Friday, April 24, 2015

Through Even These Good Times, Technology’s Effect on Jobs Continues

It’s been quiet.  The American work situation, except for conservatives emphasizing its bad aspects to carp about Obama and liberals trying to use its good side to push a minimum wage higher than latent demand for employment can justify, has just about dropped out of the press.  There haven’t been many reports of automation-related work losses.  However, we've been seeing technology-progress accounts that effectively predict them.

One was published Tuesday in The Wall Street Journal.  Titled “These Robots Serve Up Cocktails, but Can They Tell if You’ve Drunk Too Much,” it gives the status of something I predicted two years ago, in Choosing a Lasting Career, when I gave bartenders poorer prospects than other food service jobs, pointing out that automated drink mixing and serving would not only save labor but reduce legal liability, and that a “thoroughly automated bar” would be around long before the book’s 2033 horizon.  In the article, we learn about a variety of these cocktail robots (an already established name), and how they are being developed to be more interesting than just generic machines, with one a “rocket-shaped device with a large red start button that releases steam,” another with flashing lights, and one more with “plastic crystals and a fake flame.”  Giving them entertainment-machine bells and whistles may help with customer acceptance, a potential problem I noted before, and they are already in use for customer-facing applications on a Royal Caribbean cruise ship, along with, no doubt, large numbers of service bars.  The story makes clear that cocktail robots figure to proliferate, and replace many jobs.

A second article came out Thursday, in the Detroit Free Press.  It was a note on one area on which has strengthened since 2013, self-driving cars.  That year I gave taxi drivers and chauffeurs Very Good prospects for lasting 20 years, and said they had a “low automation threat.”  The story described discussion opening the Society of Automotive Engineers World Congress, in which a panel speaker, and prominent industry figure, talked about the need to reduce the one million-plus people dying annually, worldwide, due to car crashes caused by driver error.  The software for self-driving cars has been in development for over a year now, and although perfecting it has been challenging, another panelist said that laws, not technology, could ultimately hold back its use.  It is clear that driverless cars will be ready, and demonstrably an improvement over human-operated ones, before the decade is out, meaning that anyone who drives for a living may have his or her livelihood imperiled sometime in the next one.

Third, there was a story on Sunday’s New York Times, “The Machines Are Coming,” by Zeynep Tufekci.  It was not chosen for publication in the Times’s op-ed section by presenting the old issue of automation costing jobs, but for its reminder that work replaced by technology will not all be at or near the minimum-wage level.  Despite its future-leaning title, capabilities the article attributes to current machines include diagnosing medical test results, emotion detection, making hiring decisions, and providing accurate and spoken responses to people calling help desks.  There have already been many cases of what we would call good jobs being swept away by automation, especially in electronically-compatible areas of analysis that turn out to be completely algorithmic, and evidently there will be many more. 

These three articles show that further technology-caused job losses are, if not with clear timetables, on the way.  A major point Tufekci makes is that rising automation takes away workers’ power, in ways such as facilitating employers to “”optimize” worker schedules in a way that devastates ordinary lives.”  That brings us a fourth piece from the past week, a column by Robert Reich.  Its title, “America’s “flexible” economy is making worker’s lives hell,” published in Salon, nearly says it all, with its contents describing how software helps companies determine when employees are not needed – information some, such as Target and The Gap, now use to call off workers only minutes before their starting times.  The idea of what Reich said could be called “just-in-time scheduling,” “on-call staffing,” or “on-demand work” is nothing new – I experienced it, along with many of my scheduled work times being cut short, in a restaurant job 30 years ago, but the practice now seems more widespread, and fits with the ethics of the sharing economy in which income is irregular. 


What we can learn from these articles is not that automation will cost employment.  We knew that already.  It is that, despite relatively peaceful times on the number of jobs, technology, and our ways we are using it to that end, are still marching along.   

Wednesday, April 15, 2015

AJSN: As Unemployment Has Fallen, Latent Demand for Jobs from Others Has Risen

A week ago Friday, the most recent American Job Shortage Number, or AJSN, showed that almost 18.6 million more American positions could be quickly filled, if getting one were routine and easy.  I noted that, although the number of those officially unemployed dropped once again, the count of people with other statuses, other than jobless or working, was staying about even.  As a result, in March 2015, if jobs were truly plentiful, only 42% of them would be filled by people technically unemployed. 

How has that changed since the AJSN started in August 2012?  Here are a series of charts to show that. 

To start, this graph recaps the drop in official joblessness during the past 31 months.  It has been quite steep, as made clear by the trend line which evens out seasonal effects.  Note that the unemployment rates used here, as well as all other numbers in this post, are seasonally unadjusted:



The AJSN includes shares of 11 different employment statuses.  In the case of those officially jobless, it assumes about 90% would take readily available work.  Taking 90% of the actual number of unemployed, month-by-month, gets the following:



As above, the number of positions that could be filled by those technically jobless has fallen, irregularly but ultimately consistently, from over 11.8 million in January 2013 to 7.5 million in December 2014. 

Although the number of officially unemployed has fallen a great deal, that is not the case for the most significant other work statuses.  Between 600,000 and 1 million Americans, during this time period, have reported that they would like to have jobs, but think there is too little a chance of actually getting one to justify search efforts.  As per this chart, their number has also fallen, but not as much:
  



A similar and much larger category is those who do not report discouragement, but have put themselves out of the technically jobless group by not looking for at least a year.  Although unemployment as such has dropped, their numbers, and accordingly the AJSN share of 80% of them, has stayed about the same, as follows:
 


The largest employment set of people is those claiming no interest in working.  Many, especially those over 80 years old, are firmly fixed to that conviction, but some are not, as has been shown by the count of this group sometimes falling.  The AJSN’s conservative share of 5% of them taking readily available jobs shows an increase of about 275,000 more positions fillable by those in this group:
  


When we combine all the categories other than officially unemployed, the number who might work has gone up, per the trend line, almost 1 million:




Here is a pictorial view of how the shares of the AJSN from official joblessness and other statuses has changed:




That brings us to what has happened to the share of the AJSN from the above:



This trend line starts out at about 52% of American latent job demand as of August 2012, and ends up at about the 42% actually reached by last month’s data. 


It is easy to see what is happening here.  When we think about who would fill open jobs, we should know that those officially unemployed would take only a plurality of them, not a majority any more.  The need for new opportunities goes way further.  That trend is continuing, as the numbers of those with other statuses, despite more work being out there, continue to rise.  That is why we are much farther from full employment than many think, and, accordingly, why public policy should still be encouraging job creation.  Therefore, by inference, these charts have bad things to say about both higher interest rates and mandated higher wages, among other viewpoints.  The jobs crisis may have dropped off the headlines, but it is continuing – under the surface.    

Friday, April 10, 2015

3D Printing and its Jobs – Oversold, On Schedule, or Just Delayed?

Consider the following, and note they are in chronological order:

Three-dimensional printing makes it as cheap to create single items as it is to produce thousands and thus undermines economies of scale.  It may have as profound an impact on the world as the coming of the factory did...  Just as nobody could have predicted the impact of the steam engine in 1750—or the printing press in 1450, or the transistor in 1950—it is impossible to foresee the long-term impact of 3D printing.  But the technology is coming, and it is likely to disrupt every field it touches. – "Print me a Stradivarius – How a new manufacturing technology will change the world," Economist Technology, February 10, 2011.

The prediction that 3-D printers will become a part of our daily lives is happening much sooner than anyone anticipated. – Nick Bilton, “Disruptions:  On the Fast Track to Routine 3-D Printing,” The New York Times, February 17, 2013

Americans expect access to almost-free drinking water, clean air, a fine highway system, and at least some measure of personal safety – it may well come to pass that food and shelter, both possibly provided by 3D printers, and medical care, provided by software-driven robots, will be added to that list.  – I wrote this around September 2013

Are we expecting too much from these 3-D printers? – Title of article by Christopher Gregory, The New York Times, August 11, 2014

Items printed from a 3-D printer are unlikely to ever match the quality and strength and surface finish of mass manufactured goods… Printing is amazing… but do we print out our newspapers every morning on our inkjet printers? – Nick Allen, “3-D Printers are No Rival for Mass Production,” The New York Times, August 11, 2014

Gartner recently said that worldwide sales of 3-D printers will be about 217,000 units in 2015.  By contrast, in the United States alone, shipments of regular printers are typically about 24 million units a year. – Quentin Hardy, “HP Unveils Plan to Make 3-D Printing an Everyday Thing,” The New York Times, October 29, 2014

What’s happening here?

As the above sequence implies, once 3D printing began getting attention outside the technical community, it seemed that it would be so significant that it was almost scary.  Its machines could be producing just about everything, at home, in the office, and in what once were people-dense factories.  Not only price but speed would improve astronomically, and everything from hamburgers to skyscrapers, all with superb quality as well as unbeatable cost, would come out of the things.  And all of this would arrive stunningly quickly even by recent technology standards.  Although one estimate from last fall had 275 companies working on 3D printing as of four years ago, the 217,000 production could well have turned out to be close to that 24 million.

Now it is 2015, and, some real medical-related successes, a modest set of other niches, and a lot of gee-whiz stunts such as building a house notwithstanding, we’re still waiting.  Why, when 3D printers can now cost as little as $200?  I see four related reasons:

First, the printers now lack a killer app that would put them in ordinary people’s homes.  True, the novelty of making small things you can show off to people, and sometimes even use, is fun, but it isn’t enough to justify even the new lower prices, not to mention the cost of materials and supplies.  The closest they have is their physical prototype capability, which works of course superbly with CAD, but that’s not for the rank and file.

Second, large-scale use, as Allen wrote above, is not cost-effective.  Even if its expenditures were further slashed, the time 3D printing takes is still a huge problem for making large quantities of almost anything.  That and the first point mean that its major uses, for a while anyway, will be for small numbers of things, and otherwise around the fringes of routine manufacturing.

Third, while the technology is extremely promising, its real value may turn out to be, fast recent progress notwithstanding, far in the future.  Larry Niven’s science fiction stories set in 2800 or so referred to an organic soup of some sort, from which organs and other body structures were made.  That has already started, with replacement limbs, primitive though they may be by future standards, made that way.  One responder to the Gregory article above mentioned how much 3D printing will facilitate both the cost and the practice of space travel.  Those are two massive reasons to keep developing knowledge, and they provide reasons for companies to continue with it despite the return being maybe decades off, but neither offers widespread or extensive value now.

Fourth, and probably the most important to understand why expectations have lagged, is that when we were becoming aware of 3D printing we didn’t think much about exactly how it would be transformative.  I, for one, plead guilty to becoming intoxicated with the concept without projecting either what capability would arrive when, or what, after the novelty wore off, would get people to keep buying it. 


So how and when will 3D printing affect American employment?  It should contribute positions, mostly in design, maintenance, and research and development for the next 20 years or so, even with some of these and virtually all manufacturing being done in other countries.  After 2035 or so it will eliminate many, many more.   With that said, though, there is no new field, and certainly not anything else electronic, that projects to add more United States jobs over the next two decades.    

Friday, April 3, 2015

America Now Almost 18.6 Million Jobs Short as AJSN Drops Over 700,000

This morning’s employment data was expected to disappoint – and it did.  There were only 126,000 net new jobs added, the lowest since January 2014 and only enough to cover the monthly population increase.  The American Job Shortage Number, which gives how many additional new positions that could be quickly filled, improved greatly.  Why?

First, the AJSN is not seasonally adjusted, and, as virtually always, more people worked in March than in the previous February.  Second, a key set of people marginally attached to the labor force, those not working but not officially unemployed either – Americans wanting to work but not having looked for at least a year – dropped a surprising 420,000 to 3,320,000.  Not much else changed greatly, so the AJSN came out for March’s data as follows:


For other numbers, the official, seasonally adjusted, jobless rate held steady at 5.5%, while the unadjusted version fell from 5.8% in February to 5.6% in March.  The four key secondary indicators were mixed.  The number of people officially unemployed for 27 weeks or longer was down from 2.7 million to 2.6 million, and those working part-time for economic reasons, or wanting a full-time job and not finding one, rose from 6.6 million to 6.7 million.  The remaining two, the best measures of how common it actually is for Americans to work, the labor force participation rate and the employment to population ratio, were down 0.1% to 62.7% and unchanged at 59.3% respectively.  Average wages increased 7 cents per hour, for an annually compounded rate of about 4% – hardly a windfall, but still much more than inflation.

Compared with a year ago, the AJSN continued its improvement, though this time due only to a drop in official unemployment.  In March 2014, there were 1.85 million more technically jobless Americans, cutting the AJSN down almost 1.7 million, but several other statuses, especially expatriates, those wanting to work but not looking for at least a year, and those saying they did not want a job at all, went up enough to offset more than 500,000 of that.  Although fewer and fewer people indeed are contributing to those officially unemployed, many more who don’t qualify for that would take jobs if they thought they had a chance.


So what do we make of March?  As with the year-over-year comparison, the secondary numbers are not encouraging.  The gap between official unemployment and how many jobs could actually be absorbed continues to grow.  Of the almost 18.6 million new positions that could be filled quickly, only 42% would go to those technically jobless – the rest would be filled by people with other statuses, almost all of which, even in these solidly non-recession times, are steadily growing.  That is something to think about as we look at the turtle, who is now standing still.