The High Court’s recent decision in Andrews v Australia and New Zealand Banking Group Ltd [2012] HCA 30 establishes the broad reach of the common law rule and the equitable jurisdiction concerning relief against penalties and makes clear that these rules cannot be avoided through drafting alone. The case establishes that the doctrine will apply to any fee that in substance is designed to secure the performance of a contractual obligation, irrespective of whether the fee is actually triggered by a breach. 

As a consequence, a fee which secures compliance with contractual obligations will only be enforceable if it represents a genuine estimate of the loss that is likely to flow from a breach. This can prove challenging in an IT context, where the impact of a breach can be difficult to anticipate and the range of losses that may be incurred can vary widely. Nonetheless, in the wake of the Andrews decision, it is important for parties to IT contracts to take these matters into account and ensure that any fees are defensible in the event they are subject to challenge under the penalties doctrine.

Background to the penalties doctrine

As a general principle, contractual damages are intended to be compensatory rather than punitive. The penalties doctrine aims to protect this principle by preventing parties from agreeing a punitive fee which will be payable if the contract is breached. The effect of the doctrine is that a contractual provision which requires payment of a sum of money upon a breach of the contract will be unenforceable unless the amount of the payment is a genuine pre-estimate of the loss that will flow from the relevant breach. In other words, while the law allows the parties to a contract to pre-agree an appropriate amount of compensation for a breach, it does not allow them to agree upon a punishment for the breach.

Before Andrews the cases had confined the penalties doctrine to situations where the relevant liability for the payment was triggered by a breach of the contract. On this approach, the doctrine would not apply to a fee triggered by an event which was not technically a breach. This allowed informed drafters to avoid the application of the doctrine by making sure the event triggering liability to pay was not, in form at least, a breach of contract. For example, in the case of a service level credit designed to ensure compliance with contractual service levels, a drafter might avoid including an absolute obligation to achieve the relevant service levels in the contract. Instead, the contract may only require that the service provider use “best efforts” to achieve the service levels. Failure to achieve the service level would then not necessarily be a breach of contract. Accordingly any credit payable would not be subject to the penalties doctrine. However, the efficacy of this drafting approach has been undermined by the Andrews decision.

The High Court’s decision in Andrews

In Andrews, the High Court considered the enforceability of various fees and charges imposed by the ANZ bank, including “over limit” fees which applied when a customer had overdrawn their account. At first instance, the Federal Court found that these fees could not be treated as penalties, as ANZ’s customers were not under a contractual obligation to avoid overdrawing their accounts. In other words, the fees were not triggered by a breach of contract and so could not be treated as penalties. The High Court unanimously overruled the reasoning adopted in the Federal Court and said that relief against penalties is potentially available even if a fee is not payable on breach of contract. The High Court said that the proper approach to determining whether a fee is a penalty is to consider whether the substantive purpose of the fee is to secure performance of a contractual obligation. The matter was referred back to the Federal Court to determine whether, in fact, the fees charged by ANZ should be treated as penalties on this basis.

The High Court’s decision in Andrews is significant, as it means that the penalties doctrine may apply to invalidate fees that are not technically triggered by a breach. Accordingly, parties who have been relying on the drafting approach discussed above to avoid the effect of the penalties doctrine will need to revisit their standard form contracts to ensure that the service credits or other contractual fees on which they may be relying are not at risk of being found to be penalties and, therefore, invalidated. In the wake of the High Court’s decision, there remain two ways to ensure that an agreed fee is not treated as a penalty:

  • The first approach is to ensure that the fee is in substance a charge for further services or accommodation. For example, in Andrews the High Court approved an earlier decision by the NSW Court of Appeal in Metro-Goldwyn-Mayer v Greenham [1966] 2 NSWLR 717 regarding a contract for hiring of films, which only allowed one screening of the film by the exhibitor. The contract provided for a significant fee (of up to 4 times the original fee) to be paid if the exhibitor used the film for additional screenings. This fee was deemed not to be a penalty, as it was a fee for an additional accommodation in allowing the exhibitor to make additional screenings. It is possible that certain of the fees charged by ANZ could be characterised in the same way. For an example, an “over limit” fee could potentially be characterised as a fee for the provision of an additional service in the form of an overdraft facility. To rely on this approach to avoid the penalties doctrine, the party charging the fee should make clear in the contract that the counterparty will be receiving a benefit to which they would otherwise not be entitled if not for payment of the fee. If there is no additional benefit, then the fee may still be characterised as a penalty.
  • The second approach is to ensure that the fee is a genuine pre-estimate of the loss that the charging party will suffer as a consequence of a breach or non-performance by the counterparty. In other words, if the fee is genuinely compensatory, rather than punitive, in nature then it should not be treated as a penalty. The challenge with this approach is that it can be very difficult to estimate the loss that will be occasioned by a particular breach. This is especially the case in an IT context. For example, it will be a difficult task to pre-estimate the amount of loss that a customer may suffer if a service provider fails to meet its service level targets. What measurable damage will be suffered if the service is available 99% of the time rather than 99.9% of the time? Or if it takes 8 hours to fix a major defect rather than 2 hours?

Difficulties associated with estimating loss

The challenge left by the Andrews decision is that a party wishing to impose a contractual fee which in substance secures performance of an obligation under the contract must ensure that the fee represents a genuine pre-estimate of the loss that will flow from a breach or else it may be unenforceable. The case law provides some guidance as to what will constitute a genuine pre-estimate of loss for these purposes:

  • A sum that is “extravagant and unconscionable” or “out of all proportion” to the loss that may be suffered from a breach will not be a genuine pre-estimate. For example, a provision which requires a full refund of all service fees if service availability falls a few percentage points below the service level target may be out of proportion to the loss suffered, as it will suggest that the customer derived no benefit at all from the substantial time that the service was in fact available. A regime under which a partial refund is applied and increases as service availability drops (e.g. a 5% refund if availability is 0 to 5% below the target, a 10% refund if availability is 5 to 10% below the target and so on) is more likely to be enforceable as it will be easier to demonstrate some relationship between the amount of the fee and the seriousness of the breach.
  • Wider limits of discretion will be allowed in circumstances where it would be complex and expensive to establish with any sort of precision the amount of loss that is likely to flow from a breach. Unless some degree of latitude is allowed in these cases, the parties would effectively be prevented from pre-agreeing on appropriate compensation and would always need to prove actual loss (a process that can incur significant legal costs and take up valuable court time). The courts have indicated that this is not a desirable outcome.
  • Market conditions are a relevant consideration. For example, in Yarra Capital Group Pty Ltd v Sklash Pty Ltd [2006] VSCA 109 the Victorian Court of Appeal found that there was a real difference between the institutional lending market (where there are industry benchmarks and prevailing interest rates that can be used to establish the loss to the lender arising from a borrower’s default) and the short term money market (where loans are often unsecured, cost of borrowing is high and establishing the loss that is likely to flow from a repayment default is far more complex). This made it difficult to say that a high rate of interest on a short term loan was out of all proportion to the loss that may be suffered from a repayment default and, therefore, amounted to a penalty.
  • There is a presumption that a payment will be a penalty where a single sum is payable on the occurrence of one or more events, some of which may lead to substantial losses and some of which are trifling events. For example, if the same service credit applies for all service defaults, whether that be something as minor as failing to submit a regular monthly report on time or as serious as failing to remedy a major system defect, then there is a greater risk that the credit will be found to be a penalty. In this example, the lack of any direct relationship between the amount of the credit and the impact of the breach suggests that the credit is not a genuine pre-estimate of loss.
  • The fact that there is a significant difference between the fee required under the contract and the actual loss suffered will not automatically mean that the fee is a penalty. Whether or not the fee is a genuine pre-estimate of loss is to be determined at the time that the contract is made, not at the time when the loss is suffered.


The Andrews decision clearly shows that there is no easy “form over substance” way around the penalties doctrine. As a consequence, it is important that parties do not take short-cuts when designing their service credit regimes and other mechanisms under which payments may be required in order to secure performance of a contract. It is important that some attempt be made to draw a connection between the payment required and the likely consequences of a breach. Otherwise, there is a risk that the entire regime will be fatally undermined by a finding that the payments do not represent a genuine pre-estimate of loss and, therefore, are unenforceable penalties.