation of a fast drop in construction jobs does not match what has happened in past housing downswings. In prior housing downturns the loss of 700k to 1 million jobs happened over 25 to 30 months and resulted in about 30k of job losses per month. The job losses only happened late into the housing downturns.
This data can be found at Economagic: Economic Time Series Page by looking at the BLS US NonFarm Employment Construction data. For example in the housing downturn of 1973-1974 that is on your graph, housing starts peaked in Jan of 1973. But construction jobs peaked in December 1973 at 4,239,000. The low point was reached in July of 1975 at 3,501,000 - a lost of 740,000 jobs in 19 months. That comes out to 39K jobs/month.
From your graph this was a rapid downswing comparable to the current situation and it is interesting that the losses did not start until 1 year into the downswing. If you look at the other housing downturn data, you will see the exact same pattern of slow and delayed job losses. This is nothing out of the norm..... only our expectations may be out of the norm.
Fred, thanks for your input - and maybe I'm wrong. But I'd caution in looking at total construction employment, as opposed to residential construction employment.
Historically nonresidential construction follows residential construction with a fairly predictable lag. See this post on investment lags.
So if you look at overall construction employment following a housing downturn, it appears to take longer than I'm suggesting. But I'm focused on residential construction employment - and there was no data before 1985 - so we really don't know what happened in prior periods.
OTOH, it could take longer for these jobs to be lost. We will see soon!
Best Wishes.
Calmo,
I appreciate the correction, but the point I am trying to make is that the accounting for illegals should be the same on the way up as the way down. It doesn't really matter whether the adjustments included illegals on the way up or not, so long as we are using the same methodology on the way down, does it?
This is rebuttal to the argument that the graph is distorting reality, but only on the way down.