Decision desk
Find the structure you should test first
Answer one question at a time. The desk identifies a starting structure to test and the legal questions still to record — not a tax or approval outcome.
Question 1
What will the business initially do overseas?
At a glance
Compare the main overseas structures
One screening table. It identifies a route to test and a reason to escalate; it does not rank jurisdictions or predict tax outcomes.
| Structure | Best starting use case | Main advantage | Main limitation | When to escalate |
|---|---|---|---|---|
| Direct contracting | Cross-border sales only; no enduring local footprint | Lowest new entity burden | Parent remains the contracting party | Local registration, PE, importer, licence, customer or enforcement need |
| Branch / establishment | Local registration needed without a new legal person | Registered presence without a separate company | Liability usually stays close to the parent | Liability, governance, regulated activity or ring-fence |
| Local subsidiary | Staff, assets, licences, local contracts or liability separation | A defined local operating home | Does not itself clear PRC outbound or funding | Ownership, funding, investor, group or expansion needs |
| HoldCo + subsidiary | A documented group-level function exists | Separates ownership, finance, governance or exit | Extra substance, tax and maintenance cost | If the holdco’s purpose cannot be stated in one sentence |
| Multi-tier structure | Distinct countries, investors, assets, financings or exits | Separation where functions truly differ | Highest coordination and evidence burden | If any layer lacks a named owner and function |
Direct contracting
- Use case
- Cross-border sales only
- Advantage
- Lowest new entity burden
- Limitation
- Parent remains the party
- Escalate
- Local footprint, PE, licence or enforcement
Branch / establishment
- Use case
- Registered presence, no new legal person
- Advantage
- Local registration without a subsidiary
- Limitation
- Liability usually close to the parent
- Escalate
- Ring-fence, licences, governance
Local subsidiary
- Use case
- Sustained people, assets or local contracts
- Advantage
- Defined operating home
- Limitation
- Incorporation is not funding or ODI clearance
- Escalate
- Holdco or group functions
HoldCo + subsidiary
- Use case
- Documented group-level role
- Advantage
- Ownership, finance, governance or exit
- Limitation
- Substance and maintenance cost
- Escalate
- Unclear purpose
Multi-tier structure
- Use case
- Genuinely distinct functions
- Advantage
- Separation where needed
- Limitation
- Highest evidence burden
- Escalate
- A layer without a job
Operating model
Do you need an overseas entity at all?
Start with the activity, not the jurisdiction. A new overseas company is not the default answer to every cross-border sale.
Direct contracting
Where the Chinese company can sell or contract without an enduring local footprint, test the commercial route first. Registration, importer status, consumer law, permanent establishment and enforcement still need a fact-specific review. Direct contracting is a starting test, not a finding that no local vehicle will ever be required.
Branch or establishment
A branch or establishment is worth testing where a registered local presence is needed but separate legal personality is not the decisive requirement. Destination labels differ. Parent liability, tax, licensing and reporting usually remain close to the Chinese company. Do not treat a branch as a lower-risk shortcut.
Local subsidiary
A local subsidiary is commonly the next model once people, assets, licences, substantial local contracts or liability separation need a defined operating home. If the business needs a sustained operating footprint, test a subsidiary before adding extra holding layers. Incorporation does not itself resolve PRC outbound approvals, foreign-exchange, tax, data, technology or destination-law workstreams.
Holding layer
Do you actually need a holding company?
Add a holdco only when it performs a defined ownership, financing, governance, investor, asset, JV or exit function. Seat choice (including Hong Kong or Singapore) comes after that function is written down.
Functions that may justify a holdco
- Holding several operating companies
- External investors
- Joint-venture ownership
- Financing
- Regional governance
- Asset or IP separation
- Country-risk separation
- A planned sale or IPO preparation
Warning signs of unnecessary complexity
- The purpose cannot be explained in one sentence
- It exists only because “everyone uses one”
- It duplicates another entity’s role
- Funding routes are unclear
- Governance authority is unclear
- Substance has not been considered
- Compliance burden exceeds commercial value
China-side workstream
What must be cleared on the China side?
Keep this executive-level. A destination certificate of incorporation is not a PRC outbound outcome. Use the specialist ODI and funding guides for forms and clocks.
Corporate authority
Board and ownership records must support the overseas activity before anyone incorporates.
ODI classification
Direct and indirect paths still need a facts schedule that matches control and funding. See ODI filing and overseas funding requirements.
Funding and FX
Test the lawful remittance route before formation. See the outbound funding programme.
Connected controls
People, data, technology, goods and documents leaving China are separate workstreams. See cross-border data and technology.
Destination dependencies
Licensing, screening, employment and tax sit with destination counsel. Do not copy one country’s branch rule worldwide.
Coordination
Make the two-jurisdiction handoff visible
Action plan
Before you incorporate
- Define the overseas activity. Customers, people, assets, licences, data and banking.
- Identify the simplest viable operating model. Direct contracting, branch, then subsidiary.
- Document every entity’s purpose. One sentence each, or drop the layer.
- Map ownership and money. Equity, debt, guarantees, services, royalties, dividends.
- Clear regulatory workstreams. PRC outbound and destination pre-clearance on actual facts.
- Stress-test the future structure. Investor entry, financing, sale, listing, wind-down.
Triage
Common structuring mistakes
Jurisdiction first
Choosing Hong Kong or Singapore before the operating model.
“Everyone has a holdco”
Creating a holding company because it is common market practice.
Form then fund
Incorporating before the China-side funding route is tested.
Certificate as approval
Treating incorporation as PRC outbound or destination regulatory clearance.
Ignoring local law
Designing the chart without destination licensing, tax or employment advice.
Launch-only design
Ignoring investor entry, distributions and exit.
Quick answers
Frequently asked questions
Is there a best overseas holding jurisdiction for Chinese companies?
No. A jurisdiction should follow operating, ownership, funding, governance, investor, asset, JV and exit facts. This page does not rank seats or predict tax outcomes.
When does a Chinese company need an overseas subsidiary?
Usually once staff, assets, licences, substantial local contracts or liability separation require a defined operating home. That remains a fact test for destination counsel.
When is a holding company useful?
When it performs a documented group-level function that the operating company cannot. If the function is unclear, test parent-to-subsidiary first.
Is Hong Kong or Singapore better for a Chinese company’s holdco?
Neither is automatically better. Open the Hong Kong vs Singapore holding structures comparison only after the holdco role is defined.
Does forming an overseas company require ODI approval or filing?
Incorporation abroad is not itself a PRC outbound determination. Classify the investment path on the actual facts and use the ODI filing guide rather than this page as a filing manual.
Can a Chinese company fund an overseas subsidiary immediately after incorporation?
Not automatically. Test the funding and foreign-exchange route before treating the certificate as authority to remit.
Should each country have its own subsidiary?
Only where the country, activity, licence, people or liability facts require a separate operating home. Extra entities need a written purpose.
When should an existing overseas structure be reviewed?
After a material change in ownership, funding, destination, activity, people, assets, regulation, tax position, data or technology movement, financing or exit plan.
Board pack
Create a structure decision record
Record the proposed structure, the purpose of each entity, unresolved workstreams, and responsible advisers. This control writes a checklist on the page; it does not file anything.
- Business function — markets, customers, people, assets, licences, data.
- Proposed structure — entity, seat, owner, function, why needed.
- Alternatives rejected — direct contracting, branch, parent-to-subsidiary, extra layers.
- Unresolved China-side and destination-side workstreams, each with an evidence owner.
- Change triggers and contemplated exit.
Evidence library
Legal sources, evidence and implementation detail
Matrices, source stamps and diagrams stay on this page. Open a card for the citation; the desk above stays the decision surface.
01Legal framework
Scope and legal framework: the structure sits inside several legal systems
In scope: choosing among direct operations, local establishments, subsidiaries and holding-company chains for a Chinese company; documenting why each extra entity exists; mapping which PRC and destination workstreams fire before the structure can operate.
Out of scope here: generic offshore-company marketing; step-by-step NDRC/MOFCOM/SAFE forms (use the ODI filing roadmap and funding programme); Hong Kong vs Singapore ranking (use that comparison); treaty-shopping or tax-residence manuals; destination incorporation encyclopedias; financing templates; employment handbooks; data-transfer manuals; export-control treatises.
Structure selection is not a single-law exercise. Distinguish the PRC outbound framework from destination-country law. Separate regimes may still apply to funding, technology, data, personnel, tax, export controls, customs, sanctions and regulated activities. Incorporation does not itself authorise outbound funding from China.
Legal hierarchy
- PRC overarching framework — State Council outbound-investment regulation (Decree No. 837 / Provisions on Outbound Investment). See what Decree 837 changes.
- Domestic management tracks — NDRC (project classification / management route); MOFCOM (enterprise outbound / overseas-enterprise path); SAFE / handling bank (FX registration and remittance execution).
- Connected controls — technology, data, personnel, tax, trade controls / sanctions.
- Destination law — company, licence, tax, employment, screening, disclosure, sector regulation.
The overseas entity is only one layer of the analysis. Operating facts determine which additional regimes apply.
02Evidence methodology
Evidence standard for every legal claim on this page
| Level | May say | Must show |
|---|---|---|
| 1 Binding law | “The law requires…” | Primary statute or regulation, exact provision, current version |
| 2 Official administrative guidance | “The authority’s filing process is…” | Named authority and dated guidance |
| 3 Practice observation | “Banks or counsel commonly request…” | Labelled as practice; never presented as a statutory duty |
| 4 Editorial decision framework | “This structure is usually worth testing when…” | Labelled as editorial; not a mandatory outcome |
Never convert “may trigger” into “requires” without primary authority. Thresholds and day-counts that cannot be tied to a live provision are omitted, not guessed.
03Statutory matrix
Authority gateway: move the filing detail to the specialist workflow
This pillar keeps only the classification handoff. It does not reproduce filing forms, statutory clocks or bank evidence rules that are maintained on the specialist pages.
| Question raised by the proposed structure | Maintained destination |
|---|---|
| Which official China outbound instruments control? | Official primary-source register |
| Which NDRC, MOFCOM and SAFE workstreams apply? | NDRC, MOFCOM and SAFE workflow |
| How should the project and funding sequence be governed? | End-to-end ODI and funding roadmap |
| What facts and documents should be assembled? | Outbound-investment readiness checklist |
| What if an offshore entity already exists or timing has slipped? | Timing and remediation workflow |
04Primary and official sources
Primary sources and destination-law handoff
The maintained primary-source register owns the official State Council, NDRC, MOFCOM, SAFE and connected-control links. Use it to record the instrument, version, project use, evidence owner and change trigger instead of copying a static source list into this pillar.
China-side authority does not decide destination formation, investment screening, licensing, tax, employment, data or sector rules. Give destination counsel the same ownership, function, funding and governance record used for the China workstream.
05Governance authority matrix
Annotated group structure
If an intermediate entity cannot be assigned a specific function at one or more of the ten points, question why it exists.
Governance and authority matrix (fill with actual, not nominal, authority)
| Decision / right | Shareholder | Board | Local director | China parent | Bank signatory | Contract signatory | Evidence |
|---|---|---|---|---|---|---|---|
| Capital increase | |||||||
| Borrowing | |||||||
| Guarantee | |||||||
| Material contract | |||||||
| IP licence | |||||||
| Related-party service | |||||||
| Dividend | |||||||
| Asset sale | |||||||
| Share sale | |||||||
| Liquidation |
Document actual authority. Align bank mandates, delegated authority, contracts, board minutes and shareholder rights. Route tax-residence and substance questions to qualified tax advisers — this matrix does not decide them.
06Cross-border flows
What may cross the border?
Use “may” and “fact-dependent.” Structure choice does not absorb these regimes.
| What may move from China? | Why it matters structurally | Specialist workstream |
|---|---|---|
| Equity capital | ODI / FX / bank sequence | ODI + funding |
| Debt | Financing / FX / tax / authority | Funding + tax |
| Guarantee / security | Outbound investment / banking / lender issues | ODI + financing |
| Employees / directors | Employment / immigration / payroll / control | Workforce |
| Technical personnel | Export control / confidentiality / IP / tax | Technology + workforce |
| Technology / know-how | Export-control / technology-export analysis | Trade controls |
| Code / models | Technology / data / IP | Tech + data |
| Personal data | Privacy / cybersecurity / cross-border transfer | Data / SCC |
| Business data / system access | Cybersecurity / data / services | Data |
| Equipment / product | Customs / export control / origin / sanctions | Trade |
| Services | Contract / tax / FX / regulatory | Tax + commercial |
| IP licence | Ownership / tax / transfer pricing / control | IP + tax |
07China outbound workstream tree
Decision tree 3 — does the structure trigger a China outbound workstream?
Start with: will the Chinese investor directly or indirectly acquire overseas ownership, control, management rights or another relevant overseas interest — or fund or guarantee such an acquisition?
Then branch into sensitive country or region; sensitive sector or project; direct vs indirect; capital / equity / debt / guarantee; state-owned assets; technology, controlled goods, services or data leaving China; employees or technical personnel deployed; subsequent reinvestment; material ownership, capital or activity change; disposal or exit.
Outputs are routes, not conclusions: ODI specialist · FX/bank · technology/export-control · data · workforce · destination counsel.
08Detailed implementation timeline
Practical workflow — dependency timeline
This is a structural sequence, not a universal filing timeline. Statutory days appear only when a source card supports them; none are invented below.
- Fact map — people, contracts, assets, licences, IP, data, ownership, funding, destination.
- Compare operating models — direct, branch, subsidiary, holdco, multi-tier.
- PRC classification — NDRC, MOFCOM, other triggered workstreams.
- Destination pre-clearance — formation, screening, licensing, ownership, regulatory.
- Approval / record-filing / document package — sequence depends on facts.
- FX / bank / funding readiness.
- Form / capitalise / contract / operate.
- Change control — capital, ownership, country, activity, technology, data, financing, exit.
09Incremental structural burden
Calculation diagram — incremental structural burden
This page does not publish tax-saving calculations or generic jurisdiction savings claims.
Annual Structural Burden = maintenance + audit + governance + banking + tax compliance + reporting + advisers + internal time
Additional Layer Justification = documented functional benefit − incremental burden − execution risk
If the remainder is not clearly positive on written facts, test parent → opco. Do not auto-generate a legal conclusion from the arithmetic.
10Compliance risk matrix
Compliance risk matrix (triage only)
NOT A LEGAL CONCLUSION
| Issue | Green | Amber | Red |
|---|---|---|---|
| Business rationale | Entity has a documented function | Benefit speculative | No reason beyond “common market practice” |
| Governance | Actual authority matches documents | Mixed / ambiguous | Paper board only |
| Funding | Route documented before formation | Dependencies unresolved | Entity formed before funding route tested |
| Ownership | Control and economics mapped | Some rights unclear | Legal and actual control inconsistent |
| Technology / data | No relevant movement, or cleared | Classification pending | Restricted issue unresolved |
| Destination law | Key requirements scoped | Local advice pending | Licence or screening may block operations |
| Maintenance | Costs justified | Benefit uncertain | Layer adds burden without function |
| Exit | Exit path modelled | Partial analysis | Structure obstructs the contemplated exit |
11Lifecycle / change-event matrix
Lifecycle / change-event matrix
| Trigger | Why the structure needs review |
|---|---|
| Capital increase | Funding, approvals, control, dilution |
| New shareholder | Ownership, disclosure, governance |
| Acquisition | Control, financing, integration |
| Joint venture | Governance, veto, exit |
| New country | Local registration, licensing, tax, screening |
| New business line | Entity role and regulatory perimeter |
| Local hiring | Employment, payroll, control, PE |
| New premises or assets | Operating model may have changed |
| IP migration | Ownership, tax, data, technology |
| R&D activity | Technology / export-control / control issues |
| New data flows | Cybersecurity / data-transfer analysis |
| Debt or guarantee | FX, bank, authority, tax |
| Cash pooling | Funding / authority / tax |
| Dividend | Solvency / tax / FX / approvals |
| Sale or listing | Exit mechanics / consents / evidence |
| Insolvency or wind-down | Liabilities / authority / repatriation |
12Data and methodology notes
Original data charts
None on this rebuild. The site does not yet have a proprietary, methodologised dataset of structure-selection outcomes. Third-party FDI or “most popular holdco” charts are not used. A later iteration may publish anonymised first-party tool outcomes (which model users test; which facts add layers; which change events cause review) if and when that dataset exists.
13How to use this guide and change log
How to use this guide · change log
Primary authority is required for thresholds, filing duties, defined terms, exceptions and penalties. Practitioner commentary is sequencing only. Absence of a rule is not a conclusion that “X is not required.”
Change log: 24 Aug 2026 - added review boundary and contextual routes; replaced the duplicate statutory matrix and static source cards with maintained specialist gateways. Decision trees, structure comparisons, governance, cross-border flows, lifecycle controls and the decision record were retained.
17 Aug 2026 — rebuilt from a short five-section orientation note into a pre-incorporation structure decision desk (direct answer, at-a-glance table, three decision trees, legal-hierarchy and annotated-structure diagrams, statutory matrix, source cards, five-route comparison, dependency timeline, burden formula, risk / border / change matrices, mistakes, checklist, decision record, reader journeys, FAQs). Generic Frame→Plan→Execute→Review hero module removed. UK establishment material retained only as a labelled destination illustration.
