Footnotes
1. Wilcox, S. (2009) UK Housing Review 2009/10; Coventry and London: Chartered Institute of Housing and Building Societies Association - Table 19e
2. Ibid - Table 56
3. Ibid - Table 17a
4. Local authority figures from Scottish Government website 'Housing Statistics for Scotland'; RSL figures from SCORE annual reports
5. Scottish Government website 'Housing Statistics for Scotland'
6. http://www.scotland.gov.uk/Topics/Built-Environment/Housing/investment/affordable
7. http://www.scotland.gov.uk/Topics/Built-Environment/Housing/investment/affordable/casestudy
8. Limiting this exercise to medium-sized organisations was partly a logistical decision following from the relatively limited budget and short timescale for the project. In any case, in terms of dwellings in ownership, RSLs with less than 1,000 homes now account for only 17% of the sector, and a disproportionate number of these organisations operate in areas not estimated as in need of additional affordable housing (see next footnote).
9. For example, as calibrated in: Bramley, G. et al (2006) Local Housing Need and Affordability Model for Scotland - Update (2005-based)http://www.scotland.gov.uk/Resource/Doc/1035/0085202.pdf
10. Housing Corporation (2007) Unlocking the Door: Delivering More Affordable Homes from the Comprehensive Spending Review 2007; London: Housing Corporation
11. In considering financing comparisons between local authorities and RSLs it also needs to be noted that unit development costs for the grant-funded local authority programme have been somewhat lower than the typical figure for housing associations - around £120K and £130K, respectively. Underlying factors here could include contrasts in property size mix between the two sectors, as well as the possibility that municipal schemes are more likely to involve council-owned land which can - in some cases - be contributed at nil cost.
12. Tenant Services Authority (2010) 2009 Global Accounts of Housing Associations; London: TSAhttp://www.tenantservicesauthority.org/upload/pdf/Global_accounts_2009_20100324102409.pdf
13. While concerns about 'ratepayers' subsidising council tenants were quite widely voiced in the 1980s, little has been heard of such a possibility more latterly.
14. It should be noted that, in line with advice from the Scottish Government, the researchers did not actively seek research participants' views on this issue. Hence, the perspectives of other stakeholders on this matter were not explored.
15. Note that interest cover as a 'financial stress ratio' is discussed in more detail in Chapter 4.
16. Spilsbury, D. & Polhammer, M. (2010) Council as lenders? in: Chartered Institute of Housing; Investing in Affordable Housing - A Radical Rethink?
http://www.cih.org/scotland/policy/InvestAffHousingDec09.pdf
17. See p17 in: Gibb, K. & Neary, K. (2010) New Climate, New Challenges; Chartered Institute of Housing in Scotland http://www.cih.org/scotland/policy/New-Climate-Challenges-Mar10.pdf
18. It should be noted, though, that accounting rules require the Council to treat its debt principal repayment to the PWLB as a revenue charge to the General Fund, while the debt principal repayment from ELHA to the Council counts as a receipt for the Capital Fund, which can in turn only be used to fund capital, and not revenue, expenditure. In this sense, the arrangement implies some restriction on the purposes for which the Council can use its funds, although not on the overall amount of funds available to the Council.
19. Scottish Futures Trust (2009) Increasing Affordable Housing Supply from Limited Public Resources - The Proposed National Housing Trust Initiativehttp://www.scotland.gov.uk/Topics/Built-Environment/Housing/supply-demand/Proposal/Q/EditMode/on/ForceUpdate/on
20. It should be noted that this fund is not, in fact, owned, controlled or managed by the SFHA.
21. Johnson, D. (2009) Asset Realisation: Is there really any other choice? In Chartered Institute of Housing: Investing in affordable housing: a radical rethink?
http://www.cih.org/scotland/policy/InvestAffHousingDec09.pdf
22. Twinch, E. (2010) Highland Council to sell off a fifth of its homes; Inside Housing 6 August http://www.insidehousing.co.uk/news/housing-management/highland-to-sell-off-a-fifth-of-its-homes/6511044.article
23. Homes and Communities Agency (2009) Local Authorities Bidding for Social Housing Grant; http://www.homesandcommunities.co.uk/public/documents/LA_social_housing_grant.pdf
24. Scottish Executive (2005) Planning Advice Note 74: Affordable Housing, Edinburgh: Scottish Executive http://www.scotland.gov.uk/Publications/2005/03/20796/54073
25. Scottish Government (2009) Affordable Housing Securing Planning Consent 2008/09; http://www.scotland.gov.uk/Publications/2009/09/22145056/0
26. Newhaven Research (2008) All Pain, No Gain? Finding the Balance; Delivering Affordable Housing Through the Planning System in Scotland; Edinburgh: Chartered Institute of Housing in Scotland http://www.cih.org/scotland/policy/All-Pain-no-gain.pdf
27. Given Scotland's unified tenancy regime across social housing, such considerations would not, of course, apply north of the border.
28. These projections are at outturn prices, although assumptions and some results are quoted in real terms.
29. The LAs in the system are the 26 Scottish unitary authorities which have not undertaken whole stock transfer, i.e. excluding Argyll & Bute, Dumfries & Galloway, Eilean Siar, Glasgow, Inverclyde, and Scottish Borders.
30. However, our decision to omit any mechanism to 'damp' the effects of 'excessive debt' is informed by analysis confirming that loan payments as a percentage of total HRA expenditure do not appear problematic. While, on the basis of central assumptions, three authorities have loan payments exceeding 40% of total expenditure, two of these are considered to have 'zero need' (see below) and the third never generates a surplus to let it build, anyway.
31. The treatment of SHQS differs between RSLs and LAs, and between the two groups of RSLs discussed in Chapter 4, because of differences in available data and also the different typical situation of RSLs in terms of type and age of stock.
32. SHQS cost captures all major repairs capital costs. Post-2015, it is the capital cost of maintaining stock in line with SHQS.
33. The first year in which the modelled assumptions apply - previous years use the local authorities' own plans, as described above.
34. Assumptions here incorporate 2.5% for capital repayments for new build borrowing as a central assumption - which is cautious compared to the 1.66% (60-year) central assumptions for the balance of HRA debt.
35. It is also assumed that new council units add the same amount to M&M costs as the average value of these costs across the existing stock. This might be seen as a pessimistic assumption, although it balances the perhaps optimistic central assumption of a zero real terms increase in management.
36. And, in the distribution of AHIP funding to RSLs this is partly reflected in the distinction between the 'core programme' and 'topsliced' elements of the budget - see Chapter 4.
37. Using best fit of each local authority to one and only one region e.g. Fife is linked to SESPlan
38. It should, however, be noted that even under the 'high rent increase' scenario, it is highly unlikely that local authority rents (even in Edinburgh) would exceed the 30 th percentile of LHA rates for comparably sized properties.
39. In the RSL model reported in Chapter 4 we also include options to raise rents by more than the normal RPI+1 over an initial five year period, say by 10%, then revert to a stable trend increase. This has not been tested on the local authority model as yet. The rough equivalent to this for the local authority sector is the 'partial convergence' scenario as outlined above.
40. http://www.scotland.gov.uk/Publications/2004/06/19127/34802
41. However, the composite scenarios for local authorities do not precisely match those for RSLs due to the different characteristics of the local authority and RSL models and starting points. The local authority model does not include any initial uplift for all rents - just a higher rent for new build; nor does it impose a HAG envelope.
42. We have somewhat more confidence in the 30 group because we can compare our modelled projections with theirs, and in relation to some specific issues like SHQS we can incorporate their own plans for addressing this. But the analysis for the 45 is broadly comparable and we believe that the aggregate or average picture is realistic, even if this might not be true of the detailed figures for all of the individual RSLs within this group.
43. While some key stakeholders suggested that the ratios used here should replicate thresholds officially prescribed by the Scottish Housing Regulator, it is understood that no such 'official thresholds', in fact, exist. Rather, in its engagement with individual RSLs, the SHR refers to the relevant lender covenants, which will vary between RSLs and lenders and to which we have no access. We believe the chosen thresholds are representative of typical values. We report debt per unit values (averages) and these could be examined for individual RSLs. Across our 30 HAs debt per unit figures in their forward forecasts averaged £15k and showed a maximum value of £41k; in our baseline projection to 2035 both the average and maximum values remain below these levels in real terms. However, consultation with the SHR confirmed that the key constraint on borrowing capacity is the ability of the income and expenditure account to generate surpluses, rather than balance sheet numbers. Higher 'debt per house' levels can be sustained by landlords with higher rents and values alongside moderate operating costs. As reported in a footnote elsewhere, we also tested a rolling LSVT-type valuation on the 30 RSLs in that group and showed that, for all bar two of these, debt security was not a problem. It should be noted that we apply two threshold levels to allow progressive restraint to be applied. We have also tested sensitivity to threshold values and it appears that output capacity is not very sensitive to these values.
44. In some instances negative reserves appear to be generated from the data and the model. Where this is the case it is assumed that this performance criterion would be flagged as 'red' (and thus the model sets new build to zero).
45. This is informed by observed typical ratios according to the figures reported by the 30 RSLs providing business forecasts. Of this group in 2010, three were in 50-60% range, two were in 60-80% range, and three were above 80%.
46. While previous research for the Scottish Government (referred to above, see footnote 40) found that capital costs rise only in line with inflation, we are sceptical of this suggestion. Firstly, construction is characterised by relatively sluggish technical progress, which means that building at a given spec is likely to be associated with rising real cost as real wages rise. Secondly, recent data on housebuilding costs may be distorted by the tendency for dwellings to get smaller etc. And, thirdly, the previous research was for the purposes of stock transfer, and thus focused on the refurbishment and maintenance of social housing. In contrast, since the modelling in this study relates to new build, procurement costs in the central assumption include an element of land value, which is driven by rising real house prices in the long run, as documented in previous Scottish Government analyses. Nevertheless, later in this chapter we include a sensitivity test for zero construction cost increase.
47. Business plan forecasts for 30 RSLs show borrowing costs rising from around 4.75% in 2009 to 5.32% in 2013, against which the central assumption of 5.5% may be compared.
48. 'Target maximum cash & investments % of turnover' sets a figure for the total of these two items; if the amount exceeds this, 70% of the excess is directed towards defraying borrowing costs for new development (if required), subject to there not being a liquidity shortfall.
49. 5% is effectively the standard RPI+1. The rent shift is about a positive upward rent shift, so in this context 15% is low and 30% is high.
50. 'Profile growth % p.a.' is the amount by which the amount of new build increases each year for those RSLs which are not constrained by low financial performance scores.
51. Scottish Government (2010) Affordable Housing Investment Programme 2010/11 http://www.scotland.gov.uk/Topics/Built-Environment/Housing/investment/ahip/ahip2010-2011
52. Items subject to convergence in the period 2011-2015 include new build unit costs, grant rate and borrowing cost (rate of interest). Certain other parameters are taken from base period values but then constrained to lie within a range, including rent loss on voids, depreciation and short-term creditors.
53. The baseline allows continued development at past levels in low/no need areas. Excluding development in such areas is a variant option. GHA, Glasgow's largest association is a special case, because the unusual structure of its accounts mean the application of the conventional model does not allow it to develop on financial grounds. Therefore, any extra new provision it does has to be fully funded by the Scottish Government and any private finance obtainable on those units. Therefore, it would make more sense to account for its provision as part of any top slice. The numbers involved could then be added on to the numbers we model for headline totals. In our references to 'topslice' funding it is also recognised that this relates only to specific reprovisioning investment in Glasgow and not to the entirety of social housing new build funding in the City.
54. This raises questions about the relationship between debt and the valuation of the stock, and whether this would act as a further constraint on borrowing for some. For the 30 RSL set, we have explored a rolling valuation estimate based on projected financial flows without new investment, for the 30 RSL set. This suggests that borrowing security against valuation issues would be a significant constraint for only two of these organisations.
55. This is the figure advised by the Scottish Government and is also consistent with the actual 2011 grant per unit typically assumed in RSL business plan forecasts as provided to the research team.
56. Including grossed-up allowance for smaller RSLs sampled on one-in-five basis.
57. Average performance score is the average for the whole group of RSLs of the scores on five financial performance criteria where 2 is above the upper threshold, 1 is between the two thresholds and 0 is below the lower threshold. Scores of 9 or above enable RSLs to increase output, scores of 5-8 enable continued output, scores below 5 prevent any development.
58. Grant per unit is the actual average in real terms. It starts higher because of convergence, and then rises later because of assumed 1% p.a. rise in real procurement costs.
59. For this purpose RSLs are assigned to the local authority where the largest share of their stock is located.
60. The need estimate is based on Bramley et al (2006) affordability-based housing needs model for Scotland, 2007 update undertaken for Scottish Government as reported at 2008 Firm Analytical Foundations conference. The need number is the average annual positive need projected 2006-21 under assumption of no price correction, expressed as a percentage of resident households, mean value 0.49. Banded values are 0=0.0%; 1=0.01-0.24%; 2=0.25-0.49%; 3=0.50% and over.
61. Glasgow HA is included in the tables but the model does not show it as having capacity on conventional measures. Its funding structure is unique and its ability to undertake new development is effectively conditional on resources from central government and/or the prioritisation of this activity over other activities. Glasgow is also in an area shown as having zero need for additional affordable housing. A substantial proportion of new build activity in Glasgow, and certain other areas (e.g. Inverclyde) is aimed at selective re-provision to replace some demolished stock in regeneration areas, and this is funded from a top-sliced part of the budget (over and above the current £290M 'core programme').
62. The group of 30 included several providers specialising in older people and one other special needs specialist. When sampling for the 45 we did not include further examples of this kind, to keep the sample more homogeneous. It may therefore be that our estimate of capacity within this sub-sector of RSLs is low, but it must be borne in mind that any additional development by RSLs omitted from this group would need to be funded by grant, either additional to the envelope assumed or at the expense of some of the general need providers.
63. Changing one assumption at a time is simple to grasp, but some caution is needed in interpreting these results. In most/many cases, it is necessary to change the grant rate to balance out the grant claims with the available grant envelope. This balancing is achieved only in the medium term, not necessarily in a particular short period.
64. The 30 RSLs providing business plan forecasts were assuming an average annual increase in management costs of only 2.4% in money terms between 2009 and 2013, below our assumed rate of inflation.
65. That is, average annual new build output 2008-10 (30 group) or estimated new units 2009 derived from HAG allocations (45 group).
66. For the 30 group it is assumed that expenditure within the projections is sufficient to meet the SHQS target. The relevant expenditure may be in planned maintenance or capitalised maintenance. Only the figures for capitalised maintenance are 'stepped down' in 2016 in the baseline, because we do not know what proportion of planned maintenance relates to SHQS backlog. The effect is that the difference under this scenario is less than the difference with the 45 group.
67. This latter figure comprises 3,519 RSL core programme approvals and 1,342 local authority grant funding approvals.
68. The RSL model applies grant rates similar to previous actuals initially but converges these on a common rate by year 5. This can mean that initial grant costs are different from (higher than) the average, especially if there is a positive correlation between initial development rates and grant rates. In addition, although we try to reconcile HAG envelope and cumulative spending in the medium term, this is not necessarily achieved in a particular year or short period of years.
69. The headline output for RSLs is over the period 2011-35, for LAs it is for 2014-35. Grant total is at 2010 prices.
70. It should be noted that the modelling of local authority output and grant has to be modified for grant rates above 60%. Here it is assumed that the increase in output is pro rata the increase in the grant rate, and that the cost increase comprises this same pro rata element to the previous grant level plus a grant cost for new units at the specified percentage of £120k.
71. Newhaven Research (2008) All Pain, No Gain? Finding the Balance; Delivering Affordable Housing Through the Planning System in Scotland; Edinburgh: Chartered Institute of Housing in Scotland http://www.cih.org/scotland/policy/All-Pain-no-gain.pdf
72. Crook, A.D.H. and Whitehead, C.M.E. (2002) Social Housing and Planning Gain: Is this an appropriate way of providing affordable housing?, Environment & Planning A, 34, 1259-1279
Crook, A.D.H. et al (2010) The Incidence, Value and Delivery of Planning Obligations in England in 2007/08; London: Communities and Local Government
http://www.communities.gov.uk/documents/planningandbuilding/pdf/1517816.pdf
73. Intermediate provision may often be viable on the basis of zero or low land value without additional subsidy, and this can make agreements easier to reach when availability of public subsidy is uncertain, although this may also be more vulnerable to market downturns.
74. This is consistent with the fact that 2010/11 Government spending on council housebuilding will be minimal because, while £80M in funding approvals for municipal development has been approved since 2009, associated grant payments fall due only when houses are completed. In practice, it is anticipated that only a small proportion of the 3,200 homes so far approved under the programme will be handed over in 2010/11. Hence, the impact of council housebuilding commitments on the Scottish Government budget will be felt mainly in future years. It should also be noted that these estimates are based on sector-specific grant rates (25% for local authorities and 50% for RSLs) rather than a common regime.
75. It should, however, be emphasised that the £270M RSL funding envelope used in our projections includes only the core programme and not the 'topsliced' elements of the Scottish Government's Affordable Housing Investment Programme. These include social rented housing being developed in the course of reprovisioning former council stock in Glasgow and elsewhere. In 2009/10, social rented approvals funded in this way totalled some 700 homes at a HAG cost of around £50M (assuming HAG per unit at £75K).
76. Tickell, J, (2010) Call for change; Inside Housing 10 September
http://www.insidehousing.co.uk/analysis/opinion/call-for-change/6511567.article
77. A measure you can't trust; Inside Housing 5 Feb 2010 http://www.insidehousing.co.uk/analysis/in-depth/a-measure-you-can't-trust/6508428.article
78. Pawson, H. & Jacobs, K. (2010) Policy intervention and its impact: New Labour's public service reform model as applied to local authority housing; Housing, Theory and Society Vol 76(1) pp76-94
79. In Audit Scotland's 2008/09 performance indicators 10 of the 26 landlord local authorities reported that at least 80% of their homes remained non-compliant with SHQS as at April 2009 http://www.audit-scotland.gov.uk/performance/service/
80. Bramley, G. et al (2006) Local Housing Need and Affordability Model for Scotland - Update (2005-based)http://www.scotland.gov.uk/Resource/Doc/1035/0085202.pdf