Wednesday, May 10, 2017

Debunking Forbes: 14 In-demand Tech Jobs Employers Are Struggling To Fill

I'm writing this because some Congressman will inevitably cite Forbes and enter this piece of fake news into the Congressional record.

Forbes Business Contributor, Laurence Bradford, published a click-bait piece titled, "14 In-demand Tech Jobs Employers Are Struggling To Fill".  The article is 15 pages of photos, each containing a basic caption about a purported tech job. The reader has the pleasure of loading all of the Forbes advertising 15 times to get through the article.

Ms. Bradford claims that has the following, 1,076,505 positions that "...employers are struggling to fill."

Job #TitlePurported GlassDoor Postings
1IT Architect 10,487
2Security Engineer 18,686
3Data Scientist 20,498
4QA Engineer 23,543
5Front End Engineer 25,064
6Mobile Developer 40,112
7Java Developer 42,233
8Data Engineer 49,712
9Network Engineer 57,008
10Software Engineer 69,989
11Product Manager 70,488
12DevOps Engineer 155,476
13Solutions Architect 176,372
14Systems Engineer 316,837

The problem is that GlassDoor only has a total of 441,693 listings

GlassDoor: All Job Openings 441,693 Jobs

Forbes is a staunchly pro-immigration publication...

"Why Pro-Immigration and Pro-Growth Policies are"
- Forbes, Oct 25, 2011 - Economic growth is essential for progress. The broad liberal goal to bring prosperity and justice to as many people as possible relies on ...

"6 Immigrant Stories That Will Make You Believe In"
- Forbes, Oct 4, 2016 - What's also meaningful is that despite all the hot air America, a land of immigrants, remains decidedly pro-immigrant. A 2016 Pew Research ...

"What Will Donald Trump Mean For Immigration?"
- Forbes, Nov 9, 2016 - Ronald Reagan was profoundly pro-immigration and spoke of America as a “shining city on a hill.” Lyndon Johnson and other recent ...

"Illegal Immigrants Don't Lower Our Wages Or Take Our Jobs"
- Forbes, Aug 28, 2015 - Immigration has taken center stage in the race for the Republican Presidential nomination. The conventional wisdom says illegal immigrants ...

"Billionaires Rally Around Immigration, Against Trump's ..."
- Forbes, Mar 12, 2016 - Known for his continuous support for the Democratic Party, Soros' pro-immigration agenda dates back to 1996 and the Emma Lazarus Fund, ...

Archived pages:


Friday, May 5, 2017

Disinformation (brought to you) By

disinformation about computer occupations by
Click to enlarge

Disinformation IS Published by

For the latest in the computer worker shortage disinformation saga, we have a website called While promoting importance of computer literacy is important, their claims about demand for those skills are not based in reality.

Two of the claims on their Promote Computer Science page are as follows:
  • There are currently 521,444 open computing jobs nationwide.
  • Last year, only 42,969 computer science students graduated into the workforce.
The problem with these statement are:
  1. The "521,444 open computing jobs" are based upon 2024 projections. As of this writing, the current year is 2017.
  2. The "computer science students graduated" is selection bias, for the occupation(s)
We can be assured that there are not "521,444 open computing jobs." (A) The Computer occupations have never experienced that much growth in one year and (B) the Department of Labor's, JOLTS database, publishes job opening statistics. From the January 2017 JOLTS report, there were 5,625,000 openings across all sectors, we subtract 5,424,000 million hires to arrive at 201,000 open positions. Thus, it is impossible to have 500,000 open "computing jobs." You can read about the JOLTS definition of Active recruiting here. does have a document published that outlines how they arrived at these figures. That document confirms that they are using BLS 2014 to 2024 projections and also discloses the Occupational codes (SOC) they used (See Table 1).

In Table 1, I've combined their list with the BLS 2014 to 2024 projections data.

Table 1:

Computer Science codes: (Projections 2014-2024)Growth & ReplacementGrowth (Employment Level)Educational level
11-3021Computer and Information Systems Managers94,80053,700Bachelor's degree
17-2061Computer Hardware Engineers18,4002,400Bachelor's degree
25-1021Computer science teachers, post-secondary11,5003,800Doctoral or professional degree
15-1111Computer and information research scientists6,0002,700Doctoral or professional degree
15-1121Computer systems analysts191,600118,600Bachelor's degree
15-1122Information security analysts25,50014,800Bachelor's degree
15-1131Computer programmers81,000-26,500Bachelor's degree
15-1132Software developers, applications238,000135,300Bachelor's degree
15-1133Software developers, systems software107,90051,300Bachelor's degree
15-1134Web developers58,60039,500Associate's degree
15-1141Database administrators39,20013,400Bachelor's degree
15-1142Network and computer systems administrators79,40030,200Bachelor's degree
15-1143Computer network architects31,50012,700Bachelor's degree
15-1151Computer user support specialists150,50075,100Some college, no degree
15-1152Computer network support specialists36,90013,600Associate's degree
15-1199Computer occupations, all other37,7007,700Bachelor's degree
Totals 1,208,500548,300

So far, so good, we have the 548,300 in employment level growth that has referenced in their published document. Apparently, the BLS feels that 660,200 (15%) will leave the occupations by 2024.

Math & Computer Science factoid: 

Of those with bachelors degrees or higher 
(2,647,000 of 3,578,220) in 2012-13 data..
Employed% of Occupation
At least one S&E degree = 73.3%1,940,25154.2%
No S&E degrees but at least one S&E-related degree = 4.3%113,8213.2%
No S&E or S&E-related degree but at least one non-S&E degree = 22.5%595,57516.6%
Subtotal (slightly off due to rounding?)2,649,64774.0%
No Bachelors degree928,57326.0%

National Science Foundation, National Center for Science and Engineering Statistics, Scientists and Engineers Statistical Data System (SESTAT) (2013),

Source data:  and

ten years of bachelors degrees in Computer occupations
Click to enlarge

When we reference the WebCaspar database for Computer Science, in the method prescribed by, we find 60,266 bachelors degree conferred in 2015, this is quite a bit higher than 42,969 as stated by In fact, the 42,969 number does not exist in the current output.

Computer Science is not the only curriculum for the "computing jobs" designated by In fact, we can use the "SOC 2010 to CIP 2010 Crosswalk" to determine the Classification of Instructional Programs (CIP) codes for the Standard Occupational Classifications (SOC) codes in Table 2.

Table 2

SOC to CIP Crosswalk: for referenced computer occupations
01.0106 Agricultural Business Technology
11.0101 Computer and Information Sciences, General
11.0102 Artificial Intelligence
11.0103 Information Technology
11.0104 Informatics
11.0199 Computer Science, Other
11.0201 Computer Programming/Programmer, General
11.0202 Computer Programming Special Applications
11.0203 Computer Programming, Vendor/Product Certification
11.0299 Computer Programming, Other
11.0301 Data Processing and Data Processing Technology/Technician
11.0401 Information Science/Studies
11.0501 Computer Systems Analysis/Analyst
11.0701 Computer Science
11.0801 Web Page, Digital/Multimedia and Information Resources Design
11.0802 Data Modeling/Warehousing and Database Administration
11.0804 Modeling, Virtual Environments and Simulation
11.0901 Computer Systems Networking and Telecommunications
11.1001 Network and System Administration/Administrator
11.1002 System, Networking, and LAN/WAN Management/Manager
11.1003 Computer and Information Systems Security/Information Assurance
11.1004 Web/Multimedia Management and Webmaster
11.1005 Information Technology Project Management
11.1006 Computer Support Specialist
13.1321 Computer Teacher Education
14.0901 Computer Engineering, General
14.0903 Computer Software Engineering
14.0999 Computer Engineering, Other
15.1204 Computer Software Technology/Technician
26.1103 Bioinformatics
26.1104 Computational Biology
30.0801 Mathematics and Computer Science
30.1601 Accounting and Computer Science
30.3001 Computational Science
30.3101 Human Computer Interaction
43.0116 Cyber/Computer Forensics and Counterterrorism
51.0709 Medical Office Computer Specialist/Assistant
51.2706 Medical Informatics
52.1201 Management Information Systems, General

When we compare the last ten years of educational data (for permanent residents and citizens), these awards (1,235,998) are congruent with the BLS Projections (1,208,500).

Computing jobs: Degrees & Certificates conferred US Citizens and Permanent Residents (2006-15)
  • Doctorate – Professional = 9,899
  • Masters = 166,437
  • Bachelors = 544,578
  • Associate = 323,393
  • Post-grad Certificates = 11,012
  • 1 but less than 4 year Certificates = 62,541
  • Less than 1 year Certificates = 118,138

  • Total 2006-2015 = 1,235,998
Of course, the NCES colleges also educate foreign students, the capacity to exceed projections does exist and the United States grants 1.2 million employment based green-cards per decade.

Computing jobs: Degrees & Certificates conferred Temporary Residents (2006-15)
  • Doctorate – Professional = 11,805
  • Masters = 113,545
  • Bachelors = 41,312
  • Associate = 4,862
  • Post-grad Certificates = 2,076
  • 1 but less than 4 year Certificates = 974
  • Less than 1 year Certificates = 1,679

  • Total 2006-2015 = 176,253

Given the fact the NCES colleges have increased degree production by 20% of the 2006-2015 sample, it is safe to assume that there is an adequate supply of U.S. trained computer professionals.

Tuesday, March 14, 2017

H-1B: How many continue beyond the six years?

H-1B: How many continue beyond the six years?

In Computer Occupations, 22,000 second extensions granted annually.

Subjects: H-1B, Characteristics, AC21, PERM, Computer Occupations

I stumbled onto the 2015 H-1B Characteristics report, date-stamped March 17, 2016, it was embargoed from the public for almost a year. I think it is important to know what's happening within one's occupation.

On the topic of AC21 extension, I found that there are two types that are statistically significant. A seventh year extension which is annually renewed and a three year version if the intending immigrant will fall under the per-country quota backlog for PERM. There is no way i could find to tell the difference between the two numerically, but for the IT occupations (by percent), about 22,000 are filed annually. If the H-1B are primarily from India and China (3 yr. extensions), this would be 22,000 people, otherwise some portion of 22,000.

 This number is significant to me because I have to ask, does the employee accept sponsorship as a job perk? Does (s)he really want to be a citizen, or is it just more convenient to stay? In some cases, PERM requires advertising the job and a labor certification examination. Does this happen when the petition is filed, or after the H-1B holds the job for another ten years on a hopeless waiting list?  (As of January 17, 2017, DHS is giving itself 1780 days (4.8 years) to complete Labor Certifications.)

Alternately, if there is a dire shortage of these "brightest and best," why would only 1/3 be sponsored when PERM is so cheap?

Here's what the H-1B Characteristics report says about Continuing employment extensions. 

Continuing employment petitions refer to extensions, sequential employment, and concurrent employment, which are filed for aliens already in the United States. Extensions generally are filed for H-1B workers intending to work beyond the initial 3-year period up to a total of six years, the maximum period generally permissible under law. Sequential employment refers to petitions for workers transferring between H-1B employers within the 6-year period. Finally, petitions for concurrent employment are filed for H-1B workers intending to work simultaneously for an additional H-1B employer.

Neither AC21 nor prior legislation established a cap on H-1B petitions for continuing employment. Certain aliens are exempt from the 6-year maximum period of admission under the provisions of the 

American Competitiveness in the Twenty-First Century Act of 2000 (AC21), Pub. L. No. 106-313, 114 Stat. 1251. Computer Related H-1B

Computer-related occupations

Initial employment approvals
2010 = 31,661
2011 = 51,570
2012 = 83,444
2013 = 79,870
2014 = 80,877
2015 = 70,902

Total = 398,324 (66387.3 avg)

Total employment approvals (Continuing plus Initial)
2013 = 171,613
2014 = 203,425
2015 = 183,076

Total = 558,114

(Amount exceeding Initial employment approvals = 159,790)

Computer related, percent of all H-1B occupations
2013 = 59.8%
2014 = 64.5%
2015 = 66.5%

(Source data: Characteristics of Specialty Occupation Workers (H-1B): Fiscal Year 20xx_)

Concurrent and Change of employment applications
(Occupational share (percent of))
2013 = 26,493
2014 = 33,477
2015 = 34,041

Total = 94,012

Remainder of excess continuing approvals = 65,778 (21,926 avg yr.)

Employer is filing a second (or subsequent) extension of stay for an H-1B nonimmigrant.
(Occupational share (percent of))
2013 = 23,745
2014 = 28,975
2015 = 26,348

26,356 avg yr.

(Source data: FY 20xx Annual Report: H-1B Petitions)

By subtracting fee data from the Computer-related occupations in the Characteristics report, we arrive at an average of 21,926 extensions, but are unable to differentiate between single and three year duration.

By distributing the percentage of Computer-related occupations across "second (or subsequent) extension of stay" fee data, we arrive at an average of 26,356 extensions, but are still unable to differentiate between single and three year duration.


Saturday, December 3, 2016

Not in Labor Force NILF -- November 2016

Not in Labor Force (NILF) -- November 2016

Finding the hidden unemployment rate?

Subjects: Unemployment, Not in Labor Force, Working Age, Recession

Armed with the knowledge that the U-3 Unemployment Rate is subjective rather than objective, the nagging question is, "How do we compare current unemployment rates with historical unemployment rates?"

Not in Labor Force (NILF) is a top level catch-all category for those who are neither employed, nor unemployed (not actively seeking employment). The Civilian noninstitutional population includes the Labor Force, which is composed of the Employed, plus the Unemployed, everyone else considered Not In Labor Force. There will always be a certain amount of the NILF for various reasons, education, child rearing, disability. These are persons who did not work during the survey period and did not actively seek employment.

One way to model the effect of variance in the NILF would be to establish a bench-mark and distribute that percentage of NILF in all of the years going forward. With this model we effectively, establish what portion of the NILF would be considered unemployed and available for employment. In the chart below, I chose the average NILF (15.87%) for 1999, the last year we had full employment.

The chart uses the working-age population (25 through 54), to eliminate most students and baby-boomer retirements. This data is only available in the unadjusted series.

The (blue) chart data is the remainder of the NILF above 15.87%, which are then combined with the average number of unemployed for the year and then divided by the average employment level.

As expected, there is no deviation between the unemployment rate (red) and the unemployment rate plus the remainder of the NILF (blue) in 1999.

For the Bush 43 administration years, 2000-2008, we see that the headline unemployment rate would have been about 1.5% higher had these persons not been classified NILF.

For the Obama administration years, we see NILF obscure 3% of the unemployed in 2009 and over 4% in 2013, 2014 and 2015.

The St. Louis FRED is reporting a 3.8% unemployment rate in this age-group for November, 2016 (avg. 4.2% for the year), quite a departure from the 8.0% (1999 equivalent) in the chart.

See also: The plight of working age employment in the 21st century.

BLS Household Survey:
Series Id: LNU00000060
Series Id: LNU02000060
Series Id: LNU05000060