Unemployment Declined Even As Hurricanes Decreased Total Jobs

[JimBowie1958]

This report is so bizare and illustrates why the economic metrics are outdated and increasingly useless.

Jobs fall by 33,000 in September as hurricanes hammer employment

Hurricanes Harvey and Irma damaged not only Texas and Florida but also the U.S. jobs picture, as payrolls fell by 33,000 in September. That drop came even as the unemployment rate fell to a 16-year low of 4.2 percent, the Bureau of Labor Statistics reported Friday.

The jobs loss was the first monthly decline in seven years, when the economy was still pulling out of the Great Recession.

Even with the surprise jobs number, the closely watched hourly wages figure jumped higher, to an annualized rate of 2.9 percent.


Economists surveyed by Reuters expected payroll growth of 90,000 in September, compared with 169,000 in August. The unemployment rate was expected to hold steady at 4.4 percent. It declined even as the labor-force participation rate rose to 63.1 percent, its highest level all year and the best reading since March 2014....

Economists surveyed by Reuters expected payroll growth of 90,000 in September, compared with 169,000 in August. The unemployment rate was expected to hold steady at 4.4 percent. It declined even as the labor-force participation rate rose to 63.1 percent, its highest level all year and the best reading since March 2014....

In addition to September's rough month, the July number was revised lower from 189,000 to 138,000 though August got a bump higher from 156,000. In all, though, 2017 thus far has seen the slowest jobs growth in at least five years.

Revisions will bear watching in coming months, as the final payrolls number comes from the Labor Department's byzantine estimation methods. The department's household survey showed the actual level of employed Americans grew by 906,000 while the unemployment rolls fell by 331,000. The report indicated a record 154.3 million Americans at work.

Job gains for the month came from health care, at 23,000, transportation and warehousing with 22,000 and professional and business services, which added 13,000.

Harvey and Irma's effects weren't all negative when it came to jobs — Wall Street-related finance positions grew by a net 10,000 thanks to an increase of 11,000 for insurance carriers and related activities. Credit-related services lost 7,000 positions.​

 

[pinqy]

This report is so bizare and illustrates why the economic metrics are outdated and increasingly useless.

Jobs fall by 33,000 in September as hurricanes hammer employment

Hurricanes Harvey and Irma damaged not only Texas and Florida but also the U.S. jobs picture, as payrolls fell by 33,000 in September. That drop came even as the unemployment rate fell to a 16-year low of 4.2 percent, the Bureau of Labor Statistics reported Friday.

The jobs loss was the first monthly decline in seven years, when the economy was still pulling out of the Great Recession.

Even with the surprise jobs number, the closely watched hourly wages figure jumped higher, to an annualized rate of 2.9 percent.


Economists surveyed by Reuters expected payroll growth of 90,000 in September, compared with 169,000 in August. The unemployment rate was expected to hold steady at 4.4 percent. It declined even as the labor-force participation rate rose to 63.1 percent, its highest level all year and the best reading since March 2014....

Economists surveyed by Reuters expected payroll growth of 90,000 in September, compared with 169,000 in August. The unemployment rate was expected to hold steady at 4.4 percent. It declined even as the labor-force participation rate rose to 63.1 percent, its highest level all year and the best reading since March 2014....

In addition to September's rough month, the July number was revised lower from 189,000 to 138,000 though August got a bump higher from 156,000. In all, though, 2017 thus far has seen the slowest jobs growth in at least five years.

Revisions will bear watching in coming months, as the final payrolls number comes from the Labor Department's byzantine estimation methods. The department's household survey showed the actual level of employed Americans grew by 906,000 while the unemployment rolls fell by 331,000. The report indicated a record 154.3 million Americans at work.

Job gains for the month came from health care, at 23,000, transportation and warehousing with 22,000 and professional and business services, which added 13,000.

Harvey and Irma's effects weren't all negative when it came to jobs — Wall Street-related finance positions grew by a net 10,000 thanks to an increase of 11,000 for insurance carriers and related activities. Credit-related services lost 7,000 positions.​

There are two surveys in the monthly Employment Situation: The Current Employment Statistics Survey (CES), and the Current Population Survey (CPS). The CES is a survey of nonfarm business and it basically just asks how many people were on the payroll for the pay period that includes the 12th of the month. Because of the hurricanes, fewer people were on the payrolls for that month.
The CPS is a household survey with a whole bunch of question on employment and unemployment. It asks if a person worked during the week of the month that includes the 12 or if they had a job they were temporarily absent from due to vacation, illness/injury, weather, or strike.

The two surveys have different concepts and measurement of employment and they serve different purposes. The CES is a measure of nonfarm payroll jobs, and excludes agriculture, the self employed, unpaid family workers, and people who work in other people's houses (personal maids, cooks, gardeners, etc). And if someone has two jobs, they can be double counted. The advantage is that the sample is huge and the results are benchmarked to a census of company tax records. The CES is narrow, but accurate.

The CPS has a much smaller sample, but includes everyone age 16 and older who are not in the military, prison, or other institution. It's a count of people, not jobs, so no double counting, and can capture unemployment and job search data, and demographic data. So it's broader, but less accurate.

 

[JimBowie1958]

There are two surveys in the monthly Employment Situation: The Current Employment Statistics Survey (CES), and the Current Population Survey (CPS). The CES is a survey of nonfarm business and it basically just asks how many people were on the payroll for the pay period that includes the 12th of the month. Because of the hurricanes, fewer people were on the payrolls for that month.
The CPS is a household survey with a whole bunch of question on employment and unemployment. It asks if a person worked during the week of the month that includes the 12 or if they had a job they were temporarily absent from due to vacation, illness/injury, weather, or strike.

The two surveys have different concepts and measurement of employment and they serve different purposes. The CES is a measure of nonfarm payroll jobs, and excludes agriculture, the self employed, unpaid family workers, and people who work in other people's houses (personal maids, cooks, gardeners, etc). And if someone has two jobs, they can be double counted. The advantage is that the sample is huge and the results are benchmarked to a census of company tax records. The CES is narrow, but accurate.

The CPS has a much smaller sample, but includes everyone age 16 and older who are not in the military, prison, or other institution. It's a count of people, not jobs, so no double counting, and can capture unemployment and job search data, and demographic data. So it's broader, but less accurate.

Yes, true, but these methodologies are obsolete today.

IF the Federalis had their act together, they should be able to take their stats directly counted from SSN contributions from employers instead of this survey/guessing bullshit that occasionally gets contradictory results.

 

[pinqy]

There are two surveys in the monthly Employment Situation: The Current Employment Statistics Survey (CES), and the Current Population Survey (CPS). The CES is a survey of nonfarm business and it basically just asks how many people were on the payroll for the pay period that includes the 12th of the month. Because of the hurricanes, fewer people were on the payrolls for that month.
The CPS is a household survey with a whole bunch of question on employment and unemployment. It asks if a person worked during the week of the month that includes the 12 or if they had a job they were temporarily absent from due to vacation, illness/injury, weather, or strike.

The two surveys have different concepts and measurement of employment and they serve different purposes. The CES is a measure of nonfarm payroll jobs, and excludes agriculture, the self employed, unpaid family workers, and people who work in other people's houses (personal maids, cooks, gardeners, etc). And if someone has two jobs, they can be double counted. The advantage is that the sample is huge and the results are benchmarked to a census of company tax records. The CES is narrow, but accurate.

The CPS has a much smaller sample, but includes everyone age 16 and older who are not in the military, prison, or other institution. It's a count of people, not jobs, so no double counting, and can capture unemployment and job search data, and demographic data. So it's broader, but less accurate.

Yes, true, but these methodologies are obsolete today.

IF the Federalis had their act together, they should be able to take their stats directly counted from SSN contributions from employers instead of this survey/guessing bullshit that occasionally gets contradictory results.

The reference period for the CES for September was the pay period that included the week of September 10-16. So collection of the data started on 18 September. The Employment Situation Report was completed on 4 October. Do you honestly think BLS could collect and process employment, hours (regular and overtime), and wages by industry and geography,from SSN records in 2 weeks? Not remotely possible.

BLS does conduct a Quarterly Census of Employment and Wages based on Unemployment Insurance tax records filed by private businesses. It takes about 6 months to process. The CES is benchmarked to the QCEW once a year, and the discrepancy is usually less than 1% difference.

And there are no records for job search or unemployment so a household survey has to be used.

 

[JimBowie1958]

The reference period for the CES for September was the pay period that included the week of September 10-16. So collection of the data started on 18 September. The Employment Situation Report was completed on 4 October. Do you honestly think BLS could collect and process employment, hours (regular and overtime), and wages by industry and geography,from SSN records in 2 weeks? Not remotely possible.

BLS does conduct a Quarterly Census of Employment and Wages based on Unemployment Insurance tax records filed by private businesses. It takes about 6 months to process. The CES is benchmarked to the QCEW once a year, and the discrepancy is usually less than 1% difference.

And there are no records for job search or unemployment so a household survey has to be used.

As SSN payments come in from employers the data could be accumulated and tallied in reference tables for computation.

Two weeks is plenty of time as I have done similar things myself over hundreds of millions of records.

You just have to have people who know what the fuck they are doing.

Primary key the SS number as a number field and index it on a table that has the bare minimum data for the report and have it updated every time a payment is made and poof it can be queried up in a couple of hours in a good system, faster in top notch systems.

The problem is that government officials have some power by being able to spin the reports the way they prefer to and they DONT WANT accurate timely information to be available to the public as that would cost them this power.

 

[pinqy]

The reference period for the CES for September was the pay period that included the week of September 10-16. So collection of the data started on 18 September. The Employment Situation Report was completed on 4 October. Do you honestly think BLS could collect and process employment, hours (regular and overtime), and wages by industry and geography,from SSN records in 2 weeks? Not remotely possible.

BLS does conduct a Quarterly Census of Employment and Wages based on Unemployment Insurance tax records filed by private businesses. It takes about 6 months to process. The CES is benchmarked to the QCEW once a year, and the discrepancy is usually less than 1% difference.

And there are no records for job search or unemployment so a household survey has to be used.

As SSN payments come in from employers the data could be accumulated and tallied in reference tables for computation.

Two weeks is plenty of time as I have done similar things myself over hundreds of millions of records.

You just have to have people who know what the fuck they are doing.

Primary key the SS number as a number field and index it on a table that has the bare minimum data for the report and have it updated every time a payment is made and poof it can be queried up in a couple of hours in a good system, faster in top notch systems.

The problem is that government officials have some power by being able to spin the reports the way they prefer to and they DONT WANT accurate timely information to be available to the public as that would cost them this power.

But that wouldn't give hours worked, or total wages, or industry. Here are the current tables published for payroll employment:
Table B-1. Employees on nonfarm payrolls by industry sector and selected industry detail
Table B-2. Average weekly hours and overtime of all employees on private nonfarm payrolls by industry sector, seasonally adjusted

Table B-3. Average hourly and weekly earnings of all employees on private nonfarm payrolls by industry sector, seasonally adjusted

Table B-4. Indexes of aggregate weekly hours and payrolls for all employees on private nonfarm payrolls by industry sector, seasonally adjusted

Table B-5. Employment of women on nonfarm payrolls by industry sector, seasonally adjusted

Table B-6. Employment of production and nonsupervisory employees on private nonfarm payrolls by industry sector, seasonally adjusted(1)

Table B-7. Average weekly hours and overtime of production and nonsupervisory employees on private nonfarm payrolls by industry sector, seasonally adjusted(1)

Table B-8. Average hourly and weekly earnings of production and nonsupervisory employees on private nonfarm payrolls by industry sector, seasonally adjusted(1)

Table B-9. Indexes of aggregate weekly hours and payrolls for production and nonsupervisory employees on private nonfarm payrolls by industry sector, seasonally adjusted(1)

You can't categorize that many records by hours, payroll, sex, production vs supervisory, and seasonally adjust and quality check in two weeks.

• 
  

 

[JimBowie1958]

But that wouldn't give hours worked, or total wages, or industry. Here are the current tables published for payroll employment:
Table B-1. Employees on nonfarm payrolls by industry sector and selected industry detail
Table B-2. Average weekly hours and overtime of all employees on private nonfarm payrolls by industry sector, seasonally adjusted

Table B-3. Average hourly and weekly earnings of all employees on private nonfarm payrolls by industry sector, seasonally adjusted

Table B-4. Indexes of aggregate weekly hours and payrolls for all employees on private nonfarm payrolls by industry sector, seasonally adjusted

Table B-5. Employment of women on nonfarm payrolls by industry sector, seasonally adjusted

Table B-6. Employment of production and nonsupervisory employees on private nonfarm payrolls by industry sector, seasonally adjusted(1)

Table B-7. Average weekly hours and overtime of production and nonsupervisory employees on private nonfarm payrolls by industry sector, seasonally adjusted(1)

Table B-8. Average hourly and weekly earnings of production and nonsupervisory employees on private nonfarm payrolls by industry sector, seasonally adjusted(1)

Table B-9. Indexes of aggregate weekly hours and payrolls for production and nonsupervisory employees on private nonfarm payrolls by industry sector, seasonally adjusted(1)

You can't categorize that many records by hours, payroll, sex, production vs supervisory, and seasonally adjust and quality check in two weeks.

Yes, you can by maintaining a summary view of the tables for the report needed.

I have done similar things so I know what I am talking about.

The quality check can be done on the data entered, so that it is consistent with itself, but QA with Reality would require human inquiry and the report amended in revisions later, which they do anyway.