One weird trick to get data centres built
One weird trick to get data centres built
How a small tweak to business rates could be the carrot to Labour's planning stick
Labour has pledged to make decisions on important infrastructure projects, like data centres and prisons, at the ministerial level, instead of letting the local council decide. The hope is that this will allow them to approve these projects even when there is a lot of local opposition.
Overriding local objections will help. But we already use national decision making for a whole range of projects, and these projects are still regularly blocked by the intense opposition of local interests. The £1.35 billion Aquind Interconnector in Portsmouth would add 2,000 megawatts, or 16 terawatt hours a year, to Britain’s electricity supply – about as much as a medium-sized nuclear power station, and enough to power twenty large data centres. The final decision on it was made by the Business Secretary, instead of Portsmouth council, but he still ruled against it because local objections (including from Penny Mordaunt, the local MP) were so strong. (That decision was overturned by a court, but a revised decision still hasn’t been made. Presumably the Conservatives decided to leave it for Labour to handle.)
Even if locals won’t ever welcome something like a new prison near them, we might be able to soften the blow enough to make it politically possible to get national-level approval.
We propose a straightforward way to get more buy-in from locals, on top of what Labour is planning: for the kinds of infrastructure projects deemed to be nationally important, let the local council retain 100% of the project’s business rates bill, permanently, instead of up to 50% that they would retain temporarily under the status quo.
This can be done quickly, doesn’t need new legislation, wouldn’t involve any new spending or borrowing, and could be targeted at the things most urgently in need of building – things like data centres, prisons, nuclear power stations, and other green energy projects. And it would only apply to new projects, so no existing tax revenues would be affected.
Build absolutely nothing anywhere
Everyone knows the UK has a huge problem with building infrastructure of all kinds.
Last November, a £2.5 billion data centre project next to the M25 (the motorway that circles London) was blocked on the grounds that it would spoil views from pedestrian bridges over the motorway.
Earlier this month, a campaign was started to oppose what would be Europe’s largest data centre project in Havering, though the council is (so far) supportive on the grounds of the tax revenues and jobs it would bring.
In March, the BBC reported that prison overcrowding has gotten so bad that convicts could be released months earlier just to free up space for new prisoners. They will be released earlier from prison sentences that are often already notably short, even for the most violent criminals, apparently because judges are trying to avoid giving long sentences that exacerbate the overcrowding problem.
Britain has not built a new reservoir in thirty years; water shortages are being cited as a reason not to build new homes in Cambridge. Oxford MP Layla Moran has led an tireless campaign against building a new reservoir in Abingdon, just outside Oxford itself, telling the water company to fix leaks to save water before building new reservoirs.
Britain’s last nuclear power plant was built between 1987 and 1995. Two are in progress, one of which is the most expensive nuclear plant in the world, and they are only expected to come online in 2031 and around 2036.
The Boris Johnson government repeatedly promised to dual the A1 North of Morpeth, the main road between Scotland and England, repeating a promise David Cameron had made, but not delivered on, in 2014. Consent was finally given in May this year.
The Lower Thames Crossing – a proposed new tunnel under the Thames, about 10 miles east of the existing Dartford Crossing, where the M25 crosses the river – has cost £267 million just to plan, including producing six consultations and 359,000 pages of documents, largely to defray local objectors and reduce the risk of being judicially reviewed by locals. Norway’s Laerdal tunnel is six times longer and cost only £100 million to plan and build.
Traffic on the A1 North of Morpeth – the only single carriageway section of the road.
Much of the time the problem comes down to these projects having benefits that are widespread across the country, and costs that are concentrated in the local area that the project is trying to get built in.
How business rates encourage development
How did projects like these ever get built? In part, because local governments were rewarded for it. Business rates (officially ‘non-domestic rates’) are a tax paid by businesses that have existed in some form since 1601. Today, rates charge companies for the property they occupy, at 51.2 % of their rateable value, an estimate of how much they are worth annually in rent. Heathrow Airport today has a rateable value of £210 million and pays over £100 million in business rates each year, whereas a typical shop on a London high street might pay £1,500 to £3,000 annually.
This system has an obvious downside: businesses pay higher bills if they invest in their property, so there is a tax disincentive to that kind of investment. As both of us have pointed out, this can be an important disincentive to manufacturers investing in new equipment: Tata Steel’s annual rates bill at Port Talbot rose by £400,000 after it rebuilt the steelworks’s blast furnace.
But one important upside of business rates is that local communities get some benefit when valuable investments are made in their area. For example, Brixton Prison has a rateable value of £545,000, meaning it pays £279,040 every year, of which about half is retained by the local authority (Lambeth). Local residents might prefer it wasn’t there, but they do see some benefit from it being near them, since the council can then spend the money on local services like street cleaning, bin collections, upkeep of parks, and social services. Or they can use it to cut council tax.
Between the 1600s and the Second World War, local councils retained 100% of the non-domestic rates paid in their area. They also set the percentage rate it was paid at, whereas today the rates and reliefs are uniform. This meant that the construction of a new prison, shopping centre, or electricity substation, for example, would mean a new revenue stream for the local authority. So while locals would still face the frustrations of living near these projects, they’d also enjoy some of the benefits from them as well. What could be a win/lose was, in some ways, a win/win.
This started to change in 1948, when rates began to be ‘equalised’ between areas, meaning that central government paid for most of local government spending through block grants, giving local governments that raised less money through local rates more money in block grants. This meant that there was less and less point in raising money through rates, because less money raised would be offset by higher block grants. By the time the system was scrapped in 1988, £1 extra in rates tended to mean almost exactly £1 less in central funding.
The current system was brought in in 1990. Now, councils could not vary the tax rate, and the money raised from business rates went entirely into a central pot, to be redistributed back to councils according to need. Tower Hamlets council has a good explainer of how it works here.
Since 2012 local councils have in theory retained 50% of what they collect, but what could have been a positive step was fatally undermined by making the retention only temporary. The plan was for all uplift gains to be reset in 2020 and returned to central government, meaning that the benefit to councils was tiny. In practice, the Conservatives fumbled it even worse than this: they delayed the reset until 2025/6, meaning that local governments got to keep a few more years of uplift, without having been incentivised by that to accept development in the first place.
This didn’t happen to every council. The City of London got a special deal after 1990, and was allowed to set the details of, and keep a substantial portion of, its own business rates. Outside of development corporations like Canary Wharf (and partly spurred by the competition created after 1988) it has clearly been the most permissive area in London for commercial and office development.
Other similar mechanisms have also ‘greased the wheel’ of local consent. The Shetland Charitable Trust was set up in 1976 to collect ‘compensation’ paid by oil companies in exchange for the construction of the Sullom Voe Oil Terminal, which receives, stores and exports North Sea oil and gas. Since that time it has given out £320 million to local charities and the 22,990 residents of the islands.
Getting data centres built
Data centres are clusters of tens of thousands of extremely powerful computers that store and process most of the activity that takes place on the internet. They are now becoming even more important as artificial intelligence grows in power, driven primarily by the amount of computing power available to it (rather than by more sophisticated models). Data centres are also extremely energy-hungry: Brian Potter writes that:
large data centers can require 100 megawatts (100 million watts) of power or more. That’s roughly the power required by 75,000 homes, or needed to melt 150 tons of steel in an electric arc furnace. Power demand is so central, in fact, that data centers are typically measured by how much power they consume rather than by square feet.
The UK cannot grow its domestic AI sector – one of the few reasons for optimism about Britain’s economy – if we do not make it easy to build data centres and the energy supply they rely on. But the same objections that have hamstrung the country’s ability to build housing and other infrastructure are now preventing us from building these as well.
The UK’s biggest existing data centre campus is Virtus, at Stockley Park in Uxbridge. There are four data centres there, with rateable values of £4.4 million, £2.5 million, £4.2 million, and £3.4 million. Together, they will pay about £7.4 million in business rates annually. The 50% share of that going to local authority Hillingdon, or about £3.4 million, is equivalent to the council tax they get from 2,600 band D households with far less of the spillover and infrastructure costs that new housing involves (e.g. little need for new schools and hospitals).
Our proposal is to allow new projects of certain kinds – data centres, prisons, nuclear power stations, and probably some other types of infrastructure – to have their entire rates bill go to the local authority permanently, not just 50% of it temporarily.
Rates from existing data centres (and prisons, and other infrastructure included in this policy) would continue to be split between the local council and the centre as usual. Nothing would change for anything that exists right now. But a new data centre or prison would pay its rates entirely to the local council, for as long as it was around.
This would mean that a local authority that gets a new project of the scale of Stockley Park would collect the entire £7.4 million/year rates bill that it pays, not the £3.7 million it would under the current regime until the reset. In other words, the fiscal reward for accepting it would be twice as large as it is now.
To put this in context, this would completely cover Havering council’s adult social care overspend if it was built there, and the entire overspend of the average District council.
Under this approach, local councils would have a much higher incentive to accept projects of these kinds. The tax take from them would be at least twice what it is under the current regime, so the compensation to the council – and, hence, the area’s local residents – would be double what it currently is. While this would not eliminate NIMBY objections to these projects, it would weaken them.
This would create a carrot to go along with the stick Labour is proposing. And, as we suggest, we could extend it to other nationally important infrastructure like prisons and new energy supply. Along with the need for data centres, new prisons are clearly needed to allow us to imprison dangerous and repeat offenders for long enough, and do so in humane conditions. (More modern prisons might also help us to improve rehabilitation, lowering crime that way as well as via incarceration.) As well as the direct need to supply data centres with power, building more nuclear power plants and other low carbon sources of energy is vital if we are to have any hope of achieving Net Zero at all, let alone in a way that involves energy abundance rather than extremely costly demand suppression.
In fact, the national need for these kinds of infrastructure projects is so great that there is a case for greater than 100% rates ‘retention’ – ie, some kind of top-up from the rest of the country to areas that permit essential infrastructure to increase the incentive to accept these projects even more. And there may be a case for contributing to other local governments as well as districts, like neighbourhood forums, parish councils, and city mayors. But this would obviously be more costly to enact than what we’re suggesting here, and we recognise that funds are scarce.
Doing this would be surprisingly easy. It would not need any new primary legislation, other than a line in the next Budget’s finance bill.
It can be done without changing any existing revenue streams. The only loss in terms of expected future revenues would be for projects that would have happened anyway under the status quo. But if these projects wouldn’t have been built under the current system this cost doesn’t apply.
The other small downside is that there will be a that local councils will be more inclined to support these projects over ones that do not come with full rates retention, if they have to choose between them. We think this is probably very small, since it will only be a problem where a council has to decide between a project like this and another equally or more valuable project – eg, for the same site – but it is still a downside worth acknowledging.
The approach we outline would not solve every problem. Locals often still object to projects with high rateable values. People may not always see or feel the benefits of higher tax revenues going to their local authority, let alone connect the dots between a high profile project, its business rates bill, and funding of the local services they use. We realise that extremely few people see things that way. But local councillors are more likely to understand, and they may decide that the prospect of a larger revenue return from a proposed project is worth the political flak they will take from their constituents if it goes ahead. They may also be able to communicate the benefits to their voters in more tangible terms that assuage their objections enough to get it built.
Getting the delivery right
We think this approach would go well with Labour’s proposals, though we have some concerns about the details of Labour’s plans and suggest they could be modified to work better. Labour is considering applying the 'National Significant Infrastructure Projects' process to more types of development, where development is permitted through a Development Consent Order. This system is slow, and developers often deliberately avoid it to use the normal planning system. There are other options worth considering for some of the projects, like Special Development Orders or just using the Housing Secretary’s existing, nearly unlimited powers to rewrite local plans to permit the strategic development.
But whatever mechanism Labour picks, it will not fully solve the problem. Local concerns about developments can still be effective at the national level – the campaigns against the A303 Stonehenge Tunnel (which was also granted a DCO) and the Aquind interconnector in Portsmouth are proof of that. One small, easy-to-enact change to business rates could go a long way to helping Britain to build things again.