What actually makes mainland China different
Five structural facts that change the research, not five facts about the economy.
Ownership type is the biggest pay variable in the country
The National Bureau of Statistics reports average annual wages separately for urban non-private and private units, and the gap is not marginal. For 2025 the figures were ¥129,441 and ¥71,590 respectively, with nominal growth of 4.3 per cent and 3.0 per cent.
No other market we cover has a single structural variable that moves pay this far. A China benchmark that does not state whether its comparator set is state-linked, domestic private or foreign-invested is not telling you anything usable.
Three sectors sit far above the rest
On the same 2025 release, information transmission, software and information technology services led all sectors at ¥248,752, finance followed at ¥211,164, and scientific research and technical services came third at ¥182,064. Manufacturing pay grew faster than the national average, at 5.2 per cent in non-private units and 6.4 per cent in private ones.
That ordering is stable enough to design around, and it explains most cross-sector moves we are asked to research.
Cities are separate labour markets with separate policy
Beijing, Shanghai, Shenzhen, Guangzhou, Hangzhou, Suzhou, Chengdu, Hefei and Wuhan each hold different industrial capability, and municipal talent subsidy programmes actively move senior specialists between them.
The national wage release reports by province and region — Guangdong at ¥143,284 for non-private units, the eastern region at ¥149,902 — but not by city. The city-level numbers have to come from municipal yearbooks, and we say so instead of presenting a provincial figure as a city one.
Intermediary activity is licensed; information work is not
Under the Interim Regulation on the Human Resources Market, a human resources service licence is required for intermediary activity — recommending candidates to employers, running job fairs and similar. Collecting and releasing information about labour supply and demand, and providing guidance, is treated differently, and pure consulting, assessment and training generally sits under a filing rather than a licence.
That line is exactly where Audentia operates. We produce research and analysis. We do not introduce candidates to employers for a fee and we do not charge anything contingent on a hire.
Personal information has a regulated route out of the country
Under the Personal Information Protection Law and the Cyberspace Administration of China’s provisions, personal information leaves the mainland through a security assessment, a standard contract with filing, or certification, unless an exemption applies. The certification measures took effect on 1 January 2026, and the regulator published further question-and-answer guidance on administrative policy in July 2026.
Reported volume thresholds differ between secondary summaries, and data classified as important triggers assessment regardless of volume. We agree the transfer route with you in writing before fieldwork, and where the right answer is an aggregated or in-country deliverable we say so rather than designing around it.
What we deliver in China
Scoped to the decision rather than sold as a package.
Candidate identification and longlisting
Researched longlists against your brief, built from Chinese-language sources rather than exported from a database, with the evidence behind each name.
Delivered in your format, into your systems, ready for your consultants to work rather than ready to be re-keyed.
Screening and structured write-up
Structured screening against your criteria, with the answers recorded in the same shape for every candidate so a selection decision is comparable and defensible.
Conducted in the working language of mainland China and written up in English, with the original retained.
Credential, qualification and right-to-work verification
Checking qualifications and professional registrations against the Chinese registers that hold them, and establishing work authorisation position, with the date of every check recorded.
This is the part most often done badly at distance, because it needs someone who can read the register and knows which register is authoritative.
Compliance file and audit trail
A complete, consistent file per candidate: what was checked, when, by whom, against what source, and what the result was.
Built to survive an audit at your end rather than merely to look tidy. In licensed markets that is not a nicety.
Pipeline maintenance and data hygiene
Keeping your existing talent pools current: de-duplication, entity resolution across corporate name variants, re-verification on a schedule and decay flagging.
A pipeline nobody maintains is a liability that looks like an asset on the balance sheet of your ATS.
Coordination, scheduling and desk cover
Interview scheduling across time zones, candidate communication and desk cover while your office is closed — which for a European or American firm hiring into mainland China is most of the working day there.
Our out-of-hours cover exists precisely for this gap.
Most engagements use three or four of these rather than all of them. The scoping call establishes which questions you actually need answered, and the written scope says what is in and what is deliberately out.
Pay and labour cost in China
A benchmark that does not state its basis is not a benchmark.
| Component | What it is | Where the evidence comes from |
|---|---|---|
| Base salary and the payment pattern | Usually a monthly figure, frequently paid across thirteen or more instalments. The number of payments is a real variable, not a rounding detail. | Primary research. The national statistics report annual totals, which conceal the instalment pattern entirely. |
| Ownership type of the employer | Urban non-private versus private is the structural divide the official series is built around, and foreign-invested enterprises behave differently again. | National Bureau of Statistics annual wage release, which reports non-private and private separately for the nation, by sector and by region. |
| Sector | Information technology, finance and scientific research sit well above the national average; the ordering has been stable across recent releases. | National Bureau of Statistics annual wage release, by industry category. |
| City and region | The eastern seaboard sits above the national average and inland centres below it, with municipal subsidy programmes layered on top for senior specialists. | Provincial and regional figures from the national release; city-level figures from municipal statistical yearbooks, cited individually. |
| Social insurance and housing fund | Employer contribution rates and contribution bases are set locally, so the fully-loaded cost of the same salary differs by city. | Published municipal rates and bases, quoted with the city and the effective date. |
| Equity and long-term incentive | Material in listed technology companies and in the China arms of listed multinationals; largely absent elsewhere. | Company filings and primary research. Not captured in the national wage series. |
China has one of the better national wage series of any market we cover and one of the largest gaps between that series and what a hiring decision needs. The series tells you, credibly and on the record, how sectors and ownership types compare. It will not tell you what a senior autonomous-driving engineer in Shanghai is paid.
So we report in two layers. The official layer is cited, dated and reproducible. The researched layer names its comparator set, states how many observations it rests on and carries a confidence grade. We do not blend them into a single number, because the moment you do, nobody can tell which part of it is evidence.
Regional and city markets in China
We report by market, not by national average.
| Market | What concentrates there | What a researcher has to watch |
|---|---|---|
| Beijing | Internet platforms, artificial intelligence research, state-owned group headquarters, regulators, universities and the policy ecosystem. | The densest concentration of research-grade AI capability and of state-linked employers in the country. Ownership type matters more here than anywhere, and so does the difference between an academic affiliation and an operating role. |
| Shanghai | Financial services, automotive and electric vehicles, semiconductors, pharmaceuticals, and the China headquarters of most multinationals. | The deepest multinational population on the mainland, which means a large share of candidates are legible to foreign employers — and competed for by all of them at once. |
| Shenzhen | Hardware, communications equipment, consumer electronics, electric vehicles, batteries and a very large software engineering base. | The fastest-moving market in the country and the most closely linked to Hong Kong. Engineering populations here are young, mobile and poorly captured by any dataset more than a year old. |
| Guangzhou and the Pearl River Delta | Automotive, trade and distribution, chemicals, consumer goods and a dense manufacturing supplier base. | Functions as one labour market with Shenzhen and Dongguan for many engineering roles, and as a separate one for commercial roles. The boundary has to be set in the scope. |
| Hangzhou, Suzhou, Wuxi and Nanjing | E-commerce and platform technology in Hangzhou; semiconductors, precision manufacturing and biopharma across the Jiangsu cluster. | Municipal subsidy programmes compete openly for senior specialists here, so individual mobility is partly a function of policy rather than of employer action. |
| Chengdu, Xi’an, Wuhan, Hefei and the inland centres | Aerospace, optoelectronics, displays, batteries, automotive and a growing software base, supported by large university populations. | Where cost-driven relocation of engineering capability has been most pronounced. Pay sits materially below the eastern seaboard, and retention behaves differently. |
Sectors we cover in China
Deepest where the research is hardest and the stakes are highest.
- Technology, software and artificial intelligence — the highest-paid sector in the official wage series and the deepest engineering population in the market.
- Automotive, electric vehicles and batteries — vehicle makers, battery and cell manufacturers, autonomous driving software and the supplier tiers.
- Semiconductors, electronics and displays — design, equipment, materials and fabrication, concentrated in Shanghai, Shenzhen and the Jiangsu cluster.
- Industrial manufacturing and engineering — where wage growth ran ahead of the national average in both ownership categories in 2025.
- Financial services — the second highest-paid sector nationally, spanning banking, insurance, asset management and the foreign-invested joint ventures.
- Pharmaceuticals and life sciences — domestic innovators, contract research and the China operations of global pharma.
- Supply chain, logistics and trade — a large and specialised population, increasingly focused on cross-border and nearshoring capability.
- Energy, renewables and materials — solar, wind, grid and the chemicals and materials base underneath them.
- Consumer and retail — domestic brands and the China commercial functions of international groups.
What may lawfully be recorded, and moved, in mainland China
Chinese research sits under the Personal Information Protection Law, alongside the Cybersecurity Law and the Data Security Law. For talent research the binding question is usually not whether a professional fact may be recorded but whether, and how, a dataset containing personal information may be transferred out of the mainland.
The Cyberspace Administration of China operates three routes: a security assessment, a standard contract accompanied by a filing, or certification. Provisions issued in March 2024 introduced exemptions, measures governing the certification route took effect on 1 January 2026, and the regulator published further guidance on administrative policy in July 2026. Reported volume thresholds vary between secondary summaries and we do not repeat them as settled; data classified as important triggers the assessment route regardless of volume.
What this means for an engagement. We agree the transfer route in writing before fieldwork begins, with your counsel in the loop, and we design the deliverable to fit it. For a competitor structure or a market sizing, an aggregated deliverable that contains no personal information often answers the question completely and removes the transfer problem rather than managing it. Where named individuals are genuinely required, we say what route that needs and what it will take before you commit.
On the service line: the Interim Regulation on the Human Resources Market makes intermediary activity — recommending candidates to employers, running job fairs — a licensed business, while collecting and publishing labour supply and demand information is treated differently and consulting, assessment and training generally sit under a filing. Audentia produces research and analysis. We do not introduce candidates for a placement fee and we charge nothing contingent on a hire.
One further constraint worth stating because it affects what a map can promise: post-employment non-competition restrictions are enforceable in China only for a limited period and only where the former employer continues to pay compensation, and the Ministry of Human Resources and Social Security issued compliance guidelines on their implementation in September 2025. A named individual being findable is not the same as that individual being free to move.
How a Chinese recruitment support project actually runs
A written scope before any fieldwork begins.
Understand your process, not just your vacancies
- What we establish
- How your team actually works, what your systems expect, what your clients are promised and where the bottleneck really is.
- Why it matters
- Support that does not fit the existing process creates work instead of removing it.
Agree the standard and the capacity
- What we establish
- What good looks like per deliverable, what volume we are resourcing for, and what the escalation route is.
- Why it matters
- A fixed monthly fee only works when both sides have agreed what it buys.
Set up white-label delivery
- What we establish
- Your templates, your branding, your confidentiality terms, your systems and your file conventions.
- Why it matters
- Your client should see your firm’s work, not a supplier’s.
Run the research and screening
- What we establish
- Longlists, screens and write-ups produced in the working language of mainland China and delivered in English to your format.
- Why it matters
- Language capability is the thing you are actually buying in this market, and it cannot be substituted by a translation layer.
Verify and build the compliance file
- What we establish
- Credentials and work authorisation checked against authoritative registers, with dates and sources recorded per check.
- Why it matters
- In licensed markets the file is the defence, and it has to exist before anyone asks for it.
Review, measure and adjust
- What we establish
- Measured throughput and quality against the agreed standard, reviewed on a fixed cycle.
- Why it matters
- Support that is never measured drifts, and both sides notice too late.
Recruitment Support, and what it is not
These get bought as if they were interchangeable. They are not.
| Recruitment support (this page) | RPO | Additional in-house headcount | |
|---|---|---|---|
| What is outsourced | Research, screening and administration | The hiring process, end to end | Nothing |
| Who owns the candidate relationship | You | The provider, usually | You |
| Commercial model | Fixed monthly fee for capacity | Per hire or managed service fee | Salary and on-costs |
| Local-language capability | By design, per market | Varies by provider | Whatever you hire |
| Scales down again | Yes, on notice | Slowly, contractually | No |
| White-label | Yes | Sometimes | Not applicable |
| Best for | Capacity and language capability without headcount | Handing over a whole function | Core, continuous, strategic work |
Language is the other thing worth being explicit about. Mainland Chinese research runs in Simplified Chinese. The public record — company registry filings, patent and academic material, industry press, listed-company disclosure — is overwhelmingly Chinese-language, and the English layer over it is partial and lags.
The transliteration problem is sharper here than in Japan. Pinyin collapses distinct characters onto identical romanised strings, so a common name in pinyin can correspond to dozens of different people, and matching on it produces confident-looking nonsense. Corporate names compound it: a group’s registered Chinese name, its English trading name and its subsidiaries’ names frequently share no common element.
We research and reconcile on the Chinese form, hold the characters in the dataset alongside the romanisation, and deliver analysis in English. The market term for this work is 人才地图; the commercial market for contingent placement is 猎头, and the two are not the same purchase.
What recruitment support costs in China
Fixed project fees, scoped on the research rather than on the salary of a role.
| Engagement | What it covers | Typical duration |
|---|---|---|
| Scoping and feasibility | An honest read on whether the question is researchable in mainland China, what the sources will support and what the limits are. Written, before you commit to anything. | Days, at no charge |
| Single-decision project | One function or one defined population in mainland China, scoped to a named business decision. | Four to six weeks |
| Multi-function or multi-city project | Several functions, or several Chinese markets held to one definition so the comparison is like for like. | Six to ten weeks |
| Multi-market programme | China alongside our other eleven markets, with the population definition held constant across all of them. | Scoped per programme |
| Refresh of an existing dataset | Re-verification of a population we built previously, with decay reported against the original. | One to three weeks |
Every engagement is a fixed fee agreed in writing before work begins. There is no commission, no percentage of salary and no success fee on any service we sell — which in mainland China is a regulatory position as well as a commercial one, because fee-charging placement is a licensed activity there and research is not.
Fees are quoted as a fixed project fee and can be contracted in pounds sterling, US dollars or renminbi. Our published fee bands sit on the recruitment support cost page; the band that applies depends on the population size, the number of fields to be verified and the language coverage required, and the scoping call establishes all three before anything is quoted.
What drives the fee in mainland China is verification depth, not headcount. A population of 200 names corroborated on six fields costs more than 600 names corroborated on two, and the scope document says which you are buying. The fee comparison calculator shows what a fixed research fee comes to against a contingency or retained model.
Recruitment Support in China — FAQs
How do you handle candidate data for mainland China?
By agreeing the position before the engagement starts rather than discovering it mid-flow. Personal information leaves the mainland through a security assessment, a standard contract with filing, or certification, unless an exemption applies.
Where an engagement involves your own employees’ data under lawful employment rules the position differs from third-party research, so we set it out explicitly with your counsel rather than assuming one route covers both.
Can you screen in Mandarin?
Yes, and for most mainland roles it is the only way the screen produces anything reliable. We screen in Chinese and write up in English with the original retained.
Self-reported English level on a Chinese CV is among the least reliable fields on it, so where a role genuinely needs business English we test it explicitly.
How do you verify Chinese qualifications?
Against the authoritative register for that qualification, with the date of the check recorded. Degree and professional qualification verification in China has established channels and we use them.
Where verification is not possible from an authoritative source, the file says so rather than carrying a tick we cannot support.
Do you charge per hire?
No. The Interim Regulation on the Human Resources Market makes intermediary activity a licensed business, and we deliver research and administration.
Fixed monthly fee for an agreed capacity, with nothing contingent on anyone being hired.
Can you check non-compete exposure before we invest in a candidate?
We can establish and record what is knowable, which is usually enough to avoid wasted process. Post-employment restrictions are enforceable for a limited period and only where the former employer keeps paying compensation.
The Ministry of Human Resources and Social Security issued compliance guidelines in September 2025. We record the facts; the legal assessment is for your counsel.
Will you work white-label?
Yes. Your branding, your templates, your confidentiality terms, no approach to your client market in our own name, and no re-use of research done for you.
For China engagements the data handling terms are part of that agreement rather than a separate conversation.
What we will not tell you about China
Because a research supplier that never says "we do not know" is not worth buying from.
- City-level average wages for Beijing, Shanghai and Shenzhen are not in the National Bureau of Statistics national release. Where a decision needs them we cite the relevant municipal statistical yearbook, and where we have not obtained one we say the figure is not established rather than substituting the provincial number.
- Widely circulated figures for the shortage of new-energy vehicle and artificial intelligence professionals are projections, and several originate with firms that sell recruitment services. We cite them as projections, attributed, or not at all.
- There is no reliable published measure of senior executive mobility between ownership types. It is one of the most commercially interesting questions in the market and one of the least documented, and we treat it as a research question rather than a statistic.
Primary sources and how we check
- Average Annual Wages of Employees in Urban Units in 2025National Bureau of Statistics of China
- National statistics, annual data and statistical yearbooksNational Bureau of Statistics of China
- Employment, human resources market and labour policyMinistry of Human Resources and Social Security
- Personal Information Protection Law of the People’s Republic of ChinaNational People’s Congress
- Cross-border data transfer policy and guidanceCyberspace Administration of China
- Doing business, investment and regulatory guides for ChinaDezan Shira & Associates
- Investment environment and sector guidesJETRO
Figures quoted on this page were checked against the sources above on 30 October 2026. Statutory rates, thresholds, licence conditions and quotas change, and several of these series are revised after first publication: where a number drives a budget or a legal obligation, confirm it against the primary source before you rely on it.