Scotland 2045 - fourth National Planning Framework - draft housing land requirement: explanatory report

Method

Approach

33. The statutory requirement is new and has required a new approach to be developed. It is met through the provision of the MATHLR.

34. We have sought to provide a consistent approach across the development planning system in Scotland, that is simpler when compared to current policy expectations and is therefore more transparent. We have used a robust and evidence-based approach to ensure national analysis is informed by local information and policy ambition, resulting in figures that are based on current housing need and future demand for housing.

35. The approach has produced broad, reasonable and long-term requirements, as is appropriate at the national scale. It looks to provide clarity early in the development plan process and confidence about the baseline amount of land to be identified locally. Local flexibility will remain in that authorities will be able to allocate additional land through their LDPs, should this be supported by additional evidence or take-up of land at a faster pace than estimated. The approach will also provide for local needs to be met by enabling local planning and housing authorities to address place-based matters, for example location, priorities and deliverability, at the local level.

36. The approach reaffirms that it is important for the planning and housing systems to relate to each other. The important linkages between the LDP and LHS remain. The future components of the Development Plan, NPF and LDPs, will be moving to a ten-year review cycle, while the LHS will remain in a five-year review cycle. The LHS review provides an opportunity to update local evidence and test the appropriateness of the housing land requirement identified in LDPs. In carrying out this work, planning and housing interests across Scottish Government have worked collaboratively and we have sought input from planning and housing interests in local authorities, who have engaged with their local stakeholders.

Consultation

37. A discussion paper^[13] setting out a proposed methodology that could be used for setting 'targets for the use of land in different areas of Scotland for housing' was published in March 2020 and was consulted on alongside the NPF4 Call for Ideas. The period for comments was extended, as it was for the Call for Ideas, due to the pandemic. An independent analysis of the consultation responses was included as part of the analysis of the NPF4 Call for Ideas^[14].

38. 25 respondents replied directly to the consultation. This included responses from local authorities, house builders, Homes for Scotland and community councils. Other respondents made relevant comments under the housing-related themes in the broader NPF4 Call for Ideas consultation and these views were also included in the independent analysis.

39. The discussion paper introduced a number of guiding principles for a proposed methodology. Whilst there was a mix of views on these, with some respondents saying they could not support them or that the principles may have some problems working in practice, there was general support for them. A brief summary of the feedback received on the guiding principles is set out below:

• Providing early clarity and reducing conflict and complexity
  There was support for moving towards an approach that would reduce debate around housing land requirements and reduce conflict and complexity.

• Ensuring LDPs allocate sufficient land for housing
  Views indicated that it is impossible to state the quantum of land needed without first assessing the number of homes required. The process should also be outcome focused and any land identified as a result of a housing land requirement must be presented in the context of being necessary to meet a housing supply target. A minimum requirement could be open to interpretation and will increase opportunities for debate and conflict.

• An agreed proportion of this land should be deliverable
  There was broad support for the principle of focusing on deliverable, rather than effective land. However, some respondents stated their disagreement with this principle. For example, it was stated that identifying deliverable land should not be used as a reason to reject other sustainable and deliverable home building opportunities if it cannot be demonstrated that the identified land is free of barriers (market or otherwise) to the delivery of homes. It was not clear what 'an agreed proportion' meant in practical terms, and how this would be measured and established. Nor was it clear with whom agreement would be required, or what was meant as 'deliverable' in this context. There should be a clear plan of action within LDPs to show that every site is, or will go through defined steps to become deliverable. This would establish deliverability over the short, medium and longer term.

• Minimum figures should be set for all local authority areas
  Most agreed with the principle of setting minimum figures for all local authority areas and there was support for consistency, although some expressed concern that it could result in unsustainable development and the undermining of spatial and regeneration priorities. There was majority support for a single figure rather than a range.

• National approach should be informed by regional and local knowledge, analysis and input.
  Input of regional and local knowledge and analysis was welcomed and would be crucial in ensuring that the targets are realistic and result in the delivery of high-quality homes in the right locations. While a national approach requires support at the regional and local level, this should not be to the detriment of meeting Scotland's housing need and demand.

40. The discussion paper proposed a methodology for estimating the housing land requirement as:

• The Scottish Government could run the first steps, steps 1 and 2, of the HNDA Tool using the default scenario and assumptions of:

  a. Household projections - the default scenario choice is the National Records of Scotland (NRS), 2018-based, principal household projection.
  b. Existing housing need count - the default is the proxy method built into the tool which is a count of homeless households in temporary accommodation and households who are both overcrowded and concealed (HoTOC).

• Local areas could agree or propose adjustments to the scenarios and assumptions noted above. Authorities would be required to agree the scenarios and assumptions with their Housing Market Partnership (HMP) (including key wider stakeholders such as Homes for Scotland). In some areas, authorities may wish to reflect functional housing market areas that cross local authority boundaries. In this instance, they could propose adjustments at local authority level.

• The Scottish Government could apply a level of flexibility.

• Housing land figures would be set out in the Draft NPF4 and be subject to public consultation and transparent scrutiny. Draft NPF4 would be accompanied by a report supporting the housing land figures.

• A revised NPF4 would subsequently be laid before Parliament. There would be an opportunity to update the data between draft and revised NPF to reflect up-to-date information.

• LDPs would be expected to meet the housing land figure set in NPF4 as a minimum. Any adjustment to the figures would need to be agreed at the LDP gate-check stage.

41. Responses to the consultation raised a number of matters regarding the methodology. These are summarised below:

• There was a general view that the housing land requirement should be applied nationally. Although, as the requirement for housing land varies from area to area, there was a view to apply local evidence to support local requirements.

• There was general support for a single figure as this would clearly define housing land requirements and would reduce debate at the development plan gatecheck and examination, provided that local knowledge and expertise had been brought to bear.

• Using the HNDA Tool as a starting point to commence the process of setting a housing land requirement was generally supported; with the importance of considering policy drivers and their impacts being stressed.

• With regard to using NRS household projections, it was recognised that a more rounded approach is needed to avoid setting housing numbers that merely reflect trends rather than future potential.

• With regard to using HoTOC, its' limitations were noted as homelessness may be more broadly defined.

• With regard to using flexibility, this was noted as allowing for sites to come forward when others had failed or were delivering more slowly than programmed.

• It was highlighted that there should be the opportunity to reflect functional housing market areas, that cross local authority boundaries, through effective regional working.

• There was a strong call for there to be clear definitions of meanings for key terms.

Housing Advisory Panel

42. A Housing Advisory Panel chaired by the Scottish Government's Chief Planner, was established to help guide this work. It comprised a cross-section of approx. 12 members that were invited to take part on a personal basis due to their knowledge, skills and experience on housing and planning matters. Their experience was drawn from across the public and private sector and representative organisations as well as academia to reflect a broad range of views.

43. Recognising the strong and varying views of members, the remit of the panel was clarified to confirm that the role did not extend to producing outputs or recommendations, or taking decisions. It was also clarified that individuals were not representing a particular organisation.

44. The panel met in September 2020 and March 2021. Discussions focused on the approach and methodology for estimating the MATHLR. Significant points discussed and noted by the panel included:

September 2020

• It will not be easy to take conflict out the system; there will be challenges given the substantial vested interests. It is however, a worthwhile goal to move to a system with less conflict.
• Housing markets and geographies are very local, making it difficult to manage at a national level. In rural and remote areas, housing can be very different.
• Wider policy levers and budgetary context are important, for example Housing to 2040.
• Identifying numbers is the start of the process – their deliverability will be key.

March 2021

• Population and housing projections are trend based, they do not take into account ambitions for an area.
• The types of existing household need accounted for in the HNDA input data are limited, there are likely to be other areas of need relevant.
• Completions information uses an average over 10 years, over which completions generally increased to 2019 and can be looked at using Scottish Government data or local authority Housing Land Audits.
• In some areas, initial estimates are low.

Consideration of Matters Raised

45. Taking into account the responses to the discussion paper and the contributions from the Housing Advisory Panel, the approach to establish the MATHLR was refined. This built upon the approach presented in the discussion paper and represented fine-tuning and iterative amendment.

46. A Q&A^[15] document was prepared that outlined the Scottish Government position on a range of the matters raised: key points are summarised below:

• The NRS household projections are trend based however, the method proposed provides for local authorities to take account of national, regional and local policy drivers to inform their choice of projection or any alternative. As the planning process moves to the local level and LDPs are prepared, authorities will have an opportunity to reflect any updates to evidence, including policy drivers.

• The HoTOC data contains information collated nationally however, the method proposed is a starting point: it is recognised that existing need will differ locally around Scotland reflecting different housing market dynamics. Local authorities are asked to use robust, locally available information that evidences a requirement for an additional housing unit.

• HNDA is a well-established tool that uses the best available data to provide a consistent approach. The method proposed provides for the HDNA inputs to be refined using robust local information.

• Understanding Housing Market Areas (HMAs) and the dynamic of them is important. The method proposed provides the ability to work across boundaries in regional groupings and where that is preferred, HMAs will be particularly relevant.

• The completions information provided was a 10 year average from 2010 to 2019. It is based on local authority NB1 and NB2 returns (generally from building standards departments) and Housing Association new build information provided to the Scottish Government (Housing statistics quarterly update: new housebuilding and affordable housing supply - gov.scot (www.gov.scot)). These are submitted and collated systematically, as well as quality assured and formally published. Whilst there is best practice on Housing Land Audits, there is no standard approach.

47. A number of amendments and clarifications were made to the approach, including:

• Rather than requiring authorities to 'agree' the figures with their HMP, it was instead considered appropriate to require authorities to 'engage' with them and relevant interests. This reflects that it is local authorities that have responsibilities for planning and housing and preparing LDPs and LHSs.

• Rather than the Scottish Government apply flexibility at the end of the process, it would instead be considered locally with other inputs.

• Rather than flexibility potentially being used to lower figures, it would instead only be applied to increase figures to meet its intended objective.

Inputs

48. The calculation for arriving at the MATHLR is set out in Figure 1 below:

49. For household projections, the 2018-based principal projection of National Records of Scotland (NRS) data is used^[16]. This equates to step 1 of the HNDA process.

50. For existing housing need, this uses a count of homeless households in temporary accommodation and households who are both overcrowded and concealed using NRS 2011 census information, the 2016-2018 Scottish Household Survey and Scottish Government homelessness statistics as at March 2020. This equates to step 2 of the HNDA process.

51. A flexibility percentage is then applied. This represents a contingency of land to allow for changes in sites coming forward. It is in the long term public interest to ensure there is sufficient land available to meet future requirements, in the knowledge that over time there will be changes that effect whether sites come forward or not.

52. The flexibility is similar to the 'generosity' applied in current Scottish Planning Policy (SPP); in that it reflects the aim of ensuring that LDPs allocate sufficient land for housing. The percentages are higher than the SPP (10-20%) to account for the change in legislation that now requires plans to be reviewed at intervals of no more than 10 years, instead of the previous 5. As plans will likely be in place for longer, there is a greater chance of changes to sites coming forward.

53. The flexibility applied is 25% for urban authorities and 30% for rural authorities. The distinction is made to reflect the statutory requirement that the NPF must contribute to the outcome of increasing the population of rural areas. The Randall Classification is used to distinguish authorities, with those authorities with a population density below one person per hectare identified as rural and those above identified as urban.

Terminology

54. Clarity of the use of terms was considered important. They can therefore be explained as follows:

55. The figures will be a minimum amount of land to be identified within LDPs. Expressing the figure as a minimum requirement prevents this being interpreted as a limit or a cap to development: planning authorities through their LDPs will be able to have higher local housing land requirements where evidence justifies this and the home building industry will then be able to fulfil and exceed delivery of homes above the minimum national requirement.

56. The figures will be all-tenure as it is the scale of land that is relevant for national spatial planning purposes. Different tenures, for example market and affordable housing, will continue to be considered at the local level in the more detailed planning for housing processes of the HNDA, LHS and LDPs.

57. The figures will focus on the delivery of housing land. The statutory requirement of the Act relates to 'use of land' and it is land use that the planning system regulates. LDPs allocate land where housing will be located and the aim is to ensure there is enough land planned, in the right locations, to meet housing need and demand. As the amount of land to be allocated is informed by the amount of housing need and demand, expressing the requirement as a number of units is more relevant than expressing it in acres or hectares.

58. The figures are termed a requirement to better convey the intention that they are to be met through allocation of sufficient land within LDPs. Reference to targets could be inferred as aspirational amounts that authorities try to achieve. There is also the potential for confusion with the Housing Supply Target that is set out in the LHS.

Process

59. The process for arriving at the MATHLR is set out in Figure 2 below.

60. The Scottish Government ran the first two steps of the HNDA Tool using default scenarios and added the relevant flexibility allowance. This provided Initial Default Estimates for each local and national park authority.

61. Authorities, either individually or in regional groupings, with their Housing Market Partnership (HMP) and local stakeholders, then considered the Initial Default Estimates and, where relevant, they proposed Locally Adjusted Estimates based on local information and relevant policy drivers.

62. The Scottish Government then considered the Locally Adjusted Estimates and have proposed the Minimum All Tenure Housing Land Requirement for each planning authority in Draft NPF4.

63. Following engagement and scrutiny, requirements will be included in the final NPF and they must then be taken into account in preparing LDPs.

Initial Default Estimates

64. The Initial Default Estimates (IDE) are set out in Appendix A.

65. The detail of how the IDEs were arrived at is set out in a Method Paper^[17]. As well as the key inputs, amendments to the data were made to enable figures to be produced for national parks, to annualise figures to a 10 year timeframe and to round to the nearest 50. The authorities were also re-ordered to reflect regional groupings.

66. It was made clear that the Initial Default Estimates were a starting point for local consideration and to enable local input. They represented the beginning of the process and were, in general, a statistical and policy neutral figure to build on. The estimates were expected to increase through local input that took into account public and private sector ambitions to support growth in housing provision: they were not the estimates that were expected to be included within Draft NPF4.

Local Input

67. In February 2021, the Scottish Government Chief Planner and Director for Housing and Social Justice wrote to local authority Heads of Planning and Heads of Housing to seek the input of their authorities and relevant stakeholders to meeting the statutory requirement. This included:

• A letter^[18] that emphasised the Scottish Government's commitment to enabling the delivery of high quality homes and the need for positive engagement and constructive co-operation across planning and housing and leadership with local stakeholders from the public and private sectors. It confirmed the Initial Default Estimates as a starting point.

• A Method Paper^[19] outlined the methodology for the Initial Default Estimates.

• A Template^[20] was provided for authorities to complete and return.

• Guidance^[21] to support completion of the template.

• A calculator^[22] was provided to support consideration and to be completed and returned.

• A presentation^[23] and Q&A document^[24] were also made available following discussions with the Heads of Planning Development Planning sub-committee and the Housing Advisory Panel.

68. Working with their HMPs and local stakeholders, authorities were asked to consider the estimates and whether robust local information and policy drivers indicated an adjustment was appropriate. It was expected that authorities would take into account the completions data provided.

69. Responses were sought by early June 2021. A number of extensions to this timescale were sought, with responses being received during the month.

Locally Adjusted Estimates

70. Copies of submissions from authorities are available online: Responses from Planning Authorities | Transforming Planning. Information submitted provided commentary on a range of matters, summarised below.

71. Household Projections: The Initial Default Estimate was based on the NRS 2018-based principal household projection. Many authorities pointed out that these are a continuation of past trends and were based on the 2011 census which, given the time elapsed, needs to be considered carefully alongside other evidence. It was indicated that the Locally Adjusted Estimate should apply a policy interpretation to the NRS projection. Overcoming these issues, to take account of policy, was a consistent component in evidence submitted. This emphasised taking account of local economic growth/growth deals (and housing supply to support this), demographic change, affordability, Housing to 2040 and post-pandemic outcomes.

72. For some, the adoption of a high migration variation on the NRS 2018-based household projection was sufficient to accommodate limitations. Others sought to adopt their own forecast for household growth.

73. Existing Housing Need: The HNDA Tool counts two types of existing housing need - homeless households in temporary accommodation and households that are both overcrowded and concealed. The HNDA Tool and supporting guidance state that the existing housing need figure used by the Scottish Government is only a minimum and that other types of existing housing need exist. Local authorities and stakeholders highlighted a range of additional factors including higher levels of homelessness, some types of households on social housing waiting lists, affordable housing need, special housing need, Below Tolerable Standard housing and/or, where available, results of local HNDA surveys where these were quality assured.

74. The near unanimity of this expanded assessment of need was recognised. Consequently, this represented an opportunity to expand existing housing need where evidence to quantify it was available; including ensuring that it resulted in a need for a new home and that there was no double counting. This expansion of existing housing need based on local knowledge and local housing markets mirrors what many authorities can, and do, already for their HNDA.

75. Further, for local household surveys to be relevant and offer strong evidence, they need to be robust and representative of the local area and community. The questions asked need to eliminate bias and the representativeness of the surveys must be demonstrable. Any limitations of the surveys must be clearly articulated when the findings are used. A knowledge of survey design is required for this. Local household surveys can be costly and time consuming to undertake and a careful cost/benefit analysis of undertaking one should always be carried out including an assessment of what data sources already exist and whether these are sufficient in terms of the question(s) to be answered.

76. Flexibility: The application of flexibility to rural and urban areas was accepted in all but two instances. Aberdeenshire and the Cairngorms National Park indicated a preference for different percentages. Aberdeenshire sought the urban 25% as they consider the population density is skewed by the area of the authority within the Cairngorms National Park. The National Park sought a change to 10%, highlighting potential impacts on nature conservation interests. Authorities are encouraged to be more directive in where new development should take place, which can be away from areas authorities consider there will be issues with deliverability. As a result, the Randall Classification was considered appropriate despite its limitations.

77. Completions: It was thought prudent to take account of the 10-year trend in completions as a comparative benchmark. Homes for Scotland was of the view that as there is evidence of general growth of completions each year between 2010 and 2019, using an average from this period would not be suitable. It was suggested that current completion levels (2019) should be used as this reflected pre-pandemic capacity following the financial crisis of 2008. However, it was argued that the 10-year average trend was reflective of delivery and more robust than a one-off year; particularly where this was the year with highest completions for most local authorities.

78. Past completions provide evidence of past trends in the delivery of new housing units. Completions do not in themselves, create demand for a new unit of housing. The information should not therefore be a determinant of future housing requirements but is a helpful factor to use as a comparative benchmark to sense check whether a requirement is reasonable.

79. Housing Land Audit (HLA) Completions: A view emerged amongst local authorities that the HLA completions represented a robust record of completions, although many HLA completions omitted developments on small sites or made assumptions in this regard. This is because HLA completions are used to inform the preparation and monitoring of LDPs. However, research^[25] has indicated there is a lack in consistency of approach across Scotland to HLAs and given that completions, whether measured through the HLA or by returns to the Scottish Government, should not be a primary determinant for the housing land requirement estimate, but rather act as a benchmark; local authorities were encouraged to retain the quality assured Scottish Government 10-year completion figure for this purpose.

80. Delivery: A view was expressed that low housing land requirements can create complications for infrastructure planning and delivery, and uncertainty for communities. Similarly, an over-generous housing land requirement can undermine confidence in the delivery of housing. Most respondents sought to use completions data, either Scottish Government or HLA, as a benchmark against which the adjusted estimate could be compared and contrasted.

81. Housing Market Areas: Glasgow City Region (GCR) set out that the principle of housing demand being met irrespective of local authority boundaries is highly applicable in the GCR as there are strong interconnections between the outer authorities and parts of Glasgow. The GCR argued that an approach that requires a MATHLR for each local authority, disregards established policy based on functional Housing Market Areas. Planning for housing based on a local authority area, rather than on a HMA, may lead to excessive green belt planning pressures in areas of buoyant commercial market demand and reduce the effectiveness to direct development to areas requiring interventions and support, usually on brownfield sites. Consequently, the MATHLR should be for functional HMAs.

82. It is a statutory requirement for each planning authority to prepare an LDP and across Scotland, generally, each authority prepares one plan: the exceptions relate to large geographical areas (i.e. Highland) and subject specific plans (i.e. South Ayrshire). The method for arriving at the MATHLR provides for cross-border issues to be considered and addressed through co-operation with neighbouring authorities, including consideration of functional HMAs. These considerations could lead to some local authority areas having an increased MATHLR, to accommodate a higher level of homebuilding than anticipated; whist other areas would be reduced where there was strong commercial demand with environmental and infrastructure constraints necessitating restraint. However, the over-all impact on the MATHLR would be neutral across the authorities

83. Rounding to Nearest 50: It was suggested that where an estimate is low, the '50s' rounding requirement should not apply as it can have a significant distorting impact that could undermine deliverability. Consequently, for local authorities where any component of the MATHLR is 50 or less, no rounding should be applied.

Housing Market Partnerships & Stakeholder Views

84. Generally, HMPs and stakeholder engagement informed the Locally Adjusted Estimates. However, the level of engagement was mixed. Largely, most HMPs and stakeholders viewed the Initial Default Estimates as low, requiring an uplift to reflect local circumstances and aspirations. Support was forthcoming for increasing the housing land requirement through expanding the definition of need and emphasising policy drivers (particularly with regard to economic growth and social housing provision) to increase household projections. Benchmarking through comparing and contrasting with historic completion rates informed these considerations.

85. In addition to responses from authorities, comments were received from the following:

• South of Scotland Enterprise, which supported adjusted estimates from Scottish Borders Council and Dumfries and Galloway Council.

• Chair of Aberdeenshire Council's Infrastructure Services Committee and Chair of Aberdeenshire Council's Communities Committee, who were in support of the joint submission from Aberdeen and Aberdeenshire Councils.

• Methlick Community Council, who sought a joined-up approached, agreed by all, which incorporates housing need and service provision with the wishes of the local community.

• Newmachar Community Council, seeking an increase in the number of social housing units built in the future.

• Homes for Scotland (HfS), who reflected on their engagement with local authorities, provided observations on the process, highlighted household survey work as a means for refining existing need and argued for a review of the HNDA Tool. HfS retained a number of concerns about the data/material that underpins the Draft NPF4 MATHLRs.

• The Scottish Property Federation shared concerns expressed by Homes for Scotland (HfS) and sought an upwards adjustment of housing need and demand, and corresponding minimum housing land requirements. It argued for higher minimum flexibility of 30% in urban areas and 35% in rural areas. Economic development and growth was highlighted as a key consideration in house building. The Initial Default Estimates, it argued, appeared to lack any ambition to address the six statutory outcomes set out in the Planning (Scotland) Act 2019.

• Colliers, on behalf of Mactaggart & Mickel Homes, indicated concerns regarding low figures for the Initial Default Estimates and planning for housing land for Edinburgh and the Glasgow conurbation. They encourage the housing land requirement to set ambitious targets with 250,000 over ten-years being suggested.

Assessment and the Proposed MATHLR

86. The Scottish Government considered the Locally Adjusted Estimates submitted by the local and national park authorities through Summer 2021. This reviewed evidence and views presented in relation to household projections, existing housing need and flexibility. It also included looking at the information provided on housing market partnerships and stakeholder involvement and policy evidence, as well as sign off by senior officials and the statistical evidence used.

87. Where accepted, the Locally Adjusted Estimates (or a minor revision to them) have been included as the proposed MATHLR within Draft NPF4. Where further explanation or clarification was considered required, discussions with relevant authorities took place to seek further information, which was used to inform the proposed MATHLR. Additional information submitted by authorities is available online Responses from Planning Authorities | Transforming Planning. Where matters could not be resolved, the MATHLR included within the Draft NPF4 reverted to the Initial Default Estimate or a variation of it.

88. Recognising that the proposed MATHLRs promote a strategic allocation of housing land to provide a broad estimate for local authority areas and is not intended to be precise, they were benchmarked against a number of factors, including completions, HLA programme, established housing supply, existing housing stock and an area's historic dwellings growth. Annex B provides details of the benchmarking information.

89. An Assessment Report for each authority or regional grouping, as relevant, has been prepared. Annex C provides links to the Assessment Reports for each authority or regional grouping. A single report has been produced for Fife. The information covers two areas, Fife North, and Fife Central and South. This reflects that Fife was formerly part of two Strategic Development Plan areas and contributed to separate Housing Need and Demand Assessments.

90. Annex D provides an overview of the proposed MATHLR for each local and national park authority. Within the table, figures shown for Eilean Siar are not rounded. This is because of the effect rounding to the nearest 50 can have on numbers at this scale. Figures for the Glasgow City Region are provided for the MATHLR only as breakdowns were not received.

91. Annex E provides a breakdown of the Initial Default Estimates, Locally Adjusted Estimates and the proposed MATHLR by the three inputs. It illustrates the progression to the figures in Draft NPF4. As above, within the table, figures shown for Eilean Siar are not rounded because of the effect rounding can have on numbers at this scale. Also as above, figures for the Glasgow City Region MATHLR breakdown are not included as they were not received.