What are the CMA’s provisional findings in the PR19 redetermination?

On 29 September, the Competition and Markets Authority (CMA) published its Provisional Findings in the redetermination of Ofwat’s PR19 price review for four disputing companies (Anglian Water, Bristol Water, Northumbrian Water and Yorkshire Water), following a reference from Ofwat at the request of each company. This article provides commentary on three key areas: finance issues; cost assessment; outcome delivery incentives (ODIs).

Finance

The CMA has provisionally concluded on an allowed rate of return at the appointee level that is 117bp lower than the PR14 price control, but 54bps higher than Ofwat’s proposed PR19 price controls (see Table 1). This allowed return forms part of a package that the CMA considers will protect the financial and service-quality interests of both current and future consumers. The CMA considers that the provisional determinations also secure resilient services in the face of increased challenges from climate change.

Table 1 CMA Provisional Findings WACC parameter estimates CPIH-real

The CMA’s latest view on the total market return (6.20–7.21%, with a point estimate of 6.95%) and risk-free rate (-1.41% to -0.81%, with a point estimate of -0.96%) is higher than Ofwat’s and also represents an increase on the CMA’s own provisional findings for NERL earlier this year.

The CMA’s approach to estimating the risk-free rate is an important shift away from recent UK regulatory proposals. Taking account of evidence developed by Oxera on behalf of the Energy Networks Association and Heathrow Airport, the CMA concludes that although index-linked gilt yields are a useful input to the estimation of the risk-free rate, they are not a perfect proxy, and hence that it is also relevant to consider yields on AAA-rated non-government bonds.

When estimating the equity beta, the CMA uses a wide range of data frequencies (i.e. daily, weekly and monthly) and estimation periods (i.e. two-year, five-year and ten-year). Its estimated range for the unlevered beta is consistent with Ofwat’s. However, its point estimate for the equity beta is higher than Ofwat’s as a result of i) using a lower debt beta and ii) using a point estimate above the midpoint of its range.

On the cost of debt, the CMA’s provisional approach leads to higher allowances in terms of both the cost of embedded debt and the (indexed) cost of new debt, albeit the latter has been offset by market movements since Ofwat’s Final Determination. There are two major differences in the approaches taken by the CMA and Ofwat.

First, the CMA has accounted for water company debt raised in the early 2000s by using a longer-term trailing average of yields on corporate bonds when estimating the cost of embedded debt. This results in a higher range for the cost of embedded debt, as corporate bond yields were significantly higher in the early 2000s than they have been since the global financial crisis.

Second, the CMA has removed Ofwat’s ‘outperformance wedge’ adjustment on the basis that the evidence does not indicate that water companies have systematically been able to raise debt at lower yields than the market index (once accounting for credit rating and tenor). As well as increasing the cost of embedded debt, this latter decision will also result in a higher allowance from cost of new debt indexation for the four disputing companies, as the ‘true-up’ value will no longer be reduced by the outperformance wedge.

The CMA has heard ‘aiming up/down’ arguments from various parties over the course of the redetermination and has sought to apply a framework for deciding whether to deviate from the midpoint of its range estimates on each of i) the cost of equity, ii) the cost of embedded debt and iii) the cost of new debt. In doing so, it considers the relative levels of uncertainty around parameters, the need to avoid underinvestment in the water sector and the consequent consumer welfare loss, the asymmetry of returns (accounting for the negatively skewed package of regulatory incentives), and other cross-checks (including expected financial ratios). This leads it to choose point estimates that are above the midpoint for the cost of equity but at the bottom end of the estimated range for the cost of embedded debt.

Following the identification of an error in Ofwat’s calculation of the downward retail margin adjustment to the appointee WACC, the CMA has increased the adjustment from 4bps to 8bps of each company’s regulatory capital value (i.e. the wholesale WACC is now 8bp lower than the appointee WACC).

In terms of the CMA’s approach to assessing financeability (as a check to the overall package), the key aspects of the Provisional Findings are as follows.

• The CMA emphasises the importance of the WACC to the question of financeability and refers to credit ratio analysis as playing ‘a supporting role’.
• The CMA focuses on notional company financeability and appears to follow the approach adopted by companies (and Ofwat) in targeting a BBB+/Baa1 rating.
• The CMA recognises Moody’s view that changes to capitalisation rates do not improve credit ratios, and explicitly conducts its assessment of financeability on the assumption that Ofwat’s acceleration of revenues through using a higher ‘pay-as-you-go’ ratio does not improve interest cover ratios.

Cost assessment

The appellants’ total expenditure (TOTEX) allowance was adjusted upwards by the CMA, as set out in Table 2 below.

Table 2 CMA Provisional Findings on TOTEX, £m

In PR19, Ofwat assessed base costs¹ separately from enhancement costs. For base costs, this was primarily undertaken by benchmarking costs across the industry. In its provisional findings, the CMA has generally endorsed Ofwat’s base cost models and base cost assessment approach.

In order to establish efficient base cost levels for the four appellants, the CMA used Ofwat models with three adjustments.

1. The CMA excluded Ofwat’s alternative models, which were used in the Final Determinations by Ofwat as a sensitivity test on its core models. Ofwat only used the results of these models to adjust Anglian Water’s (AWS) allowed costs. The CMA determined that these models should not be used on the basis of modelling and data concerns—removing the additional £50m base cost adjustment for AWS.
2. The CMA also excluded one of Ofwat’s sewage collection models, SWC1, from its suite of core models. The CMA considered that this model was counterintuitive.
3. On forecasting the cost drivers to predict the efficient cost levels over the next price control period, the CMA followed Ofwat’s approach in most cases, but updated the forecasts with new ONS forecasts on connected properties and population density.

The reasons that the CMA offered for not changing the models (primarily with alternative cost drivers proposed by appellants) included: a lack of data, a lack of statistical significance in some of the proposed alternative cost drivers, a lack of operational evidence, and a lack of robust alternatives being presented.

In addition to suggesting changes to Ofwat’s base cost models, the appellants also argued that there was a cost–service disconnect in Ofwat’s PR19 framework, whereby service improvements were being set while the costs related to these improvements were assumed to be allowed for within the base expenditure models. Overall, the CMA agreed with Ofwat’s approach in general but agreed with appellants in allowing for a cost increase to fund service improvements in specific areas.

A key part of Ofwat’s cost assessment framework involved identifying an appropriate benchmark for setting efficient cost levels for companies. With regards to this aspect of Ofwat’s framework, the disputing companies challenged Ofwat’s choice of third or fourth company as the benchmark. Ofwat introduced this choice of benchmark at the Final Determinations, having previously used the upper quartile (UQ). The CMA considered that the quality of the models was the key factor in determining the appropriate benchmark. Having reviewed the evidence, the CMA decided to move the benchmark back to the UQ. This was one of the main reasons for the movement on base expenditure allowances across appellants.

The above concerns base expenditure. Again, the CMA largely endorsed Ofwat’s PR19 approach in general, but with some amendments to individual assessments. In particular, while the CMA left Ofwat’s assessment broadly unchanged where it used econometric models to assess enhancement expenditure, it revisited Ofwat’s ‘deep dives’ (detailed, bottom-up assessments). Where the CMA increased the allowed enhancement expenditure of disputing companies, it was mostly in deep dives, including metaldehyde—an additional £144m across disputing companies—and leakage—an additional £93m for Yorkshire Water (YWS).

As well as identifying efficient cost levels by benchmarking companies’ costs, Ofwat applied an ongoing efficiency assumption to capture technological progress. On this issue, the CMA reduced Ofwat’s ongoing efficiency assumption from 1.1% p.a. to 1.0% p.a., but applied this to all wholesale costs² (in contrast, Ofwat had only applied its ongoing efficiency assumption to base costs, growth costs, WINEP, and some metering enhancement costs). The CMA reasoned that the frontier benchmark was based on the total cost base of comparator sectors.

Outcome delivery incentives

IN PR19, most of the performance commitments (PCs) included in the Final Determinations were accompanied by financial outcome delivery incentives (ODIs). These were designed by companies but amended by Ofwat. Some ODIs were symmetrical, whereas others were penalty-only. Some ODIs included caps on the level of outperformance rewards to protect customers, some included penalty collars to limit company risk, and some included ‘deadbands’.

Ofwat’s approach to PCs and ODIs at PR19 included:

• assessing companies’ bespoke PC and ODI proposals;
• setting three common PCs on the basis of forecast UQ performance, with the remaining 12 common PCs set with reference to the ranges of anticipated performance included in companies’ business plans;
• setting a minimum reduction of 15% for leakage across the sector (itself a common PC);
• limiting any outperformance or underperformance ODIs to 3% gross RORE.

The CMA made adjustments to the Ofwat Final Determinations of bespoke company PCs and ODIs in only a very limited number of cases. Indeed, the main points of contention raised by the companies concerned the common PCs and ODIs adopted by Ofwat, and its interventions at the Final Determinations. The 15 common PCs were:

• the three upper-quartile (UQ) measures—supply interruptions, pollution incidents and internal sewer flooding;
• reducing water demand—leakage and per capita consumption;
• statutory measures—the compliance risk index (CRI) and treatment works compliance;
• asset health measures—mains repairs, unplanned outages, and sewer collapses;
• resilience measures—risks of sewer flooding in a storm, and severe restriction in a drought;
• vulnerability measures—the priority services register;
• customer experience—the customer measure of experience (C-MeX) and the developer services measure of experience (D-MeX).

The appellant companies had argued that:

• there were difficulties in comparing companies on a like-for-like basis in a way that took adequate account of topographical differences;
• Ofwat had ignored the link between the service performance targets that were set and the increased costs of meeting these targets;
• Ofwat’s approach gave insufficient weight to each company’s own engagement with its customers.

However, the CMA stated:

• Ofwat was right to intervene in company business plans to take account of comparisons between companies;
• there was no simple cost–service relationship whereby more demanding PCs should always be accompanied by higher allowed costs (although given the particularly stretching PC target for leakage, the CMA recognised the need for additional funding here—see below);
• while the extensive engagement undertaken by companies had gone a long way to reflect the specific priorities of customers, the CMA considered that there were limits to the weight that can or should be placed on customer research evidence in this area.

The CMA judged that the PC levels for the three common performance measures (the forecast UQ) were appropriate, although the collar for pollution incidents was increased in the case of AWS, and the collar for internal sewer flooding was increased in the case of Yorkshire Water. Moreover, the CMA made adjustments to some of the other common PCs and ODIs:

• the CMA proposed some adjustments to the ODI rates, caps, and collars for the common PCs relating to unplanned outages and mains repairs (with deadbands to limit downside exposure to factors that are outside companies’ control);
• for leakage, the CMA retained the 15% minimum PC commitment required by Ofwat in the Final Determinations, but determined that some of the companies could require an additional allowance.

On this latter point, the CMA has provisionally concluded that there is a link between maintaining higher performance on leakage and costs—one that is not adequately compensated for in the base cost modelling for companies that are currently high performers (i.e. above the UQ). The CMA has therefore adjusted the base cost allowance for the two higher performers—AWS and Bristol Water.

In addition, the CMA is minded to provide additional enhancement funding to AWS, Bristol, and Yorkshire of £68m, £4m, and £93m respectively, to achieve the future required level of performance.³ The CMA intends to undertake further work to refine these funding numbers. At the same time, it has removed enhanced ODIs for leakage (for all four appellant companies) and has increased the companies’ penalty rates for underperformance (for Anglian, Bristol and Yorkshire).

In sum, therefore, the CMA is in agreement with much of Ofwat’s overall approach to PCs and ODIs. However, the CMA has sought to reduce the risk exposure of the companies to certain limited aspects of the PC and ODI package. In addition, the CMA recognises that for at least three of the companies, attaining the ambitious leakage PCs will require further enhancement funding.

¹ Base costs represent operating expenditure and capital maintenance expenditure. Ofwat also modelled growth expenditure in its base cost models.

² This is except in cases where there is strong evidence that 1% has already been included in projections.

³ However, the net adjustment for AWS was £1.2m for base and -£3.4m for enhancement expenditure, as the CMA had also removed other leakage-based allowances for AWS (by dropping Ofwat’s alternative base cost models and Ofwat’s allowance of £71.4m for leakage enhancement expenditure).