About that slight pickup in residential construction employment: I wonder if the builders aren't working overtime to get projects completed (and sales settled) before the pricing situation gets worse.
I am also curious whether anyone has numbers on the percent of residential construction employment that involves illegal immigrants. Perhaps we're seeing worried contractors moving to "on the books" employees?
If U.S. yearly population growth--according to census reports--is approximately 3 million, the average monthly addition to the workforce would be 125,000 only if only half of that new workforce is employed.
Yet we know that both women and men are entering the work force, of necessity.
Consequently, just to keep unemployment even, the workforce should grow somewhere around or between 175,000 to 250,000.
Could you clarify your statement that "Job growth is about in the middle of the expected range" in terms of how the expected range was determined? Do you mean that this range is what we would expect to see given the average US job growth over the past 30-50 years?
If that is so, then job growth during Bush's second term is a non-issue for either party.
The construction employment number is likely reflects builders rushing to try to finish projects that have been in process before the bottom completely drops out of the housing market.
I'm not surprised - companies layoff VERY slowly until they are certain they are screwed then drop people like soiled undies.
In the case of the large builders, many of the actual 'employees' are office overhead & supervison and the actual construction site workers are contract workers - ie 1099 subcontract employees or contractors who then hire 1099s or employees.
So as work slows they might be cutting back hours but not necessarily letting people go completely... not officially... not yet.
So maybe there is a dip in the build but it doesn't drive true unemployment per se but rather more 'underemployment'... which goes unreported.
That is until things get really ugly and the direct employees are laid off by the builders AND the 1099s throw in the towel & close up shop.
I am certain this is why UE rates are pretty meaningless right now - so many 1099s out there - yours truly included, though not in RE. It far more common everywhere.
AS for the Mexican workforce per tanta's entry, I would expect them to be near the last people let go - screw the 'books', no one checks them and its easy to fudge anyway. My guess is due to 'fudging' they are often counted as legit - except for the onesy twosy day laborers hired by the home owners for cash. But if they are on REAL job sites they are counted - fake documents and all. Counted in the paper work as 'somebody' anyway... if not as 'Juan Coneja' then as 'John Doe'.
I hate to be a party pooper but I don't think we'll get a clear picture of where the housing market employment is until next spring at the soonest. Those 1099s are NOT going to throw in the towel OFFICIALLY until (1) after next spring to see if the market rebounds and (2) until they have someplace else to go - which is where? Until then they will be self described self-employed or partially employed employees of contractors and won't count among the UE. Not yet anyway.
So I don't look for big changes in the official BLS stats immediately. This too will also take a long time to play out.
This whole RE thing trucks along like a speeding glacier.
Tanta, I am a little confused as to why residential construction employment has held up so well (only off 1% since Feb). Perhaps builders are rushing projects to completion - that would make sense. It's just a matter of when ...
Stormy, the U.S. adds about 3 Million people per year, but only half are working age.
Scott Peterson, I came up with that range early last year - I'll dig up the rational again. The high end (10 million over 4 years) means that employment participation would have to increase and the unemployment rate drop further. The low end is basically treading water with the growth in working age population. I'll post my calculations on how I determined that range - maybe for next month's numbers.
I don't think job growth is an issue in this election (it might be in '08 ). But that is only part of the employment picture - wages and benefits is also important.
Economic models often have a difficult time picking up trend changes. This report is probably an anomoly.
Anecdotally, I am in the development industry in SoCal and have yet to hear of any construction layoffs. I have heard of many builders cancelling future new projects. This implies that once the projects that are currently in construction finish the construction workers will have far fewer new projects to work on.
Figuring on average a construction project lasts 12-18 months. And, in a healthy market as we have had the past few years projects are continuously starting and finishing. Therefore, the median project is 6-9 months from completion. Therefore, I anticipate major construction layoffs will lag builder's decisions to cancel new deals by about 6-9 months. FYI, I heard builders cancelling new projects starting in May, 2006. Therefore I don't anticipate substatial layoffs until the end of the year.
CR, thanks for the info. There is clearly a lag from when the housing market slows to when the layoffs start. Maybe it will be a few more months before the bust shows up in employment.
BTW, I posted some recent layoff announcements from the Inland Empire - so I think we are getting close.
A couple of people have mentioned modeling as the problem in the jobs report. There is a model involved, but this is, at its heart, a counting exercise. The model part, the birth/death plug, has been rejiggered to improve its responsiveness to changes in the economy. Hiring overall has been very steady at its new, lower pace. Shouldn't we at least suspect that's a sign that the data are pretty clean?
I realize that our host has devoted lots of attention to the residential construction sector, which helps explain the focus today on residential construction jobs, but there were other oddities. Education and health care jobs accounted for nearly half the hiring in the August report, adding 60k jobs. The 6-month average was just 35k. My first thought was that this was a seasonal problem due to teachers showing up early in the school jobs tally, but that doesn't seem to be it. Hospital hiring and day care seem to account for the big pop. If I were going to doubt something, I'd doubt the 25k or so "extra" jobs from health and day care.
In the construction sector, there is, after all, a mountain of unfinished homes, as Tanta points out. Starts have slowed a lot. That means unfinished inventory will pretty soon fall quickly, as well. Construction hiring has certainly responded to the reduction in demand, turning flat. Be patient. The layoffs are probably coming once inventories of unfinished houses look more like starts.
Over on BigPicture Mr. Ritholz has quite a rant on employment and the appropriate figure of merit on net monthly jobs debate. The Big Picture
IF you'll pardon my quoting myself perhaps the numbers from my comment are useful in this discussion:
Just in case anybody's interested in backing up the rant with some #'s the first time new jobs was seriously less than appealing was in Q300 when only 84.3K new jobs were created on a monthly average. If we want to assume that the figure of merit (the old benchmark) is 150K jobs/month required for steady-state breakeven then we had a net deficit of about 66K jobs/month or 197K for the quarter.
If you run that logic forward, which is in effect saying two really important things - which have been and are implicit in many of Barry's arguments: a healthy economy creates at least 150K jobs/month and second, a growing economy creates much more than that - around 250K/month. Furthermore a recovering economy will go thru a period of 18months or more when it's creating 300-350K jobs/month.
Hopefully that's not too painful too follow-thru on the screens but following the logic the cumulative jobs deficit reached a nadir of 7.97 MILLION jobs in Q104. It is still about 7.47 million at the end of Q2 this year. In other words we've been loosing ground and never made up for lost jobs. Bear in mind that's out of an employed population of 135-136 million.
There in a nutshell is the driving engine and explanation for why people don't feel good about the economy (or Administration), why GDP is slowing and why business spending won't pick up the slack as the Housing ATM and construction fall off. Organic, self-sustaining economic growth happens when businesses invest in future demand and accelerate the economy thru employment growth in a virtuous cycle.
We've disguised the lack of organic growth with a high fever via Housing and low rates. You need to watch employment very carefully and anytime it's below 150K then trouble is building up.
Even ADP wasn't that high, considering how political things have become, their "model" will be even more under attack. I expect a downward revisement closer to ADP.
Of course the main layoffs from real estate are yet to come, CR was a little gun hoe, wasn't he?
No point drawing too many conclusions from a month's data when it has yet to be revised. Next month, this may look totally different.
That said...Ive been looking at this issue of moving the goalposts, how now 125k jobs is good enough to keep up with the labor force growth. Well, first off, there is some amount of adjustment, but I dont think it is that substantial. Working age population growth has slowed from about 1.4% a year in the late 90s to about 1.2% a year last year, on a steady slow decline since 1999. On that basis alone, if the old rule of thumb of about 175k needed a month to keep up with labor force growth, is reduced by about 10% due to the slower working age population growth, then you might be at 160k. However, the incidence of labor force participation for 65+ is rising steadily, so, you can pare back some of that reduction. Also, the lfprs for women were rising steadily til the recession, but since 2001 have declined slightly to about 59.2% from 60% and males about 75% to 73.5%. That might shave off about another 5%.
so, all in all, to make the claim that now we are 120k (KC fed) to 100k, are both way on the low side. Convenient for spinnng the numbers, but basically screwed up. Id be surprised if it was any lower than 140k.
Geoff - thanks for working thru that analysis. Helpful. Based on that think I'm gonna stick with my 150K/month figure of merit as a reasonable benchmark. Reactions ?
CR - Based on this exchange and your explanation of your envelope analysis can see where you're coming from but have you considered some comparative analysis ? That is you're running forward from a given starting point without looking backward at net job creation and GDP/business-cycle growth. If Geoff's right and I'm not smokin too much funky stuff in my pipe then just running forward doesn't get us ahead .
BTW, if you chart out since the late 80s, residential construction employment, vs nipa residential building, you'll see that the lag is about 2-3 quarters in terms of when residential construction turns, and when the job cuts come. That said, we should be looking at major reductions related employment probably in the October report on employment. How big? Well, I just did a really basic model of residential construction employment vs resid building...with a 2 qtr lag. If you put in a pretty reasonable qtrly construction forecast, with the qtrly downdraft at between 10-20% at an annual rate the next four quarters, which, for perspective is against -9.8% which we had in q206, I show employment falling by 07q2, by just shy of one million jobs, which is about a -85k drag a month on job growth the next year. That's scary stuff. These quarterly declines in resid bldg could easily be worse, in that this downturn could well be more severe than the last two, at least based on building permit trends so far. I'll keep updating this each quarter.
One more thing to add here. If you look at this relationship over time vs more recently, youll see that the runup of the last few years in resid construction did not coincide with as big of a runup in construction employment as would be expected. Im thinking this is largely the influence of ballooning prices of homes, which inflate the value of what is built as well. That said, the extent to which prices fall and that influences the reported NIPA res bldg number, it will mitigate the fall in employment. So, instead of 1 million of these jobs disappearing in the next year, it is probably more like half to three quarters that amount, Im guessing. Still, a pretty sizeable hit to employment. Add to that the related RE jobs in mortgage, home related retail, and the incomes from commissions, etc. It's really hard to see how we dont end up with about zero net job gains in a few months, and once that happens, I cant imagine how the rest of the economy is going to perceive that as a sign that more investment is needed - either in non res bldg, or machinery and equipment.
Someone please tell me how this ends in anything other than recession. I just cant see it anymore.
About that slight pickup in residential construction employment: I wonder if the builders aren't working overtime to get projects completed (and sales settled) before the pricing situation gets worse.
I am also curious whether anyone has numbers on the percent of residential construction employment that involves illegal immigrants. Perhaps we're seeing worried contractors moving to "on the books" employees?
I am puzzled.
If U.S. yearly population growth--according to census reports--is approximately 3 million, the average monthly addition to the workforce would be 125,000 only if only half of that new workforce is employed.
Yet we know that both women and men are entering the work force, of necessity.
Consequently, just to keep unemployment even, the workforce should grow somewhere around or between 175,000 to 250,000.
"Construction industries added 17,000, including 4,000 in residential construction."
The biggest suprise? Oh come on. This is very hard to believe. The BLS loses big time credibility on this repotr.
Could you clarify your statement that "Job growth is about in the middle of the expected range" in terms of how the expected range was determined? Do you mean that this range is what we would expect to see given the average US job growth over the past 30-50 years?
If that is so, then job growth during Bush's second term is a non-issue for either party.
The construction employment number is likely reflects builders rushing to try to finish projects that have been in process before the bottom completely drops out of the housing market.
Not buying the "likely reflects" bit at all. Construction sector slowed heavily. Sometimes models lie.
I'm not surprised - companies layoff VERY slowly until they are certain they are screwed then drop people like soiled undies.
In the case of the large builders, many of the actual 'employees' are office overhead & supervison and the actual construction site workers are contract workers - ie 1099 subcontract employees or contractors who then hire 1099s or employees.
So as work slows they might be cutting back hours but not necessarily letting people go completely... not officially... not yet.
So maybe there is a dip in the build but it doesn't drive true unemployment per se but rather more 'underemployment'... which goes unreported.
That is until things get really ugly and the direct employees are laid off by the builders AND the 1099s throw in the towel & close up shop.
I am certain this is why UE rates are pretty meaningless right now - so many 1099s out there - yours truly included, though not in RE. It far more common everywhere.
AS for the Mexican workforce per tanta's entry, I would expect them to be near the last people let go - screw the 'books', no one checks them and its easy to fudge anyway. My guess is due to 'fudging' they are often counted as legit - except for the onesy twosy day laborers hired by the home owners for cash. But if they are on REAL job sites they are counted - fake documents and all. Counted in the paper work as 'somebody' anyway... if not as 'Juan Coneja' then as 'John Doe'.
I hate to be a party pooper but I don't think we'll get a clear picture of where the housing market employment is until next spring at the soonest. Those 1099s are NOT going to throw in the towel OFFICIALLY until (1) after next spring to see if the market rebounds and (2) until they have someplace else to go - which is where? Until then they will be self described self-employed or partially employed employees of contractors and won't count among the UE. Not yet anyway.
So I don't look for big changes in the official BLS stats immediately. This too will also take a long time to play out.
This whole RE thing trucks along like a speeding glacier.
If that is so, then job growth during Bush's second term is a non-issue for either party.
Yup. And if you listen carefully, in a mid-term election year, the silence has been deafening.
Gas prices, Iraq, budgets - all on the 'A' List. Employment/unemployment... pretty quiet. That's on the agenda for 2008.
Tanta, I am a little confused as to why residential construction employment has held up so well (only off 1% since Feb). Perhaps builders are rushing projects to completion - that would make sense. It's just a matter of when ...
Stormy, the U.S. adds about 3 Million people per year, but only half are working age.
Scott Peterson, I came up with that range early last year - I'll dig up the rational again. The high end (10 million over 4 years) means that employment participation would have to increase and the unemployment rate drop further. The low end is basically treading water with the growth in working age population. I'll post my calculations on how I determined that range - maybe for next month's numbers.
I don't think job growth is an issue in this election (it might be in '08 ). But that is only part of the employment picture - wages and benefits is also important.
Best Wishes.
Economic models often have a difficult time picking up trend changes. This report is probably an anomoly.
Anecdotally, I am in the development industry in SoCal and have yet to hear of any construction layoffs. I have heard of many builders cancelling future new projects. This implies that once the projects that are currently in construction finish the construction workers will have far fewer new projects to work on.
Figuring on average a construction project lasts 12-18 months. And, in a healthy market as we have had the past few years projects are continuously starting and finishing. Therefore, the median project is 6-9 months from completion. Therefore, I anticipate major construction layoffs will lag builder's decisions to cancel new deals by about 6-9 months. FYI, I heard builders cancelling new projects starting in May, 2006. Therefore I don't anticipate substatial layoffs until the end of the year.
CR, thanks for the info. There is clearly a lag from when the housing market slows to when the layoffs start. Maybe it will be a few more months before the bust shows up in employment.
BTW, I posted some recent layoff announcements from the Inland Empire - so I think we are getting close.
Best Wishes.
A couple of people have mentioned modeling as the problem in the jobs report. There is a model involved, but this is, at its heart, a counting exercise. The model part, the birth/death plug, has been rejiggered to improve its responsiveness to changes in the economy. Hiring overall has been very steady at its new, lower pace. Shouldn't we at least suspect that's a sign that the data are pretty clean?
I realize that our host has devoted lots of attention to the residential construction sector, which helps explain the focus today on residential construction jobs, but there were other oddities. Education and health care jobs accounted for nearly half the hiring in the August report, adding 60k jobs. The 6-month average was just 35k. My first thought was that this was a seasonal problem due to teachers showing up early in the school jobs tally, but that doesn't seem to be it. Hospital hiring and day care seem to account for the big pop. If I were going to doubt something, I'd doubt the 25k or so "extra" jobs from health and day care.
In the construction sector, there is, after all, a mountain of unfinished homes, as Tanta points out. Starts have slowed a lot. That means unfinished inventory will pretty soon fall quickly, as well. Construction hiring has certainly responded to the reduction in demand, turning flat. Be patient. The layoffs are probably coming once inventories of unfinished houses look more like starts.
Over on BigPicture Mr. Ritholz has quite a rant on employment and the appropriate figure of merit on net monthly jobs debate. The Big Picture
IF you'll pardon my quoting myself perhaps the numbers from my comment are useful in this discussion:
Just in case anybody's interested in backing up the rant with some #'s the first time new jobs was seriously less than appealing was in Q300 when only 84.3K new jobs were created on a monthly average. If we want to assume that the figure of merit (the old benchmark) is 150K jobs/month required for steady-state breakeven then we had a net deficit of about 66K jobs/month or 197K for the quarter.
If you run that logic forward, which is in effect saying two really important things - which have been and are implicit in many of Barry's arguments: a healthy economy creates at least 150K jobs/month and second, a growing economy creates much more than that - around 250K/month. Furthermore a recovering economy will go thru a period of 18months or more when it's creating 300-350K jobs/month.
Hopefully that's not too painful too follow-thru on the screens but following the logic the cumulative jobs deficit reached a nadir of 7.97 MILLION jobs in Q104. It is still about 7.47 million at the end of Q2 this year. In other words we've been loosing ground and never made up for lost jobs. Bear in mind that's out of an employed population of 135-136 million.
There in a nutshell is the driving engine and explanation for why people don't feel good about the economy (or Administration), why GDP is slowing and why business spending won't pick up the slack as the Housing ATM and construction fall off. Organic, self-sustaining economic growth happens when businesses invest in future demand and accelerate the economy thru employment growth in a virtuous cycle.
We've disguised the lack of organic growth with a high fever via Housing and low rates. You need to watch employment very carefully and anytime it's below 150K then trouble is building up.
Even ADP wasn't that high, considering how political things have become, their "model" will be even more under attack. I expect a downward revisement closer to ADP.
Of course the main layoffs from real estate are yet to come, CR was a little gun hoe, wasn't he?
No point drawing too many conclusions from a month's data when it has yet to be revised. Next month, this may look totally different.
That said...Ive been looking at this issue of moving the goalposts, how now 125k jobs is good enough to keep up with the labor force growth. Well, first off, there is some amount of adjustment, but I dont think it is that substantial. Working age population growth has slowed from about 1.4% a year in the late 90s to about 1.2% a year last year, on a steady slow decline since 1999. On that basis alone, if the old rule of thumb of about 175k needed a month to keep up with labor force growth, is reduced by about 10% due to the slower working age population growth, then you might be at 160k. However, the incidence of labor force participation for 65+ is rising steadily, so, you can pare back some of that reduction. Also, the lfprs for women were rising steadily til the recession, but since 2001 have declined slightly to about 59.2% from 60% and males about 75% to 73.5%. That might shave off about another 5%.
so, all in all, to make the claim that now we are 120k (KC fed) to 100k, are both way on the low side. Convenient for spinnng the numbers, but basically screwed up. Id be surprised if it was any lower than 140k.
Geoff - thanks for working thru that analysis. Helpful. Based on that think I'm gonna stick with my 150K/month figure of merit as a reasonable benchmark. Reactions ?
CR - Based on this exchange and your explanation of your envelope analysis can see where you're coming from but have you considered some comparative analysis ? That is you're running forward from a given starting point without looking backward at net job creation and GDP/business-cycle growth. If Geoff's right and I'm not smokin too much funky stuff in my pipe then just running forward doesn't get us ahead
.
BTW, if you chart out since the late 80s, residential construction employment, vs nipa residential building, you'll see that the lag is about 2-3 quarters in terms of when residential construction turns, and when the job cuts come. That said, we should be looking at major reductions related employment probably in the October report on employment. How big? Well, I just did a really basic model of residential construction employment vs resid building...with a 2 qtr lag. If you put in a pretty reasonable qtrly construction forecast, with the qtrly downdraft at between 10-20% at an annual rate the next four quarters, which, for perspective is against -9.8% which we had in q206, I show employment falling by 07q2, by just shy of one million jobs, which is about a -85k drag a month on job growth the next year. That's scary stuff. These quarterly declines in resid bldg could easily be worse, in that this downturn could well be more severe than the last two, at least based on building permit trends so far. I'll keep updating this each quarter.
One more thing to add here. If you look at this relationship over time vs more recently, youll see that the runup of the last few years in resid construction did not coincide with as big of a runup in construction employment as would be expected. Im thinking this is largely the influence of ballooning prices of homes, which inflate the value of what is built as well. That said, the extent to which prices fall and that influences the reported NIPA res bldg number, it will mitigate the fall in employment. So, instead of 1 million of these jobs disappearing in the next year, it is probably more like half to three quarters that amount, Im guessing. Still, a pretty sizeable hit to employment. Add to that the related RE jobs in mortgage, home related retail, and the incomes from commissions, etc. It's really hard to see how we dont end up with about zero net job gains in a few months, and once that happens, I cant imagine how the rest of the economy is going to perceive that as a sign that more investment is needed - either in non res bldg, or machinery and equipment.
Someone please tell me how this ends in anything other than recession. I just cant see it anymore.