Understanding Demand Theory

Understanding Demand Theory - Transcript

Understanding Demand
Information about demand is essential for making pricing and production decisions Empirical estimation of demand needed Decisions about product demand regarding pricing and production requires both judgment and quantitative analysis Need to employ statistical tools regression analysis to achieve objective Statistical analysis will help us estimate function and demand elasticities

Steps to Demand Estimation
Specify a Demand Function use knowledge of consumer theory Specify a functional form How to choose between one form and another form How to estimate demand when firms make price or take price

Estimation of Demand
Demand Projections for NISSAN car Demand Projections for Nestle Milk Demand Projections for TATA Steel

Empirical Demand
Derived from actual market data price quantity price of related product income advertising They allow us to get quantitative estimates of the effect on sales of change in price change in income change in competiting products They allow us to attempt some demand projections forecasting a long term pattern of demand for a product Assist us to price effectively as we get a hang of elasticity also

Demand Function
How to determine which variables to include in the demand function Look up theory of consumer behavior to understand what determines demand in general There may be some variables which may be difficult to quantify expectations and taste Thus we specify a general demand function

Demand Function
Recognize geographic boundaries Western region versus Northern region of IndiaINCOME Recognize to whom the product is targeted Recognize as to how many related products you would like to consider Recognize market size

Demand Function
Q a bP cM dP2 eN
P Price M Income P2 Price of related product N Number of buyers A b c d and e are called parameters and measure change in quantity demand that would result from a one unit change in price Q is called dependent variable and P M P2 independent variables

Demand Function
b is assumed negative law of demand c is defined as ZQ ZM and if it is greater than zero normal good and if less than zero inferior good d is defined as ZQ ZP2 and if it is greater than zero substitute and if less than zero complements n is assumed positive

Demand Function
We can also calculate elasticity price income and cross price from estimated demand functions HOW We know that b c and d measure slopes and we have from data different data points P Q just use elasticity formula

Estimation of Demand
Linear or Non Linear form of Demand Function Suppose you want to have a demand model where elasticity s are constant Suppose you know that for a certain product as income gets higher a small percentage of income goes to consuming that product and more is saved Suppose the HR department wants to study the annual employee earnings as a function of the agea young worker will get more income with age but increase in age will not increase income beyond a point

Estimation of Demand
Model Specification Consider some agricultural products since it takes time to grow decisions on how many acres to plant must be handled years in advance before the product is actually supplied to the market There may be a lagged relationship between price of cotton and production of cotton If cotton prices change farmer s may not be able to respond immediately interpretation of coefficient is different

Estimation of Demand
Model Specification

Some times it may be difficult to specify a variable Suppose Y represents the salary of an Executive and salary levels depends on the Institution granting MBA X 1 if the employee is IIM MBA O other wise

Estimation of Demand
Which one to choose What is the purpose of the demand estimation In what ways are the chosen variable required to relate to demand or dependent variable or you expect them to relate Between L and NL forms we have to consider in which data set you expect elasticities to vary or not vary P Q data spread over a wide range of values or P Q data clustered over a narrow price quantity range

Estimation of Demand
STEP1 Review the Literature Consumer Theory and develop a theoretical Model Demand Equation with assumptions and signs STEP 2 Specify the Model Select the independent variables price1 price2 income and the Functional Form Non Linear or Linear

Estimation of Demand
STEP 3 Set up or HYPOTHESIZE the signs that you expect of the coefficients 1 Own Price to be negative 2 Cross price to be positive 3 Income to be positive

Estimation of Demand
STEP 4 Collect Inspect and Clean DATA Some comments on data
The more observations in the sample size the better as long as the observations are from the same population All variables must have the same number of observations All variables should have the same frequency monthly quarterly or annual Unit of measurements Millions Rs lakhs Conclusions of signs and significance are independent of units of measurements

Estimation of Demand
Inspection of DATA is important Look for OUTLIERS and check and in extreme cases if the correct number is not found or number is not from the same population drop LOOK for possible relationship Scatter PLOT Look at mean maximum and minimum values to check out inconsistencies GPA of 7 in a scale of 4

Estimation of Demand
Steps 1 to 4 can take months STEP 5 D Q a bP cM dP2 eN E P Price M Income P2 Price of related product N Number of buyers E Error term which captures all other variables not captured by the model

Estimation of Demand
What does E represent
All variations in D that cannot be explained by included independent variables It is a symbol of econometricians ignorance or inability to model all the movements in the dependent variable D We say E consists of 2 segments deterministic and stochastic deterministic component is E Y given X if the average height of all 14yr girls is 5 feet then 5 feet is the expected value of a girl s height given that she is 14 yr old The value of Y in real world is unlikely to be like this Thus we need to add a stochastic component

Estimation of Demand
What does E represent
All variations in D that cannot be explained by included independent variables It is a symbol of econometricians ignorance or inability to model all the movements in the dependent variable D We say E consists of 2 segments deterministic and stochastic deterministic component is E Y given X if the average height of all 14yr girls is 5 feet then 5 feet is the expected value of a girl s height given that she is 14 yr old The value of Y in real world is unlikely to be like this Thus we need to add a stochastic component

Estimation of Demand
Suppose aggregate consumption in a economy is a function of disposable income There may be many reasons why actual consumption may be less or more Uncertainty of future and it is hard to measure thus the impact of this variable will be land up in E thus some random events might increase or decrease aggregate consumption

Estimation of Demand
Run the econometric softwareSTATA E Views LIMDEP GAUSS

Estimation of Demand
Document the results Check how well the equation fits the datacheck R2 Check the signs and magnitude of the coefficients and did you expect them THUS we are into evaluations of econometric equations

Estimation of Demand
Q 103 20 6 38P 2 56M 6 72P2 0 88 1 26 0 93 t value 7 22 N 100 R2 89
How to interpret this equation Number in parenthesis is the estimated standard error of the estimated coefficient

T value is used to test if the true value of the coefficient is different from zero WE check the quality of the regression results

Estimation of Demand
Method of estimation depends upon whether price of the product is manager determined or market determined Price taking firms versus Price setting firms in the First one price is being set within a system of demand and supply variable is being determined by a system of equations and is an endogenous variable In the second case price is being determined outside the system demand and supply manager of the firm determines it

Estimation of Demand
Market determined demand involves variations in price and quantity arising from both supply and demand all factors that can either shift supply and demand Use 2 stage least square estimators Manager determined demand involves variations in Q arising from variations in P Use Ordinary least square estimators