II TheForecastIsWrong

Risk Management & Real OptionsII. The forecast is always wrong : 

Risk Management & Real Options II. The forecast is always wrong Stefan Scholtes Judge Institute of Management University of Cambridge MPhil Course 2004-05

NPV – the industry standard: 

NPV – the industry standard Aim: Value a project that requires an investment now and generates future cash flows over several periods, say several years Naïve: Value = sum of cash inflows – cash outflows Treat initial investment as cash outflow in period zero But: $ 1 today is worth more than $ 1 in a year’s time Inflation – time value of money Could invest $1 elsewhere – opportunity cost of capital Opportunity cost of capital: “Best” rate of return on alternative investment What does “best” mean? Bank account 2% p.a. Government bond 5% p.a. Stock market 15% p.a. Venture capital 25% p.a.

Rewards for risk taking: 

Rewards for risk taking Three main risks involved in investment Macro-economic risk (exchange rates, GDP growth, oil price, etc.) Shared throughout the economy Default risk Risk to debt holders: Company defaults Equity risk Risk to equity holders: Future cash flows are uncertain Want reward for taking risk Taking macro-economic risk is rewarded by government bond rate Taking default risk is rewarded by corporate spread = corporate debt rate – government bond rate Equity risk is reflected in company’s “beta” (CAPM  Finance textbook)

Discounted cash flow models: 

Discounted cash flow models Annual compounding Continuous compounding

Discounted cash flow models: 

Discounted cash flow models Discount rate d reflects the annual opportunity cost of capital for a project with a similar “level of risk” (whatever that means…) The higher the discount rate, the lower the present value of a future cash outflow (positive cash flow) The higher the discount rate, the lower the present value of a future cash inflow (negative cash flow) Present value (PV) of a project: sum of all discounted future cash flows Net present value (NPV) = PV minus today’s investment See Parking Garage.xls for an example of an NPV sheet (without tax considerations)

How is NPV used?: 

How is NPV used? Is this project economically sensible? “YES if NPV>0” Advise changes with discount rate Which of several projects should we do? “Choose the ones with larger NPV first, until budget is exhausted” Ranking changes with discount rate

What discount rate?: 

What discount rate? Practice: Discount rate is the return expectation of the capital owners, debtors and equity investors (“weighted average cost of capital”) BUT: Cambridge Antibody Technology (or the likes) : “…We know as a relatively young biotech company we should have a discount rate of 20%+. But if we were applying discount rates of this order, we wouldn’t do a single project…” BP (or the likes): “…We don’t discuss discount rates. We apply a 10% hurdle rate to all our projects…” Boeing (or the likes): “…Any sensible discount rate would wipe out all our returns beyond a 15 year time horizon – but our aircraft projects have a product life of 30+ years…”

NPV: Plus and minus: 

NPV: Plus and minus Plus: Setting up an NPV model forces you to think about the logic of a system’s value generation What are the key ingredients: costs, revenues What are the key drivers: demand, prices, unit costs, fixed costs, etc. What are the major tax implications: depreciation, etc. Minus: Which discount rate? Even more important: NPV calculation is based on projections of uncertain future demand, prices, unit costs, fixed costs, etc., and THE FORECAST IS ALWAYS WRONG Secondly, NPV is based on a fixed plan of action Does not account of deviating from plan if uncertainties unfold different from expectations (come to this later)

Cost forecast: 

Cost forecast Ratio of actual to estimated costs for routine airport resurfacing of runways Source: R. de Neufville, MIT

Oil price forecast: 

Source: U.S. Department of Energy, 1998 120 100 80 60 40 20 0 1975 1980 1985 1990 1995 2000 2005 Year 1982 Trend predicted 1981 1984 1985 1986 1987 1991 1995 Actual Dollars per Barrel Oil price forecast US DEO oil price forecasts 1983

Demand forecasts: 

Demand forecasts In the early 1980's McKinsey were hired by AT&T to forecast the growth in the mobile phone market until the end of the millennium. They projected a global market of 900,000 handsets Today, 900,000 handsets are sold every three days

A first cure: Sensitivity analysis: 

A first cure: Sensitivity analysis Simplest model: Numbers-in-numbers-out Need number calculator Sensitivity analysis = what-if analysis Improved model: Range-in-range-out Calculate the range of output values corresponding to a range of input values of one uncertain variable Vary one variable at a time Easily done in a spreadsheet

Typical graphical output: 

Typical graphical output

Tornado diagrams: 

Tornado diagrams Input ranges typically specified by Base value (“most likely”) Pessimistic value Optimistic value Base-case: Calculate base value for the output measure (e.g. NPV) on the basis of base values for inputs Tornado bar: For each input variable Determine the highest and lowest value of the output measure as the input variable varies over its range Extremes of output measure typically achieved at either end of the input range Determine the range of percentage deviations of the output from its base value as input varies over its range Rank the input variables by their impact on percentage deviation of the output variable from the base value

Tornado Diagram: 

Tornado Diagram Illustrates the effect of a RANGE of values of one input variable on the performance measure E.g.: The “variable cost range” can change the performance measure by more than 100% to either side of its base value

Problems with sensitivity analysis: 

Problems with sensitivity analysis Vary uncertainties one-by-one Varying many inputs simultaneously over their ranges and recording the highest and lowest value of the output measure leads to a huge range of the output measure “End-range” scenarios are overly pessimistic or overly optimistic Difficult to incorporate dependencies of variables (e.g. dependence of price on demand) Scenarios are played out for us but we don’t know how likely they are! Need to enter the world of probability to understand the notion of “likelihood”

Where are we going?: 

Where are we going? Introduction The forecast is always wrong The industry valuation standard: Net Present Value Sensitivity analysis The system value is a shape…