The cost of capital is of major importance for infrastructure and utility businesses because they are capital-intensive. The Australian Energy Regulator (AER) estimates that the return on capital is typically about half of the total revenue of the electricity and gas networks it regulates.
Furthermore, small changes in the cost of capital lead to very large changes in regulated revenues. The AER estimates that a 1% change in the cost of capital results in an 8.2% change to regulated revenues. Regulated revenues for electricity networks in Australia were $12 billion in 2022, so a 1% change in the cost of capital would shift revenue by about $1 billion.
Consequently, estimates of the cost of capital are strongly contested. Consumers and regulated businesses try to persuade regulators to employ lower or higher values. This has generated debate across almost all elements of the cost of capital and there is rich and extensive literature on the subject, including litigation in the courts.
In more recent times (at least in Australia and New Zealand), the field of debate has narrowed as approaches have become more settled. There is still active engagement on a range of hot topics including:
Table 1: The three decisions this paper draws on. Three regulators, three recent reviews, and three different starting points on several of the questions that follow.
The ten below, chosen because they are where regulators, networks and consumers are still actively disagreeing. Others would pick differently.
There are many topics that could be included in the top 10, and different analysts are likely to have different choices. The following are the topics that seem most important and relevant in my view.
So far, not by changing the numbers. New Zealand and Australia have both written climate objectives into their frameworks, but neither regulator has adjusted the cost of capital in response, preferring to rely on flexibility elsewhere in the determination.
Many countries now include climate change objectives in their utility regulation. In New Zealand, the Climate Change Response (Zero Carbon) Amendment Act (the Zero Carbon Act) applies. The NZCC explicitly considered this legislation in its 2023 review of the Input Methodologies.
The NZCC considered this aspect from two perspectives:
In conclusion, the NZCC decided that specific adjustments to the cost of capital were not necessary. Instead, it pointed to the flexibility in the regulatory determinations to respond appropriately. It highlighted adaptability in demand and expenditure forecasts (including expenditure to increase resilience) and pass-through mechanisms and re-openers. It did make some adjustments to increase the flexibility available in these mechanisms.
In Australia, the National Electricity, Gas and Retail Laws were amended in September 2023 to introduce an energy emissions objective. The addition is included in section 7(c):
The AER’s most recent decisions were almost finalized before the legislation came into effect, but the amendments were anticipated in the proposals it received and its decisions. The results are most evident in forecasting and resilience expenditure.
Enough that it can be a reason not to change. The AER conducted a full review of its rate of return approach and concluded that continuing it was the better course, on the grounds that minimising change itself supports efficient investment.
It is common to see different regulators taking different approaches to key elements of the cost of capital. At times these differences can be quite material. Nevertheless, once an approach is employed there is considerable momentum to continue the chosen approach. The AER provides a clear example of this momentum. In its process, both consumers and regulated businesses submitted there should be a “high bar for change”. The AER went on to explain the outcome of its considerations as follows:
Regulated utilities are attractive to investors because they are stable, low-risk businesses that offer consistent revenues and returns. This preference for stability extends to the regulatory framework and approach.
They cross-check it, while accepting that the cross-checks are weaker than the estimate they test. The data behind them is no better than the data behind the components, and the indicators are heavily influenced by things other than the cost of capital. Cross-checks nonetheless sit at the centre of decision making, and have been reviewed by the courts.
Typically, when regulators set the cost of capital, they do so by exercising regulatory judgment about a range of individual inputs, which are then brought together to generate a result. This is most prominent in the cost of equity, where it is not possible to observe expectations directly.
Consequently, once the result is derived, regulators tend to want to step back and ask themselves whether the result seems reasonable. However, there are considerable difficulties in assessing overall reasonableness. Firstly, the data available is no better (and more likely worse) than the data used to estimate the individual components. Secondly, and more importantly, indicators of reasonableness are heavily influenced by factors outside of the cost of capital. For example, the profitability of regulated businesses depends on the impact of incentive schemes and the ability to outperform the regulator’s opex and capex determinations. Therefore, any use of reasonableness checks is subject to a high degree of judgment.
Nevertheless, reasonableness checks play a central role in regulatory decision making. In 2012, the NZCC undertook a range of reasonableness checks of its final WACC. These checks were then reviewed by the New Zealand High Court on appeal. The High Court observed that:
The Court then went on to conclude that the Commission’s cost of capital satisfied the reasonableness test, but the appellants would not if they were successful:
Similarly, the AER undertook substantial cross-checking in its 2022 rate of return Instrument process. The cross-checks included: regulated asset base (RAB) multiples, financeability tests, scenario testing, historical profitability, investment trends, other regulators’ rate of return and analysts’ discount rates. The AER concluded:
Among the cross-checks, RAB multiples are the most controversial, with some participants arguing that they have no value while others say they have great value. In respect of RAB multiples, the AER concluded:
The AER observed that cross-checks have also been employed by Ofgem and Ofwat:
For example, Ofgem uses the following cross-checks for its cost of equity estimate:
However, in its most recent decision, Ofgem made no adjustment to the rate of return based on the findings of cross-checks. Further, Ofgem’s latest draft decision (ED2 from June 2022) shows that CAPM is used without adjustment even though cross-checks supported lower values, stating that: ‘…we do not adjust the results […] because we are mindful that no cross-check is perfect, and we are confident that CAPM should remain the primary model.’
The Water Services Regulation Authority in the UK (Ofwat) has recently discussed the use of cross-checks in its draft methodology for PR24, stating that it intends to ‘set the allowed return on equity on the basis of the CAPM [and does] not envisage departing from the CAPM-derived estimate of the allowed return, unless there is strong and compelling evidence from market-based cross-checks.’ Further, Ofwat notes that ‘there should be a high evidential bar for moving away from [a] central estimate [and] expect that any adjustment would be modest and would in any case lie within the endpoints of the CAPM derived cost of equity range.’
Ofwat proposed to use the market-to-asset ratio analysis (that is, RAB multiples) as the main cross-check in determining the allowed cost of equity, noting that ‘[Ofwat does] not envisage departing from the CAPM-derived central estimate of the allowed return, unless there is strong and compelling evidence from market-based cross-checks such as MAR analysis.’
Ofwat’s recently published final methodology for PR24 proposes that there is not ‘sufficiently strong evidence from our MAR-based cross-check to choose a point estimate other than our central CAPM-derived point estimate.’
Regulators fall back on older data or on firms that are less comparable. Australia’s two best comparators, Spark Infrastructure and AusNet, have both been acquired and delisted, and the NZCC’s energy comparator sample fell from 74 companies in 2016 to 51 in its 2023 decision.
In recent years, we have seen a trend where publicly listed utility firms have been acquired by private equity firms and subsequently delisted, resulting in the unavailability of data on their traded equity betas.
Most importantly, in Australia, the two best comparator firms, Spark Infrastructure and AusNet, have been acquired and delisted. Typically, these acquisitions have taken place at values well exceeding the regulatory asset base.
With this diminishing data set, regulators are faced with using increasingly old data or drawing on data from firms that are less comparable, including international firms.
In 2016, the NZCC drew on an energy comparator sample that included 74 companies from New Zealand, Australia, United States, and United Kingdom. In its final decision in 2023, 51 comparator firms were included. The resulting equity betas were like those calculated in 2016.
Unless there is a change in the trend of acquisitions, regulators are likely to face a comparator set that is smaller and less representative. This decline is felt most acutely when estimating equity beta, but also impacts credit ratings and gearing.
The AER has signalled that it intends to undertake further work on the comparator set:
Regulators are split, and both of the most recent reviews decided to stay where they were. The AER considered moving from ten years to five and kept ten; the NZCC came under pressure to adopt the Australian approach and kept matching the regulatory period.
Table 2: The term used to estimate the cost of equity, by regulator
There has been a long-standing question about whether the cost of equity should be estimated based on the length of the regulatory period or a period more consistent with the life of the underlying assets. Regulators have taken different approaches in the past and have changed their approaches. For example, the Queensland Competition Authority and Economic Regulation Authority in Western Australia previously employed a term matching the length of the regulatory period but switched to 10-year terms in 2021 and 2022 respectively.
The AER considered changing from a 10-year term to a 5-year term in 2022, but ultimately decided to stay with a 10-year term. Even though the shift would only have a small impact on the total return (once consequential amendments to other parameters were made), the shift was strongly opposed by regulated networks. The networks indicated their opposition was because the shift would undermine the stability and predictability of the current framework “if the AER makes this change, what other changes might be coming in the future?”.
In its review of its Input Methodologies, the NZCC was under considerable pressure to switch from its practice of matching the term of the regulatory period to match the approaches employed in Australia. Ultimately, it decided to maintain its current approach. The NZCC concluded that:
Badly in principle, and regulators have so far decided not to fix it. A trailing average can sit well below the prevailing cost of debt just as a network needs to raise large new capital, but the AER’s proposed weighted trailing average found almost no support and has been deferred.
There is a tendency among regulators to employ long-term averages in their cost of capital. For example:
In respect of the cost of debt, utility businesses typically raise debt in a portfolio that spans several years. This means a relatively small proportion of the total debt falls due for renewal each year. In this way, the business can mitigate its financing risk. However, if the regulator sets the cost of debt at each determination, then the business faces interest rate risk unless it can hedge a large proportion of its debt portfolio during the determination window.
If the regulator employs a trailing average approach where a proportion of the cost of debt is updated each year, then the business faces much less pressure to hedge its interest rate risk. In addition, the trailing average approach tends to smooth variations in interest rates, so consumers face a smoother price path across regulatory periods.
These approaches work fine when regulating a business with a stable regulatory asset base and steady year-on-year capital additions. However, long-term averages can be far from the prevailing cost of capital at points in time. In the current circumstances where very large investments in transmission links are needed to support new renewable generation, investment incentives can be distorted.
In Australia, the NSW electricity business, Transgrid, is in the process of constructing several new links that have the potential to double its regulatory asset base over a short period. The trailing average cost of debt employed by the AER is currently below the prevailing cost of debt and this may inhibit the new investment.
The AER recognized this potential issue in its draft Rate of Return instrument and proposed introducing a weighted trailing average based on expected additions to the capital base. The AER set out detail on how the weighted average would operate. However, there was very little support for making the change from stakeholders, including Transgrid. The AER noted that it expects the change in regulatory asset bases to be gradual, with the new projects going through regulatory investment tests and early stages of construction and this would minimize pressure on the trailing average. The AER intends to undertake further work in advance of the next rate of return instrument.
The NZCC employs a prevailing cost of debt approach and received submissions that it should switch to a trailing average. The NZCC decided to maintain its current approach noting:
Figure 1: Regulated suppliers’ debt portfolios: tenor against interest rate repricing period, 2022
There is no agreed method, so most fix a value and revisit it periodically. The AER considered blending historical excess returns with a dividend growth model, found little support, and kept a fixed premium; the NZCC settled on a fixed value.
The market risk premium (MRP) is the most challenging element of the cost of capital for regulators. Part of this challenge is that there are very different potential approaches depending on the cost of equity model or models employed. The major classes of models include:
Most regulators employ the CAPM. As such, they need a way to estimate the MRP. A common approach is to set a fixed MRP for a period and revisit it every so often. Under this approach the cost of equity varies one-for-one with the underlying risk-free rate. This contrasts with the Wright approach, where the cost of equity is fixed and does not move with the risk-free rate.
The AER has considered how to estimate the MRP. The approaches it has considered include historical excess returns, DGMs, surveys, conditioning variables and total market returns. In its draft decision it considered using a blended approach with an estimate derived 50% from historical excess returns and 50% from the 3-stage DGM. There was little support from stakeholders for this approach, so the AER continued with a fixed MRP for the life of its rate of return instrument.
The NZCC uses three main approaches to estimating the MRP including:
The NZCC then determined a fixed tax-adjusted market risk premium of 7.0% for its Input Methodologies.
Professor Aswath Damodaran publishes an annual update for estimating the equity risk premium that provides further details.
The Independent Pricing and Regulatory Tribunal in NSW (IPART) uses a blended approach which estimates the cost of equity as the midpoint between its estimates of the current and historic cost of equity. The historic cost of equity employs a historical estimate of excess returns as the MRP. The current cost of equity uses the median value of the MRP derived from six methods:
The first four of these methods are variations of the DGM. They differ in detail, but all infer a forward-looking market average return on equity based on expected dividends.The fifth is another variation of the DGM, which uses stock market analysts' forecasts for individual stocks and a dividend discount model. The sixth method uses four economic indicators to derive an indirect estimate of the MRP.
By returning capital earlier, and the courts have upheld it. The NZCC shortened asset lives for gas pipelines and the High Court of New Zealand dismissed the appeal against it, concluding that stranding risk had to be addressed somewhere in the framework.
Controversially, in its 2022 Input Methodologies decision, the NZCC allowed for the shortening of asset lives for gas pipelines, in effect, bringing forward depreciation to reduce asset stranding risk. This decision was appealed by the Major Gas Users’ Group to the High Court of New Zealand. The appeal was dismissed in favour of the NZCC’s approach. The Court concluded that stranding risk needed to be addressed somewhere within the regulatory framework. Typically, for a firm in a workably competitive market, that risk would be compensated up-front through expected returns. However, in the case of regulated gas networks, that approach is not feasible, and some ex-post accommodation is required. The NZCC’s approach of an early return of capital was viewed as reasonable.
The Court considered a range of interesting questions including: Is stranding risk already addressed through the weighted average cost of capital component – or otherwise provided for in the input methodologies? Is there provision for asymmetric risk in the WACC? Is there double counting? And is the 2022 Input Methodologies Decision premature?
Like the NZCC, the AER has introduced a degree of accelerated depreciation for gas pipelines. In the case of Evoenergy in the Australian Capital Territory, the AER concluded that accelerated depreciation via shorter asset lives was warranted:
The AER noted that it would not be prudent for the network owner to add substantial new discretionary investment capital in the current environment.
But what if gas network closure advances far more rapidly than anticipated? It depends on whether customers are locked in or can exit the network. If customers are locked in, they will pay higher prices to recover the outstanding capital over a shorter period. On the other hand, if demand and customer numbers decline, which is more likely, the remaining customers will face rapidly increasing prices. At this point, it may no longer be possible for the network to charge the regulatorily determined prices because it will accelerate the “death spiral”. Options at this stage include deregulating the network, leaving the risk to be managed by the equity holders, or government intervention to support customers or the asset owners.
Where the asset is discrete enough to tender, yes, and the auction reveals a cost of capital as a by-product. The UK has auctioned offshore transmission rights and New South Wales is tendering its first renewable energy zone network.
With new large-scale links being built to support renewable energy generation, there is an option for these links to be built and operated through a competitive tender process. Provided the competitive tender is robust, this is an excellent opportunity to avoid the need for regulation altogether. Further, the information revealed in these processes could be used by regulators to cross-check the cost of capital they are employing in their regulatory decisions.
There are 2 recent prominent examples: UK offshore wind transmission and NSW renewable energy zones (REZ).
In the UK, the right to build, own and operate offshore wind transmission lines has been auctioned. From the auction results, it is possible to estimate an implied cost of capital. Ofgem has published the results of its analysis.
Figure 2: Offshore transmission owner auctions: average nominal post-tax equity IRR. The implied return falls 3.7 percentage points across six rounds, from 10.2% to 6.5%, with the sharpest step between 2015-16 and 2017-19. Twenty-six projects in total.
In NSW, EnergyCo is leading the development of the renewable energy zone (REZ) transmission network infrastructure as the Infrastructure Planner. EnergyCo is responsible for developing and overseeing the planning and approval processes for the REZ network infrastructure. Its first major project is the Central-West Orana REZ which will be serviced by new transmission network infrastructure, including transmission lines and energy hubs. EnergyCo is currently running a competitive tender process to appoint a Network Operator to design, build, finance, operate and maintain the REZ transmission network.
It became routine, cost consumers $6.5 billion, and never once reduced revenue. Reviews were sought on 32 of 51 decisions between 2008 and 2016, amendments in 2013 failed to change that, and the regime was abolished.
While merits review is not strictly a cost of capital issue, its influence on cost of capital is of the highest importance.
Regulatory determinations embody the exercise of judgment. Otherwise, regulatory outcomes could be coded in legislation and there would be no need for regulators. When exercising judgment, it is possible for regulators to err. Merits review provides a mechanism for regulatory judgements to be tested and corrected if made in error. On face value, this logic is compelling. However, in practice, merits review has the potential to result in highly undesirable outcomes. This was the case in Australia, and ultimately led to the removal of merits review.
In Australia, the use of merits review of regulator decisions was commonplace. Between 2008 and 2016 reviews were sought on 32 out of 51 regulatory decisions (62%). Even after amendments were made in 2013 to rebalance and minimise reviews, 12 out of the 19 decisions from 2015 onwards were taken to review (63%). Of all the matters that were varied or remitted back to the primary decision maker, none resulted in a decrease in revenues for the regulated business compared to the original decision.
Figure 3: AER regulatory decisions appealed under limited merits review, 2008 to 2016. Appeals were routine throughout: 32 of 51 decisions. The 2014 reforms, designed to curb them, made no difference, 12 of the following 19 decisions were still taken to review.
In the second reading speech on the Bill to abolish limited merits review the Minister noted a range of problems with the framework:
Since the removal of limited merits review, the regulatory framework in Australia has fundamentally changed. Instead of regulation being “something that is done to consumers”, regulated networks are undertaking extensive and meaningful engagement with consumers, giving them a real say in the proposals they submit to the regulator.
The Powerlink Queensland electricity transmission decision in 2022 was a watershed moment in the cycle of regulatory determinations. Powerlink approached its determination with the objective of putting forward a proposal that was capable of acceptance by consumers, the regulator, and itself.
In its final decision, the AER noted:
This trend has continued. It is brilliant to see consumers at the centre of the regulatory process rather than lawyers. In the most recent round of regulatory proposals, consumer engagement and influence have been a dominant feature. In almost all cases, the AER has been able to accept the proposals put to it, rather than intervening.
Table 3: Outcomes of the AER’s most recent regulatory determinations
While not making our top 10, the following issues are worth a mention.
It is possible for regulators to very accurately measure and then estimate the forward cost of debt. The AER uses an approach where it obtains details of each debt instrument issued by the private businesses it regulates. This data is sensitive, so the process is conducted in confidence. The AER then publishes a summary of its analysis which shows how the actual cost of debt is tracking against the yield curves it employs for estimating the cost of debt.
This analysis shows that there is a very close correlation between the actual cost of debt and the AER’s estimates.
Figure 4: Energy infrastructure credit spread index against AER cost of debt estimates, June 2015 to June 2023
This issue arose in a Competition and Markets Authority decision in 2021. Since then, it has been raised in front of multiple regulators, but to date accepted by none. The argument, as put to the NZCC by Oxera, is “in contrast to the highest-quality non-government bonds, government bonds have special properties that create additional demand for these instruments, which pushes the government bond yield to fall below a ‘true’ risk-free rate based on a zero beta asset”.
The proposed solution is to add a convenience yield to the risk-free rate or include high quality non-government bonds in the estimate of the risk-free rate.
The NZCC did not agree to depart from using government bonds to estimate the risk-free rate. It noted:
Similarly, the AER considered a convenience yield in depth and decided not to employ the approach.
The “W” in WACC stands for weighted. Regulators typically estimate the cost of debt and equity separately and then bring them together using a ratio. A common ratio is 60% debt and 40% equity. This ratio has tended to be relatively uncontroversial.
The AER updates its estimate at each rate of return review. The data shows that businesses employ a ratio that is largely stable. The AER received a report from Professor Graham Partington and Stephen Satchell that concluded:
It is not uncommon to find submissions that advocated for the use of financeability testing. The NZCC concluded:
The mix of opex and capex used in a regulatory determination is not, of itself, a cost of capital issue. However, the mix has a material impact on cash flow and measures of financeability.
Ofgem employs a totex approach where a total expenditure of opex and capex is determined and the regulated business then has a degree of flexibility in the balance of opex and capex going forward. Under the totex approach, there is a risk that regulated businesses will favour opex over capex to increase cash flows in the short term.
The NZCC received some submissions that carbon reduction required further incentives for capex, such as the totex framework employed by Ofgem. It concluded:
A split cost of capital would involve applying different WACC values to existing “sunk” assets and capital on new assets.
The NZCC considered that its approach provides suppliers with incentives to invest without providing for a different WACC. Non-systematic risks can be more appropriately dealt with through measures outside of the WACC.
The NZCC employs a peculiar approach where it estimates a standard error for the cost of capital and then adjusts the final value of the cost of capital based on its judgement on where the probability of error should sit.
When exercising this judgement, the NZCC considers potential asymmetry in the consequences of setting an incorrect cost of capital.
The NZCC ultimately decided to set the cost of capital for electricity business at the 65th percentile (down from the 67th percentile). This approach seems to have had its genesis in a major electricity cable failure into New Zealand’s largest city (Auckland) in 1998. The failure resulted in electricity outages and major disruption for 5 weeks.
Previously, the NZCC did not index the regulatory asset base for Transpower. When the regulatory asset base is indexed, investors receive part of their total return through capital accumulation rather than cash.
The NZCC decided that it would change its approach for Transpower and index the regulatory asset base going forward. It considered “the benefits of indexation (protection from inflation and promoting pricing profiles that are more likely to be consistent with allocative efficiency) justify the change”.
Estimates of actual and forecast inflation are employed in many places within the regulatory models and cost of capital estimates. The issues are extensive and complex. The AER undertook an extensive review of inflation, publishing its final position paper on the regulatory treatment of inflation in December 2020.
I was delighted that MCC's work was completed on time, and within budget, helping us deliver important changes and improvements, to the benefit of our stakeholders. MCC's report is published on the CCC website.
I am delighted to recommend MCC Economics. Specifically, I worked closely with PJ, who helped us with our Nuclear and CCUS projects. PJ helped us develop new policies and answer questions from our stakeholders. His support helped us deliver important changes and improvements, to the benefit of our stakeholders.
MCC Economics has helped us better understand the most important issues for our stakeholders, including: charges, shareholder returns, debt payments and inflation impacts.
I am delighted to recommend PJ and his team at MCC Economics. We've been working together on National Policy Statements to help meet net zero targets for 2030 and 2050. We initially appointed MCC Economics to support us on offshore wind consultation analysis and have recently reappointed MCC Economics to undertake a larger consultation analysis role across all sectors, including hydrogen, CCUS and networks. I can confirm that PJ and his team have shown excellent spreadsheet skills, alongside very good project management, planning and analysis skills, helping us deliver important changes, and continuous improvements, to the benefit of our stakeholders.
I am delighted to recommend PJ and his team from MCC Economics. They helped us with our price controls for Heathrow airport and for NATS (En Route) plc (the air traffic services provider). Specifically, the MCC team helped us deliver important changes and improvements to our financial models and supporting policy documents, to the benefit of our stakeholders.
I am delighted to confirm that I worked with PJ on a retail project in 2015. The project helped stakeholders understand electricity costs and charges. Specifically, the project helped us explain to stakeholders, internally and externally, why electricity charges differed across the regions (GB, NI & Ireland). PJ was a key member on the project team, which helped deliver changes and improvements in the understanding of energy retail.
.png)
Evidence-based analysis for transparent, defensible, and effective decision-making.
© All rights reserved – MCC Economics 2026.
.jpg)
.jpg)
