How a proposed new model would use guarantees to reprofile debt costs and make grant go further

Social Finance proposes that homes acquired into the scheme would be forward purchases from the existing development pipelines of housing associations, but suggests that purchases could also target other new build homes or existing stock.

In response to the financial constraints on housing associations, and citing the fact that in 2024 the cost of servicing debt for private registered providers exceeded net earnings for the first time in 15 years, the model is designed to sit off-balance-sheet.

However, while AHAS would raise 70 per cent of debt, the remaining 30 per cent is expected to be provided “in the form of equity largely from participating housing associations and potentially partially funded by capital grants”.

Modelling in the report assumes that this would comprise 10 per cent equity and 20 per cent grant.

The publication also states that “initial feedback is that Asset Co would not be required to be a registered provider”.

‘Step up’ structure

Repayments over the bond or loan term would have a ‘step up’ structure, meaning that these are higher towards the end of the proposed 30-year term and that the terminal payment is much higher than for a conventional bond.

As an example, the report compares an illustrative step-up bond with a conventional bond. Both concern a principal amount of £100m over a 30-year term, and deliver an investor annualised return, or internal rate of return, of six per cent.

The conventional bond might see debt servicing costs of six per cent in each year and a final repayment of the £100m principal. By contrast, the step-up structure would see lower initial debt re-servicing costs of 1.75 per cent per year, increasing by 0.5 per cent every five years, and would culminate in a terminal payment of £365m.

Taking the net rent assumptions of the properties delivered over the period, the report’s modelling suggests that the AHAS is able to generate a cash surplus from inception using just 20 per cent of grant, whereas a conventional bond is not.

Risk to guarantor

However, the debt-to-assets ratio would be much higher at the bond term (70 per cent), compared with homes delivered through the Affordable Homes Programme (10 per cent), albeit while using a smaller proportion of grant (20 per cent vs 50 per cent). The report therefore notes that the AHAS would be “much more exposed to residual asset values than a typical AHP model”.

This would also mean exposing the government as guarantor to “much greater risk” than its current guarantee scheme – the Affordable Homes Guarantee Scheme – where loans are backed by registered provider covenants with additional security from charged assets.

The report also acknowledges a need to determine how much of the scheme obligation would be classified as on the government’s balance sheet for its own accounting purposes, with regard to public sector net financial liabilities. But it believes that the government guarantee underpinning the bonds issued would “remain a contingent liability with only a small portion of the guaranteed debt value falling on the government balance sheet based, as is likely, on a default risk assessment.”

Developing the model

Speaking to Social Housing, Jonathan Flory, co-author of the report and a senior advisor at Social Finance, said the consultancy had been through “several chapters” of the project to develop this model, starting with work with the government and a group of housing and investment stakeholders to consider “how guarantees can work a bit harder in the system”. More recent work has been to refresh that earlier research, with support from the Joseph Rowntree Foundation.

Early iterations of the model involved looking at index-linked financing, Will Damazer, report co-author and associate director at Social Finance, says. However, at current interest rates and pricing, index-linked funding looks less attractive and so the step-up bond has been proposed to provide some of the benefits while not adding to the stock of index-linked securities, Mr Damazer adds.

Asked whether housing associations might find it challenging to free up equity capital of their own to invest in the scheme, Mr Flory refers to a “useful precedent” in the form of the Welsh Housing Partnership, set up in 2011 and involving four registered social landlords. The initiative has seen the housing associations invest equity, including capital grant from the Welsh government, into a joint venture alongside private debt finance to develop homes. These are then leased back to the housing associations to let to tenants at sub-market rents.

Mr Flory adds: “So part of the answer is there, which is that you could set up a structure where the housing associations have some skin in the game and they have to put a bit of the equity in themselves, but potentially a large part of the equity could come from reprocessing capital grant, so I think the model is flexible.”

Social Housing also asked about the house price inflation elements baked into the model and whether there is an expectation that homes would need to be sold at market value at the end of the scheme.

The report refers to an assumption that the AHAS would rely on “house price growth and refinancing or, if needed, portfolio sales to pay down the scheme debt raised.”

It notes: “In our central scenario, house price growth is required to be c.2.9% per annum. Historically the minimum annualised house price growth over a 30-year period is 5.9%. Accordingly, we believe that the AHAS structure offers [the government] as guarantor an acceptable risk of low default and meanwhile offers a much more efficient use of capital grant.”

Mr Damazer tells Social Housing: “Regardless of whether it comes in as EUV-SH or market value, the assumption is essentially that how it exits is how it comes in. So we might be able to get a discount on schemes as we purchase, and invested homes will grow at that inflation rate over the period of the scheme. Or if we buy at market [value], we can value it at market at the end of the scheme,” Mr Damazer says. 

He adds: “Over a 30-year period there could be scope to convert a portion of the homes if needed to market at the end of tenancies. This is one area where further detailed work is required.” 

Mr Flory and Mr Damazer are both keen to emphasise that they see the proposed model as providing a part of the funding mix that could be available to the sector, to free up capacity to develop, rather than a replacement for grant or existing guarantee schemes.

The £500m pilot proposed (or £2bn scaled-up model) represents the entire value of the scheme, of which 70 per cent is the guaranteed debt. This could fund around 2,500 to 3,000 homes, the authors believe.

Next steps for the research project would include establishing a “cross-governmental working group” and engaging stakeholders to identify and address practical challenges, the paper sets out.