A foreign supplier negotiates hard and secures a clause: on late payment, the Chinese buyer pays liquidated damages of 0.5% of the contract price per day. Two years later the buyer defaults, the supplier sues for the accrued figure — and the court awards a fraction of it.

Nothing went wrong procedurally. The clause was valid. What happened is Article 585 of the Civil Code of the People’s Republic of China (in force 1 January 2021), which gives Chinese courts and arbitral tribunals an express power to adjust an agreed damages figure. Understanding that power is the difference between a clause that deters breach and a clause that merely looks like it does.

What Article 585 says

Article 585 has three limbs:

  1. The parties may agree that a defaulting party pays a fixed sum of liquidated damages, or agree a method for calculating damages.
  2. Where the agreed figure is lower than the loss actually caused, the aggrieved party may apply to have it increased.
  3. Where the agreed figure is manifestly higher than the loss actually caused, the defaulting party may apply to have it appropriately reduced.

Two structural features follow.

Adjustment is application-driven. The court does not adjust on its own initiative. A defendant who does not ask for a reduction will not get one — which makes the reduction application a standard, and often decisive, item in Chinese defence pleadings.

The benchmark is actual loss, not the agreed figure. Chinese law does not adopt the common-law penalty/liquidated-damages distinction, under which a genuine pre-estimate of loss is enforceable as agreed. It asks a different question: what loss did the innocent party actually suffer, and is the agreed figure manifestly out of proportion to it?

How far will a court go?

Judicial practice has long worked from a reference point: an agreed figure exceeding actual loss by around 30% may be treated as “manifestly higher”. That benchmark originated in judicial interpretation of the former Contract Law and has been carried into the current framework through the Supreme People’s Court interpretation on the general provisions of the contract part of the Civil Code (Fa Shi [2023] No. 13, effective December 2023).

Verification note. The 30% reference point is a starting position for judicial reasoning, not a formula. It is not applied mechanically, and the exact article number of the current interpretation should be confirmed before it is cited in a pleading. Confirm the operative provision and any subsequent guidance before relying on it.

More importantly, the reference point is not the whole test. Courts weigh:

  • the degree of performance achieved before breach;
  • the degree of fault of each party;
  • the expected benefit the innocent party would have obtained from performance;
  • the bargaining position of the parties and whether the clause was negotiated or imposed in a standard form; and
  • the general principles of good faith and fairness.

A daily rate that is standard in the market, agreed between two commercially sophisticated parties with equal bargaining power, is far more likely to survive than the same rate embedded in a standard-form contract imposed on a small counterparty.

The evidential dynamic nobody plans for

Here is the practical trap. On a reduction application, the tribunal needs to know the actual loss in order to judge whether the agreed figure is manifestly disproportionate to it. The party best placed to prove actual loss is the claimant — the innocent party.

So the claimant who drafted the liquidated damages clause precisely to avoid having to prove loss ends up proving loss anyway, defensively, to resist the reduction. And a claimant who arrives at the hearing with no loss evidence, relying entirely on the clause, is exposed: the tribunal has nothing to weigh against the agreed figure except its own sense of proportion.

The operational conclusion: a liquidated damages clause reduces the burden of proving loss; it does not eliminate the need to be able to prove it. Keep the evidence anyway — funding cost records, replacement purchase invoices, documented lost margin on downstream contracts, storage and handling costs.

Article 584 — the foreseeability rule. Damages for breach cover losses caused by the breach, including expected profits, but may not exceed the loss that the breaching party foresaw or ought to have foreseen at the time of contracting. This caps “actual loss” itself, and therefore indirectly caps what a liquidated damages clause can defensibly recover.

The drafting implication is direct: if you expect an unusual category of loss on breach — a downstream penalty, a bespoke tooling write-off, a specific financing cost — recite it in the contract. A loss recorded in the recitals is a loss the counterparty foresaw.

Article 577 — the remedies for breach: continued performance, remedial measures, or compensation for loss. Liquidated damages sit alongside these, not instead of them.

Drafting for survivability

  1. Anchor the figure to something. A clause that recites the commercial basis for the rate — cost of funds, documented storage cost, a specific downstream exposure — gives the tribunal a reason to accept it. A bare percentage invites scrutiny.
  2. Keep the rate within market range. Very high daily rates attract reduction almost automatically. A rate that is defensible by reference to prevailing commercial practice, and to statutory ceilings on interest for delayed payment, is far more robust.
  3. Cap the aggregate. An uncapped daily rate on a long default can reach a multiple of the contract price. Tribunals reduce those. A stated aggregate cap, at a level that is plainly proportionate, signals restraint and survives better.
  4. Recite foreseeable losses expressly. This engages Article 584 in your favour.
  5. Do not rely on the clause alone at trial. Plead and evidence actual loss in the alternative.
  6. Consider a deposit (定金) where appropriate. The deposit regime operates on a separate statutory basis with its own statutory ceiling, and interacts with liquidated damages — the innocent party may generally elect between them rather than recover both. Whether a deposit or a liquidated damages clause is the better instrument depends on the transaction, and the two should be considered together at the drafting stage rather than bolted on separately.

The takeaway

Chinese contract law is not hostile to agreed damages clauses. It is hostile to agreed damages clauses that operate as a windfall. A clause that reflects a genuine, articulable commercial exposure, that is capped, and that is backed by loss evidence when it matters, will generally be enforced close to its terms.

A clause drafted purely as a deterrent, at a rate nobody expected to actually pay, will be adjusted — and the party who relied on it will discover the adjustment at the least convenient moment.

Instruments referred to in this article: Civil Code of the People’s Republic of China (in force 1 January 2021), Articles 577, 584 and 585; Supreme People’s Court Interpretation on the General Provisions of the Contract Part of the Civil Code (Fa Shi [2023] No. 13).