CEPR - Center for Economic and Policy Research


En Español

Em Português

Other Languages

Home Publications Blogs Beat the Press How Many Jobs Does It Take to Keep Pace With the Growth of the Labor Force?

How Many Jobs Does It Take to Keep Pace With the Growth of the Labor Force?

Thursday, 01 September 2011 19:55

I have been saying that it takes roughly 90,000 jobs a month to keep pace with the underlying growth rate of the labor force. This means that more rapid growth should lead to declines in the unemployment rate while less rapid growth would lead to increases.

Many other analysts have used higher numbers. For example, in her NYT blognote, Catherine Rampell suggested that the necessary number for keeping pace with the growth of the labor force is 150,000 jobs a month. People have often asked me to explain the difference.

I can't say where others are getting their higher numbers from, but I know where I get my numbers. The Congressional Budget Office projects that the labor force will grow 0.7 percent annually for the next several years. If we go back to the pre-recession level of payroll employment (140 million), this implies 980,000 jobs a year or 82,500 a month.

There is another way to back out a growth number. The Bureau of Labor Statistics estimates that the civilian non-institutionalized population over age 16 increased by 1,781,000 people over the last 12 months. If we assume that 64 percent of this increment to the above age 16 population is employed (roughly the 2000 percent) then this would imply an increase in employment of 1,140,000. 

However, this would overstate payroll employment slightly since roughly 6 percent of the workforce is self-employed. If we assume that 6 percent of the increment is also self-employed, this implies that roughly 1,070,000 payroll jobs are needed to keep pace with the growth of the labor force, or just under 90,000 a month. 

Anyhow, that is how I get my job growth estimate.

Comments (11)Add Comment
written by PeakVT, September 01, 2011 10:01
Rampell's 150,000 is higher than the numbers typically thrown about. But the population and labor force growth rate has slowed down significantly from just a few years ago, when annual increases of 1.1-1.3% were common. So she's just repeating slightly dated information. OTOH it seems unlikely that the growth rate will stay as low as 0.7%, so 90,000 could be inaccurate pretty soon as well.
written by BKM817, September 01, 2011 10:19
I have noticed the number cited to keep up with growth has steadily and rapidly declined over the last decade. From 250000 during the first Bush recession, to 125000 during the financial meltdown and now down to 90000 with the Obama recession.
Why are we even arguing about this?
written by SM, September 01, 2011 11:19
Here's my question: why are we even arguing about the "number of new jobs" we need? Who cares what the "unemployment" rate is?

Even you say that we would be better off taking the current amount of jobs and having more work sharing - ie, 10 people split 9 jobs, 3 split 2, etc. But why is this?

When i look at the total GDP - would it not be the same? Whether we have 60% or 65% of the people working in this country, would we not create the same amount of goods?

Doesn't this just come down to an arcane way of distributing the goods that are made? That we have somehow conceded that someone must "work" to get what they "deserve"? Could we just leave it at 60% of people working, and distribute the goods in the same manner as if 65% were (ie, through tax policy, transfer policy, etc)? No one seems to care that "full employment" still means, what, 30% of the country is "not employed" at all? Why is that "acceptable" but 35%, or any other arbitrary number, is not?

Why are we arguing for more work to do? As far as i can tell, jobs are only useful so as to get work done - ie, create more wealth for the population. I think its stupid that we're still using them as a proxy as to how to distribute said wealth. And when you argue for technocratic solutions like "work sharing" - aren't you just supporting this idiocy? Who cares if 0 or 150k or 90k or whatever arbitrary number of "jobs" are "created"? Would we really be better off with more people doing "busy work" if we found 90k positions for this? Sure, we could pay 90k more people to dig holes then fill them up again, but what would be the difference if we just paid them to watch TV in their homes - ie, be unemployed?

Why not just come out and say, "we don't need more 'jobs' we need a better system of distribution because this one is broken"?

And after this, get to work putting people to work actually creating USEFUL goods and actual wealth for this country so that we are all better off, even if this means only 50% of people work.

Otherwise, why not just pay half the people to smash windows and employ the other half as glaziers? Imagine how high GDP would go!
written by Kat, September 02, 2011 11:22
"The Congressional Budget Office projects that the labor force will grow 0.7 percent annually for the next several years. If we go back to the pre-recession level of payroll employment (140 million), this implies 980,000 jobs a year or 82,500 a month."

But pre-recession employment is not the size of the pre-recession work force, so you appear to be taking a percentage of the wrong number. Please either elucidate or reconsider.
What about those not in the labor force?
written by Bill H, September 02, 2011 2:25
Are they not giving birth? Or are they only giving birth to babies that do not intend to work for a living, and who therefor will never become part of the work force?
SM's comment
written by Ian, September 02, 2011 3:39
SM, good comment. In the 1970's it was far more common for one parent in the household to stay home and raise a family. If we returned to that paradigm, unemployment would disappear. Perhaps there just aren't a need for many people in the workforce.
written by skeptonomist, September 02, 2011 3:51
The media, pundits and bloggers are in a tizzy about the latest jobs report, but there is little that is surprising in it. Private employment actually turned around at the beginning of 2010 and is still going up, though last month was a little weak:


while government employment has been continually going down


The decline is state/local, not federal, which has been steady. Given the way several states have been cutting jobs, the continued decline is entirely predictable. The recovery of private employment has hardly been spectacular, but it has actually been quicker than after 2001. As Dean says, there is little evidence of a double-dip. Barring another financial catastrophe, there is still a slow recovery going on, not nearly fast enough to make a dent in unemployment.
non-institutional civilian popuation
written by Robert Oak, September 02, 2011 7:45
It's completely dependent upon the growth of non-institutional population and the problem with this is they are still using the 2000 base and not the 2010 Census base as the starting point to extrapolate out growth projections.

Yet monthly we routinely see non-institional population increased by 200k or so, annual is lower.

Annual for the past year has that strange yearly adjustment in and it's not spread out annually, by taking august to august, so you're 1.7 million growth rate divided by 12 months is artificially low as a result.

Also, we've got almost 4 million in increase for non-civilian population in 2003,3 million 2007 and around 2 million for 2008, 2009, 2010. So, assuming less illegals popped by I would think taking a yearly of at least 1.9 million increase is more on target...

but if the Census is publishing anything on non-institional population from the 2010 Census data, that would be ideal to me.

It drives me crazy we don't have accurate non-institutional populations and then that 1 time yearly adjustment between December and January throws everything off.

To see what I'm talking about see this graph of the monthly change and notice the variance in the yearly adjustment:


This I think sums up the problem, getting a real base number, non-institutional civilian population upon which all of these calculations depend.

I wouldn't trust the CBO to save my soul on projections. :) I'll bet they are using our beyond belief low non-institutional population to employment ratios too in their labor force growth rate.

Also, self-employed, not incorporated, was 6.7% and you're missing agriculture workers as a subtraction from non-farm payrolls, but then again, we have so many part-timers at this point, might have to estimate those "two jobs" to get a job into the "jobs per month" projections.
written by Some guy, September 02, 2011 8:31
It's not completely dependent on population growth. It's dependent on population growth in the 90's less people leaving the labor force (retirements and deaths). Considering the average net monthly growth in the number of Social Security Disability and Old Age beneficiaries in current pay is about 170-200K a month and population growth in the 90's was about 275K a month, the 90K figure seems plausible.
To skeptonomist
written by In Hell's Kitchen (NYC), September 02, 2011 11:28
if you chart PAYEMS on a log scale you'll see the exponential growth trend broke in 2001 and nonfarm payrolls have been stagnant since then. In other words, the economy is missing about 39 million jobs. There is no way unemployment will ever fall again.

written by urban legend, September 03, 2011 1:13
A simpler way: annual increase in working age population = 1.8 million = 150,000 per month X 58.2% (current employment to population ratio) =

Write comment

(Only one link allowed per comment)

This content has been locked. You can no longer post any comments.


Support this blog, donate
Combined Federal Campaign #79613

About Beat the Press

Dean Baker is co-director of the Center for Economic and Policy Research in Washington, D.C. He is the author of several books, his latest being The End of Loser Liberalism: Making Markets Progressive. Read more about Dean.