CEPR - Center for Economic and Policy Research


En Español

Em Português

Other Languages

Home Publications Blogs Beat the Press How Many Jobs Do We Need: Teaching Arithmetic To Economists

How Many Jobs Do We Need: Teaching Arithmetic To Economists

Sunday, 09 October 2011 08:10

I hate to harp on a seemingly small point, but it does offend me that professional economists and people who write on economics have such an aversion to simple arithmetic. One of the numbers that frequently in appears in discussions on the state of the economy is the number of jobs that we need to keep pace with the growth of the labor force. I have consistently been using 90,000 a month, however I see considerably higher numbers routinely thrown out, sometimes as high as 150,000 a month. For example, this Post article on the September jobs numbers told readers that the economy has to create 125,000 jobs a month to keep pace with the growth of the labor force.

There are two different ways to get to my 90,000 a month number. The first is to take the Congressional Budget Office's (CBO) estimates of the potential growth of the labor force. Its latest reports put this at 0.7 percent annually. Payroll employment peaked at just under 138 million before the downturn. Assuming normal growth, we would be at 142,000 million today Seven tenths of a percent of 142 million translates in 994,000 jobs a year, or roughly 83,000 jobs a month.

Of course CBO is not God, they could be wrong. But on the other hand, since their projections on the deficit are treated with such enormous reverence in the same news outlets, it would be absurd to just dismiss the economic projections that provide the basis for the budget projections. In other words, if we take CBO's budget projections seriously, then we must take their economic projections seriously. The latter are the basis for the former.

The other way we can get to the 90,000 a month number is by taking the Bureau of Labor Statistics (BLS) estimates of the growth in the non-institutionalized population and multiply by the employment to population ratio (EPOP). According to the BLS, the non-institutionalized population grew by 1,750,000 last year. If we apply assume a pre-recession EPOP of 63 percent, this implies an increase in employment over the last year of 1,103,000. If we assume that 6 percent of these workers will be self-employed (the average for the current workforce), this implies that we would have needed 1,036,000 payroll jobs to keep pace with the growth of the labor force over the last year, or 86,000 jobs a month.

This is how I get my 90,000 jobs a month, I don't know where others get their higher numbers. Again, this is not an especially important point in the context of an economy that is missing 10 million jobs, but the lack of respect for arithmetic is. It was a lack of respect for arithmetic that caused almost all economists and economics reporters to miss the housing bubble and the stock bubble before it. If we can't prod the people at the top of the profession to do the simple arithmetic that underlies the claims they make about the economy then we are in serious trouble.

Comments (11)Add Comment
Chickens, Eggs and the Labor Force: Which Comes First
written by izzatzo, October 09, 2011 10:39
One of the numbers that frequently in appears in discussions on the state of the economy is the number of jobs that we need to keep pace with the growth of the labor force.

Any economist knows demand creates supply. Supply does not create demand.

The labor force doesn't just arrive on the scene and demand the number of jobs necessary to keep it fully employed. If that was the case large families would get more income than small ones, spurring a perverse incentive to increase the labor force even more until it implodes from massive unemployment.

Arithmetic indeed. For it to add up reverse the question and ask how much labor is necessary to satisfy the existing level of demand.

Don't count job entitlement chickens before they hatch which only creates an unnecessary labor surplus falsely labled as 'unemployment' via those willing to work but can't find it.

Let chicken demand come first and egg supply will follow to maintain an organically balanced equilibrium free of surpluses and shortages.

Stupid liberals.
written by Sandwichman, October 09, 2011 11:35
I'm afraid economists got their 150,000 number from the same place they get most of the rest of their ideas: from groupthink.
And graphs, for the people whose silly leaders follow them
written by Rachel, October 09, 2011 11:55
Part of the problem with the housing bubble was the economists who knew that it was likely to burst one day, but (unlike Dean) didn't take the trouble to say so very often. Too busy talking about whatever more emotive issues were occupying their followers at the time. And their followers, without adequate data about where things were going (which could have been partially remedied by a little education and a few dozen big graphs) were easily distracted by emotive issues.

And now many people are putting all the blame for our troubles on Wall Street, which admittedly deserves a great deal of blame. And so they neglect the role of the health care system in damaging the economy. And all the while the various interest groups are pressuring Congress to produce new health plans, all likely to inflict needless pain on our fellow citizens. And the citizens, without the necessary knowledge (or even basic graphs), certainly short of adequate leadership, can do nothing.
125,000 jobs
written by Robert Oak, October 09, 2011 2:53
I think they are "doing math", just not your math. If one takes civilian non-institutional population growth from Jan 2010 to Dec. 2010, one gets 2,057 million increase. Divide by 12 is 171.42k per month. Take Dec 2007 civilian non-institutional populationt to employment ratio, 62.7%, one gets 107.47 jobs. Take the monthly reported increase in non-institional civilian population for September 2011, you get 200k. Take that times 62.7%, you get 125.4k jobs needed to keep up with population growth.

I have no idea why you harp on this each month. It's completely dependent upon civilian non-institional population estimates, which frankly vary way too widely for my tastes. The fact is the BLS does an annual adjustment, but only between Dec-Jan, instead of averaged out going back for that year. This mon-mon adjustment that's really for the year throws off the base non-institional civilian population number, depending which start and end data point (date) one uses. We have not even gotten to age categories, such as the massive influx of immigrants below age 30, or people working longer or baby boombers.

I personally never give an absolute number on the number of jobs needd each month and then declare "this is the number". I show a few calculations and give ranges, estimates.

Really, all hell is going to break loose for January 2012, when finally, the 2010 Census data is incorporated, now that should create a slew of fictional press reports complete with bad math and where the real fun will start.
margin of error
written by Robert Oak, October 09, 2011 3:13
Why not talk about the margin of error? The reality is there is a 100k monthly margin of error on payrolls, CPS is 400k. I did an overview on that http://www.economicpopulist.or...ld-survey

Having monthly non-farm payrolls below the margin of error, 4 of the last 5 months, is scary as hell to me.
Sorry on previous comment,no spell check in IE forms, non-institutional.
written by urban legend, October 09, 2011 9:49
It's simply old data that was roughly correct before 2010. Since then, that population growth rate has dropped precipitously, as has the employment-to-population ratio. 85 or 90K is the better estimate now, and 135-150K is outdated info suitable for lazy reporters.
BLS on Population growth
written by Dean, October 09, 2011 11:20
I was looking at the estimate of population growth over the last year, but if we want to just take a shorter period we can look at 3-month interval I was using as the basis for comparison. (I would argue that annual projections are somewhat more accurate than the projection for any specific month, but we could ignore than for now.)
BLS projected that the non-institutionalized population grew by 582,000 between June and September or 194,000 a month. Assuming an EPOP of 63 percent, we should expect to see employment growth of 122,000 a month. Roughly 6.0 percent of these would be self-employed, leaving payroll growth of just under 115,000 a month.

Of course there are confidence intervals around these numbers, but there are also confidence intervals around the budget numbers. The latter are huge and never ever reported, why would we make a big point of reporting anything other than the central estimate with the labor market data?
written by urban legend, October 10, 2011 12:06
Why assume EPOP of 63%?mptame Keeping up with non-institutional 16+ population growth at any given time merely means applying the EPOP at that time. Right now, it's just a little over 58%.
wish the Non-institutional Pop was more accurate
written by Robert Oak, October 10, 2011 1:47
Dean: I think the monthly report which separates out those reading the BLS report and numbers versus reporters aping some number they got somewhere is the February release, with the January data. Since that's where the yearly population adjustments are. Your overall premise, that many reporters do not read the actual government reports has to be true from so many "headline buzz" articles that miss completely what the data actually implies. Not just on the unemployment report, but manufacturing, trade, CPI, you name it.

What bothers me about the BLS is that yearly adjustment done on just December-January boundary, it makes no sense to me and why they don't go back over the monthlies and modify each month.

Then, using the 2000 Census base for 11 years has to be putting in some bias, regardless of statistical algos for yearly adjustments.

I think the Census/BLS should track immigration status too. What's wrong with just having some data in this globalization/labor arbitrage age? We have illegals, alphabet soup of NIVs (temporary foreign guest worker Visas) that are included in the employment statistics and it biases the numbers, esp. when looking at occupations. STEM, for example, is notorious, where NIVs are used to labor arbitrage, offshore outsource.
Urban Legend:
My point really is if one just looks at the monthly and takes an EPOP of 62.7%, that's the ratio for 2007 unemployment rates. If you take the current one, 58.2%, considering the "not in the labor force" has so dramatically increased, you hide the # of needed jobs by throwing people who just arrived, increased overall pop., into "not in the labor force".

Also, considering the unemployment rate is based on the monthly Non-institutional population as shown in the report, is it more accurate to go with just the monthly increase and estimate "jobs needed" or the yearly? I get your use of yearly, esp. with that large end of year adjustment, plus the 2000 Census base being used, on the other hand, the report is deriviing ratios and unemployment rates and so on from the monthly number, exclusively.

To sum: Sure I want a better press corp., although it would kill my hobby here, but what I really want is more precise, detailed and accurate labor statistics and part of the reason we can't get it, is Congress, funding, politics.
lifeaholic of america political historian
written by clarence swinney, October 11, 2011 1:52
1921-2003 Democratic presidents created 1,800,000 jobs per year to Republican 800,000.
Carter + Clinton created 222,000 per month to 99,000 per month by Reagan + Bush I + Bush II

Wealth creates jobs? Ha!Try Bush 8.
clear explanation of assumptions
written by scott moore, October 11, 2011 5:07
Nice job on explaining your labor force assumptions. As to the BLS and why they do what they do- they are beholden to statistics. Here is a link to how they actually calculate unemployment, including labor force. It is a project, and they do it every month. http://connectmooredata.com/20...us-method/

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.