Post on 12-Jan-2016
The Determinants of Demand for The Determinants of Demand for Hybrid CarsHybrid Cars
Shad AhmedShad AhmedMark BaldwinMark BaldwinKelly FogartyKelly Fogarty
Michael KendraMichael Kendra
OverviewOverview ObjectivesObjectives Hypotheses / Variable ExaminedHypotheses / Variable Examined SoftwareSoftware ApproachApproach ModelModel VariablesVariables StatisticsStatistics ResultsResults Policy ImplicationsPolicy Implications
ObjectivesObjectives To develop an econometric model and analyze historical To develop an econometric model and analyze historical
data sets to determine which variables explains what data sets to determine which variables explains what factors drive the demand for hybrid vehicles in order to factors drive the demand for hybrid vehicles in order to confirm or deny public speculation.confirm or deny public speculation.
To maximize the statistical significance of the model in To maximize the statistical significance of the model in order to provide forecasters with a working model that can order to provide forecasters with a working model that can be used to make predictions about future demand for be used to make predictions about future demand for hybrid cars.hybrid cars.
To provide the environmentally conscience public, To provide the environmentally conscience public, automobile industry, and law makers a foundation upon automobile industry, and law makers a foundation upon which to make policy decisions based on objective which to make policy decisions based on objective reasoning.reasoning.
To provide a solid foundational model upon which future To provide a solid foundational model upon which future research projects can build on.research projects can build on.
HypothesesHypotheses
HH11: The demand for hybrid cars is explained : The demand for hybrid cars is explained by gas prices.by gas prices.
HH22 : The demand for hybrid cars is : The demand for hybrid cars is explained by the Producers Price Index for explained by the Producers Price Index for automobiles.automobiles.
HH33: The demand for hybrid cars is explained : The demand for hybrid cars is explained by the personal consumption on by the personal consumption on expenditures for automobiles by the US expenditures for automobiles by the US populationpopulation
Variables examinedVariables examined
Dependent variable: Demand for hybrid vehiclesDependent variable: Demand for hybrid vehicles
Economic IndicatorsEconomic Indicators– PPI for motor vehiclesPPI for motor vehicles
– Personal consumption Personal consumption on motor vehicleson motor vehicles
– Bank loan rateBank loan rate
– Consumer credit Consumer credit outstandingoutstanding
– Unemployment rateUnemployment rate
Energy IndicatorsEnergy Indicators– Price of gasolinePrice of gasoline
– Barrels of gasoline Barrels of gasoline consumedconsumed
– Total energy Total energy consumption of US consumption of US populationpopulation
Variable Identification and DefinitionVariable Identification and Definition
VariableVariable TypeType Hypothesized SignHypothesized Sign Demand for Hybrid CarsDemand for Hybrid Cars DepDep PPI for motor vehiclesPPI for motor vehicles End End NegNeg Personal consumption on Personal consumption on EndEnd PosPos
motor vehiclesmotor vehicles Bank loan rateBank loan rate ExoExo NegNeg Consumer credit outstandingConsumer credit outstanding ExoExo PosPos Unemployment rateUnemployment rate ExoExo NegNeg Price of gasolinePrice of gasoline EndEnd PosPos Barrels of gasoline consumedBarrels of gasoline consumed EndEnd PosPos Total energy consumption of Total energy consumption of EndEnd PosPos
US populationUS population
SoftwareSoftware
WinORSfx was used to develop the model.WinORSfx was used to develop the model. Availability of Economic data from Availability of Economic data from
EconomagicEconomagic Extensive ability to determine statistical Extensive ability to determine statistical
significance.significance.
ApproachApproach
Monthly data sets were used from 2004 – 2007.Monthly data sets were used from 2004 – 2007. Stepwise regression was run to determine which variables Stepwise regression was run to determine which variables
to eliminate from the model.to eliminate from the model. Remaining variables were examined for practicality.Remaining variables were examined for practicality. Ordinary Least Square method was used to test the Ordinary Least Square method was used to test the
remaining variables for multicollinearity, remaining variables for multicollinearity, homoscedasticity, explainability, and serial correlation.homoscedasticity, explainability, and serial correlation.
First Difference was run to attempt to eliminate serial First Difference was run to attempt to eliminate serial correlation.correlation.
A final model was assembled.A final model was assembled.
Determinants ModelDeterminants Model
Qx = -167376 + 758*P + 163.20*PQx = -167376 + 758*P + 163.20*Pgasgas + .104C + .104C
Qx = Demand for hybrid vehiclesQx = Demand for hybrid vehicles
P = PPI for automobilesP = PPI for automobiles
PPgasgas = Price of gasoline = Price of gasoline
C = Personal consumption of automobilesC = Personal consumption of automobiles
Predictive Ability of ModelPredictive Ability of ModelPredictive Ability (OLS)
Dependent Variable: Total Hybrid Sales
Actual Predicted
Observation454239363330272421181512963
Act
ual &
Pre
dict
ed
32,500
29,250
26,000
22,750
19,500
16,250
13,000
9,750
6,500
3,250
F-statisticF-statistic
The P-value 0.00001 is significantly below The P-value 0.00001 is significantly below the critical value, 0.05 the critical value, 0.05
The model is statistically significant above The model is statistically significant above the 95% confidence intervalthe 95% confidence interval
F value: 44.63F value: 44.63
P value: 0.00001P value: 0.00001
Coefficient of DeterminationCoefficient of Determination Demonstrates that a high degree of variability in Demonstrates that a high degree of variability in
hybrid sales can be explained by variation in the hybrid sales can be explained by variation in the independent variablesindependent variables
Root MSERoot MSE 3489.6003489.600 SSQ(Res)SSQ(Res) 426205816.031426205816.031 Dep.MeanDep.Mean 15818.92315818.923 Coef of Var (CV)Coef of Var (CV) 22.060%22.060% Multiple RMultiple R 89.038%89.038% R-SquaredR-Squared 79.278%79.278% Adj R-SquaredAdj R-Squared 77.502%77.502%
MulticollinearityMulticollinearity
No evidence of multicollinearity is present No evidence of multicollinearity is present in the model (VIF<10)in the model (VIF<10)
Average VIF = 1.037Average VIF = 1.037
Parameter VIFsParameter VIFs
Variable:Variable: VIFVIF
Price of gasolinePrice of gasoline 1.0401.040
PPI of automobilesPPI of automobiles 1.0531.053
Personal consumption on Personal consumption on 1.0181.018
AutomobilesAutomobiles
Constant VarianceConstant Variance
White’s Test shows that the model is White’s Test shows that the model is homoskedastistichomoskedastistic
White’s Test = 8.32White’s Test = 8.32
P-Value for White’s = .502P-Value for White’s = .502
Constant Variance GraphConstant Variance GraphConstant Variance Test (OLS)
Dependent Variable: Total Hybrid Sales
Predicted28,95626,54324,13021,71719,30416,89114,47812,0659,6527,2394,8262,413
Res
idua
l
14,700
12,600
10,500
8,400
6,300
4,200
2,100
0
-2,100
-4,200
Auto CorrelationAuto Correlation
Durbin Watson test shows evidence of Auto Durbin Watson test shows evidence of Auto CorrelationCorrelation
Ho: Rho = 0Ho: Rho = 0 Rho: Pos & NegRho: Pos & Neg RejectReject Rho: PositiveRho: Positive Do Not RejectDo Not Reject Rho: NegativeRho: Negative RejectReject
First difference solution attempted; resulted First difference solution attempted; resulted in a new R-squared value of .277in a new R-squared value of .277
Normality of Error TermsNormality of Error TermsNormal Probability Chart (OLS)
Dependent Variable: Total Hybrid Sales
Expected Residual9,0007,5006,0004,5003,0001,5000-1,500-3,000-4,500-6,000-7,500-9,000
Sor
ted
Res
idua
l
14,700
12,600
10,500
8,400
6,300
4,200
2,100
0
-2,100
-4,200
ElasticitiesElasticities
VariableVariable Parameter Parameter EstimateEstimate
Price of gasolinePrice of gasoline 2.892.89
PPI of automobilesPPI of automobiles 8.388.38
Personal consumption on Personal consumption on 3.873.87
AutomobilesAutomobiles
Elasticity ImplicationsElasticity Implications
Income elasticityIncome elasticity– Hybrids are a “luxury” itemHybrids are a “luxury” item– Elasticity is >1Elasticity is >1– As income increases, QAs income increases, Qxx increases increases
Cross price elasticityCross price elasticity– Gasoline and other automobiles are substitutesGasoline and other automobiles are substitutes– Elasticity is >1Elasticity is >1– As prices of gasoline and other autos increases, As prices of gasoline and other autos increases,
QQxx increases increases