Innovating Towards Unemployment
In my day job I work as an analyst with a very large Fortune 500 company. The part of our business that I work for involves helping clients manage their supply chain. As part of that work we maintain nearly 8 million square feet of warehouse space here in Louisville. That space obviously requires staffing and we employ about 3,000 people in various warehouse positions. These kinds of jobs are fairly blue-collar, with an overwhelming majority of our employees having a high school education only. The pay is competitive but it is still a lower-middle class wage. The work they do does not require any special skill set in advance and they are all trained on the job. What this also means is that because of their lack of skills and their lack of education they are unfortunately among the most vulnerable members of America’s workforce.
About three years ago I visited one of our buildings where we were in the first year of a ten-year contract with a customer I will call Acme Cellphones. As the account was ramping up we were slammed with work. In this early stage many of our processes were very manual which required large pools of labor to perform. Not wanting to take on a large burden of permanent payroll, benefits, etc for employees we would not need in the future, we relied heavily on temporary labor. This is S.O.P. for us in all of our client implementation projects. We have to be conservative on hiring because as the project moves forward we get faster and more efficient at doing the job for the customer which means we will need fewer employees. A typical implementation takes about 18 months for a large customer and we will see as much as a 50% reduction in workforce from the highest-point of labor during implementation to the start of normal operations. At the time I visited this account they were at the high-point. The operation reminded me of a beehive with employees scurrying about everywhere. It was organized chaos. Of course this meant a higher failure rate for servicing the customer which is also very normal for the early stages of a project like this.
This customer has been somewhat unique in that chose to make a large investment in automation as part of giving us their business. We like these kinds of investments because they signal a longterm commitment but they also require hundreds of hours in planning and installation. This was all completed about a year ago and when I toured the facility again yesterday I was amazed at the change. There are now roller lines and conveyor belts traveling all over the building. Boxes of product zip around using automated systems to direct them to the right location. Efficiency has reached a high level and errors have reduced to one of the best performance percentages in our entire operation.
The temps are gone now. We can flex them back in when the next hot phone hits the market and demand soars for a temporary period, but right now things are steady. Our estimation is that the workforce has reduced by nearly 70% from three years prior. We are doing more with less and making fewer mistakes in doing it. Technology is giving the customer a better experience and helping us with our bottom line. But it also feels a little bit scary.
I am, generally speaking, a believer in the free market. I don’t like the ‘living wage’ movement and I was certainly no fan of the Occupy movement. But as a parent and someone who will need to remain employed for the next 30 years or so, I can’t help but fear for the day when innovation might someday make me obsolete. This isn’t a new problem though. John Henry was replaced by the steam hammer. They say the trick is to stay ahead of technology or a find a profession that can never be automated. This is the dynamic facing so many American workers today. When the economy dips companies figure out how to make do with less. When better times return they realize they can get by without those employees and innovation helps them get to that decision. It is the Western way but I’m not 100% sure it’s always the corect way. In our Taiwan facility the joke is that if there is a job we can do with one person in America, they will do it with three. They do less real work (and make less) but they take home a paycheck. Combine that with a few relatives’ income and you can afford an apartment and good food on the table. The American independent streak will not allow for that. Each man is an island and our pay must represent our status as king of that island. The problem is when there is a computer that can easily replace us. Then all bets are off.
“The trick is to stay ahead of technology or a find a profession that can never be automated.”
Are there enough of those? Will there still be 50 years from now? And what do we do if there aren’t (i.e. 8.5% unemployment (15% underemployment) becomes the new norm).Report
We all end up plumbers?Report
We always need more plumbersReport
Unless you’re already a plummer.Report
The original monkey wrenchers – the Luddites – were opposed to technological progress eliminating their jobs. I have a lot of sympathy for the view. There is no upper limit to how many jobs can be eliminated by mechanization, and the solace we have is believing that job creation from innovation and creativity can keep pace – well even that’s not enough, right? – with job loss so that employment conditions don’t fall through the floor.
I don’t see it happening.Report
There is a upper limit – All of them.Report
You need a bit more than just increasing automation for 8.5% unemployment to become normal. One of the following would also have to be true:
1. 8.5% of the population have a marginal product of zero. ie. There is nothing useful they can do for anyone else. Unlikely, in my view.
2. In some way or another, the cost to the person or their new employer is larger than the potential benefit of their having the job. Conservatives like to point out that minimum wages or unememployment benefits can have this effect, but search costs and very short lived opportunities also can.
3. There is some kind of long-lasting disequilibrium that means even though people can do useful work and other people are willing to employ them. there’s no way to structure the transaction so it actually happens.Report
see sticky unemployment on the econ blogs.Report
I’ve been studying the employment trends over the last 10000 years and one pattern has become clear. We always need hookers.
And, since the left believes that employment is pretty much the financial equivalent of being screwed by one’s employer, I think my argument is lock tight.Report
only if you’re working for Edison or Disney, capiche?Report
since the left believes that employment is pretty much the financial equivalent of being screwed by one’s employer
In just the same way employers believe any particular wage or salary is too high – some positive sum transactions are better than others! – so they’re getting screwed by their workers.
{{{If I can just get those over-privileged Americans to work for 90 cents an hour, I can save money on shipping costs!!!}}}Report
We always need hookers.
And congressmen. But I repeat myself.Report
Oh Roger. You’re always good for a laff. This Lefty believes employment is for dreamers and rubes and sissy boys who still entertain the fatuous notion whereby management and workers are working toward the same ends: getting paid.
Of course workers get screwed, especially salaried tree huggers, sleepwalking through their lives, thinkin’ their jobs are going to be there tomorrow and the next day. They work their little Dockers-clad asses off for chump change just so they can get some fucking health insurance for their kiddies and the premiums keep going up and their pediatricians haven’t been paid because they haven’t resubmitted the EDI at least four times and talked to some piggy-eyed temp named Latisha who’s working off a laminated page of canned responses and wouldn’t know one end of an EDI837 transmission from the other.
Now, perhaps you can convince me employment these days is anything but getting screwed. Let’s see you try.Report
You can’t outsource management. At least upper management. That’s why they get paid ludicrous salaries and bonuses, but not employies. Got to remain competitive with India, dontcha know.
Although in my experience, at least in software engineering, you don’t actually SAVE money offshoring jobs. It tends to cost more and take longer, but that comes during someone else’s tenure as upper management.
In the short term, look at the savings! Two quarters and out is a pox upon American business.Report
Outsourcing to India has been my personal salvation since about 1999. Doubled my billing rate. Started with GE Appliances receiving six fucked-up Java Servlet-based projects. GE set me and two other guys to re-interviewing every consultant in the place, about 200 of them. About a dozen survived the process.
If it wasn’t for TCS and Infosys and Wipro fucking up everything, I swear I’d be back to making an honest buck.Report
Outsourced stuff…*ugh*. You can’t outsource code like that unless you outsource the team requiring it (so they can do back-and-forth), OR write ridiculously rigid design docs in a Waterfall style (which most companies can’t do at all, much less do right).
Which is expensive and hard to do correctly.
And frankly, I’ve never been impressed with the end results of the work, even when the prep is done properly. (Which, as I said, it never ends. Companies outsource to save money quickly. Spending a ton of money doing front-end work doesn’t pretty up the bottom line. Instead it’s all done at the end, for several times the cost, and the final result is shit).
I’ve seen Waterfall done right. I’ve seen big projects that are widely seperated from end-users work. But the stuff shipped overseas always seems to end up generic and — at best — just an approximate solution that’s better than nothing.
At worst? It’s an utter shit fest that requires expensive talent brought in to fix.Report
These firms are run by idiots. The fish rots from the head. If these systems were designed by the people who used them, they’d be simple and obvious and extensible.
But since these systems are controlled by a collection of empty-headed frat boys with their heads stuffed full of MBA horse shit, all glad-handing each other and giving each other big raises, we shouldn’t be surprised to see them treat enterprise software development exactly like they treat their employees.
Really, I have nothing but contempt for most of the people who pay me to fix this shit. They’re ruining this country. I just can’t seem to get away from it, this fixing broken shit. Just picked up a gig this week, 90% of the tables go unused and none of them are normalised and there are no indexes. The front end is implemented in the worst PHP you ever saw. Username and password hanging out in the open, naked as a baboon’s ass.
When I read stuff like I don’t like the ‘living wage’ movement and I was certainly no fan of the Occupy movement. But as a parent and someone who will need to remain employed for the next 30 years or so, I can’t help but fear for the day when innovation might someday make me obsolete. I just secretly laugh. The reason he’s afraid in the second sentence is because he sorta understands the consequences of the first sentence. Oh, Mike may not like the idea of a living wage. He may think the Occupy folks are full of shit. But that’s okay. People like me are the logical outcome of that sort of thinking.
The profits which flow to the top (and certainly not into his pocket) will shortly end up in my pocket or someone exactly like me. Some talk around here about loan sharking and high interest rates. People pay me the wages of sin. I don’t come cheap. I punish the stupid where it hurts and they’re glad to pay me to put out the fires they started.Report
Damn, that’s worse than what I’m used to dealing with. I mean, sure, I was dealing with tables written by an aerospace engineer turned coder turned web-developer for internal tools (first normal form, unneccessary unique ID’s — basically what a coder used to spreadsheets would make. Easy for him to use. Luckily space wasn’t a concern, but I tightened it up a bit and changed a few things that wouldn’t impact already developed work too much but would make it easier going forward…).
There is a problemn with designing tightly to end users — you can’t grow your end user market if you’re not careful. The current POS I’m slowly cleaning (every change I make reveals a half-dozen bugs I have to stop and fix) is basically written around a sole user’s particular work path.
He’s an important user, half the calculation end of the software was based on his math (and some of it wrtten by him in fortran) and he’s been the biggest user for a decade AND he uses it to produce data that is basically the backbone of the rest of the system.
But nobody else uses that particular set of GUI’s because it’s so wrapped around HIM and his needs. Anything he doesn’t use is buggy or broken, output is often formatted specifically for his needs (god forbid he read it in via CSV file and use an a macro) which changes….
Doesn’t help that the previous developer threw up his hands and rewrote it in a bullshit “My first OOP” way over what couldn’t have been more than a week or three. Obvious cut and paste jobs, completely broken OO design, and the data and GUI seperate is non-existant.
So trying to make things work I’m reduced to harassing the GUI to give me back data that I fed it early (it’s out of scope), it’s easy as hell to corrupt, and you don’t even want to know about some of the black box violations it’s got.
Like passing the pointer to the main class to other class methods as a backdoor hack of global variables. I can’t sell them on a “Let me nuke it from orbit” approach so I’m reduced to teasing out the data, locking it off from the GUI code (because right now they’re so tightly dependent on each other that you can trigger bugs by moving backwards in the GUI or altering data out of order when it shouldn’t MATTER until you execute something) so that the GUI and data are, you know, seperate.
GUI should display data, manipulate data, examine data — but not BE the data in the weird-ass way that’s there.
I’m not a consultant though. 🙂 I do have like 2 years worth of work on just two GUI’s to bring them up to the professional quality of the rest of the suite, and that’s not including the upgrades they want to add.
I looked at a more contract-oriented approach — what sounds like what you’re doing — but I don’t like being without health insurance (my wife’s is crap), and I’m not a huge fan of interviewing.
I prefer lengthier jobs and a feeling I’ve done more than clean up some mess or solve a quick problem.
Plus, gives me time to futz around with side projects. 🙂 I really should publish some of my genetic programming stuff, but honestly…I wanted to see what I could do. Telling other people about it is work. 🙂 All that writing and editing and pulling out data….ugh. I just want to play.Report
MVC can solve a lot of these problems. I had a gig once, US Dept of Agriculture. Two main offices, one in DC and the other in St Louis. Each thought the other office was staffed exclusively by madmen and incompetents.
Solution: design two separate Views operating against a common Controller against the same Model.Report
Simon is right, automation doesn’t cause unemployment at a macro level.Report
There’s always work to do. The question is whether there is work enough to command or justify the salary required to deem what a worker and society believes is an acceptable standard of living. And, of course, whether they’re willing to do it.Report
Not on its own, no. But a deficiency of retraining resources can lead to displacements of labor that are hard to fix.Report
I’m glad we settled that.Report
Simon is right, automation doesn’t cause unemployment at a macro level.
How about underemployment? Or employment at below subsistence levels? I mean, I get the theory: automation entails job losses in company or sector X and the displaced workers find employment in sector Y if there is productive value in hiring them. That is, there is some wage such that some person will hire them to do something useful.
Sure. But what is the wage relative to the cost of living? Sure, it logically follows that someone could hire a US worker to wash windows at 50 cents an hour, but there are legal, cultural, and practical reasons why this doesn’t happen. The only one you have a solution for is eliminating the legal hurdle (by eliminating the minimum wage). But really, if the theory works so well, why aren’t all the unemployed finding the best deal they can under the table? (And maybe they already are, right?) I submit that it’s because of the two things conveniently left out of the equation: cultural influences and the practical effects of taking a wage lower than a person wants to, independently of the reasons he might have for doing so.Report
And to amplify on that a bit: culture is sticky. It’s pretty rigid. People have an expectation of how things should go for them and the situations they can tolerably find themselves working in. This is part of a broader concept of human dignity, I suppose, where self-identification as a type of person are built into the cultural fabric of our societies. Letting go of those concepts and identification markers is a big constraint on the choices a person could make, or is willing to make.
So, supposing that a tradesman lost his job to a machine and he’s unwilling to take a job washing windows for 50 cents an hour because his self-concept as a productive tradesman won’t permit it, it is incoherent – it seems to me – to say that automation didin’t cause his unemployment.
And to say that he needs to get over it (his conception of who he is???) both misses the point, but leads to unfortunate FYIGM type responses from those who disagree.Report
This is quite possible, Stillwater. If I grew up expecting to be an auto-worker in Detroit in the 1970s and lost my job, there’s probably no way, now, that I could recover my expected relative standard of living (although my absolute standard of living may be better, it probably doesn’t feel that way because so is everyone else’s). In theory, nothing prevents people having a marginal product less than the income needed to survive, although I think in practice is quite unlikely.
It would be helpful, when discussing these kinds of macro issues if we abandoned the idea of things having single causes. Your hypothetical tradesman’s unemployment was caused by the machine, his unwillingness to take a new job, inadequate retraining programs, and (probably) insufficiently expansive monetary policy. In some ways the machine is the least interesting part of the explanation …
This is a very different thing from saying “8.5% unemployement is the new normal”. At the macro level that requires something in addition to automation to explain it. In your example, unwillingness to take a lower wage job is an example of wage stickiness – a cause of disequilibrium. Obviously that’s totally unhelpful to those people actually in that position. I don’t know what you do about that.Report
Your hypothetical tradesman’s unemployment was caused by the machine, his unwillingness to take a new job, inadequate retraining programs, and (probably) insufficiently expansive monetary policy.
The direct cause of his unemployment was replacement by a machine. On that we agree. But how are the other things causes? By hypothesis, he’s not unwilling to work, he’s unwilling to take a job he finds beneath him. Inadequate job training isn’t a cause of his remaining unemployed, it seems to me, even if retraining may be causally sufficient for gaining future employment as long as some other conditions are met. And if austerity-mania is driving our monetary policy, then 8.5% unemployment may in fact be new normal, at least for a while.
I don’t mean to harp on these issues, because to you they make perfect sense and all my errors are crystal clear. But supposing that monetary policy doesn’t change for the foreseeable future, all the stuff about wage stickiness as a cause of disequilibrium comes into play. That is, if wages are sticky – and I would say that in a closed labor pool they will be – why isn’t it accurate to say the resulting disequilibrium is caused by our monetary policy? Or by mechanization itself? Or capital flight? Or etc.?
Surely some of those variables mentioned can be justified on utilitarian grounds, but wage stickiness can as well, it seems to me.Report
Here’s another way to say it: it seems to me your views of macro economics are overly reliant on a concept of wage flexibility in the labor pool, one which may not be the case, and which I’m suggesting ought not be the case.Report
From a practical standpoint, there are limits to the extent to which we can expect people to take jobs out of their skillset that net barely more than not working. It may be the case that we shouldn’t have an expected standard of living above and beyond basic food and shelter, but we do.
(In case there is any confusion, that was me not disagreeing with you.)
My question remains, what are we going to do with the people whose new economic value is below the standard of living we and they are willing to accept simply because their line of work and potential lines of work have been automated?
This is one of the walls I keep banging my head against. As someone that does not believe that knowledge-work and work that can’t be outsourced are universally obtainable, and who is skeptical of continuous lateral retraining (especially when training appears to be increasingly on the employee’s dime), and who is sympathetic to a reluctance to downgrade to service class wage… I haven’t the faintest clue what we do.Report
Nor do I. Maybe we can nudge Simon and James K into figuring something out for us. I still believe two things about it tho. The first is that economic forces aren’t inexorable, or beyond human control. Humans create those forces, so humans can change them. The second is that there has to be a better model for political economy than the assumption that humans are nothing more than points of exchange defined by their preferences. For one reason, a preference is made apparent only when given options and a choice. This leads to problems. For example, suppose you get layed off by a machine (negative utility but no choice) and are presented the option of doing shitty work for half the previous pay (positive utility with a choice). According to preference utility based on voluntary choices, you’re way ahead: every choice you’ve made is positive sum! But your objective utility is much lower. Is there an upper limit beyond which this phenomenon couldn’t happen? The only constraint is imagination and implementation at the right price, tho at some point labor might become so cheap and abundant that mechanization itself isn’t profitable. (That’s a nice picture, no?)
Another is that culturally determined expectations will continue to define preferences even as we slide down the rabbit hole. That means culturally determined self-identification markets will create labor stickiness even as the market demands increasingly flexible labor markets. Unions are a political loser at this point, and even if they have pragmatic utility in slowing or even reversing the downward slide, the very people who we’d need to champion them – economists – are generally (not in toto) opposed to them because they create the very thing they’re intended to create: labor rigidity. In fact any potential solution appears to be rejected on the grounds of increasing rigidity. But the only solution to the problem, at least as we see the problem, is to stiffen up a few of the markets.
So yeah, what to do?
(And one other thing. I heard a liberal say “Democrat politician” the other day. I owe you a beer.)Report
The first is that economic forces aren’t inexorable, or beyond human control.
I would probably agree with this, though it depends on what you mean by “economic forces.” We can always stand in the way of automation and globalization, though I think the harms outweigh the benefits on the whole. The more primary question to me is what we can do with the proceeds of these advancements. And what we do with the regard to the displaced.
The second is that there has to be a better model for political economy than the assumption that humans are nothing more than points of exchange defined by their preferences.
I agree that the “preferences” part of this is misplaced. We make decisions with limited information and – as you point out – a lot of cultural influences.
But the only solution to the problem, at least as we see the problem, is to stiffen up a few of the markets.
And here we have to ask what the costs of stiffing up those markets are. Not just the costs to the CEO’s and corporations.
(And one other thing. I heard a liberal say “Democrat politician” the other day. I owe you a beer.)
Tsall good. We agreed on what it *mostly* meant, just disagreed on the totality of it. Oh, and the grammar. Let’s not get into the grammar! 🙂Report
Any one interested (or concerned) about what all this means for the future should read the book “The Lights in the Tunnel: Automation, Accelerating Technology and the Economy of the Future” …. it’s all about this specific issue.
You can find it on Amazon (Kindle version is only $4) or a free pdf is at http://www.thelightsinthetunnel.comReport
Thanks for posting that. I’ll check it out.Report
THE WEST WING
5X19 – “TALKING POINTS”
JOSH
You knew we were for free trade. You knew it when you endorsed us five
years ago.
PARSONS
Yeah, ’cause you told us we might lose old economy jobs – shoe manufacturing
– to some dirt-poor country, but if we trained ourselves we’d get better
jobs. Now they’re being vacuumed out of here, too.
JOSH
We’re going to fight for more job training, more transition assistance…
PARSONS
I have members on their third and fourth career. What are they supposed to
train for now, nuclear physics? Cello playing? Or should they just give up
and bag groceries for minimum wage?
Report
I haven’t collected a regular pay check in almost 30 years. I came out Army and university, drove a cab then did some consulting, simply because I was going to school.
Got married, decided I’d climb the ol’ Corporate Ladder, got a full time job writing software. Decided to buy a house on the strength of that job. The two guys who owned the company went on a fishing trip, got sold a bill of goods, came back and fired half the programmers.
I was two weeks from closing on a house.
My lawyer and parents told me to stay quiet about it, close on the house, get another job. I did. But I’ve never trusted an employer since.
I’m there to do a job. I’ll charge by the hour and I’ll charge you more than three times the going rate for a full time employee for the privilege of you telling me not to come in tomorrow or me taking a day off because I feel like it. I don’t want to waste time at your idiotic staff meetings. I’ll bill you for every minute I spend on the premises of your godforsaken little installation. I come cheaper if I don’t have fly in and rent a car and get a hotel so I can sit in your conference room and ssh into some box in Pune India to deal with your incompetent offshore talent and with the morons who hired them. If you lie to me or try to fuck me over, I will hand in my badge and leave and if you do not pay my invoice, my lawyer will tear you a new one.
I do not want to participate in the Reindeer Games of corporate politics. I do not enjoy the company of the folks in Carpet Land. The only people I like are the blue collar people who do the work and they walk on sealed concrete.
Newsflash for all those folks in Carpetland: all those blue collar people you hire when the need arises and dispose of like so much used toilet paper when the job is done? You’re next. Your jobs were never safe. The morons in the corporate suite do not love you.Report
It’s pretty funny to hear this in the voice of Michael Keaton in “Mister Mom” talking to the guys on the assembly line.Report
I would have and will read your thoughts on these questions to triple the length of what you had to say here and beyond, Mike. This kind of ground-level reckoning with how ideas work out in practice is highly valuable, and I appreciate the time you take to put it down.Report
They say the trick is to stay ahead of technology or a find a profession that can never be automated.
Nope, it ain’t gonna be. Every last one of us can be replaced by something, the question is only at what cost? None of us can say what future innovation will obsolete.
Excellent post, Mike. I had a lot of similar thoughts rolling around in my head along a lot of these lines throughout the Inequality Symposium. There are no easy answers.Report
Don’t forget, if you try that then it’s your own fault when you graduate college or training to find a glut of people who ALSO did that out there.
On the other hand, if you’re friends with a few well-connected sorts, a job is always just a phone call away.
Which is the real lesson: The one indispensible job skill for continuous employment — knowing some rich guys that own companies. Or run them.Report
Nah. The solution is to hyper-specialise and develop friendships with fellow travellers who know more than you do about specific areas of expertise. Tell you about my little gang.
There’s this friend of mine from SoCal, just the most laid-back, friendly sorta guy, perfect for doing face to face stuff with the client, managing expectations, always laughing and smiling, puts everyone at ease. Project Manager.
Then there’s the Message Guy, intense, wiry, the most productive interface guy I’ve ever known. He can sit there for weeks, give him a message to compose, he’ll wind up to several thousand RPMs in a few milliseconds and he’s got the work done. Does a few minutes of coding every week. Also does porn. Drives a Harley.
There’s Rev. Object. Devoted Christian guy, suffers from ulcers, can automagically go through metadata and compose Controllers. His code generates code which generates code. Deeply committed to excellence. Will tell the truth when it hurts. Eats at Denny’s every day when he’s on site.
Then there’s Dr. Data. Teradata Master. Drives around in a big motor home with his wife and dog. Drinks Gentleman Jack.
And Rules Guy. Drives a cherry red Corvette. Biggest bullshitter in the history of the industry. Must be kept away from clients. But he can instinctively build efficient rules engines and pricing trees like the Rain Man can count cards.
Microsoft Dude. Hygiene issues. Knows more about Microsoft than is good for anyone. Must also be kept clear of clients. Independently wealthy, hugely unreliable. Will only work on things which interest him. But if they do, he will work nonstop, days on end, to solve a Windows internals problem.
Me, I’m the Plumber. I connect things.
Clients come, clients go. The one indispensable job skill in this world of sin and error is knowing who’s good at what they do.Report
The question, though, is whether the innovation is making everything so cheap that being a minimum-wage grocery bagger is enough.
“Oh, but that’s not a fulfilling career!” Right, like shelf-stacking and box-filling was?
The fact is, that civilisation requires slaves. The Greeks were quite right there. Unless there are slaves to do the ugly, horrible, uninteresting work, culture and contemplation become almost impossible. Human slavery is wrong, insecure, and demoralizing. On mechanical slavery, on the slavery of the machine, the future of the world depends. — Oscar WildeReport
Technology has always been a part of the equation. The wheel started it all. The scary part is that today people think they are done with education at High School or college. Alzheimer’s is on the rise because people quit challenging their brains. Evolving is done by learning. Do it for your job, you stay ahead of your fellow workers and make yourself more valuable. Do it for life and be a more interesting person.Report