The recurring call sits at 11pm your time. Nobody chose that. Someone in London picked an afternoon slot because that was what the calendar showed as free, and the fact that it lands after your dinner never appeared on their screen at all.
That is the part worth getting straight before anything else. What you are dealing with is not a boundary you failed to hold. It is the shape of the org chart, expressed as a clock.
What time shifting actually looks like
Jasmina Chauvin, Prithwiraj Choudhury and Tommy Pan Fang looked at communication data from 12,038 employees of one large multinational, spread across 167 cities and 48 countries. Their question was what happens to a team when time zones pull it apart.
The answer is that people absorb it. Across the sample, 43 percent of live, synchronous communication took place when at least one of the two people was outside their own local business hours. When an extra hour of separation opened up between two colleagues, their overlapping business hours shrank by about 19 percent, while the volume of live conversation between them fell only about 11 percent.
The gap between those two numbers is the hours somebody moved.
The same study found who moves them. Employees based in countries with stricter legal limits on working hours did significantly less time shifting. So did women. Which is another way of saying that the capacity to be reachable at 11pm is not evenly distributed, and that whoever has the most of it ends up looking like the most committed.
Nothing in that data records who was inconvenienced. It only records that the conversation happened.
Why nobody upstream is going to fix this
Overlap with head office is worth real money to the firm. Dany Bahar, working with data on multinational headquarters and their affiliates, found that subsidiaries with more working hour overlap with their headquarters carry out more knowledge intensive work, and that the effect appears sharply right at time zone borders, where an affiliate one hour closer sits measurably higher on that scale.
Read that from where you sit. The overlap you personally supply is not a side effect of the arrangement. It is part of what makes your office useful to the centre.
That does not make it fair. It does explain why it has never come up as a problem to be solved.
The advice that does not survive contact
The standard answer is a policy. Switch notifications off after seven, agree a company wide right to disconnect, and the problem is handled.
Larissa Barber, Alecia Santuzzi and Xinyu Hu tested that directly. They surveyed two samples of full time workers, 235 in one and 482 in the other, measuring telepressure, the urge to respond to messages immediately combined with a preoccupation with how fast you respond. The presence of a disconnection policy was not associated with telepressure in either sample. Whether the policy was consistently enforced made no difference either.
What did predict it was implicit. The felt expectation inside your immediate team that you will be available after hours, and a general sense that work takes priority over home. Most of the workers surveyed did not think a formal policy would benefit them, and close to half expected one to create problems.
The lever is not the policy page. It is what the eight people you actually work with have quietly come to expect.
What is worth doing anyway
Two honest limits first, because writing that skips them is doing something dishonest.
Personal practice does not fix a structural problem. If your company runs on one office's clock, the real repairs are rotating call times and a genuine asynchronous default, and those get decided several levels above you.
And detachment is not free. In a meta-analysis pooling 91 samples and 38,124 employees, Johannes Wendsche and Andrea Lohmann-Haislah found detachment associated with lower exhaustion, as expected, but also with slightly lower contextual performance and slightly lower creativity. Small effects. Real ones. The version of you that has genuinely stopped thinking about work is a little less likely to be the one who catches the thing at 11pm.
With that on the table, three moves that fit inside a normal week and need nobody's permission.
Close the call rather than leaving it open. The same meta-analysis found that what most strongly gets in the way of detaching is not how many hours you worked, it is doing work during non-work time. The association is close to twice as strong for the second as for the first. So the fifteen minutes of notes you would otherwise write tomorrow morning are better spent in the fifteen minutes after the call ends. The point is not the notes. It is that the thing is finished, rather than still open in your head at 1am.
Put the handover in their clock, not yours. You already send something at the end of your day. Add one line: what moved, what you need from them, and the time you will next be at your desk, converted into their local time. It reads as coordination, because that is what it is. It also puts the asymmetry on the page in a form nobody can argue with, without a single sentence of complaint.
Keep a private count. Date, time, which meeting. Nowhere official and nowhere shared. The reason the flexing stays invisible is that nothing records it, including your own memory, which will round a year of Tuesday evenings down to it has been a bit busy lately. A count turns a feeling into a number, and a number is the only version of this that can be raised usefully later.
Detachment can be improved deliberately, though modestly. A meta-analysis of 30 studies covering 3,725 people found a small to moderate average effect for interventions designed to build it, with longer and higher dosage programmes working better than short ones. Worth knowing the ceiling before you start, so a partial result does not get read as a personal failure.
The geometry is not yours to change. What you can change is whether the evening you gave up is recorded anywhere, and whether it ends when the call does.