A government truly committed to economic growth would make a plan to build one million houses in this area. Cities and areas can grow quickly - for example, the population of Seoul grew by more than 3 million people in both the 1960s and 1980s, the Dallas metro area grew by 1.5 million in the last decade, and by a million in the previous, while Phoenix Arizona metro area grew by a million in the 2000s. That is how fast cities can grow in dynamic economies, and the UK should aim to add a million houses - say 2.4 million people - within a maximum of 25 years from announcing the plan. 

That plan should look to do the following.

This would involve deciding how many of the new houses will be in and around each of Oxford, Cambridge and Milton Keynes - all successful economies, and how many will be in new settlements. It is generally the case that expanding existing settlements has the highest returns, but new places can be valuable - at least when they are of sufficient scale to generate agglomeration economies of their own. We do not want lots of small dormitories.

The government should use “community land auctions” to assemble land. Rather than designating land for development and then negotiating over the price of the land with the landowners from a position of weakness, the government should ask landowners in each of the chosen, broadly defined areas how much they want for their land. Thus, for example, the government might be broadly indifferent as to whether Cambridge expands to the north, south, east or west. Landowners in those areas might be told that they have a roughly 1 in 10 chance of being chosen and that the price they name will materially affect their likelihood of being chosen. In effect this is competitive tendering - exactly the process that the government uses to buy everything from ambulances to school buildings. It seems likely that the government will be able to buy land for twice agricultural value and certainly for five times (under this approach, there is no other route to getting planning permission, increasing landowner incentive to participate).

To ensure progress is not stalled, the government could use compulsory purchase order (CPO) to overcome ransom strips - small parcels of land that prevent wider development - and for land needed for new infrastructure, as necessary, but would not aim to use CPO intensively.

Having assembled the right amount of land, the government or the local mayoral region can declare a development corporation, taking away the power for local areas to say no. This is how the Docklands got built. 

The new corporation would be given loans by the government and should hire planners to design the new areas. In all cases, a strong premium should be placed on providing not only housing, schools, parks and so on, but also office and lab space.

Ancient woodlands, sites of special scientific interest (SSSIs) and other high-value conservation areas would be protected as usual.

To streamline development, no environmental surveys would be taken of ‘typical’ land. Instead, reference values will be taken from the average value of previous developments, and biodiversity net gain will be applied on that basis. It is silly to spend a fortune on surveys, rather than using the money to enhance the environment. The net gains would all be within or near the newly developed areas - for example, by providing high-quality parks.

Utility firms would be required to provide water, sewage, electricity, broadband and 5G, but not gas. The regulators would allow the relevant rates of return for their investments.

The houses would be built to high levels of insulation, would be heated with heat pumps, contain domestic batteries and have air conditioning. They would be both a delight to live in and cost little to run.

Densities will be relatively high - cores of 350 homes per hectare, suburbs at 60-80, ensuring that lots of local shops and services are viable, and making these ‘walking places’. This also increases the viability of the settlement in construction terms - cities that are smaller per person need fewer roads, public transport and so on, reducing the total cost to build. 

The government can ensure that the houses are sold very quickly by selling them on a community land trust (CLT) basis. This means that people will pay only the direct costs of construction, including a fair contribution towards infrastructure. As a broad indicator of price, new-build four-bed, two-bath detached houses with garages start at about £235,000 in the Midlands (taking into account the 5% deposit offer). Given land acquisition via community land auction, it should be possible to build and sell houses for this sort of price. 

For example, the cheapest recent new-build four-bed houses in Oxford was £700,000, for an end-terrace with no garage. The equivalent is £580,000 in Cambridge, while Milton Keynes is cheaper, but prices of £450,000 for a semi without a garage are still far above the levels that we would expect to charge. If the CLT were offering houses at 40-50% of current prices, we would expect them to sell like hot cakes - as quickly as they can be built. Owners would only be able to sell them back to the CLT in due course, they would receive the original price plus inflation. This means there is no chance of negative equity - making these great purchases for first time buyers and very low risks for banks. 

Some houses will be sold conventionally as freeholds initially, according to demand (some people will want the possibility of appreciation) and over time the CLT houses will be resold on that basis. In the long run, therefore, the government will recoup all that it invests, and plenty more besides. A typical amount of social housing will be provided. Since the land is acquired relatively cheaply, the cost of social housing will be lower than would typically be the case. 

Office and other commercial space, including shops, labs and so on will be sold at full commercial prices. This will generate a significant surplus and will be used to pay down some of the money provided up front by the government for land acquisition and infrastructure provision. 

The government would ensure that services are provided. Given the scale, that means not only the standard offer of schools and GPs, but a full range of new hospital provision, including mental health, a new university or significant university expansion and so on. We should note that some of these costs are additional, but not all. The number of children in England is falling and therefore the cost of schools is additional - they would not otherwise have to be provided. That is not true of hospitals, however. We will need more and upgraded hospitals come what may, a plan like this simply means that they will be needed here rather than elsewhere. Indeed, to some extent this plan will lower cost pressures elsewhere: reducing pressure on other hospitals means that they are more likely to be operating at a sensible level, rather than bulging at the seams.

New transport links would be provided. What's needed will depend on location, but the shape is clear enough. At the Oxford end, a Reading–Heathrow connection would let the OxCam arc train reach the airport with at most one change: either a tunnel from Langley, or, more cost-effectively, a western element to the Stockley Flyover. At the Cambridge end, links to Stansted would be improved. Alongside these, the arc needs better connections within its own core, and outwards to both north and south.

People will, of course, come up with many objections, because this is Britain in the 2020s. They will say that we are short of water or workers. None of these objections is valid. Reservoirs take time to build, but are not intrinsically difficult to construct. Milton Keynes - which has built more houses than pretty much anywhere else in the country for many years - finds the workers, because when good jobs are available, people train for them. With the right ambition and a definitive plan for housing, the government can harness the unique potential of the OxCam arc, building the vibrant, high-growth economy that our future demands.