Question: On March 4, 1990, the New York Times reported "Wine Equation Puts Some Noses Out of Joint." In this problem you will estimate an equation that predicts the quality of wine. Typically wine is rated long before it is ready to drink by having an "expert" taste the premature wine from the barrel. Their pronouncements on the quality of the wine help determine the wholesale price. Often it will take many years before the wine can be ranked with any precision. Professor Ashenfelter, however, has devised a regression model that predicts wine quality by incorporating as explanatory variables the weather that prevailed when the wine was grown. The belief that weather affects wine quality is widespread, but Ashenfelter offered an explicit equation to assess the quality of red wine from the Bordeaux region of France. The data you will use contains four variables: an index of auction prices for wine that has matured (measured in logs), the level of winter rain (October through March-measured in millimeters), average temperature during the growing season (April through September-measured in degrees celsius), and the level of harvest rain (August through September-measured in millimeters). These data are available in http://www.wiley.comicollege/ashenfelter.
a. Estimate the wine equation through the period 1987. Interpret the coefficients.
b. What proportion of the variation in wine prices does the model explain?
c. Explain how you would use this model to predict the quality of wines that are not yet mature.