What work can be done from home? What work needs to be done at an office?

(written by lawrence krubner, however indented passages are often quotes). You can contact lawrence at: lawrence@krubner.com, or follow me on Twitter.

(UPDATE from 2022-05-13 — this post made people angry. I am not trying to upset anyone. I want to emphasize, where I make a personal assertion, my experience is limited to New York City. If you live somewhere else, then you might be seeing a situation very different than what I describe here.

Some people thought I was saying that less work should be done from home, but that is incorrect. I think businesses are happy to have workers work from home, where that business thinks they can pay a bit less in exchange for the workers working from home. I am not saying that is good or bad, I’m only saying it is what I’ve seen in New York City. I believe that in the future more and more jobs will be work-from-home jobs. However, I think some of these jobs will pay less than jobs done in an office. But my experience is limited to New York City.)

The Journal Of Public Economics had an article that suggested that 37% of all jobs can be done at home:

On Reddit I wrote:

The problem is that any work that can be done from home is also work that can be outsourced to India, Vietnam, Romania or Brazil, so this becomes yet another force that pushes wages down in the USA, by making American workers compete directly with people overseas. Here in New York City, I’ve been helping companies hire software developers, and companies have been offering as much as an extra $30,000 to get people back in the office. But that also means that everyone who works from home is effectively taking a pay cut. Which doesn’t matter much when it happens for just one year, but over 5 years this will lead to huge differences. I can imagine two workers who start off at the same level, making the same money, but after 5 years, one is a manager making $250,000, and the other is still a mid-level software engineer working from home for $110,000.

On Reddit, xyrrus asked:

Are you suggesting though that the workers of these 37% should return to the office to maintain the illusion that their job cannot be outsourced and by extension keep wages high as a result? Because a job that can be done from home doesn’t change the fact that it can still be outsourced whether a worker is doing it from home or at the office so the only conclusion I can draw from what you’ve said is that they should definitely go into the office to keep the illusion alive so that they don’t ultimately lose their job which is a bigger problem IMO because you’re effectively trading actual benefits that working from home provides for a unspoken threat that will always hang over our heads and that in order to prevent it from ever happening, we need to keep up appearances by playing pretend.

This is a complex topic, but here is a summary of what I would say in response:

The pattern that I am seeing this year in New York City is that the founders of a startup come to the office everyday, unless they are out in the field meeting customers. When entrepreneurs talk to investors, they prefer to do so in-person. When CEOs talk to their CTO, CFO, and CMO, they prefer to talk in-person. The best salespeople prefer to talk to their customers in-person. If there is a legal problem, and so the CEO needs to meet with the company lawyer, they strongly prefer to have that meeting in-person. Less important workers are allowed to work-from-home, at least some of the days of the week. Your average software developer, working for $110,000 a year and simply cranking out some code, is often allowed to work-from-home. Their manager, making $200,000 a year, prefers to go to the office, to meet with the CEO or CTO. Why do the people at the top want to be meeting people in-person? Apparently they feel there is some vague, hard-to-define, intangible value to meeting in-person. The argument, then, is that there is vague, intangible, extra work that happens in-person, that is not done remotely. Therefore, when someone stops going to the office, and starts working from home, they are actually doing less work. For instance a software developer who starts to work 100% from home probably has more time to focus on the task of writing software, but is excluded from informal, spontaneous brainstorming sessions between the CEO and CTO. Whereas, previously, when the CTO met with the CEO then the CTO might have invited certain senior engineers to participate in the meeting, but now those work-from-home engineers are left out of the meeting (it is awkward to include one person via Zoom when everyone else is meeting in-person). The actual job of the software engineer shrinks: they are no longer invited to participate in spontaneous and creative work that they would have participated in if they were at the office, and because of this, they end up more directly competing with software engineers in India, Vietnam, Romania or Brazil.

That’s the summary. You can stop reading right now. Below I add in more details, but you already know the gist of the argument.

(UPDATE: on Hacker News someone wrote “The current generation of senior leaders came up before working from home was really possible, therefore are by definition going to be the people who thrive in face to face environment.” To be clear, I work with a lot of startups. I was recently working with PairEyewear.com, founded by 2 people who were 26 years old, and who are now 28 years old. They go to the office 4 days a week. I do not see a generational shift, I see leaders of all ages going to the office.)

A note for managers

Whenever I’m leading a tech team, I invite members to lunch as often as possible. Productivity isn’t the goal, so a bigger group can be invited. This is especially important when someone is new to the team — I want to make them feel welcome, so inviting them to join a few of us for lunch is a great way to get to know them, away from the pressures of a specific agenda.

You’ve probably heard the saying “People don’t quit jobs, they quit managers.” Take your team to lunch and get to know them as people. Then you won’t be one of the managers that they want to quit.

Having lunch or dinner with your team can blur the line between one’s work and one’s personal life. Sometimes this is a negative thing, and sometimes it’s positive. On the positive side, your spouse or significant other is typically interested in what your work is like, and such after-hour social events are a moment when they can join you and meet the people you work with.

Fundamentally, humans work, work is human, humans are social, and so all work is social. Having lunch or dinner together is a chance to focus on each other as humans rather than workers. At such social gatherings, I’ve forged many friendships that long outlasted our temporary allegiance to any particular employer. Indeed, a successful career is often based around those friendships rather than any particular company. At the same time, I’ve found that building those friendships allowed me to enjoy a particular job much more than I would have otherwise, and every worker who works with you will probably feel the same.


A few more extraneous details:

About work-from-home, both its benefits and also its downsides, I devote a whole chapter to this in my book One-on-One Meetings Are Underrated, Group Meetings Waste Time. If the subject really interests you, that is where I go into detail regarding the good and the bad.

I’ll mention, of course, that most entrepreneurs now make use of both outsourcing and work-from-home freelance contractors. Entrepreneurs love this because they can hire very talented people for less money than if those entrepreneurs were trying to get together a team in a particular office. Outsourcing and work-from-home: entrepreneurs are lucky to function in an era when they have so many options. But most of these entrepreneurs still have co-founders who they meet with in-person, suggesting that they still feel in-person meetings have some vague, intangible benefits, which the entrepreneurs value. And also, I’m here talking about startups in the early stages, with less than 100 workers. I suspect that as these entrepreneurs grow their companies, the entrepreneurs will want to have an office where, at a minimum, the top leadership can meet, especially as the top leadership becomes a large group of people at a large company.

I’ve managed outsourced teams, so I know how difficult it is, and there are many reasons to use workers who are in the USA. In my book I have a whole chapter devoted to reasons why a company might want to avoid outsourcing. But it remains true, once management says “This work can be done from home” that management will start thinking “Maybe this work can be done in Brazil or Romania?” That’s simply a fact: I’ve had that conversation many times. With a few exceptions, workers who work in an office are seen, by management, as more essential than workers who work-from-home.

(There are laws that sometimes force American companies to hire American workers, but I’ll leave that issue to some future essay.)

When can work be outsourced? There is a famous economist named Coase who wrote a famous thesis on this very subject. (A good summary is on Wikipedia). He asked the question, why do employees exist? That is, why isn’t everyone an independent free agent, a corporation of one? He concluded that there is a lot of work that is not easy to define, not easy to write down, and where “quality” is multivariate and composed of intangibles — in all such cases it is better to hire an employee than outsource the work to another company. But the vagueness of definition that leads to the existence of employees also makes it difficult to allow work-from-home, for all of the same reasons. The vague, intangible aspects of work that lead to the invention of the concept of “the employee” also lead to the invention of the modern office.

According to Coase, there is a large category of work that is somewhat difficult to define, perhaps because the work involves a creative or spontaneous element. Or possibly because the work is multi-variate such that it is impossible to achieve quality on every variable, and so the evolution of quality in particular variables both limits and opens the door to the future of the company (in other words, Google cannot outsource the creation of the algorithm of its search engine, as that algorithm defines the future of Google). In such cases, the manager cannot easily define the work, nor can the manager easily check if the work is being done, nor can the manager easily evaluate the quality of the work. In such cases, the corporation will want to in-source the work. For such work, it is important to build long-term, trustful relationships with those doing the work.

That much is I think a straight re-statement of Coase.

A brief history of outsourcing, as it affected the tech industry:

Between 1990 and 2010 about 3 million tech jobs were outsourced from the USA to nations such as India and Vietnam and Romania and Brazil. None of those jobs are coming back.

There is a large category of very boring, standard, straightforward work where outsourcing is the best, and also the cheapest, option. So long as the work is tedious and highly standardized, it can easily be outsourced. By contrast, it is the work that contains a creative element, then that typically cannot easily be outsourced.

This is a complex issue, but let’s stick with the official statistics from the Bureau Of Labor Statistics. For the category of “computer programmer” the peak was in the 1980s, and then afterwards there was a long decline:

1990 Number of Jobs 565,000

2010 Number of Jobs 363,100

2012 Number of Jobs 343,700

These are absolute numbers. In relative terms, the decline was much worse, since the population was expanding.

However, this decline was offset by the creation of new BLS categories, such as “web designer”.

Very roughly speaking, there was no increase in tech jobs between 1990 and 2010. (There was a bubble during the Dot Com years but it later deflated.) A combination of outsourcing, and then the crisis of 2008, meant that the tech industry saw 20 years of stagnation.

Keep in mind, there was a tremendous transformation of what “tech” meant during this time. Up till 1990, computer programming jobs were largely tied to heavy industry. Deindustrialization therefore meant a loss of hundreds of thousands of computer programming jobs. These computer programming jobs were mostly in Ohio, Michigan, New York, New Jersey, Illinois, Indiana, Minnesota and Wisconsin. As late as 1980 it was still possible to think of New Jersey as the center of the American tech industry, since Bell Labs was there, and Bell Labs had created the Unix operating system.

After 1980, and especially after 1990, a completely new tech industry arose, largely based in Silicon Valley, plus a few other tech centers such as Seattle, Redmond, Austin, New York City. The paradox is that, if you look at job numbers, the tech industry saw 20 years of unchanging stagnation, but if you look at the type of work being done, there was a complete revolution of activity, with both the geography of the tech industry and the focus of tech industry changing radically.

Since 2010 or 2015 the total number of tech jobs has been growing, but we can still see the aftermath of the long years of stagnation. You can see this in the wages: programmers were making $100 an hour in the 1990s just to build simple corporate dashboards using simple tools such as VisualBasic, and nowadays many average software engineers make roughly $100 an hour now, the same nominal wage, but adjusting for inflation this amounts to a 50% wage cut. Only since 2015 have we seen strong wage gains for tech workers, and they have not yet completely caught up with where they were 25 years ago, adjusting for inflation. (But if the current boom lasts another 2 or 3 years, then tech workers should see the kind of wages they last saw during the peak of the Dot Com mania of the late 1990s.)

The freedom to work-from-home is an enormous perk for many workers. Many of them don’t want to go to the office because they face a long commute. Why should they spend an hour going and an hour returning, every day, two wasted hours? And why should they sit at an office and waste their time in meetings? Why can’t they just sit at home, and focus on their real work? Here in New York City, over the last year, I saw employers offering software engineers an extra $30,000 bonus to get people back into the office, in addition to very high salaries. Put differently, this amounts to a $30,000 pay cut for those who work-from-home. For many workers, in the short run, this trade-off is worth it. Many software engineers have told me that childcare and commuting costs them more than $30,000 a year, so they are actually better off if they work-from-home for $30,000 less. This is, of course, a personal decision, and every worker should make the best decision for themselves and their family. However, it clearly suggests that the new pattern means that the ambitious people show up in an office, and the unambitious people work-from-home. There is nothing wrong with this, either way, just so long as the workers realize that working from home for a few years is probably going to limit their career in the long-run. So long as everyone considers this, and makes a rational decision about their long-term goals, then everyone should end up with the situation that is happiest for them and their families.

So work-from-home has happy results for everyone, the workers and the businesses, so long as both sides are correctly estimating what they need and what sacrifices they are willing to make

If we wanted to simplify this discussion as much as possible, we could say:

“Any job that can be done from home can also be done from India or Brazil.”

Sometimes I say this and people say “India is a different culture and it is a different time zone.”

Well, okay, let’s say that’s true, but there is still Brazil, which is only one time zone away from New York City. Brazil has a Western culture, and thanks to a recent political crisis its currency devalued so it now seems like a low cost place to outsource work.

This is the fundamental force that puts downward pressure on those workers who work-from-home — their work resembles the work done by those who work from other countries, and so eventually work-from-home wages will converge towards the wages of those in other countries. By contrast, the top leadership of each company that I work with sees in-person meetings as providing a vague but real dose of spontaneous creativity that is impossible to recreate in any situation other than an in-person meeting.

I’ve often heard workers ask “Why should it matter if I’m at the office? All that matters is whether I get my work done. I should be judged on that and that alone. If I get my work done, on time and with excellent quality, then I should be rewarded for that. And I don’t need to be in the office for that.”

But what is the work? Who defines it? If I was dealing with a very junior software engineer, I will take on the burden of defining all of the work for them, in great detail, then I will watch them closely, stepping in often to correct any mistakes. Such workers can easily work-from-home, or from Brazil. But if I am dealing with a senior engineer, then I’m typically dealing with someone who has a specialty where they know more about the subject than I do. As such, I cannot easily define the work for them. I need them to educate me, and only then can we, working with a high level of feedback and cooperation, jointly define what the work really is. When I am being educated, I prefer to meet in-person with my teacher. My teacher should be able to see when I’m confused, and possibly they can try different strategies to help me learn whatever I am ignorant about. If these experts insist “I’m only willing to educate you on a Zoom call, I’ll educate you via video” then I have the option of hiring any expert in the world, I don’t need to hire an expert in the USA. If I need to educate myself on a set of technologies, so as to evaluate competing software architectures, then I would prefer to meet with experts in-person, but if for some reason I had to rely on communication via Zoom or Slack, there is no reason for me to hire someone in the USA. In such cases, the only reason I would hire a remote worker in the USA is if they were the greatest expert in the world on a subject that I wanted to learn. This suggests that super elite workers can work-from-home, if they wish, and they will make very good money doing it. But those workers who are a bit more average are the ones who will likely face difficulties if they work-from-home. Either they go to the office, and try to offer a vague but important extra spontaneous creativity, or they end up in direct competition with average software engineers from all over the world.

Anyone who wants to dismiss the vague, intangible benefits of meeting in-person then needs to explain what are the vague, intangible benefits of hiring a work-from-home American rather than someone in Brazil

Of course, in between outsourcing and work-from-office there are infinite shades of gray, which can be paid at an infinite sliding scale of values. I don’t think that an American software engineer will ever be paid exactly the same as a software engineer in Brazil. There are vague, intangible but important factors that probably offer some benefit to sometimes hiring a work-from-home American worker instead of a hiring someone in Brazil.

This is important: anyone who wants to dismiss the vague, subjective, intangible benefits of meeting in-person then needs to explain what are the vague, subjective, intangible benefits of hiring a work-from-home American rather than someone in Brazil. If you simply say “vague, intangible benefits should not be considered when hiring people” then there is no need to hire any American for remote jobs, but rather, you are saying that all of the work should be sent overseas. Vague, subjective, intangible, factors have to be considered important, or there is no need for the work to stay in America. In fact, according to the above article from the The Journal Of Public Economics, 37% of all the jobs in the USA can be sent overseas, unless we acknowledge that vague, subjective, intangible factors sometimes create real benefits to hiring someone in America, and if we are putting that much importance on vague, subjective, intangible factors, we might as well talk seriously about the vague, subjective, intangible factors that cause top leadership to prefer in-person work.

So the hierarchy is:

1. Vague, subjective, intangible factors make Americans who work from home more valuable than workers in Brazil or Vietnam (for work that is formally and nominally the same, but really isn’t)

2. Vague, subjective, intangible factors make Americans who work at the office more valuable than American workers who work from home (for work that is formally and nominally the same, but really isn’t)

And we know this hierarchy is real because free market forces have mostly validated it:

1. American workers who work from home are paid more than workers in Brazil (for work that is formally and nominally the same, but really isn’t)

2. American workers who work in the office are paid more than American workers who work from home (for work that is formally and nominally the same, but really isn’t)

Your manager does not want extra work

Again, some people insist “It’s the job of the manager to define the work. Once the work is defined, if the worker can do it from home, on-time and with good quality, then that worker should be rewarded.” But that puts all of the burden of defining the work on the manager. If the manager actually defines the work in such detail, then it can be embedded in a legal contract and outsourced to a 3rd party outsourcing firm. But this top-down, hierarchical idea is false: the burden cannot rest 100% with the manager because the manager cannot fully define the work. The manager needs the active participation of the workers and it is this creative and educational work that often goes best when done in-person. (Again, when I’m dealing with a complete novice, I will define all of their work for them, but when I’m dealing with a senior engineer, they are typically an expert on some subject of which I am ignorant, so I need them to educate me.)

I have often seen 5 people spend 5 hours in a room and still walk away with unanswered questions and unresolved conflicts. And that is with an intense, good faith, in-person effort. Trying to make progress on such complex issues, while working remotely, is more difficult and takes more time. No doubt, given an infinite amount of time, all things can be explained via Zoom or Slack or WhatsApp, but it takes more time and effort, and time and effort cost money. Therefore it is often better for the company to require certain workers to come to the office, so that the meetings can be in-person. For the company, in-office worker can be cheaper and more effective, even when you count in the expense of the office itself. (How much these cost savings should be shared with the workers is, I think, the hottest issue of class-conflict in the USA this year.)

As an aside: when I give a talk to 40 or 50 people, it can seem like a one-sided conversation. I have the microphone, so I am the one is talking, and everyone else needs to be silent. And yet, it is not really a one-sided conversation. I am hyper-sensitive to every reaction that I see in the audience. If I say something contrarian, and someone gasps in shock, I know that I need to slow down and explain why I’m being contrarian. If I tell a joke and people laugh, I know the audience is understanding my cultural references. If I tell a joke and no one laughs, then I know I’ve lost the audience and I need to figure out why (in such cases I’ll sometimes stop my talk and asked members of the audience what they think I am trying to imply). If I tell an anecdote and I see people nodding their heads, then I know my story resonates with the audience. If I tell an anecdote and people look confused, then I know the audience has had experiences different from mine, and I need to stop and ask questions to figure out the heart of our differences. But it is a two way conversation, even if I am the one talking. All of this call-and-response communication is gone on a large Zoom call. Typically in a large Zoom call everyone is on mute except for the person who is talking. I cannot hear them, I cannot pick up on subtle communication. But at least I can see their faces? Sometimes. On large Zoom calls the image of any one person is very small and difficult to see. Even worse, many people turn off their video, in which case I’m talking to a blank, black square, with zero information about how that person is reacting. Obviously in-person communication is easier than this. I find it exciting to talk to a group, but almost terrifying to talk to a bunch of blank, black squares.

On Hacker News, Matt Palmer asked:

So some people claim that vague, intangible but fundamentally valuable parts of a job do not happen at home. What evidence is there for this supposition?

To develop evidence, you would need to generate examples of vague, intangible but fundamentally valuable work. Since that is difficult, research on this subject is going to proceed slowly. Since it takes effort, on the part of the leadership, to define vague, intangible but fundamentally valuable work, doing so clearly has a cost associated with it. This extra cost is the starting point of Coase’s theorem. Rather than define this cost exactly, most workers, and most management, proceed on instincts. So management says “I’m willing to pay an extra $30k a year to get workers into the office, because defining their work will cost me more than $30k a year.” And then some workers say “Working in an office costs me more than $30k a year, so I won’t go to the office” or some say “I’ll make a profit if I take the $30k and go to the office.” Rather than work out this value exactly, most are proceeding on instincts.

Still, we can reason about vague intangible factors by looking at the behavior of the top leadership. Top leadership prefers to meet other top leadership in-person. In New York City right now, the pattern I’m seeing is that leadership meets each day in the office, while less essential workers are allowed to work-from-home. Likewise, with the pandemic ending, most meetings of the Boards Of Directors have gone back to in-person meetings. And the best salespeople I know greatly prefer in-person sales calls rather than Zoom calls. In every case, those at the top behave as if in-person gatherings give them some extra benefit, which might be difficult to define, but whose importance is revealed by the behavior of those at the top. When a CEO meets with their Board Of Directors, they prefer to do so in-person.

On Hacker News, Matt Palmer asked me:

I’m not sure I follow your argument. You seem to be saying that if a job can be done from home you are competing with people in cheaper countries. Therefore you should work in the office, to avoid that competition. But if the job can be done at home, what does it matter if you do that same job in an office? Are you just hoping that being in an office is some kind of protective camouflage?

Very simply, the argument is that the vague and intangible aspects of the job are probably not happening when the work is done at home. Until there is better peer-reviewed and long-term studies on work-from-home, we do need to rely on anecdotal data, and careful observation of people’s real behavior. And again, when we look at the top leadership, their behavior reveals that they put a high premium on the vague and intangible benefits of in-person work. Therefore, we assume that a job in an office is simply not the same job when it is done from home. When done from home, the top leadership feels there is something missing, which is why companies are paying in-office workers more than they pay work-from-home workers.

Many people have written to me to assert some variation of “I work-from-home and I work just as hard as when I am at the office and I actually get more done and I am more productive.” Maybe that is true, but in that case, it is possible that such people as those who have written to me are defining their jobs in terms that are different compared to the way that the top leadership defines those jobs. Is it possible that some workers love to work-from-home exactly because they have an easier time imposing their own definition of their job when they are at home?

All of this is to say, the most important and creative work, at any company, will probably continue to happen at an office, so anyone who is ambitious needs to balance the short-term comfort of working from home against the long-term benefit of building a reputation at an office.

But for entrepreneurs, we can celebrate how many options there are nowadays for finding great talent, all over the world, much more cheaply than would have been possible in earlier eras.

I cover all this in my book One-on-One Meetings Are Underrated, Group Meetings Waste Time and I also talk a bit about the social aspects of work, which is another big subject that every worker needs to think about carefully when they think about their future.

Post external references

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7