Energy Demand Analysis and Forecasting
The following model (in Box 1) was developed by Company X in 2006 to inform the relationship between electricity consumption, gross state product (GSP), and electricity price for New South Wales. Data for the period 1980-2005 was used to develop this model.
Box 1 Electricity demand model for New South Wales
|
Dependent Variable: LOG(Electricity Consumption)
|
|
Method: Least Squares
|
|
|
|
|
|
Sample: 1980 2005
|
|
|
|
|
|
Included observations: 26
|
|
|
|
|
|
Variable
|
Coefficient
|
Std. Error
|
t-Statistic
|
Prob.
|
|
Constant
|
1.091484
|
0.553990
|
-7.067906
|
0.0000
|
|
LOG(GSP)
|
0.795888
|
0.032037
|
41.35211
|
0.0000
|
|
LOG(Electricity Price)
|
-0.049489
|
0.069514
|
-9.682326
|
0.0000
|
|
R-squared
|
0.995855
|
Mean dependent variance
|
11.06420
|
|
Adjusted R-squared
|
0.994640
|
S.D. dependent variance
|
0.121828
|
|
S.E. of regression
|
0.008919
|
Akaike info criterion
|
|
-6.433828
|
|
Sum squared residual
|
0.001034
|
Schwarz criterion
|
|
-6.288968
|
|
Log likelihood
|
54.45359
|
Hannan-Quinn criterion
|
-6.426410
|
|
F-statistic
|
1392.803
|
Durbin-Watson stat
|
|
1.796665
|
|
Prob(F-statistic)
|
0.000000
|
|
|
|
|
|
This model has been used by:
i. policy makers and regulators in the government, to develop policies that targeting change in electricity demand pattern, and
ii. utility planners and market operators, to plan for future electricity demand.
The drastic changes in socio-economic conditions and government policies over the recent years prompt Company X to revise the "current" model. In view of the renowned UTS graduates, you are approached by this company to update the model.
With curiosity, your first attempt is to assess the model's predictive performance. This is done by using the "current" model (Box 1), and data on exogenous variables published by the Australian Bureau of Statistics, to develop estimates for backcast, ex-post simulation, and ex-post forecast. These estimates are then compared against actual electricity consumption, as shown in the figure in Annex 1. You also calculate the deviation between the estimated demand trend and actual consumption, in terms of Root Mean Square Percentage Error (RMSPE), and Theil Inequality Coefficients - also shown in the figure in Annex 1. You have noticed that there is a substantial deviation between electricity demand estimates and actual electricity consumption over the two time periods: i) prior to 1980, and ii) post-2005.
TASKS
1. Provide a philosophical critique of the "current" model. (no more than two dot points with each dot point no more than five lines)
2. Comment on the robustness of the "current" model in terms of its applications for:
- policy makers and regulators, and
- utility planners and market operators.
(no more than five lines for each of the above two dot points)
3. Consider the following statement: The applied econometrician soon discovers from experience that a useful model is not one that is true or realistic but one that is parsimonious, plausible and informative.
Based on your answers in Tasks 1 and 2, comment on the relevance of this statement in the context of updating the model for Company X. (no more than five lines for each dot point - parsimonious, plausible and informative)
Annex 1 Electricity consumption trends in New South Wales (TWh)
