CH3 QuantativeDemand
CH3 QuantativeDemand
Course outline
Managerial Microeconomics
Basic Model
Managerial perspective
Additional topics
application
Demand
2
Question-PRS
According to an FTC Report by Michael Ward, AT&Ts own price elasticity of demand for long distance services is -8.64. AT&T needs to boost revenues in order to meet its marketing goals. To accomplish this goal, should AT&T raise or lower its price?
1.
2.
3.
Lecture Outline
1.
1. 2. 3. 4.
Elasticity of Demand
Own-price elasticity of Demand Income elasticity of demand Cross-price elasticity of demand Advertising elasticity of demand
2. 3.
Measures responsiveness of demand (or supply) to changes in an underlying factor (e.g. price, income, other prices).
Cross-price elasticity of demand = % change in quantity demanded % change in price of other commodity Positive if substitutes Negative if complements
Question-2
In July 1995 the Consumer Council of Hong Kong published a report on the gas industry in the territory. It contained annual estimates for the price elasticity of town gas: -1.384 in 1991, +0.706 in 1993. The report concluded: electricity and liquefied petroleum gas are not close substitutes for town gas with consumers having limited liability to switch in response to price changes.
Market
Suppose a firm sells two related products, X and Y. The impact of a small percentage change in the price of product X (%Px= Px/Px) on the firms total revenue is:
10
11
2. ESTIMATING DEMAND
1.
Demand functions
Linear Log-linear
2.
3.
Demand functions and elasticity Estimate demand function and understand regression results
13
Question-3
A New York Times article contained a table summarizing the price paid per (airwaves) license in 10 different regions, the number of licenses sold, and the population of each region. The average price is $70.7 million for a single license. lnP=2.23-1.2lnQ+1.25lnPop A regional telephone company: region population 7% higher than average of the 10, the FCC plans to auction the average number of licenses. How to estimate the amount of money needed to buy a license?
14
15
P
Elasticity differs along the demand curve
Constant Elasticity
Linear
Q Log-Linear
16
Suppose at P=30, A=446, C=30, D=90 What is own price elasticity? -0.16 What is advertising elasticity? 0.15 What is cross price elasticity? 0.42
17
1.53 ; -0.053
18
19
20
Regression analysis is concerned with the study of the dependence of one variable, the dependent variable, on one or more other variables, the explanatory variables. Used to estimate demand functions
21
Important terminology
22
23
,b ,b ,b ) (b Find best fit the data in the sense 0 P A I that it minimizes the sum of squared deviations.
) Minimize ( D D
i 1 i i
i n
24
R-square (R2): called the coefficient of determination: measure of the goodness of fit for a multiple regression model;
Variation Explained by Regression
R2 =
Total Variation of Y
Percentage of variation explained by the variable (Between 0 and 1)
If the number of estimated coefficients equals or exceeds the number of observations, R2 is always 100%. Degrees of freedom of residue: the number of observations minus the number of coefficients to be estimated. (including constant.)
26
Corrected coefficient of determination (adjusted R2 ): Downward adjustment of R2 in the light of the number of data points and estimated coefficients. R2 = R2 ((k-1)/(n-k)) (1-R2) K is the number of coefficient to be estimated.
27
n is the number of observations, k is the number of estimated coefficients. Fk-1, n-k = R2 / (k-1) (1-R2)/(n-k)
28
2.3.3 F Statistic
F statistic ranges from 0 to infinity. The F statistic is calculated under the hypothesis that the coefficients (excluding the constant) are all 0. If the hypothesis is true, F statistic follows F distribution. If the F statistic is large enough (> 3 or 4 depending on n, k), then we can reject the hypothesis, which means the regression estimates are statistically significant.
29
F4,30
2.69
4.02
Significance (P-Value) gives the probability of obtaining the value of F statistic this extreme (large) or even more extreme (larger) if the Null hypothesis is true, i.e., all coefficients (excluding constant) are 0. If this probability falls below a specific benchmark, then we say that the regression estimates are statistically significant. The conventional benchmark is 1% and 5%.
31
t stat of bp^
t statistic measures individual variable significance It ranges from negative infinity to positive infinity. tn-k = b^ / standard Deviation of b^ t statistic is calculated under the hypothesis that the coefficient is 0. If the hypothesis is true, t statistic follows t distribution.
32
2.576
2.3.4 t statistic
If the absolute value of the t statistic is large enough ( > 2.5 or 3), then we can reject the hypothesis, which means the estimated coefficient is statistically significant. The significance (P-Value) gives the probability of obtaining the value of t statistic or more extreme values if the coefficient is 0, i.e., the hypothesis is true.
34
(bp^-bp)/sd of bp^ ~ t n-k Prob{|(bp^-bp)/sd of bp^ | <=2} = 95% Prob{-2<= (bp^-bp)/sd of bp^ <=2} = 95% Prob{bp^-2sd of bp^<=bp<=bp^+2sd of bp^} = 95%
35
New coke: In April 1985, Coca-Cola announced its plan to change to an improved formula.190,000 taste tests conducted by the company. Three months later, apologized, and Coke Classic.
36
3. Creating demand
What (goods or services) else can my customers buy so that they will value my product (service) more?
Real estate Supermarkets
37
3. Thinking Complements
Auto insurance Auto repairs: parts and service. Alfa Romeo and Fiat
had trouble selling their cars in the U.S., because people knew that theyd have trouble finding spare parts and qualified mechanics.
38
Boeing bid for U.S. Post Office to deliver airmail Automakers own car rental agencies:
Chrysler owns part of Avis GM owns part of Avis and owned National Mitsubishi owns part of Value Rent-a-car
39
40