Employment Shrinkage: Plus ca Change?

This last spring, I posted on how I thought that there were basically two recessions going on. That is that there were two sectors that made up a disproportionate share of the employment losses. From that post:



What I've done here is calculate what it would look if employment shrinkage had been spread equally across economic sectors weighting for their percent of employment.

For example, if employment shrinkage had been distributed equally 275,862 construction and mining jobs would have been lost. In reality over a million jobs disappeared in this sector. In contrast, if employment shrinkage had been distributed equally 600,984 education and health jobs would have been lost across the country. In reality this sector gained 487,900 jobs.

Same contrast with manufacturing and government jobs. If employment shrinkage had been distributed equally 467,980 manufacturing jobs would have been lost. In reality over a million jobs disappeared in this sector. In contrast, if employment shrinkage had been distributed equally 758,984 government jobs would have been lost across the country. In reality this sector gained 487,900 jobs.

What I plan to do in this post is a sort of update on the one I did in the spring.

First, the structure of employment is continuing to change. With these numbers remember that a 0.1% of employment is over 139,000 jobs. No small change.

As can be seen, traditional blue collar sectors (Manufacturing, Construction, Trade et al) continue shrink relative to the new service sector (Leisure and other services). White collar employment as a whole (Info,Financial, Professional, Education and Health, Government) is up, however private white collar employment (Info,Financial, Professional) is down, and growth is being driven by public employment (Education and Health, Government). Given the growing fiscal problems at the federal, and in particular state and local government, that should be a sign for concern. On the up side, a very smart policy of avoiding putting public employment to the knife probably helped avoid a more serious recession (aka a depression).

Another aspect of the employment shrinkage between March and August of this year has been that Manufacturing and Construction continue to be hit disproportionately hard. Employment shrinkage in these sectors is over 3 times what we would expect if all sectors were equally affected.

Overall between the March and August of this year employment shrunk by 1,977,400. On a monthly basis that means that employment shrank by 329,567. In comparison, in the 16 months between the start of the recession and March 2009, employment shrank by 4,610,100. Or on a monthly basis 288,131.

Another way to look at this is compare changes in the employment level to the size of the labor force. In the 16 months between the start of the recession and March of this year, employment fell by 4,610,100, on the other hand the labor force actually grew by 1,368,300. Obviously this means that unemployment is going to rise, because there are less jobs available, and more people looking for them.

Compare this to the six months between March and August. Employment fell by 1,977,400, but the labor force also shrank by 523,900. Less jobs, but also less people looking for them. So the unemployment rate falls.

The question that needs to be asked is why these people are leaving the labor force. If they are so discouraged that they've just given up hope of ever finding a job, then they are arguably still unemployed, but they just aren't counted.

If we hold the labor force (50 states+DC-Territories) constant, and then show the effect of employment shrinkage on the unemployment level, then unemployment increases by 0.3 points from 9.67% to 9.97%. In short, the shrinking labor force conceals around 3-4% of total unemployment.

The lesson, beware of the raw unemployment rate. The unexpected is at work here concealing a portion of the unemployed from view.




I’m returning to a question I raised a couple of days ago with another writer who made a noble effort (twice) to help me.

Please forgive me, but in college I had an elective to fill and I made the mistake of taking a philosophy course. I haven’t been thinking the same since. I picked up this habit of needing precise definitions.

In employment discussions there are so many terms that I can’t keep them straight: UNEMPLOYMED (various categories like ‘first time’, ‘longer that x# of weeks’, ‘used up benefits’, ‘extended benefits’, ‘discouraged’, ‘continuous’, have I missed any?); JOBS LOSS; JOBS GAINED; EMPLOYMENT SHRINKAGE; EMPLOYMENT LEVEL; UNEMPLOYMENT LEVEL; JOBS AVAILABLE’…

More specifically, I’m really interested in the ‘Jobs Loss’ category. Does the term ‘job loss’ (singular – one job loss) denote a 'person' who gets the proverbial ‘pink slip’ or does it mean that the 'job' per se that the person worked no longer exists. For example, if there is a secretary pool of five secretaries (i.e. five jobs) and one gets ‘laid off’, there are still five jobs, but only four are filled and the expectation is that eventually the fifth job will be filled again (i.e. no 'job loss', the job still exists; the desk, computer, file cabinet, etc. are still there waiting to be used again).

But, if the company eliminates one job, i.e. there are now only four jobs and there is no expectation of hiring a fifth secretary, then the job has been LOSS (i.e. the desk, computer, etc. are removed hence forth there will not be more than four secretaries).

If you have the time and are off the mind, perhaps you could indulge me… If not, I understand. Socrates was a pain too.

Thank you

Interesting question

Unemployment and 'Job Losses' typically measure two different things. Both are based on surveys. Unemployment surveys households as to status of employment. This from the Bureau of Labor Statistics (BLS):

People are classified as employed if they did any work at all as paid employees during the reference week; worked in their own business, profession, or on their own farm; or worked without pay at least 15 hours in a family business or farm. People are also counted as employed if they were temporarily absent from their jobs because of illness, bad weather, vacation, labor-management disputes, or personal reasons.

People are classified as unemployed if they meet all of the following criteria: They had no employment during the reference week; they were available for work at that time; and they made specific efforts to find employment sometime during the 4-week period ending with the reference week. Persons laid off from a job and expecting recall need not be looking for work to be counted as unemployed. The unemployment data derived from the household survey in no way depend upon the eligibility for or receipt of unemployment insurance benefits.

The civilian labor force is the sum of employed and unemployed persons. Those not classified as employed or unemployed are not in the labor force. The unemployment rate is the number unemployed as a percent of the labor force. The labor force participation rate is the labor force as a percent of the population, and the employment-population ratio is the employed as a percent of the population.

As for job losses, private nonfarm businesses such as factories, offices, and stores, as well as federal, state, and local government entities are surveyed. Again from the BLS:

Employees on nonfarm payrolls are those who received pay for any part of the reference pay period, including persons on paid leave. Persons are counted in each job they hold. Hours and earnings data are for private businesses and relate only to production workers in the goods-producing sector and nonsupervisory workers in the service-providing sector.

It comes down to two different perspectives on employment: household and business. So, when a job is lost it usually means it was cut from a businesses payroll.

I hope that answers the question.

RebelCapitalist.com - Financial Information for the Rest of Us.

Thank you etc.

Thank you. As I said before: “Getting answers to questions on this blog is like going to graduate school.”

If I may, I would like to probe this a little further in the context of my secretary pool example.

First, let me say that I am currently reporting on Italian American census data in my blog. A question was put to me about the difference between ‘ethnic’ and ‘race.’ To answer the question, I went to the questionnaire that the census taker uses. The difference is very obvious just by reading the questions and increasingly understandable with the supporting documentation.

In unemployment statistics, I know the question that is asked. Something like: “Did you work, were you available for work, did you look for work, etc.” A former mason, I’ve answered that question more times than I care to remember. If you did not work, were available for work and looked for work, then you are counted as an unemployed person. All those people are aggregated into the unemployment statistics.

What I can’t find, and I have been though numerous BLS pdf documents, is the exact question that is put to the employer about ‘number of jobs.’ There are many pages of “concepts”, but no questions. Concepts such as your BLS quote: “Employees…are those who received pay…including paid leave. Persons are counted in each job that they hold”

Then JOB = EMPLOYEE; i.e. the number of jobs equal the number of employees. Then the number of JOB LOSSES per unit time (should) = INITITAL UNEMPLOYED per same unit time. In my secretary pool as soon as the secretary is laid off (assuming it is not temporary) the number of jobs in the firm is reduce from 5 to 4. One job loss = one unemployed?

If this is the case than ‘job losses’ should be roughly equal to unemployed. This is not true. Job losses for the whole month of September were roughly 250,000, while weekly average of initial claims was roughly 550,000.

Again, I don’t mean to belabor a Platonic definition. But, the ideological debates (‘green shoots’ or not), ergo political debates, turn on such equivocal definitions of employment and various other economic variables. X says things are getting worse ‘look at increasing rate of unemployment’. Y says things are getting better ‘look are decreasing rate of job losses.’ Over at Bonddad, New Deal Democrat is promising us, like the Old New Deal Democrats, a “chicken in every pot”. Others disagree. Both are armed with lots of statistics.

Thank you again,

Remember these are surveys.

One job loss will not equal one unit of unemployed because they are not surveying the same sets - exact workers to exact employees. You really have to read the definitions and methodology of the survey. These are just indicators there will always be sampling errors and certain biases in these numbers.

There will always be that disagreement between green shoots vs. green weeds in some form or another. And it's hard to figure out who is right. But I guess in a way this is human nature.

The predictive quality of economics is highly questionable particularly in our current situation. In my opinion, what matters are the levels of unemployment, debt and other things. Small or some other incremental change means very little to me since we are in such a huge hole. But that is just my opinion.

RebelCapitalist.com - Financial Information for the Rest of Us.

Let me just add

This is not like a typical recession. This recession was caused by a financial crisis. History, according to Prof. Ken Rogoff, tells us that recessions caused by financial crisis are deep and nasty. Does this mean our Great Recession will be any different? I don't know and neither does New Deal Democrat (despite the well researched and reasoned analysis).

RebelCapitalist.com - Financial Information for the Rest of Us.


My acerbic reply isn't, of course, directed at you, RebelCapitalist, but your mention of Rogoff does provide my opening.

I happened to turn on NPR the other night and Rogoff was on, laying forth on his recent econ history text and prosyletizing on the recovery and the end of the recession!

He sounded amazingly ignorant and stupid, but then, Rogoff is a member of the Group of Thirty (same as Summers, etc.) and the Bretton Woods Committee (whose influence greatly watered down the "Buy American" clause in the stimulus which appears to be going more towards the multinationals then American workers) and the Peterson Institute (whose long-term agenda is the offshoring of as many American jobs as possible, the privatizing of Social Security and proclaiming anything and everything which might aid the American worker - such as the auto industry aid - to be against the WTO's Financial Services Agreement - whil dutifully ignoring that all those bankster bailouts were directly against the WTO's Financial Services Agreement).

James, I understand.

I certainly don't agree with Rogoff on everything (probably very little) but he and a colleague did publish an interesting paper about history of past economic crises caused by financial sector.

RebelCapitalist.com - Financial Information for the Rest of Us.

Just a wayward thought on the matter.

I just had to take this opportunity to spew forth on Rogoff, but please consider, if he has at least equal access to the same information I do -- and as I've seen the undergrad library at Harvard University and imagine he probably has much better access to data given his position and placement -- and he has arrived at the "learned" conclusion that the financial crisis is over --- while many of the rest of us consider that it has just begun --- then what scholarly accuracy and historicity do you really consider his writings to possess?

Off to Chicago for the union protest against that American Bankers Association meeting there -- hope not to encounter any trouble from the Chicago PD, but if it's combat they want, they'll find a number of combat vets there, from across several generations, at least.

The Magical Mystery Tour

You are logically seeking answers to stats which have been purposefully obfuscated. Employers only above a certain size have to report actual unemployment numbers or laid off workers, and many times will ignore layoffs at subsidiaries as they can frequently get away with listing them as a unique, small business entities.

Likewise, sometimes they will count their foreign (as in offshore) job site employees, sometimes not. (They may also include them in their productivity stats, while not including them in their actual worker count.) The most difficult stat to arrive at is actual number of the unemployed for these reasons and numerous others. Just arriving at the number of offshored jobs is impossible from straight data, as it is considered "voluntary" information when sought by the GAO and Congress, and the last request in the early part of this century '00 or '01, yielded only a 1% response (I repeat ONLY 1% response) from corporations inquired of on this topic. From that, various stink tanks made erroneous extrapolations which were obviously statistically incorrect and invalid.

The crucial items are the BLS study, reported here and in the NY Times some weeks back, which indicated that there had been effectively zero job creation in the Private Sector from July 1999 to July 2009, extraordinarily indicative of having reached critical mass in the offshoring of American jobs (as in, there's hardly any offshorable jobs left - thus bringing the unemployment jobs loss multiplier into effect) -- as well as 9 consecutive months of rising unemployment in the 372 American cities surveyd by the BLS -- as well as 9 consecutive months of downward trending in payroll hours of the employed.

It is similar to that phony process they use to "record" exports, counting everything shipped to any offshore factory or production site as an export, even though it isn't receiving any payment in return, but will go into that creation process, thus artificially (and fraudulently) driving up the number of exports astronomically, when in fact they are actually quite paltry.

what is the source of your raw data?

census, bls ... how about a URL?