Appraisal & Viability 08 min read

Sensitivity analysis in housing development

A development appraisal shows a scheme's viability at one point in time, on one set of assumptions about build costs, sales values, programme, interest rates and grant. Some of those assumptions will turn out wrong. Sensitivity analysis tests what happens to viability when they change, so the risk in a scheme is understood before the organisation commits to it.

For housing associations it serves three purposes: managing risk internally by showing what could go wrong and how much it would matter, supporting board decisions by showing that risk has been identified and bounded, and feeding the wider stress testing that the Regulator of Social Housing expects of private registered providers. SDS makes Proval Vantage, and we refer to it where it illustrates a point.

SDS · Updated October 2026

What is tested in a sensitivity analysis

The variables below are those most often tested in social housing development appraisals. The size of each test is a matter for the organisation's risk appetite; the percentages here are illustrations, not a standard.

Build costs

Construction is usually the largest cost in a residential appraisal, and build cost inflation driven by materials prices, labour supply or supply chain disruption can move quickly. Testing increases above the base case, for example of 5% and 10%, shows how much cost headroom a scheme has.

Sales values

Where a scheme includes open market sale or shared ownership, lower sales values reduce income and therefore viability. Tests typically model falls of several percentage points, with larger movements for schemes in more volatile markets.

Interest rates

Finance costs rise with interest rates, and the effect grows with the size of the scheme and the length of the development period. For rented homes, the rate also affects whether long-term net rents can repay the debt, which is central to housing association appraisal (National Housing Federation, Understanding Development Appraisal, 2015).

Programme delay

A delay of six or twelve months extends the period over which interest accrues, can push sales into a different market, and can affect when grant is paid. Homes England pays grant against scheme milestones, so slippage can move grant receipts as well as costs (Capital Funding Guide, chapter 6).

Grant

For grant-funded schemes, test what happens with less grant than applied for, or none. Homes England treats grant as the minimum balancing sum after the applicant's own contributions have been maximised, and assesses value for money mainly on grant per home against benchmarks (Capital Funding Guide, chapter 6). Knowing how viability responds to grant helps set a realistic bid.

Rents and running costs

For homes the association will own and let, long-term assumptions on rent increases, voids, bad debts, management and maintenance costs, and the discount rate drive the net present value. An NHF guide shows NPV can be highly sensitive to the discount rate chosen.

How to present sensitivity analysis

Sensitivity analysis is most useful when decision-makers can see the range of outcomes at a glance and identify which assumptions the scheme depends on most. A clear sensitivity table shows the base case result, typically NPV for rented homes or residual land value, beside the result under each scenario, with the change from the base case shown alongside.

Three further techniques help:

Break-even tests show how far a variable can move before the scheme stops being viable, for example the build cost increase at which NPV reaches zero.
Combined scenarios move several variables together, such as higher costs and lower sales values at the same time, which reflects how risks tend to arrive in practice.
Ranking variables by their impact directs attention to the assumptions that most need evidence and monitoring.

Proval lets teams stress-test a scheme against changes in grant, build cost and sales price and see where the viability threshold sits. It includes sensitivity analysis among its outputs alongside NPV, IRR and payback (G-Cloud 14 listing).

Scheme-level and organisation-wide testing

Sensitivity analysis works at two levels.

Scheme-level testing focuses on the assumptions that matter most for a particular scheme: its build cost benchmark, the sales values in that market, the contractor's programme.

Organisation-wide testing applies the same scenarios across every scheme, such as the same build cost increase or interest rate movement. Development and finance directors then see the combined exposure of the programme, and schemes can be compared on a like-for-like basis. Proval lets administrators set development assumptions and defaults centrally, so every appraisal starts from the same basis.

Connection to business plan stress testing

Scheme sensitivity analysis also feeds the organisation's own stress testing. The Regulator of Social Housing's Governance and Financial Viability Standard requires private registered providers to carry out detailed and robust stress testing against identified risks and combinations of risks across a range of scenarios, and to put mitigation strategies in place (Governance and Financial Viability Standard). The regulator's code of practice treats stress testing as broader than simple sensitivity testing, including multi-variate analysis across financial, economic and operational risks (Anthony Collins Solicitors). Consistent scheme-level results make that organisation-level work easier and more credible.

Sensitivity analysis and board governance

A development committee needs to know whether a scheme is viable on the base case and whether it stays viable if things go moderately wrong. A common format sets out:

the base case;
a downside scenario, with a modest adverse movement in the two or three variables the scheme is most sensitive to, applied together;
a stress scenario, with larger adverse movements;
the break-even point for the key variables.

The exact scenarios and thresholds should come from the organisation's own risk appetite and be applied consistently across schemes, so committee members can compare like with like.

Proval produces consistent, board-ready outputs for governance sign-off (SDS), and SDS states that it stores appraisals centrally, with the assumptions, defaults and overrides behind each one held in one place. That avoids rebuilding the appraisal by hand for each scenario and gives the board a traceable basis for the figures in front of it.