Is Working Remotely Less Productive?
This is my daughter helping me write my dissertation. She helped by crinkling papers, pulling my beard, and teaching me that one should never, ever attempt to play peek-a-boo while changing a diaper.
Like most doctoral students, my work arrangement back then could be described as hybrid: Sometimes I worked at home, sometimes I went to the office.
I would not have wanted to be in the office all the time, though I assumed I'd be more productive there. Now I'm not so sure. The time saved by not having to un-crinkle papers and change diapers would've been offset by the time needed to get my daughter ready for infant care and commute more frequently to campus. Then, I'd be sitting at my desk trying to work while missing her and wondering why career success seems to require that parents spend less time with their children.
More could be said about my dissertation years, but you get the idea: The relationship between where you work and how productively you work can be complicated.
A new paper from the Stanford Institute for Economic Policy Research (SIEPR) has gotten me thinking about these complications. It's an illuminating paper on recent trends in working from home. One of the conclusions, reported in business-oriented media like Bloomberg and Forbes, is that remote work diminishes productivity.
Why is this important? Most people prefer working remotely, at least part of the time. In recent Gallup polls, for instance, roughly 60% of full-time employees prefer hybrid schedules, while just over one-third want to be fully remote. The SIEPR paper suggests that if these folks get what they want, at least some of them would become less productive.
In this newsletter I'll discuss SIEPR's conclusion and how strongly it's warranted by the data. I'll demonstrate that strong methods and statistics don't guarantee strong conclusions. And I'll suggest that employers should be as responsive as possible to the workplace preferences of individual employees.
Background
Since the mid-1960s, thanks to advances in computer technology, working from home (WFH) has become increasingly common. Following a dramatic spike during the pandemic, WFH has declined but remains far above pre-pandemic levels.
The figure above shows the percentage of days worked at home among people earning at least $20K per year. We can also simply ask how many full-time employees work on-site vs. hybrid vs. fully remote. The breakdown now is roughly 60%, 30%, and 10%, respectively, with historical trends similar to those in the figure.
Most experts expect WFH, especially hybrid schedules, to become more prevalent moving forward. The traditional, on-site, 9-to-5 workday is becoming merely one option in an increasingly flexible workplace.
Is WFH desirable? That depends on who you ask, which WFH schedule you're asking about (remote vs. hybrid), and what outcomes interest you (productivity, motivation, retention, etc.). The most striking claim in the new SIEPR paper pertains specifically to remote work and employee productivity.
Remote work and productivity
The new paper, available now on the SIEPR website, was co-authored by Drs. Jose Barrero (Instituto Tecnológico Autónomo de México), Nicholas Bloom (Stanford University), and Steven Davis (University of Chicago).
Along with recent trends in WFH, the SIEPR team looked at perceptions of remote work as well as its actual impact.
Regarding perceptions, the SIEPR team found a disconnect between employers and employees. On average, employers believe that WFH diminishes productivity, while employees believe that productivity increases.
As for actual impact, the SIEPR team notes that recent debates often stem from confusion between the two types of WFH. They claim that whereas hybrid schedules have no impact or a slightly positive one, fully remote work diminishes employee productivity.
Why would remote work make you less productive? According to the SIEPR team, it's a combination of losses and gains: You lose easy access to colleagues (whom you may need for everything from time-sensitive communications to mentoring), while gaining distractions from your home environment. As Dr. Bloom noted in an email to me this week, "the old joke was that 'the fridge, the television and the bed are the three great enemies of working from home'".
Based on recent studies, Bloom and co-authors suggest that fully remote work lowers productivity by an average of 10 to 20%.
This is the conclusion that attracted media attention. Although productivity can be difficult to assess, the studies they reviewed relied on sophisticated methods of measuring seemingly objective variables (e.g., number of phone calls fielded per hour by call-center employees).
What does the data really show?
In my newsletters I often note that good statistics can't save weak methods. Good statistics are also unable to guarantee strong conclusions.
The studies reviewed in the SIEPR paper all rely on powerful statistical models. But one big, honking detail these studies all have in common is that they were conducted following the onset of the COVID-19 pandemic.
To be able to speak in causal terms – to say that remote work causes a decline in productivity – you have to do two things that these studies cannot do:
1. You have to show that the culprit is remote work per se and not the shift to remote work.
All of the studies looked at how productivity changed when employees shifted to remote work. During workday hours, most of these people suddenly found themselves quite close to their fridges, TVs, and beds – not to mention their pets, hobbies, and vehicles. That's a lot of temptation, and perhaps some of them succumbed more often than they should've. (I know I did.)
In contrast, people who've worked remotely throughout their careers have probably become more accustomed to managing such temptations.
In short, the studies may only show that productivity declines immediately after a shift to remote work. Perhaps the effect is temporary. It's not clear that remote work per se permanently diminishes productivity.
2. You have to show that the culprit is remote work per se and not remote work during the pandemic.
We can't generalize from the pandemic to much of anything – except another pandemic.
COVID-19 did more than just change work environments. Some people who were forced to work from home didn't like it. No surprise if their productivity diminished. For others, productivity was impaired by pandemic-related stress. Working from home was, for many, a constant reminder that they and their loved ones were under siege by the coronavirus. Arguably, it would be easier to temporarily forget the pandemic and concentrate on one's work in the familiar surroundings of the office or cubicle.
All this by way of saying that the effects of shifting to remote work have not been distinguished from the effects of the pandemic. In technical terms, the onset of the pandemic and the shift to remote work are confounded.
In sum, based on the current data, we can't tell whether one or more of the following is true:
–Remote work is generally less productive.
–Shifting to remote work merely causes temporary declines in productivity.
–Pandemic-related stress causes declines in productivity when working from home.
Generalizability issues
Even if you set aside the ambiguities I just described, the studies reviewed by the SIEPR team focused on the productivity of call-center employees, data-entry workers, and IT professionals. This is a very narrow sampling of remote work.
In my email exchange with co-author Nicholas Bloom, I asked whether remote work might actually not impact productivity in certain industries. Dr. Bloom responded, in effect, that if a particular industry allows for remote work, your productivity will tend to diminish if you pursue this option:
"Some industries are well set-up for [remote work]. The most obvious are call centers and data-entry as these are very easy to measure tasks remotely, have little team-work and can be done from home. Interestingly the studies on this find fully-remote tends to reduce productivity but hybrid increases it."
(The comment about hybrid work is, I think, shorthand for the Bloom and his co-authors' contention that hybrid arrangements either have zero impact on productivity or slightly increase it.)
Although Dr. Bloom's comment makes sense, I think you could still question the generalizability of the conclusion. People who work for themselves, people who work on commission, and people whose immediate supervisors make concrete, time-sensitive demands will tend to maintain their productivity regardless of whether their work is face-to-face or remote. Interior designers have increasingly worked from home since 2020 (you send them photos and measurements of your rooms; they send back visuals and video-chat with you about the details) and so have accountants, but I suspect their productivity remained constant, because it had to.
What is productivity anyway?
Even setting aside the ambiguities and the limitations on generalizability that I've discussed so far, there's still one more interpretive issue.
We all know what productivity is (sort of), but can we readily measure it?
In the simplest case, you can just count outputs, like numbers of calls answered, words translated, or garments sewn, but I doubt you'd want to stop there. It doesn't seem very productive if the calls are answered unhelpfully, the translations sound awkward in places, or the garments are stitched together poorly. Productivity is a matter of quality as well as quantity.
Studies on the topic acknowledge this. For instance, in the call-center study that SIEPR reviewed, productivity was defined as the average number of calls answered per hour, but the researchers also considered the quality of support, as determined by customer satisfaction with each call (expressed via the familiar one-to-five star rating system.)
The researchers found that once employees shifted to a remote schedule, the number of calls they handled per hour decreased, but customer satisfaction remained unchanged. The conclusion was that productivity had declined. Employees were working as well as they always had; they were just doing less.
At the same time, as the researchers themselves admit, customer satisfaction ratings are far from a reliable guide to quality:
"A satisfied customer might not leave a review (the participation rate is 11.5%); a dissatisfied one might leave a 5-star review to be polite (the mean review is 4.9 out of 5); and an irate customer might leave a 1-star review regardless of what the worker says."
Yikes. Apart from the fact that 11.5% participation is a small and likely non-representative sample, mean customer reviews of 4.9 out of 5 make it impossible to know whether quality of support remained constant. If quality improved after the shift, for instance, we wouldn't know it, because ratings were already about as high as they could possibly be. This is called a ceiling effect. If workers handled fewer calls following the shift because they were providing better, more thorough service per call, you wouldn't say their productivity had diminished.
In sum, through no fault of the researchers, we may know less than we think we do about the impact of remote work on productivity, because productivity is inherently difficult to measure.
Individual differences
All of the studies (plus informal observation and first-hand accounts) tell us that the impact of remote work varies a lot from person to person. Some people thrive, some do not, some exhibit no discernible changes.
I asked Dr. Bloom about his views on who tends to be be more vs. less productive in remote work environments. I added that I was specificially interested in in knowing whether individual beliefs about impact on productivity are linked to actual productivity. Here's his reply:
"Great and hard to say... I think self-control is definitely a factor. Much like some students study very well in their bedroom others need to go to a library or coffee shop. Broadband connectivity and having your own room that’s not your bedroom are also big factors. Having no folks interrupting – think kids – also matters. Finally, I think some folks are more gregarious and dislike the isolation that WFH can bring at times."
All of this makes sense to me. What's missing though is that connection, if any, between what people believe and how they actually peform. Data from a variety of sources (Gallup, SIEPR, Work Trend Index, etc.) tell us that that a substantial percentage of on-site workers would prefer hybrid or remote arrangements. A smaller but not inconsiderable percentage of remote workers would rather be back on site, at least part of the time. This leads to a simple prediction: Productivity will be influenced by consistency between where people actually work and where they'd like to work.
My prediction is based on the assumption that happier employees tend to be more productive. This is a common-sense assumption supported by independent research, but it hasn't been adequately tested yet.
Conclusion
This week, without intending to be ironic, Zoom announced that all employees within 50 miles of an office would need to start showing up in person, at least on a part-time basis.
Zoom's decision is understandable. Employers want employees in their chairs, because office space is costly, and because on-site work is thought to foster employee communication, engagement, and productivity.
Remote work indeed affects employee interactions – recent studies show that remote workers receive less feedback and mentorship than their on-site colleagues do. What's not clear from the research is how productivity is affected.
This line of research is new – most of the studies discussed in this newsletter are working papers by prominent experts – but two findings have already bubbled up with some consistency: A slight majority of people prefer hybrid schedules (around 60 to 70%, depending on population and study), but hybrid work per se doesn't undermine productivity. In some cases, slight benefits may accrue.
Hybrid work may foster productivity insofar as it draws on the best of both worlds: Face-to-face interactions with colleagues on some days; on the other days, freedom to work when one is feeling most productive. For instance, Microsoft researchers have noticed a pattern they call a "triple-peak day." Traditionally, productivity tended to peak immediately before and immediately after lunch. Microsoft examined keyboard activity (with employee consent) and found an additional peak in the evening.
However, only about 30% of employees exhibited that third peak. This illustrates what I take to be the surest conclusion from the literature: People vary a lot in how they react to hybrid and remote work schedules. I've talked to people who swear by on-site work, because their home environments are too distracting (the fridge, the comfy couch, the kids, etc.). Others strongly prefer WFH. My daughter, an energy policy expert who no longer crinkles my papers, recently shifted to a primarily remote work arrangement and tells me that she's more productive now, because she doesn't have to spend time preparing for the commute and then actually commuting.
In my opinion, employers would be best served by allowing, to the greatest extent possible, each employee to choose among on site, hybrid or fully remote work options.
Naturally, employers shy away from the logistical hassles and perceived injustices that would arise if each employee chose their own schedule. But, as I've argued in this newsletter, the connections between where people work and how productively they work aren't clear yet. There's no particular reason to doubt that greater flexibility on the part of employers would lead to net gains in employee satisfaction and productivity.
Thanks for reading!