The close of Chapter 19 pinned the test of a site on a single point: whether the most ordinary person, who has no intention of consuming and wants nothing, can stay on without hurry or anxiety. This conclusion can be pursued one layer further — who is able to be unhurried depends in part on whether someone at home is waiting to be cared for, on whom that care falls, and on the terms under which it falls. The caregivers with children at the banyan plaza appear only in the morning, because the afternoon holds another round of pickups and cooking; the building chat group's administrator who "cannot be away from her" (Chapter 17) never appears on any site at all. Beneath the material conditions of re-encounter there lies a more hidden condition: how care arrangements distribute the available margin of society's members. This chapter takes up that condition — asking not who should care, but how each kind of arrangement depletes, replenishes, and transfers the caregiver's margin.
As a thought experiment, we set four care arrangements side by side within the fictional setting of Chengjiang: a retired teacher caring alone for her mother with dementia (family); an engineer household employing a live-in nanny (market); a building-wide shift-exchange mutual aid group (mutual aid); an elderly person entering a nursing home and being cared for by scheduled care workers (institution). The four arrangements serve comparable care needs, and the point of the analysis is not a ranking of better and worse but the direction of margin flow in each — who is depleted, who is replenished, and on whom the exhaustion finally settles. All persons and plots are fictional and correspond to no real institution, person, or data.
How Care Occupies Margin
On the surface, care consumes time: a few hours a day of washing, feeding, accompanying to appointments. But a time ledger cannot compute the real occupation involved in care. What is genuinely occupied is possibility — the inability to accept a posting abroad, the not daring to fall ill, the not daring to silence the phone, the inability to plan a departure longer than half a day. What care locks in is the openness of "what may still happen next"; time is merely the line item under which this locking is booked. Anyone who treats margin as a time account will misjudge her situation with "she actually still has a few free hours each day" — the conclusion of Chapter 14, that "you have time" does not equal available margin, holds equally at the site of care.
We must first set the concept straight. The original meaning of margin conservation is this: no locking ever exhausts the ground, and the unlocked possibility always remains. It is not the conservation of some total of time, money, or social resources, and it does not guarantee that any particular person can always obtain bearable conditions of exit. Applied to care, it yields only two directions of analysis: first, the portion of a caregiver that is locked in has not vanished from the world but has been transferred elsewhere — deferred, shifted onto others, borne by someone else; second, locking is reversible, and when the arrangement changes, margin can regenerate. Tracking transfer and regeneration comes closer to the problem itself than computing any total.
This chapter organizes the analysis with three verbs. Deplete: the burden of care locks in the caregiver's possibility. Replenish: substitutes, rotation, respite, payment, and recognition release the portion that was locked. Transfer: when the arrangement changes, the burden has not disappeared but has only changed bearers — from daughter to nanny, from family to care worker, from one generation to the next. Chapter 9 analyzed the chain by which responsibility shifts among family, market, community, and institution; this chapter asks after something else on the same chain: with each transfer, to whom is the loss of margin left. To judge a care arrangement we must ultimately look at whose margin it depletes, whose margin it replenishes, and whether the transfer is acknowledged and paid for.
How Care Locks in the Caregiver's Margin
In solo caregiving the gap between booked labor and real locking is immense. Dementia care cannot be scheduled by the hour — the mother may rise at any moment, wander off, turn on the gas — so "being present" expands from a time slot into a state: not six hours of care a day, but readiness to arrive within twenty-four hours at any time. Planning collapses first: appointments must carry a pre-arranged excuse for withdrawal, long trips are simply canceled, and even falling ill must be timed to the mother's naps. The depletion of margin therefore far exceeds labor time — what is locked down is the very matter of "being able to be otherwise engaged."
Sleep is kept half-awake, the phone is carried outdoors, groceries are bought at the nearest stall. This state of bodily presence with attention on standby is the most hidden form of margin depletion: it enters no tally of working hours, and often the caregiver does not acknowledge it herself, because "I wasn't doing anything." But a person on standby cannot truly rest — rest is defined not by stillness but by not being interruptible at any moment. Prolonged standby rewrites bodily rhythm and emotional baseline, so that exhaustion often keeps deepening even after the labor has stopped. To recognize this depletion we cannot rely on timesheets; we must ask a more precise question: how long has it been since this person was last in a state of "cannot be reached."
Chapter 2 argued that boundaries are what allow care to be sustainably chosen; here we add the formulation in terms of margin: boundaries are necessary because, under an exclusive responsibility, the very absence of a substitute is a continuous drain. A person who knows that "no one can take my place" must hold margin in reserve for every contingency — unable to spend all her strength, because tomorrow must proceed as usual; unable to collapse, because collapsing means the supply is cut. This self-restraint looks like a sense of duty but is in fact the system allocating the cost of redundancy to the individual. The caregiver in Chapter 17 who "cannot be away from her" is the limit of this form: she is no longer living on her own margin — she has become part of other people's margin.
A Comparison of Four Care Arrangements
The Chengjiang record begins with the first: the retired teacher lives with her mother with dementia; an elder brother remits money monthly from another city; the community makes a phone check-in once a month. Her margin is triply locked — the state of presence, the vigilance of standby, the exclusiveness of having no substitute; replenishment comes only from the remittances and the one week each Spring Festival when the brother returns. Transfer is occurring as well: the brother converts care into money, the mother entrusts her old age to the daughter, the public institution leaves responsibility inside the phone call. This arrangement is supremely efficient — all coordination costs are absorbed by one person's margin; its fragility lies in the same place: she cannot fall ill, and the whole arrangement halts the moment she does.
The second: the engineer household employs a live-in nanny, with a contract specifying cooking, cleaning, and attending the elderly parent. The family's margin is markedly replenished — the couple keeps their full-time jobs, the school pickup no longer clashes, someone is with the elder day and night. But the care has not disappeared; it is transferred intact onto the nanny: presence, standby, no substitute, nothing missing, and more thorough than that because she lives in — she does not have even an after-work stretch of "cannot be reached." What the price mechanism accomplishes here is converting this depletion of margin into a monthly wage that can be bargained over, and that is bargained steadily downward. The market's replenishment is real, and so is the transfer; looking at either side alone will mislead.
The third: the building's mutual aid group is formed of seven or eight households who take turns by day watching one another's elders and children and pool purchases on weekends. Its mechanism is mutual replenishment of margin: the portion locked in each household shrinks through rotation — three hours of watching today buys half a day out tomorrow; with substitutes present, contingency becomes bearable again, and whoever falls ill, the group can find someone to cover the shift. But mutual aid has a precondition: the households' margins must not be strained in sync; only among households with similar care needs and similar structures of time does the cycle turn. Chapter 7 analyzed the transfer of migrants' support networks, and the mutual aid group is its local version: it can absorb fluctuation, but not a crisis in which everyone is strained at once.
The fourth: the nursing home receives an elder who can no longer be cared for at home. The family's margin rebounds immediately and on a large scale — sleep, work, marriage, and time with children recover month by month; the institution allocates care to care workers on a schedule, and the shift system ensures that no one holds exclusive responsibility. On the surface the distribution of margin is the most even, but beneath the roster there are further layers: night-shift density, the ratio of one worker to many residents, procedures timed to the minute — these compress the care workers' margin toward the floor set by the price of labor. The family's relief is real, and so is the care workers' exhaustion; between the two lies a transfer concealed by price and class.
The Extraction of the Family Arrangement and Its Real Value
There is no day that can be called the day of "deciding to exhaust oneself." The solo caregiver's margin contracts month by month: first the long trips are canceled, then the gatherings with friends are dropped, then the re-employment is resigned, and finally even the two hours that "belong to her after Mother falls asleep" are taken up by insomnia. Each step has ample reasons, and each step is irreversible — the social contact withdrawn does not automatically return because free time later appears. Graduation conceals the exhaustion: at every stage the people around see her as "still coping," while the locking of margin compounds, and the later the stage, the more external conditions recovery requires. To recognize it we cannot ask "can she still hold on"; we must ask "over the past year, what things she never intended to give up have been lost."
Care arrangements can also use virtue narratives to conceal extraction: pressing a caregiver's margin to zero can be completely beautified as filial piety — giving up social contact is called duty, resigning employment is called sacrifice, exhausting sleep is called not forgetting the debt of upbringing. The mechanism is a narrative substitution — the distributive question of whose margin is locked in and who bears the cost is rewritten as a question of the caregiver's character; once the distribution is rewritten, complaint becomes a blemish on virtue, and exit becomes a family scandal. The beneficiaries are clearly identifiable: the household that saves the cost of care, the institutions spared responsibility, the care arrangement itself that receives a stable supply. The evidentiary threshold is equally clear: does the praise come together with rotation, payment, and substitutes — if the narrative of virtue always appears where bearing is demanded and never appears where conditions are given, it is the lubricant of an apparatus of extraction.
We must not pre-judge every discourse of virtue as extraction. Gratitude in care is a real relational event: the narrative "she has done so much for my mother" sustains recognition between daughter and mother, between family and care worker; abolish this language and care regresses into the coldness of pure labor. The difference lies not in the discourse itself but in whether it is used in place of conditions: gratitude plus rotation is a relationship; gratitude in place of a substitute is requisition. The method of discrimination is the same as in Chapter 18 — follow the direction of time and benefit, rather than censoring the warmth of the words.
At the same time the other side must be written in full: family care has a component that institutions cannot replicate. A familiar voice can call back the dementia patient's residual memory; a daughter can read the half-sentence her mother never finished — this is not a question of efficiency but a capacity for interpretation underwritten by shared history, and Chapter 2's conclusion that "need is first of all something concretely happening" holds here. To presuppose institutionalization as the solution is to declare that this relational dimension can be discarded at will; just as to presuppose home care as filial piety is to declare that the caregiver's margin can be drawn upon at will. The two presuppositions commit the same error: taking the need of one arrangement as a verdict on another.
The Market Arrangement Converts Margin into Working Hours
The market care contract purchases slices of time: from what hour to what hour, doing what, to what standard. Slices of time can be priced, but the core of care is precisely what cannot be sliced — whether the elder is afraid at night, who remembers his odd habits around medication, who is the first to notice when something anomalous appears. This continuity across time slices is either borne without pay by a care worker long in the relationship (relational labor folded into the price), or borne by no one (the anomaly waits to be discovered at the next visit). Chapter 7 already showed that the temporal structure of connection determines the quality of care at a distance, and market slicing cuts through that structure: payment covers presence, but not the keeping-in-mind.
The care market is a low-barrier labor market, and the side with weak bargaining power fills the price gap with its own margin: accepting live-in work and thereby losing a private life, accepting night attendance and thereby losing whole stretches of sleep, accepting "while you're at it, one more thing" and thereby extending the contract's boundary without pay. What institutions save in price competition is not efficiency but the portion deducted from the care workers' possibility. This is isomorphic with Chapter 4's analysis — education continuously extracting available margin in the name of developing potential — only the object of extraction has changed to the care worker, and the name to professional devotion.
The second stern scenario thereby comes into view: institutionalization and purchased services can outsource the family's exhaustion to low-paid care workers, and the loss of margin descends along class lines. The chain runs: a middle household purchases relief at an affordable price; institutions compete for contracts by cutting prices; the low prices are paid for in compressed staffing ratios, rest, and substitutes for care workers; those able to bear this cost are the group with the fewest employment options, a considerable part of whom are the migrants analyzed in Chapter 7 — they leave home to look after other people's parents, leaving their own parents to an even older spouse or more distant relatives. Margin conservation here shows its cold side: the possibility unlocked for one household has not disappeared but is stored in the losses of others. No conspiracy is required; the price mechanism's systematic draw on the margin of those without bargaining power suffices.
The Mutual Replenishment and the Ceiling of the Mutual Aid Arrangement
Mutual aid replenishes margin through one mechanism only, but it is decisive: the presence of substitutes makes contingency bearable again. The solo caregiver must hold margin in reserve for every contingency, because contingency means the supply is cut; a person in a shift-exchange group can spend all her strength in the present, because someone will stand in tomorrow. Rotation also restores care from identity to time slot — the person on duty today is not "a born caregiver" but simply the one whose turn it is. Chapter 18 said that the continuation of tradition depends on maintainers being able to carry, rotate, and refuse, and the mutual aid group is one of the few arrangements that brings these three down into care.
But mutual aid has a premise that is often romanticized: the households' margins must not move in sync. When everyone in the group is strained, the same crisis — influenza, a power cut, a rent increase — makes all withdraw their availability at once, and mutual aid fails at the moment of greatest need. The concentrated fragility recorded in Chapter 17 is the other face of this mechanism: everyday mutual replenishment conceals a dependence on the few who always fill in. To assess a mutual aid group we cannot look at its everyday activity, but at who can still come out at the tensest moment — and why it is always the same few.
Here the third beautification enters: community warmth. Once mutual aid becomes visible to the outside, it risks requisition — the administrative department writes the group up as a governance achievement, the reports cast the shift-takers as moral exemplars, and what follows is more expectation and less support. Chapter 18's requisition by warmth applies fully here: in the name of "one big neighborhood family," the mutual aid participants' margin becomes a resource that a public image can draw upon again and again. The direction of repair is not to dissolve mutual aid but to refuse to write it as obligation. The moral value of mutual aid rests precisely on the possibility of not participating; turn voluntary replenishment into a default duty, and it degenerates from a mechanism that replenishes margin into a mechanism that extracts it.
The Margin Accounting of the Institutional Arrangement
The institution answers the problem of exclusiveness with the shift system: no one is irreplaceable, and exhaustion is spread out. This is a real advance — care workers go off shift, and in principle still have whole stretches of sleep. But beneath the averaging there is a downward push: cost constraints make night shifts dense, staffing ratios compressed, procedures timed to the minute, pressing the care workers' margin toward the floor of labor prices, and the positions that can bear this floor are, as before, filled by those with the fewest employment options. The institution's margin accounting is therefore double: to the families it sells relief; internally it turns the cost of that relief into a roster.
The cost of institutionalization does not fall on the care workers' side alone. The care recipient's margin — the hour of rising, the route of the walk, the autonomous rhythm of drinking water and using the toilet — is narrowed item by item in the uniform procedure. These look trivial, but they are the last space of possibility the incapacitated person retains. Chapter 19 said that a public site must be judged by whether the most ordinary person can stay on without hurry; the nursing home can test itself with the same question: whether the most ordinary elder, who has no intention of cooperating with the procedure, can continue at her own rhythm. The narrowing proceeds not from malice but from procedure's native exclusion of the anomalous; yet the yardstick of margin reminds us: the excluded "anomaly" is precisely the care recipient's residual freedom.
The family's side of the ledger must also be honestly kept. Entering the institution restores the family's margin on a large scale, and this is a real relief that should not be moralized; at the same time, visiting frequency declines with an intricate mixture of distance and guilt, and everyday mutual watching becomes a weekly ritual. Chapter 13 said that a change of form does not directly equal the disappearance of function; the symmetric formulation here is that the recovery of margin also does not directly equal the intactness of relationship. Acknowledging the relief and mourning the rupture can both be true at once — that is the honesty this section means to keep.
Virtue, Compounded Loss, and Crisis
Filial piety, professional devotion, community warmth — three discourses serving three arrangements, with one and the same grammar: rewriting a state of distribution (whose margin is locked in) as a state of character (this person is good). Once the rewriting is complete, the question of conditions can no longer be raised — asking "can she rest" sounds like profanation, because the narrative has already defined not resting as her virtue. Discourse does not create exhaustion; it renders exhaustion immune to questioning. Every culture of care contains this grammar, including the one that looks warmest.
The beneficiaries are stratified by arrangement: households save on the cost of care, institutions use devotion narratives to lower the cost of retention and substitution, public departments use warm stories to postpone the necessity of institutional support. The discourse evidence to check is not complicated: does the praise appear on occasions where bearing is demanded, or on occasions where conditions are given; after the commendation, have rotation, substitutes, and payment changed; can the praised one say "I am tired" without damaging her image. If the narrative of virtue appears precisely wherever cost can be saved, the extraction structure is more credible than any benevolent explanation.
Extraction ultimately recoils upon its beneficiaries. A caregiver at zero margin fails at some point — a fall, a missed medication, a collapse — and at that point praise cannot buy back the continuity of care; virtue is not the antidote to exhaustion but its deferral. But the counter-case must equally be written in full: gratitude, respect, and honor are real needs for recognition within care relationships. The care worker wants her skill to be seen, the daughter wants her giving to be known by her family; abolish recognition and care regresses into labor service. Recognition's true position is as an addition after conditions, not a substitute before them — with rotation and payment already in place, praise is an ornament; used in place of rotation, it is a receipt for extraction.
The Compounding of Margin Loss
Care's locking of margin exceeds the period of care itself: the years withdrawn from employment hollow out the pension, the interrupted career makes re-entry a demotion, and the long-absent social contact does not automatically recover once care ends. What the caregiver spends is not only present margin but the future distribution of possibility — this is precisely the difference between available margin and commutable resources: a loss of money can be compensated, but the years locked away change the starting point of every subsequent choice. To evaluate care arrangements by present burden alone is to systematically underestimate their long-term extraction.
Chapter 18's gender distribution reappears in the dimension of margin: the default caregiver is still the daughter, the wife, the mother, the female care worker; "she is more attentive" and "Mother is at ease" are the everyday grammar of conscription. Chapter 9's transfer chain falls upon those with the fewest options, and gender has distributed "fewest" before any statistic occurs. The triple locking of presence, standby, and no substitute therefore has a consistent gendered face, and the virtue discourse that underwrites it is thickest: virtuous, great, born knowing how to care. This is not to say that men do not care; it is to say that the current arrangement need not wait for them.
The loss of margin is also booked across generations. Chapter 18 pointed to the time debt transmitted in education; the version within care arrangements is this: children who grow up in a household where "Mother cares for Grandmother until she collapses" learn not the text of filial piety but a specimen in which care equals exhaustion. They either replicate the specimen or flee all obligation of care — both directions convert the previous generation's loss of margin into the next generation's starting conditions. The sustainability of care is therefore not this generation's private affair; it determines what distribution of possibility is handed to the next generation.
The Margin Run in the Moment of Crisis
Chapter 17's rainstorm has an exact correspondence in the dimension of margin: the strains that are normally unsynchronized across households synchronize in crisis — water and power cut off, a family member fallen ill, income interrupted, all at once. The mutual aid group that absorbs fluctuation in normal times has its own margin drawn away at this moment; the market's home services stall when need is highest; the institution enters emergency scheduling. The moment a care system most needs redundancy is precisely the moment when all redundancy is simultaneously occupied.
Care in crisis therefore falls back upon the most readily deployable person. Chapter 17 gave this structure, and the dimension of margin supplies its cost: deployability does not equal sufficiency of margin — quite the opposite; those who already bear exclusiveness and standby before the crisis have the thinnest margin, and the crisis pushes them past zero. The analogy to a bank run is limited: if everyone turns to the same reliable bearer at once while his capacity is fixed, the stacked demand may exceed what he can bear; this does not mean the two processes share a model or a probability of occurrence. A crisis report that merely tallies "someone provided care" cannot see this layer.
The conclusion of stability continuation gathers here: redundancy, low loss, multiple paths, rotatability — these are not merely the comforts of normal times; they are what remains in crisis. A system that presses its caregivers' margin to zero in normal times is not an efficient system but one that has spent its crisis capacity in advance. The reminder of margin conservation is therefore double — possibility is always there, but if it has been locked away cleanly in normal times, then in crisis the only path left for unlocking it is collapse, the most costly path of all.
The Distribution of Margin and the Boundary of Care
Let us gather the foregoing mechanisms into verifiable questions. For any care arrangement: who is doing the caring — the person whose turn it is, or the person assigned by default; is there a substitute — can one fall ill, leave, say "not this time" without care being interrupted; to what degree is one on standby — how long since being in a state of "cannot be reached"; on whom does exhaustion fall — is the transfer acknowledged and paid for; how much everyday autonomy does the care recipient retain; which sentence of virtue discourse is currently standing in for conditions; and who can still come out in crisis. These seven questions share the method of Chapter 18's labor audit and Chapter 19's inquiry into sites: first unfold what has been folded, so that the trade-offs can be discussed.
The audit must guard against one error: converting margin into a summable account — points, time banks, care savings. Chapter 9 already analyzed how forms and accounts create new care labor; margin is especially unsuited to management by account, because its core is precisely the part that cannot be commuted: the openness of possibility, the freedom not to be found, the permission to fall ill. Measurement also thins margin — once rest becomes an expenditure that must be reimbursed, rest itself carries a sense of debt. What the audit offers is questions, not a new ledger.
The same record of care costs can support different decisions: switching to other services, or continuing care at home under conditions of full information and bearable strain. Whichever is chosen, the care recipient and the primary bearer should both participate, and it should be stated how the division of labor will be redone the next time conditions change.
Respite Conditions and the Boundary of Care
Foremost among all respite conditions: making "the caregiver may be excepted" not equal "care is interrupted." A substitute need not be all-capable — someone who can cover one night, take one week, sit through one afternoon at the hospital is enough to release the greater part of exclusiveness. Rotation, respite, the right to refuse — Chapter 2's right of refusal and Chapter 18's "able to carry, rotate, refuse" converge here into one thing: restoring the caregiver's margin from the foundation of the system to its reserve. The foundation cannot move; the reserve can be periodically rebuilt — this difference of position is what decides whether care is sustainable.
The second: extracted margin needs real channels of replenishment — payment, guaranteed continuity, paths of re-entry after interruption, recognition and advancement of care skills, and an institutional floor under the care workers' margin: whole stretches of sleep, a private life with boundaries, the right not to be requisitioned beyond the contract. The principle on which this volume relies applies here as well: the transmission of pressure should be dissolved on the advantaged side, not transferred downward along the gradient. The standard of repayment is therefore the same — whoever benefits from the arrangement bears the cost of replenishing the caregiver's margin; pressing the cost back onto the person with the thinnest margin is to use the arrangement to replicate the very problem it was meant to solve.
The final boundary: family, market, mutual aid, institution — the four arrangements each have their profile of extraction and replenishment, and none can be pre-judged as the solution. The family preserves the relational dimension while extracting from the one who is exclusive; the market replenishes the family while pressing down the care worker; mutual aid replenishes margin across households while fearing synchronized strain and being turned into obligation; the institution averages exhaustion while narrowing the care recipient's everyday autonomy. Discrimination always lies in the mechanics of use: whether substitutes exist, whether transfers are paid for, whether virtue has replaced conditions, whether the care recipient can still live at her own rhythm. Whether a care arrangement is sustainable comes down in the end to this one point — whether the person carrying it can one day say "not this time" and care continues as before; just as with Chapter 19's site, which is judged by whether the most ordinary person can stay on without hurry.