It is now May 1 of year 0, and GM is deciding whether to produce a new car. The following information is relevant.

■ The fixed cost of developing the car is incurred on January 1 of year 1 and is assumed to follow a triangular distribution with smallest possible cost $300 million, most likely cost $400 million, and largest possible cost $700 million. The fixed cost is depreciated on a straight-line base during years 2 to 5. The tax rate is 40%.

■ The car will first come to market during year 2 and is equally likely to sell for 6, 7, or 8 years.

■ The market size during year 2 will be between 20,000 and 90,000 cars. There is a 25% chance that the market size will be less than or equal to 50,000 cars, a 50% chance that it will be less than or equal to 70,000 cars, and a 75% chance that it will be less than or equal to 80,000 cars. After year 2, the market size is assumed to grow by 5% per year.

■ The market share during year 2 is assumed to follow a triangular distribution with most likely value 40%. There is a 5% chance that market share will be 20% or less and a 5% chance that it will be 50% or more. The market share during later years will remain unchanged unless R&D makes a design improvement.

■ There is a 50% chance that R&D will make a design improvement during year 3, a 20% chance that it will make a design improvement during year 4, and a 30% chance that no design improvement will occur. There will be at most one design improvement. During the year (if any) in which a design improvement occurs, GM’s market share will increase to 50% above its current value. For example, suppose GM’s market share at the beginning of year 3 is 30%. If a design improvement occurs during year 3, its market share during year 3 and all later years will be 45%. 

■ The car sells for $15,000 each year.

■ The cost of producing the first x cars is 10,000x^0.9 dollars. This builds a learning curve into the cost structure.

■ During year 2 and later years, cash flows are assumed to occur midyear.

■ GM discounts its cash flows at 15% per year.

Use simulation to model GM’s situation. Based on the simulation output, GM can be 95% sure that the NPV generated by the car is between what two values? Should GM produce this car? Explain why or why not. What are the two key drivers of the car’s NPV? The way the uncertainty about the market size in year 2 is stated suggests using the Cumul distribution, implemented with the RISKCUMUL function. Look it up in @RISK’s online help.