One of the central features of the debate about working from home is that it leads to the loss of random, but productive, encounters with colleagues. I’ve responded with the observation that some of my best research ideas have come from largely unplanned encounters on the Internet.
It’s just struck me that there is a conflict here between the interests of workers and those of firms and managers.
A lot of universities (or, more precisely university managers), think of themselves as developing and promoting a corporate brand. In this context, research collaboration within the university (particularly if it is trans-disciplinary) is viewed very positively, while collaboration with other universities is less well-regarded. But for individual academics, the big rewards come from high-profile work within tightly defined fields, which implies a desire for collaboration with other people in the same field who will, in general, be located elsewhere. While intra-university collaboration may be rewarded in internal promotion decisions, the outside opportunities are greatest for people with external collaborators. Those outside options are routinely used as a bargaining chip in negotiations over salary.
This issue isn’t specific to universities. Labor market theory distinguishes between firm-specific skills and general skills (which are of value to any employer). Back in 1964, Gary Becker made the argument that firms would be willing to pay the cost of firm-specific training for their workers, but not for general training which increases their outside opportunities. (This seems entirely convincing to me, although the empirical evidence I found on a quick search is both limited and inconclusive).
What applies to training also applies to serendipitous encounters. Collaboration with co-workers can enhance productivity within the firm, but doesn’t do much for your market value outside. Conversely, if workers enhance their productivity at home by making more use of industry discussion groups, Skype chats with people in other firms who are addressing similar problems, and so on, that enhances their bargaining power relative to their employers.
In this context, it’s striking that the hardest push for a return to the office is coming from the finance sector, led by JP Morgan. Even though textbook finance is all about hard numbers on earnings, risk and so on, the industry actually operates largely on personal contacts, networks and exchanges of favours, particularly information. That’s why it’s concentrated in a handful of global cities, and why so much attention is paid to issues like “poaching” of staff, no-compete clauses and the like. It’s obviously in the interests of employers to build up internal networks and control external interactions.
As with all these issues, my ideas here are provisional and almost certainly wrong in some respects. So, feel free to correct me.
John Quiggin 
Nice to know that we, the “armée de l’ombre” of JQ’s unpaid blogerati, are strengthening his hand in pay bargaining with his university. If you win, please remember us in a small donation to a suitable Good Cause.
IME managers are often social people who like face to face meetings for that reason. Getting them to produce an agenda or even announce a reason for the meeting can be a challenge, because for some reason they’re shy about saying “I like touching people”.
Geeks who work for them are often not, so prefer short meetings with firm agendas and definite outcomes. Or even better, online discussions leading to decisions.
Speaking of which, in geek world it’s often assumed if not explicitly required that people have extensive online/outside the company interests and experiences. At the extreme are companies who will only hire people who do extensive work for free… in the form of contributing to open source projects. That is a whole other discussion, involving both feared-by-the-right terms like privilege and feared-by-the-left terms like unpaid 🙂
From my experience, initial training is general, and more specific (usually more expensive) training is concentrated on the more talented later on.
One flaw here is the usual economist’s assumption that pay will directly relate to skill. It more usually relates to some combination of industry norms, patronage and whatever profit (distinct from contribution to value) the area generates. That last is easy to see in finance, but finance is atypical.
Also – firms in technical areas collaborate, and share the same knowledge base – and a need to share knowledge if their products are to fit together – if our new alloy needs to be machined in some new way to get the best result, you need to tell the users.
Just on the University experience, I think it’s quite different with teaching and research. The teaching side most definitely requires co-ordination within the University. It benefits greatly from focusing on the interactions between colleagues with the odd networking outside the institution to generate and implement new ideas. Whether it be the best way of handling different students, ensuring appropriate scaffolding through the course to ensure skills are built over an appropriate time frame, just observing other good teachers, it’s the contact with academics within the institute that help the most.
With research, your experience matches mine. The people with the skills and resources that are needed are often not in your institution and forcing collaborations tends to lead to square pegs in round holes.