9781118041581

The regression output associated with estimated Equation 4.17 shows that the explana-

tory variables are all statistically significant (according to their t-values) at a 99 percent

degree of confidence. The variables N, S, and C exhibit a moderate degree of multi-

collinearity (summer films tend to have starry casts and very wide releases), but, given

the large sample (204 observations), this does not unduly affect the validity of the indi-

vidual coefficients.

However, the R^2 of the regression equation is only .31—that is, the equation explains

only 31 percent of the variation in revenues for the films released that year. This should

not be very surprising. As noted, film revenues are inherently unpredictable. To explain

31 percent of these variations is a solid achievement. Clearly, the theater executive could

make a better forecast if she knew the magnitude of the studio’s advertising and promo-

tion budget, the reviews the film will receive, and the strengths of competing films for

release during the same time period.^12 However, at the time she must contract for films,

this information is unavailable. Given the large standard error of the regression, the mar-

gin of error surrounding AR is in the neighborhood of plus or minus 33 percent.

Predicting movie revenues will always be a risky proposition.

SUMMARY

Decision-Making Principles

1. Decisions are only as good as the information on which they are based.
   Accurate demand forecasts are crucial for sound managerial decision
   making.


2. The margin of error surrounding a forecast is as important as the
   forecast itself. Disasters in planning frequently occur when management
   is overly confident of its ability to predict the future.


3. Important questions to ask when evaluating a demand equation are the
   following: Does the estimated equation make economic sense? How well
   does the equation track past data? To what extent is the recent past a
   predictable guide to the future?

Nuts and Bolts

1. Demand estimation and forecasting can provide the manager with
   valuable information to aid in planning and pricing. Ideally, the
   forecasting process should provide (1) the forecast, (2) an estimate of its
   accuracy, and (3) an explicit description (an equation) of the
   dependency relationships.

168 Chapter 4 Estimating and Forecasting Demand

(^12) Conspicuously missing as explanatory variables in Equation 4.16 are price and income. These
variables have no demand effects because both are essentially fixed over the one-year time period
(and theaters do not vary ticket prices across films).
c04EstimatingandForecastingDemand.qxd 9/5/11 5:49 PM Page 168