As we start the New Year, we face what is perhaps the most unpredictable and bizarre labor market I can remember. This morning the Conference Board released the December Help-Wanted Online report, which apparently shows a sharp increase in labor demand over the past year in most occupations. However, the BLS employment by occupation data shows no corresponding gain, even in occupations with soaring want ads. Nor does the unemployment by occupation data show any corresponding movements.
I have my own thoughts about what the data means, which I’ll share below. But first let me present the full array of data, by occupation, so you can make your own judgments.
The first column of data in the table below is the Supply/Demand Rate, as calculated by the Conference Board. That indicates the ratio of unemployed workers to ads, so a small number is better. The second column is the change in online ads over the past year, as measured by the Conference Board, so a big number shows that demand has ramped up. The third and fourth columns are the changes in occupational employment and unemployment over the past year, respectively, as measured by the BLS.
The ordering of occupations by supply/demand rate feels more or less right. But when it comes to the link between changes in demand, employment, and unemployment, there’s little consistency. We’ve got occupations with soaring demand and no gains in employment (management, transportation). We’ve got occupations with good supply/demand ratios and no gains in demand (health practioners). And so forth and so on.
*************************************************
| A Bizarre Labor Market | |||||
| Supply/Demand | yr/yr percentage change** | ||||
| Rate* | Want Ads | Employment | Unemployment | ||
| Healthcare practitioner and technical | 0.3 | 2% | -1% | 8% | |
| Computer and mathematical | 0.4 | 27% | -3% | 21% | |
| Life, physical, and social science | 0.8 | 40% | 4% | -3% | |
| Architecture and engineering | 1.0 | 47% | 1% | -18% | |
| Management | 1.4 | 56% | -3% | -5% | |
| Legal | 1.5 | -6% | 1% | -32% | |
| Business and financial operations | 2.1 | 6% | 2% | 10% | |
| Community and social services | 2.3 | 19% | -8% | 6% | |
| Arts, design, entertainment, sports, and media | 2.6 | 9% | -1% | 12% | |
| Healthcare support | 2.6 | 2% | 0% | 7% | |
| Sales and related | 3.5 | -2% | 2% | -1% | |
| Office and administrative support | 3.9 | 21% | 0% | 3% | |
| Installation, maintenance, and repair | 4.0 | 36% | 3% | 3% | |
| Education, training, and library | 4.3 | 22% | -1% | -7% | |
| Personal care and service | 5.5 | 8% | 5% | 14% | |
| Transportation and material moving | 7.4 | 61% | 2% | -2% | |
| Protective service | 8.0 | 34% | 1% | 36% | |
| Food preparation and serving | 9.2 | 26% | 2% | 6% | |
| Production | 10.8 | 59% | 10% | -6% | |
| Building and grounds cleaning and maintenance | 14.1 | 39% | -2% | 0% | |
| Farming, fishing and forestry | 28.4 | 38% | 2% | 52% | |
| Construction and extraction | 29.7 | 31% | -8% | -15% | |
| *Supply/demand rate, calculated by the Conference Board, is the number of unemployed workers | |||||
| divided by number of want ads based on latest data. | |||||
| **December 2009-December 2010 for online wants ads, November 2009-November 2010 for other data | |||||
| Data: The Conference Board, Bureau of Labor Statistics | |||||
| Calculations: South Mountain Economics LLC | |||||
*************************************************
My interpretation: The labor market is getting ready for a massive rise in employment over the next year, as companies finally start hiring for positions they’ve been advertising for.
We’ll find out soon.
Mandel on Innovation and Growth 
I hope you’re right.
Is there enough historical data to show how these vary over the cycle? I am reminded of a company that held a large job fair attended by over 500 people. A month later the number of jobs openings they listed on their site had increased from 52 to 57.
No, not enough historical data.
Are “multiple hat” job ads counted in each category to which they apply? For example jobs combining the skill sets/job responsibilities of IT analyst, database admin, project manager, and IT support liaison?
Of late there seems to be an increasing number of such multiple-role ads. There *are* organizations with not enough work for a full-time position (or an integral number of such) in each of the categories. But it looks more like companies trying to hire jacks-of-all-trades to get more “value” out of fewer people, and possibly more “flexibility” in the hope of multiple-hat people being more fungible. Unfortunately the more job roles and “skills” you pile on, the more unrealistic the ad, and the more it repels qualified applicants who are not looking for failing in a high stress job or wasting their time on an engagement that will likely not work out.
I usually assume that ads with long and implausible laundry lists were mostly to have grounds on which to reject “imperfect” candidates, or created by lazy managers/recruiters trying to cut down the candidate pool. But the example I gave above is genuine. The group that had this ad really wanted to staff such a position.
Then you have the phenomenon of recruiting being farmed out to recruiting agencies, or enterprising recruiters scraping ads, which leads to hard to quantify ad duplication (often with typos/grammar errors or signature phrases intact – and you stand a chance of finding the original ad by googling those).
Then there are “fishing” ads to collect or refresh resume/candidate pools, etc.
But the multiple-hat roles phenomenon seems to have gotten broader recognition lately, not least as part of the “structural skills mismatch” chatter.
That’s an interesting point to consider. “We can’t find enough good people” really might mean “we can’t find enough supermen to maintain unsustainable productivity numbers”.
I’m not sure “supermen” even enters into the picture. For the most part it’s probably “this is a time to ask for more” (regarding worker “performance”). But on top of what I mentioned many employers who complain they cannot find people will stupidly exclude a large part of the candidates most likely willing to make concessions or take lowball offers.
While at the same time expecting the “prime” candidates they will consider, who often have jobs or can obtain better offers (or are (still?) confident they can), to make concessions and then wondering they don’t get takers.
The methodology is supposed to eliminate ad duplication
I’m not a real fan of the structural skills mismatch phenomenon, because the data for young workers shows the same pattern.