Delivering affordable housing through Section 106: outputs and outcomes

A significant proportion of affordable housing provision is delivered through the land use planning system using Section 106 (S106) agreements. However, little is known about whether such agreements are being fully implemented. As government statements suggest the increasing importance of S106, this research seeks to address current knowledge gaps regarding the outcomes of S106 planning agreements.

Through detailed analysis of a range of evidence and a sample of sites, this report:

• examines whether the affordable housing element in S106 agreements is being delivered, and the reasons for any delay or non-delivery;
• assesses whether negotiated changes to initial agreements and non take-up of planning permission result in the loss or modification of plans for affordable housing;
• identifies common patterns in the delivery of the affordable housing through S106 agreements.

At a time when planning permissions for affordable homes provided through S106 are increasing rapidly, this research provides valuable insights for the future monitoring and implementation of S106.

Example 2: Renegotiation of minor details

This example covers some minor renegotiations of the S106 but not of the affordable housing element. The total number of homes agreed was 385 of which 97 (or 25 per cent) were affordable. Of the affordable flats, 70 were for rent with the aid of social housing grant, while 27 were shared ownership which did not require grant. The two types of tenure were built in separate blocks of flats, but close to one another. The affordable housing was totally separate from the market housing: “Effectively there was a party wall between the two types of housing, almost making it two different sites.” The housing association was brought into the negotiations after the initial discussions but before the S106 was signed. The developer delivered both market and affordable housing as originally agreed. There was some renegotiation over the type of public art to be provided as part of the S106 but this did not affect the housing.

What was found on the ground

The research examined 39 sites to explore any differences between the outcome and the original S106 agreement and why these occurred. These sites had S106 agreements signed between 1998 and 2005 and were fully or partly completed.

For 24 sites all had been built as agreed. The only variations to S106 agreements related to non-housing matters with no changes to the numbers or sizes of homes produced. One RSL commented that the standard of the social rented houses was lower than expected.

On five sites, despite planning permission being granted prior to 2000, the site was still being developed at the end of 2005. These sites account for 576 of the 2,174 (27 per cent) of affordable homes in the research database granted planning permission prior to 2000 but still incomplete at the end of 2005. This suggests that, although long delays between permissions and completions may only affect a small proportion of sites, they often affect very large sites producing a high proportion of the total number of affordable housing units granted permission.

By far the most common outcome was that, where construction had started, the S106 housing was delivered, i.e. the S106 approach is working well. Some sites were still being built, but local authorities and housing associations confirmed that they expected that the full quota of affordable housing would be met.

A deed of variation was also fairly common, especially on larger sites. This may allow the local authority to increase the proportion or numbers of affordable housing delivered. Deeds of variation often relate to changes in site phasing rather than to the delivery of affordable housing as a whole: research on specific sites found that most amendments to S106 agreements do not affect the affordable housing provision.

Renegotiation can produce fewer affordable homes. One S106 specified 30 per cent affordable homes scattered throughout the development. However, this was an up-market block of flats with access to a sports and pool complex: no social landlord would take on the high service charges. The developer negotiated a commuted sum and the on-site affordable homes were lost (although the commuted sum will go towards another development).

Sometimes renegotiation can deliver more affordable homes: for example, where it provides a higher number of units overall. In such cases, the numbers of affordable units (though not the proportion) may increase through higher densities.

Cases where all parties accepted the change but no formal agreement was felt necessary were relatively rare: most parties prefer to ensure official agreement. In one example, the affordable homes were specified as shared ownership in the S106 but, after handover, the housing association received additional funding and turned them into social rented homes. The developer was not involved in the decision.

Where one or more parties did not achieve what they expected, this was usually because the relevant elements were not fully specified in the S106. For example, one party may be disappointed with the quality of the homes provided or their precise location on the site.

It is difficult to assess the extent of cases where the planning permission is not implemented. Data obtained for other projects suggests that approximately 20 per cent of permissions are not taken forward. However, as these were pre-2004 permissions, there is a five-year delay before it can be stated with any certainty that the planning permission has expired and the units will not be built.

Cases where, having implemented the permission, the developer fails to comply with the S106 affordable housing obligation are extremely rare: although stakeholders talked about this, there was no evidence in the research database. Only one in five authorities consider failure by developers to deliver planning obligations (including affordable homes) to be a problem.

Of the 39 developments, in 10 cases both the RSL and the local authority were unable to identify the site from their records and the outcome was unknown. Some local authorities and housing associations seemed unable to access any records which would tell them whether or not the S106 agreement was in fact delivered. Problems in retrieving data included: multiple names for each site; several sites to which information might have applied; staff turnover; poor record-keeping; and loss of records following office moves. Clearly, monitoring is a concern: better monitoring by local authorities and housing associations would reduce the incidence of cases where the outcome cannot be ascertained.

Conclusion

Once development starts on a site, in most cases, S106 pretty much delivers what was agreed. The two most important issues are:

• the general planning permission not going ahead; and
• delays associated with large, complex or brownfield sites which often involve renegotiation.

There remains an issue over whether the outputs are as good as the original stakeholders had expected.

However, this growth and compliance has occurred during a period of economic expansion. A downturn in the market might significantly change both attitudes and development behaviour.

Example 3: A large complex site

A developer specialising in brownfield redevelopment bought a former hospital site. The development was in phases. Phase 1 was for 400 homes, of which 21 were affordable, 11 social rented and 10 low-cost home-ownership. Phase 1 delivered all that was originally agreed and is popular with all residents. However, after Phase 1, the developer divided the site and sold on parcels to other developers. As a result, the development has not progressed as quickly as planned. There is only one housing association on site. It is now involved with multiple developers who inherited the S106 agreement. It also bought one parcel which it has developed itself. The affordable housing is spread throughout the site, although it is described as "'chunk pepper-potting', because there were three different phases to it, but it is not all shoved in one corner". The site is now in its third and final phase, which forms a demonstration project for the Urban Village Task Force. The housing association has taken over all the S106 affordable housing as well as its own development. The developers have all met their S106 obligations in terms of the numbers, but they did not produce homes of the same standard as those built by the housing association.

About the project

The research was undertaken by Sarah Monk, Diane Lister, Roland Lovatt, Aoife Ni Luanaigh and Christine Whitehead from the Cambridge University with Tony Crook and Steven Rowley from Sheffield University. Evidence on quality issues is from a related study by Jon Watson.

After updating secondary data, a scoping survey of key stakeholders was used to develop a typology of possible scenarios. Local authorities, housing associations and in some cases the developers involved in these sites were then telephoned to find out whether there were any differences between the outcome and the original S106 agreement and why these occurred. The site-specific analysis drew on earlier analyses together with a detailed re-examination of the team's database of 39 sites. Six sites were then investigated in more detail to gain a better understanding of what happens after the original S106 agreement was negotiated.