Jobs paradox?

Some readers have asked how it’s possible for unemployment to fall when the economy is still losing jobs, albeit at a slower rate. The answer is a bit annoying.

First, the jobs number and the unemployment number are based on different surveys — a survey of establishments in the first case, a survey of households in the second. Sometimes employment rises by one measure while falling by the other, although it happens that this month there isn’t much difference in the jobs number. (The establishment survey is considered a more reliable measure of month-to-month job changes.)

Second, how do we measure unemployment? Contrary to what some correspondents think, it doesn’t have anything to do with receipt of unemployment insurance. It comes, instead, from a survey in which people are asked whether they’re working and, if not, whether they’re looking for work. And what this month’s data show is a relatively large rise in the number of people “not in labor force” — neither working nor looking for work. That’s how the unemployment rate can fall even with fewer people working.

Isn’t U6, the broadest measure of unemployment, supposed to include people who are discouraged and stop looking? Yes — but at least according to the survey, that’s not the reason more people have dropped out of the work force.

Basically, though, what you need to bear in mind is that these are imperfect measures, subject to a fair bit of noise. When the trend in the labor market is very strong in either direction, the measures move together. But when you have the kind of scene we have now — the employment situation is drifting down, but not plunging — occasional mixed signals are likely. No big deal.

The basic story is that things are sort of stabilizing — but they’re definitely not improving yet.

Comments are no longer being accepted.

The government compose the jobless rate by counting people who claim for unemployment checks. Those people usually are from companies that are big enough to pay unemployment insurance. 75% percent American business are related to service and when the lay off from the big companies happen the consumption are decrease. The impact then reach to the small businesses and causing lay off. The problem is that many workers from small business are not entitle to get unemployment. In another word we are facing the unemployment situation that we couldn’t see.

I think the household survey is the better number since it takes a random sample rather then a biased sample of known existing establishments. Coming off a recession there are bound to be new establishments that are not included in the old list. The old list is overweighted with companies that were failing.

If 34,000,000 people are, without debate, getting food stamps and the population of the US is roughly 300,000,000 people (or 340,000,000 for arguement sake) then isn’t the true unemployment rate AT LEAST 10%.

I’m just an ignorant Engineer but I can divide the food stamp and SSI numbers by the general population.

In my mind…for what it’s worth…isn’t the unemployment rate even more extreme when you use a much smaller working-age population in the denominator???

How am I wrong?

Last month’s worse than expected jobless numbers led this blog to call for a second stimulus package, with a doubt.

This month’s better than expected jobless numbers lead to a comment that this measure is imperfect.

I think what it really shows is that a second stimulus package may not be needed.

As Will Rogers once said, “there are lies, damn lies, and statistics”. It’s all in what you measure and how you measure it. Still I believe that all of the evidence taken together suggests that the economic ship is starting to turn.

My mom was forced into a early retirement. Despite the fact she is looking for a job, she is considered retired not unemployed. I wonder how many other people like her are skewing the numbers.

the job market is stabalizing?? do people really believe what you write?? i would invite all of this blog’s readers to do their own research.

its like this: the job market is either growing or it is contracting… there’s not a lot of middle ground… when have you ever seen a jobs curve that has flatlined?? yes, its true at the bottom of a curve it is neither up nor down but is it possible to be in the exact bottom of a curve for months on end?? no, this is still downward sloping with no end in sight despite what The Professor or Obama tell us.

The employment to population ratio fell from 59.5% to 59.4% (labor force participation fell even more from 65.7% to 65.5%). Justin Y: the continued weakness in the labor market tells me that we do need more aggregate demand stimulus.

Dr. Krugman,

These mixed — yet no longer disasterous — unemployment statistics suggest that we either have hit a bottom or are approaching the bottom of this sever economic downturn. If so (and that may still be a presumptuous assumption), here are some preliminary thoughts as to why stabilization — if not outright recovery — is occurring on this somehat faster-than-anticipated schedule:

1. Although the explicit stimulus was too small relative to the lost aggregate private demand, and was weighted too much toward tax cuts, the “hidden stimulus” of the bailout to financial institutions and markets (a fiscal policy not a monetary policy) may well have been of sufficient magnitude to allow the financial sector to reume its intermediation role. I had previously written on this blog comment page that we would get the stimulus we need, not the stimulus we want. Well, the old-fashioned percolate-up stimulus played a part (the stimulus we wanted), and so may have the financial bailout (the stimulus we needed).

2. The bailout and changes in financial accounting mark-to-market rules may have sufficiently improved the balance sheets of financial instituions at a speed greater than occurred in Japan during its supposed “Lost Decade.” Thus, we may have attenuated the duration of the “balance sheet” recession/depression of which Richard Koo has written.

3. This “balance sheet recession” is a phenomenon that seems to have been expressed by other economists in other ways. For example, we could as accurately say that the Administration’s actions have walked us away from the Minsky Moment which occurred last fall. We could also say that the Administration’s actions have unleashed some of the “animal spirits” of which Keynes spoke, and of which Akerlof and Shiller have recently written.

On balance, perhaps these results from these policies indicate that President Obama’s non-ideological and pragmatic approach, while imperfect in certain perspectives, was a way (the only political way?) to solve the macroecnomic problem he faced.

@Pia Fields, that’s not how they measure unemployment, at least not in the BLS survey. The unemployment figures come, as Prof. Krugman mentioned, from the household survey, so they actually call people up and ask them if they’re looking for a job.

@morgan, I’m not sure your mother would be considered ‘retired’ either, as long as she’s looking for a job.

@Marcus Pratt, food stamps are open to low-income families, not just no-income families. They might be a better proxy for the poverty rate than the unemployment rate.

I would be very interested in hearing how the “not in labor force” numbers could rise so much without discouraged workers. The month’s entire population growth, according to the BLS figures, would only account for about a third of the change. Anyone have any other ideas?

morgan,

If the Bureau of labor statistics calls your mom and she tells them she is both unemployed and acactively seeking employment then she will be included in the unemployment figure.

Pia Fields,

The unemployment number is based on a random survey and is not related to unemployment insurance.

//www.bls.gov/news.release/empsit.faq.htm

Re #1/Pia: “The government compose the jobless rate by counting people who claim for unemployment checks. ”

Did you not read what the author said?: “Contrary to what some correspondents think, it doesn’t have anything to do with receipt of unemployment insurance. ”

Also:
//www.bls.gov/news.release/empsit.tn.htm
“The unemployment data derived from the household survey in no way depend upon the eligibility for or receipt of unemployment insurance benefits.”

And:
//www.bls.gov/news.release/empsit.faq.htm
“Is the count of unemployed persons limited to just those people receiving
unemployment insurance benefits?

No; the estimate of unemployment is based on a monthly sample survey of households. All persons who are without jobs and are actively seeking and available
to work are included among the unemployed. (People on temporary layoff are included even if they do not actively seek work.) There is no requirement or question relating to unemployment insurance benefits in the monthly survey.”

I understand and appreciate statistics having worked with them a lot. But whether or not the economy is getting better, I have not been able to find paid work although I have stellar education and work experience. To add insult to injury, although qualified for unemployment in California, I am not receiving it, due I believe, to a system that is totally
overwhelmed with applicants. Fortunately I have some savings, but they will eventually run out. And then??

Joseph O’Shaughnessy August 7, 2009 · 11:45 am

247,000 is a long way from home, baby. That isn’t even second base. Not with 14,000,000 on the dole and another year of 3 million foreclosures.

We are a remarkably resilient country, that’s for sure, and a nice thing to see we still have some of that. But we’re in a big hole and no one should be fooled by climbing markets or slightly lower bad numbers.

This will be a long haul. We’re coming out of a bad recession this time, “bad” bad, and we have no manufacturing to speak of. Maybe we’ll catch on; maybe someone will get it; but I just don’t see the leadership except for one man on horseback and Sancho Biden.

We interrupt the cheering for the better-than-expected, less-bad-than-previous-months unemployment increase for the following announcement:

Seasonally adjusted private nonfarm employment, August 1999: 108,959,000.

Seasonally adjusted private nonfarm employment, July 2009: 108,924,000(Preliminary).

A loss of 35,000 private jobs in TEN (10) years.

Dr Krugman:
I majored in Mathematics in college with a minor in Statistics. For many years I was a computer programmer but those days are gone in part due to the trend of “offshoring” IT jobs. My question is since the current recession has shed 6.7 million jobs and there is a need, as you insightfully pointed out, for around 1.5 million new jobs per year to account for additions to the labor force wouldn’t the following be needed to truly stabilize the economy:

In the next 4 years, the creation of 6.7 million jobs to absorb those who have lost jobs + 2.3 million jobs to absorb the additions to the labor force since 2008 + 6 million to absorb the additions to the workforce over the next 4 years for a total of 15 million.
So wouldn’t we need at least 15 million new jobs in the next 4 years-something as big as what we had from 1996-2000 just to break even?
I guess what I am asking is after this complete economic meltdown, will job growth really matter if it fails to be extraordinary job growth?

Marcus,

You don’t have to be unemployed to be on food stamps. About 1 in 5 jobs in the US pays below the poverty line for a person trying to support a family. We work 2-3 weeks longer per year than in the 50’s and make less in proportion to cost of living. Executive compensation? Not so much. Thanks Corporate America!

Aaron,

You were right to correct Pia and Morgan on the methodological issues above and you explained it very well. I am assuming you are the same Aaron from an earlier post at 10:22. In that post, you said that the establishment survey is biased because it doesn’t take into account new establishments. The BLS is aware of the downward bias in estimation that is created by sampling existing reporters and imputes for that using trended universe figures in from preceding years. They call it the Net Birth/Death model. Since there is no way to sample new units as they are born, they must use trends from the universe, which has been shown to be fairly accurate. Here is the link explaining the methodology:

//www.bls.gov/web/cesbd.htm

OK, take 20% out of 34,000,000 and you get you get 27,200,000 people. Now, compare that with the official number of unemployed and tell me which is bigger.

So let me see if I have this right. They use the telephone to call and ask random people whether they are employed? Isn’t there a huge bias here concerning the act of answering the phone? You might not answer the phone if you are at work or if your are homeless. You won’t answer your work phone if you got fired. Maybe after getting fired, you cancelled your cell phone or decided to move in with your folks. How do they expect to know that their sampling is fair by calling people on the phone. Sounds ridiculous to me.

Andrew,

It is true that you don’t have to be unemployed to get food stamps. it is not true to say that the real wage has declined since the 50s, if that is what you meant by “make less in proportion to cost of living.” The real wage has risen by about 50% since 1960 to end 2008. (Even if you use an inflation rate that is biased to the upside by taking the maximum inflation rate from the CPI, CPUI Core, and PCE core in each year.)

This increase in wages is roughly 1.5% a year – in line with long-term productivity gains measured in a consistent fashion. Exactly what economic theory predicts.

It is nice that the headline number is now declining for the first time in 17 months.

But, the employment ratio declined to 59.4 percent of the population. That number, along with hours worked and wage growth, is the key number to determine our collective ability to service debt. We saw only slight (near zero) increases in hours worked and wages. So overall income is still apparently contracting.

As for debt it seems we moved leverage to the public balance sheet but I really do not see any net major deleveraging despite all the pain.

So are things getting better or worse? Maybe both at the same time?

Robert,

Can you collect SSI and be employed? So, why not use SSI plus continued unemployment claims as a minimum for unemployed people in the US?

This smoke and mirrors approach to what the unemployment situation is can be used as a tool of manipulation.

There isn’t anyway that 34,000,000 people can be on food stamps with an unemployment rate under 10%.

I am aware of some studies that purport to show a decline in real wages between 1960 and the present. However, I think they are nonesense. (Even if you did believe in them, they suggest that the bulk of the decline was between 1960 and the start of the reagan administration, whence the decline was halted and real wages have grown slightly.)

Marcus,

Continued unemployment clains could be used as a minimum, indeed there is such an index. it currently suggests a ‘minimum’ unemployment rate of 4.7%.