Plaintiff v Defendant [2024] VSC 788 | Case summary

Plaintiff v Defendant [2024] VSC 788 | Case summary | Zoe

The plaintiff ('ConnectEast') owns and operates the EastLink tollway ('EastLink'). The defendant ('CML') owns and operates the CityLink tollway ('CityLink').
This case concerns the interoperability arrangements between the parties and, more specifically, alleged overpayments by ConnectEast to CML under agreements providing for that interoperability. ConnectEast's claim is for restitution for monies had and received.
Interoperability is a word used to describe the ability of account holders of one toll road operator to travel (or 'roam') on any other toll road in Australia with a single account and have the accrued tolls charged to that account. These arrangements allow for CML account holders who have a Linkt e-tag in their cars to drive on EastLink without having to open a separate account with ConnectEast.
On 11 January 2008, CML and ConnectEast entered into agreements which provide for interoperability between them ('Roaming Agreements').
The Roaming Agreements provide, relevantly and amongst other things, that when CML's account holders travel on EastLink, CML will collect the applicable tolls and fees from their account holders, then pay those tolls and fees to ConnectEast, less an amount referred to under the Roaming Agreements as the 'Discount Amount'.
The Discount Amount is also referred to by the parties as the 'roaming fee', as it represents a fee paid for the provision of the service of collecting the applicable tolls and paying those tolls to the operator of the toll road travelled on. CML provides that service to ConnectEast, and ConnectEast pays a roaming fee to CML, when CML account holders travel on EastLink. In that circumstance and in this proceeding, that service is referred to as the 'Link roaming service'.
As a consequence of interoperability, CML (or Transurban, as will be explained) undertakes billing and payment functions for more than half of all toll road journeys (referred to as 'trips') taken on EastLink. Further, travel on EastLink accounts for around 30% of the total trips taken by CML account holders. Between 1 January 2009 and 31 October 2020, CML collected almost $2 billion of toll revenue on behalf of ConnectEast.
By the Roaming Agreements, the parties agreed the roaming fee in the sum of $0.18 per trip excluding GST, adjusted on 1 July each year by reference to CPI.
The roaming fee that may be imposed by CML for the Link roaming service provided to EastLink is capped by s 93AB(1) of the Melbourne City Link Act 1995 (Vic) ('MCL Act'). In short, that section provides that the roaming fee 'must not exceed the amount that represents the net incremental marginal cost to [CML] of providing that Link roaming service'.
The net incremental marginal cost is referred to as the 'NIMC' in these reasons.
By statement of claim filed on 23 November 2020 ('Statement of Claim'), ConnectEast alleges that at all times since 1 January 2009, CML has withheld a Discount Amount from (or charged a roaming fee to) ConnectEast in an amount that exceeds the legislative cap imposed by s 93AB(1).
By concise statement of defence filed on 12 February 2021, CML denies that the amount it withheld exceeded the legislative cap at any relevant time. In respect of claims concerning amounts paid on or before 23 November 2014, CML pleads defences based on the Limitation of Actions Act 1958 (Vic) ('LA Act').
The key issues in dispute are:

(a) in around 1995, technologies were developed such that electronic tolling through the use of an e-tag was possible. This introduced the concept of interoperability across toll road operators;

(b) CityLink opened in August 1999. The e-tag was introduced as a mechanism to collect tolls on CityLink using an electronic toll collection system known as 'GATe';

(c) in or about 2001 the then owners and operators of toll roads in Australia executed a document known as the 'Memorandum of Understanding Electronic Toll Collection' (the 'MoU'), which established the basis and framework for interoperability between toll road operators in Australia. CML has been a signatory since around 2001 or 2002;

(d) in or about 2003 or 2004, Transurban commenced planning for the replacement of GATe;

(e) on 14 October 2004, the concession agreement for EastLink between the State of Victoria, ConnectEast and others came into effect;

(f) in late 2006 to early 2007, CML and ConnectEast began negotiating interoperability arrangements for the CityLink and EastLink toll roads, including the Discount Amount;

(g) the Roaming Agreements between the two parties (discussed below) were executed on 11 January 2008;

(h) EastLink opened in July 2008, and ConnectEast commenced tolling one month later. The Interoperability Commencement Date under the Roaming Agreements was 27 July 2008;

(i) in the same year, ConnectEast also became a signatory to the MoU;

(j) on 1 January 2009, ConnectEast commenced paying the roaming fee to CityLink;

(k) in December 2011, the GATe system used by Transurban was replaced with a new tolling and customer accounting system known as 'GLIDe'. GLIDe is now used by Transurban for all its Australian toll roads and customers and is in the process of being rolled out across all its North American roads and customers;

(l) as at December 2016, 18 parties had executed the MoU; and

(m) in July 2018, Transurban rebranded all its Australian retail brands as 'Linkt'. There are around six million Linkt customers across Australia.

(a) the activities relating to finance (including tax, internal audit and treasury), risk, legal, executive, and other corporate functions of CML are carried out by Transurban;

(b) Transurban arranges physical space for its total workforce of 9,000 people and leases commercial spaces in Melbourne, Sydney and Brisbane. Transurban recruits and develops its staff;

(c) Transurban manages the daily processing and reconciliation of payments to and from ConnectEast that relate to roaming trips by Linkt customers;

(d) the incursion of merchant fees by CML customers who travel on CityLink and EastLink are incurred by Transurban;

(e) Transurban is responsible for the operation of the GLIDe tolling system across all Transurban's toll roads which involves activities to maintain Transurban's IT systems including cybersecurity of Transurban's systems, hardware and software to support Transurban's employees, managing the operation of Transurban's IT systems (including responding to GLIDe production issues such as system changes or major incidents), managing storage for Transurban's systems and providing network connectivity for Transurban's systems;

(f) Transurban's retail business has a single brand (called Linkt, as noted above) for its supply of electronic tolling services to motorists in Australia. That is, irrespective of where customers live and travel in Australia, including CityLink, their account is with 'Linkt' - they are 'Linkt' (i.e., Transurban) customers;

(g) Transurban built and now operates and maintains the Linkt app, which is the principal means by which Linkt customers (including CML account holders) communicate and interact with Transurban in relation to their accounts. It also developed and operates other Linkt digital platforms, including the Linkt website; and

(h) a call centre is operated by Transurban and staffed by its employees (in addition to an offshore team based in Manila which is contracted by Transurban) to receive contacts from Linkt customers. Customers receive support irrespective of the toll road used by the customer.

The parties, referring to and relying on the opinions of their respective experts, also agreed that costs that can be described as either incremental costs or avoidable costs should be taken into account in the estimation of the NIMC.
Incremental costs to an entity of providing a 'service' refers to the additional cost incurred in providing the service in question, compared to the counterfactual where that service is not provided but the entity continues to provide all of its other services. Specifically in relation to the provision of the Link roaming service, the incremental cost to CML comprises those costs that CML would not otherwise have incurred, and so excludes any costs that CML would have incurred in any event to provide retail services to its account holders or roaming services to other toll road operators.
Avoidable costs are essentially the same as the concept of incremental cost, the difference being that it focusses on the extent to which costs would be avoided if the service in question was no longer provided.
This characterisation of costs as avoidable and incremental was relevant to the experts' consideration of the most appropriate methodology for determining the amount representing the NIMC. As set out in section F.1.5, the experts agreed that methods applicable to the estimation of incremental or avoidable costs were appropriate to estimate the sum representing the NIMC. However, the parties also agreed that those methodologies may fail to account for some common costs that should be included in the NIMC and that a different methodology would need to be applied. The only example of such common costs that arose in this case were the costs incurred by CML adding roaming capability to its tolling system. The assessment of those additional costs is addressed in section G.3.
Against that background and in my view, regardless of how the parties may choose to characterise a particular cost - e.g., direct, indirect, common, incremental or avoidable - it will only be included in the NIMC:

Professor Lyon describes the methodology he employed to estimate the NIMC in detail in his first report. He provided further information in his reply report and by his presentation to the Court during trial.
A summary of the methodology employed by Professor Lyon is set out in the following paragraphs.
As set out above, the experts agreed that to identify the additional costs incurred by CML (which they tended to refer to variously as the 'avoidable' or 'incremental' costs) it was necessary to determine the costs which would be incurred in a counterfactual scenario where the Link roaming service is not provided. The difference in costs between the factual scenario where it is provided and the aforementioned counterfactual scenario represents the 'incremental cost' of the Link roaming service.
Professor Lyon explained that notwithstanding that he was asked to estimate the costs to CML, it was necessary to have regard to the national cost accounts of Transurban which he says account for a significant portion of shared costs that are directly relevant to the Link roaming service.
He further explained that as Transurban did not maintain discrete accounts for the costs of the Link roaming service, the methodology he adopted required the application of judgment to identify and assess the costs to be included in the estimate of NIMC. In his first report, he opined that because the Link roaming service is a 'small, "add on" function performed for another tollway company' the costs of performing that service are not separately accounted for, and therefore, the necessary task is 'to identify and evidence particular portions within larger shared cost accounts [which are] driven by the Link roaming service.'[34] For this reason, he identifies 'cost drivers' as the necessary mechanism to isolate the additional cost, acknowledging that this 'introduces complexity and invites potential subjectivity in the estimation of costs'.[35]
While Professor Lyon considered the Link roaming service to be a small add on function from Transurban's perspective, CML submits that the Link roaming service was a significant component of CML's business.
Professor Lyon described his methodology as a 'bottom-up' approach,[36] involving the following three stages, in summary:

The Transurban Model refers to the Shared Retail/Asset Service driver as 'Interoperability Retail Shared Service' which reflects the percentage of retail trips as a proportion of total trips (being between 58% and 60% in the relevant years). The Lyon Adjusted Transurban Model similarly refers to the driver as 'Interoperability Retail Shared Service'. However in his reports Professor Lyon refers to it as 'Share Retail/Asset Service'.
Professor Lyon modified this driver. He opined that a more appropriate driver took into account both the proportion of retail trips compared to total trips and the proportion of full-time employees engaged in retail functions compared to those engaged in asset functions.
Of the 1,542 full-time employees engaged in 2019,[43] 218 only worked on the retail side of the business, 284 only worked on the asset side of the business, and 1,040 worked across asset and retail functions. These figures were provided by Transurban in a spreadsheet, and were relied on by Professor Lyon. I note that no explanation was provided about the actual activities carried out by the employees.
In order to estimate the proportion of each cost centre that should be included as a retail cost, Professor Lyon apportioned the 1,040 employees evenly across retail and asset functions (i.e., 520 in each). Professor Lyon then used those figures to determine the proportion of employees working on retail functions: ((520+218)/1,542) × 100 = 48%.
Professor Lyon then 'blended' the 48% figure with 59% (Transurban's Interoperability Retail Shared Service driver for 2019) to arrive at 54%: Professor Lyon's allocation for the 'Shared Retail/Asset Service' driver for 2019. In other words, he applied the half-way point between Transurban's figure of 59% (derived by reference to the proportion of retail trips to total trips) and his own figure which factors in the headcount within retail as compared to asset functions (48%).
Professor Lyon did not clearly set out in either of his reports how he arrived at the percentages he allocates for this cost driver in each year. In its closing submissions, CML sets out the precise formula it says Professor Lyon adopted, which is reflected in the description given above. That formula accords with Professor Lyon's comments during his presentation at trial that the calculation involved a 'blending' of two percentages and I am satisfied that he adopted that process.
His adjustment resulted in a driver that retained between 53% and 55% of the costs to which it was applied in the relevant years.

(a) Customer driven engagement with Transurban is typically provoked by trips.

(b) The services provided by the billing team have a direct link and scale with the Link roaming service.

(c) The activities of the payments function are directly related to trip volume and the number of customer accounts. As trip volume increases, the frequency of account top-ups and payments increases. An increase in trip volume also leads to an increase in the number of accounts that become overdrawn, which leads to an increase in customer communications.

(d) The costs of managing overdue accounts and counterparty risk[59] are directly related to the number of trips and road users. It is submitted that as trips increase, Transurban's credit risk increases, together with the risk of unpaid tolls, which generates work.

(e) There is a close correlation between digital interaction and trip volumes.

(f) The costs associated with providing customer-facing digital services scale with the volume of trips taken by CML account holders (regardless of which road they travel on), because more trips means more toll payments generated through the website and mobile app, which in turn requires more processing power and storage.

(g) As Transurban's business grows, its cloud and hosting requirements also grow. If there are more trips taken by CML account holders (including on EastLink), more cloud and physical storage is needed.

(h) While baseline software licence costs are set by contracts with software providers, these costs are impacted by trip and customer volumes. More trips produce more data. Over time, this means greater processing power is required and therefore, software licensing costs may increase.

(i) Additional staff are required to support the increased trip volumes that flow from roaming, which means additional IT services and equipment are required to support those staff.

(j) Increased trips by CML account holders means increased call centre activity, which in turn increases the IT activity required to support the call centre.

(k) Generally, the cost of maintaining and managing CML technology and infrastructure is directly proportional to the number of customers and the number of trips they take. For example, the storage that is required to record these trips, the computational power needed to process transactions and the network bandwidth needed to transmit information increase with the number of trips. Each of these elements determine costs. This means that an increase in each of these elements increases cost. If CML account holders did not take any trips on EastLink, CML could reduce those costs.

(a) GLIDe development costs. GLIDe is the tolling and customer accounting system used by Transurban for all of its Australian and North American roads and customers. It was developed and operational by 2011. ConnectEast submits that the costs of development are therefore fixed in relation to the four sample years (2015, 2017, 2019 and 2020), regardless of the number of trips taken in those years.

(b) Software licences. Mr Chaitow gave evidence that Transurban has many different software licences and the basis on which the licence fees are set varies. Most commonly the fee varies based on the number of Transurban employees using the software. For some there is a fixed cost that does not vary with number of employees. There are some that are wholly unconnected with the number of trips taken on EastLink or other roads.

(c) Websites and apps. The cost of developing the website and apps that customers use are fixed. As Mr Manrique explained, they require the same features and functionality, regardless of the number of customers that use them, or the number of trips that customers undertake. It is only the cost of processing power that may increase with trips. There is no evidence as to whether the increase in processing power is material.

(d) E-tags. CML's evidence about the relationship between costs associated with e-tags and trips was inconsistent. Professor Lyon opined that battery life of an e-tag was depleted due to trips and so there is a direct, proportionate relationship between trips and e-tag costs. Mr Robinson, Billing & Credit Manager at ConnectEast, gave evidence that trips contribute to the diminution of battery life in e-tags. On the other hand, Mr Jackson gave evidence that while there is a relationship between trips and e-tags costs, it is not direct and proportional. Specifically he explained e-tags deteriorate with age, and the more trips account holders take without a working e-tag, the more fees they pay and therefore the more likely they are to ask for a new e-tag.

(e) Call centre costs. ConnectEast submitted that the evidence supported the proposition that the number of contacts through the call centre may not vary in proportion to trips:

(i) Transurban responded to a question about the cost centre '110315 Call Centre VIC', stating:

[n]o, [c]ontact centre drivers are not all related to trips, customers reach out to the contact centre for various reasons and trip related contacts are one of many. All contact centre staff capture contacts based on the region, asset and the contact driver and [Transurban has] used the same data to calculate the allocation drivers.[63]

(ii) Transurban also contacts account holders proactively with a view to keeping customers engaged, ensuring that accounts are in good health and that customers credit card details are up to date.

(f) Processing and reconciliation of payments. According to Ms Hebrard, Transurban has approximately one full-time employee working on the daily processing and reconciliation of payments in respect of roaming trips on EastLink. Key tasks include checking that interoperability files have been received and ingested into Transurban's systems, reviewing requests for daily payments to ConnectEast and amounts to be paid, and going back and forth with finance to confirm the payments and bank reconciliations. None of those tasks vary in size based on the numbers of trips. The work undertaken to process payments, and reconcile them, is a batch process unconnected with the number of trips.

(g) Privacy obligations. According to Mr McKay, Transurban incurs legal costs connected to privacy of information received from customers who travel on EastLink and CityLink. He said Transurban must comply with credit card legislation on privacy. However, Mr McKay accepted that Transurban would incur such expenditure whether or not there was any interoperability. Mr McKay was not aware of there being any ratio between trips (of any kind) and costs incurred by Transurban to maintain its privacy obligations.

(h) Internal audit. Mr McKay gave evidence that his 'expectation' is that the provision of roaming services generally 'may' add 'some' time to the internal audit process, but he did not suggest that providing the Link roaming service would do so.[64] It is not connected at all to the number of trips that occur (on any road).

(i) Risk Management, health and safety. Mr McKay described some of the services provided by the risk team:

(i) Regular checks that CML is complying with the Roaming Agreements.

(ii) Ensuring compliance with the MCL Act.

(iii) Ensuring compliance with the Payment Card Industry Data Security Standard.

(iv) Preparing a risk management framework for GLIDe.

(v) Managing the risk associated with Transurban's call centres.

(a) The activities the subject of the graphs comprise a fraction of the activities undertaken by the many cost centres to which the Transurban Model is applied and accordingly, the graphs say nothing about the relationship between trips and the vast majority of the costs to which the Transurban Model is applied.

(b) The graphs do not show that costs increase in proportion with trips:

(i) None of the graphs depict a relationship between trips and costs. They compare trips to some other activity. Some costs of the activity may be fixed. There may be significant economies of scale or costs may move in some more nuanced way.

(ii) Mr Balchin explained that when there are multiple variables that might cause some activity to occur, the only way to discern whether there is a relevant causal relationship is through statistical analysis.

(iii) There are clearly causes other than trips for the activities referred to in the graphs. For example, customer contacts. Transurban has said that customer contacts occur for various reasons other than trips, such as when vehicles are bought and sold, or when customers move house, change phone numbers or require a new e-tag. Another example is the expansion of a Transurban program known as 'Linkt Assist' that Transurban undertook in response to the COVID-19 pandemic and the bushfires. It resulted in increased proactive communications and a 25-fold increase in requests for financial support assistance through the Linkt Assist team.

(iv) Each graph presents data for a limited window of time over the unusual COVID-19 pandemic period. From data briefed by Transurban to Professor Lyon, ConnectEast was able to prepare graphs that extended earlier in time (prior to the COVID-19 pandemic). Some of those graphs suggest that the correlation between trips and the relevant activity outside the COVID-19 pandemic period was weaker and perhaps non-existent. Under cross-examination, Mr Jackson agreed that he could not see a correlation between trips and customer communications, trips and e-tag volumes or trips and digital interactions.

(v) The graphs only show trips on Transurban roads. They do not show any roaming trips on EastLink. It is unknown whether any relationship they show would remain the same or vary in some way if they were to include or show only trips on EastLink.

ConnectEast submitted that contrary to that approach, for the cost centres that pass Professor Lyon's incremental cost 'screen' at Stage 1, he thereafter simply allocated a share of the total cost in that cost centre to the Link roaming service (by application of cost drivers in Stage 2 of his analysis). No analysis or evidence is presented to suggest that the allocators or drivers that he employs will isolate the incremental cost associated with the Link roaming service. That is, Professor Lyon has not actually sought to quantify the incremental cost but rather has applied 'drivers' that are (subject to a small number of exceptions) arbitrary and unrelated to why the cost centre in question may experience an increase in cost. For example, one of Professor Lyon's justifications for his application of the Trips Driver at Step 3 is that it indicates the 'proportional use and benefit derived by ConnectEast'.[68] But that is distinct from a method that will provide an estimate of the incremental effect on that cost item of providing a particular service. Cost items that are fixed may be used by the various services including the Link roaming service, and so provide a benefit to the Link roaming service, but may not be incremental to the Link roaming service (or in other words, are not 'avoided costs' with respect to the Link roaming service).
Mr Balchin did not think it was possible to estimate the incremental cost of providing a small, add on service by commencing with the full management accounts of an entity (and especially one the size of Transurban) and then applying cost drivers.
Further, it was observed by Mr Balchin that Professor Lyon's first criterion only considers if a cost relates to CML's roaming services as a whole and not with specific regard to the service provided to ConnectEast (i.e., the Link roaming service). He opined that, as a consequence, more costs are included in the calculation of the roaming fee than are incremental to the Link roaming service.
Mr Balchin further observed that it is therefore likely that Professor Lyon materially over-estimated the incremental cost of providing the Link roaming service because he does not exclude costs that might be used by and for the benefit of the Link roaming service but are not avoided costs with respect to the roaming function overall and specifically with respect to providing the Link roaming service.
Third, it was submitted that the third criterion applied by Professor Lyon at Stage 1 of his analysis was to ask whether there was a 'strong' driver to isolate and attribute a portion of each cost centre in question to the Link roaming service.
ConnectEast submitted that for the reasons that follow, Professor Lyon should not have been satisfied that the drivers employed by Transurban (including those few that he modified) were capable of isolating relevant costs from any cost centre perfectly or even approximately.
At Step 1 of Stage 2 of his methodology, Professor Lyon adopted the Transurban Model with almost no modification. ConnectEast submits this aspect of the Transurban Model and the Lyon Adjusted Transurban Model is inadequate for the task in two material respects.

(a) on 2 December 2021, Mr Balchin requested ongoing access to certain documents within CML's books, access to the equivalent documents in Transurban's books and the documents relied upon by Professor Lyon in his first report;

(b) on 8 December 2021, Transurban noted in relation to Mr Balchin's request for documents in Transurban's books:

... we understand that the requirement is for access to be provided to the 'books of the defendant' (being CML, as opposed to the 'books' of Transurban Limited or any other entities in the Transurban Group). However, we will try to accommodate any reasonable and relevant requests to the extent possible. Accordingly, if you require any additional information from a non-party Transurban Group entity, we ask that you please provide a more detailed explanation of what specific information you require, and why it is required to calculate CML's NIMC for the four identified years. That will enable me to seek input from the relevant persons in respect of that request and ensure you have what you reasonably need to prepare your report. In case it assists, I note that the Transurban Group level figures for the relevant cost centres to which Mr Lyon referred in his calculation of the NIMC are already included in the material provided to Mr Lyon and are available on the database hosted by KWM.[74]

(c) on 14 December 2021, Mr Balchin provided a more detailed memorandum to Transurban requesting, among other things:

(i) the Lyon Adjusted Transurban Model which had not been included in Professor Lyon's report;

(ii) the underlying values used to determine the percentage values of the cost drivers in each of the Transurban Model and the Lyon Adjusted Transurban Model and information about the differences between them;

(iii) documents recording the total number of trips taken by CML customers and customers of other toll roads;

(iv) the original and current functional specifications for GLIDe (Transurban's electronic tolling system) and information about any functionality added after GLIDe became operational;

(v) information concerning the implementation of roaming on other toll roads in Queensland and New South Wales; and

(vi) information about customer statements and invoicing as well as call centre interactions, such as the number of call centre interactions for various categories of callers (including account holders and toll invoice customers);

(d) on 4 February 2022, Transurban provided its responses to the categories of information requested and made some general observations, including that:

... a number of requests relate to information at a level of granularity that simply does not exist at CML or Transurban due to the way in which the business is structured and operates.[75]

(e) specifically, Transurban said the business does not organise itself or operate in a way to enable the following information to be tracked:

(i) all operational costs associated with roaming fee activities;

(ii) roaming trip interactions;

(iii) roaming trip customer escalations; and

(iv) the debts associated with roaming trips compared with CML asset trips;

(f) furthermore, it said:

(i) CML and Transurban do not price operations or technology on a functionality basis (so the interoperability functionality cannot be neatly separated from the overall system functionality);

(ii) neither CML nor Transurban captured records in relation to the cost of adding the roaming function to the GLIDe system;

(iii) neither CML nor Transurban capture information regarding the organisation of staffing for the roaming function, what activities staff members perform for that function, or how many hours each employee spends on a daily basis to discharge CML's roaming function with ConnectEast as well as other toll road operators; and

(iv) neither CML nor Transurban capture the number of call centre interactions for the categories of callers identified in Mr Balchin's request;

(g) on 16 February 2022, Mr Balchin commented on Transurban's responses, identified issues for further discussion and requested a 'follow-up discussion to clarify the information [he had] received, and for those areas where information is unavailable, to discuss what alternative information may exist and be sufficient.'[76] He noted that a transcription of the discussion would be helpful to provide a comprehensive record of matters discussed which he could use and rely upon to finalise his report; and

(h) further letters were exchange on 22 March 2022, 30 March 2022 and 7 April 2022.

(a) Roaming-specific activities which he characterised as activities that would not be undertaken if the Link roaming service were not provided. The two activities he identified as falling within this category are:

(i) creating the capacity for CML's billing and customer management systems to provide the roaming service. Most significantly, the capacity required to be able to import information about trips undertaken by CML account holders on toll roads other than CityLink (such as EastLink) and for that information to appear on customer bills; and

(ii) the daily transfer of information between CML and ConnectEast about valid accounts, trips undertaken, and total amounts of money transferred. Mr Balchin observed that while this task is automated, it requires staff oversight to address potential errors, check the outcomes of the automated systems, effect the daily payments to other toll roads and to confirm receipts.

(b) Expansion of customer retail activities. Mr Balchin opined that the Link roaming service leverages off customer service activities and functions that CML would provide in any event, however, there are five areas where a material increase in cost may be expected as a consequence of providing the Link roaming service, namely:

(i) additional customer communications and statements;

(ii) additional costs incurred to receive payments (i.e., commissions payable) associated with the additional billings;

(iii) additional contacts to call centres, and resulting customer issues to be resolved;

(iv) additional e-tags required to be issued, maintained and replaced; and

(v) additional credit management costs associated with the additional billings.

(c) Consequential wider firm activities which he characterised as activities outside those areas of the firm that are more directly involved in providing the Link roaming service or customer retail services but which may cause additional costs, such as:

(i) the wider costs associated with the additional employees caused by the provision of the Link roaming service (i.e., IT, IT support and human resources); and

(ii) corporate costs caused by the provision of the Link roaming service.

Fifth, and related to the fourth reason, his methodology was inappropriate for the task. Having regard to the information that was available and its limitations, he should have adopted an alternative methodology. Instead when it is revealed to him that the data he is looking for doesn't exist because Transurban 'doesn't structure its accounts that way' and it 'doesn't have to, he adheres ... resolutely and improperly to the methodology for which he now knows he has no data'.[82] Further his methodology was not suited to the case as pleaded. That is, the plaintiff pleaded that CML had taken steps to reduce costs since the Discount Amount was agreed and so what used to be a fair fee has ceased to be so. This case has not been conventionally run. Instead of reviewing the information about cost accounts provided by Transurban and determining if ConnectEast's pleaded assertion, that things got cheaper, was right, ConnectEast 'started from scratch'.[83] Senior counsel for CML characterised ConnectEast's approach as effectively saying if Transurban hasn't organised its affairs and structured its accounts in a way that permits ConnectEast to adopt the methodology it wants to, well 'that's [Transurban's] fault and therefore [ConnectEast will] go skinny on the NIMC'.[84] CML submitted that such an approach is grossly unfair, but equally CML made a point of submitting that it does not say it was impermissible for ConnectEast to adopt an approach which departed from its pleaded case.
Sixth, Mr Balchin did not apply his methodology comprehensively, invariably or consistently. Rather it is a patchwork of approaches. In the context of Mr Balchin's approach to estimating the attributable cost of the GLIDe system, CML says 'it is not a methodology at all'.[85]
There were other criticisms of Mr Balchin's estimate, criticisms that were directed to the application of his methodology (specifically to the validity of particular components of his calculation) and therefore the submission that his NIMC estimate cannot be relied upon and cannot be resurrected. These are addressed in detail in section G. In summary, it was submitted:

(ii) there is no basis upon which I could assess the reliability of the customer drivers employed in the Lyon Adjusted Transurban Model. As submitted by ConnectEast, there does not appear to be any instruction provided to Professor Lyon that explains how those drivers were created to enable him to assess them as apt to apply to any particular cost centre;

(iii) the key driver employed by Professor Lyon at Step 2 of Stage 2 was the 'interoperability shared services driver'. It accounted for around 75% of Professor Lyon's NIMC estimate for 2020. That driver, as modified by Mr Balchin, was based on a combination of relative trips and headcount in the Transurban business. There is no clear reasoning expressed by Professor Lyon as to how that allocator is suitable for the purpose of isolating incremental costs of providing the Link roaming service. As expressed by Mr Balchin, '[s]uch an allocator bears no relationship to the specific requirements of the roaming service or to the characteristics of the relevant cost centre (and, specifically, how those costs may change in response to roaming)'.[88]

(iv) the use of the Trips Driver at Step 3 of Stage 2. Professor Lyon opined that at Step 3, the Trips Driver 'reflect[s] the proportional use and benefit derived' by ConnectEast.[89] However, the task undertaken by the experts is not to assess the proportional benefit derived but rather to estimate the incremental cost to CML of providing a particular service. Further, and in any event, using that driver in relation to every cost centre assumes that all costs within every cost centre to which it was applied (i.e., all those that have survived the application of Steps 1 and 2 of Stage 2) vary based on the number of trips. I am not satisfied that is the case. While the evidence did reveal there may be a relationship between trips and some activities undertaken by Transurban, it fell short of establishing that all costs within the cost centres to which the trips driver was applied had a direct and proportional relationship to trips. As Mr Balchin observed in his first report 'it is self-evident that many of the costs that CML incurs as a retailer do not in fact scale with the number of trips and so applying such an allocator cannot hope to provide an estimate of incremental cost'.[90]

Further, and in any event, I prefer the methodology proposed by Mr Balchin.
First, I found Mr Balchin to be a most impressive expert witness. His report set out his opinions in a clear and logical manner. He was considered in his responses to my questions and under cross-examination, taking time to reflect on each question before answering. He was clearly frustrated at times by the lack of available information about costs incurred by Transurban and the resulting imprecision in his estimate of the NIMC. He made appropriate concessions about that imprecision and where relevant, these are addressed in section G. Those concessions were to his credit.
Second, Mr Balchin's extensive experience in relation to estimating incremental costs is not in dispute. The methodology he identified and implemented is clearly directed to that task.
While Mr Balchin's methodology is simply stated, the application of it to each category of activities he identified did vary. It varied depending on the information available to him about the costs incurred. As set out in some detail in section G below, there was a dearth of information about the costs incurred by CML in relation to the activities identified by Mr Balchin. This was a consequence of the manner in which Transurban managed its accounts and the very limited information it was able to produce in response to the reasonable requests made by Mr Balchin. In relation to each activity Mr Balchin explains how, drawing on his expertise and in the absence of more specific information, he has estimated the additional costs to CML. In section G below I have addressed Mr Balchin's estimates in relation to each of the activities identified by him and how they were undertaken. For the reasons set out in that section I accept his methodology and his application of it as appropriate for the purpose of ascertaining the amount that represents the NIMC. His approach was not arbitrary. It involved the formation of an opinion based on such evidence as there is that bears on the question in each case. His reasoning was clearly articulated. There is a level of imprecision in his estimate as a result of the limited information available. However, I am not persuaded that it is so imprecise that his method should be rejected.
I reject CML's submission that NIMC should be construed in a way that makes it regularly calculable, without excessive inconvenience and that 'it is improbable that the legislature intended to impose a regulatory regime that required CML to account for the roaming costs on such a granular or separated basis, as contemplated by Mr Balchin.'
The legislature chose to impose a cap. It expressly provides two methods by which the NIMC may be calculated in s 93AC of the MCL Act being:

(a) He requested any documents describing the system requirements needed to implement roaming or any documents or data that record the actual costs incurred by CML in relation to CML's original implementation of roaming on Sydney toll roads in around 2000. Transurban responded that the requested information does not exist stating:

[b]ecause of the way our tolling systems (including the GLIDe platform) are designed, we are not aware of any situation in which the implementation of interoperability for a specific toll road (or even a region) would need to be described or scoped specifically.

Further to this, CML and Transurban do not price operations or technology on a functionality basis (so the interoperability (i.e., roaming) functionality cannot be neatly separated from the overall system functionality).[101]

(b) He requested the same information in relation to CML's subsequent implementation of the function required for roaming on toll roads in Queensland and then EastLink. Transurban again responded that the requested information does not exist and added:

[t]he roaming capability is inherent in the GLIDe system design (i.e., the GLIDe platform was built with the requirement that roaming services be provided to other toll roads including EastLink and therefore, forms part of the overall cost of that system).

Accordingly, we can confirm that there is no separate or incremental cost incurred by Citylink that is attributable to the extension of interoperability to any specific toll road.[102]

(c) He requested the functional specification for the GLIDe system including the original specification and changes made during the development of the GLIDe system and changes made after the commencement of its operation. Transurban responded that having undertaken reasonable searches it had not located any documents in relation to the original specification and change requests. In relation to the request for details of any additional functionality added since that time, Transurban noted that:

as a key technology asset for its business, the GLIDe platform receives regular routine technology and software upgrades (including 'bug fixes'), some of which would 'relate' to the roaming functionality. It would be extremely time-consuming (if even possible) to identify each of these upgrades that in some way 'related' to the roaming functionality, which is an integrated part of the GLIDe system.

Further and in any event, any additional functionality added to GLIDe has no impact on CityLink's costs of the platform, as the licence fee was fixed in 2010 and escalated at CPI.[103]

(a) Calculation of depreciation ('Annualised cost of GLIDe system - depreciation only'). He took the original cost of the GLIDe asset, as recorded in the Project Aston Report prepared by Ernst & Young (as it was then known). On the basis of the information contained in the Project Aston Report that Transurban proposed to charge an annual licence fee over a 10-year period to recover the cost of the asset, Mr Balchin assumed that the payment that CML makes to Transurban for the use of the GLIDe system represents depreciation on the original cost of the asset. He therefore allowed for the amount to be depreciated over a 10-year period with the depreciation allowance escalating by 3% per annum.

(b) Calculation of commercial return on capital ('Return on capital for GLIDe system'). He then sought to determine the commercial return on capital which he calculated by multiplying the '[c]apital base of GLIDe system' (which he also referred to as the 'opening nett book value' as it existed at the points in time for each relevant year[108]) by a percentage referred to in the below table as the '[c]ost of capital for GLIDe system' (which he variously referred to as the 'commercial rate of return' and 'profit line'[109]). That percentage is derived from nine inputs including risk premiums, costs of equity and debt and the corporate tax rate.

(c) Calculation of the annualised cost of GLIDe ('Annualised cost of GLIDe system - depreciation and return'). To determine the annualised cost, Mr Balchin added the outputs of (a) and (b) above.

(d) Calculation of 'Proportion incremental to roaming'. He estimated the proportion of the annualised cost of GLIDe referred to in (c) that represents the costs of developing the roaming capability in GLIDe (i.e., the 2.00% figure in the table below). The derivation of the figure is addressed in paragraphs 341 to 348 below.

(e) Calculation of 'Proportionate use by CML'. He shared the costs of GLIDe between CML and other toll roads noting that over the relevant period of analysis GLIDe was extended to all of Transurban's toll roads. His method is explained in Appendix A of his first report. In summary, prior to 2015 CityLink was the only Transurban toll road that had been onboarded to GLIDe and therefore the relevant proportion he attributes to CML is 100% albeit that this ignores the toll roads that were onboarded to GLIDe during that year. By the end of the period, all of Transurban's toll roads had been onboarded. Mr Balchin assumed a 50% share for 2020 which is based on CML's share of retail trips for that year of 44% but reflects that the final three toll roads were onboarded during the course of the year. He then assumed a linear decrease in CML's share of trips processed in the GLIDe system during the intervening years.

(f) Calculation of 'EastLink's share of roaming by CML customers'. Mr Balchin then attributed what he considered to be a reasonable share of CML's costs (identified in (e) above) to ConnectEast by reference to EastLink's share of roaming trips by CML customers.

Mr Balchin considered Transurban's description of the functions that GLIDe performs and the associated diagram. He opined that the material suggested 'roaming (interoperability) is a small component of the overall capability of the GLIDe system'.[111] However, he noted that those materials were of 'a very high level in nature and do not provide a basis for estimating quantitively the incremental cost of adding the roaming capability'.[112] During cross-examination he explained that he 'relied upon it as a cross-check as to how extensive interoperability appeared to be ... as a proportion of the overall capability'.[113] He said it provided 'assurance that the numbers should be a reasonably small proportion of the overall total' but otherwise the information was 'not particularly useful'.[114]
The second document he referred to, entitled 'GLIDe System Capabilities - Final', was produced by CML in response to Mr Balchin's request for a functional specification of GLIDe and which CML described as a 'general and high-level description of some of the functionality of the GLIDe system'.[115] Mr Balchin considered the document served to 'reinforce his belief that the addition of the roaming functionality was only a small component of the cost of GLIDe'. By way of example, he noted that of a list of 85 functionalities of GLIDe contained in the document, only two relate to roaming (i.e, 2.4%).
The third document was a letter from ConnectEast to CML dated 21 June 2007 which set out ConnectEast's understanding of the terms on which CML was offering to provide a roaming service to ConnectEast. It contained a table of the fees and charges proposed by CML, including a fee of a around $5 million representing the '[u]pfront fee to cover CityLink's pre-operational costs'.[116] Incorporated into that fee was a sum of $2 million for 'IT Costs'. Mr Balchin relied on that one line entry as an estimate by CML of the cost to upgrade GATe (not GLIDe) to provide for roaming. Mr Balchin noted that the $2 million estimated by CML to upgrade GATe amounted to approximately 1.3% of the GLIDe development costs.
Having regard to those matters Mr Balchin estimated that the roaming functionality would have added, conservatively and approximately, 2% to the cost of the GLIDe system. He reasoned that this was higher than the $2 million estimated by CML in 2007 as being required to upgrade the GATe system for EastLink and consistent with the extent of coverage of the roaming service in the document entitled 'GLIDe System Capabilities - Final'.
He noted that 'very little information that is relevant to the cost of adding the roaming capability has been made available to assist with this part of my task'.

(a) The 2% multiplier is flawed for the following reasons:

(i) Mr Balchin relied on three documents for this 2% figure.

(ii) In relation to the first document Mr Balchin relied upon (referred to a paragraphs 342 to 344), Mr Balchin conceded during cross-examination that it provided no basis for coming up with a percentage, let alone 2%.

(iii) As to the second document (referred to a paragraph 345), Mr Balchin conceded that the Court could place no reliance on this document, he realised this after reading the transcript of ConnectEast's opening submissions and before his presentation and he should have informed the Court that he had changed his view.

(iv) As to the third document (referred to in paragraph 346 above), Mr Balchin's reliance on a one-line entry in a letter stating that $2 million was estimated in 2007 as being required to upgrade GATe (not GLIDe) for EastLink is not the application of economic theory but a piece of supposed logic that is either acceptable or not. It was further submitted that I should not accept it because Mr Balchin is not an IT expert and he does not know what was involved with the $2 million worth of work done on GATe**.**

(b) The second percentage multiplier (60% for 2019) is intended to '[scale] down the incremental-to-roaming portion of the annualised cost ... by an amount that reflects the proportionate use of the GLIDe system that is made by CityLink, noting that GLIDe was progressively extended to all of Transurban's toll roads over the period of the analysis'.[117] CML submitted that if the cost of GLIDe is to be spread across all of the roads that use GLIDe, the cost of onboarding those roads must also be included. Those costs are set out in Mr McKay's supplementary witness statement. The figures are taken primarily from Transurban's fixed asset register, totalling $163,990,411, and were excluded in their entirety from Mr Balchin's NIMC estimate.

(c) Mr Balchin conceded that the evidence was abysmally weak and ultimately admitted that he was unable to provide an accurate estimate for GLIDe.

(d) Mr Balchin adhered resolutely and improperly to a methodology for which he knows he had no data.

(e) CML submitted that the Court should have pause when considering Mr Balchin's insight into his own approach having regard to the fact that Mr Balchin eschewed a multi-step approach to the calculation of the incremental cost during his presentation, yet his figure with respect to GLIDe was the product of four multiplicative steps.

(a) The context in which Mr Balchin developed and applied his methodology is important.

(b) Mr Balchin's task of estimating the cost of developing the roaming functionality within GLIDe was particularly difficult because CML provided almost no information on which he could rely.

(c) He tried repeatedly to obtain information from CML about GLIDe that might assist that task.

(d) Transurban declined to identify and produce any materials to Mr Balchin that had any information about the cost of adding roaming functionality to GATe or GLIDe.

(e) When Professor Lyon later asked for information about the cost of adding roaming functionality to GLIDe for the purposes of his reply report, Transurban was able to and did locate some relevant information. Specifically, when it served its lay evidence, CML introduced, through Mr Bourke's witness statement dated 31 May 2023, precisely the kind of information about the cost of adding roaming functionality to GATe that Balchin had requested, including board papers that set out estimates of that cost. Specifically, he gave evidence that at the time of the negotiations an amount in the order of between $4 million and $5 million was needed to cover some, but not all of CML's upfront costs to facilitate interoperability with ConnectEast. His recollection was supported by several documents including a draft letter from him to ConnectEast dated 20 July 2007 providing further information on CML's start-up costs to prepare for interoperability and attaching a table revealing the expected cost of 'IT Development' and 'System Capacity Increase' would be close to $2 million.[118]

(f) The cost of adding roaming functionality to GLIDe is a matter that is peculiarly within CML's knowledge. ConnectEast has no ability to assess it, other than through information it obtains from CML. In those circumstances, the evidence that ConnectEast needs to adduce on this issue is slight.

(g) The Court should reject Transurban's assertions of its inability to adduce evidence relevant to identifying the cost of developing roaming functionality within GLIDe. They are inherently implausible given that Transurban developed both GATe and GLIDe over many years and at considerable expense. CML emphasised, throughout the hearing, that there are hundreds of IT staff and engineers working at Transurban. It had ample ability and opportunity to lead evidence from persons with knowledge to contradict Mr Bachin's estimate. Yet, it did not.

(h) Mr Balchin's estimate of the cost of adding roaming functionality to GLIDe was based principally on Transurban's estimate of the cost of adding roaming functionality to GATe. The introduction of interoperability functionality into GATe provided a form of natural experiment; a real world example of a case in which a toll road operator - indeed, CML itself - has had to develop roaming functionality within its billing and operations system. It therefore provided Mr Balchin with a meaningful insight into the costs that CML is likely to have incurred when CML subsequently installed roaming functionality within GLIDe.

(a) He estimated the percentage of account holders that CML would not have had but for roaming. He considered that estimate to be the variable that best explains the incremental cost caused by the Link roaming service. He assumed this percentage was equal to the proportion of CML account holders that drove on EastLink but not CityLink in a given year. Mr Balchin reasoned that if CML did not provide the Link roaming service, those customers would cease to be CML account holders, whereas any customer who travelled on CityLink at least once would continue to hold their account. For the year 2019, he calculated the proportion of CML account holders who travelled only on EastLink to be 2.55% based on information provided by CML. Specifically, he relied on a letter from CML where it was noted that in FY2019, approximately 43,000 of a total of 1,686,000 CML account holders travelled only on EastLink. That proportion equates to 2.55%.

(b) For 2019 and 2020, he then assumed that the customer communications expense lines from the cost centre '110380 Customer Communications Vic' would move in proportion to the number of account holders. He excluded other expense items (like employee benefits or marketing) on the basis that they would either be unlikely to be affected by a small change in volume of communications or were irrelevant to the Link roaming service. He also excluded customer communication costs associated with enforcement activities. Professor Lyon had made this exclusion and Mr Balchin agreed it was appropriate.

(c) For 2015 and 2017, Mr Balchin applied the same method. However, CML's accounts did not separate the communications expenses between enforcement and non-enforcement activities for those years. Accordingly (and again adopting a method from Professor Lyon), Mr Balchin used the proportion observed between enforcement and non-enforcement activities in 2020 to split the earlier years. Professor Lyon and CML both reported that non-enforcement customer communication expenses (i.e., the relevant portion) were 28% of the total in 2020 whereas Mr Balchin calculated its share was 17%. Mr Balchin used 17% in his calculation of the relevant communication costs.

It was then put to Mr Balchin that his assumption may be incorrect, principally because to open a pre-paid CML account, customers need to pay a deposit of $20, and to open a post-paid account, customers pay a fee of $3.50 online or $5 by phone. It was put to him that for customers who rarely travel on CityLink, this may create a disincentive to opening an account and that he had not thought through the likely migration or possible migration of CML customers to ConnectEast accounts in that circumstance.
Mr Balchin accepted that he had not considered this issue.
He accepted that the postulated disincentive may exist and that it may affect his estimate of the NIMC. To put it another way, he accepted that in a world with no roaming, the $20 deposit required to open a pre-paid account with CML or opening a post-paid account and paying a fee of $3.50 (online) or $5 (by phone) may discourage some infrequent users of CityLink from having an account and that it may only be because of roaming that they set-up and/or maintain an account with CML. Therefore, there may be some infrequent users of CityLink who should have been, but were not, accounted for in Mr Balchin's estimate of CML's additional account holders.
On the basis of that concession Mr Balchin was pressed to accept, which he did, that he cannot say the 2.55% is a valid figure as it may be understated.
It was submitted that I cannot accept Mr Balchin's estimate of the additional costs of this activity (or any of the other estimates relying on this figure) as a consequence of that concession.
I reject that submission.
The task being undertaken by Mr Balchin is one of estimation, not a precise calculation. Mr Balchin's concession, made in the heat of cross-examination, was based on a potential disincentive about the costs associated with being an account holder of CML.
Mr Balchin relevantly observed that '[t]here is no arrangement that's costless from [CML's] point of view'.[129] While he did not expand on what he meant by 'costless', I note that Mr Jackson's statement reveals that if an infrequent user of CityLink chose not to become an account holder of CML, on the occasions they did travel on CityLink they would be required to:

No evidence was drawn to my attention by CML which would allow the quantification of customers that may fall into that category. The histogram relied upon does not quantify the number of trips actually taken by those identified as infrequent users of CityLink. It only identifies the percentage of trips taken by CML account holders on each toll road. During cross-examination Mr Balchin was asked about those customers that fell in the bars of the histogram marked '>90% EL Trips' and '>80% EL Trips'. It may be that some of those who travel over 90% of the time on EastLink take many trips on CityLink, but take significantly more trips on EastLink. For example, a customer living in Wheelers Hill who travels to and from work in Doncaster via EastLink three times per week and makes a monthly return trip to visit her in-laws in East Melbourne via CityLink. Over 90% of that hypothetical customer's trips are on EastLink and yet that person still travels on CityLink two times per month, and 24 times per year. Such a customer could hardly be regarded as an infrequent user of CityLink notwithstanding she is counted in the second bar of the histogram, and there is no basis to say she would be disincentivised in the manner postulated by CML.
It is possible that most of the customers within the bar '>90% EL Trips' take at least the number of trips sufficient that the costs they incur in opening a pre-paid or post-paid account outweigh the costs and burden of obtaining single-trip passes, or travelling on EastLink without any arrangement in place. The number of customers for whom the disincentive would cause them to no longer be a CML account holder may or may not be material for the purpose of estimating of NIMC.
The substance of Mr Balchin's concession is that there is a theoretical possibility that, on further analysis, Mr Balchin's 2.55% figure might increase by some amount. It may not increase at all.
Reflecting on the cross-examination, the concessions made and the evidence of Mr Jackson concerning the cost of non-arranged travel and travelling with a pass, I have come to the conclusion that while the possibility of a disincentive might have some bearing on Mr Balchin's estimate, it would be de minimus. The task I am concerned with is one of estimation. Precision is not possible. In the absence of any evidence substantiating the impact of the alleged potential disincentive, I am of the view that the 2.55% remains the best estimate of the additional customers.
Whether Mr Balchin applied the wrong counterfactual
CML submitted that if I was minded to accept Mr Balchin's use of customer accounts as the basis for estimating additional costs, which I am, then I should find that the relevant number is not 2.55%, but 30-40%.
This percentage range was an estimate given by Mr Balchin on the assumption that the correct counterfactual applicable to the determination of additional costs was the one-way roaming counterfactual (where ConnectEast customers can roam on CityLink but CML customers cannot roam on EastLink). When asked during cross-examination to assume that was the correct counterfactual, Mr Balchin agreed that, based on the information depicted in the histogram contained in Mr Jackson's statement, 30-40% of CML account holders would cease to be CML customers and would instead become ConnectEast account holders based on the frequency with which they drove on EastLink.
For reasons set out in sections F.1.5 and F.1.6, I have rejected the one-way roaming counterfactual and instead accepted the opinion expressed by Mr Balchin that the only sensible counterfactual for the purpose of identifying additional costs is the scenario where there is no roaming arrangement between CML and ConnectEast. It follows that I also reject CML's submission that the multiplier Mr Balchin ought to have applied was between 30-40%.

Having assumed that the provision of the Link roaming service may cause CML to have additional account holders, Mr Balchin was concerned with estimating the cost to CML of e-tags for those additional account holders. Logically therefore, he directed his attention to estimating the number of additional account holders and to the relationship between the number of account holders and the costs of e-tags.
CML submits that a trips driver is a better basis for determining additional costs to CML of providing the Link roaming service arising in relation to e-tags. Specifically, CML submits 'the trips driver is not perfect. It doesn't purport to be. It's a valid, reliable, least worse way of estimating'.[142]
Again, I understand CML's reference to the trips driver to be the Trips Driver (31%) employed by Transurban at Step 3 of the Transurban Model and adopted by Professor Lyon at Step 3 of Stage 2 of the Lyon Adjusted Transurban Model. In any event, the effect of CML's submission is that the application of a driver based on trips to CML's costs is better than the approach taken by Mr Balchin of determining the proportion of customers that are additional to CML as a result of roaming (2.55%) and applying that proportion to CML's costs relating to e-tags.
It is clear to me that Mr Balchin's reliance on his estimate of additional account holders (2.55%) is a better way of estimating the cost of e-tags for those additional account holders than the application of a trips driver. To put it simply, Mr Balchin's proxy of 2.55% is directed to identifying only those costs that are additional.
By comparison the application of a trips driver operates to allocate all costs in proportion with volume of roaming trips without any consideration of whether or not the costs allocated to the Link roaming service are additional in the sense of that they would have been avoided if the Link roaming service had not been provided. The application of such a driver falls short of the task required by the legislation. Further and in any event the application of a trips driver assumes all costs in the relevant cost centres vary in direct proportion with trips. The evidence falls short of establishing such a relationship.

(d) Mr Balchin asked Transurban to identify whether it captures data about the number of interactions with account holders related to a roaming trip, and for EastLink in particular. Transurban responded that '[n]either CML nor Transurban capture this data. The business does not organise itself or operate in such a way to enable roaming trip interactions to be tracked'; [146]

(e) Mr Balchin requested the number of tier 2 customer escalations that were dealt with during each subject year, separating these into escalations by account holders and escalations by others. In response, Transurban provided some data for financial years 2017, 2019 and 2020. The 2015 data was not available;

(f) Mr Balchin requested the number of tier 2 customer escalations by account holders that relate to roaming trips, separated into trips on EastLink and other roaming trips. Transurban responded that '[n]either CML nor Transurban capture this data. The business does not organise itself or operate in such a way to enable roaming trip customer escalations to be tracked'; [147]

(g) Mr Balchin requested the number of tier 3 customer escalations that were dealt with during each subject year, separating these between account holders and escalations by others. Transurban provided some data for financial years 2019 and 2020. Data for 2015 and 2017 was not available. Mr Balchin said he would like to discuss what information is kept regarding the nature or cause of tier 3 customer escalations and how it may be possible to identify the proportion of these escalations that may be caused by roaming trips. Transurban responded that:

... asset and toll collector names are not part of the list of available options in Tier 3 contacts reporting. Neither CML nor Transurban captures information to identify the proportion of escalations that may be caused by roaming trips. [148]

(h) Mr Balchin requested the number of tier 3 customer escalations by account holders that relate to roaming trips, separated between trips on EastLink and other roaming trips. Transurban responded that '[n]either CML nor Transurban capture this data. The business does not organise itself or operate in such a way to enable roaming trip customer escalations to be tracked'; [149]

(i) Mr Balchin sought an explanation of how 'retail' contacts (as distinguished from 'asset' contacts) have been derived for the following three drivers employed in the Transurban Model and adopted in the Lyon Adjusted Transurban Model:

(i) 'Driver 2 - Contact Centre - Retail -Vic'.

(ii) 'Driver 2 - Customer care - Retail Vic'.

(iii) 'Driver 2 Resolutions - Retail - Vic'.

(a) He took what he considered to be a conservative approach and assumed that each of the 1,791 trips was subject to a separate contact notwithstanding that those trips related to only 361 account holders and that one contact may have addressed multiple trips.

(b) Again, taking a conservative approach, he assumed that each contact triggered an action by the resolutions team.

(c) On the basis of those assumptions he calculated that the 1,791 contacts amounted to 0.16% of the total contacts made to the contact centre and 29.01% of the resolutions cost centre contacts in 2019.

(d) He assumed those proportions were approximately representative of the situation for all years.

(e) He assumed that a disputed toll would come via a contact centre contact, then be managed by the resolutions team, and require the advice of billing staff.

(f) The cost associated with the billing staff was included in his estimate of costs associated with the staff responsible for daily data transfers and so was not addressed further in relation to this activity.

(g) He assumed that the costs of the relevant cost centres would scale in proportion to the number of issues being addressed.

(h) He then calculated the incremental cost of the activities associated with roaming by multiplying the relevant cost centre costs by the assumed proportions in all years except 2015:

(i) Between 2015 and 2017, there was a substantial change in the functions performed by the resolutions team (costs in 2017 were approximately 30% of the costs in 2015).

(ii) As a consequence, he considered it would be incorrect to apply the 29% proportion to that earlier year (the true proportion would be much lower, reflecting the much larger number of issues being managed by the resolutions team in 2015). However, he could not calculate a substitute proportion because CML could not provide information on the total number of resolutions contacts in 2015 (this information was only available for 2019 and 2020).

(iii) Given the information gaps, he instead assumed that the expenditure in 2015 for the subset of activities undertaken in 2017 and onwards was equal to the 2017 expenditure.

(a) Mr Balchin relied on the 0.16% figure because he had no other information on which to estimate the proportion of total customer contacts that involved roaming-specific enquiries;

(b) the approach of CML to this estimate was to put to Mr Balchin a series of speculative propositions about types of contacts that hypothetical customers might have made. In neither CML's cross-examination, nor its evidence more generally, has it identified any example of any actual contact from a customer that is contrary to the assumptions that Mr Balchin made. Transurban is the entity that receives (all of) the calls in question. There is any number of ways in which it could have identified and adduced evidence of the kinds of calls it receives: illustratively, surveying staff and instructing staff to keep certain records for a period. It chose not to do so, presumably forensically. That choice was made notwithstanding the statutory ceiling on the fee that could be charged. Mr Balchin's evidence should be assessed in that context. That is not to suggest that CML has the burden of proof in the trial. It is to propose that the Court can put to one side speculative propositions that are unsupported by evidence, which it could have but forensically chose not to obtain and then lead;

(c) if the Court considers that the 0.16% figure is too low, it should make a finding as to an alternative figure. An appropriate course would be to make some adjustment to the figure to allow for the uncertainty in it. The alternative figure need have no precision to it at all. The evidence does not permit precision. But, if the Court finds that Mr Balchin's methodology broadly is apt, a finding should then be made. ConnectEast has obtained, and led, all evidence that is reasonably available to it concerning customer contacts. To find that more precise evidence is required would be to countenance CML defeating a claim based on a breach of its statutory obligation to charge no more than the NIMC, on the basis that it has chosen not to keep or produce records that enable the NIMC to be determined with some degree of precision;

(d) CML say phone calls represent 10% of contacts and so it is either stated or inferred that the figure of 1,791 is out by a factor of 10. But there is no evidence on the way complaints occur and are recorded which supports the proposition that this figure is out by a factor of 10. CML could have led evidence as to other methods or routes by which complaints are made and their quantity, but didn't; and

(e) CML criticised the 29.01% figure on the basis that it was calculated by reference to the 1,791 figure. As Mr Balchin said in his first report, that number looks implausibly large given the vast array of issues that could arise across the entire toll road service provided by CML. It is, therefore, conservative. Further, even if one were to use a trips driver to determine the number of '110330 Resolutions VIC' cost centre contacts that were caused by the Link roaming service, the number would be little different. In those circumstances, the Court should be satisfied that the 29.01% figure is a reasonable estimate of the number of '110330 Resolutions VIC' cost centre contact in 2019 that were likely to have been caused by the Link roaming service.

(a) Mr Balchin's estimate was not arbitrary. It involved the formation of an opinion, based upon a reasoned approach, as determined by Mr Balchin as an expert economist, having regard to the limited evidence available to him that bears on the question.

(b) Having identified an activity undertaken by CML which could increase CML's costs as a consequence of providing the Link roaming service Mr Balchin went to great lengths in his pursuit of relevant information and to derive an estimate based on that which was provided.

(c) CML was the party best placed to adduce further evidence about customer contacts but did not adduce any. Instead Mr Balchin was cross-examined about the kinds of calls that customers might make that may not have been included in the 1,791 integer. The cross-examination was not persuasive. CML did not produce evidence of a single contact actually made that should have been included in Mr Balchin's estimate but which was not. Consequently, there is no evidence that would allow Mr Balchin or the Court to determine whether the number of contacts that may have been excluded from his calculation was material or vanishingly small.

(d) The only evidence available to Mr Balchin to estimate the proportion of total customer contacts that involved roaming-specific enquiries was adduced by ConnectEast. I am satisfied that ConnectEast led all evidence that was reasonably available to it concerning customer contacts.

(e) The parties agreed that the task to be undertaken by Mr Balchin was to estimate, not calculate, those additional costs. While it may not be permissible to make an estimation where there is no evidentiary basis, where, as here, there is some kind of evidentiary foundation an estimation can be made, even where precision is impossible.[166]

(f) Mr Balchin took a very a conservative approach, in favour of CML, in relation to the only evidence available to him in that he assumed that:

(i) each of the 1,791 trips the subject of complaint by only 361 CML account holders were also each the subject of a separate contact; and

(ii) each of the 1,791 contacts triggered an action by the resolutions team.

Mr Balchin explained that he considered that the cost centre and the activities recorded within it (as described by Mr McKay and in part by senior counsel for CML during cross-examination) were less likely to be directly related to (or scale with) the number of employees.
CML was the party best placed to adduce evidence to contradict this assessment by Mr Balchin that was made clear in his first report dated 5 July 2022, but it did not do so. In the absence of any contradictory evidence I reject CML's submission.
CML submits that Mr Balchin's calculation of the employee benefit costs (e.g., the figure of $[REDACTED] in 2019) was flawed because it only included the costs accounted for as 'employee benefits'. CML submits there are other costs captured by the 10 Balchin Cost Centres relevant to the employment of employees but did not adduce evidence of those costs. In the absence of such evidence and in any event, I accept Mr Balchin's reliance on Transurban's accounting of costs attributed to 'employee benefits' as appropriate for the purpose of estimating total employee benefit costs.
CML submits that Mr Balchin's calculation of the employee benefit costs is flawed as he has applied multipliers that CML has submitted are invalid. For the reasons I have already given I have rejected CML's submissions that the application of such multipliers is invalid. Although I note that I have adjusted the percentages employed in relation to customer contacts from additional account holders as set out in section G.7.
Finally I reject CML's submission that Mr Balchin took a crude broad brush approach when other alternatives were available to him. The only alternative approach put forward by CML as available was that employed by Transurban and adopted by Professor Lyon. For the reasons I have given I have rejected that approach.
In my view the amounts estimated by Mr Balchin as representing the additional cost to CML in relation to the wider employee costs, in each of the sample years, are properly included in the calculation of the NIMC, subject to the necessary recalculation having regard to adjustments I made to Mr Balchin's estimate of the additional costs associated with contacts by additional account holders.
The figures that require adjustment are the employee benefit costs included in Mr Balchin's NIMC calculation as follows:

The cost centres that survived that process were then subjected to Stage 2 of his methodology where he applied Transurban's three step process, allocating costs at each of Steps 1, 2 and 3 using, in almost all cases, the same drivers as Transurban had used in its model.
I have rejected Professor Lyon's methodology. It follows that the fact that Professor Lyon included costs from the 61 centres in his NIMC estimate does not persuade me that there are in fact costs in those cost centres that should be included.
Instead I have considered each of the 61 cost centres to determine whether or not I am satisfied on the balance of probabilities that there are costs within them, that are additional costs to CML of providing the Link roaming service, that ought to be included in the amount representing the NIMC in any of the four sample years.
For the reasons that follow I am not so satisfied.
CML's position in relation to each of the 61 cost centres was clearly articulated in Appendix C and I have adopted the structure of that appendix in these reasons.
Appendix C splits the 71 cost centres into two sections. Section 1 comprises 41 cost centres which CML described and I will describe in my reasons as the 'Section 1 Cost Centres'. Section 2 comprises the balance of 30 cost centres, described as 'Section 2 Cost Centres'. The key distinction between these sections is that Professor Lyon either did not address, or gave very little attention to, the Section 2 Cost Centres in his reports and model, despite having included costs from them in his NIMC estimate.
CML has set out its submissions and the evidence it relies on in relation to each of the 61 cost centres in a methodical and formulaic way. I make that observation without criticism. It was a sensible and helpful approach to the collation of that material. My reasons adopt a similar form and are set out in the schedule to these reasons. The schedule comprises 45 pages.
I did not want to clutter this document with a further 45 pages of reasoning addressing the 61 cost centres which can, instead, be summed up in a few short paragraphs as follows:

(a) There is insufficient evidence to satisfy me on the balance of probabilities that any additional costs were incurred by CML providing the Link roaming service that are captured in any of the 61 cost centres.

(b) The evidence relied on by CML is directed to supporting CML's submission that activities are accounted for in those cost centres that relate to roaming and that therefore costs within those costs centres are likely to increase with the provision of the Link roaming service. That submission is in line with the first criterion applied by Professor Lyon at stage 1 of his methodology. Had I accepted Professor Lyon's methodology and applied it, such evidence may have been enough to support the inclusion of some costs from those cost centres in the NIMC estimate. But I have rejected it.

(c) My analysis of CML's submissions and evidence revealed that in relation to each and every cost centre there is insufficient evidence (if any at all) to support a finding that:

(i) there are activities within the cost centre related to the provision of the Link roaming service; or

(ii) those activities have in fact been undertaken; or

(iii) if undertaken, have in fact resulted in costs being incurred by CML; or

(iv) any costs actually incurred are additional costs in the sense that but for the provision of the Link roaming service, they would have been avoided.

(d) If I am wrong and CML has incurred additional costs of providing the link roaming service captured in one or more of the 61 cost centres, there is a total absence of evidence and therefore no proper basis for me to estimate those costs and I allow nil for them. In reaching this decision I note:

(i) I have rejected Professor Lyon's evidence and there is no other evidentiary basis (expert or lay) for me to estimate those costs. There is not even enough evidence for me to assess if the additional costs are significant or immaterial;

(ii) the evidence necessary to support the estimation or determination of the additional costs incurred is solely in the possession of CML; and

(iii) Mr Balchin's report was filed on 23 May 2022. By that report it was plain that he had not allowed for any costs accounted for in these 61 cost centres. CML has had ample opportunity to adduce evidence in relation to the additional costs it asserts it incurred and that are captured in these costs centres but it did not do so.

(a) 'reasonable diligence' is not the doing of everything possible to uncover the mistake. The standard of reasonable diligence is but the doing of that which, under ordinary circumstances and with regard to expense and difficulty, could be reasonably required;[179]

(b) the construction to be put on the words 'reasonable diligence' must vary with the particular context in which they are to be applied.[180] In some circumstances, a person will not have failed to exercise reasonable diligence if they were not 'alerted to the utility of uncovering information, especially where to do so would be onerous and costly';[181] and

(c) recently, in Test Claimants in the Franked Investment Group Litigation v IRC ('Test Claimants'),[182] the UK Supreme Court considered the meaning of 'reasonable diligence' in the context of a claim for restitution of money paid under a mistake of law arising from an unlawfully levied tax. The decision of Test Claimants suggests that:

(i) the standard of reasonable diligence is how a person carrying on a business of the relevant kind would act, on the assumption that they desired to know whether or not they had made a mistake, if they had adequate but not unlimited staff and resources and were motivated by a reasonable but not excessive sense of urgency;[183]

(ii) the test of discoverability is that a mistake is discoverable when the claimant knows, or could with reasonable diligence know that they made a mistake with sufficient confidence to justify embarking on the preliminaries to the issue of a writ' or 'in the sense of recognising a worthwhile claim arises';[184] and

(iii) the question is not whether the claimant should have discovered the mistake sooner, but whether they could with reasonable diligence have done so. The burden of proof is on the claimant, who must establish on the balance of probabilities that they could not have discovered the mistake without exceptional measures which they could not reasonably be expected to take.[185]

I have set out a short chronology of events relevant to the parties' submissions in the following paragraphs. I do not purport to capture every event or detail described in the parties' tender bundles and witness statements. I have endeavoured to provide a context so that submissions made (and summarised below) may be understood.
In late 2006, CML and ConnectEast began negotiating interoperability arrangements, including a roaming fee.
By letter dated 30 November 2006, ConnectEast informed CML that it had reached a stage in its customer service rollout program where it would like to settle the terms of a roaming agreement. Accordingly, ConnectEast provided a draft roaming agreement for CML to consider. ConnectEast pointed out that the draft provides for the Discount Amount (or roaming fee) to be calculated and agreed by reference to the NIMC incurred in processing roaming transactions. The elements of that cost were described in clause 7.5 of the draft agreement. ConnectEast stated it understood that the roaming fee was a matter for further commercial discussions in the new year.
The draft roaming agreement provided for the roaming fee to be reviewed six months after the 'Commencement Date' and provided for further reviews 'not before the date 6 months from the date of the last review'.[186]
Negotiations around the roaming fee commenced in earnest in 2007.
Mr Bourke became the main representative of CML in these negotiations and had delegated authority to negotiate an agreement with ConnectEast on behalf of CML. At that time he held the position of CEO of CML.
His initial contact at ConnectEast was with Peter Bentley, ConnectEast's General Manager - Operations and Marketing and then Mr Gardiner, variably described as the Chief Executive Officer and a Managing Director of ConnectEast during these negotiations.
An internal memorandum sent on 22 December 2006 by Peter Bentley to other ConnectEast personnel records, amongst other things, ConnectEast's understanding at that time that the NIMC would be 'only the additional costs incurred in relation to CML [e-tags] roaming on [EastLink]' and would be approximately 5 cents per trip including a variable merchant fee.[187] The memorandum further records that ConnectEast 'argued that for toll roads such as CML and [ConnectEast] all of the customer service and [e-tag] costs are already recovered as part of the toll...'. [188]
By a letter dated 22 May 2007, Mr Bourke provided to Mr Gardiner 'information related to the up-front costs that will be incurred by CityLink in its efforts to be ready for interoperability with EastLink on the current anticipated early schedule.'[189] Mr Bourke further stated:

You say: 'I was made aware by John Gardiner, managing director of ConnectEast, that representatives from ConnectEast and CityLink were engaging in commercial negotiations to determine the terms of a roaming agreement.' Now just to orientate yourself with dates, this is June 07, and if you cast your eye down to paragraph 45, you'll remind yourself that around about six months later, the roaming agreements were signed? ---Yes. > > So you've arrived at an interesting time?---Yes. > > And you say, 'I was made aware by Gardiner that these commercial negotiations were going on.' And you say: 'I was not directly involved in these negotiations, but was on occasions consulted by Mr Gardiner regarding the reasonableness of the costs asserted by CityLink to ensure interoperability with EastLink and the proposed roaming fee.' You must've been horrified, then, at the idea that ConnectEast was going to pay Transurban 18 cents per trip?---Um, perhaps wouldn't use the term 'horrified'. I was um, surprised that it was um, it was going to be that amount, given ah, what I understood the industry wide amount to be at 12.5 cents - - - > > Don't worry too much about the given. The answer you've given is, you weren't horrified. Surprised?---Surprised. > > And if your view was anywhere approximately aligned with what you now know Mr Balchin's view to be as to what the per-trip cost was, your view would've been, 'this is like, 700 per cent more than we should be paying them.' It must've horrified you?---Um, again ah - perhaps not horrified, but ah, yeah. > > Well, surprise?---Yes. Yes. > > And you knew that what ConnectEast was proposing to do was to lock itself into a 10 year contract?---Yes. > > You must've formed the view that this is going to impose a burden on my client - on my employer - for a decade, that will see it paying tens or hundreds of millions of dollars more than is reasonable?---Yes. > > And where are the memos where you set that out? Where are they referred to in your statement? Where are the emails where you said, 'wow. What are we thinking of doing'? ---Ah, well it wasn't so much emails. It was - - - > > No. Reference in your statement to the conversations you had with Mr Gardiner where you said, 'this is lunacy', or 'I'm surprised'?---Um, I haven't made such a statement. > > No you haven't. And the truth of the matter is that you knew that they were proposing to strike a fee of 18 cents per trip, correct?---Yes. > > And you knew it was in the range, correct?---Ah, in the range of? > > In the range of what was reasonable?---Ah, no. Can I clarify the - - - > > No, you can answer the question please. You knew it was in the range of what was reasonable?---Ah, no I - I would not say that I thought the - - - > > Not even in the range - - - > > MR SOLOMON: Your Honour, my learned friend twice asked the witness to answer his question, and then cut off the answer. > > MR GLEESON: Go ahead?---Um, None of us within ConnectEast. > > I didn't ask about anyone else in ConnectEast?---Sure. > > I said not even in the range, in your view. > > MR SOLOMON: No, Your Honour, my learned friend asked the witness whether he considered the fee reasonable. The witness was answering that and was cut off, and he's now affording himself the opportunity to answer that question. > > MR GLEESON: Was it, in your view, not even in the range? ---Yes, it was not within the range. I - I expected - based on standards elsewhere. > > That doesn't reconcile very well with your evidence that you weren't horrified, you were just surprised, does it? ---Um, no I don't - don't accept that. > > You had a pretty senior position in 07, didn't you?---Ah, I - I wouldn't say I was in a um, highly senior position when I arrived at, at ConnectEast. > > You were directly reporting to the CFO?---Ah, not back then, no. > > No?---I was reporting to ah, a person with the title of 'toll monitoring manager'. > > You were directly consulted on the reasonableness of costs by the managing director?---Yes. > > It was your moment, wasn't it?---Ah, no. > > No. That's the answer, thank you - - -?---The - - - > > No. The evidence you give in paragraph 42 just stops at the point of saying that you were consulted. You don't say a word about what you said to Mr Gardiner, do you?---No, I don't say that. > > And we're not favoured with Mr Gardiner. And we're not favoured with your calculations that presumably you made to determine that this was not only a little bit high, but was completely outside the range. Where are those calculations?---Ah, I don't have those calculations. > > And you never made any calculations, did you?---No. No - no calculations were - were made by myself. > > To your knowledge, no calculations made by anyone?---Ah, at the time, there were attempts um, by others within the company to put a number on what was reasonable. But I wasn't involved except through consultation. > > Was it all just a bit of a back of the envelope thing, was it? ---The calculation? > > Yes?---Um, no I - I've subsequently seen um, what was back then detailed analysis by others within our finance team. Ah - and legal team, as to what they thought a reasonable number might be. > > I see. And you don't refer to any of that in your statement? ---No. > > No. Do you know that roaming fees agreed by commercial parties in other Australian jurisdictions for equivalent services are at about 24 cents?---Ah, no - no, that's not my understanding. It's 12 and a half cents, plus .75 of the sum of the tolls. [194]

(a) First, the two integers used to calculate the alleged overcharge were the roaming fee and the statutory cap in s 93AB of the MCL Act. Plainly, the quantum of the roaming fee and the fact of the statutory cap were matters within ConnectEast's knowledge.

(b) Second, the evidence has shown that much attention was devoted to the question of the cost of providing the Link roaming service during negotiation of the Roaming Agreements in 2007. It is implausible that ConnectEast could not have discovered the fact that it had a worthwhile cause of action if it exercised reasonable diligence shortly after the negotiation of the Roaming Agreements, or the commencement of interoperability. It clearly had sufficient information and/or suspicion to warrant embarking on the preliminaries to the issue of a writ, such as taking advice and collecting advice.

(c) Third, and allied to the point above, Mr Robinson, gave evidence that he was 'surprised' by the prospect of a roaming fee of $0.18 per trip during negotiations.[203] He confirmed that he formed the view that a roaming fee of $0.18 per trip would 'impose a burden on [CML] for a decade, that will see it paying tens or hundreds of millions of dollars more than is reasonable'[204] and was not even 'in the range of what was reasonable'.[205] It is also clear that Mr Robinson was directly consulted about the reasonableness of the fee by the managing director, Mr Gardiner. This evidence is entirely inconsistent with the theory that ConnectEast could not, with reasonable diligence, have discovered that it had made a mistake as to the quantum of the NIMC at the time the Roaming Agreements were concluded, or shortly thereafter.

(d) Fourth, as Mr Gardiner has not given evidence on these matters despite his availability, this Court should infer that his evidence would not have assisted ConnectEast on reasonable discoverability.[206]

(e) Fifth, and consistent with the above, Mr Griplas, ConnectEast's CEO, stated that he became aware in July 2020 that the Roaming Fee may have been higher than permitted under statute. But he did not say how he found out or what he knew prior to July 2020, or what steps he took from August 2014 to find out.

(f) Sixth, no evidence was given as to what, if any, steps were taken by ConnectEast after execution of the Roaming Agreements. ConnectEast's failure to take any action within the period of 12 years from the completion of the Roaming Agreements until ConnectEast's purported awareness of the mistake around July 2020 does not satisfy the standard of what is reasonably required by a person seeking to challenge the validity of the Roaming Fee. ConnectEast has not done that which, under ordinary circumstances and with regard to expense and difficulty, could be reasonably required.[207]

(a) the clue to reaching the correct outcome is to focus on the correct proposition: what reasonably was required. The correct proposition here is whether in this case ConnectEast, with reasonable diligence, ought to have known that the NIMC of providing the Link roaming service (for the purposes of s 93AB) was less than the contractual amounts withheld;

(b) that proposition speaks to the internal functions, activities and operation of CML and/or Transurban in providing the Link roaming service. All relevant information (to the extent it was captured) was held in the possession of CML and/or Transurban. There was no occasion (reasonably evaluated), in advance of the 'price review' contractual negotiation (occurring on the tenth anniversary of the Roaming Agreements of which Mr Griplas speaks in his evidence), for ConnectEast to have known or enquired as to CML's NIMC. Prior to the price review, carrying on its business as usual, there were no circumstances that existed in which ConnectEast were 'alerted to' the mistake;

(c) this is, in truth, a paradigm example of a situation where ConnectEast could not have discovered the mistake without 'exceptional measures' - reviewing or otherwise having access to the internal financial information of a competitor, CML. Nor could ConnectEast have known what arrangements were in place between CML and its parent company, Transurban (the absence of such arrangements has only become known during this proceeding). As already identified, and in this context, upon ConnectEast initiating the 'price review' contractual negotiation, it sought from CML details on its underlying cost structure and costs incurred to provide the Link roaming service; and

(d) the singular feature, here, is the asymmetry of information. Only CML possessed the necessary information, for a comparison to occur, year on year, between the contractually withheld amount, on the one hand, and the statutory NIMC, on the other. It cannot fairly be proposed that ConnectEast lacked reasonable diligence. Nor was this proposition put to ConnectEast's senior executive, Mr Griplas, in relation to the unchallenged evidence on the discovery of the mistake.

(a) Data centre hosting incurs costs, including fees for use of the facility, power consumption and hardware costs, for example, servers and chassis, storage disks, memory, power, physical appliances and hardware destruction costs (i.e., when hardware reaches the end of its life).

(b) Cloud storage is purchased on a consumption basis, which is charged on a per hour basis, with costs scaling depending on the amount of storage used. Managing these platforms requires a team of at least 15 engineers within Transurban.

(c) As Transurban's business grows, its cloud and hosting requirements also grow. If there are more trips taken by Linkt customers (including on EastLink), more cloud and physical storage is needed.

(d) Based on my experience in my current role overseeing infrastructure, architecture and technology procurement, generally, the cost of maintaining and managing Transurban's technology infrastructure is directly proportionate to the number of customers it has and the number of trips they take. For example, the storage that is required to record these trips, the computational power needed to process transactions and the network bandwidth needed to transmit information increase with the number of trips. Each of these elements determine costs. This means that an increase in each of these elements increases cost. If Transurban customers did not take any trips on EastLink, Transurban could reduce these costs.

(e) In relation to this function, more customers would be a bigger correlation than more trips but both would influence costs.

(f) Electricity is correlated to the consumption of resources in the data centre. So primarily it would be the cooling function, because you need to keep servers from overheating but also powering those servers which are basically large computers that run the GLIDe system, for example.

(g) If there were fewer customers there are elements of this function that we could cut back on and they would certainly incur a lower bill from Amazon for their consumption of its services in hosting those cloud applications and websites.

Plaintiff v Defendant

At a glance

Court

Decision date

Before

Source

Judgment (1055 paragraphs)

Parties

Legislation Cited (3)

Cases Cited (15)