A joint venture is the only China structure that is expensive to unwind. A supply contract ends with a notice. A distribution agreement ends with a transition. A joint venture ends with a negotiation, a valuation, a tax clearance, a registration and, if the parties cannot agree, a lawsuit.
That asymmetry explains why the governance and exit clauses deserve more attention than the commercial terms at term-sheet stage. The commercial terms describe the deal the parties hope for. The governance and exit clauses describe what happens when the hope fails.
Most foreign companies form a joint venture in China for one of three reasons: the sector requires a Chinese partner, the partner brings a licence, land, customers or a government relationship the foreign company cannot obtain alone, or the partner brings capital at a stage when the foreign company cannot raise it elsewhere. Each of those reasons implies a dependency, and a dependency is what a deadlock clause has to survive. A joint venture between parties who do not need each other is a joint venture that should have been a supply contract.
The practical conclusion, stated at the outset: the exit mechanism is not a clause to be negotiated at the end of the deal as a concession. It is the clause that determines whether the joint venture is an investment or a trap. A foreign company that cannot describe, before signing, the precise event on which it can require a buy-out, the price it will receive and the approvals the buyer will need has not yet decided to enter the joint venture. It has decided to enter a relationship it cannot leave.
What the structure actually is
A Chinese joint venture between a foreign investor and a Chinese partner is, in almost every case, a limited liability company with two or more shareholders. The company is the legal person. The shareholders hold equity interests recorded in the articles and on the register maintained by the market regulator. The joint venture does not exist as a separate legal form with its own statute any more: the Foreign Investment Law brought foreign-invested enterprises into the same company law framework as domestic companies, with additional rules on foreign investment entry and reporting.
Three documents govern the relationship, and they are not interchangeable.
The articles of association are the constitutional document filed with the market regulator. They bind the company and, in practice, the shareholders. A provision that is not in the articles may still bind the shareholders as a contract, but it may not be registrable and may not bind the company.
The shareholders’ agreement is the contract between the investors. It can be more detailed than the articles, and it usually is. The risk is divergence: where the shareholders’ agreement and the articles say different things about the same matter, the company’s officers, the registrar and a court will look to the articles, and the shareholder’s remedy is a claim for breach of contract against the other shareholder rather than an order that the company act differently.
The joint venture contract, where one is used, is a legacy of the pre-Foreign Investment Law regime, and some transactions still produce one because a partner’s internal approval process expects it. Where a joint venture contract exists, it must be reconciled with the articles, or the parties create a third source of inconsistency.
The governance architecture sits inside those documents: the shareholders’ meeting, the board, the general manager, the legal representative, the supervisor or supervisory board, and the seals. Each is a point at which control can be exercised or blocked, and each needs to be designed deliberately.
Registered capital and capital contribution are the fourth element. The amount of registered capital, the form of contribution and the deadline for payment are all registered. A shareholder that has not paid its subscribed contribution is exposed to the company, to creditors and to the other shareholder, and its voting and distribution rights may be affected. In a joint venture that needs cash, the contribution schedule is the first thing to check when a capital call is proposed, because a partner that has not paid its original subscription is unlikely to fund a new one.
Why deadlock happens
Deadlock is rarely caused by a disagreement about strategy. It is caused by a disagreement about something the agreement did not allocate. Four issues account for most cases.
Capital calls. The joint venture needs more money, one parent is willing and the other is not, and the agreement says nothing about dilution, default interest or forced contribution.
Related-party transactions. The joint venture buys components from one parent and sells through the other. When pricing is disputed, the entire operation stops, because neither parent will accept a margin it considers unfair.
Control of the legal representative and the company chop. In Chinese practice, the legal representative signs and the chop authenticates. Where one parent appoints the legal representative and holds the chop, the other has a veto by inertia.
Technology and business transfer. One parent wants to move a product line or a customer relationship into a joint venture subsidiary; the other sees it as a transfer of value out of the joint venture.
A 50:50 structure turns any unresolved disagreement on these issues into a stoppage. Even a 51:49 structure only converts deadlock into a majority decision, which the minority may then resist through information rights, litigation or refusal to cooperate.
There is a fifth cause that appears in AI and technology joint ventures more often than in traditional manufacturing: the allocation of intellectual property developed inside the joint venture. Where the joint venture develops a model, a dataset or a product, both parents will claim the result, and the agreement frequently says only that the joint venture owns what it creates. That answer is incomplete, because it does not say what licence each parent receives, whether a parent may use the joint venture’s output outside the joint venture’s field, and what happens to the output if the joint venture is dissolved. A joint venture whose only valuable asset is an unallocated licence is a joint venture that cannot be valued on exit.
Writing the deadlock trigger
A deadlock clause that says “in the event of a deadlock, the parties shall consult in good faith” is a clause that produces a meeting, not an exit. The trigger has to be objective, so that a party can establish it without the other party’s agreement.
The convention in well-drafted agreements is to define deadlock by reference to a failed decision process rather than to a state of relations. Three formulations are common, and they can be combined.
The first is a board failure: two consecutive board meetings at which a matter requiring a board resolution is not passed, whether because the votes are tied or because a quorum is not achieved. It has the advantage of being evidenced by the minutes.
The second is a shareholder failure: a matter reserved to the shareholders’ meeting is not passed at two consecutive meetings, or a written resolution is not signed within a stated period.
The third is a specific failure: a defined event, such as a failure to agree a capital call within thirty days of a funding notice, a failure to approve the annual budget before the start of the financial year, or a failure to appoint a general manager within a stated period after a vacancy.
The definition should also exclude the ordinary disagreements that the parties intend to resolve elsewhere. A clause that treats every failed resolution as a deadlock will be triggered by a routine dispute and will exhaust the terminal mechanism on a matter that did not deserve it. The reserved matters list is the filter: deadlock should attach to the matters on that list, not to every board decision.
Two further definitional points are worth drafting.
Who can declare deadlock. If either party can declare it, the clause becomes a tactical weapon, and the party that is content with the status quo will be declared into a buy-out it did not want. If only the chair can declare it, the chair’s appointer controls the trigger. The usual compromise is that either party may declare, but the declaration has no effect until the escalation ladder has run and a stated period has expired.
What survives a declaration. A deadlock declaration should not suspend the joint venture’s operations. The agreement should state that the business continues under the current budget, that the management team continues to be paid, and that neither party may take unilateral action to wind down operations while the mechanism runs. Without that provision, the declaration itself becomes the event that destroys the value the parties are arguing about.
Escalation: a ladder with deadlines and a terminus
The conventional structure runs: chief executives within five business days, boards within ten, mediation, then the terminal mechanism. The structure is right. What decides whether it works is the detail.
The notice. The escalation should begin with a written notice that states the matter in dispute, the resolution sought, and the provision of the agreement said to be breached or unfulfilled. A notice that says “the parties are in deadlock” gives the other side nothing to respond to and gives a later tribunal nothing to assess.
The information pack. Each escalation meeting should be preceded by a short written pack: the financial position, the operational impact of the deadlock, the options considered and the consequences of each. The purpose is to move the conversation from positions to consequences. In practice, the pack is also the document that a mediator or tribunal will read first, and preparing it well shortens everything that follows.
The attendees. The chief executive level should be attended by people with authority to settle, not by the deal team that negotiated the original agreement. A deadlock escalated to the same negotiators who created it rarely resolves, because their positions are the ones in dispute.
The standstill. During the escalation period, neither party should be permitted to commence proceedings, other than for urgent injunctive relief. A standstill clause is what makes the ladder meaningful; without it, one party files first and the escalation becomes a negotiation about the litigation.
The interim operating protocol. The agreement should state how the joint venture is run while the dispute is unresolved: who signs, who approves payments, how the budget is set, and what happens to customer commitments. The protocol is the operational counterpart of the standstill, and it is the provision that most often saves a joint venture that later settles.
Mediation. Mediation in China is available through commercial mediation institutions, through the courts’ pre-litigation mediation process, and through ad hoc mediation agreed by the parties. A mediated settlement can be given enforceability through a court’s confirmation procedure, which converts the agreement into an enforceable instrument. That step is worth taking, because a settlement that depends on goodwill is not a settlement.
The terminus. The ladder must end somewhere. The last rung should be the terminal mechanism: a put option, a buy-sell, a drag-along or, in the last resort, a dissolution petition. An escalation clause with no terminus is a delay mechanism, and the party that benefits from delay will use it.
Governance design that prevents deadlock
Prevention is cheaper than exit. Five design choices do most of the work: a reserved matters list with a hierarchy, an escalation ladder with deadlines, a tie-breaker for operational matters, real information rights, and a deadlock trigger with a terminal consequence. Each deserves more detail than a list can carry.
The reserved matters list. Some matters require unanimity, some require a supermajority, some are delegated to management. The list should be short and specific; a long unanimity list is a recipe for paralysis. A workable list covers changes to the business scope, the annual budget, capital expenditure above a threshold set in the agreement, borrowing above a threshold, related-party transactions, the appointment and removal of the general manager and the finance controller, the declaration of dividends, the issuance of new equity, the transfer or licence of material intellectual property, and any transaction that would change the nature of the business. Eight to twelve items is manageable. Thirty is not.
The quorum and the voting rule. A board quorum that requires the attendance of a director appointed by each parent gives each parent a blocking right by absence. That may be intended, but it should be stated. A voting rule that requires a simple majority of directors present converts a 50:50 board into a contest for the independent director’s vote, and the independent director’s appointment and removal terms then become the most important clause in the agreement.
The chair’s casting vote. A casting vote for the chair on defined operational matters prevents day-to-day paralysis without conceding strategy. The definition of “operational” is the negotiation. A clause that gives the chair a casting vote on all matters converts the joint venture into a controlled subsidiary with a profit share, and the minority will resist it, correctly.
The independent director. Where the parties want equality and a tie-breaker, an independent director appointed by both, or by a nominating body agreed by both, supplies a vote that neither controls. The appointment mechanism, the removal mechanism and the independence criteria all need drafting, because an independent director who is in fact aligned with one parent is worse than no independent director at all.
Information rights that are real. Monthly management accounts, board papers, access to the finance system and a right to appoint an observer. Minority shareholders who feel informed litigate less. The rights should be backed by a remedy: if information is withheld beyond a stated period, the minority may appoint an auditor at the joint venture’s expense, or the withholding counts as a deadlock event.
The budget default. A budget that cannot be approved is the most common cause of operational paralysis, because the joint venture cannot spend, hire or commit. The clause should provide that if the board has not approved the next year’s budget by a stated date, the current year’s budget continues with an adjustment for inflation or for a defined index, until a new budget is approved. This is a time-limited budget veto, and it is one of the cheapest protections available.
The deadlock definition and the terminal mechanism. The two must be drafted together. A definition without a mechanism produces a declared deadlock with no consequence; a mechanism without a definition produces a buy-out that can be triggered at will.
The legal representative and the chop
In Chinese practice the legal representative has statutory authority to bind the company, and the company chop authenticates documents. Control of those two instruments is often more valuable than a board seat.
The articles should state who appoints the legal representative, how that appointment can be changed, and who holds the chop and the finance seal. Best practice is to separate the roles: the general manager appointed by one parent may be the legal representative, while the chop is held by a finance controller appointed by the other, with a dual-signature rule for payments above a threshold.
The separation has to be written into the internal approval matrix as well as the articles, because a bank will act on the signature and seal it has been instructed to accept. The joint venture’s bank mandate should require two authorised signatories for payments above a threshold, and the mandate should be reviewed whenever the finance controller changes.
Where a dispute arises, the parent that physically holds the chop and the business licence can continue to operate the company, open bank accounts and file registrations. The other parent’s remedy is a court application, which takes months. The time to prevent that is at formation, not at deadlock.
A practical addition: keep a certified copy of the business licence, the articles and the register of shareholders in the foreign parent’s files, and record the chop number. If the chop is later replaced, the record shows what was used before.
Exit mechanisms, and what each is actually worth
| Mechanism | How it works | Main weakness |
|---|---|---|
| Put option | Minority can require the majority to buy at a defined price | Buyer funding and approvals; price disputes |
| Call option | Majority can buy out the minority on a trigger | Minority may resist valuation |
| Buy-sell (shoot-out) | One party names a price, the other chooses to buy or sell at that price | Requires both parties to be able to fund |
| Drag-along | Majority can force the minority to sell in a third-party sale | Needs a valuation floor to be accepted |
| Judicial dissolution | Court orders dissolution where management is deadlocked | Slow, last resort, liquidation value |
| Secondary sale | Sale to a third party with tag-along rights | Buyer availability and regulatory approvals |
Each mechanism has a determinacy problem. A put option is only as good as the valuation method and the funding obligation behind it. A buy-sell is elegant in theory and depends on both parties having the liquidity to perform. A drag-along protects the majority but is only accepted by a minority if it carries a floor price and a tag.
The practical answer is usually a combination: a put option on defined triggers, a drag-along with a floor, and a buy-sell as the terminal mechanism, with judicial dissolution as the statutory backstop.
The put option is the minority’s instrument. It should specify the trigger events, the notice period, the valuation method, the payment terms and the closing conditions. The trigger events commonly include a deadlock that has run through the escalation ladder, a change of control of the other shareholder, a material breach that is not cured, a failure to pay dividends for a defined period, and a failure to obtain a required regulatory approval. Two years without dividends is a common example; the number matters less than the fact that it is stated.
A put option that is not backed by funding is a right to sue. The clause should require the buyer to demonstrate the ability to pay, through a bank guarantee, an escrow, a parent company guarantee or a payment schedule secured on the shares. Where the buyer is a Chinese operating company with limited liquidity, the guarantee of its parent may be worth more than the put itself.
The call option is the majority’s instrument, and it is the mirror image. It is easier to enforce because the buyer wants to buy, and it is more dangerous for the minority because the price is usually set by a formula that the majority influences. The minority should insist on an independent valuation or an expert determination for the call price, and on a minimum price where the trigger is a breach by the majority.
The buy-sell is the mechanism that resolves a true deadlock between equals. One party names a price per unit, and the other chooses whether to buy at that price or sell at that price. The elegance is that the party naming the price must name one it would accept in either role, which disciplines the number.
There are two common variants, and the difference matters. In the version usually called a buy-sell, one party serves a notice stating a price, and the other elects within a period to buy the first party’s shares at that price or to sell its own shares at that price. In the variant sometimes called a shoot-out, both parties submit sealed bids, and the higher bidder buys the other out at its own bid. The first is simpler and depends on one party’s judgement; the second is more robust against a party that names an absurd price, because the other party can simply accept the role the price implies.
Three drafting points decide whether a buy-sell works.
The election period must be long enough for the recipient to raise finance and short enough to produce a decision. A period that is too short favours the party with cash on hand.
The funding obligation must be real. The clause should require the buyer to deposit the price, or to provide a guarantee, before the seller is obliged to transfer. A buy-sell that allows the buyer to elect and then fail to pay leaves the seller with a claim and the buyer with the shares.
The default consequence must be stated. If the electing buyer fails to complete, the seller should have the right to buy the buyer’s shares at the same price, or to specific performance, or both. Without that provision, the mechanism can be used to freeze the other party out at no cost.
The drag-along allows the majority to force the minority to sell in a third-party sale. It is the mechanism that makes a joint venture saleable to a strategic buyer, because a buyer will not pay for a company with a minority shareholder who can refuse to sell. The minority will accept it only with a floor price, a tag-along right and a requirement that the minority receives the same price and terms as the majority. A drag-along without a floor is a right of expropriation.
The tag-along protects the minority when the majority sells. It is cheap to grant and it changes the majority’s incentive, because a buyer who must acquire 100 per cent will price the whole company rather than the majority stake.
Valuation: the clause that decides the dispute
The valuation clause decides more disputes than the exit clause. Methods in common use are a multiple of EBITDA with a defined adjustment mechanism, net asset value with an independent appraisal, the price of the last financing round with a time adjustment, or an independent expert determination with a baseball-style outcome.
Whatever the method, three elements should be fixed in advance: the valuation date, the treatment of intercompany balances, and the mechanism for resolving a disagreement between valuers. An expert determination clause with a defined timetable is more useful than a right to litigate the valuation.
The method should be chosen for the business the joint venture actually operates. A manufacturing joint venture with stable margins is suited to an EBITDA multiple. A joint venture whose value is a licence or a property is suited to a discounted cash flow or an appraisal. A joint venture that has not yet generated revenue is suited to a cost-based or last-round method, with a mechanism for adjusting for progress. The mistake is to adopt the method from a template without checking that it fits.
The valuation date is the second decision. The date of the trigger, the date of the notice and the date of closing each produce a different number, and the choice matters most when the business is deteriorating. A seller wants the earliest date; a buyer wants the latest. The usual compromise is the date of the trigger, with an adjustment for value transferred out after that date.
The adjustments are where most of the argument lives. Net debt, working capital, intercompany balances, shareholder loans, unpaid dividends, and assets that one parent has contributed but not transferred all change the equity value. A clause that says “the price shall be the enterprise value less net debt” and leaves the definition of net debt to the parties is a clause that will be litigated. A schedule that defines each adjustment and states the accounting policy to be applied is worth the drafting time.
The discount question should be addressed expressly. A minority stake in a private Chinese company is less marketable than a controlling stake, and a valuer may apply a discount for lack of control or lack of marketability. A seller should negotiate that the valuation is on a controlling, marketable basis, or that no discount applies, because the buyer is acquiring a stake that completes its control. A buyer should resist, and the negotiation should happen at signing rather than at exit.
A deadlock discount is a further point. A buyer may argue that a company in deadlock is worth less, and that the price should reflect the distress. A seller should negotiate a clause that the valuation assumes the business continues as a going concern, without a discount for the deadlock that triggered the sale. Otherwise the mechanism punishes the party that uses it.
The expert determination mechanism is the procedural answer. The parties appoint an independent expert, each submits its case, the expert determines the value, and the determination is final except for manifest error. The clause should set the timetable, the qualifications of the expert, the information each party must provide, and the cost allocation. A right to litigate the valuation converts a commercial exit into a two-year dispute.
Capital calls and dilution
Capital calls are where good relationships end. The joint venture needs more money, one parent can fund it and the other cannot or will not, and the agreement is silent on what happens next.
A workable capital call clause answers four questions in advance. Who decides that the call is needed, and by what majority? How much notice must be given, and what information must accompany the notice? What happens to a shareholder that does not pay: dilution, default interest, loss of governance rights, or all three? And is there a cap on the number of calls, so that the funding obligation is finite?
The dilution mechanism is the most contested element. A formula that dilutes the non-paying shareholder at the pre-call valuation is generous to the payer; one that dilutes at a heavily discounted valuation is punitive and may be challenged as a penalty. The clause should state the formula, the valuation basis and the consequences clearly enough that a court would enforce it.
A second mechanism is worth adding: a loan alternative. A parent that cannot subscribe can lend to the joint venture on defined terms, with the loan convertible at a later date. That keeps the joint venture funded without forcing a dilution fight at the worst moment.
A third is the defaulting shareholder’s governance consequence. A shareholder that has not funded may lose its right to appoint a director, or its voting rights may be suspended, until it pays. That consequence is often more effective than dilution, because it addresses the control the non-paying shareholder is trying to preserve.
The cap matters as much as the mechanism. A joint venture with an unlimited capital call obligation is a joint venture in which the foreign parent has written an open-ended cheque. The agreement should state the maximum aggregate contribution, or the number of calls, or the circumstances in which a call may be made. Beyond the cap, the funding has to be voluntary, and the consequence of not funding is dilution or exit rather than default.
Related-party transactions and transfer pricing
A joint venture that buys from one parent and sells through the other is a transfer pricing problem before it is a governance problem.
The agreement should require related-party transactions to be on arm’s length terms, supported by contemporaneous documentation, and approved by the board with the interested director abstaining. Where the joint venture has minority shareholders, an independent fairness opinion for material transactions is a practical protection.
Chinese tax law requires related-party transactions to be at arm’s length and imposes documentation and filing obligations on enterprises above defined thresholds. A joint venture that buys components from its foreign parent at a margin that suits the parent may be adjusting taxable income in China, and the tax authority can make a transfer pricing adjustment with interest.
The governance and tax issues converge. A pricing formula that is defensible to the tax authority is usually also defensible to the minority shareholder, because both require evidence that the price reflects what unrelated parties would agree. A joint venture that keeps a contemporaneous benchmarking file serves both purposes at once.
The practical drafting is a pricing policy annexed to the agreement: the categories of related-party transaction, the method for each (comparable uncontrolled price, resale price, cost plus), the review interval, and the documentation to be retained. A policy is more useful than a clause, because it tells the finance team what to do on the first transaction rather than after the dispute.
Regulatory and practical constraints on exit
A share transfer in a Chinese company requires registration with the market regulator, and the shareholders’ agreement cannot override mandatory company law. Pre-emptive rights of other shareholders, the articles and any regulatory approval for the sector all affect whether a transfer can complete.
The sequence is more complex than a foreign investor expects, and each step can delay or defeat a closing.
Pre-emptive rights. The other shareholders usually have a right of first refusal over a transfer to a third party. The right can be waived, but the waiver must be obtained in the form the articles require. A transfer that ignores the right may be challenged, and the registration may be refused.
The articles. A transfer that conflicts with the articles cannot be registered until the articles are amended, and an amendment may itself require a supermajority. A deadlock that prevents the amendment also prevents the exit, which is why the exit mechanism should include a power of attorney or an irrevocable undertaking from the other shareholder to vote in favour of the amendments needed to give effect to the transfer.
Sector approval. Where the joint venture operates in a sector subject to foreign investment restrictions, a transfer that changes the foreign shareholding may require approval or filing under the foreign investment regime. The negative list in force at the time of the transaction determines whether the sector is prohibited, restricted or open, and the list changes. A transaction structured on the list as it stood at formation may not be registrable at exit.
Merger control. A change of control may trigger a merger control filing if the turnover thresholds are met. The thresholds change, and the analysis should be refreshed at the time of the transaction rather than assumed from the entry. A share transfer that should have been filed and was not can be unwound.
National security review. A foreign buyer in a sensitive sector may face a national security review. The review applies to exits as well as entries, and a joint venture formed in a sector that has since been added to a review list may find that its exit requires a review it did not anticipate.
State-owned partners. Where the Chinese shareholder is a state-owned enterprise, the transfer of its equity may be subject to asset valuation filing and, in many cases, to a requirement that the stake be offered through a property exchange. That process has its own timetable and its own valuation rules, and it can override a price agreed between the shareholders. A foreign investor negotiating a put option against a state-owned partner should establish at the outset whether the partner can perform the obligation, and if not, what alternative security is available.
Tax clearance. A transfer by a foreign shareholder of a Chinese company may be subject to enterprise income tax on the gain, and the tax authority will not issue the clearance needed for the registration until the tax is settled or the exemption is confirmed. The clearance is a closing condition in practice, whatever the agreement says.
Foreign exchange. Repatriation of the after-tax proceeds requires the transfer to be registered, the tax to be settled and the bank to be satisfied that the documentation is complete. That process takes weeks, and it should be built into the closing timetable rather than discovered afterwards.
Tax and repatriation on exit
The exit price is not the amount the seller receives.
A foreign shareholder that transfers its equity in a Chinese company is subject to enterprise income tax on the gain, generally withheld by the buyer before remittance. Where the transfer is of an offshore holding company whose value derives mainly from Chinese assets, the indirect transfer rules may apply, and the tax authority can look through the structure.
Stamp duty applies to the transfer document. Where the consideration is not arm’s length, the tax authority may assess the gain by reference to the fair value of the equity.
The indirect transfer point deserves emphasis because it is the one that foreign groups most often miss. A sale of the shares of a Hong Kong or Cayman holding company that owns a Chinese subsidiary can be treated as a transfer of Chinese taxable assets where the offshore company has little substance and its value derives mainly from the Chinese assets. The consequence is a Chinese tax liability, and a withholding obligation on the buyer, in a transaction the parties thought was offshore. The structure should be reviewed before the sale process begins, not after the term sheet is signed.
The practical sequence is: agree the price, obtain a valuation that supports it, calculate the gain, settle or provide for the tax, obtain the clearance, register the transfer, and remit. Each step has its own documentation, and the seller should assemble the file before the buyer’s advisers ask for it.
Judicial dissolution
A shareholder with at least ten per cent of the voting rights may petition a court to dissolve the company where serious management difficulty has caused substantial loss and cannot be resolved by other means. The remedy is real but narrow: courts require evidence that other routes have been exhausted, the process is slow, and the outcome is liquidation at a value below a negotiated sale.
In practice, the threat of dissolution is more useful than the remedy. It changes the negotiating position because it threatens both parties with the destruction of value, and it is most effective when the alternative, a funded buy-out, is on the table.
Three points are worth knowing before relying on it.
The petition is against the company, and the other shareholder is usually joined. The evidence is the corporate record: board minutes showing the failure, correspondence showing the attempts to resolve, and financial evidence showing the loss. A shareholder that has not kept records has a weak petition.
The court will ask whether other means have been tried. Mediation, a buy-out offer, a proposal to amend the articles, and an offer to appoint an independent director are all evidence that the parties have not exhausted the alternatives. A petitioner should document the attempts.
The outcome is liquidation. The company’s assets are sold, the creditors paid, and the shareholders receive the residue. For a company whose value is a licence, a team or a customer relationship, the residue is close to nothing. That is why the remedy is leverage rather than a plan.
Operating the joint venture during deadlock
Deadlock is a process, not an event. The parties usually know months before the formal trigger that the relationship is failing.
Three practices reduce the damage. Keep the operational relationship separate from the governance dispute, so that customer deliveries and supplier payments continue. Use the escalation ladder early, before positions harden. And document the disagreement in writing, because a negotiated settlement or a later arbitration will turn on the contemporaneous record.
The worst outcome is a joint venture that neither operates nor dissolves, with both parents funding a company that produces nothing. The exit mechanism exists to prevent that outcome, and it only works if it can be triggered without the consent of the party that is causing the deadlock.
A fourth practice is worth adding: protect the assets that can walk. The key engineers, the customer contracts, the source code repository and the data are all capable of leaving the joint venture during a dispute. Employment contracts with the joint venture, assignment clauses, access controls and a clear record of what belongs to whom reduce the damage. A parent that has seconded its engineers to the joint venture should confirm that their work product belongs to the joint venture and that their access can be revoked.
What deadlock looks like in practice
A typical case: a 50:50 joint venture between a European technology company and a Chinese manufacturer. The joint venture buys components from the Chinese parent and sells to customers introduced by the European parent. The European parent wants to move a new product line into the joint venture; the Chinese parent wants to keep it in its own subsidiary.
The board meets twice without a decision. The budget is not approved, so the joint venture continues on last year’s numbers. The Chinese parent’s finance director controls the chop and declines to sign a bank facility. The European parent’s seconded engineers continue to work, but their access to the factory is restricted.
Six months later, the European parent invokes the deadlock clause. The agreement defines deadlock as two consecutive board meetings without a required resolution, which is satisfied. The put option requires the Chinese parent to buy at an EBITDA multiple, but the Chinese parent has not agreed the EBITDA adjustment mechanism, and the valuation dispute takes another year.
The lesson is not that the clause failed. It is that the clause was drafted without the operational detail that determines whether it can be used: the budget default, the chop control, the access rights, and the valuation mechanics.
The case also shows the value of a second, cheaper exit. If the agreement had contained a right to sell to a third party with the other shareholder’s tag-along, the European parent could have run a sale process in parallel with the put negotiation. The competitive tension would have produced a price, and the Chinese parent would have had to choose between buying at that price and accepting a new partner. A single exit route gives the other party a veto by delay. Two routes do not.
Documents, filings and timeline
| Step | Authority or counterparty | Typical timing | What foreign companies miss |
|---|---|---|---|
| Joint venture structure and negative list check | NDRC, MOFCOM system, local district | Before term sheet | Assuming the sector treatment at formation still applies at exit |
| Articles and shareholders’ agreement | Parties, market regulator | At formation | Leaving the exit mechanics out of the articles |
| Registered capital and contribution schedule | Market regulator, bank | At formation and per schedule | Not linking voting rights to paid-up capital |
| Foreign investment information report | Market regulator system | After establishment and on change | Treating the report as a formality and missing the update on change |
| Bank mandate and seal custody | Bank, company | At formation | Leaving one parent with sole control of both |
| Board and shareholder resolutions for a transfer | Company, shareholders | At exit | Missing the pre-emptive rights waiver |
| Equity transfer registration | Market regulator | After signing and tax clearance | Sequencing the tax clearance after the registration |
| Tax filing and clearance on the gain | Tax authority | Before registration completes | Not modelling the indirect transfer position |
| Merger control or national security review, if applicable | SAMR, NDRC/MOFCOM | Before closing | Assuming the entry analysis still holds |
| Foreign exchange registration and remittance | Bank, SAFE system | After registration | Building the timetable without the bank’s requirements |
Where foreign companies get this wrong
Signing a shareholders’ agreement that the articles do not reflect. The exit mechanism is in the contract and not in the constitutional document. When the parties come to implement it, the registrar requires an articles amendment, the other shareholder refuses to vote for it, and the mechanism is unenforceable in practice. The fix is to put the transfer mechanics in the articles, or to take an irrevocable voting undertaking at signing.
Accepting a put option with no funding behind it. The foreign investor negotiates a put at a fair price and discovers at exit that the Chinese partner has no cash, cannot obtain approval to spend it, or is a state-owned entity that must sell through an exchange. The put becomes a claim, and the claim takes years. The fix is security: a guarantee, an escrow, or a charge over the partner’s other assets.
Leaving the valuation method to be agreed later. “The parties shall agree a fair price” is not a valuation clause. It is an agreement to negotiate, and the negotiation happens when one party is under pressure. The fix is a method, a date, an adjustment schedule and an expert determination procedure.
Assuming the entry approvals cover the exit. The negative list, the merger control thresholds and the national security review list change. A joint venture formed when the sector was open may need approval to transfer when the sector is restricted. The fix is to re-run the analysis at the time of the transaction and to make the approval a condition with a defined consequence if it is not obtained.
Treating the legal representative as a formality. The foreign parent accepts the Chinese partner’s appointment of the legal representative and the finance controller, and discovers that the partner controls every signature. The fix is to separate the roles, to write the separation into the bank mandate, and to keep copies of the corporate documents.
Not documenting the deadlock. The parties negotiate for a year without minutes, without written proposals and without a record of the escalation. When the arbitration begins, the record shows a disagreement but not a deadlock, and the tribunal declines to find that the trigger was met. The fix is a correspondence protocol from the first sign of trouble.
Cost and commercial impact
The cost of a joint venture exit is driven by four items, and only one of them is legal fees.
The first is the valuation gap. The difference between what the seller thinks the stake is worth and what the buyer will pay is usually larger than the legal cost of the exit, and it is resolved by the valuation clause or by a negotiation conducted under the threat of dissolution.
The second is the time to closing. A transfer that requires tax clearance, registration, a pre-emptive rights waiver and possibly a merger control filing takes months. During that time the business continues to consume management attention, and the value may deteriorate.
The third is the tax. Enterprise income tax on the gain, stamp duty and the indirect transfer exposure can take a material share of the proceeds. The exit should be modelled on an after-tax basis before the price is agreed.
The fourth is the operating cost of the deadlock itself. A joint venture that is not investing, not hiring and not approving a budget loses customers, and the loss is not recovered when the dispute settles.
The commercial case for spending on the governance and exit clauses at formation is that each of those four costs is reduced by a clause that was drafted before the parties fell out. The clause is cheap when it is written and expensive when it is needed.
Practical checklist
- Check the negative list, the merger control position and the national security review exposure before signing, and diarise a review for the exit.
- Put the transfer mechanics and the board composition in the articles, not only in the shareholders’ agreement.
- Define deadlock by objective reference to failed board or shareholder resolutions on reserved matters.
- Keep the reserved matters list to the eight to twelve decisions that genuinely matter.
- Provide a budget default so that a failure to approve the budget does not stop operations.
- Separate the legal representative from the holder of the chop, and reflect the separation in the bank mandate.
- Fix the valuation method, the valuation date, the adjustment schedule and the expert determination procedure.
- Back the put option with a guarantee, an escrow or a charge, and state what happens if the buyer does not pay.
- Provide a drag-along with a floor and a tag-along for the minority.
- Model the tax, the indirect transfer position and the repatriation timetable before negotiating the price.
- Keep a contemporaneous written record from the first sign of disagreement.
- Review the articles and the shareholders’ agreement together, because a mechanism that conflicts with the articles may not be registrable.
The commercial logic is simple. A joint venture is a marriage with a prenuptial agreement written by the parties who are still in love. The value of the agreement is tested only when they are not. A foreign company that treats the exit clauses as boilerplate has bought an equity stake it may not be able to sell, in a company it may not be able to control, under a valuation it may not be able to determine. The clauses are the investment.
Geason Law Offices advises foreign companies on Chinese joint venture structures, governance and deadlock mechanisms, buy-out and valuation design, and exit execution including regulatory approvals and tax. The joint venture service page sets out how we structure these arrangements, and the related guides on contributing IP to a Chinese joint venture and arbitration or Chinese courts cover the two questions that most often decide whether an exit can be enforced.
Frequently Asked Questions↑ Back to article
What causes deadlock in Chinese joint ventures?
Most deadlocks arise from four issues the agreement never allocated: additional capital calls, related-party transactions between the JV and one parent, the appointment of the legal representative and control of the company chop, and the transfer of technology or business to JV subsidiaries. A 50:50 structure turns any unresolved disagreement on those issues into a stoppage.
Does a 50:50 joint venture always deadlock?
No, but it has no tie-breaker. Where the parties want equality, the governance design has to supply a mechanism — a casting vote on defined matters, an independent director, escalation to chief executives with a deadline, and a terminal buy-sell. Without one, a disagreement that the parties cannot resolve becomes a dispute by default.
Can a foreign shareholder force a buy-out of its stake?
Only if the shareholders' agreement provides a put option, or if a statutory route applies. A put option is enforceable as a contract, but it must specify the valuation method, the payment currency and the approvals required; a put option that cannot be funded or approved is not an exit. Statutory routes include the shareholder's right to demand purchase in defined circumstances and judicial dissolution.
What is judicial dissolution and when is it available?
A shareholder holding at least ten per cent of the voting rights may petition a court to dissolve the company where serious difficulty in management has caused substantial loss and cannot be resolved by other means. It is a last resort: courts are reluctant, the process is slow, and the outcome is liquidation rather than a sale. It is leverage more often than a plan.
How is a JV stake valued on exit?
By whatever method the agreement specifies. Common approaches are a multiple of earnings, net asset value, an independent valuation, or the price of the last financing round with adjustments. The method matters less than its determinacy: a clause that says the parties will agree a fair price is a clause that produces a dispute.
Can a foreign investor repatriate proceeds from a JV exit?
Yes, where the transaction is properly documented and tax is settled. A share transfer to a Chinese or foreign buyer requires registration with the market regulator, and the proceeds can be remitted after payment of applicable taxes. The practical constraints are the buyer's funding, the tax clearance and the registration timetable rather than a general prohibition.
Should deadlock be resolved by arbitration or litigation?
Arbitration is usually better for governance disputes, because it is confidential and the tribunal can be chosen for commercial and technical expertise. Chinese courts are necessary for certain remedies, such as judicial dissolution and compulsory enforcement of registration changes. Many agreements provide for arbitration of disputes and reserve specific matters for the courts.