If you have been following the news over the last few months, you may be aware that one of the most interesting developing stories is the significant shortage of workers, and more specifically, of low-skilled workers in areas such as restaurants, warehouses, and hotels.
So, last week’s article in the New York Times, "Pandemic Wave of Automation May Be Bad News for Workers”, may have come as a surprise:
“An increase in automation, especially in service industries, may prove to be an economic legacy of the pandemic. Businesses from factories to fast-food outlets to hotels turned to technology last year to keep operations running amid social distancing requirements and contagion fears. Now the outbreak is ebbing in the United States, but the difficulty in hiring workers — at least at the wages that employers are used to paying — is providing new momentum for automation.”
The article, as well as the statement, raise many questions regarding the implications for the future of employment and technology. The immediate one, however, is “how is it possible to have a shortage of employees if firms have been adding more automation?”
One might assume that this shortage is precisely what is driving the increased adoption of automation. However, the first observation I want to make is that adopting technology is not something that firms can do instantaneously. It requires significant capital, resources, time, and has considerable chances of disrupting business. So, while the adoption of automation may be a response to the scarcity and the increase in wages (or as some like to call it, inflation), what we see now is not a response to the current shortage.
In fact, in January, well before we knew the efficacy of the vaccines and as the pandemic raged, we saw a surge in orders for automation technologies, as documented by the Bloomberg article: “The Spread of Covid-19 Led to a Surge in Orders for Factory Robots.”
One also has to consider that this trend of adopting automation (and by automation, I mean anything that takes human activity and replaces it, or any part of it, with technology), has been ongoing. It's not that there was no automation until now and we are suddenly seeing robots popping up all over the place. Some have already identified this trend and named it the Fourth Industrial Revolution. A paper by Daron Acemoglu and Pascual Restrepo has shown that “between 50% and 70% of changes in the US wage structure over the last four decades are accounted for by the relative wage declines of worker groups specialized in routine tasks in industries experiencing rapid automation.”
Before we dive deeper into the reasons behind this new surge, you may be asking yourself if this is just a clickbait article from the New York Times, or if it’s real. The article reports the situation from the business owner’s point of view; the one adopting the robots. We see similar news from the realm of firms developing these technologies. Over the last few months, there has been a significant surge in investments in firms that develop automation and robots, e.g., for warehouses. Just this past week, The Information reported that Hai Robotics, a Chinese startup that develops warehouse robots, is raising more than $150 million at a valuation of $1.3 billion, almost triple its last valuation from March. So, this new surge is definitely real.
So, this brings me to the question of "why now”. Why has Covid increased the usage of automation?
When I hear about automation, I always think of my favorite book, The Goal by Eli Goldratt.
In one of the famous scenes from the book, the protagonist meets his professor and tells him, enthusiastically, about the robots they just adopted and how they help improve operations. The professor then asks three simple questions:
Did you sell more?
Did you fire anyone?
Did you reduce inventory?
When the professor receives a negative answer to all three questions, he tells the protagonist that he has not improved anything since he has not created more value for his stakeholders. Mere improvements to efficiency are not real improvements unless they translate to higher value: either by selling more (or selling at a higher price) or by reducing costs. Efficiency for the sake of efficiency is a mirage.
So, let’s see what the business owner in the New York Times article is saying:
“Ms. Gonzales, the Checkers franchisee, isn’t looking to cut jobs. She said she would hire 30 people if she could find them. And she has raised hourly pay to about $10 for entry-level workers, from about $9 before the pandemic. Technology, she said, is easing pressure on workers and speeding up service when restaurants are chronically understaffed.”
While the framework above may sound a little simplistic, it will serve well in order to analyze the quote.
Is “Selling More” really what’s going on? I can see how during the pandemic, the scarcity of employees may have forced firms to shut down, so it may well be that the adoption of automation allows the owner to stretch the hours of operations with limited staffing. However, I find it hard to believe that Ms. Gonzales is actually selling more due to automation. For that to happen, she must increase demand due to the increased automation, which in turn, should allow her to offer a better service.
It is a possibility if it indeed allows her to provide a better and faster service, but most likely both of these are driven by demand for “Checkers”, rather than this specific restaurant. In other words, I do not see “selling more” as a viable reason to adopt automation. At least not in the timeline we discussed.
Cost reduction? In the article, all the business owners quote that they are not planning to lay people off. However, it is clear that, as wages continue to rise, this will become more of a reason and firms will have significant incentives to adopt the technology. So, it’s possible that cost reduction is an important reason, but I don’t think it’s the main reason for the current surge. We are in the middle of a period of prosperity (the re-roaring twenties?), which is usually not the time when cost-cutting is the main focus.
Reducing Inventory Costs? We already celebrated the death of Just-in-Time. No one is discussing inventory anymore.
The New York Times article cites Katy George from McKinsey:
“For the first time, we’re seeing that these technologies are both increasing productivity, lowering cost, but they’re also increasing flexibility,” she said. “We’re starting to see real momentum building, which is great news for the world, frankly.”
While “lowering cost” and “increasing productivity” are points that have been highly debated and contested, the point about “increasing flexibility” is interesting. Maybe we need to add a fourth question to the framework from “The Goal”: Does automation allow you to be more flexible and handle variability and risk better?
Primarily, it is interesting because, traditionally, humans are considered to be more flexible and thus a better hedge against variability.
Robots (and automation tools in general) function better in a specified environment. When this is not the case and specified or predefined actions/processes are not available, humans are a better fit for the task. In other words, because we, as humans, are not built to be good at something specific, we have the ability to adapt and excel in anything. Apple has been trying to use robots for many years. The firm tried to automate the production line of the 12-inch MacBook, but the conveyor system had issues with erratic movement and the robot installing the keyboard "kept malfunctioning”, requiring workers to come to its rescue and ultimately delaying the final launch date by 6 months.
Similarly, most of the picking and packing at Amazon’s warehouses is done by people. It’s easier to add workers at peak demand periods and they are better at adjusting the grip between products with different levels of dexterity. Traditionally, humans have been considered more flexible.
It seems, however, that things are changing.
Why? Most of the new automation is driven by software and even robots that are hardware-based are driven, to a large extent, by more advanced software. While it’s hard to adjust the hardware, it’s much easier to keep software learning in the background through modern machine-learning methods. Waze becomes better at navigating over time. Amazon’s Alexa learns how to understand us better over time. The implication is that automation is beating people at being flexible.
So, one reason for the new surge in automation is that technology has advanced to the point where it is more flexible than humans, making the investment more palatable. In the past, you had automated tools for a specific purpose and calibrated everything for that purpose. This is no longer the case. These tools are now able to adjust and improve over time.
But how is all of this related to Covid? Is this just sensationalist writing?
Share Gad’s Newsletter
Over the last few weeks, I have written about how Covid is causing many people to question many aspects of their life. Apparently, it is not only Covid but pandemics in general.
A very interesting research paper shows that every pandemic that occurred over the last two decades caused firms to invest in automation.
“The results show that pandemics lead to an increase in robot adoption over time, with some lag. In the second year, we estimate that about 0.35 more new robots are installed per 1000 employees and 0.7 more new robots in four years after a pandemic event.”
This is extremely interesting since it shows that pandemics (as extremely disruptive events) cause firms to evaluate their operations and infuse more automation into them. The main implication is that firms use pandemics as a catalyst for soul searching, both in terms of their cost structure but also in terms of the risk and variability (and lack of flexibility) they have. Once you factor in risk and flexibility, these long-term decisions about robots and automation make much more sense, even if they don’t lower costs or offer the promised boost of productivity immediately.
Personally, the central learning is that maybe businesses actually think about risks more than we give them credit for.
But my second and more important concern is the long-term implications on the labor market:
“Automation in tandem with the Covid-19 recession is creating a ‘double-disruption’ scenario for workers. In addition to the current disruption from the pandemic-induced lockdowns and economic contraction, technological adoption by companies will transform tasks, jobs and skills by 2025. Forty-three percent of businesses surveyed indicate that they are set to reduce their workforce due to technology integration, 41% plan to expand their use of contractors for task-specialized work, and 34% plan to expand their workforce due to technology integration. By 2025, the time spent on current tasks at work by humans and machines will be equal. A significant share of companies also expect to make changes to locations, their value chains, and the size of their workforce due to factors beyond technology in the next five years.”
The surge in the adoption of automation is driven by the fact that robots are becoming more adaptable and flexible through machine learning. Since these are traits that have traditionally kept humans one step ahead, this should be a warning sign for most of us. There are only a few more things we are better at than robots:
Decision-making in ambiguous situations
Come to think of it...not that many.
Thanks Gad, nice discussion of this topic. Looking at last-mile delivery robots as an example, they were coming on the scene before the pandemic, either to allow retailers to add new services (same-day delivery) or to reduce delivery costs through automation. Cost per delivery seems to be the main variable to viability. But then pandemic-related concerns added another justification to using robots: social distancing and hygiene. Robots won’t breathe on you when delivering your package or pizza.
I do wonder about another question: robots may “replace” workers in some cases, but for now they seem to require an increase in human labor — robot handlers, chaperones, drone pilots, mechanics/engineers, etc. Every delivery robot I have seen seems to have an army of humans walking behind it. This is not a permanent feature, but full automation of robots seems to be a challenge that for now still requires a lot of human effort.