You are Marcela Irías, chief financial officer of Hidroeléctrica Montecierzo, a 15 MW hydroelectric project in western Honduras. A pension fund holding 30% of the equity has asked the question your single-number model cannot answer: how likely is the project to lose money? Before you run a single scenario, you must decide how wide to draw the range of each uncertain input.
The situation
Tegucigalpa, Tuesday 12 May 2026. At the shareholders' meeting Elske van Dijkhuizen, who represents the European pension fund, interrupts you: "What is the probability that this project loses money? And what is the worst reasonable case?" You answer with what you have: a net present value of US$12.4 million at an 8.2% cost of capital over 15 years. "That is a number," she replies. The board gives you until 30 June, the day of the financial-close vote.
Your model assumes the plant produces on average 65% of its capacity, that construction costs US$45 million and takes 24 months, and that running it costs US$1.2 million a year. All output is sold to Eléctrica Lencala, the state distributor, which has a record of paying late, under a 15-year contract at US$95 per MWh rising 2% a year. The project is financed 70% with debt and 30% with equity.
None of those figures is certain. The hydrologist, Saúl Andino, has 30 years of rainfall: in the driest years the plant would have run near 40% of capacity, in the wettest near 85%. The cost engineer thinks construction could come in 5% cheaper or up to 35% dearer if caverns appear under the dam or the rules change. Legal gives 85% to Lencala honouring the contract, 10% to a partial default, and 5% to it stopping payment, leaving the plant to sell on the spot market.
The general manager, Fausto Rivera, twenty years in hydro plants, sees no urgency: "The project makes 12.4 million. Every insurance policy we buy comes out of that number." Van Dijkhuizen wrote to you that afternoon: "The average is no use to our pensioners if the project cannot pay its debt."