1Overview: Demand• Demand for Anti-Ulcer Drugs– Data, Estimation– Interpretation of Empirical Results• Modeling of Consumer ChoiceDemand Analysis• Today– Empirical Demand Analysis– Modeling of Consumer Demand •Later– Demand and Product Characteristics– Discrete Choice Models– Network Effects– Dynamic Adjustments in Demand2Market Demand for Anti Ulcer Drugs• Typical Empirical Application– Rapid Growth Market– Changes in Market Structure• Focus on Price Effects and Advertising Effects– Price Elasticity and Advertising Elasticity• Example of Practical Modeling ConsiderationsMarket Demand for Anti-Ulcer Drugs• Monthly Observations, Aug. 1977 - May 1993• Four Drugs: Tagamet, Axid, Pepcid, ZantacPrescription (prior to OTC versions)• VariablesLQ - Log (Market Quantity)LP - Log (Market Price)LADV - Log (Cumulative Advertising)T - Time Trend • Data Source: Berndt, Bui, Reiley and Urban (1994)3Market Demand for Anti-Ulcer Drugs• Model is of the Form:LQ = α + ElasPLP + ElasADVLADV + τ T + ε• Basic EstimatesLQ = -10.04 - 1.16 LP + .88 LADV + .001 T + ε• Namely ElasP= −1.16, ElasADV= .88 (Not bad!)•But….Model Specification AnalysisYikes!!! Something is wrong!!!• What Could Cause the Residual Pattern? • What should we do about it ???4Introduction of Anti-Ulcer Drugs• Drugs Did Not Appear Simultaneously• Introduction DatesTagamet (Smith Kline): August 1977Zantac (Glaxo): June 1983Pepcid (Merck): October 1986Axid (Lilly): April 1988• How is this incorporated?Final Estimation Results• Refined Model is of the Form:LQ = α + ElasPLP + ElasADVLADV + εwhere ElasticitiesPADVOne Drug -1.67 0.77Two Drugs -0.65 0.4Three Drugs -0.65 0.4Four Drugs -0.65 05Retail Pricing and Promotions• Retail pricing involves list prices as well as promotions, or temporary price reductions• Modeling and estimation applied here too, in early growth phase• Household models and targeted couponing6Consumer Choice Models• How Do Consumers Make Decisions?• Behavioral Assumption: Consumers Act in Their Best Interests– Interests = “Preferences” or “Utility”– Best Interests = “Maximize Utility” over available choices. • Utility MaximizationChoose Quantities of Goods: A, B, .... bymaximizing Utility Function U(A,B,...) subject to Budget Constraint: pAA+ pBB + ... = IConsumer Choice Models (2)Optimal: MRSAB = pA/pB, etc., namely MUA/pA= MUB/pB= (all goods); “Equal Bang per Buck”Result: Individual Demand FunctionsA = DA(pA,pB,...,I) B = DB(pA,pB,...,I)Empirical Models Build in Consumer Differences, Demographics, etc.78Take Away Points• Demand curves are real: they can be estimated!• Estimation involves– Model specification– Estimation– Interpretation and modification• A basic understanding of regression output allows you to critically assess claims based on it.• Utility maximization models are the workhorse of economics and finance, among
View Full Document