When you are building a company, your contracts are your shield. You draft clauses to protect cash flow, secure supply chains, and ensure that if a vendor, developer, or enterprise client walks away, your startup is not left holding the bag. To achieve this, founders frequently rely on liquidated damages (LD) clauses, clauses that fix in advance the sum payable if a party breaches the agreement.
But there is a trap waiting for the unwary. If your LD clause is deemed too high or improperly structured, a court will strike it down as an unenforceable penalty, leaving you with nothing but legal bills and the hard task of proving your actual financial losses from scratch.
For companies operating in or scaling through Singapore, a critical hub for many Asian startups, this boundary was redrawn by the Singapore Court of Appeal in Denka Advantech Pte Ltd v Seraya Energy Pte Ltd.[1] The decision confirmed that Singapore remains firmly anchored to the century-old Dunlop test, even as other major common law countries have moved in different directions. Understanding this decision is not an academic exercise — it is a practical necessity for anyone drafting or negotiating commercial agreements.
Distinction between Liquidated Damages and Penalties
Liquidated damages are a pre-agreed, fixed or formula-based sum payable upon a specific breach of contract. Their core purpose is to provide commercial certainty and to save both sides the expense, delay and uncertainty of litigation over loss. They are especially useful when damages are inherently hard to quantify, such as lost profits from a delayed software launch, reputational harm, or supply chain disruption.
A penalty, in contrast, is a sum designed not to compensate but to punish or deter a breach, typically an extravagant, unconscionable figure that bears little relationship to the harm that would actually be suffered. The common law has long refused to enforce penalty clauses, because the purpose of contract damages is to compensate, not to punish. If a clause is struck down as a penalty, the innocent party must fall back on proving its actual loss, which can be a heavy burden.
Crucially, the distinction does not depend on what the parties call the clause. A label like “genuine pre-estimate of loss” will not save a penalty; a clause headed “penalty” is not automatically void. The court looks at the substance, the commercial context, and the proportionality of the sum at the time the contract was made.
The century-old Dunlop Test
For over a century, the global common law standard for distinguishing between these two categories came from the 1915 House of Lords case, Dunlop Pneumatic Tyre Co Ltd v New Garage & Motor Co Ltd.[2] Lord Dunedin established four classic guidelines to determine whether a clause is a genuine pre-estimate of loss or an unenforceable penalty:
- The Extravagance Test: The stipulated sum is a penalty if it is "extravagant and unconscionable" compared to the greatest conceivable loss that could actually flow from the breach.
- The Non-Payment Trap: If the breach consists simply of failing to pay a specific sum of money (e.g., an invoice), and the contract mandates that defaulting parties must pay a significantly larger sum as damages, it is automatically a penalty.
- The Single Lump-Sum Presumption: There is a strong presumption that a clause is a penalty if a single, uniform lump sum is made payable across a wide variety of breaches ranging from trivial infractions to total catastrophic defaults.
- The Estimation Difficulty Exception: A clause is not a penalty merely because precisely estimating the potential losses ahead of time is almost impossible. In fact, that difficulty is precisely why liquidated damages exist.
For generations, this framework meant that drafting enforceable damages required looking into a crystal ball, forecasting real financial losses, and ensuring the math remained modest and defensive.
UK’s shift to Legitimate Interests’ Test
In the UK Supreme Court's decision in Cavendish Square Holding BV v Talal El Makdessi[3] (paired with ParkingEye Ltd v Beavis[4]), the court refined the test. A clause is a penalty only if it imposes a detriment on the party in breach "out of all proportion to any legitimate interest" that the innocent party has in enforcing the contract.
This broader "legitimate interests" test allows clauses that protect commercial interests beyond pure compensation, such as maintaining a brand's integrity or business model. In Makdessi, a clause forfeiting earn-out payments upon breach of restrictive covenants was held valid, not as a pre-estimate of loss, but because it protected the buyer's legitimate interest in the seller's goodwill. In ParkingEye, an £85 parking charge was valid because the landowner had a legitimate interest in efficient use of parking spaces that went beyond mere compensation for a lost £2 fee.
For founders, this initially seemed liberating. A seemingly punitive clause in a founder-shareholder agreement might be saved by its legitimate purpose. However, the Cavendish test also introduced uncertainty: what is a "legitimate interest," and when is a detriment "out of all proportion"?
Singapore’s position in Denka Advantech
This was the question facing the Singapore Court of Appeal in Denka Advantech Pte Ltd v Seraya Energy Pte Ltd[5]. The case is a textbook example of how a seemingly sophisticated liquidated damages formula can still fail the classic test, and why Singapore, a critical jurisdiction for many founders in Asia, explicitly refused to adopt Cavendish.
The case concerned three electricity retail agreements entered into in 2012 between Seraya Energy Pte Ltd (Seraya), an electricity retailer, and its customers Denka Advantech Pte Ltd and Denka Singapore Pte Ltd. The electricity retail agreements were due to expire on 31 January 2021 and each contained a liquidated damages clause, linked to Seraya's right to terminate the agreements in various circumstances, including where Denka breached its obligations under the agreements. The liquidated damages were to be determined by a formula: A × B × 40%, where A was the number of months between the date the contract was terminated and 31 January 2021, and B was the average amount payable by Denka in the period before termination.
In August 2014, Denka indicated that it no longer wished to purchase electricity under the ERAs. Seraya thus terminated the ERAs and in December 2014 sued Denka for breaching their respective ERAs, claiming liquidated damages or, in the alternative, common law damages.
The High Court at first instance found the LD clauses unenforceable, ruling they were not a genuine pre-estimate of loss but were instead designed primarily to deter any breach by Denka. Seraya appealed.
The Court of Appeal reversed, and its reasoning contains several lessons worth unpacking.
- First, Singapore held the line on the Dunlop test. The Court had a choice: follow England's more flexible Cavendish approach or maintain the older Dunlop standard. It chose Dunlop. The Court found that the "legitimate interest" concept was too general and "protean"—capable of being used too flexibly, leading to excessive uncertainty for commercial parties at the time of contracting and for courts interpreting agreements later. The only "legitimate interest" the penalty rule is concerned with is compensation.
- Second, the scope of the penalty rule was confirmed as narrow. The Court held that the rule applies only where a clause, in substance, requires a party to pay a liquidated sum for breaching the contract. If your contract requires payment on a trigger event that is not itself a breach such as exercising an option or making an election, the penalty rule does not engage. This preserves freedom of contract for sophisticated commercial parties.
- Third, the 40% formula survived. Despite yielding a very large sum, the Court concluded that the amounts were not extravagant and unconscionable when compared to the greatest loss that could conceivably be proved to have followed from the breach. Expert evidence supported the calculation. The key takeaway: size alone does not kill a clause, proportionality does.
- Fourth, a large lump sum creates only a presumption, not a verdict. The Court emphasised that while a large sum payable gives rise to a presumption of penalty, the "greatest loss test" is of overarching importance. If the clause is not out of all proportion to the greatest conceivable loss, it is a genuine pre-estimate of loss.
Key takeaways
Given Singapore's strict adherence to the Dunlop test, founders must be meticulous when drafting liquidated damages clauses. The goal is always to ensure the clause is a genuine pre-estimate of loss and not a disguised penalty.
- Genuine Pre-estimate, Not Punishment
A valid LD clause must be built on a genuine, reasonable pre-estimate of your expected loss at the time of contracting. If you are a startup delaying a launch, don’t pluck a daily rate from thin air. Estimate lost revenue, wasted internal time, and direct project costs. Create a short internal memorandum or board note that records your calculation and the assumptions behind it. That paper trail could be decisive in a later dispute.
- Consider the Nature of the Breach
Different types of breaches may lead to different levels of loss. If a single sum is stipulated for various breaches of varying gravity, it risks being deemed a penalty. Where possible, tailor the liquidated damages to specific breaches or use a formula that reflects the potential loss for each type of breach, similar to the formula used in Denka Advantech .
- Consider Caps and Exclusive Remedies
A cap on liquidated damages can signal that the clause is a reasonable risk allocation rather than a punishment. Similarly, clarifying whether liquidated damages are the exclusive remedy for the relevant breach or whether common law damages remain available can prevent disputes over double recovery.
- Challenge penalty claims early
Founders often hesitate to dispute an LD claim, fearing legal costs. But as Denka Advantech shows, if the mathematics of the clause clearly diverges from the other side’s actual probable loss, you have a powerful defence. The penalty rule is a mandatory rule of law; you cannot contract out of it by writing “This is not a penalty.” The court will always look through the label to the substance.
Conclusion
The line between a liquidated damages clause and a penalty is not a bright one, but it is navigable. The core principle across jurisdictions remains consistent: compensate, don't punish. For founders, the practical takeaway is straightforward i.e. document your loss estimate, calibrate the sum to the breach, and show that the clause reflects commercial reality rather than coercion.
Denka Advantech reinforces that in Singapore, the century-old Dunlop framework endures. It is a test that rewards careful drafting and genuine commercial reasoning. Founders who understand this distinction will write contracts that hold up when they matter most in a dispute.
Denka Advantech Pte Ltd v Seraya Energy Pte Ltd [2021] SGCA 19 ↑
Dunlop Pneumatic Tyre Co Ltd v New Garage & Motor Co Ltd [1914] UKHL 1 ↑
Cavendish Square Holding BV v Talal El Makdessi [2015] UKSC 67 ↑
ParkingEye Ltd v Beavis [2015] EWCA Civ 402 ↑
Denka Advantech Pte Ltd v Seraya Energy Pte Ltd [2021] SGCA 19 ↑