demand estimation
DESCRIPTION
Demand Estimation. Specifying a Demand Equation General Form: Q = F(P, M, P R ,T, P e , N) Empirical (Regression) Form: Q = a + bP + cM + dP R + eN You estimated a demand equation like this in the demand project. The Identification Problem. Estimate Q = a + bP + cM + dP R + eN - PowerPoint PPT PresentationTRANSCRIPT
Demand Estimation Specifying a Demand Equation General Form: Q = F(P, M, PR,T, Pe, N) Empirical (Regression) Form:
Q = a + bP + cM + dPR + eN You estimated a demand equation
like this in the demand project
The Identification Problem Estimate Q = a + bP + cM + dPR + eN
If P and Q are determined by supply and demand, how do you know that you are estimating a demand relationship?
Answer: You don’t! If you have not estimated a demand
relationship, you have an identification problem. Warning signs:
Coefficient on price (b) is positive Coefficient on price (b) is not statistically significant
Solution to Identification Problem Use a technique called 2-stage Least Squares
Specify Demand and Supply Equations Qd = a + bP + cM + dPR + eN Qs = f + gP +hPI
Run a 2-stage least squares program When is Identification not a problem?
Data are for a single firm setting it’s own price Price not set by market (regulated prices)
Electricity prices are regulated by the states Identification not a problem for demand project
Calculating Elasticities for estimated demand equations Linear equation – demand project
Q = a + bP + cM + dPR + eN E = b(P/Q) = (ΔQ/ΔP)(P/Q) EM = c(M/Q); EPR = d(PR/Q)
Log-linear equation – constant elasticity Ln(Q) = g + h(lnP) + j(lnM) + k(lnPR) Coefficients (h, j, k) are elasticities No calculation needed
Calculate elasticities at the sample means Milkwh= 10800 – 3581(Pkwh) +
0.004(Pop) + 2252(PGas) Elasticity = (coeff.)(value/Milkwh) Sample Means: Milkwh=25365
Pkwh=9.0 Pop=5,756,577 PGas=11.4 E = -3581(9.0/25365) = -1.27 Epop = 0.004(5,756,577/25365) = 0.91 Epgas= 2252(11.4/25365) = 1.01
Exercise: Elasticities Milkwh= 10800 – 3581(Pkwh) +
0.004(Pop) + 2252(PGas) Milkwh=38,526 Pkwh=6.92
Pgas=10.6 Pop=5,900,962 Elasticity = (coeff.)(value/Milkwh) Calculate electricity demand
elasticities with respect to Pkwh, Pop, & Pgas
Forecasting Demand Using Elasticities
Multiply elasticities by projected % changes in explanatory variables
Add the results to get projected % change in demand
Using linear regression equation Multiply coefficients by projected values for
explanatory variables in future period Add results and intercept to get forecast of
demand
Forecasting with elasticities Estimate a log-linear equation
LMilkwh = 0.04 – 0.92(LPkwh) + 1.0(LPop) + 0.4(LPgas) – 0.4(Linc)
Get projected % changes Pkwh:10% Pop:1% Pgas:20% Inc:2%
Calculate the projected % change in Milkwh -0.92(10%)+1.0(1%)+0.4(20%)-0.4(2%) = -
1% Do not use the intercept! It doesn’t change.
Suppose Pgas goes down by 20%, not up?
Exercise: Forecasting with linear demand Estimate a linear demand equation
Milkwh= 10800 – 3581(Pkwh) + 0.004(Pop) + 2252(PGas)
Get forecasts of explanatory variables Pkwh=10 Pop=20,000,000 Pgas=10
Calculate a Forecast for Milkwh Substitute forecast values for explanatory
variables and do the arithmatic