Land is the fundamental component of new housing yet it is probably the least transparent part of the market. I have looked at the value of land in a previous note (http://sav.li/45e) but data on its ownership, use and sale remains limited. However, the DCLG have recently released data on the change in land use across England that provides some useful insights.

The data shows that 60% of new residential address built in 2013-14 were constructed on previously developed land. It falls to 45% for net additional addresses and only 23% of land used for housebuilding was previously developed (reflecting higher densities). This is a significant change from the last reported figures in 2011 as per Fig 1. Some of this will reflect changes in the market post NPPF but most of it is due to the new methodology. It takes a more detailed approach to measuring change* and unfortunately makes historical comparison (& the chart below) useless.

The DCLG release also provides some context on existing land use and its designations. It indicates that ~11% of England is developed. In terms of land with development constraints: ~13% is covered by Green Belt while National Parks, Areas of Outstanding Natural Beauty and Sites of Special Scientific Interest cover 29%. Adding them together while allowing for overlaps indicates that ~40% of England is covered by these restrictions. The map below highlights how much of the south-east is covered by these and other development restrictions/risks (e.g. high flood risk).

Brownfield Land

Developing on brownfield land has re-emerged as a priority for Government. They have committed to statutory registers of brownfield land for housing but currently the most comprehensive data on brownfield land was released in 2010. That data suggested there was brownfield land with the housing capacity for just over one million homes. Many of the issues arising from a reliance on brownfield land have been well discussed in other research. These include issues such as the actual capacity of the land, its location relative to housing need and that many of the sites are already in use. We also need to consider whether it is financially possible to build on these sites.

As a very rough guide, it costs around £100 per square foot to build a new home on a simple development-ready site (so excluding abnormal costs) and a typical new home needs to sell for at least £200 per square foot once developer profit, payment for land and other costs including finance have been included. Using £200 per square foot as a price threshold suggests that around 40% of the potential homes in the brownfield database are in areas where likely new build house prices may make development financially difficult if not impossible (Fig 3).

Realistically, the cost of development on brownfield land may be even higher due to factors such as site remediation and higher densities in urban locations. Reducing the risks and costs involved in planning will help deliver some additional sites but the housing capacity of already identified brownfield land may well disappoint.

Previous analysis by Savills has found that the upper quartile price per square foot of existing housing in a local market provides a good proxy for new build prices in a normal competitive market.

In markets with less competition, new build prices tend to be closer to the upper decile of the existing market.

Green Belt

It is unlikely that we will see top-down reform of Green Belt during this parliament but that does not mean no new houses built on Green Belt. Instead, we will continue to see locally led reviews releasing or swapping Green Belt land for development. Some local authorities have seen a large proportion of their new homes built on the Green Belt, albeit at low numbers. At a national level, the DCLG statistics indicate that 3% of new addresses were on Green Belt designated land and that equates to ~3,500 homes in 2013-14. However, it is worth noting that building on the Green Belt does not necessarily mean building on green fields, 62% of those addresses were built on previously developed land.

77% of Three Rivers local authority (covering Rickmansworth, west of Watford) is designated Green Belt and 76% of new addresses in 2013-14 were built in the Green Belt according to the DCLG data. However, given the low number of total new homes built, that only equates to ~100 homes built on the Green Belt.

*the method now considers multiple previous uses on a single development site. The example given is a hospital that historically would have been considered as wholly ‘previously developed’ but is now broken into component parts with buildings etc classified as ‘previously developed’ and the grounds as ‘not previously developed’.

Taken from
Housing Market Notes:

• Caught in chains
• The housing crisis
• House prices burning bright
• (Social) home truths
• The value of land
• Approaching peak loneliness?
• Size matters
• Housing delivery
• Pension reform and the housing market
• Housing market activity
• Rental Britains
• First time buyer affordability
• Local house price indices
• Measuring housing need
• Magnetic London