The Half-Sentence I Left Hanging Last Time
Last time I wrote about myself: outsourced, a shade over ten thousand yuan a month, and a liquidated-damages clause in the contract reading three hundred thousand.
Halfway through I realized that telling only my own story left a lot unsaid. Because the real object of outsourcing isn’t just the people who got outsourced. It presses down on the ones who haven’t been outsourced yet, at the same time.
This one finishes the other half of the sentence: why the regular-employee seats keep getting scarcer, what a degree is actually doing inside this structure, and — if working for a wage is genuinely narrowing — where the exits are.
One Desk, Two Contracts
The person at the desk next to mine is a regular employee. We do the same work, in the same system, at the same morning stand-up, and we eat lunch at the same table.
We are not the same kind of person.
- Payday differs: theirs is the 15th of the month, ours the 25th.
- Housing fund differs: theirs is matched on both sides at actual salary; ours is at the minimum base, or simply doesn’t exist.
- Year-end bonus differs: theirs is two to four months; ours is zero.
- Annual leave differs: theirs accrues with years of service; ours is given at the statutory floor.
- Physicals, supplemental medical insurance, team-building, training budget: they have them, we don’t.
- System permissions differ: certain data they can see, we can’t.
- Layoffs differ: they get N+1; we get “returned” to the staffing company.
That isn’t a gap in treatment. It’s a gap in status.
Treatment can be negotiated; status cannot. I can fight my supervisor over one raise, one project, one slot on a business trip. I cannot fight my way into “signing the contract a different way.” That isn’t within his authority — it’s the base-layer setting of the whole employment structure.
Article 63 of the Labor Contract Law says equal pay for equal work. But what it governs is the “method of distributing labor remuneration.” It doesn’t reach the year-end bonus, the housing fund ratio, supplemental medical insurance, the promotion track, or whether you qualify to walk into that meeting room.
And it asks you to prove the positions are “of the same kind.” All a company has to do is rename the role — “Operations Specialist” versus “Operations Support Post,” “Development Engineer” versus “Technical Delivery Engineer” — and on paper they’re no longer the same kind.
Equal pay for equal work has been on the books for seventeen years. The part that has actually been enforced is the base-salary column.
The Two-Track System: The Core Shrinks, the Fringe Grows
If outsourcing is this convenient, why not outsource everything?
Because there are things outsourcing cannot do.
- Some positions have to sign. When something goes wrong, there has to be a person who can be held liable in law.
- Some knowledge has to settle. Outsourced staff turn over too fast; every handoff loses half.
- Some clients deal with faces. What they signed is with this company, not with the small print on my badge.
- Some credentials have to be registered. The certificate must sit under a regular employee’s name.
So the stable state isn’t “all outsourced.” It’s two tracks: a small, hard core, plus a wide, replaceable fringe.
And the smaller the core gets, the higher the bar to enter it. Degree, age, résumé, connections, whether anyone will vouch for you — layer upon layer, the price keeps rising.
Regular employees aren’t being laid off to the last one. They’re being compressed into a very small circle. People outside look in and call it stability; people inside know it’s only a matter of time before they’re pushed out.
That is the real mechanism behind “regular employees are being eaten too.” It isn’t turning regular employees into outsourced ones — that’s only part of it — it’s making the regular-employee seats themselves fewer. Fewer seats means whoever is left has to pay more to keep one.
Forced Conversion to Outsourcing Is a Very Tidy Layoff
For the past few years a lot of companies have been doing the same thing: they don’t lay you off, they have you “transfer your contract.”
Years of service reset to zero, the open-ended contract disappears, the social insurance base drops, and the severance that was “termination of the labor contract” becomes “nothing to do with the original employer.” The official name for the whole process is “organizational optimization,” “business restructuring,” “flexible employment.”
Run the numbers.
Someone with 8 years of service at 20,000 yuan a month. A normal termination means N+1, which is 180,000.
Convert them to outsourcing and you pay nothing at all. “Optimize” them the following year and it’s calculated on one year of fresh service: two months, 40,000.
Same event, different name, 140,000 saved.
This is the most honest translation of “cutting costs and improving efficiency”: not making the cake bigger, but swapping out the people who get to slice it.
“Win-Win” in a Growth Phase, “Win-Win” in a Stock Phase
The word “win-win” means two completely different things in two different eras.
In a growth phase the cake is expanding. You don’t have to take anything from anyone; your own slice gets bigger on its own. There, “win-win” is a description. It’s a fact.
In a stock phase the cake has stopped growing. Distribution turns zero-sum: one more piece for you is one less for me. There, “win-win” is rhetoric — because nobody is going to agree to “you take a bit less,” so it has to be phrased as “let’s all strive together.”
So when I look at a company, I don’t listen to what it says. I look at three things: how the contract is written, what day payday falls on, and how many months it hands over in a layoff.
The collaboration, cooperation and win-win talk, and what’s actually written in the contract, are two separate languages. The first is for meetings; the second is for litigation.
A Degree Isn’t Discrimination, It’s a Distribution Ritual
A lot of people read the degree requirement as “the company looks down on people without one.” I don’t see it that way.
Degrees do two jobs in this structure.
First, they act as a cheap filter. A diploma doesn’t prove ability, but it does prove that this person could tolerate four years of discipline, hand things in on time, and survive inside a rulebook. For the recruiter, that’s the lowest-cost way to do a first pass. Whether the role actually needs that degree — nobody has ever really calculated it.
Second, they tell a story to outsiders. “Such-and-such a share of our staff come from 985 universities, such-and-such a share hold master’s degrees” — that sentence isn’t addressed to job seekers. It’s addressed to investors, clients, regulators and local government. It’s a spec sheet for the talent structure, proof that the company is impressive.
But here I want to correct a line people repeat.
It isn’t that “a lot of the work can’t be done by these people.” It’s that a lot of the work was designed so they wouldn’t be needed for it.
Job downgrading and degree inflation happen together: the work gets chopped into repeated motions, judgment is pulled out of it, only execution is left behind; meanwhile the entry bar is raised, because someone has to be filtered out. The result is master’s graduates doing bachelor’s work, bachelor’s graduates doing associate-degree work, and associate-degree holders with no work at all. Plenty of people finished a 985 or 211 bachelor’s and master’s — the state’s tiers of elite universities — and still end their careers in the execution layer.
The real function of degree screening is to supply a reason that looks objective for deciding who gets into the inner circle, at a time when there aren’t enough positions to go around.
A Profit Chain That Nobody Owns
Outsourcing thrives not because it’s especially good for anyone, but because every link in the chain gets something out of it.
- The staffing company: takes a management fee per head — one to two tenths is common in the industry. It produces nothing. What it carries is the legal identity of “employer.”
- HR: hits the headcount target, brings labor cost down, makes the reports look good.
- The department head: their actual span of control grows, and the reporting line becomes “I manage fifty people.”
- The client’s finance department: employee compensation turns into outsourcing service fees. The first is a liability owed to human beings; the second is a procurement line that can be cut.
- The worker: less money, weaker protection, shorter tenure.
Everyone is only responsible for their own cell.
So when something actually goes wrong, nobody owns it. The staffing company says “we executed to the client’s requirements,” the client says “that’s the vendor’s business,” the department says “I was only carrying out company policy.”
Responsibility evaporates in the middle layer. That’s also why this thing can’t be pushed back: there are too many beneficiaries, they’re too scattered, and not one of them thinks of himself as a bad actor — the staffing company says it “creates employment,” HR says it “optimizes the structure,” the department head says it’s “flexible employment.”
Doing harm doesn’t require bad people. It only requires a crowd of people each responsible for exactly one cell.
What Wage Work Actually Accumulates
I’ve been working for wages for more than a decade. The hardest part isn’t that the money is small. It’s looking back and finding nothing left in my hands.
Later I worked out that there are two kinds of capital.
One is firm-specific: things that are only worth money inside this one company. Internal processes, internal systems, internal jargon, the trust of one particular leader, the client relationships specific to this company, the toolset only used here.
The other is general: things that work wherever you go. Transferable skills, a portfolio, clients, a public record, cash flow, assets.
Almost everything wage work accumulates is the first kind.
And the first kind depreciates the day you hand in your notice — some of it goes straight to zero. Being “very proficient” at this company means proving yourself from scratch somewhere else; being “very important” at this company doesn’t produce one line on a résumé.
So the problem isn’t that wage work offers no growth. It’s that the growth from wage work is designed to be non-portable.
A company that lets its employees accumulate transferable capital is training its own competitors. So it doesn’t do that. What it gives you is “very proficient in this position,” not “very valuable in the market.”
My feeling after more than a decade isn’t an illusion. It’s the design objective.
Exit Costs: They Don’t Have to Keep You, Only Make Leaving Unaffordable
The last section said wage work mostly accumulates things you can’t carry out. There’s a harsher layer on top of that: even the little you could carry, they cut off physically.
The day you resign — or before you’ve been fired, when you’ve merely “made the list” — the computer gets locked. No copying files, no deleting anything, and even the documents you wrote yourself can’t be taken out. The company’s explanation is preventing leaks and preventing someone from wiping the database and running.
It sounds reasonable. But what actually gets cut off isn’t the “capacity to damage.” It’s bargaining chips. Client lists, work samples, records of what you did, correspondence — these are the evidence for your next job search and the basis for negotiating price with the next employer. With the chips in your own hands, leaving is a negotiation; with the chips in the company’s hands, leaving is a plea.
The second layer is corporate IM.
WeCom, Feishu, DingTalk — it looks like just switching tools to chat in. But in personal WeChat, the contacts are yours and the history sits on your phone; in corporate IM, colleagues, clients, group chats and message history are all hosted under the company’s tenant. The moment the account is suspended, those people vanish from your address book.
What it does isn’t swap a tool. It re-books “your network” from your asset column into the company’s.
Line these up and they form one line:
- Device locks → cutting off what you can carry out
- Corporate IM → cutting off the relationships you can carry out
- Non-competes and liquidated damages → cutting off the opportunities you can carry out
- Vesting options, a mortgage, consumer loans → cutting off the time you can carry out
All four put together have one name: exit cost.
A company doesn’t need to retain you. It only needs to make leaving unaffordable. Retaining people costs money; making them unable to leave costs far less, and it doesn’t require your consent.
Only then does the “cog” question arrive.
What decides whether you’re a cog isn’t how important you are. It’s how expensive you are to replace. People in the core circle aren’t “irreplaceable,” they’re “expensive to replace”; people on the fringe aren’t “useless,” they’re “cheap to replace.”
So degree, position and title all determine replacement cost, not replaceability. What the company calculates has never been “is this person important” — it’s “how much and how long would it cost to swap him out.”
Even a CTO works the same way. His irreplaceability comes from board trust and the capital-markets narrative, not from technical skill. When the trust shifts, the technical skill goes to zero overnight.
And there’s one boundary a lot of people don’t know, which is why they get frightened into silence:
A company has the right to manage its own devices and its own data. It does not have the right to hold a person back, withhold wages, confiscate personal belongings, or refuse to process departure paperwork when you leave. The equipment is theirs. Your person and your pay are yours.
Involution: The Man Paying Out of Pocket to Work
A few days ago I came across a post. The author complained that the AI tools the company provides are slow, and said he was buying his own quota with his own money to use at work. The justification was one line: all for efficiency.
The poll below it went mostly against him. Seeing the result, I actually felt relieved.
But what I cared more about was something else: why was this even a thing that could be put up for discussion, and still pull nearly half the votes in favor.
He isn’t necessarily a bad person. He simply hasn’t worked out the arithmetic — the money he spends ends up being paid by everyone.
Today he funds his own efficiency; tomorrow the supervisor asks: why can the others do it without spending? So everyone has to start paying too, and wages don’t rise by a cent. The competition didn’t raise total output at all. It only raised the price of participating. That’s the plainest definition of involution.
A lot of people call this kind of person a scab. The original meaning of a scab is someone who breaks collective action. What he’s actually doing is something else — raising the bar for everybody, and then tripping over that bar himself.
The question worth asking isn’t whether he should spend the money. It’s whose money it was supposed to be in the first place.
I’ve read my own contract. Everything produced during working hours belongs to the company. Patents belong to the company, code belongs to the company, and so do the documents I’ve written and the proposals I’ve built.
Article 6 of the Patent Law: the right to apply for a service invention belongs to the employer. Article 13 of the Regulations on the Protection of Computer Software: the copyright in software developed as part of one’s duties belongs to the employer. Service works under copyright law are a bit more convoluted, but once you land on code and drawings the outcome is much the same.
If one hundred percent of the return goes to the company, then the cost should too. That’s consideration, not morality.
Buying your own computer, your own software, your own quota to work for the company is, in essence, the employee fronting the means of production while the company collects the surplus. It sounds like efficiency-mindedness. In practice it’s giving yourself a pay cut, and pressing down your colleagues’ price at the same time.
Then there’s “doing it well or badly makes no difference.”
That line is usually thrown at employees as an accusation — lying flat, no professional spirit. But it’s an outcome the system produced.
When pay is decoupled from output, the rational actor’s optimal strategy is the minimum compliant effort. Economics calls it the principal–agent problem; put plainly: whatever you pay me for is whatever I’ll put effort into.
There’s another layer, called efficiency wages. A wage above market rate never buys only hours; it also buys the “I’m willing to go a bit further” part of the willingness. I came in at half the salary, which means the company gave up that part of its own accord on the day I signed.
It broke the implicit contract first, then demands that I perform my side unilaterally. That isn’t dedication. That’s welshing.
So “working only my eight hours” isn’t lying flat. It’s strict performance of the contract.
The contract specifies hours. It doesn’t specify “extra enthusiasm.” The company has performed the minimum of its side — reduced pay, thin benefits, dismissible at any time — and I perform the minimum of mine. That’s symmetry.
A labor contract is an incomplete contract by nature. It sets time, place and basic obligations; the rest gets filled in by reciprocity. Reciprocity presupposes fairness. When fairness goes, reciprocity stops. What’s left is one sentence: however much money, that much work.
As for “dying at your post.”
Article 15 of the Regulations on Work-Related Injury Insurance: where a person dies of a sudden illness during working hours and at the workplace, or dies within forty-eight hours despite rescue efforts, it is treated as a work-related injury.
The phrase “treated as” matters. This was never standard work-injury coverage to begin with — it’s compensation the law grants on top.
And in an outsourcing or labor-dispatch arrangement, you first have to establish whose labor relationship it was, then go through the recognition process, then through compensation. Every step can be dragged out; every step can be litigated.
That programmer’s widow fought several lawsuits. Not because she liked litigation, but because every door required her to prove the same thing again: my husband was this company’s person.
When the company wanted him to work overtime, he was “our person.” When he died, he was “the vendor’s person.”
Last, there’s the line “worst case, I’ll sell the apartment and go back home.”
That isn’t defeatism. It’s an exit option.
Someone who has an exit option naturally holds more bargaining power; someone without one can only accept “sign or lose the job.” Read the earlier list of exit costs backwards and you get this: the more able you are to walk, the less it dares to mistreat you.
But “Control Your Desires” Isn’t the Answer
Once you see all this clearly, one conclusion comes easily: spend less, save up, get out early.
The direction is right. The sequence is very often reversed.
Run the numbers.
- Ten thousand a month, a 30% savings rate, that’s 36,000 a year.
- To reach 360,000 — roughly three years of living expenses — takes 10 years.
- At the same 30% savings rate on twenty thousand a month, you save 72,000 a year, and the time is cut straight in half.
The savings rate multiplied by the slope of your income curve is the real speed of primitive accumulation. Frugality has a floor — living costs can’t be squeezed to zero. Income has neither a floor nor a ceiling.
Then look at that oft-quoted 4% rule: spending 60,000 a year requires 1.5 million in principal. Saving 3,000 a month takes 40 years.
This isn’t meant to discourage anyone. It’s meant to be clear: completing primitive accumulation out of money you scrape together doesn’t work arithmetically.
Thrift is defense, not attack. Defense keeps you from falling further; it can’t move you up.
And there’s a trap here.
“The way out of wage work is to control your desires” happens to be the version of the story this system likes best.
Because it translates a structural problem into a problem of personal discipline: you failed to accumulate because you drank bubble tea, upgraded your phone, and weren’t self-controlled enough. The blame for failure lands back on you, and the system bears none of it.
It’s the same translation trick as calling shadow teachers “a teaching-staff reserve.”
The Real Core Is Distribution
At this point the thing has to be said all the way down: everything above is phenomenon. Underneath the phenomena there’s only one thing — distribution.
Look at two sets of numbers.
On the basis used by the UBS (formerly Credit Suisse) Global Wealth Report, the top 10% of Chinese households hold about 70% of household wealth, and the bottom 50% hold about 6%.
On the basis of the flow-of-funds accounts published by the National Bureau of Statistics, labor compensation as a share of GDP in China has long sat around 50%, while most advanced economies sit around 60%.
Put the two sets together and they say the same thing: more and more of what growth produces lands in the capital column instead of the labor column.
Distribution has three layers.
- Primary distribution: the market — the shares labor and capital negotiate out.
- Secondary distribution: taxes, social insurance, transfer payments — adjusting the result of primary distribution once more.
- Tertiary distribution: charity and donation, voluntary.
The problem sits mainly in the second layer, and it runs in the reverse direction.
Individual income tax is only around seven percent of national tax revenue, and the main body paying it is salaried workers — wages are withheld at source, not a cent escapes. Genuine property taxes are essentially absent: the real estate tax is still in pilot, there’s no inheritance tax, and capital gains bear a very low effective burden in practice. The top marginal rate on wages is 45%; dividends are 20%. Social insurance contributions have a capped base, and a large number of people contribute at the minimum base, so the actual effect is regressive.
So in many places secondary distribution isn’t “taking from the fat to feed the lean.” It’s taking from the lean to feed the fat.
This isn’t something individual effort can bend back. The rules themselves lean in one direction, and they lean very steadily.
The Late-Ming Ledger, and the Price of Not Reforming
The problem is this obvious, so why can’t it be fixed?
Because reforming distribution means making the people who write the rules cut their own flesh.
The end of the Ming is one long ledger entry.
In the seventeenth year of Chongzhen, military pay was in crisis and the court organized a subscription drive. Pressed all the way up to Zhou Kui, the emperor’s father-in-law, for a hundred thousand taels, he pleaded poverty and finally put up thirteen thousand. The entire drive is said to have raised somewhere around two hundred thousand taels.
After Li Zicheng entered the city, he ran a campaign of recovering “stolen” wealth to fund the army, with Liu Zongmin in charge and the ankle press applied. The figure squeezed out of the nobility, imperial in-laws and officials is commonly given as seventy million taels.
Two orders of magnitude.
This passage is usually told as a morality play: the powerful were rich and heartless, they watched the dying and did nothing, and that’s how the Ming ended.
Moral condemnation isn’t wrong, but it fails to explain one thing: why everyone knew the city would fall and they’d all die together, and still nobody paid.
The real problem was fiscal extraction capacity: imperial-clan stipends consumed the annual revenue of several provinces; the gentry enjoyed exemptions and degree holders didn’t pay; the Three Levies surcharges fell entirely on owner-farmers without degrees; and on top of that came the global silver contraction of the seventeenth century. The money was there. It simply couldn’t be collected.
The reason it couldn’t be collected isn’t moral. It’s that the power to tax was held by the people who were supposed to be taxed.
One layer further in is the collective-action trap. If I pay and others don’t, my money is wasted when the city falls; if I don’t pay and others do, I get it for free. So for every individual, the most rational choice is not to pay. The result is that nobody pays, and then everyone is finished together.
What gets called “human nature” here isn’t the cause. It’s the effect. The rules are what made the bad choice into the rational choice.
But the most powerful part of this story is the ending.
Those officials believed they were protecting their money. Then Li Zicheng entered the city, the ankle press came out, and the money was taken anyway — taken in far greater amounts, and in a far more humiliating way. Zhou Kui’s entire household was confiscated and he himself was tortured to death.
They didn’t preserve their wealth. They only exchanged a worse way of losing it.
That’s the real cost of “not reforming”: not successfully defending vested interests, but trading one controlled concession for one uncontrolled reckoning.
Their Refusal to Reform Doesn’t Mean the Ledger Stops Changing
So I’m not counting on anyone’s conscience awakening. Concessions have never come from enlightenment. They’ve been forced out.
Historically there are only three routes to changing distribution:
- Crisis compulsion: the finances collapse, the system can’t hold, and change is forced. The cost is the highest — that’s the late-Ming road.
- Institutional competition: an alternative is on the table, people and capital can leave, and the rules have to follow.
- Rising bargaining power: the people at the bottom organize and make the cost of not reforming exceed the cost of reforming. This is the only bloodless path, and it’s the hardest.
The first two don’t require anyone to become better. They only require the result of the calculation to change.
“No hope” is honest as a feeling. As a judgment it leaves out one link: their refusal to reform doesn’t mean the ledger stops changing.
“Getting Out of Wage Work” Needs Redefining
Getting out of wage work doesn’t mean not working. It means no longer depending on a single employer.
To actually be able to walk, three things have to be in place at once.
First, a buffer. Six to twelve months of cash flow without a job. This is the base of your courage to say “no,” and it’s your only chip against “sign or lose this job.” People without a buffer have no seat at the negotiating table.
Second, an exit. A line of income that doesn’t come from an employer. Even a small one — as long as it’s real and repeatable, the nature of your position changes: you go from standing on one leg to standing on two.
Third, transferable capital. Work samples, clients, reusable skills, a public record. Whatever can’t be carried out, invest in sparingly; whatever can be carried out, do it even if it doesn’t pay in the short term.
The order is: frugality buys a buffer, the buffer buys optionality, optionality buys income, income buys capital.
The hardest of the three is the second. Because it has to start growing while you still have a job; you can’t wait to be laid off and begin from zero. And early on it will definitely be slow, ugly and shapeless — everything portable grows that way.
Last
Back to that desk. The person next to me signed a regular contract; I signed an outsourcing contract. We do the same work, and we are not the same kind of person — at least not on paper.
The strongest thing about this system isn’t how far it pushes wages down. It’s that it makes every single person believe the problem is inside themselves: I wasn’t hardworking enough, I wasn’t disciplined enough, my degree wasn’t good enough, my luck was bad.
It doesn’t even need you to believe in it. It only needs you to stop believing in everyone else.
In the last piece I wrote: give me back the sheet of paper I signed, and I’ll stand up on my own.
This one adds a line: one person standing up doesn’t count as standing up. The contract is handed to one person, but the weight comes down on a whole layer of them. What has ever changed that piece of paper is never the people who signed it. It’s the number of people who refuse to sign becoming large enough.
评论
0 条评论