§7.8Downside Mode & Sunset Economics

If the campaign stalls at the first close ($12M): both Lighthouses keep running and Tekka waits. The plan can run indefinitely on the two Lighthouses alone. Their Year 3 spend, about $5.6M ($5,624,989), covers thirteen Master Fellowships, 23 relief cooks at Chinatown Complex and 6 at Geylang Serai, and the Regulars' List, within reach of the draw, the operating match and a single foundation partner: from Year 4, at about $3.38M a year, no year runs short at the central assumptions. Over the four years the two cost $16,224,542. A centre keeps what it was already given: the Package is given, not traded. Nothing in the architecture requires scale to be honest; scale only makes it consequential.

From Year 4, on the heritage logic: the ledgers used to keep each centre's Year 3 budget for ever, so Chinatown Complex would have funded ten new Fellowships a year long after its 27 heritage-worthy stalls had each been offered one. That made no sense, and I have changed it. Once a centre's planning count has been offered – by the end of Year 3 at all three – its Fellowships continue at a maintenance rate fed by the yearly at-risk count, which comes from the signals hawkers give voluntarily and from public facts (§3.6). Assumption: about 10% of the planning count a year, at least one, until the count gives real numbers: three new Fellowships a year at Chinatown Complex and one each at Geylang Serai and Tekka. The cover, the full pools of relief cooks, the centre staff and the national core, the Regulars' List with it, stay at their Year 3 level. From Year 4 the programme at three centres then costs about $4.13M a year in Year 3 prices, indexed, against about $5.32M from the draw, the operating match and continuing gifts. At the central assumptions no year runs short, and what is left of the gifts joins the corpus (the Budget Book, B.9).

To be sure, the rate is the assumption that matters most. It sits below the 15–25% of listed stalls the plan expects the count to find at risk (§3.6), and the two measure different things. The count is a snapshot: the listed stalls whose hawkers say that no successor is in view and that they mean to stop within three years, or that a public fact flags. Spread over those three years, 15–25% means about 5–8% of listed stalls reaching that point each year (derived). The 10% is a share of each centre's planning count, offered a new Fellowship each year, and not every stall counted will want one: a Fellowship goes only to a hawker who wants to teach a dish the Registry finds fragile, and many were offered one in Years 1–3 already. Still, the yearly rate could be faster than 10%. With the higher wages, the margin is tighter than it was. At 20% a year the three centres run short in just one year up to Year 10, by about $0.06M, and by about $5.8M over Years 4–20, and the corpus ends Year 20 at about 87% of its target in today's dollars; at 30%, thirteen new Fellowships a year across the three, they run about $6.9M short over Years 4–10; at 40%, about $16.0M (the Budget Book, B.9.4). So the yearly count decides. If the count finds more stalls at risk than the plan assumes, a larger endowment, a second campaign or partners who adopt a pilot centre would each pay for more Fellowships, and the choice is better made on the count and the evaluator's first reports than ahead of them. Until then, the figure stays printed.

If the sunset triggers: the Seed Bank returns unspent donations to the donors, as each gift's terms require. The endowment's capital, held for hawker-culture purposes, goes to another charitable purpose agreed with the Commissioner of Charities, the Seed Bank board and the Hawker Council. And the archive – every Genome file, oral history and reference taste – passes to the National Heritage Board under the same non-exclusive licences, with each master's name kept on his work. Even in failure, the knowledge stays saved and stays open. That is the minimum promise of a seed bank.