15 min read
There is a sentence people say to dependable colleagues that is meant kindly and lands like an invoice: I knew I could count on you. It is true, it is warm, and it is also a statement about routing. Somebody had a choice about where to send something. Your track record made that choice for them, and the making of it took no effort and cost them nothing at all.
Over one quarter that is pleasant to hear. Over four years it produces a situation most reliable people will recognise immediately: two colleagues at the same grade doing the same nominal job, with workloads that no longer resemble each other, and no document anywhere accounting for the gap. One of them is quietly paying for the other’s certainty.
A tax, and the word is literal
We call it The Reliability Tax™: the compounding operational cost extracted from dependable performers when their consistency becomes load-bearing infrastructure.
Four properties come with the word tax. Two of them fit what happens to dependable people exactly. Two of them do not, and the places where the comparison breaks are where the only workable moves come from.
This is not a reward for being good. This is a rate for being certain.
Why certainty is the thing being charged
The instinctive explanation is that capable people get handed more because they are better at the work. That is close, and it is wrong in a way that matters. Quality is not the variable being priced here. Variance is.
Look at the decision from the sender’s side. Sending an item to somebody who might deliver it, or might come back with questions, or might miss the date, carries an overhead. You have to check. You have to hold a fallback in mind. You have to keep some of the item in your own head. Sending the identical item to somebody who always delivers has no overhead whatsoever. You hand it across and you stop thinking. That gap — the price of holding an open loop — is what makes dependable colleagues the cheapest available destination for anything at all.
Which explains why the load lands on consistency rather than on brilliance. An outstanding performer who is unpredictable attracts none of this. A merely competent performer who never drops anything attracts all of it. The organization is not seeking the best outcome. It is minimising its own uncertainty, one small routing decision at a time, and consistency is the only thing that reliably reduces uncertainty. Your steadiness stops being a personal quality and turns into infrastructure that other people’s plans run across. Infrastructure, once it is being depended on, is rarely asked how it is holding up.
The four properties, and the move each one leaves you
A real tax has an assessor you can write to and arrives on a statement you can hold up. This has neither, and those two absences do more work here than the two properties that match. Each of the four also settles what can be done about it, so the move that answers a property is set out beside it rather than saved for the end of the page.
One. The capacity levy
An income tax is charged on income, whether or not you wanted the income. This behaves the same way. What triggers it is a demonstrated capacity to deliver without drama, and that capacity is not something you can decline to have. You cannot un-deliver the last two years. Every item you handled cleanly is evidence still sitting in other people’s working assumptions, and it is that evidence, rather than any decision of yours, that determines where the next unowned thing goes.
This property leaves you no move at all, which is worth saying plainly rather than working around. There is nothing here to be more disciplined about, and that is why advice about willpower bounces off. The leverage is in the three that follow.
Two. The absent assessor
Here is the first place the comparison breaks. A real tax has an assessor: an office that set the rate, a form, a person who can be written to and occasionally persuaded. There is none of that here. Nobody sets this rate, no meeting agreed it, and nobody could lower it on request. What exists instead is a large number of separate people, each making one sensible routing decision, none of them able to see the others’ answers. So the usual instinct — find who did this and take it up with them — has no target and will burn weeks.
The move: price the next one aloud. With nobody to appeal to, the correction has to be made at the routing, one item at a time, and not through a conversation about fairness. When something arrives because you are the safe destination, the move is not refusal. It is a question about displacement: this will take about three hours, and the forecasting work is what moves — is that the trade you want? Asked that way, the charge gets decided by the person who benefits from it rather than by the person paying for it.
Three. The rising rate
This is the compounding property, and it runs in a short loop. An item is routed to you because you are the low-risk destination. You handle it cleanly, which confirms the rating, which lowers the perceived cost of sending you the next one. Reliability demonstrated is reliability charged. The additional load then has to fit inside the same hours, so it comes out of your slack — the unstructured buffer you used for the harder, slower, more valuable work. That is where most of the real damage sits and it is the part nobody sees. The visible cost is extra tasks. The actual cost is the disappearance of the stretches in which difficult, career-defining work gets done. That work is never dropped loudly. It is postponed indefinitely, one reasonable week at a time.
The move: defend the slack, not the hours. Since the charge falls on unstructured time, guarding your total hours guards the wrong quantity. Block the stretches in which the hard work happens and treat them as immovable in the way a meeting with somebody senior is immovable. Slack is not the spare capacity in your week. It is the capacity your future runs on, and it goes first here because it is the only thing nobody else can see.
Four. The missing statement
Here is the second place the comparison breaks, and it is the one that hurts. A tax, whatever else it does to you, arrives on a statement: the amount is written down by somebody and can be produced on request. Ad hoc work is exempt from every measurement system an organization runs, because it arrives without a ticket, a brief or an owner. So the elevated load quietly becomes what normal looks like for you, and a new joiner would learn it as the shape of the role. There is no return to the previous level, because no record exists that a previous level was ever different. You end up describing something real to somebody who has no way of seeing it, and unmeasured costs are not disbelieved so much as never considered.
The move: make the load countable. If no statement is issued, write one. For a fortnight, log every item that turned up without a ticket, a brief or an owner, with the minutes it took, and do not editorialise the log. The number is the argument, and the number is the only part of this that survives being repeated by somebody else in a meeting you are not in.
None of that is answered by becoming less reliable on purpose, and advice to start missing deadlines is not advice. What the four properties leave you is narrower and better: with nobody to appeal to and nothing arriving on a statement, every correction has to be made at the routing, in writing, at the moment the charge is made, by the only person in the building who can currently see it.
An illustration: two analysts at the same grade
Two analysts join a team in the same month, at the same grade, with the same job description.
Neither analyst exists. The figures below are picked so that the arithmetic can be checked, not because anybody measured them, and none of it should be quoted as data.
Early on, both are asked to turn round an ad hoc data request. The first delivers quickly and cleanly. The second takes longer and comes back with a question, which turns out to be a good question but which cost the requester twenty minutes. Nothing is said. Nobody rates anybody. A preference has nevertheless formed, and it formed on speed and certainty rather than on the quality of the thinking.
A year in, suppose the first analyst is fielding about three ad hoc requests a week and the second about one. The difference between them is two requests. If each takes somewhere between forty minutes and two hours, then two extra requests is between one hour twenty minutes and four hours of additional work every week — and that is the honest figure, not the first analyst’s total. It is worth being precise about, because the gross load is the number people reach for and it overstates the gap by half. Even the honest figure is invisible in every system the organization runs, because ad hoc requests do not have tickets.
Two years in, the second analyst has finished a piece of forecasting work that changed how the team plans, because they had the uninterrupted stretches to do it. The first analyst has not, and has instead built a reputation for being the person who never drops anything. At review, the second is described as strategic and the first as dependable. Both descriptions are accurate. Only one of them is a promotion case, and that is the moment the cost stops being about hours and starts being about a career.
Later still, the elevated load is simply what the first analyst’s job looks like, and there is nothing left to point at. No decision was taken that could be appealed. No memo, no reassignment. There was a sequence of sensible routing choices made by different people who could not see one another’s answers, and nobody anywhere was adding them up.
Three explanations that do not survive contact
Three explanations get offered in place of this one. None of them survives being checked, and each one aims a remedy at the wrong thing.
It is not ordinary workload. Ordinary workload rises and falls with the business, and it arrives with the things that scale alongside it: acknowledgement, headcount, budget, authority, or at the very least a conversation. Unlike normal workload, it increases without corresponding recognition, compensation, or support. That absence is not an oversight inside an otherwise fair arrangement. It is the defining property, and it is why this cannot be waited out. There is no far side of the busy period, because the level does not come back down when the period ends.
It is not appreciation. Being depended on is genuinely something, and mistaking it for recognition is the most costly confusion available here, because the mistake is comfortable. The two behave differently and the difference is checkable. Recognition moves at least one of three things in proportion to what you hold: your pay, your authority, your support. Dependence moves none of the three and raises what you hold anyway. So run the check over three years. If the amount you carry has grown and not one of those three has moved with it, the warm language is describing a dependency rather than an appreciation.
It is not a pay problem. A raise settles a past account. This is a rate on future flow. Being paid more does not alter where the next unowned item gets routed, so the same charge continues against a larger salary. That is the trap inside negotiating your way out of it: you can win the negotiation and keep the position, because the thing that needed to change was never the number. It was the routing, and money does not touch routing.
Working out what you are actually paying
Four questions, and all four want a count rather than an impression.
- Over the past fortnight, how many items reached you specifically because you would certainly handle them, rather than because they were yours?
- Set that count beside a colleague at your grade. Ask them the same question and compare the two numbers rather than your impressions of them.
- In the last two years, has your pay, title, authority, headcount or support moved to match what you now carry? If none of the five has moved, the charge is being made and not assessed.
- What is the piece of work you have been meaning to get to for six months? That piece is the currency this is actually paid in.
The second question is the one that can falsify the whole account, which is why it should not be skipped. If a peer at your grade returns roughly the same count of unowned items, then the routing is not differential, and whatever you are dealing with is a team-wide volume problem rather than the mechanism described on this page. The claim being made here is narrow and testable: the load concentrates unevenly, on the most predictable person, and it does so without the support concentrating alongside it.
A long count here is not evidence that anybody treated you badly. There is no assessor, which was the second property, so there is nobody to take it up with and nothing to reverse. It accumulated out of many small and individually reasonable decisions, and it will keep accumulating until the routing changes — which is what the move under the second property is for, and why it has to be made item by item.
How this term connects to the others
This is the pricing step in a set of ten mechanisms, sitting with The Load-Bearing Person™, Competence Gravity™ and Silent Assignment™ at the responsibility end, none of them outranking the others. Work routed by track record raises what dependability costs, which pulls more work down the same path, which hardens into a dependency the structure plans around, and the sum of that sequence is The Misallocation Effect™.
For how this feels from the inside rather than how it is built, read the surcharge you pay for being reliable. For the question most readers arrive with, there is you are not overworked, you are over-relied-on.
Frequently asked
What is the reliability tax?
The Reliability Tax™ is the compounding operational cost extracted from dependable performers when their consistency becomes load-bearing infrastructure. It is levied by many small routing decisions and never by an assignment, which is why the person paying it can rarely produce evidence that it exists.
Why am I punished for being reliable?
You are not being punished, and that is part of why it is so hard to raise. Organizations route work to whoever costs the least uncertainty to send it to, and consistency is the cheapest form of certainty on offer. Each separate decision to send you something is reasonable. Nobody is adding them up, so the load concentrates without the support concentrating with it.
Does a raise or a promotion cancel the reliability tax?
No. A raise settles a past account, while this is a rate charged on future flow. Unless the routing changes, the same charge continues against a higher salary, and a promotion often increases it by widening the range of work for which you are the obvious destination. The variable that has to move is where unowned work goes, not what you are paid.
Where does The Reliability Tax sit inside The Misallocation Effect?
The Reliability Tax™ is the pricing step: the point at which being dependable starts to determine where unowned work goes next. The Misallocation Effect™ is what the full set of named mechanisms produces, and no single one of them is the whole of it.
Where to start
If those four questions produced a number you did not like, the answer is not a conversation about salary. It is a change to what happens when the next item arrives, and a plan for the pile already sitting behind you.
For the pile, we publish the Responsibility Reclaim System™. The part of it that does this particular job is the script bank: six pre-written reassignment conversations, covering the professional, leadership, peer and personal versions, so that the hardest one you will have does not have to be improvised in the moment you are least able to improvise it. What it is for is returning what is already on your back to where it belongs, rather than declining the next thing, which is the easy half.
Being dependable is not the mistake. It is the most valuable thing you own, and at the moment it is being spent at a rate nobody set and nobody reviews. A rate that has been written down can be argued with. An unwritten one only goes up.
— HCOS