NEMS Forecast Evaluation Methodology
This
is a working document prepared as a job of work (DE-AP01-06EI38129.A000) on
behalf of the Energy Information Administration (EIA) in order to solicit
advice and comment on methods for evaluating EIA forecasts.
Summary
The purpose of the methodology proposed here is to propose a way to assess the accuracy of National Energy Modeling System (NEMS) projections. NEMS is configured to project future energy product production and consumption in a fashion that accounts for the wide range of detailed circumstances that result in multi-market economic equilibria for energy products. In addition to market forces NEMS accounts for technological change and the impact of government actions and policies. The evaluation methodology presented here calls for constructing statistical approximations of important energy market relationships implicit to NEMS, e.g., residential and commercial sector energy consumption. The approximations will be derived from solutions designed to isolate important influences and regression analyses of the NEMS solutions prepared in support of the Annual Energy Outlook (AEO). The approximations are to be specified to account for important, explanatory relationships, e.g., the elasticity of sectoral energy consumption to the sectoral price of energy. Based on this, “differences” between NEMS projections and the actual values of the variables projected can be partitioned among general uncertainty, “errors” in projecting explanatory variables, structural changes in market behavior, and transitory influences such as the weather. An illustration of the approach is provided for residential and commercial sector demand for delivered energy, utilizing the versions of NEMS used in support of the 1998 – 2000 AEO’s.
January 2007
NEMS Forecast Evaluation Methodology
I. Background. In
principle, an evaluation of NEMS forecast accuracy that accounts for the
various sources of differences between forecast and actual values can be
readily accomplished. The form of the model, in brief, involves projecting a
variety of important influences on energy markets, e.g., technology and the
parameters of consumer behavior. Based on the projections of these “conditional
variables” the model then determines the corresponding multi-energy-market
equilibria with the associated projections of energy product prices and
quantities. When a forecast period actually occurs, and the actual values of
the conditional variables are known, the model can be re-run using the actual
values for model assumptions and the resulting projections compared to the
actual prices and quantities. The impact of individual conditional variables
can be determined by differentially using the original projections and the
actual values, ceteris paribus. Routines
within the model might also be amended to account for changes in government
policies, technologies, or consumer behavior. Extra-model influences such as
weather could also be accounted for in the historical period, independent of
NEMS, or using NEMS’ own routines for
processing the impact of weather variables. In general, the size of the model
and need to maintain, and then run,
non-current versions of the model over many years appear to make this
approach impracticable. In any event, this approach has not been attempted. To
do so, among other requirements, would entail archiving and re-implementing
versions of NEMS years and even decades after their current use. A multitude of
problems suggest that this approach is too expensive, and perhaps not possible
as a practical matter.
In
spite of these difficulties, it remains desirable to assess the degree to which
NEMS projections differ from the eventual, historical values, as related to
differences between the assumed and actual values of important conditional
variables. The reason for this is that NEMS structure is explicitly intended to
enable detailed, conditional evaluations of many distinct influences upon
energy consumption and production. This model structure is intended to support
detailed analyses of alternative energy policies and contingent future
circumstances. Accordingly, the merits of the model’s approach requires an
assessment of model performance that embodies, to the degree practicable, the
same level of detail as it relates to projected versus actual outcomes for the
variables forecast by the model.
II. Proposed Method. An alternative to running NEMS in a fully integrated fashion, using actual values for the conditional variables, is to isolate the explanatory variables at issue for each model component before the fact, at the time a current version of NEMS is formulated. . Model solutions can be constructed particular to each NEMS component that embody changes in important assumptions. Given this, the underlying interrelationships, e.g., price elasticities of supply and demand, can be determined for a current version of the model and then retained. In general, the method would be to selectively change important assumptions, ceteris paribus, and catalogue the corresponding sensitivities. Later, these sensitivities could be applied to actual data to determine the basis for forecast differences without having to archive and re-run the corresponding version of NEMS (This was done for RSTEM in Costello (2006) Reduced Form Energy Model Elasticities from EIA’s Regional Energy Model (RSTEM) , released 5/9/2006 as a one time report, for price and weather elasticities) A differential accounting for the impact of differences in projected versus actual “conditional variables” could still be approximated from the sensitivities. As before, extra-model, transitory influences on the historical values could also be accounted for. The basis for the approach is to detect the important market sensitivities implicit to NEMS’ representation of energy markets for each version of the model. This approach was implemented in support of the preparation of the methodology with respect to the effects of changes in the weather on residential and commercial sector consumption. The elasticities derived from these, stand alone solutions in comparison to a reference case (in this case the AEO2007 base case), provide a means to assess the impact of actual versus assumed weather for the period forecast by the model without having to archive the AEO2007 version of NEMS for later use. An illustration of accounting for weather using these elasticities is given below with details in Appendix A. In general, the construction of a set of NEMS solutions relative to a base case with variations specified to account for important influences is the preferred starting point for the evaluation methodology proposed here.
There is an alternative method for extracting the underlying sensitivities implicit to NEMS. The approach noted above would be to solve NEMS components, relative to a base or reference case, with each important assumption, e.g., the residential sector price of energy, changed, individually (with all other assumptions held constant), and compare results. The alternative method is to estimate the market sensitivities implicit to NEMS based upon the solutions prepared each year for the AEO. Among other ways, NEMS solutions are saved in a binary format and can be processed by the PC-based graphic interface Graf2000. Solution data have been saved in this format starting with those prepared for the 1998 AEO. This utility includes a regression component that enables regression analysis to be conducted using resident data; and, a data extraction routine that enables any collection of solution series to be extracted and input to other statistical procedures. Initially, this method entails no additional resource requirements in terms of running NEMS or archiving versions of NEMS for use at a future time.1 Instead, the solution sets for the AEO versions of NEMS can be pooled for the projections to be evaluated at a future time.
The basic approach is to specify the underlying energy market supply and demand relationships in terms of their important, explanatory variables; and, given this, to estimate the relationships based upon NEMS solution data. The results of the estimates provide a description of the NEMS model version in terms of how energy markets are represented. The actual specifications utilized would be guided by the expertise of the EIA staff responsible for developing and maintaining individual model components. Since the solution sets can be readily archived, the actual regression analyses need not be conducted until the time that a model version is to be evaluated, although the outcomes of the regressions can have immediate diagnostic use in NEMS development.2 A demonstration of the general success of representing NEMS components via regression analysis is given in: Buck and Lady, “Approximation of Large, Computer-Based Economic Models,” presented at annual meetings of International Atlantic Economic Association on October 9, 2005 in New York City, New York. A copy of the paper can be downloaded from the link:
http://optima-com.com/buck_lady/AES_Paper.htm
It is proposed to configure the means of performing the differential analyses, as based upon specialized NEMS runs or the regression analysis of NEMS solutions, in a fashion that can be routinely conducted and maintained by EIA staff. The goal of the statistical analysis is to enable the errors in EIA forecasts to be explicitly decomposed with respect to influences such as the following:
Transitory Influences, e.g., weather, strikes, accidents, embargoes not accounted for in the projections.(* for weather)
Institutional Influences, e.g., changes in laws and regulations and changes in data series definitions compared to model assumptions.
Structural Influences, e.g., changes in resource availability or energy use technology compared to model assumptions.*
Errors in Projecting Conditional Variables, e.g., differences in the eventual values of activity drivers and other exogenous factors such as GDP and population.*
Errors in Behavioral Parameters, e.g., changes in consumer price sensitivities compared to those assumed by the forecasting methodology.*
Uncertainty, e.g., the residual error of
the projection method.*
The methodology for partitioning forecast differences among (such as) the influences outlined above is as follows for the items indicated by “*”, given the availability of actual data for previously forecast values.
In the case of elasticities derived from specialized, NEMS runs, the results will be applied directly to the differences between the assumed and actual values of the conditional variables. This approach is outlined below in the case of accounting for the weather. In the case of the regression equations, the equations derived to represent the important relationships of supply and demand are re-run using the actual values for the explanatory variables. The actual values are substituted for the values used (or solved for) in the original projections, one explanatory variable at a time. This enables the identification of the influence of each explanatory variable separately. The equation is then re-run with all explanatory variables assigned their actual values. For this case, the residual error is due to general forecasting "uncertainty" or other, structural changes. Structural change is the issue of whether or not the values of the coefficients in the forecasting equation have changed for the forecast period compared to the model version to which the equation had been "fit." The equation is re-estimated and the results compared to the outcome of the estimation used to approximate the characteristics of the model. One method to assess if there were significant "differences" between the original, and revised, estimate is the Chow test (Chow, Gregory, "Tests of Equality Between Sets of Coefficients in Two Linear Regressions," Econometrica, 28, (July 1960), pp. 591-605.). Events influencing the actual data not accounted for by the forecasting equation will be identified and evaluated by EIA staff as appropriate.
NEMS design is intended to enable the differential assessment of changes in many influences on energy consumption and production. The present methodology for evaluating NEMS projections, e.g., as presented in:
http://www.eia.doe.gov/oiaf/analysispaper/forecasteval/index.html),
assesses the accuracy of projections
of seventeen aggregate variables, e.g., total energy consumption. A review of
this method was presented in: Winebrake, James and Denya Sakva (2006), "An
evaluation of errors in
III. An Example for Residential and Commercial
Sector Consumption of Delivered Energy. The example presented here is
intended to be illustrative and indicative of the methodologies proposed,
rather than a definitive evaluation of NEMS projections of residential and
commercial sector energy demand.
III.1 Weather Elasticities.
III.I.1 Sources: The effects of weather on energy consumption are estimated by comparing actual weather to “normal” weather using the measure Heating Degree Days (HDD) for the fall/winter season and Cooling Degree Days (CDD) for the spring/summer season. Given these measures, the weather effect is then derived by employing “weather elasticities.” These measures, however derived, represent the percentage change in energy consumption per one percent change in HDD or CDD from “normal” values. There is more than one source for both the HDD and CDD measures as well as the associated elastiticies. The sources used here are given below.
(1) Historical values for the HDD
and CDD measures, Tables 1.7 and 1.8, in the Annual Energy Review (DOE/EIA-0384(2005)) posted on the EIA website
(2) Historical values for the HDD and CDD measures in 2005, with estimated values for 2006 and 2007, Table 1, in the Short-Term Enery Outlook, December 12, 2006 release posted on the EIA website.
(3) Historical values for the HDD and CDD measures in 2001 – 2005 with estimated values for 2006 and “presumably normal” values for 2007-2030 in the AEO 2007 NEMS base case solution file aeo2007.1121a.ran.
(4) Weather elasticities for selected fuels for energy consumption in the residential and commercial sectors given in: Final Reduced Form Energy Model Elasticities from EIA's Regional Short-Term Energy Model (RSTEM) May 2006 (PDF file) derived from a comparative statics analysis of the Regional Short-Term Energy Model (RSTEM)..
(5) Weather elasticities for the
residential and commercial sector as derived from a comparison of the AEO2007
base case cited in (3) above to two special NEMS solutions prepared, ceteris paribus, with, HDD values
increased by 10% (rsaeo07.1205a.ran) and CDD values increase by 10%
(rsaeo07.1205b.ran).
III.I.2 HDD and CDD Measures. An assessment of the method of deriving HDD and CDD measures by each of the three sources cited above has not been done. Small differences can be seen in the measures as presented below.
Historical HDD (with estimates
for 2006 and 2007)
Year AER NEMS RSTEM
1995 4531 * *
1996 4713 * *
1997 4542 * *
1998 3951 * *
1999 4169 * *
2000 4460 * *
2001 4223 4157 *
2002 4284 4257 *
2003 4460 4432 *
2004 4290 4254 *
2005 4228 4272 4315
2006 * 4094 4124
2007 * 4370 4451
Normal 4524
Historical CDD (with estimates
for 2006 and 2007)
Year AER NEMS RSTEM
1995 1293 * *
1996 1180 * *
1997 1156 * *
1998 1410 * *
1999 1297 * *
2000 1229 * *
2001 1245 1287 *
2002 1393 1406 *
2003 1290 1313 *
2004 1232 1258 *
2005 1444 1421 1395
2006 * 1436 1381
2007 * 1293 1239
Normal
1215
III.I.3 NEMS Weather Elasticities. Weather elasticities were calculated by comparing the NEMS solutions cited in (5) above for the years 2007-2030. Three methods, roughly equivalent, were used to compute the value of the elasticity. Let HDD1 and HDD2 be, respectively, the values for the HDD measure in the NEMS base case and cold winter cases. Let CDD1 and CDD2 be defined correspondingly for the warm summer case. Let Q1 and Q2 be the consumption projections for a selected fuel for the base case and weather case. The three methods for computing the HDD elasticity were:
Arc Elasticity = %∆Q/%∆HDD =
((Q2-Q1)/(HDD2-HDD1))((HDD2+HDD1)/(Q2+Q1)).
The arc elasticity uses the average of the two cases for the percentage base.
Base % Elasticity = %∆Q/%∆HDD = ((Q2-Q1)/(HDD2-HDD1))(HDD1/Q1)).
The base % elasticity uses the
base case values as the percentage base.
Constant Elasticity.
The constant elasticity assumes that the functional relationship that describes
energy consumption as related to the HDD measure is given by Q = A(HDD)E,
where E is the elasticity and A accounts for all other influences on
consumption. For the base case and weather case the value of E is found by
solving the linear system:
Log(Q1) = log(A) +
E(Log(HDD1));
Log(Q2) = Log(A) +
E(Log(HDD2)).
Given this,
E =
dLog(Q)/dLog(HDD).
The elasticities with respect to
the CDD measure were found in the same way. Below are the elasticities from the
NEMS scenarios cited in (5) above for the year 2007 (for purposes of
illustration) compared to the RSTEM elasticities cited in (4).
Weather Elasticities
|
|
HDD |
CDD |
||
|
Fuel |
NEMS |
RSTEM |
NEMS |
RSTEM |
|
Residential: Liquids |
.601 |
* |
* |
* |
|
Natural Gas |
.554 |
.880 |
* |
-.01 |
|
Electricity |
.078 |
.180 |
.120 |
.263 |
|
Total Delivered |
.385 |
* |
.048 |
* |
|
|
|
|
|
|
|
Commercial: Liquids |
.269 |
* |
* |
* |
|
Natural Gas |
.451 |
.526 |
.01 |
-.017 |
|
Electricity |
.042 |
.015 |
.127 |
.110 |
|
Total Delivered |
.212 |
* |
.07 |
* |
III.I.4 Weather Impacts. For purposes of illustration, the NEMS elasticities above (for the year 2007) were applied to historical data for the period 1995-2005. The elasticity used was the “constant elasticity” case. Although there are small differences compared to NEMS (for the years that can be compared), the AER data cited above in (1) for actual and “normal” weather were utilized. The “weather impact” estimated is the degree to which actual consumption is different from “if normal weather” consumption. The method utilized is as follows for the elasticity E.
First, the value for “A” is computed in the formula:
Actual Q = A(Actual HDD or CDD)E.
Then, given “A,”
if normal Q = A(Normal HDD or CDD)E.
The weather impact or adjustment is then computed as,
adjustment = (if normal Q) – (actual Q).
For example, in 2005 actual delivered energy consumed by the residential sector was 11.6 quads. In that year, as measured by HDD, the fall winter season was only 93.46% as cold as a “normal” season. Using the corresponding HDD elasticity (=.3848), if weather had been “normal,” then the consumption of delivered energy would have been 11.9 quads. Accordingly, the weather impact, or adjustment, is .3 quads, i.e., consumption was lower by .3 quads due to warmer weather than it otherwise would have been had weather been as cold as “normal.” Accordingly, a NEMS projection of residential energy consumption in 2005 would be .3 quads high, due to the warmer actual winter compared to the assumed to be colder normal winter (the results are only indicative since the HDD and CDD assumptions for normal and actual weather for NEMS and the AER exhibit some differences).
In the tables below, the weather adjustments are presented for the residential and commercial sectors for delivered energy and selected fuels consumed. In the appendix following this section, the elasticities calculated from the solutions cited in (5) above are presented.
HDD-Related Adjustments
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Residential: Delivered
Energy
HDD Weather Elasticity = .3848
Year
Actual %d(HDD) Adjustment If
1995 10.46122 100.15 -.0062 10.455
1996 11.12847 104.18 -.1739 10.9546
1997 10.66744 100.4 -.0162 10.6512
1998 10.23326 87.33 .5474 10.7807
1999 10.64985 92.15 .3403 10.9901
2000 11.1721 98.59 .0614 11.2335
2001 10.91906 93.35 .2931 11.2122
2002 11.16996 94.69 .2367 11.4067
2003 11.52878 98.59 .0634 11.5922
2004 11.39384 94.83 .2353 11.6291
2005 11.59748 93.46 .3059 11.9034
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Commercial: Delivered
Energy
HDD Weather Elasticity = .2118
Year
Actual %d(HDD) Adjustment If
1995 7.329521 100.15 -.0024 7.3271
1996 7.594991 104.18 -.0656 7.5294
1997 7.77284 100.4 -.0065 7.7663
1998 7.65573 87.33 .2228 7.8785
1999 7.781741 92.15 .1359 7.9176
2000 8.170951 98.59 .0246 8.1956
2001 8.11114 93.35 .1192 8.2303
2002 8.21455 94.69 .0954 8.3099
2003 8.388131 98.59
.0254 8.4135
2004 8.398893 94.83 .095 8.4939
2005 8.461396 93.46 .1221 8.5835
CDD-Related Adjustments
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Residential: Delivered
Energy
CDD Weather Elasticity = .0483
Year
Actual %d(CDD) Adjustment If
1995 10.46122 106.42 -.0314 10.4298
1996 11.12847 97.12 .0157 11.1442
1997 10.66744 95.14 .0257 10.6931
1998 10.23326 116.05 -.0733 10.16
1999 10.64985 106.75 -.0336 10.6163
2000 11.1721 101.15 -.0062 11.1659
2001 10.91906 102.47 -.0129 10.9062
2002 11.16996 114.65 -.0736 11.0964
2003 11.52878 106.17 -.0333 11.4955
2004 11.39384 101.4 -.0076 11.3862
2005 11.59748 118.85 -.0963 11.5012
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Commercial: Delivered
Energy
CDD Weather Elasticity = .0702
Year
Actual %d(CDD) Adjustment If
1995 7.329521 106.42 -.0319 7.2976
1996 7.594991 97.12 .0156 7.6106
1997 7.77284 95.14 .0272 7.8
1998 7.65573 116.05 -.0795 7.5762
1999 7.781741 106.75 -.0356 7.7461
2000 8.170951 101.15 -.0066 8.1644
2001 8.11114 102.47 -.0138 8.0973
2002 8.21455 114.65 -.0785 8.1361
2003 8.388131 106.17 -.0352 8.3529
2004 8.398893 101.4 -.0082 8.3907
2005 8.461396 118.85 -.102 8.3594
III.2 Using Regression Results From NEMS Solutions. A method preferred to that outlined below would be to utilize elasticities from NEMS solutions specially designed to isolate the associated sensitivities. This was the method used for the weather elasticities above. When this is not possible, the regression approach outlined here can be readily applied to NEMS solutions. To illustrate the proposed method, solution data from the 1998-2000 AEO versions of NEMS were assembled. A simple relationship was used to represent each sector’s demand:
Qt = A + B(Sector Price)t + C(Driver)t + DQt-1,
where Qt = sectoral consumption of delivered energy in year t, (Sector Price)t = the average price of energy delivered to the sector in year t ($ per million Btu), and (Driver)t = a sectoral “activity” variable. The driver for the residential sector was millions of households and the driver for the commercial sector was billions of square feet. The lagged endogenous variable picks up secular trends in such as energy intensity, the average size of a household, and other infrastructure or behavioral characteristics. Since the AEO hi/lo GDP cases represent shifts in energy demand, to accommodate issues of identification, the regressions were performed on data pooled from the AEO Base Case and the hi/lo WOP cases. The regressions were run on the solution data for the years 2000-2020. Generally, the fit of the regressions was exceptionally good (regression results are in Appendix C below). Below are the results for the residential sector. The regression equations were then used with “actual” data (the solution values in the AEO 2006 NEMS solutions) for projections for the years 2000-2005 and the differences compared to the original forecast values. An example is provided below for the projections of residential consumption of (delivered) energy in the year 2005 as projected by the 2000 AEO version of NEMS.
Model
= N_ResAll_00 for Year = 20054
Residential
Energy Consumption: NEMS - Actual =-.28533 (as % =-2.44)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 13.02821 17.24 1.0614 9.08 -.277041 -.093418
Driver 111.3056 115.3573 -.2825 -2.42 .685312 .57847
Lag 11.33194 11.43521 -.016 -.14 .154078 .240553
HDD 4524 4228 .2932 2.51 .385 n/a
CDD 1215 1444 -.0949 -.81 .048 n/a
Total .9612 8.22
Uncertainty -1.2465 -10.67
The NEMS projection for 2005 was .28533 quads (or 2.44%) low as made for the 2000 AEO base case. The first five rows of the table evaluate the difference in the NEMS projection associated with actual, rather than assumed, values for each of the explanatory variables. The first three are from the regression equation while the assessment of the weather using heating degree days (HDD) and cooling degree days (CDD) was made independently (see section III.1 above and Appendix A). In each row the value assumed for the 2000 AEO is compared with the actual value in 2005. The impact, in quads, is the difference in the projected versus actual value associated with the explanatory variable. For example the average price of energy assumed for the 2000 AEO was over 24% lower than the actual price in 2005.3 As a result, given the estimated residential sector price elasticity of -.277, the forecast value figured to be 1.0614 (9.08%) high. In contrast, the NEMS assumption for residential households was 3.5% low with an estimated impact (from the regression approximation of residential sector demand) of 2.42%. When the lag and weather effects were added in, the NEMS projection figured to be 8.22% high. But in fact it was 2.44% low. The corresponding uncertainty, or “unexplained” variation is, therefore, -10.28%.5
This differential analysis of forecast differences was performed for each of the years 2000-2005. A table summarizing the outcome of the analysis for each year is provided below.
Residential Sector Projection % Differences: Model =
N_ResAll_00
Year
2000 2001 2002 2003 2004 2005
NEMS%
-2.04 .91 -.84 -2.61 -.9 -2.44
Adjustments%
1.8 4.57 1.16 1.08 4.27 8.22
Uncertainty%
-3.83 -3.66 -1.99 -3.69 -5.17 -10.67
Inspection of the table reveals
that NEMS projections were generally low for each year (with 2001 the
exception) while the backcast adjustments suggested that NEMS projections
should have been high. A more detailed analysis would accompany a regression
equation with more explanatory variables. One hint for the example used is that
the “updated” price elasticity, derived by running the same regression equation
for the 2006 AEO version of NEMS shows a significantly lower price elasticity.
As given above, a formal analysis of structural change across NEMS versions will
be conducted using the Chow Test, or related statistical procedure. A plot
summarizing the comparison of the NEMS projection compared to the backcast
projection is provided below.
In the above plot “NEMS SIM” is the projection provided by the statistical approximation of the 2000 AEO NEMS version residential sector demand function. The “Backcast” is that provided by the regression equation using actual values for the explanatory variables. As noted, the projected values tended to be low, but the impacts of errors in projecting the corresponding explanatory variables led to the estimate that the NEMS projection should have been high, i.e., the backcast projections should be low. These results imply the significance of features of the NEMS residential sector demand relationship not accounted for in the regression equation utilized for this example; and/or, other, extra-NEMS impacts upon consumption not accounted for in the illustration developed here.
Appendix A presents a summary of the results of the NEMS solutions prepared to isolate weather impacts. Appendix B presents the year by year workup for each of the approximations of NEMS residential and commercial sector demand for the 1998-2000 versions of NEMS. Appendix C provides the regression results for the sectoral demand approximations. Appendix D provides remarks by OIAF staff on the data entries in the AEO2007 NEMS solutions for the years 1995-2005 compared to historical values for these years.
Notes
1. A fair number of NEMS scenarios are run in support of each AEO, e.g., around forty were run for the AEO2006, in order to reveal important sensitivities and uncertainties. As the use of regression analysis of solution data is developed for the purposes proposed here, some number of additional runs might be formulated to facilitate the isolation of important influences for the ultimate evaluation of NEMS projections. As noted below, the formulation of weather scenarios might facilitate an accounting for weather effects as they cause actual variable values to deviate from forecast values.
2. Large changes in year to year implicit sensitivities could detect anomalies in model development.
3. The monetary units used for prices in the AEO lag the current period by two years. Accordingly, prices for the 2000 AEO are in $1998 while those in the 2006 AEO are in $2004. Actual prices for each AEO version are converted to the monetary units used in the contemporary AEO. The price index used to adjust monetary units to those used in each NEMS version was the “GDP Chain-Type Price Index,” as taken from Table B-3 (Appendix B) of the 2006 Economic Report of the President.
4. In the automated simulations of NEMS solutions supporting the examples presented here, “N_ResAll_##” stands for the regression-based approximation of NEMS residential sector energy demand with ## ranging from 98-00 for the 1998-2000 AEO version of NEMS. “N_ComAll_##” similarly stands for the approximation of NEMS commercial sector energy demand.
5. (Total Impact) + Uncertainty = NEMS – Actual.
Appendix A: Derivation of the Weather
Adjustment Impact Multiplier
The tables below give weather adjustments derived using the 2007 HDD and CDD weather elasticities as computed from a comparison of the AEO2007 base case and two custom NEMS solutions one with HDD increased by +10% and the other with CDD increased by +10%.
HDD-Related Adjustments
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Residential: Liquid Fuels
Subtotal
HDD Weather Elasticity = .6013
Year
Actual %d(HDD) Adjustment If
1995 1.38331 100.15 -.0013 1.382
1996 1.48805 104.18 -.0361 1.4519
1997 1.42806 100.4 -.0034 1.4247
1998 1.31383 87.33 .1115 1.4253
1999 1.47267 92.15 .0741 1.5468
2000 1.56307 98.59 .0134 1.5765
2001 1.53862 93.35 .0651 1.6037
2002 1.4625 94.69 .0487 1.5112
2003 1.53884 98.59 .0133 1.5521
2004 1.55437 94.83 .0504 1.6048
2005 1.53552 93.46 .0638 1.5993
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Residential: Natural Gas
HDD Weather Elasticity = .5538
Year
Actual %d(HDD) Adjustment If
1995 4.98352 100.15 -.0042 4.9793
1996 5.39054 104.18 -.1208 5.2697
1997 5.12495 100.4 -.0113 5.1137
1998 4.67104 87.33 .3638 5.0348
1999 4.85683 92.15 .2249 5.0817
2000 5.0998 98.59 .0404 5.1402
2001 4.90693 93.35 .1907 5.0976
2002 4.994331 94.69 .1531 5.1474
2003 5.229481 98.59 .0414 5.2709
2004 5.016002 94.83 .1497 5.1657
2005 4.984002 93.46 .1903 5.1743
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Residential: Electricity
HDD Weather Elasticity = .0781
Year
Actual %d(HDD) Adjustment If
1995 3.55701 100.15 -.0004 3.5566
1996 3.69353 104.18 -.0118 3.6817
1997 3.67091 100.4 -.0011 3.6698
1998 3.85594 87.33 .041 3.8969
1999 3.90649 92.15 .025 3.9315
2000 4.06864 98.59 .0046 4.0732
2001 4.098272 93.35 .0221 4.1204
2002 4.317502 94.69 .0184 4.3359
2003 4.345447 98.59 .0049 4.3503
2004 4.41363 94.83 .0184 4.432
2005 4.65656 93.46 .0246 4.6812
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Residential: Delivered
Energy
HDD Weather Elasticity = .3848
Year
Actual %d(HDD) Adjustment If
1995 10.46122 100.15 -.0062 10.455
1996 11.12847 104.18 -.1739 10.9546
1997 10.66744 100.4 -.0162 10.6512
1998 10.23326 87.33 .5474 10.7807
1999 10.64985 92.15 .3403 10.9901
2000 11.1721 98.59 .0614 11.2335
2001 10.91906 93.35 .2931 11.2122
2002 11.16996 94.69 .2367 11.4067
2003 11.52878 98.59 .0634 11.5922
2004 11.39384 94.83 .2353 11.6291
2005 11.59748 93.46 .3059 11.9034
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Commercial: Liquid Fuels
Subtotal
HDD Weather Elasticity = .2691
Year
Actual %d(HDD) Adjustment If
1995 .73206 100.15 -.0003 .7318
1996 .7513301 104.18 -.0082 .7431
1997 .70388 100.4 -.0008 .7031
1998 .66083 87.33 .0246 .6854
1999 .6613801 92.15 .0147 .6761
2000 .75635 98.59 .003 .7593
2001 .74168 93.35 .0138 .7555
2002 .68099 94.69 .0101 .6911
2003 .7708901 98.59 .003 .7739
2004 .758963 94.83 .0109 .7699
2005 .7722571 93.46 .0141 .7864
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Commercial: Natural Gas
HDD Weather Elasticity = .4506
Year
Actual %d(HDD) Adjustment If
1995 3.1165 100.15 -.0022 3.1143
1996 3.25094 104.18 -.0594 3.1915
1997 3.30639 100.4 -.0059 3.3005
1998 3.09815 87.33 .195 3.2931
1999 3.13194 92.15 .1175 3.2494
2000 3.25438 98.59 .0209 3.2753
2001 3.11164 93.35 .0981 3.2097
2002 3.22363 94.69 .0802 3.3038
2003 3.33061 98.59 .0215 3.3521
2004 3.226009 94.83 .0781 3.3041
2005 3.14611 93.46 .0974 3.2435
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Commercial: Electricity
HDD Weather Elasticity = .0421
Year
Actual %d(HDD) Adjustment If
1995 3.25205 100.15 -.0002 3.2518
1996 3.34397 104.18 -.0058 3.3382
1997 3.50285 100.4 -.0006 3.5023
1998 3.67799 87.33 .021 3.699
1999 3.76624 92.15 .013 3.7792
2000 3.95569 98.59 .0024 3.9581
2001 4.06351 93.35 .0118 4.0753
2002 4.1115 94.69 .0094 4.1209
2003 4.08484 98.59 .0025 4.0873
2004 4.194 94.83 .0094 4.2034
2005 4.32186 93.46 .0123 4.3342
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Commercial: Delivered
Energy
HDD Weather Elasticity = .2118
Year
Actual %d(HDD) Adjustment If
1995 7.329521 100.15 -.0024 7.3271
1996 7.594991 104.18 -.0656 7.5294
1997 7.77284 100.4 -.0065 7.7663
1998 7.65573 87.33 .2228 7.8785
1999 7.781741 92.15 .1359 7.9176
2000 8.170951 98.59 .0246 8.1956
2001 8.11114 93.35 .1192 8.2303
2002 8.21455 94.69 .0954 8.3099
2003 8.388131 98.59 .0254 8.4135
2004 8.398893 94.83 .095 8.4939
2005 8.461396 93.46 .1221 8.5835
Base Solution = aeo2007.1121a.ran
HDD+10% Weather Solution = rsaeo07.1205a.ran
CDD-Related Adjustments
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Residential: Electricity
CDD Weather Elasticity = .1191
Year
Actual %d(CDD) Adjustment If
1995 3.55701 106.42 -.0263 3.5307
1996 3.69353 97.12 .0129 3.7064
1997 3.67091 95.14 .0218 3.6927
1998 3.85594 116.05 -.0677 3.7882
1999 3.90649 106.75 -.0303 3.8762
2000 4.06864 101.15 -.0055 4.0631
2001 4.098272 102.47 -.0119 4.0864
2002 4.317502 114.65 -.0697 4.2478
2003 4.345447 106.17 -.0308 4.3146
2004 4.41363 101.4 -.0073 4.4063
2005 4.65656 118.85 -.0948 4.5618
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Residential: Delivered
Energy
CDD Weather Elasticity = .0483
Year
Actual %d(CDD) Adjustment If
1995 10.46122 106.42 -.0314 10.4298
1996 11.12847 97.12 .0157 11.1442
1997 10.66744 95.14 .0257 10.6931
1998 10.23326 116.05 -.0733 10.16
1999 10.64985 106.75 -.0336 10.6163
2000 11.1721 101.15 -.0062 11.1659
2001 10.91906 102.47 -.0129 10.9062
2002 11.16996 114.65 -.0736 11.0964
2003 11.52878 106.17 -.0333 11.4955
2004 11.39384 101.4 -.0076 11.3862
2005 11.59748 118.85 -.0963 11.5012
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Commercial: Natural Gas
CDD Weather Elasticity = .0096
Year
Actual %d(CDD) Adjustment If
1995 3.1165 106.42 -.0019 3.1146
1996 3.25094 97.12 .001 3.2519
1997 3.30639 95.14 .0016 3.308
1998 3.09815 116.05 -.0045 3.0937
1999 3.13194 106.75 -.0019 3.13
2000 3.25438 101.15 -.0004 3.254
2001 3.11164 102.47 -.0007 3.1109
2002 3.22363 114.65 -.0042 3.2194
2003 3.33061 106.17 -.0019 3.3287
2004 3.226009 101.4 -.0004 3.2256
2005 3.14611 118.85 -.0052 3.1409
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Commercial: Electricity
CDD Weather Elasticity = .1272
Year
Actual %d(CDD) Adjustment If
1995 3.25205 106.42 -.0257 3.2264
1996 3.34397 97.12 .0124 3.3564
1997 3.50285 95.14 .0222 3.5251
1998 3.67799 116.05 -.069 3.609
1999 3.76624 106.75 -.0311 3.7351
2000 3.95569 101.15 -.0058 3.9499
2001 4.06351 102.47 -.0126 4.0509
2002 4.1115 114.65 -.0709 4.0406
2003 4.08484 106.17 -.031 4.0538
2004 4.194 101.4 -.0074 4.1866
2005 4.32186 118.85 -.0939 4.228
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Commercial: Delivered
Energy
CDD Weather Elasticity = .0702
Year
Actual %d(CDD) Adjustment If
1995 7.329521 106.42 -.0319 7.2976
1996 7.594991 97.12 .0156 7.6106
1997 7.77284 95.14 .0272 7.8
1998 7.65573 116.05 -.0795 7.5762
1999 7.781741 106.75 -.0356 7.7461
2000 8.170951 101.15 -.0066 8.1644
2001 8.11114 102.47 -.0138 8.0973
2002 8.21455 114.65 -.0785 8.1361
2003 8.388131 106.17 -.0352 8.3529
2004 8.398893 101.4 -.0082 8.3907
2005 8.461396 118.85 -.102 8.3594
Base Solution = aeo2007.1121a.ran
CDD+10% Weather Solution = rsaeo07.1205b.ran
Appendix A
(Continued)
HDD
Elasticities
Table
#2 Energy Consumption by Sector and Source (quadrillion Btu, unless otherwise
noted)
Sector
and Source: Residential: Liquid Fuels Subtotal
year Arc Elasticity Base % Elas. Constant Elas.
2007 .6016 .5898 .6013
2008 .5947 .5829 .5944
2009 .5855 .5736 .5852
2010 .5781 .5661 .5778
2011 .5709 .5589 .5706
2012 .5641 .5521 .5638
2013 .5582 .5461 .5579
2014 .5525 .5405 .5522
2015 .5467 .5346 .5464
2016 .5402 .5281 .54
2017 .5353 .5231 .535
2018 .5293 .5171 .529
2019 .5219 .5097 .5216
2020 .5135 .5013 .5132
2021 .5037 .4915 .5034
2022 .4934 .4812 .4931
2023 .4825 .4703 .4822
2024 .4709 .4588 .4707
2025 .4592 .4471 .4589
2026 .4471 .4351 .4469
2027 .4348 .4229 .4346
2028 .4214 .4095 .4211
2029 .4064 .3947 .4062
2030 .3903 .3787 .39
Average .5126 .5006 .5123
Table
#2 Energy Consumption by Sector and Source (quadrillion Btu, unless otherwise
noted)
Sector
and Source: Residential: Natural Gas
year Arc Elasticity Base % Elas. Constant Elas.
2007 .5541 .542 .5538
2008 .5567 .5447 .5565
2009 .5568 .5447 .5565
2010 .5583 .5463 .5581
2011 .56 .5479 .5597
2012 .5616 .5496 .5613
2013 .5633 .5512 .563
2014 .566 .5539 .5657
2015 .5674 .5554 .5671
2016 .5705 .5585 .5702
2017 .5749 .5629 .5746
2018 .5782 .5663 .5779
2019 .5802 .5682 .5799
2020 .5809 .5689 .5806
2021 .5799 .568 .5796
2022 .5774 .5655 .5772
2023 .574 .5621 .5737
2024 .5697 .5577 .5694
2025 .5641 .552 .5638
2026 .557 .5449 .5567
2027 .5497 .5376 .5494
2028 .5404 .5283 .5402
2029 .5296 .5175 .5294
2030 .5171 .5049 .5169
Average .562 .55 .5617
Table
#2 Energy Consumption by Sector and Source (quadrillion Btu, unless otherwise
noted)
Sector
and Source: Residential: Electricity
year Arc Elasticity Base % Elas. Constant Elas.
2007 .0782 .0747 .0781
2008 .0817 .0782 .0817
2009 .0817 .0781 .0816
2010 .0824 .0788 .0823
2011 .0837 .08 .0836
2012
.0846 .0809 .0845
2013 .0853 .0815 .0852
2014 .0864 .0827 .0864
2015 .0874 .0835 .0873
2016 .0879 .084 .0878
2017 .0882 .0843 .0881
2018 .0884 .0846 .0884
2019
.0884 .0845 .0883
2020 .0882 .0844 .0882
2021 .0878 .084 .0877
2022 .0872 .0834 .0871
2023 .0864 .0826 .0863
2024 .0859 .0821 .0858
2025 .0842 .0805 .0841
2026
.0825 .0789 .0824
2027 .0808 .0773 .0807
2028 .0791 .0756 .079
2029 .0772 .0738 .0771
2030 .0754 .072 .0753
Average .0841 .0804 .084
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Residential: Delivered
Energy
year
Arc Elasticity Base %
Elas. Constant Elas.
2007 .3851 .3736 .3848
2008 .3853 .3738 .385
2009 .3823 .3708 .382
2010 .3806 .3691 .3803
2011 .3795 .3681 .3793
2012 .3784 .3669 .3781
2013 .3772 .3658 .377
2014 .3767 .3654 .3765
2015 .3756 .3642 .3753
2016 .3748 .3635 .3746
2017 .3746 .3632 .3744
2018 .3739 .3626 .3737
2019 .3724 .3611 .3722
2020 .3702 .3589 .3699
2021 .3672 .3559 .3669
2022 .3632 .352 .363
2023 .3587 .3475 .3584
2024 .3539 .3428 .3537
2025 .3481 .3371 .3479
2026 .3418 .3309 .3415
2027 .3354 .3246 .3351
2028 .3281 .3174 .3278
2029 .3199 .3094 .3197
2030 .3111 .3007 .3108
Average
.3631 .3519 .3628
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Commercial: Liquid Fuels
Subtotal
year
Arc Elasticity Base %
Elas. Constant Elas.
2007 .2693 .2598 .2691
2008 .2827 .2729 .2825
2009 .2751 .2655 .2749
2010 .2806 .2709 .2804
2011 .2815 .2717 .2813
2012 .2834 .2736 .2833
2013 .2821 .2724 .282
2014 .2861 .2762 .2859
2015 .2855 .2757 .2853
2016 .2849 .2751 .2847
2017 .2855 .2757 .2853
2018 .286 .2761 .2858
2019 .2865 .2766 .2863
2020 .2867 .2768 .2865
2021 .2869 .277 .2867
2022 .288 .2781 .2878
2023 .2882 .2783 .288
2024 .2895 .2795 .2893
2025 .2889 .279 .2887
2026 .29 .28 .2898
2027 .2909 .2809 .2907
2028 .2915 .2815 .2913
2029 .2912 .2813 .291
2030 .2926 .2826 .2924
Average
.2856 .2757 .2854
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Commercial: Natural Gas
year
Arc Elasticity Base %
Elas. Constant Elas.
2007 .4509 .4388 .4506
2008 .4494 .4373 .4491
2009 .4482 .4361 .4479
2010 .4489 .4368 .4486
2011 .4475 .4354 .4472
2012 .4463 .4343 .4461
2013 .4447 .4327 .4444
2014 .4442 .4322 .444
2015 .4397 .4277 .4394
2016 .4404 .4284 .4402
2017 .4387 .4267 .4384
2018 .4378 .4258 .4375
2019 .4371 .4251 .4368
2020 .4357 .4238 .4355
2021 .4342 .4223 .434
2022 .4324 .4205 .4322
2023 .4317 .4198 .4314
2024 .4307 .4188 .4305
2025 .4286 .4167 .4284
2026 .4261 .4143 .4259
2027 .4256 .4137 .4254
2028 .4232 .4114 .423
2029 .422 .4101 .4217
2030 .4206 .4087 .4203
Average
.4369 .4249 .4366
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Commercial: Electricity
year
Arc Elasticity Base %
Elas. Constant Elas.
2007 .0421 .0402 .0421
2008 .0405 .0387 .0405
2009 .0386 .0368 .0385
2010 .0389 .0371 .0388
2011 .0386 .0369 .0386
2012 .0376 .0359 .0376
2013 .0371 .0353 .037
2014 .0371 .0354 .037
2015 .0363 .0346 .0362
2016 .0356 .034 .0356
2017 .0348 .0332 .0348
2018 .0344 .0328 .0344
2019 .0341 .0325 .034
2020 .0335 .032 .0335
2021 .033 .0314 .0329
2022 .0325 .031 .0325
2023 .0329 .0313 .0328
2024 .0336 .032 .0335
2025 .0312 .0298 .0312
2026 .0311 .0297 .0311
2027 .0314 .03 .0314
2028 .0317 .0303 .0317
2029 .0312 .0298 .0312
2030 .0309 .0295 .0309
Average
.0349 .0333 .0349
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Commercial: Delivered
Energy
year
Arc Elasticity Base %
Elas. Constant Elas.
2007 .2119 .2039 .2118
2008 .2113 .2033 .2112
2009 .2092 .2012 .209
2010 .2099 .2019 .2097
2011 .2095 .2015 .2094
2012 .2087 .2007 .2085
2013 .2077 .1997 .2075
2014 .2074 .1995 .2073
2015 .2049 .1971 .2048
2016 .2043 .1965 .2041
2017 .2025 .1948 .2024
2018 .2014 .1937 .2013
2019 .2006 .1929 .2004
2020 .1993 .1916 .1992
2021 .198 .1903 .1978
2022 .1965 .1889 .1963
2023 .1956 .1881 .1955
2024 .195 .1874 .1948
2025 .1923 .1848 .1921
2026 .1908 .1834 .1906
2027 .1902 .1828 .19
2028 .1887 .1814 .1886
2029 .1873 .18 .1871
2030 .186 .1787 .1859
Average
.2004 .1927 .2002
CDD Elasticities
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Residential: Electricity
year
Arc Elasticity Base %
Elas. Constant Elas.
2007 .1191 .1141 .1191
2008 .1179 .1129 .1178
2009 .1174 .1124 .1173
2010 .1186 .1136 .1185
2011 .1207 .1156 .1206
2012 .1234 .1182 .1233
2013 .1266 .1213 .1265
2014 .1316 .1261 .1315
2015 .1365 .1309 .1364
2016 .1398 .1341 .1397
2017 .1444 .1384 .1442
2018 .1487 .1427 .1486
2019 .1518 .1456 .1517
2020 .1554 .1491 .1553
2021 .1586 .1522 .1585
2022 .1625 .156 .1624
2023 .165 .1584 .1649
2024 .1683 .1616 .1682
2025 .1707 .1639 .1706
2026 .1714 .1645 .1712
2027 .1727 .1658 .1725
2028 .1721 .1653 .172
2029 .1734 .1665 .1733
2030 .1731 .1662 .173
Average
.1475 .1415 .1474
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Residential: Delivered
Energy
year
Arc Elasticity Base %
Elas. Constant Elas.
2007 .0483 .0461 .0483
2008 .048 .0459 .048
2009 .0481 .0459 .0481
2010 .0491 .0468 .049
2011 .0502 .048 .0502
2012 .0517 .0494 .0517
2013 .0532 .0508 .0532
2014 .0557 .0532 .0557
2015 .0574 .0548 .0573
2016 .0594 .0567 .0593
2017 .0618 .0591 .0618
2018 .0643 .0614 .0642
2019 .0662 .0632 .0661
2020 .0683 .0652 .0682
2021 .07 .0669 .07
2022 .0721 .0689 .0721
2023 .0738 .0705 .0738
2024 .076 .0727 .076
2025 .0776 .0742 .0775
2026 .0781 .0746 .078
2027 .0792 .0757 .0791
2028 .0794 .0759 .0793
2029 .0805 .077 .0804
2030 .0809 .0773 .0808
Average
.0646 .0617 .0645
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Commercial: Electricity
year
Arc Elasticity Base %
Elas. Constant Elas.
2007 .1273 .122 .1272
2008 .1243 .1191 .1242
2009 .1225 .1174 .1224
2010 .1221 .117 .122
2011 .1205 .1154 .1204
2012 .1184 .1134 .1183
2013 .1159 .111 .1158
2014 .1152 .1103 .1151
2015 .1138 .109 .1138
2016 .1121 .1074 .1121
2017 .1109 .1062 .1108
2018 .1101 .1055 .1101
2019 .1089 .1042 .1088
2020 .1082 .1036 .1081
2021 .1068 .1022 .1067
2022 .1064 .1018 .1063
2023 .1059 .1013 .1058
2024 .1068 .1022 .1067
2025 .1045 .1 .1044
2026 .104 .0996 .104
2027 .1047 .1002 .1046
2028 .1045 .1 .1044
2029 .1044 .0999 .1043
2030 .1037 .0992 .1036
Average
.1117 .107 .1117
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Commercial: Natural Gas
year
Arc Elasticity Base %
Elas. Constant Elas.
2007 .0096 .0091 .0096
2008 .0092 .0088 .0092
2009 .0093 .0089 .0093
2010 .0101 .0097 .0101
2011 .0102 .0097 .0102
2012 .0104 .0099 .0104
2013 .0099 .0094 .0099
2014 .0107 .0102 .0107
2015 .0075 .0071 .0075
2016 .0105 .01 .0105
2017 .0099 .0095 .0099
2018 .0104 .0099 .0104
2019 .011 .0104 .0109
2020 .011 .0105 .011
2021 .0109 .0104 .0109
2022 .0108 .0103 .0108
2023 .0116 .011 .0116
2024 .0124 .0118 .0124
2025 .012 .0114 .012
2026 .0112 .0106 .0112
2027 .0125 .0119 .0125
2028 .012 .0114 .012
2029 .0125 .0119 .0125
2030 .013 .0124 .013
Average
.0108 .0103 .0108
Table #2 Energy Consumption by Sector and Source
(quadrillion Btu, unless otherwise noted)
Sector and Source:
Commercial: Delivered
Energy
year
Arc Elasticity Base % Elas. Constant Elas.
2007 .0703 .0672 .0702
2008 .0697 .0666 .0697
2009 .0681 .0651 .068
2010 .0683 .0653 .0683
2011 .0676 .0646 .0675
2012 .0667 .0637 .0666
2013 .0652 .0623 .0652
2014 .0654 .0625 .0653
2015 .0636 .0607 .0635
2016 .064 .0611 .0639
2017 .0634 .0606 .0634
2018 .0634 .0606 .0634
2019 .0631 .0603 .0631
2020 .063 .0602 .0629
2021 .0623 .0595 .0623
2022 .0623 .0595 .0623
2023 .0626 .0598 .0625
2024 .0636 .0608 .0636
2025 .0624 .0596 .0624
2026 .062 .0593 .062
2027 .0631 .0603 .0631
2028 .0631 .0603 .0631
2029 .0634 .0606 .0634
2030 .0634 .0606 .0634
Average .0646 .0617 .0645
Appendix B: Backcast Results And Workup
In the tables below, N_ResAll_## and N_ComAll_## refer to
NEMS solution data sets for the residential and commercial sectors for the 1998, 1999, and 2000 AEO versions of
NEMS, identified, respectively by ## = 98, 99, or 00.
Residential Sector Projection
% Differences: Model = N_ResAll_98
Year 2000 2001 2002 2003 2004 2005
NEMS% 1.83 5.06 3.05 1.08 2.67 .92
Adjustments% 4.04 5.93 3.52 2.72 5.43 8.72
Uncertainty% -2.21 -.87 -.48 -1.64 -2.75 -7.79
Model
= N_ResAll_98 for Year = 2000
Residential
Energy Consumption: NEMS - Actual = .20424 (as % = 1.83)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.45476 13.44 .2043 1.83 -.20489 -.093418
Driver 105.3386 105.189 .0075 .07 .475539 .57847
Lag 11.2109 10.66378 .1837 1.64 .333017 .240553
HDD 4524 4460 .0623 .56 .385 n/a
CDD 1215 1229 -.0063 -.06 .048 n/a
Total .4515 4.04
Uncertainty
-.2473 -2.21
Model
= N_ResAll_98 for Year = 2001
Residential
Energy Consumption: NEMS - Actual = .55232 (as % = 5.06)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.46056 14.48 .4188 3.83 -.20489 -.093418
Driver 106.4929 109.0325 -.127 -1.16 .475539 .57847
Lag 11.38 11.17576 .0686 .63 .333017 .240553
HDD 4524 4223 .3002 2.75 .385 n/a
CDD 1215 1245 -.0134 -.12 .048 n/a
Total
.6472 5.93
Uncertainty
-.0949 -.87
Model
= N_ResAll_98 for Year = 2002
Residential
Energy Consumption: NEMS - Actual = .34203 (as % = 3.05)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.37254 13.33 .1986 1.77 -.20489 -.093418
Driver 107.6285 110.472 -.1422 -1.27 .475539 .57847
Lag 11.47643 10.92411 .1854 1.65 .333017 .240553
HDD 4524 4294 .2298 2.05 .385 n/a
CDD 1215 1393 -.0761 -.68 .048 n/a
Total
.3955 3.52
Uncertainty
-.0535 -.48
Model
= N_ResAll_98 for Year = 2003
Residential
Energy Consumption: NEMS - Actual = .12474 (as % = 1.08)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.38588 13.98 .3306 2.87 -.20489 -.093418
Driver 108.7612 112.0135 -.1626 -1.41 .475539 .57847
Lag 11.55409 11.21206 .1148 1 .333017 .240553
HDD 4524 4460 .0636 .55 .385 n/a
CDD 1215 1290 -.0335 -.29 .048 n/a
Total
.313 2.72
Uncertainty
-.1883 -1.64
Model
= N_ResAll_98 for Year = 2004
Residential
Energy Consumption: NEMS - Actual = .30526 (as % = 2.67)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.30915 14.89 .5352 4.68 -.20489 -.093418
Driver 109.9131 113.6467 -.1867 -1.63 .475539 .57847
Lag 11.63279 11.50805 .0419 .37 .333017 .240553
HDD 4524 4290 .2376 2.08 .385 n/a
CDD 1215 1232 -.0078 -.07 .048 n/a
Total
.6202 5.43
Uncertainty
-.3149 -2.75
Model
= N_ResAll_98 for Year = 2005
Residential
Energy Consumption: NEMS - Actual = .10744 (as % = .92)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.31398 16.77 .9241 7.91 -.20489 -.093418
Driver 111.0799 115.3573 -.2139 -1.83 .475539 .57847
Lag 11.74047 11.43521 .1025 .88 .333017 .240553
HDD 4524 4228 .3033 2.6 .385 n/a
CDD 1215 1444 -.0982 -.84 .048 n/a
Total
1.0178 8.72
Uncertainty -.9104 -7.79
Residential Sector Projection
% Differences: Model = N_ResAll_99
Year 2000 2001 2002 2003 2004 2005
NEMS% .77 3.88 1.82 -.42 .97 -.98
Adjustments% 3.74 5.8 3.52 2.56 4.92 8.83
Uncertainty% -2.98 -1.92 -1.71 -2.99 -3.94 -9.8
Model
= N_ResAll_99 for Year = 2000
Residential
Energy Consumption: NEMS - Actual = .08649 (as % = .77)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.97353 13.66 .1601 1.43 -.24661 -.093418
Driver 105.0142 105.189 -.0062 -.06 .344404 .57847
Lag 11.12915 10.66378 .2101 1.88 .447881 .240553
HDD 4524 4460 .0616 .55 .385 n/a
CDD 1215 1229 -.0062 -.06 .048 n/a
Total .4194 3.74
Uncertainty
-.3329 -2.98
Model
= N_ResAll_99 for Year = 2001
Residential
Energy Consumption: NEMS - Actual = .4236 (as % = 3.88)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.957 14.72 .4113 3.77 -.24661 -.093418
Driver 106.1805 109.0325 -.1011 -.93 .344404 .57847
Lag 11.26225 11.17576 .039 .36 .447881 .240553
HDD 4524 4223 .2969 2.72 .385 n/a
CDD 1215 1245 -.0133 -.12 .048 n/a
Total .6328 5.8
Uncertainty
-.2092 -1.92
Model
= N_ResAll_99 for Year = 2002
Residential
Energy Consumption: NEMS - Actual = .20367 (as % = 1.82)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.85189 13.55 .1629 1.45 -.24661 -.093418
Driver 107.345 110.472 -.1109 -.99 .344404 .57847
Lag 11.34771 10.92411 .1912 1.71 .447881 .240553
HDD 4524 4294 .227 2.02 .385 n/a
CDD 1215 1393 -.0752 -.67 .048 n/a
Total
.3951 3.52
Uncertainty
-.1914 -1.71
Model
= N_ResAll_99 for Year = 2003
Residential
Energy Consumption: NEMS - Actual =-.04873 (as % =-.42)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.94609 14.22 .2972 2.58 -.24661 -.093418
Driver 108.525 112.0135 -.1237 -1.07 .344404 .57847
Lag 11.41573 11.21206 .0919 .8 .447881 .240553
HDD 4524 4460 .0627 .54 .385 n/a
CDD 1215 1290 -.033 -.29 .048 n/a
Total
.2951 2.56
Uncertainty
-.3438 -2.99
Model
= N_ResAll_99 for Year = 2004
Residential
Energy Consumption: NEMS - Actual = .11147 (as % = .97)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 13.00933 15.14 .4971 4.35 -.24661 -.093418
Driver 109.7295 113.6467 -.1389 -1.21 .344404 .57847
Lag 11.45932 11.50805 -.022 -.19 .447881 .240553
HDD 4524 4290 .2337 2.04 .385 n/a
CDD 1215 1232 -.0077 -.07 .048 n/a
Total
.5622 4.92
Uncertainty
-.4507 -3.94
Model
= N_ResAll_99 for Year = 2005
Residential
Energy Consumption: NEMS - Actual =-.11463 (as % =-.98)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 13.04349 17.05 .9347 8 -.24661 -.093418
Driver 110.9643 115.3573 -.1558 -1.33 .344404 .57847
Lag 11.54668 11.43521 .0503 .43 .447881 .240553
HDD 4524 4228 .2976 2.55 .385 n/a
CDD 1215 1444 -.0963 -.82 .048 n/a
Total
1.0305 8.83
Uncertainty
-1.1451 -9.8
Residential Sector Projection
% Differences: Model = N_ResAll_00
Year 2000 2001 2002 2003 2004 2005
NEMS% -2.04 .91 -.84 -2.61 -.9 -2.44
Adjustments% 1.8 4.57 1.16 1.08 4.27 8.22
Uncertainty% -3.83 -3.66 -1.99 -3.69 -5.17 -10.67
Model
= N_ResAll_00 for Year = 2000
Residential
Energy Consumption: NEMS - Actual =-.22746 (as % =-2.04)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 13.30314 13.81 .1277 1.14 -.277041 -.093418
Driver 105.3671 105.189 .0124 .11 .685312 .57847
Lag 10.70899 10.66378 .007 .06 .154078 .240553
HDD 4524 4460 .0599 .54 .385 n/a
CDD 1215 1229 -.006 -.05 .048 n/a
Total
.201 1.8
Uncertainty
-.4285 -3.83
Model
= N_ResAll_00 for Year = 2001
Residential
Energy Consumption: NEMS - Actual = .099 (as % = .91)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 13.17413 14.89 .4324 3.96 -.277041 -.093418
Driver 106.539 109.0325 -.1738 -1.59 .685312 .57847
Lag 10.9483 11.17576 -.0353 -.32 .154078 .240553
HDD 4524 4223 .2884 2.64 .385 n/a
CDD 1215 1245 -.0129 -.12 .048 n/a
Total
.4987 4.57
Uncertainty
-.3997 -3.66
Model
= N_ResAll_00 for Year = 2002
Residential
Energy Consumption: NEMS - Actual =-.09372 (as % =-.84)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 13.06647 13.7 .1597 1.42 -.277041 -.093418
Driver 107.7027 110.472 -.1931 -1.72 .685312 .57847
Lag 11.02311 10.92411 .0154 .14 .154078 .240553
HDD 4524 4294 .2211 1.97 .385 n/a
CDD 1215 1393 -.0732 -.65 .048 n/a
Total .1299 1.16
Uncertainty
-.2236 -1.99
Model
= N_ResAll_00 for Year = 2003
Residential
Energy Consumption: NEMS - Actual =-.30046 (as % =-2.61)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 13.06551 14.37 .3287 2.86 -.277041 -.093418
Driver
108.8805 112.0135 -.2184 -1.9 .685312 .57847
Lag 11.11834 11.21206 -.0146 -.13 .154078 .240553
HDD 4524 4460 .0613 .53 .385 n/a
CDD 1215 1290 -.0323 -.28 .048 n/a
Total .1247 1.08
Uncertainty
-.4252 -3.69
Model
= N_ResAll_00 for Year = 2004
Residential
Energy Consumption: NEMS - Actual =-.10327 (as % =-.9)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 13.08315 15.31 .5612 4.91 -.277041 -.093418
Driver 110.0828 113.6467 -.2485 -2.17 .685312 .57847
Lag 11.20759 11.50805 -.0467 -.41 .154078 .240553
HDD 4524 4290 .2294 2.01 .385 n/a
CDD 1215 1232 -.0076 -.07 .048 n/a
Total
.4878 4.27
Uncertainty
-.5911 -5.17
Model
= N_ResAll_00 for Year = 2005
Residential
Energy Consumption: NEMS - Actual =-.28533 (as % =-2.44)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 13.02821 17.24 1.0614 9.08 -.277041 -.093418
Driver 111.3056 115.3573 -.2825 -2.42 .685312 .57847
Lag 11.33194 11.43521 -.016 -.14 .154078 .240553
HDD 4524 4228 .2932 2.51 .385 n/a
CDD 1215 1444 -.0949 -.81 .048 n/a
Total
.9612 8.22
Uncertainty
-1.2465 -10.67
Commercial Sector Projection
% Differences: Model = N_ComAll_98
Year 2000 2001 2002 2003 2004 2005
NEMS% -4.28 -2.97 -3.49 -3.59 -1.48 -2.7
Adjustments% 4.95 4.7 2.84 2.1 3.18 3.84
Uncertainty% -9.23 -7.67 -6.33 -5.69 -4.67 -6.53
Model
= N_ComAll_98 for year = 2000
Commercial
Energy Consumption: NEMS - Actual =-.348716 (as % =-4.28)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.3309 13.18 .0395 .49 -.063328 -.132454
Driver 74.89623 68.70575 .364 4.46 .565757 .318224
Lag 7.722946 7.76739 -.0168 -.21 .374241 .649403
HDD 4524 4460 .0235 .29 .212 n/a
CDD 1215 1229 -.0063 -.08 .07 n/a
Total
.404 4.95
Uncertainty -.7527 -9.23
Model
= N_ComAll_98 for year = 2001
Commercial
Energy Consumption: NEMS - Actual =-.241118 (as % =-2.97)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.20863 14.43 .1035 1.27 -.063328 -.132454
Driver 75.71932 70.45488 .3095 3.81 .565757 .318224
Lag 7.803784 8.1525 -.1315 -1.62 .374241 .649403
HDD 4524 4223 .1142 1.41 .212 n/a
CDD 1215 1245 -.0135 -.17 .07 n/a
Total
.3822 4.7
Uncertainty
-.6233 -7.67
Model
= N_ComAll_98 for year = 2002
Commercial
Energy Consumption: NEMS - Actual =-.287438 (as % =-3.49)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.1045 13.38 .0594 .72 -.063328 -.132454
Driver 76.51044 72.17918 .2547 3.09 .565757 .318224
Lag 7.880362 8.12148 -.0909 -1.1 .374241 .649403
HDD 4524 4294 .0875 1.06 .212 n/a
CDD 1215 1393 -.0765 -.93 .07 n/a
Total
.2342 2.84
Uncertainty
-.5216 -6.33
Model
= N_ComAll_98 for year = 2003
Commercial
Energy Consumption: NEMS - Actual =-.299084 (as % =-3.59)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.05031 13.72 .0778 .93 -.063328 -.132454
Driver 77.33833 73.67056 .2157 2.59 .565757 .318224
Lag 7.957394 8.244832 -.1084 -1.3 .374241 .649403
HDD 4524 4460 .0242 .29 .212 n/a
CDD 1215 1290 -.0338 -.41 .07 n/a
Total .1756 2.1
Uncertainty
-.4747 -5.69
Model
= N_ComAll_98 for year = 2004
Commercial
Energy Consumption: NEMS - Actual =-.122073 (as % =-1.48)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 11.94473 14.24 .1069 1.3 -.063328 -.132454
Driver 78.19029 75.03583 .1855 2.25 .565757 .318224
Lag 8.03783 8.336914 -.1128 -1.37 .374241 .649403
HDD 4524 4290 .0909 1.1 .212 n/a
CDD 1215 1232 -.0079 -.1 .07 n/a
Total .2626 3.18
Uncertainty
-.3847 -4.67
Model
= N_ComAll_98 for year = 2005
Commercial
Energy Consumption: NEMS - Actual =-.2273 (as % =-2.7)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 11.83853 15.85 .1868 2.22 -.063328 -.132454
Driver 79.01656 76.20448 .1654 1.96 .565757 .318224
Lag 8.122841 8.244914 -.046 -.55 .374241 .649403
HDD 4524 4228 .1169 1.39 .212 n/a
CDD 1215 1444 -.0998 -1.18 .07 n/a
Total
.3233 3.84
Uncertainty
-.5506 -6.53
Commercial Sector Projection
% Differences: Model = N_ComAll_99
Year 2000 2001 2002 2003 2004 2005
NEMS% -2.84 -1.41 -1.82 -2.02 0 -1.31
Adjustments% -1.75 -1.09 -3.44 -3.92 -2.58 -.07
Uncertainty% -1.09 -.32 1.63 1.9 2.58 -1.24
Model
= N_ComAll_99 for year = 2000
Commercial
Energy Consumption: NEMS - Actual =-.231366 (as % =-2.84)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.57405 13.4 .104 1.28 -.172156 -.132454
Driver 62.89052 68.70575 -.2929 -3.59 .397353 .318224
Lag 7.820432 7.76739 .0287 .35 .536537 .649403
HDD 4524 4460 .0239 .29 .212 n/a
CDD 1215 1229 -.0064 -.08 .07 n/a
Total
-.1427 -1.75
Uncertainty
-.0887 -1.09
Model
= N_ComAll_99 for year = 2001
Commercial
Energy Consumption: NEMS - Actual =-.114683 (as % =-1.41)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.45059 14.67 .2794 3.44 -.172156 -.132454
Driver 63.5958 70.45488 -.3455 -4.25 .397353 .318224
Lag 7.921134 8.1525 -.1252 -1.54 .536537 .649403
HDD 4524 4223 .116 1.43 .212 n/a
CDD 1215 1245 -.0137 -.17 .07 n/a
Total
-.089 -1.09
Uncertainty -.0257 -.32
Model
= N_ComAll_99 for year = 2002
Commercial
Energy Consumption: NEMS - Actual =-.149878 (as % =-1.82)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.28535 13.6 .1655 2.01 -.172156 -.132454
Driver 64.25379 72.17918 -.3992 -4.84 .397353 .318224
Lag 8.006797 8.12148 -.0621 -.75 .536537 .649403
HDD 4524 4294 .0891 1.08 .212 n/a
CDD 1215 1393 -.0778 -.94 .07 n/a
Total
-.2846 -3.44
Uncertainty .1347 1.63
Model
= N_ComAll_99 for year = 2003
Commercial
Energy Consumption: NEMS - Actual =-.168493 (as % =-2.02)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.33296 13.95 .2036 2.44 -.172156 -.132454
Driver 64.93594 73.67056 -.4399 -5.28 .397353 .318224
Lag 8.094954 8.244832 -.0811 -.97 .536537 .649403
HDD 4524 4460 .0246 .3 .212 n/a
CDD 1215 1290 -.0343 -.41 .07 n/a
Total
-.3271 -3.92
Uncertainty
.1586 1.9
Model
= N_ComAll_99 for year = 2004
Commercial
Energy Consumption: NEMS - Actual = .000088 (as % = 0)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.34807 14.48 .2684 3.26 -.172156 -.132454
Driver 65.61822 75.03583 -.4743 -5.75 .397353 .318224
Lag 8.168421 8.336914 -.0912 -1.11 .536537 .649403
HDD 4524 4290 .0923 1.12 .212 n/a
CDD 1215 1232 -.008 -.1 .07 n/a
Total
-.2128 -2.58
Uncertainty
.2129 2.58
Model
= N_ComAll_99 for year = 2005
Commercial
Energy Consumption: NEMS - Actual =-.110758 (as % =-1.31)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.33938 16.12 .476 5.64 -.172156 -.132454
Driver 66.28476 76.20448 -.4996 -5.92 .397353 .318224
Lag 8.245002 8.244914 0 0 .536537 .649403
HDD 4524 4228 .1185 1.41 .212 n/a
CDD 1215 1444 -.1012 -1.2 .07 n/a
Total -.0063 -.07
Uncertainty
-.1045 -1.24
Commercial Sector Projection
% Differences: Model = N_ComAll_00
Year 2000 2001 2002 2003 2004 2005
NEMS% -4.16 -2.41 -2.67 -2.8 -.72 -1.8
Adjustments% -1.97 -2.56 -3.39 -3.8 -2.61 -.78
Uncertainty% -2.19 .15 .72 .99 1.9 -1.02
Model
= N_ComAll_00 for year = 2000
Commercial
Energy Consumption: NEMS - Actual =-.33897 (as % =-4.16)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 13.08415 13.55 .0281 .35 -.085136 -.132454
Driver 63.34077 68.70575 -.166 -2.04 .249884 .318224
Lag 7.710082 7.76739 -.0397 -.49 .686408 .649403
HDD 4524 4460 .0236 .29 .212 n/a
CDD 1215 1229 -.0063 -.08 .07 n/a
Total
-.1603 -1.97
Uncertainty
-.1787 -2.19
Model
= N_ComAll_00 for year = 2001
Commercial
Energy Consumption: NEMS - Actual =-.195572 (as % =-2.41)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.85606 14.83
.1192 1.47 -.085136 -.132454
Driver 64.20343 70.45488 -.1935 -2.38 .249884 .318224
Lag 7.81353 8.1525 -.2346 -2.89 .686408 .649403
HDD 4524 4223 .1148 1.41 .212 n/a
CDD 1215 1245 -.0135 -.17 .07 n/a
Total
-.2076 -2.56
Uncertainty
.012 .15
Model
= N_ComAll_00 for year = 2002
Commercial
Energy Consumption: NEMS - Actual =-.219833 (as % =-2.67)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.60261 13.75 .0693 .84 -.085136 -.132454
Driver 64.93169 72.17918 -.2243 -2.72 .249884 .318224
Lag 7.925908 8.12148 -.1354 -1.64 .686408 .649403
HDD 4524 4294 .0883 1.07 .212 n/a
CDD 1215 1393 -.0772 -.94 .07 n/a
Total
-.2793 -3.39
Uncertainty
.0595 .72
Model
= N_ComAll_00 for year = 2003
Commercial
Energy Consumption: NEMS - Actual =-.233694 (as % =-2.8)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.52698 14.1 .095 1.14 -.085136 -.132454
Driver 65.62161 73.67056 -.2491 -2.99 .249884 .318224
Lag 8.024999 8.244832 -.1522 -1.83 .686408 .649403
HDD 4524 4460 .0244 .29 .212 n/a
CDD 1215 1290 -.034 -.41 .07 n/a
Total
-.3159 -3.8
Uncertainty .0822 .99
Model
= N_ComAll_00 for year = 2004
Commercial
Energy Consumption: NEMS - Actual =-.05913 (as % =-.72)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.46017 14.64 .1316 1.6 -.085136 -.132454
Driver 66.34257 75.03583 -.269 -3.26 .249884 .318224
Lag 8.10322 8.336914 -.1618 -1.96 .686408 .649403
HDD 4524 4290 .0916 1.11 .212 n/a
CDD 1215 1232 -.008 -.1 .07 n/a
Total
-.2155 -2.61
Uncertainty
.1564 1.9
Model
= N_ComAll_00 for year = 2005
Commercial
Energy Consumption: NEMS - Actual =-.151769 (as % =-1.8)
Source NEMS Actual Impact Percent Elasticity Updated Elas
Price 12.3208 16.29 .2396 2.84 -.085136 -.132454
Driver 67.10787 76.20448 -.2815 -3.34 .249884 .318224
Lag 8.185784 8.244914 -.0409 -.49 .686408 .649403
HDD 4524 4228 .118 1.4 .212 n/a
CDD 1215 1444 -.1007 -1.19 .07 n/a
Total
-.0656 -.78
Uncertainty
-.0862 -1.02
Appendix C: Regression Results
Endogenous
Variable:
Table
#2 Energy Consumption by Sector and Source (Quadrillion Btu per Year, Unless
Otherwise Noted)
Sector and Source: Residential: Delivered Energy...........
Exogenous
Variables:
# 1)
Table #3 Energy Prices by Sector and Source (1996 Dollars per Million Btu)
Sector and Source: Residential....................:
# 2)
Table #4 Residential Sector Key Indicators and Consumption (Quadrillion Btu per
year, Unless otherwise noted)
Key Indicators and Consumption: Households (millions): Total......................
# 3)
Lagged Table #2 Energy Consumption by Sector and Source (Quadrillion Btu per
Year, Unless Otherwise Noted)
Sector and Source: Residential: Delivered Energy...........
Exogenous
Variable Mean Coefficient Elasticity t-statistic
Variable#
1 12.16087
-.20739 -.20489 -5.242288
Variable#
2 117.0589 .050005 .475539 5.686485
Variable#
3 12.20893 .335754 .333017 2.896522
Constant 4.878567
Endogenous Mean SER R-sq LR-Multiplier
Variable 12.30925 .030533 .997571 1.505466348310
Data
pooled for the years 2000 to 2020 for the solutions given below:
aeo98b.ran hwop98.ran lwop98.ran
Endogenous
Variable:
Table
#2 Energy Consumption by Sector and Source (Quadrillion Btu per Year, Unless
Otherwise Noted)
Sector and Source: Residential: Delivered Energy...........
Exogenous
Variables:
# 1)
Table #3 Energy Prices by Sector and Source (1997 Dollars per Million Btu)
Sector and Source: Residential....................:
# 2)
Table #4 Residential Sector Key Indicators and Consumption (Quadrillion Btu per
year, Unless otherwise noted)
Key Indicators and Consumption: Households (millions): Total......................
# 3)
Lagged Table #2 Energy Consumption by Sector and Source (Quadrillion Btu per
Year, Unless Otherwise Noted)
Sector and Source: Residential: Delivered Energy...........
Exogenous
Variable Mean Coefficient Elasticity t-statistic
Variable#
1 12.78631 -.233292 -.24661 -6.478882
Variable#
2 117.4799 .03546 .344404 6.662139
Variable#
3 12.0023 .45137 .447881 5.39698
Constant 5.49542
Endogenous Mean SER R-sq LR-Multiplier
Variable 12.09579 .024701 .998236 1.822722053114
Data
pooled for the years 2000 to 2020 for the solutions given below:
aeo99b.ran hwop99.ran lwop99.ran
Endogenous
Variable:
Table
#2 Energy Consumption by Sector and Source (Quadrillion Btu per Year, Unless
Otherwise Noted)
Sector and Source: Residential: Delivered Energy...........
Exogenous
Variables:
# 1)
Table #3 Energy Prices by Sector and Source (1998 Dollars per Million Btu)
Sector and Source: Residential....................:
# 2)
Table #4 Residential Sector Key Indicators and Consumption (Quadrillion Btu per
year, Unless otherwise noted)
Key Indicators and Consumption: Households (millions): Total......................
# 3)
Lagged Table #2 Energy Consumption by Sector and Source (Quadrillion Btu per
Year, Unless Otherwise Noted)
Sector and Source: Residential: Delivered Energy...........
Exogenous
Variable Mean Coefficient Elasticity t-statistic
Variable#
1 13.06701 -.252014 -.277041 -4.669263
Variable#
2 116.8374 .069721 .685312 5.080185
Variable#
3 11.78577 .155396 .154078 .95813
Constant 5.202166
Endogenous Mean SER R-sq LR-Multiplier
Variable 11.88658 .031538 .997149 1.183986815122
Data
pooled for the years 2000 to 2020 for the solutions given below:
aeo2k.ran hwop2k.ran lwop2k.ran
Endogenous
Variable:
Table
#2 Energy Consumption by Sector and Source (quadrillion Btu, unless otherwise
noted)
Sector and Source: Residential: Delivered Energy
Exogenous
Variables:
# 1)
Table #3 Energy Prices by Sector and Source (2004 dollars per million Btu)
Sector and Source: Residential:
# 2)
Table #4 Residential Sector Key Indicators and Consumption (quadrillion Btu)
Key Indicators and Consumption: Households (millions): Total
# 3)
Lagged Table #2 Energy Consumption by Sector and Source (quadrillion Btu,
unless otherwise noted)
Sector and Source: Residential: Delivered Energy
Exogenous
Variable Mean Coefficient Elasticity t-statistic
Variable#
1 17.20498 -.066236 -.093418 -5.26068
Variable#
2 122.5325 .05759 .57847 5.694203
Variable#
3 12.07313 .243057 .240553 1.942199
Constant 3.347288
Endogenous Mean SER R-sq LR-Multiplier
Variable 12.19881 .110425 .977698 1.321103438435
Data
pooled for the years 2000 to 2020 for the solutions given below:
aeo2006.1119a.rlp2006.1201a.rahp2006.1130a.ra
Endogenous
Variable:
Table
#2 Energy Consumption by Sector and Source (Quadrillion Btu per Year, Unless
Otherwise Noted)
Sector and Source: Commercial: Delivered Energy...........
Exogenous
Variables:
# 1)
Table #3 Energy Prices by Sector and Source (1996 Dollars per Million Btu)
Sector and Source: Commercial.....................:
# 2)
Table #5 Commercial Sector Key Indicators and Consumption (Quadrillion Btu per
year, Unless otherwise noted)
Key Indicators and Consumption: Total Floorspace(bill. sq. ft.): Total......................
# 3)
Lagged Table #2 Energy Consumption by Sector and Source (Quadrillion Btu per
Year, Unless Otherwise Noted)
Sector and Source: Commercial: Delivered Energy...........
Exogenous
Variable Mean Coefficient Elasticity t-statistic
Variable#
1 11.60786 -.046574 -.063328 -12.408086
Variable#
2 82.13966 .0588 .565757 28.645436
Variable#
3 8.4741 .377014 .374241 18.558989
Constant 1.052853
Endogenous Mean SER R-sq LR-Multiplier
Variable 8.536895 .003931 .999913 1.605172507889
Data
pooled for the years 2000 to 2020 for the solutions given below:
aeo98b.ran hwop98.ran lwop98.ran
Endogenous
Variable:
Table
#2 Energy Consumption by Sector and Source (Quadrillion Btu per Year, Unless
Otherwise Noted)
Sector and Source: Commercial: Delivered Energy...........
Exogenous
Variables:
# 1)
Table #3 Energy Prices by Sector and Source (1997 Dollars per Million Btu)
Sector and Source: Commercial.....................:
# 2)
Table #5 Commercial Sector Key Indicators and Consumption (Quadrillion Btu per
year, Unless otherwise noted)
Key Indicators and Consumption: Total Floorspace(bill. sq. ft.): Total......................
# 3)
Lagged Table #2 Energy Consumption by Sector and Source (Quadrillion Btu per
Year, Unless Otherwise Noted)
Sector and Source: Commercial: Delivered Energy...........
Exogenous
Variable Mean Coefficient Elasticity t-statistic
Variable#
1 11.94725 -.125904 -.172156
-10.667412
Variable#
2 68.93118 .050367 .397353 14.026527
Variable#
3 8.663424 .541122 .536537 17.853183
Constant 2.081833
Endogenous Mean
SER R-sq LR-Multiplier
Variable 8.737453 .008466 .999703 2.179228465953
Data
pooled for the years 2000 to 2020 for the solutions given below:
aeo99b.ran hwop99.ran lwop99.ran
Endogenous
Variable:
Table
#2 Energy Consumption by Sector and Source (Quadrillion Btu per Year, Unless
Otherwise Noted)
Sector and Source: Commercial: Delivered Energy...........
Exogenous
Variables:
# 1)
Table #3 Energy Prices by Sector and Source (1998 Dollars per Million Btu)
Sector and Source: Commercial.....................:
# 2)
Table #5 Commercial Sector Key Indicators and Consumption (Quadrillion Btu per
year, Unless otherwise noted)
Key Indicators and Consumption: Total Floorspace(bill. sq. ft.): Total......................
# 3)
Lagged Table #2 Energy Consumption by Sector and Source (Quadrillion Btu per
Year, Unless Otherwise Noted)
Sector and Source: Commercial: Delivered Energy...........
Exogenous
Variable Mean Coefficient Elasticity t-statistic
Variable#
1 12.22253 -.060376 -.085136 -10.431859
Variable#
2 69.99389 .030945 .249884 10.613828
Variable#
3 8.595922 .692154 .686408 30.985085
Constant 1.290166
Endogenous Mean SER R-sq LR-Multiplier
Variable 8.667882 .01009 .999555 3.248377435470
Data
pooled for the years 2000 to 2020 for the solutions given below:
aeo2k.ran hwop2k.ran lwop2k.ran
Endogenous
Variable:
Table
#2 Energy Consumption by Sector and Source (quadrillion Btu, unless otherwise
noted)
Sector and Source: Commercial: Delivered Energy
Exogenous
Variables:
# 1)
Table #3 Energy Prices by Sector and Source (2004 dollars per million Btu)
Sector and Source: Commercial:
# 2)
Table #5 Commercial Sector Key Indicators and Consumption (quadrillion Btu)
Key Indicators and Consumption: Total Floorspace (billion square Feet: Total
# 3)
Lagged Table #2 Energy Consumption by Sector and Source (quadrillion Btu,
unless otherwise noted)
Sector and Source: Commercial: Delivered Energy
Exogenous
Variable Mean Coefficient Elasticity t-statistic
Variable#
1 16.49366 -.073351 -.132454 -5.652417
Variable#
2 82.44836 .035254 .318224 4.778979
Variable#
3 8.997299 .659264 .649403 8.872552
Constant 1.505515
Endogenous Mean SER R-sq LR-Multiplier
Variable 9.133919 .085399 .990365 2.934823441021
Data
pooled for the years 2000 to 2020 for the solutions given below:
aeo2006.1119a.rlp2006.1201a.rahp2006.1130a.ra
Appendix D: Data Issues
An issue not addressed in the methodology above concerns the availability of “historical” data for the endogenous and exogenous variables in NEMS projections. In some cases year by year data for some conditional variables are simulated based upon modeling observations made less frequently. Some data are proprietary and acquiring all of the historical data may not at present have been budgeted for by EIA. A number of OIAF staff responded to a request to evaluate the availability of historical data as reported for the years 1995-2005 in the AEO 2007 NEMS solution files. Their responses are given below with respect to the main solution tables reported in the ftab solution report input to Graf2000.
NEMS Historical Data Availability For
ftab Tables 1 – 18, Years = 1995 - 2005
Summary
Skelly (
Table #1 Total Energy Supply and Disposition Summary (quadrillion Btu,
unless otherwise noted)
Skelly (
Table #2 Energy Consumption by Sector and Source (quadrillion Btu,
unless otherwise noted)
Skelly (
Table #3 Energy Prices by Sector and Source (2005 dollars per million
Btu, unless otherwise noted)
Skelly (
Table #4 Residential Sector Key Indicators and Consumption (quadrillion
Btu, unless otherwise noted)
Skelly (
Table #5 Commercial Sector Key Indicators and Consumption (quadrillion
Btu, unless otherwise noted)
Boedecker (
The NEMS representation of the commercial sector is based on the latest available EIA Commercial Buildings Energy Consumption Survey (CBECS), while floorspace growth is projected via equations in the macroeconomic module. The macroeconomic equations were developed using the proprietary F.W. Dodge commercial floorspace database. The latest Dodge update was received in late 2004, guaranteeing that the 2005 values are projections rather than "actual" values.
Given the proprietary nature of the Dodge data,
CBECS estimates are the only publicly available "actual" values for
Even using CBECS estimates presents
complications, due to discontinuities between surveys. Changes in survey
coverage and sample variability prevent a smooth transition when updating the
NEMS commercial floorspace from one CBECS survey to the next. This
is clearly illustrated in the "dynamic profile" for commercial energy
intensity (energy use per square foot) mentioned in your example (see below).
2005 Commercial Sector Energy Intensity (Thousand Btu per Sq. Ft.)
|
AEO Year |
1998 |
1999 |
2000 |
2006 |
|
Measure |
103.8 |
125.6 |
123.4 |
110.7 |
I
also have a slight qualification for the commercial table.
Table
5 Commercial fuel by end use - 2004 values represent projections one year after
the latest CBECS. That is, the latest CBECS provides estimates for 2003,
the first year the commercial module runs is 2004.
The
2004 values would be the closest to CBECS available from the restart file.
Table #6 Industrial Sector Key Indicators and Consumption
Honeycutt (
There is a similar problem with the industrial energy data. While the aeo2007 restart file has what we currently believe to be "correct" history values, previous aeo restart files also had data thought to be correct. You will notice very large changes in the aeo99 to aeo2001 period just because EIA's "history" changed (e.g., the "history" value for industrial biomass changed by 50 percent).
Table #7 Transportation Sector Key Indicators and Delivered Energy
Consumption
Skelly (
Maples (
1)
Non-EIA published historical transportation energy data (USDOT fuel
sales data collected for tax collection purposes) doesn't match EIA reported
energy data by mode/fuel type, which creates problems for us modelers
particularly when benchmarking to EIA published values. As EIA/DOT
addresses these discrepancies, I assume EIA reported historical energy use data
by sector will be updated.
2)
Historic vehicle travel and on-road fuel economy data published in
the AEO are taken from the FHWA Highway Statistics. As FHWA gains access
to new survey data, they often revise previously published historic data to
reflect the newly derived information. For example, for 2004 FHWA revised
both travel and on-road fuel economy estimates for light duty trucks and single
axle commercial trucks, which lowered the fuel economy and increased the travel
for light duty trucks while doing the opposite for the single axle commercial
trucks. FHWA has indicated that they will be updating some portion of the
time series data for light duty trucks, single axle commercial trucks, and
heavy trucks to reflect the updates made in the 2004 reported data.
3)
Not all historical data included in the ran file are updated
annually. Many data items in the transportation tables are derived from
purchased databases (that we can not afford to buy annually) and/or periodic
surveys (e.g. VIUS). Vehicle stocks currently reflect actual historic
data through 2001 for light duty vehicles, 2000 for heavy duty trucks, 2003 for
fleet vehicles, and 2005 for aircraft, the remaining historic years are model
estimates. The light, heavy, and fleet vehicle stocks will be updated for
the next AEO, which will change the historic values.
Table #8 Electricity Supply, Disposition, Prices, and Emissions (billion kilowatthours, unless otherwise noted)
Skelly (
Table #9 Electricity Generating Capacity (gigawatts)
Probably
not to be considered in the first round of NEMS forecast evaluations.
Table #10 Electricity Trade (billion kilowatthours, unless otherwise
noted)
Probably
not to be considered in the first round of NEMS forecast evaluations.
Table #11 Liquid Fuels Supply and Disposition (million barrels per day,
unless otherwise noted)
No
response.
Table #12 Petroleum Product Prices (2005 cents per gallon, unless
otherwise noted)
No
response.
Table #13 Natural Gas Supply, Disposition, and Prices (trillion cubic
feet, unless otherwise noted)
Skelly (
Table #14 Oil and Gas Supply
Benneche (
Table #15 Coal Supply, Disposition, and Prices (million short tons,
unless otherwise noted)
Skelly (
Mellish (
There
are some definitional issues that mean that our historical data differs from
numbers published in EIA data publications. One such exception is that we
publish an average minemouth price number that includes both captive and open
market coal mines, whereas, beginning in 2001, EIA data publications shifted to
publishing an average minemouth price number that excludes production from
captive coal mines.
Table #16 Renewable Energy Generating Capacity and Generation
(gigawatts, unless otherwise noted)
Skelly (
Table #17 Carbon Dioxide Emissions by Sector and Source (million metric
tons carbon dioxide equivalent, unless otherwise noted)
Skelly (
Table #18 Macroeconomic Indicators (billion 2000 chain-weighted
dollars, unless otherwise noted)
Skelly (
Unruh (