By Carl Benedikt Frey, Director of the Oxford University Future of Work program
Remote work isn’t a new phenomenon. Before the Industrial Revolution, which took off in England around 1750, the vast majority of the population worked from home. The domestic system, whereby workers make products in their own homes, was predominant in Europe and elsewhere. The typical artisan lived in a cottage with often only one room, which served as both home and workshop. Much like a gig worker today, he decided for himself when a day’s work began and when it ended. In this light, the repulsion many felt toward the factory system is easier to understand. As historian David Landes puts it, the mechanized factory, which gradually replaced the domestic system, “required and eventually created a new breed of worker, broken to the inexorable demands of the clock.”
Before the COVID-19 pandemic, the factory model of working, which the Industrial Revolution created, was still largely intact. Despite striking advances in communications technology, telework remained rare. Across the European Union, the share of the population working from home hovered between 4% and 6% throughout the 2000s and 2010s. And in the US, a recent survey by the Bureau of Labor Statistics found that only 15% of respondents ever had a full day working from home. More striking still, only 2% of employees reported ever having worked from home full time.
Over the past year, however, COVID has forced millions of workers to set up home offices. Right now, working from home accounts for a staggering more than 60% of economic activity in the US. Thus, it is hardly an exaggeration to say that we are currently undergoing the largest work experiment in history.
While the switch to remote work has been enormously disruptive for many businesses, it is likely to boost productivity growth over the medium term. As an analogy, consider the strike on the London Underground in February 2014. As many lines were closed, Londoners were forced to rethink their commutes to work. Such disruption meant that many came in late to the office, but it also brought unexpected efficiency gains over the following years.
The economists Shaun Larcom, Ferdinand Rauch and Tim Willems, found that while only 5% of commuters stuck to their new route even after the strike, the benefits from that change were long lasting, exceeding the costs produced by the strike, which was a one-off event. In similar fashion, the pandemic has forced businesses to rethink their routines and work processes, albeit on much greater scale. Virtually all large companies that can, have switched to remote work.
A key question is to what extent will we work remotely when the pandemic subsides. Indeed, it could be argued that we have seen this movie before. With the arrival of the Web in the 1990s, many thought that remote work would become the new normal. Writing in 1997, economist Frances Cairncross argued that, “In half a century’s time, it may well seem extraordinary that millions of people once trooped from one building (their home) to another (their office) each morning, only to reverse the procedure each evening ... Commuting wastes time and building capacity. One building … the home … often stands empty all day; another … the office … usually stands empty all night. All this may strike our grandchildren as bizarre.”
Others were more skeptical. In 1998, economist Edward Glaeser argued that if we want to understand the future of cities and offices, we need to understand the forces of agglomeration, and how they interact with changes in technology. Glaeser provided two key reasons for why we won’t see the end of offices and cities: digital technologies, he argued, provide poor substitutes for face-to-face meetings and sporadic interactions. However, digital technologies have made a great leap forward since Glaeser published his article. They have become much better substitutes for in-person meetings. What digital technologies are still unable to do, however, is substitute for the “watercooler moments” at the office.
There are good reasons to think that remote work will be the new normal for many. In addition to saving real estate costs and commuting time, many studies have demonstrated that people working from home are more productive. For example, one recent experiment conducted by Stanford’s Nicholas Bloom and collaborators found that remote work increased performance by 13% due to fewer breaks and sick days, and a quieter work environment. The dilemma facing businesses around the world is that while remote work can bring tangible efficiency gains, it makes innovation less likely to happen, which is what drives productivity and business performance over the long run. The most creative ideas aren’t going to come when people work productively in front of their monitors at home. The drive to improve efficiency, in other words, could imperil innovation, which is fundamentally about exploration.
This trade-off is well known to artificial intelligence (AI) researchers. How often should an AI algorithm take actions that it hasn’t tried, versus already tried actions that are expected to lead to some reward. This is a question they constantly have to grapple with. For example, when the computer program AlphaGo beat world champion Lee Sedol in 2016 in the board game Go, it did so by exploring entirely new moves that most human players had never seen being played before. In the second match against Sedol, the AI algorithm made a move so calculated that there was only a one-in-ten-thousand chance that a human player would make it. And it turned out to be a winning move.
Human innovation entails a similar process of exploration. That is why innovating industries have always been highly clustered. From Renaissance Florence, to Manchester during the Industrial Revolution, to contemporary Silicon Valley, cities have acted as “collective brains” by facilitating knowledge transmission and innovation. To be sure, the internet, together with other digital technologies like Zoom and Slack, allows people to work remotely more effectively than ever before. That is why the average geographical distance between co-applicants on the same patent has grown exponentially since the 1990s. And as digital technologies continue to improve, it stands to reason that more work will be done remotely. However, people need to meet somewhere to decide to collaborate in the first place.
Although e-meetings occur increasingly in the post-COVID era — thanks to the miracles of technology — nobody lives in cyberspace. Thus, our virtual interactions mirror our networks in the physical world. Indeed, numerous studies show that many new projects are launched when people meet randomly in various physical settings. For example, we know that collaboration and innovation suffer when important conferences are cancelled.
To realize the efficiency gains from remote work while boosting innovation, business leaders must assess which jobs and tasks are best done remotely, and when the company benefits from employees coming in to the office. To assess this in a systematic way, the remote work matrix to the right offers valuable guidance. As the matrix reveals, whether a job should be done remotely or not is best assessed along two dimensions. The first dimension relates to the importance of knowledge spillovers. In particular, tasks that entail exploration, like developing new ideas and artifacts, benefit from sporadic in-person interactions that do not happen when employees work remotely. Examples of occupations where exploration is important include the jobs of art directors, product developers and software engineers, just to name a few. It is important to remember that even if these jobs can be done remotely, it does not mean that they should be. If companies overdo remote work, innovation and productivity will suffer in the long run.
The second dimension relates to the location specificity of a task. At one extreme, a biolab technician can only do her job in a laboratory. A telemarketer, on the other hand, can work from almost anywhere. To be sure, there are many cases in between. A derivatives trader, for example, can easily work from home, but nonetheless benefits from better and faster computer systems at the office. And while a relationship manager may benefit less from in-person meetings because of knowledge spillovers, she typically needs to be close to her client base.
This matrix has been adopted and modified from Matthew Clancy (2020): The Case for Remote Work, Working Paper.
Thus, in short, tasks or roles that do not benefit from knowledge spillovers and that mustn’t be done at a specific location are best performed remotely. Indeed, they might even be offshored to low-income countries. And as these tasks are digitized and/or offshored, they will become increasingly automatable over time: AI algorithms are capable of learning from digital interactions and will gradually be able to perform more of the tasks previously done by human labor.
All the same, looking forward, the key challenge facing companies and managers will be how to harness the benefits of remote work in ways that don’t imperil innovation. Innovation can happen almost anywhere. Research shows that prohibition reduced innovation as fewer people met sporadically. Whether people work remotely or not, companies must facilitate the sporadic interactions that foster innovation, which is ultimately what drives business performance over the long run.
Giving employees the opportunity to work from home can increase productivity and well-being, as many studies have shown.
In-office hours, however, will be needed to drive innovation in many jobs. Moreover, office space must be reconceptualized to maximize the chance of unplanned encounters. If people work from home more, which they will, companies must get their employees to “collide” when they are at the office. For example, at Sky Central in London, a variety of stairs and ramps connect all spaces, which have purposefully been constructed to create multiple routing options that encourage journey variation throughout the office space. This, in turn, facilitates sporadic encounters and augments exploration, allowing both ideas and people to flow on a continuous basis.
This article was written by Carl Benedikt Frey, Director of the Oxford University Future of Work program.
For more insight from Dr. Frey, tune in to the three-part podcast series (part 1, part 2, part 3) “Redesigning Work for the Post-pandemic Age,” where he’s joined by Ben Pring, who leads Cognizant’s Center for the Future of Work, and Mariesa Coughanour, AVP & Head of Cognizant’s Intelligent Automation advisory practice.