Bài giảng Managerial Economics - Chapter 07: Demand Estimation and Forecasting

Chapter 7: Demand Estimation and ForecastingMcGraw-Hill/IrwinCopyright © 2011 by the McGraw-Hill Companies, Inc. All rights reserved.Direct Methods of Demand EstimationConsumer interviewsRange from stopping shoppers to speak with them to administering detailed questionnairesDirect Methods of Demand EstimationPotential problems with consumer interviewsSelection of a representative sample, which is a sample (usually random) having characteristics that accurately reflect the population as a wholeResponse bias, which is the difference between responses given by an individual to a hypothetical question and the action the individual takes when the situation actually occursInability of the respondent to answer accuratelyDirect Methods of Demand EstimationMarket studies & experimentsMarket studies attempt to hold everything constant during the study except the price of the goodLab experiments use volunteers to simulate actual buying conditionsField experiments observe actual behavior of consumersEmpirical Demand FunctionsDemand equations derived from actual market dataUseful in making pricing & production decisionsSimple regression analysisSimple linear regression assumes one-way causationInappropriate for competitive marketsPrice and output are simultaneously determined in competitive marketsAdvanced regression techniques are available for estimating demand in competitive markets7-6Empirical Demand FunctionsIn linear form, an empirical demand function can be specified aswhere Q is quantity demanded, P is the price of the good or service, M is consumer income, & PR is the price of some related good REmpirical Demand FunctionsIn linear formb = Q/Pc = Q/Md = Q/PRExpected signs of coefficientsb is expected to be negativec is positive for normal goods; negative for inferior goodsd is positive for substitutes; negative for complementsEmpirical Demand FunctionsEstimated elasticities of demand are computed asIn this form, elasticities are constantNonlinear Empirical Demand SpecificationWhen demand is specified in log-linear form, the demand function can be written asTo estimate a log-linear demand function, covert to logarithmsDemand for a Price-SetterTo estimate demand function for a price-setting firm:Step 1: Specify price-setting firm’s demand functionStep 2: Collect data for the variables in the firm’s demand functionStep 3: Estimate firm’s demand using ordinary least-squares regression (OLS)Checkers Pizza7-12Linear Regression7-13Time-Series ForecastsA time-series model shows how a time-ordered sequence of observations on a variable is generatedSimplest form is linear trend forecastingSales in each time period (Qt ) are assumed to be linearly related to time (t)If b > 0, sales are increasing over timeIf b < 0, sales are decreasing over timeIf b = 0, sales are constant over timeLinear Trend Forecasting Use regression analysis to estimate values of a and b Statistical significance of a trend is determined by testing or by examining the p-value forA Linear Trend Forecast(Figure 7.1)Estimated trend lineSalesQt199719981999200020012002200320042005200620077201212Linear Trend Estimation7-17Forecasting Sales for Terminator Pest Control (Figure 7.2)Seasonal (or Cyclical) VariationCan bias the estimation of parameters in linear trend forecastingTo account for such variation, dummy variables are added to the trend equationShift trend line up or down depending on the particular seasonal patternSignificance of seasonal behavior determined by using t-test or p-value for the estimated coefficient on the dummy variableSales with Seasonal Variation(Figure 7.3)2004200520062007Dummy VariablesTo account for N seasonal time periodsN – 1 dummy variables are addedEach dummy variable accounts for one seasonal time periodTakes value of one (1) for observations that occur during the season assigned to that dummy variableTakes value of zero (0) otherwiseEffect of Seasonal Variation(Figure 7.4)SalesTimeQttQt = a′ + bta′aQt = a + btcQuarterly Sales Data7-23Dummy Variable Estimates7-24Dummy Variable Specification7-25Some Final WarningsThe further into the future a forecast is made, the wider is the confidence interval or region of uncertaintyModel misspecification, either by excluding an important variable or by using an inappropriate functional form, reduces reliability of the forecastSome Final WarningsForecasts are incapable of predicting sharp changes that occur because of structural changes in the marketConfidence Intervals7-28