China compensation benchmarking: a 2026 pay structure guide

What the national series reports

The National Bureau of Statistics annual wage release is the backbone of any credible China benchmark. For 2025 it reported average annual wages of ¥129,441 for employees in urban non-private units and ¥71,590 in private units, with nominal growth of 4.3 and 3.0 per cent respectively.

It ranks industries on a consistent basis: information transmission, software and information technology services first at ¥248,752, finance second at ¥211,164 and scientific research and technical services third at ¥182,064. Manufacturing pay grew faster than the overall average in both ownership categories, at 5.2 and 6.4 per cent.

Regionally it reports Guangdong at ¥143,284 for non-private units and the eastern region at ¥149,902. All of this is dated, citable and reproducible, which distinguishes it from most China pay commentary in circulation.

Ownership type dominates every comparison

The gap between urban non-private and private units is the single largest structural pay variable in the country, and the official statistics separate them precisely because of it.

Foreign-invested enterprises form a third category with different pay structures, different equity practice and different mobility patterns from either domestic grouping, and they are the comparator set most international employers actually care about.

A benchmark that does not state which of the three its comparator set is drawn from is not reporting a market rate. It is averaging three different markets into a figure that is wrong for all of them, and the error is large rather than marginal.

City, region and where the numbers come from

The national release reports by province and by region. It does not report by city, and the cities are where the hiring happens.

City-level pay therefore has to come from municipal statistical yearbooks, cited individually, or from primary research with its comparator set named. Substituting the provincial figure for a city one is the most common silent error in China benchmarking, and it is directional: it flatters inland cities and understates the coastal ones.

Municipal talent subsidy programmes add a further layer that no wage series captures, because they attach to individuals rather than to roles and they move senior specialists between cities for policy reasons rather than market ones.

The components the series cannot see

Payment pattern. Chinese salaries are usually monthly figures paid across thirteen or more instalments, and the number of payments is a real variable rather than a rounding detail. The national statistics report annual totals, which conceal it entirely.

Social insurance and housing fund. Employer contribution rates and contribution bases are set locally, which means the fully-loaded cost of an identical salary differs by city. Quoting a salary without the local on-cost understates the difference between two apparently similar locations.

Equity and long-term incentive. Material in listed technology companies and in the China arms of listed multinationals, largely absent elsewhere, and in no national wage series. It is established from filings and primary research or marked unestablished.

What a defensible China benchmark looks like

Two clearly separated layers. The official layer, cited to the release and its reference year, covering ownership type, sector and region. The researched layer, with its comparator set named and its confidence and observation count stated, covering city, role level, payment pattern and equity.

Ownership type stated on every comparator. City named rather than province. Local social insurance and housing fund rates quoted with the city and effective date where a fully-loaded cost is given.

And anything not established named as unestablished. In this market that will frequently include a city figure, and a benchmark that quietly uses the provincial number instead has made a choice it did not disclose.