Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
When digging up roads and fixing sewers, regulated utilities are not known for taking coordinated action. Show them a price cap, however, and the synergy is balletic.
Thames Water yesterday delivered its response to a draft determination for the 2025 to 2030 regulatory regime. To no great surprise, Thames argues that Ofwat’s proposals are untenable because they would make water infrastructure uninvestable.
What may be more interesting is the way Thames’s mates had its back.
Wessex Water, Yorkshire Water, Anglian Water, Affinity Water and United Utilities have all fired out statements saying they’ve reached a nearly identical conclusion, in nearly identical language. Here’s Wessex:
Having carefully considered the overall risk and return package, we believe that the draft determination puts our ability to retain and attract investment at risk and in our response have explained and provided evidence for this view.
Yorkshire:
The Board considers that significant changes to the draft determination are required before the sector can be considered investible.
UUs:
The draft determination risk and return balance has been set punitively and we believe Ofwat has not secured that companies face a “fair bet” opposite allowed returns and the incentive package.
And Anglian:
[T]o achieve the scale of work needed, our Final Determination, and indeed the sector as a whole, needs to be investable. [ . . . ] We strongly urge Ofwat to re-think this point, and support the industry in attracting long term, quality investment to drive economic growth.
Only Macquarie-owned Southern Water is a bit more circumspect, saying . . . .
In its current form, Ofwat’s Draft Determination will not support the sheer size and complexity of investment needed to run the business sustainably, to meet either our legal obligations or our customers’ ambitions.
Other operators echo nearly word-for-word the line taken by Water UK, the trade association. It says price caps “would likely make it impossible for the water sector to attract the level of investment that it needs”.
Though stage management is an apparently essential part of lobbying, it’s rare to see a retaliatory front so united around one simple message.
The shared working? Not so simple.
Hat-tip to MainFT’s Robert Smith for highlighting the cost-of-equity presentation KPMG has prepared on behalf of the UK water industry’s main players. It’s . . . dense.
The report’s authors, professors Alan Gregory and Alex Edmans, present in granular detail all the ways in which investors might take lower returns than those Ofcom’s proposal assumes. They highlight, for example, an assumed risk-free rate that’s probably unobtainable (because of the gap between borrowing and saving rates) and argue that recent volatility is not typical (because water stocks were flight-to-safety defensives during the recent inflation spike). The arguments all sound reasonably plausible once put in plain English.
What it’s not is plain English. Here’s a typical section, from page 34, on finding the right risk-free savings rate:
Diamond and Van Tassel (2024) estimates CY(NG) using the put-call parity relationship on European FTSE100 options. It finds 2Y CY(NG) of 29bps.
In the PR24 FM, Ofwat inferred 2Y CY(ILG) from the 2Y CY(NG) in Diamond and Van Tassel by applying the following formula from Liu et al. (2015):
CY(NG) – CY(ILG) = Gilt BEI – Swap BEI (breakeven inflation)
The September 2023 CoE report indicated that this formula assumes the entire gap between gilt BEI and swap BEI is due to higher CY for 2Y NGs relative to 2Y ILGs. However, it should reflect that the gap could also be due to the illiquidity of inflation swaps. The modified formula becomes:
CY(NG) – CY(ILG) = Gilt BEI – Swap BEI + inflation swap illiquidity premium
Ofwat highlighted the inflation swap illiquidity issue in the DD and did not disagree with it. Moreover, Ofwat recognises in the DD that inflation swap rates incorporate an illiquidity premium.
In the PR24 FM, Ofwat used an estimation window for CY of 18/06/2007 to 27/07/2020 which broadly aligns with that in Diamond and Van Tassel. 2Y CY(NG) less 2Y CY(ILG) is 27bps over Ofwat’s estimation window based on the modified formula above. This implies a 2Y CY(ILG) of 2bps.
The September 2023 CoE report considered that 2Y CY(ILG) is likely to lie between the estimate from the modified Ofwat analysis and the 2Y CY(NG) estimate from Diamond and Van Tassel. This approach reflects a key finding from the report that the majority of CY factors cited in academic literature appear to apply similarly to NGs/ILGs but NGs may be more liquid than ILGs.
The result of the approach is a range for 2Y CY(ILG) of 2-29bps. The bounds of 2bps and 29bps are both likely to be higher based on a more recent data cut-off as explained in section 4.4.3.
As such, it does not appear appropriate to place excessive weight on the lower bound. The midpoint of the range of 15.5bps is selected as the point estimate for 2Y CY(ILG).
It is reasonable to assume this 15.5bps holds for longer-dated ILGs based on section 4.4.2. This means that CY(ILG) of 15.5bps needs to be added to the 20Y ILG yield to derive rs.
And from page 104, here’s a bit about how one size of beta doesn’t fit all:
CCZ use a pooled Ordinary Least Square (OLS) regression, which assumes that the average elasticity is the same across firms. If the assumption of uniform average elasticity across firms does not hold, alternative models, such as the fixed effect model, should be used. The fixed effect model incorporates firm-specific, time-invariant effects, relaxing the assumption of uniform elasticity and accounting for individual heterogeneity across firms that affects elasticity.
The pooled OLS regression can be expressed as follows, where the intercept term α is fixed across firms.
The fixed effect model can be expressed as follows, where the term 𝑢𝑖 represents the firm-specific, time-invariant effects.
The fixed effect model can also be expressed as:
This alternative expression may be more intuitive, as the firm-specific, time-invariant effect is represented by a firm-specific intercept 𝛼𝑖 , rather than a constant intercept term (α) as in the pooled OLS regression.
Great stuff!
It can be difficult to see how 116 pages of decomposed equity betas and Fama-French factors will lead in any direct way to fewer jobbies in rivers. Permitted returns under previous pricing regimes were hardly onerous. The required infrastructure investment hasn’t happened. The jobbies keep coming.
But of course, we should be blaming the game as well as the players. It’s silly to think regulators can determine ideal long-term levels of risk and return, since investors have limitless ways to cut the calculation. Anyone with a knack for regression analysis will more likely be working for Brookfield or KKR. (Ofwat salaries are Civil Service banded.)
And ultimately, no amount of water industry lobbying around permitted returns will fix a reputational problem born of past performance. One of the easier-to-understand charts in KPMG’s report is this one:
Further reading
— It’s time to pull the plug on Thames Water (FTAV)
— Fix your insolvent UK water company with this one simple trick (FTAV)
