There was a lot of chatter about work from home last week. A lot of these were the same old debates about whether work from home is good or bad. Is work from home increasing productivity or hurting it?

It started with Marc Andreessen’s Q&A with Noah Smith

It’s hard to overstate the positive shock that remote work works. Remote work isn’t perfect, there are problems, but virtually every CEO I’ve talked to over the last year marvels at how well it works. And remote work worked under the extreme duress of a pandemic, with all of the human impact of lockdowns and children unable to go to school and people being unable to see their friends and extended families. It will work even better out of COVID. Companies of all shapes, sizes, and descriptions are retooling their assumptions on geographic footprint, where jobs are located, where employees are located, how offices are configured, and if there should be offices at all.

Marc Andreessen

And then another with the recently cancelled Antonio García Martínez in which he said:

Any company that was running a physical plant had to shut down that plant, but any company running on knowledge work, and all of the companies that have any knowledge-work component, kept working. Car companies had to stop making cars, but they were able to keep designing cars. The Hollywood studios could no longer film on set, but they could continue to do post production and animation. Basically there is not a single case I know of a company whose knowledge work operations were disrupted by this. The entire financial system kept working, the stock exchanges have worked, the banks worked, the Internet kept working, all these things just kept working.

Marc Andreessen

Then Dustin Moskowitz, the co-founder of the work management tool published a thread on why hybrid work arrangements won’t work

And today, The Ken published a piece on how work from home is leading to increasing working hours, decreased productivity, attrition and workplace disillusionment.

A study on professionals at a large Asian IT services company, conducted in May 2021 by the US-based Becker Friedman Institute for Economics, supports this. The total hours worked increased by roughly 30%, including a rise of 18% in working after normal business hours. Average output, however, did not change significantly. Consequently, productivity fell by about 20%.

Charlie Bilello had a couple of polls on the same subject, the results of which confirm that the shift towards remote work isn’t a passing phenomenon and is here to stay.

Which begs the question? Is work from home good or bad? Will it increase productivity or decrease productivity? I’ll take the cowardly path to answer this. I don’t think there are binary answers here. Taking a reductive approach to thinking about this problem will lead to wrong conclusions and, by extension, misguided prescriptions.

For some industries, remote work is a boon, and for some, not so much. Moreover, in jobs that can be quantified, measuring productivity is easy, but not all jobs, particularly those in the knowledge work domain, can be measured.

We’re still just a year into this experiment, and the data that is being analysed can be quite noisy. The bigger problem in all these debates is the binary phrasing of whether remote work is good or bad. There’s no right answer, and it’s context-dependent. You could reasonably make an argument that it is a net positive and a net negative too.

For every Ken article that shows remote work is hurting productivity, I can cite a paper that shows the opposite. Here’s a concluding remark from this paper by Jose Maria Barrero, Nicholas Bloom & Steven J. Davis:

We also estimate that higher levels of WFH will boost productivity by about 4.6 percent. Over half of this productivity gain reflects the savings in commuting time afforded by WFH. These true productivity gains will go largely unrecorded in conventional productivity statistics, because they do not encompass the effects of reduced commuting time. Indeed, when we mimic the conventional approach, we estimate a comparatively paltry 1.0 percent productivity boost from higher WFH levels in the post-pandemic economy.

While it’s tempting to jump to conclusions, I think we’re still very early in this experiment. Steven Sinofsky, in this brilliant piece, said it best:

At the broadest level the next years will be about experiments, but I don’t like that word because it feels very lightweight and ephemeral. There was time when something like “buy a PC for a marketing executive” was the kind of experiment I’m talking about. I don’t literally mean an A/B test, but rather an alternate model of working — something new that is tried with all of the effort and zeal that any existing process would have been applied. The only thing experimental about it is that many different companies will be trying many different approaches all at once. Over time as people move between companies, and as companies rise and fall, a new set of norms and practices will be established. These will never be as prescriptive as some would like them to be — business and management are social sciences and for every solution there is an abundance of context that makes all the difference.

Then there’s the question of the economic impact. Again, I think this needs more research, but there was some preliminary research by Kristian Behrens, Sergey Kichko, Jacques-François Thisse, which gives you some sense:

First, we find that the long-run effects of telecommuting are all described by bell-shaped curves: Telecommuting first increases skilled and unskilled workers’ productivity and GDP up to some threshold. Beyond that level, a higher share of home-workers reduces the strength of the knowledge and information spillovers which, therefore, do not produce desirable effects. Too much WFH may thus be detrimental to long-run innovation and growth due to limitations of information and communication technologies as well as foregone agglomeration economies in the form of face-to-face contact and knowledge spillovers.

We find that GDP rises slowly and monotonically with the WFH share in the short-run; it is bell-shaped in the long-run. Hence, the short-run benefits may send inappropriate signals about the long-run effects of WFH when all markets have adjusted.

Lastly, Figure 3 shows that telecommuting will exacerbate economic inequality by making skilled workers better off than unskilled workers.

This will be a fascinating debate to watch out for. As more companies and more data comes out, we’ll have a better picture of the impact of remote work at large.