The rule underneath the rules
What one benefit does to another
Care Entitlement rules are written one program at a time and are read one program at a time, but they do not operate one program at a time. An award can reduce another award. Money that arrives once can count against a household for months. A raise can cost more than it pays. None of that is hidden, and almost nobody is told it before they act.
A screen answers one question at a time. A household lives in all of them at once.
Every program was designed against a picture of a household. No program was designed against the other programs that household is already in.
Asking whether a household may be owed something is a single-program question, and it has a single-program answer. Deciding whether to take it is not. The second question is whether the award changes anything the household already has, and that answer lives in a different rulebook belonging to a different agency that was never asked.
This is the failure most likely to be caused by good advice. Somebody helpful identifies a real entitlement, the household applies, the award is correct, and something else moves. Nobody involved did anything wrong and the household is worse off.
Four shapes that interaction takes
One award counts as income to another
Money is income to a program unless that program's rules exclude it, and the exclusions differ. Cash from one source may be fully counted by a second program, partly counted by a third, and ignored by a fourth. Support given in kind rather than in cash, somebody else paying rent or buying food, is treated as income by some programs and not by others.
One award unlocks or closes another
Some entitlements are categorical: holding one confers access to a second automatically, without a separate test. The useful and uncomfortable consequence is that losing the first can end the second, even though the household still meets the second program's own rules. A household can be cut off from something it plainly qualifies for because the door it came through closed.
Money that arrives once is counted twice
A lump sum, back pay, a settlement, a small inheritance, is commonly treated as income in the month it arrives and as a resource from the following month onward. A household can be over a resource limit in month two because of money it already spent in month one. Programs differ in how long they allow for it to be used or set aside, and that period is the whole question.
A small rise costs more than it adds
Where a rule is a threshold rather than a slope, crossing it removes the whole award rather than part of it. Extra hours worth a little can end support worth much more. This is the one families are most often warned about and least often given a number for, because the number depends on which programs the household holds and how each counts the same dollar.
What to establish before accepting anything new
Three questions, asked of the agency that runs the award the household already has rather than the one offering the new thing. The offering agency is not required to know, and usually does not.
Does this program count that money, and how?
Counted, partly counted, disregarded, or counted only above a floor. The answer belongs to the existing program, not the new one, and it is worth getting in writing.
Is anything here categorically linked?
If the household holds something that automatically confers something else, the linkage needs to be known before either is given up, changed or allowed to lapse.
Is there a protection that already covers this?
Many programs carry provisions for exactly these situations: periods in which a lump sum does not count, continued eligibility while earnings rise, disregards for particular kinds of income. They exist, they are frequently unmentioned, and they are almost never applied automatically to a household that did not ask.
Why this one is worse than it looks
The other failures in this series are losses of something the household was owed. This one can be caused by claiming something the household was owed, which makes it the only case where acting on correct information produces the harm. That is enough to make a family stop trusting the next correct piece of information, and it should be, because nobody warned them the first time.
It is also the interaction an assessment is least able to see. A tool that screens against a catalog is answering what may be available. It knows nothing about what the household already holds unless it asks, and asking well enough to model the interaction means asking for far more than a screening tool should hold.
What we hold, and what we do not
Where a program's rules state how it treats income from another named source, that treatment is recorded with the rule, sourced and dated. Where they do not state it, we record that they do not, rather than inferring it from a similar program. Two programs that look alike routinely count the same dollar differently, and that is precisely the assumption this page exists to break.
What we do not do is model a household's interactions. Doing it honestly needs every award currently held, every amount, and the date each was granted, which is more than we ask for and more than we want to hold. Where an interaction turns on a fact we do not have, the answer is unknown, and unknown is never rendered as a no. A caseworker at the agency holding the existing award can answer this and we cannot.
This page covers what an award does to another award. The other ways entitlement goes wrong each have their own page: who gets counted when a limit is applied, in whose income counts; qualifying without receiving, in when the rule says yes; awards that end with no decision behind them, in when a benefit stops; difference by place, in same facts, different answer; and what a refusal actually decided, in what a denial decided. All six sit under what Care Entitlement means.