The High Cost of Automating Jobs
Donald Trump’s new Treasury Secretary, Steve Mnuchin, has impressed me this week with his grasp of the economic reality of the manufacturing world. He’s right--technology won’t be a huge factor in overall job levels for many decades. If ever.
Wild-Eyed Fantasies
If you could go back in time a hundred years, I wonder if there were sensationalist headlines about the future with new technology like automobiles and airplanes. Journalists could have written stories about how in the future any trip over 50 miles would be easy because everyone would just fly there in their personal aircraft, or children wouldn’t have to walk anywhere because their parents would all buy them their own automobiles.
Stories like these would have been confusing what was possible with what was economically feasible. Would it be possible for every family to own an airplane? Yes, but for the few times we’d actually use it, would it makes sense? Airplanes are too expensive for most families, and there is already investment sunk in the family’s autos. You don’t just buy an airplane. You also have to pay to maintain and store it. It’s cheaper just to get in your car and do your day drive. This dynamic holds true for corporations as well.
The same thing is happening en masse today. Journalists are getting a whiff of what technology is making possible, but they aren’t tempering their projections with sensible economic principles. There is no such thing as a free lunch.
A few years ago, I was invited to speak at Oracle’s annual OpenWorld convention by Anil Thomas who was a VP and founding member of MetaChain Consulting (now Avata). Anil and I had done some great work around “cluster cost analysis” which was a way of breaking down a supply chain’s true freight costs down to the product level. When a trucking fleet has a variable schedule that fluctuates often, you need to identify the cost of your city “clusters”. This allows you to break down freight rates to your product unit level.
I never ended up participating in OpenWorld because my employer considered this information too valuable to share with competitors. Most companies that operate their own trucking fleet don’t really understand their true delivery costs at the product level. So if they don’t get it, I’m sure academics in a think tank or journalists sitting behind a laptop don’t either.
The LA Times completely distorted the PWC study.
The Los Angeles Times postulated that “Robots could take over 38% of U.S. jobs within about fifteen years.” Emphasis on “could”. That’s like saying that your iPhone could theoretically get up to 400Mbps of speed on LTE throughput. Not gonna happen in the real world.
My background in logistics analysis and over twenty years of manufacturing analysis on automating work makes me find click-bait headlines like these by the Los Angeles Times quite humorous. A more accurate headline would have been something along the lines of 3-5%, and that is assuming that the law of unintended consequences doesn’t kick in. I actually liked the PWC study. They gave credence to the reality that automation has an opposing downward force exerted by the high cost. I was happy to see their assertion that the economic justification of automation was a very real hurdle.
PWC was giving an upper range estimate on the potential job losses if everything went the wrong way. Kind of like if someone asks you what you think it would cost to build a house in a way that you’ve never done before. You give a wild range to cover yourself. That wasn’t meant to be a precise forecast.
Think about this rationally for a minute. 15 years ago was 2002. PWC would never try to say that in less than the time since the 9-11 terror attacks that 40% of Americans will lose their jobs to artificial intelligence. We still have companies running Windows XP for crying out loud.
What About Trucking?
Even if you take trucking, the big headliner profession that all the doomsday Chicken Littles love to bring up, it’s not nearly as bad as people think. That’s because most of the cost allocated to the product has nothing to do with the driver. The cost of capital and fuel are the big-ticket items.
The real boon of self-driving trucks is the fact that it is hard to find drivers willing to do cross-country routes. People don’t like those long routes where they are away from home for a week at a time. If you’re not familiar with the logistics business, know this one thing: there is an awful driver shortage that hurts everyone.
But only very large corporations with lots of volume are going to be able to make the necessary return on investment for multi-million dollar automated fleets. That’s because their capital costs would go up significantly and they’ll need to spread those costs over greater volume. As a result of these greater costs, they will give up delivering product lines that don’t cross the volume threshold they need.
The fact that smaller companies will pass on new technology is not a novel new idea. Even today, there is all manner of currently available software packages or hardware devices that could streamline their operations that don’t get purchased. Why? Because it doesn’t make financial sense given their volumes.
The large companies that do take the plunge will ultimately focus their product portfolio. The product lines that are shed would fall down to the non-automated companies. These in turn would need to hire more drivers. But since the big guys are automating, they might actually be able to find the drivers that they need.
The Sky Isn’t Falling
The two most important facts to remember when critically thinking about the impact of technology on jobs are these:
As I’ve written before, jobs will go away at large corporations. But the unintended consequence of this movement is that large corporations will by necessity be forced to narrow their product offerings. This creates more opportunity that falls down to medium to small businesses.
I can say that in my twenty-five years of manufacturing analysis I have never seen a manufacturing plant shrink due to automation. That’s because automation has always allowed us to do more with the same number of people. Automation did reduce our costs and make us more competitive on price, which led to more business and often resulted in hiring more people.
The real downside to automation is to the corporations themselves. As corporations focus their product portfolios and move up in volume, they find that the market may not be able to support all the companies that are expanding their production capacity. There will be winners and losers. The winners will take the business from the losers. Jobs will be lost by the losers but gained by the winners for a net impact that is relatively small.
There are far too many people writing about something that they don’t understand, manufacturing. The problem isn’t that a finance guy like Steve Mnuchin misunderstands technology. The problem is that a bunch of journalists and technology guys misunderstand economics.
Related:
Elon Musk Is Wrong About Factory Automation
Is Automation to Blame for Jobless Recoveries
The Impact of Self-Driving Vehicles Is Overstated
If you could go back in time a hundred years, I wonder if there were sensationalist headlines about the future with new technology like automobiles and airplanes. Journalists could have written stories about how in the future any trip over 50 miles would be easy because everyone would just fly there in their personal aircraft, or children wouldn’t have to walk anywhere because their parents would all buy them their own automobiles.
Stories like these would have been confusing what was possible with what was economically feasible. Would it be possible for every family to own an airplane? Yes, but for the few times we’d actually use it, would it makes sense? Airplanes are too expensive for most families, and there is already investment sunk in the family’s autos. You don’t just buy an airplane. You also have to pay to maintain and store it. It’s cheaper just to get in your car and do your day drive. This dynamic holds true for corporations as well.
The same thing is happening en masse today. Journalists are getting a whiff of what technology is making possible, but they aren’t tempering their projections with sensible economic principles. There is no such thing as a free lunch.
A few years ago, I was invited to speak at Oracle’s annual OpenWorld convention by Anil Thomas who was a VP and founding member of MetaChain Consulting (now Avata). Anil and I had done some great work around “cluster cost analysis” which was a way of breaking down a supply chain’s true freight costs down to the product level. When a trucking fleet has a variable schedule that fluctuates often, you need to identify the cost of your city “clusters”. This allows you to break down freight rates to your product unit level.
I never ended up participating in OpenWorld because my employer considered this information too valuable to share with competitors. Most companies that operate their own trucking fleet don’t really understand their true delivery costs at the product level. So if they don’t get it, I’m sure academics in a think tank or journalists sitting behind a laptop don’t either.
The LA Times completely distorted the PWC study.
The Los Angeles Times postulated that “Robots could take over 38% of U.S. jobs within about fifteen years.” Emphasis on “could”. That’s like saying that your iPhone could theoretically get up to 400Mbps of speed on LTE throughput. Not gonna happen in the real world.
My background in logistics analysis and over twenty years of manufacturing analysis on automating work makes me find click-bait headlines like these by the Los Angeles Times quite humorous. A more accurate headline would have been something along the lines of 3-5%, and that is assuming that the law of unintended consequences doesn’t kick in. I actually liked the PWC study. They gave credence to the reality that automation has an opposing downward force exerted by the high cost. I was happy to see their assertion that the economic justification of automation was a very real hurdle.
PWC was giving an upper range estimate on the potential job losses if everything went the wrong way. Kind of like if someone asks you what you think it would cost to build a house in a way that you’ve never done before. You give a wild range to cover yourself. That wasn’t meant to be a precise forecast.
Think about this rationally for a minute. 15 years ago was 2002. PWC would never try to say that in less than the time since the 9-11 terror attacks that 40% of Americans will lose their jobs to artificial intelligence. We still have companies running Windows XP for crying out loud.
What About Trucking?
Even if you take trucking, the big headliner profession that all the doomsday Chicken Littles love to bring up, it’s not nearly as bad as people think. That’s because most of the cost allocated to the product has nothing to do with the driver. The cost of capital and fuel are the big-ticket items.
The real boon of self-driving trucks is the fact that it is hard to find drivers willing to do cross-country routes. People don’t like those long routes where they are away from home for a week at a time. If you’re not familiar with the logistics business, know this one thing: there is an awful driver shortage that hurts everyone.
But only very large corporations with lots of volume are going to be able to make the necessary return on investment for multi-million dollar automated fleets. That’s because their capital costs would go up significantly and they’ll need to spread those costs over greater volume. As a result of these greater costs, they will give up delivering product lines that don’t cross the volume threshold they need.
The fact that smaller companies will pass on new technology is not a novel new idea. Even today, there is all manner of currently available software packages or hardware devices that could streamline their operations that don’t get purchased. Why? Because it doesn’t make financial sense given their volumes.
The large companies that do take the plunge will ultimately focus their product portfolio. The product lines that are shed would fall down to the non-automated companies. These in turn would need to hire more drivers. But since the big guys are automating, they might actually be able to find the drivers that they need.
The Sky Isn’t Falling
The two most important facts to remember when critically thinking about the impact of technology on jobs are these:
- Automated systems are extremely expensive and very high maintenance.
- Most jobs are provided by small- to medium-sized companies. The least likely purchasers of automation.
As I’ve written before, jobs will go away at large corporations. But the unintended consequence of this movement is that large corporations will by necessity be forced to narrow their product offerings. This creates more opportunity that falls down to medium to small businesses.
I can say that in my twenty-five years of manufacturing analysis I have never seen a manufacturing plant shrink due to automation. That’s because automation has always allowed us to do more with the same number of people. Automation did reduce our costs and make us more competitive on price, which led to more business and often resulted in hiring more people.
The real downside to automation is to the corporations themselves. As corporations focus their product portfolios and move up in volume, they find that the market may not be able to support all the companies that are expanding their production capacity. There will be winners and losers. The winners will take the business from the losers. Jobs will be lost by the losers but gained by the winners for a net impact that is relatively small.
There are far too many people writing about something that they don’t understand, manufacturing. The problem isn’t that a finance guy like Steve Mnuchin misunderstands technology. The problem is that a bunch of journalists and technology guys misunderstand economics.
Related:
Elon Musk Is Wrong About Factory Automation
Is Automation to Blame for Jobless Recoveries
The Impact of Self-Driving Vehicles Is Overstated
RSS Feed