January 03, 2004

Notes: Elise Gould on the Household and Payroll Surveys

The EPI's Elise Gould writes about the differences between the household and the payroll surveys of employment:

Measuring employment since the recovery: ...Some have speculated that the household survey provides a better indication of the trend in employment at and around turning points in the business cycle. These critics question whether the payroll survey accurately and fully picks up new businesses, known as "firm births." This problem may be especially exacerbated when measuring employment in a recovery. In its estimates of employment, the BLS addresses the problem of firm births and deaths using past history and various estimation techniques to provide an adjustment factor to the current series. In addition, updates to the payroll survey are conducted annually to adjust for any discrepancies....

However, the BLS announced in October that its analysis of detailed tax records through March 2003 would result in a downward revision of total nonfarm payroll employment by approximately 145,000 for the March 2003 reference month (BLS 2003b).

A second critique of the payroll survey is that it leaves out self-employment. However, because the household survey employment reports do not distinguish between the self-employed who are gainfully employed and those who are searching for work--and because the numbers of self-employed nonearners would be expected to increase during tough economic times--the omission of self-employment numbers from the payroll survey may more accurately reflect overall employment trends.

The BLS periodically revises the household survey to account for new Census Bureau population estimates. In the last four years, there have been two population adjustments: one in January 2000 and one in January 2003. The shift in January 2000 was based on the new population estimates from the decennial Census and added about 1.5 million persons employed. The shift in January 2003, based on new estimates of faster than expected population growth since the 2000 Census, added another 576,000. At each shift, a discontinuity occurs in the series, reflective of only new population estimates and not an actual jump in employment. To make valid comparisons with the numbers since January 2003, previous employment numbers must be adjusted upward to account for new population estimates. The BLS warns that use of the household survey employment numbers without making these adjustments makes any estimates of trends since January 2003 not comparable with those for earlier months (Bowler et al. 2003)...

Nevertheless, enormous differences between the two surveys remain. Even after adjusting the household survey by "subtract[ing] agriculture, self-employment, private households, unpaid family workers, and those on unpaid leave, and add[ing] multiple job holders... seasonally adjusted," the payroll survey shows a 2.4 million fall in the number of jobs since March 2001, while the [fully adjusted] household survey shows a fall of only 0.2 million. This gap is not due to statistical sampling variability: workers are giving different answers to those conducting the household survey than employers are giving to those conducting the payroll survey: a lot of workers think they work (or say they work) for employers, but the employers don't think the workers work for them.

A look back at the longer time series tells us that (i) the ratio of payroll to household employment moves around a lot (and not just because of discontinuous changes in the estimated population used for the household survey, as happens--for example--at the very start of 2000), and (ii) there is no strong and obvious reason for this ratio to move around so much.

A better sociologist than I could probably write an interesting article on just how it can be that people claim to work for employers who do not claim to employ them, and how the changing numbers of such people reflect both the material base and the cultural superstructure of society.

Posted by DeLong at January 3, 2004 08:49 AM | TrackBack


>> how it can be that people claim to work for >> employers who do not claim to employ them

Consider for instance the case of George W. Bush, who claims to be working for the American people.

Posted by: P O'Neill on January 3, 2004 09:27 AM


perhaps the employers are trying to be consistent with other data thy are giving the government.

Posted by: big al on January 3, 2004 09:45 AM


Prof. DeLong asks, "A better sociologist than I could probably write an interesting article on just how it can be that people claim to work for employers who do not claim to employ them..."

A person who owns and operates a one-person or family business can claim not to be an employee on the 941 form and yet state that he is employed by that company for the purpose of health insurance, credit, etc. I doubt this is what is happening, but since so many professionals have been laid off in this recession, it's not impossible.

Posted by: Charles on January 3, 2004 11:00 AM


No, it's not impossible. And some are, I'm sure. But the swings in the numbers of employees-without-employers are, I think, the most interesting...

Posted by: Brad DeLong on January 3, 2004 11:12 AM


Could this not have something to do with emloyers sending away employees telling them "we'll call you when there is work again" and so employees considering themselves still employee of that employer and saying so when survey people come visiting, while the employer thinks of the number of employees currently on the payroll at the time someone inquires about number of employees at the enterprise?

Posted by: bulent on January 3, 2004 11:36 AM


One factor is 'independent contractors' who aren't really independent. Sometimes businesses will do this to avoid immigration restrictions (hire a foreign consulting company consisting of one person,) or to avoid paying payroll taxes. They won't necessarily even tell the employee-- a friend of mine's employer went under, and then as a kicker she got a 1099 instead of a W-2, and that was how she found out she was going to have to pay her own payroll taxes. Lawsuits followed of course.

Posted by: Verbal on January 3, 2004 03:45 PM


Couldn't the discrepacancy between the two surveys amount to a measure of the informal and underground economy, with millions of people working off the books for cash, whether because of their immigration, disability or retirement status or simply to avoid taxes?

Posted by: john c. halasz on January 3, 2004 04:28 PM


Wow, 2+ million is quite a difference. I wonder what this means for the credibility of the unemployment numbers, as they are coming from the survey.

I don't fault the BLS, it is inherently difficult to measure these numbers correctly and making adjustments that reflect the truth and actually enhance the result.

I do however fault them for not disclosing more detail on how the numbers are collected and processed. Increased transparency will allow interested people to figure in shortcomings of the measurement; otherwise your only choice is to distrust the result.

Regarding self-employed people, what I'm reading out of the article is that there is no distinction between having a contract or being "between contracts"? That can certainly mask a lot of effective un- or underemployment.

Posted by: cm on January 3, 2004 04:37 PM


Hmmm. Payroll was consistently about 90-91% of household for twenty years to about 1988-89, then shot up to 96% by 2000. What's going on there?

Posted by: David Thompson on January 3, 2004 05:17 PM


>>Hmmm. Payroll was consistently about 90-91% of household for twenty years to about 1988-89, then shot up to 96% by 2000. What's going on there?<<

You'd think this should be something I know, wouldn't you?

I think that this should be something I know.

I don't.

Posted by: Brad DeLong on January 3, 2004 05:23 PM


EHC (Economic Harmonic Convergence) of 1990's gets severe shock from 9-11, loss of high paying service sector jobs leads to decline. 9-11 shock also leads to disruption of EMCF (Economic Mind Control Factor), that is, citizen's confidence in economist's ability to control our economic minds, and thus the economy, leading to further job loss.

Was there change in composition of jobs in 90's, that is, more health care and other service jobs to replace manufacturing in late 80's? Does loss of certain types of jobs have worse ripple effects than others? I'm just thinking of some things that expanded in the 90's - health care, travel, government services (windfall capital gains taxes collected).

Posted by: northernLights on January 3, 2004 07:48 PM


Re: decrease in the ratio in the 1990s. I'm self employed-S corporation (since 2001). Currently working for Lockheed Martin. I pay myself the smallest salary that the accountantn will allow (minimize FICA). I think I am probably captured in the payroll survey based on my company's 941. I wonder if the IRS tightened up on this in the 1990s. I recall that 1099 arraqngements were under assault (engineers, over the road truckers). What this meant is, I think, in order to be a contractor, you had to incorporate and have business liability insurance (not a trivial expense).

Posted by: Robert Monical on January 3, 2004 10:46 PM


Robert: unfortunately the discussion of a common theme has been split into multiple threads. Chances are you are not in the employment survey as it covers only 60,000 households. One thing I learnt from one of the referenced articles that I was not aware of was that _apparently_ people who are self-employed but between "gigs" are counted as employed. Can you offer some insight? For example, can you apply for unemployment benefits between jobs? (Hint: if your employer/client has a shutdown, you may be eligible for unemployment during the period if it does not contain paid holidays or such, as I heard. As a contract person you shouldn't have such.)

Posted by: cm on January 4, 2004 12:00 AM


if John C. Halasz´ suspicion is even remotely plausible, I would expect sociologists to have researched the topic. At least that is what sociologists in Germany did. Their work proved to be useful indeed, since almost half a dozen studies independently arrived at the same conclusion: 20-25% of unemployment benefit recipients in Germany have a job in the shadow economy. That is why I have often claimed in the past that unemployment is actually not very much higher in Germany than in the U.S. Of course, if you you report back to us that your colleagues in the sociology department have come up with similar findings for the U.S., my theory is kaput.

Posted by: Joerg Wenck on January 4, 2004 08:27 AM


CM: You are correct, I meant the payroll survey would catch my when I was pulling a paycheck. Since I am an employee of my corporation (and I pay unemployment on me when I am drawing a paycheck) I am never technically unemployed even when I cannot pay myself.

I was specifically addressing the change in the ratio of payroll to household data in the 90s. This is when I think that the 1099 “abuses” were being squeezed out of the system. Many 1099s failed to pay FICA. A heinous sin as FICA grew to cover ever more of the Federal budget in the 90s. There are still many people earning significant money in home businesses that are not paying FICA on the income. For example, my wife's Social Security benefit would never equal what whe would get by basing her SS retirment income on my SS history. So even though she works part time in the business, she does not draw a paycheck.

One other possible explanation for the discrepancy relates to the notion of “churn”. If one is between jobs (for example, many people take a few weeks when transitioning from one technical job to another) they would be technically unemployed from a payroll perspective but would consider themselves employed.

One other dynamic: my late Uncle was in the Merchant Marine. They union had a rotation where he worked about 6 months out of the year. He would consider himself employed even though nobody was paying him. He would ship out when his number came up again. Made very good money in the 6 months that he worked.

Posted by: Robert Monical on January 4, 2004 10:02 AM


Robert: Thanks for the explanantion. In the cases you cite (your uncle's part-time arrangement, taking time off between jobs) the time off is voluntary, or at least welcome. That's not what any reasonable person would consider unemployment, and it is not part of the ILO's or anybody else's definition.

But with people waiting for callbacks it becomes more murky, or if for a self-employed it is so hard to find assignments that they have paychecks (and this is finally where is comes down to, is it not?) only 1/3 of the time. Then the time "off" is not voluntary, and in general people who are willing to work but can't are considered at least "not employed", and probably "unemployed", the latter being perhaps qualified by some notion of "seeking work", although the person waiting for a callback may not be technically seeking.

I suspect that maybe for cultural reasons or pride some people may view an ongoing "business relationship" (waiting for callback) or time spent hunting for clients as "employment", where I would not, and by the spirit if not letter of employment definitions one should not.

In Germany for example people who work less than 15 hours per week and who seek an additional or a full-time job (through the DOL) are considered unemployed.

Posted by: cm on January 4, 2004 02:48 PM


Consider the problem of Zoe Baird and her nanny.

The nanny certainly would consider herself an employee. Ms Baird, however, would incur subtantial expense if she considered herself an employee.

Posted by: Pouncer on January 4, 2004 04:42 PM


The medicare payroll tax is universal isn't it?

Shouldn't it be possible, on at least a lagging basis, to see how many unique SS # paid taxes in a given year? I can see how it would be difficult to use the payroll tax data in real time, but it seems like it should provide insight into total payrolls.

Is the aggregated data restricted or does the government software / computer system lack the capability to aggregate the data?


Posted by: Doug on January 5, 2004 03:24 PM


Post a comment