The Economic and Financial Impact of Converting Water Companies into Mutuals
The broken nature of our water industry is not just a symbol of economic failure; it is also a contributing factor. This briefing note explores the economic impact of water failure on the government's growth ambitions, and calculates what the financial implications would be for households under a different ownership model.
The government's key growth priorities are all competing for the same scarce water supply. Housebuilding, AI infrastructure and new energy capacity each depend on reliable water, yet they are concentrated in exactly the regions where supply is most stretched. In Cambridge, thousands of homes and significant commercial development are often stalled because water supply cannot be guaranteed, while the Culham AI Growth Zone was sited in an area of “serious water stress”. Left unresolved, these overlapping pressures on an already deteriorating network risk holding back the very sectors the government is counting on for growth.
Reversing this will require substantial investment over the coming decades, which under the current ownership model could push household bills to levels far above where they stand today. Using bond-level data from recent bond issuance across the sector, this briefing note compares the cost of borrowing for shareholder-owned water companies with what we could expect for a mutual company in this sector. It finds that, even with the level of investment required by 2050, a mutual model would deliver material savings through lower interest costs on the debt needed to finance the investment programme, and the removal of dividend payments. The combined effect could meaningfully reduce the burden on household bills.
Key findings
Unreliable water supply has implications on growth: water shortages already constrains housebuilding, and is set to act as drag on future AI adoption and energy production.
20bn in growth is ‘at risk’ from water shortages[1]: the leaders of the six ‘Fast Growth Cities’ name water shortages as a key risk to their sustained growth.
There is enough water for just 420,000 of the government's 1.5m new homes target: as Ofwat estimate that 36% of planned housing in Cambridge has already been lost to water scarcity.[2]
£300bn[3] is needed to reverse the “failed” investment model: the NAO described the industry as having "failed to drive necessary investment".[4] Ofwat have now outlined that a significant investment program is needed as a result.
Water bills are set to rise to £2,000 a year for the average household under the current model to finance this investment[5]: this 213% rise is to cover much needed investment in the network.
A mutual model could save between 2.2bn and 3.7bn in lower interest costs on the debt needed to reverse underinvestment: ratings agencies have historically viewed the mutual model in the water sector favourably compared to their peers due to it being “significantly more flexible than its peers.”[6]
Ending dividends alone would save households £58 a year: water companies have distributed 53.5bn in dividends since privatisation, an average of 1.5bn a year. This is money the mutual model would save.
By 2050, households could save between £182 and £329 a year: this represents a 9% and 17% off their respective bills as a result of lower debt interest costs and the ending of dividends
A new valuation framework may be needed to better capture the strengths of the mutual model: The standard regulatory gauge, net debt to RCV, cannot see the retained-earnings buffer a mutual builds by not paying dividends. Assessing resilience on cash-flow coverage, as rating agencies already do, would align regulation with how the market prices mutual risk.
[2] https://www.ofwat.gov.uk/publication/economic-impact-of-water-supply-infrastructure-investment/
1) The Economic Case
Water functions as a foundational input to economic activity in much the same way as energy, meaning both the reliability and cost of its supply can have substantial implications for growth inputs.
Water supply often struggles to keep pace with demand, especially during summer, which is mostly driven by the lack of reservoirs and an unreliable network. No major new reservoir has been completed in England since privatisation in 1989, and the only active proposal in Abingdon was first promoted in the 1970s, formally submitted in 2006, rejected by the Secretary of State in 2011 for having "no immediate need," and revived in 2018, with an operational date no earlier than 2040. While this is in part a failure of the planning system and previous government decisions, it is also a failure of the water sector to finance the pursuit and construction of new reservoirs on which the sector depends.
The costs of water supply are broadly comparable with european nations,[7] however because of the level of investment required, the water regulator, Ofwat, estimate that around £300bn of investment will be needed over the next 25 years, which could push average household bills towards £2,000 a year by 2050,[8] more than three times their current level. While commercial water bills are not as costly as energy bills, these higher costs can be significant for water intensive industries.
The main economic impact of water supply is through constraints on 1) housebuilding, 2) AI adoption and 3) energy production, which are all critical growth channels for the UK.
Building more homes is an essential part of the UK’s growth puzzle. With constrained supply, people’s ability to move to economic opportunity is limited, and employers in high-productivity regions cannot grow their workforce to meet demand. Ofwat estimate across seven of the most productive local authorities, up to half of planned new homes are at risk each year because of water scarcity, and existing water supply can only cover 400k of the governments of 1.5 million homes target by 2029.[9] Cambridge is one of the UK's most important growth regions with a strong technology sector, leading universities and proximity to London. Despite this, 9,000 homes and 300,000 square metres of commercial development are currently stalled because water supply cannot be guaranteed.[10] Ofwat estimated that 36% of planned housing in Cambridge has been lost to water scarcity, and the Greater Cambridge Local Plan delayed because the water supply issues remain unresolved.[11]
AI adoption and development will be a critical factor in the UK’s economic future. Most economists agree that it can be the sole factor that ends the near two-decade long stagnation in productivity and growth. The government has set an ambition of making the UK the fastest AI adopter in the G7.[12]
It is well known that the data centres needed for AI are water intensive. However, Ofwat are concerned about the ability of the water network to meet this AI induced demand; as an example, water usage from data centres in the United States tripled in the 10 years to 2023. Further, to harness the benefits of AI, the government will want domestic centres to be built near end users, as real-time AI applications require low-latency connections that depend on physical proximity between servers and end users. This will add a compounding problem on water supply, as these end users are often already in high water demand areas. For example, the Culham’s AI Growth Zone was specifically chosen due to close proximity to leading science centres. However, it’s in a location of “serious water stress”.[13] The alternative, Ofwat say, is“in water constrained areas, data centre developers are likely to face higher costs associated with sub-optimal locations & increased latency”.[14] This will mean compromising on the benefits of AI adoption.
Lastly, getting energy supply right is critical for growth, given its comparatively high cost on businesses and households. This will depend on the UK expanding energy infrastructure and adopting frontier technology that can end up producing cheaper and cleaner energy. The government, as a result, wants to quadruple nuclear capacity by 2050 to ensure greater base load clean energy.[15] Small Modular Reactors (SMRs) are part of the plan to achieve this, as these are new small scale nuclear reactors with a standardised design meaning they can be more easily installed across the country. Unlike traditional reactors which are based on the coast so it can use seawater as an input, SMRs will be inland so will rely on reservoirs and likely the mains water network. These SMRs will likely be placed near important energy intensive infrastructure, like AI, compounding on the water supply issue given the above. Similarly, the government has plans for an expansion in hydrogen production for energy supply and as a substitute for natural gas, combined with carbon capture. Both however, are water intensive.
Ofwat say that the total water consumption from energy production will grow between 2- to 10-times existing levels.[16]
Taken together, these pressures mean that the government's growth ambitions across housing, AI and energy are all drawing on the same constrained water resource, often in the same water-stressed regions of southern and eastern England. Without a step change in how water infrastructure is planned, funded and delivered, the cumulative effect of these competing demands risks becoming a binding constraint on the key elements of UK's long-term growth strategy.
[10] https://www.ofwat.gov.uk/publication/economic-impact-of-water-supply-infrastructure-investment/
[11] https://www.cambridge.gov.uk/news/2024/03/05/greater-cambridge-local-plan---timetable-update
[12] https://www.gov.uk/government/news/uk-will-win-ai-race-as-chancellor-sets-out-economic-big-choices
[14] https://www.ofwat.gov.uk/publication/economic-impact-of-water-supply-infrastructure-investment/
[16] https://www.ofwat.gov.uk/publication/economic-impact-of-water-supply-infrastructure-investment/
2) The Financial Impact of Mutualisation
This section explores the financial implication of GGF’s mutualisation model for water ownership. We estimate whether this transition would create sector wide savings, through reduced borrowing costs and a reduction in leakage through the ending of dividends.
The analysis takes the Ofwat estimate that £300bn of investment is needed over the next 25 years. This estimate highlights that water bills would have to rise to £2,000 per year for an average household to in part facilitate this investment. The analysis in this briefing note tests whether this corresponding rise in bills could be lower if water companies were mutuals instead.
To do so, we estimate the cost of raising debt to finance this through the typical cost of borrowing for existing shareholder-owned water companies and compare it to the cost of borrowing for a mutual.
This analysis is a demonstration of what a future water sector made up of mutually owned companies would look like, and illustrates what this could mean for household bills assuming all savings are passed through. It is not a full cost-benefit analysis, its estimates are illustrative.
Dividends
In a shareholder-owned industry in a competitive market, dividends are the return to shareholders earned for putting capital at risk to fund investment if that investment is successful. They are not a payment for the supply of capital, but the return from an investment that produces earnings. We therefore would expect dividends to broadly track post-investment profit in a competitive market.
In the water industry, dividends have not followed this pattern. Heavy investment needs mean free cash flow has been consistently negative, yet companies have continued to pay out a high proportion of their regulated equity as dividends, even in years of weak performance.[17] Distributions on this basis cannot be a share of genuine surplus; they have been sustained by borrowing, which is reflected in the near one for one tracking of dividends and rising sector debt.
This implies that dividends in the water sector are not a return on successful investment, since there has been no genuine surplus to distribute. Funded instead through borrowing, they represent an extraction from the sector rather than a reward for value created within it.
On this record, the sector's dividends are better understood as a transfer funded through the balance sheet than as a cost of doing business.
We can use Welsh Water as the counterfactual, as since mutualising in 2001 it has paid no dividends – given there is no ‘shareholder’ in a mutual - and the retained funds have reduced its gearing (that is debt as a percentage of equity, also known as ‘leverage’) from 93 per cent to 58 per cent, the lowest in the sector, while also reducing consumer bills in real terms between 2000-2020.[18]
Dividend payments have fallen in recent years. Ofwat tightened its licence conditions in 2023 to link distributions to performance, several companies – including Thames Water -suspended dividends under regulatory and public pressure. But this reduction reflects enforcement against the model, not a change in the model itself. The companies remain structured as dividend vehicles: their capital structures were built up precisely to gear returns to shareholders, and the debt taken on to sustain past distributions remains on their balance sheets, serviced by customers.
By ending dividends, we can expect the sector to save 1.5bn per year. This is based on the average total yearly dividend distribution since privatisation.
[17] National Audit Office, April 2025
[18] https://www.sbs.ox.ac.uk/sites/default/files/2021-04/welsh-water-case-study.pdf
The cost of borrowing
Ofwat's submission to the Independent Water Commission[19] estimates that the sector will need to invest up to £300bn over 25 years. Given the scale of the programme relative to the sector's revenues, the great majority of this will be debt financed. It is impossible to say precisely how much, but the alternative of utilising the shareholder model to finance this through equity is not realistic. The sector raised no significant new equity in over three decades after privatisation. Today, the AMP8 framework requires them to raise £7-12.7 billion, but have so far raised only £2.1 billion, with Pennon's £490 million rights issue requiring a 35% discount to attract investors.[20] With returns declining and clear affordability limits on how far bills can rise, it is reasonable to assume almost all of this investment will need to be debt financed.
We illustrate below what the cost of this debt would be if water companies have to raise the full amount in debt.
The cost of debt by water company
With adjustments, bond prices can illustrate the market price for a particular company’s risk profile, meaning we not only estimate the cost of debt for that company, but understand whether a company is deemed riskier than others.
The credit ratings provided by independent ratings agencies can give a condensed insight into the risk profile of a company. The sector's current ratings are set out in Ofwat's Monitoring Financial Resilience Report 2024-25:
[20] https://www.hl.co.uk/shares/share-research/pennon-announces-490mn-rights-issue
Most of the sector sits two to four notches above the investment grade floor. Thames Water sits far below it. A Caa3 rating is the second lowest on Moody's scale, reserved for issuers in or near default, and it puts Thames in breach of its licence condition to maintain an investment grade rating.
To capture a more precise picture of borrowing costs for each water company, we can compare the credit spread each company paid on its recent bond issues: the coupon on each bond minus the yield on the equivalent maturity gilt on the day of issue, taken from the Bank of England's nominal spot curve for the exact day the bond was issued. Netting off the gilt strips out the general economy-wide level of interest rates, which no company controls, and leaves the premium the market charges for that company's risk.
Key term: Gilts and Spreads
In this briefing we calculate the ‘spreads’ of bonds held by water companies. A spread generally refers to the difference between the cost of different bonds which helps investors evaluate relative risk between two different bonds. In this context, spreads compare the cost of a bond for a water company to gilts, which are bonds issued by the UK Government. Gilts represent the floor for debt risk, given governments will not plausibly default on loans. Netting off gilts from a bond isolates the ‘risk premium’, in this case for a water company. The ‘spread’ is the difference between the two.
We restrict the sample to bonds issued from 2023 onwards, a three-year window in which interest rates and market conditions have been broadly stable, avoiding comparisons across too dissimilar rate environments. We look at only those that mature within 20-30 years, to balance against the noise at both extremes: short dated bonds whose pricing reflects imminent repayment rather than credit risk, and very long dated bonds whose yields are inflated by the premium lenders charge for time itself rather than for the borrower. This also allows us to compare bonds that are of a similar length.
This gives 22 bonds, 20 issued by the ten non-mutual companies and two by Welsh Water, the sectors only mutual.
In this information, we can gain some insight into what the market's verdict on each business at the time of issue, by seeing what remains after the gilt is subtracted.
To understand what the likely cost of borrowing for a mutual would be by comparison, we assess the average spread of a non-mutual to the recent bonds issued to Welsh Water, a mutual. In one scenario, we take the average of all three recent bonds issued to Welsh Water, and in the other we take just the one issued when it held an A3 credit rating, it’s best rating which is sustained for over two decades. This latter scenario therefore represents the ‘best case scenario’ for a mutual. We consider below the likelihood of Welsh Water or other new mutual water companies achieving this rating.
In both scenarios, the average spread is lower than it is for non-mutual water companies. The reason is partly explained in the S&P's (a ratings agency) report on Welsh Water, stating: "[the mutual model] eliminates the pressure for shareholder distributions and increased leverage" and that S&P judges its financial policy "as significantly more flexible than its peers, in not being influenced by shareholder considerations, which led to significant deleveraging over the past 20 years".[21]
What the mutual model saves
Averaging the spreads gives the cost of new borrowing under each model. The non-mutual companies issued at an average of 1.82 percentage points over gilts. Welsh Water issued all recent bonds on average at 1.1. And in 2018, when Welsh Water held the A3 rating it sustained for over two decades before sector-wide distress repriced all water debt, it issued at just 0.58 over gilts.
Applied to £300bn of investment defined above, the differences compound into the following:
These are annual interest savings, recurring for the life of the debt, not one-off amounts. The coupon column reports the average rate at issue and reflects the interest rate environment of each issue date; the spread column is the comparable measure.
Spread across the 26 million households of England and Wales, the saving on financing costs is worth:
Total Savings
Combining the estimated savings from ending dividend payouts with the lower borrowing costs available to a mutual, the total annual saving ranges from £111 to £201 per household. By 2050, adjusted for inflation at Ofwat's own 2% CPIH assumption, this rises to between £182 and £329 per household per year. Set against Ofwat's projection that bills could reach £2,000 a year by 2050 under the current ownership model, these savings represent a reduction of between 9 and 16 per cent.
These figures are illustrative rather than a full cost-benefit analysis. They rest on a scenario in which the £300bn of investment goes ahead, dividends are removed, and the cost of new debt falls to mutual rates, with all savings assumed to be passed through to customers rather than retained. Because the scenario holds investment fixed and the two savings channels act on financing and distribution rather than on the investment programme itself, there is little for the savings to flow to other than bills. That pass-through assumption is a simplification, but it serves to demonstrate the scale of what an ownership change could be worth. The estimates are also conservative in one important respect, given they cover only the cost of new borrowing and exclude any gain from refinancing the sector's existing £82.7 billion debt stock at mutual rates as it matures.
Likelihood of achieving the best-case scenario
The current shareholder-owned model, however, offers little practical advantage in this respect. Shareholder-owned water companies have raised almost no meaningful new equity for more than three decades, AMP8 equity commitments have already fallen well short of expectations, and Pennon's £490 million rights issue required a 35 per cent discount to attract investors (CITE). As a result, although mutuals are formally unable to issue equity while shareholder-owned companies retain that option in principle, access to new equity capital is in practice extremely limited under both ownership models. This means that both shareholder-owned and mutual companies face constraints on their ability to strengthen balance sheets through new equity, increasing reliance on debt finance and making that debt more expensive than it would be if companies were able to achieve stronger credit ratings.
For a mutual model, however, there is a potential route to resolving this challenge without requiring a relaxation of credit standards or a change in the assessment criteria simply to accommodate a different ownership structure. The experience of Welsh Water demonstrates that a mutual can achieve and sustain a strong credit rating within the existing framework, having previously maintained an A3 rating for more than two decades and given the above analysis shows they are currently able to achieve lower bond spreads than many shareholder-owned peers even without an A3 rating. The issue is therefore not that mutuals require a more favourable methodology to reach an artificially higher rating, but rather that the existing RCV-based approach creates a structural drag on their ability to demonstrate and sustain the strongest possible credit profile. Addressing this limitation would allow credit assessments to better reflect the underlying financial strength of a mutual model rather than changing the goalposts in its favour.
A more significant constraint on the financing of mutual water companies arises from the methodology used in credit assessment, which has the effect of overstating their leverage and therefore increasing their cost of debt relative to their underlying financial resilience. In regulated water, credit assessments are dominated by the Regulatory Capital Value (RCV), which measures the value of the regulated asset base. The two quantitative financial measures applied to the sector are based on the RCV. Although RCV provides a sensible and consistent measure of the assets underpinning regulated revenues, relying on it almost exclusively means that credit assessments focus primarily on the stock of debt relative to assets, while giving comparatively little weight to a company's ability to generate and retain the cash needed to service that debt.
There is precedent for wider metrics that incorporate more than just asset value to compare debt against. For example, Moody's scores automakers on five separate leverage and coverage measures, including Debt to EBITDA and Retained Cash Flow to Debt, the latter defined as cash flow from operations net of dividends.[22] Regulated water is therefore an outlier in resting its financial assessment so heavily on the asset base alone, and the omission matters most for a mutual, since retained cash flow is where the mutual model is strongest. A company distributing nothing to shareholders retains all of its operating surplus and would score materially better than a shareholder-owned peer with identical operating performance on any dividend-adjusted cash-flow measure; on Net Debt to RCV the two would appear identical.
GGF therefore recommends that Ofwat broadens the financial assessment applied to water companies to include cash-flow measures alongside RCV-based gearing, as is standard practice in comparably asset-heavy sectors. This would give a more accurate reading of financial resilience across the sector as a whole, and for a mutual it would materially improve the prospect of reaching and sustaining a rating at A3 or above.
