The HUD financial model is a submission instrument. It tells HUD what you are proposing. It does not tell the organisation whether the proposal is viable. The Strategic Analysis identified this gap as the centrepiece risk: the absence of a pre-submission feasibility platform that stress-tests assumptions before commitment. This simulation engine is that platform.

The three layers operate sequentially. Multi-Criteria Analysis screens locations before any financial modelling, preventing capital allocation to locations that cannot be defended on non-price grounds. The deterministic model produces per-location financial outputs in HUD’s prescribed format. Monte Carlo provides probability distributions across 10,000 iterations, replacing single-point estimates with confidence intervals.

Input assumptions — synthetic vs actual

Every figure in this simulation is based on synthetic assumptions derived from public data and the HUD application pack. The following table shows what is synthetic and what actual data is needed to convert the simulation to submission-ready output.

HUD scores non-price criteria at 80% of total marks before the cost envelope is opened. An organisation that builds a competitive financial model for a location it cannot defend on non-price grounds has wasted its modelling effort. The Multi-Criteria Analysis imposes screening discipline: no location enters the financial model without passing a weighted assessment against six criteria mapped to HUD’s own assessment priorities.

South Auckland carries half the programme and is the location where capability evidence is strongest. The financial model for South Auckland must be the strongest in the submission. Hamilton is the expansion play — strong on strategic alignment, weaker on delivery capability because there is no current housing stock. Christchurch is the consolidation location — financially competitive but smaller in scale.

The Strategic Analysis identified the $46,000 per-place benchmark as the number against which every submission will be evaluated. The simulation confirms and quantifies the challenge. At synthetic development cost assumptions, every location exceeds benchmark. The programme weighted average is $62,019 — 35% above.

The benchmark gap is almost entirely driven by one variable: development cost. Debt servicing alone — the PMT of development cost, equity rate, and interest rate — consumes $30,600 to $34,979 per unit per year. That is 67–76% of the entire $46,000 benchmark before a single dollar of operating cost or contingency is added. The path to benchmark is through development cost, not through operational efficiency. Operating cost reductions, even aggressive ones, cannot close a gap driven by debt servicing. An applicant’s strongest competitive lever is existing programme cost data — actual acquisition costs that may be materially lower than the synthetic estimates used here.

The Monte Carlo analysis runs 10,000 iterations with six variables sampled simultaneously: interest rate, construction cost, operating costs, R&M escalation, vacancy rate, and index lag drift. It is not a stress test. It is a probability map — asking not “what happens in the worst case” but “across 10,000 plausible scenarios, what is the range of outcomes and how likely is each?”

The contingency governance lever

This is the finding that could only emerge from running the simulation. At Medium (50%), the HUD contingency formula is self-sustaining across all 10,000 scenarios — zero contingency exhaustion probability. But contingency adds $15,000–$17,500 per unit to the Year 1 cost. At Low (25%), approximately $8,700 per unit per year is saved across the programme, taking the weighted average from $62,019 to approximately $53,300. Still above benchmark, but materially closer.

The trade-off is real. Lower contingency means less buffer against the exposures documented in the Strategic Analysis: index lag, legislative change, and capital replacements from Year 10. The Monte Carlo shows zero exhaustion at Medium, meaning the current setting over-provisions relative to modelled risks. Whether that over-provision is prudent insurance or unnecessary cost is a risk appetite question, not a financial modelling question.

The equity available for the Flexible Fund is residual — what remains after Budget 2025 commitments are absorbed. The central estimate of $6 million is derived from public financial statements and accounts for prior capital commitments. The CFO’s confirmed figure replaces all modelled scenarios.

The central case falls 4 units short. At 15% equity and $6M available, the programme supports 76 of 80 units. The levers are: reduce equity rate to 10% (supports 115 units at central), reduce programme scope to 76, increase equity, or reduce development cost. The optimal rate depends on the actual capital available and the organisation’s debt capacity — both CFO inputs.

The Strategic Analysis devoted an entire section to this risk. The simulation now attaches programme-scale numbers to the gap. At 80 homes, annual wraparound cost ranges from $552,000 to $840,000 depending on service intensity. At 2% CPI indexation, the cumulative unfunded exposure over 25 years is $17.7M–$26.9M.

The board must resolve — before signing — how this annual gap will be sourced for 25 years. Whether through dedicated MSD contract negotiation, endowment funding, philanthropic partnerships, or a deliberate programme scaling strategy, this is a governance decision. Organisations that describe aspirational wraparound capability without a funding architecture will be identified as such by the evaluation panel.

The gap between this synthetic illustration and a submission-ready package is four inputs from the organisation: aspirational locations, total capital envelope, Budget 2025 capital position, and existing programme cost data. The simulation engine re-runs with actual data in under an hour.

The simulation engine produces outputs that map directly to every section of HUD’s prescribed Cost Response Form: equity structure (Section 1.1), programme finance (Section 1.2), Year 1 cost per place (Section 2.1a), cost accuracy evidence (Section 2.1b), contingency justification (Section 2.1d), and interest rate assumptions (Section 2.1e). The engine does not produce analysis for the sake of analysis — it produces the specific outputs that populate the specific fields in the specific forms HUD requires.

The structural findings will not change when actual data replaces the synthetic inputs: the benchmark gap is driven by development cost, the contingency level is a governance lever, the capital allocation is tight at central equity, the wraparound gap is structural and unfunded, and the path from simulation to submission runs through four conversations that can happen this week.