In years 1 and 2, the product will be sold only in the United States, but starting in year 3, Lilly might sell the product overseas. The year 1 market size in the United States is assumed to be between 500,000 and 3,000,000 units. A market size of 1,000,000 units is assumed to be twice as likely as a market size of 700,000, and a market size of 2,000,000 units is assumed to be three times as likely as a market size of 700,000.

■ Lilly’s year 1 market share is assumed to follow a triangular distribution with worst case 10%, most likely case 20%, and best case 30%. Lilly assumes that its market share will remain the same unless a competitor enters the market.

■ The growth rate in market size in later years is assumed to be the same each year. In year 1, it is assumed to follow a triangular distribution with worst case 5% annual growth, most likely case 12% annual growth, and best case 14% annual growth.

■ A single competitor might enter the market. Each year, the competitor has a 30% chance of entering the market, assuming it has not already entered.

The year after entering the market, a competitor causes a permanent loss of 40% of Lilly’s market share. For example, suppose the competitor enters in year 2, and Lilly’s share was 20%. Then in the years 3 to 6, its market share will be 12%.

■ At the beginning of year 3, Lilly will decide whether to sell Dialis overseas. If no competitor has entered the market by the end of year 2, there is a 70% chance that Lilly will sell the product overseas.

If a competitor has entered the market by the end of year 2, there is only a 30% chance that Lilly will sell the product overseas. Lilly’s market share overseas will equal its market share in the United States. It estimates that the overseas market is 25%

of world sales for drugs of this type. (The other 75% is U.S. sales.)

■ Each year the product sells for $120 and incurs a unit cost of $80.

■ Cash flows are discounted at 15% annually, and profits are taxed at 40%.

■ Cash flows for years 1 to 6 take place midyear.

Use simulation to model Lilly’s situation. Based on the simulation output, Lilly can be 95% sure the NPV for this project is between what two numbers? Would you go ahead with this project? Explain why or why not. (Hint: The way the uncertainty about the market size in year 1 is stated suggests using the General distribution, implemented with the RISKGENERAL function. Look it up in @RISK’s online help.)