Klima-, Energi- og Bygningsudvalget 2013-14
KEB Alm.del Bilag 1
Offentligt
1285018_0001.png
Confidential
July 2013
|
E3grid2012
1
NRA specific report –
Summary Data e3grid2012 for Energinet.DK
Background and structure of report
Electricity transmission system operators are regulated by national and European
directives. Revenue allowances are set by national regulatory authorities (NRAs).
One task of NRAs in many countries is to assess that the regulated revenues are
based on efficient costs. Such analysis is often based on cost benchmarking
among network companies. Given the limited number of national transmission
system operators (TSOs) many European NRAs have decided to collaborate to
develop an international sample of comparator companies.
E3grid2012 is the project to benchmark the cost efficiency of a set of European
electricity transmission operators. The purpose of this NRA specific report is to
provide an overview of included company data as well as efficiency scores. In this
NRA specific report, we present more detailed information on the efficiency
scores and sensitivities of your company. We note that the information in this
NRA specific report is based on the analysis undertaken and documented in the
main e3grid2012 report
1
.
The summary is structured as follows:
Part A: DEA Scores
– The note provides information on the individual
efficiency score of your company as well as an indication of how your
company is positioned within the set of companies. In addition, we report
the efficiency scores of your company for variants of the model.
Part B: Data summary
– The note provides a summary of confidential data
submitted by your regulated company in e3grid2012 benchmarking and as
used for calculation of the efficiency scores. Further details on the
calculation of these costs are included in two separate excel files attached to
this NRA specific report.
Part C: Descriptive statistical analysis
– The note also provides some
descriptive statistical analysis of your regulated company in comparison to
other companies. This does not substitute for the actual formal
1
This NRA specific report does not include calculations based on data which were not included in
the e3grid2012 data set e.g. due to delay in data submissions. In addition sensitivity analysis, which is
not covered in the main report, e.g. returns to scale variants, are not included in this report. We note
that NRAs have the option to request such specific runs in a separate study.
NRA specific report –
Summary Data e3grid2012 for Energinet.DK
 PDF to HTML - Convert PDF files to HTML files
2
E3grid2012
|
July 2013
Confidential
benchmarking analysis, but provides further background for your regulated
company.
Part A: DEA efficiency scores – Summary
In the following, we provide information for your regulated company on
Base model results;
sensitivity analysis; as well as
dynamic results.
Base model results
In this section we describe the results for our base model for e3grid2012. The
final efficiency scores for the base model are influenced by various factors. In
order to make the impact of these factors visible we show the development of
the efficiency scores step by step starting from a simplified model.
Table 1
illustrates the development of the base model and the various steps. In addition
we describe the reasons and interpretation of the changes in the efficiency scores
for your regulated company for the single steps.
NRA specific report –
Summary Data e3grid2012 for Energinet.DK
 PDF to HTML - Convert PDF files to HTML files
1285018_0003.png
Confidential
July 2013
|
Frontier Economics
3
Table 1.
Development of base model
Model
Unit Cost (before
Call Z) excluding
outliers
2
Description
In the unit cost approach we simply compare the total costs (totex) of the TSO
with the technical assets reflected by the NormalisedGrid (Unit Cost =
Totex/NormalisedGrid)
Unit Cost scores may serve as a first rough indication on the cost position of your
regulated company with regard to the key cost driver NormalisedGrid
Unit Cost (after Call
Z) excluding outliers
In this step we illustrate the impact from the Call Z cost adjustments on the (unit)
costs of your regulated company compared to the other TSOs. The cost
adjustment from Call Z serves only as a compensation for TSO-specific costs
which are not otherwise reflected in the model specification (i.e. which are not
included following the regression based cost-driver analysis). Therefore, the
incremental impact of the call Z adjustment is relatively low
Nevertheless, for single TSOs the additional correction for cost impacts outside
the adjustments for densely-populated area and value of weighted angular
towers e.g. due to certain other topographical characteristics can be substantial
and can have an (improving) impact on efficiency scores
This measure may serve as an indication of the impact of TSO specific cost
adjustments from Call Z on the efficiency score of your regulated company
DEA NDRS
(NormalisedGrid)
without selected
capex break
In principle the Unit Cost approach can be described as a DEA with one single
output (NormalisedGrid) and an assumption of constant returns to scale
In this step we introduce the non-decreasing returns to scale (NDRS) approach
and calculate a DEA with only one output: NormalisedGrid
This measure – compared to the “Unit Cost (after Call Z) excluding outliers”
results – serves as an indication of the impact of company size on the efficiency
of your regulated company. If the score increases, then this means that size has
an impact on your regulated companies efficiency
DEA NDRS
(composite variable)
without selected
capex break
In this step we acknowledge that (i) NormalisedGrid although a key cost driver
does not explain all cost differences between companies and include two further
cost drivers reflection environmental factors: (ii) densely-populated area; and (iii)
share of angular towers
Cost-driver analysis indicated the relative average importance of these three
costs drivers. We use this information on the average importance of each cost
driver to create a composite variable made up of the weighted sum of
NormalisedGrid, densely populated area and value of weighted angular towers
This measure serves as an indication on the impact from densely-populated area
and share of angular towers on your regulated company based on the average
importance of these cost drivers over the whole sample
Base model without
selected Capex
break
In this step we recognise that the relative average importance of the three cost
drivers (NormalisedGrid, densely populated area and value of weighted angular
towers) may vary between the companies
Hence, we let the importance of the three cost drivers on the efficiency scores
vary within -50% and +50% of the statistical estimates for the respective
2
For further details on the model specifications we refer to: Frontier/Sumicsid/Consentec,
e3grid2012
- European TSO Benchmarking Study,
Section 7, July 2013.
NRA specific report –
Summary Data e3grid2012 for Energinet.DK
 PDF to HTML - Convert PDF files to HTML files
1285018_0004.png
4
E3grid2012
|
July 2013
Confidential
coefficient (cost driver)
This measure illustrates the importance of the two parameters covering
environmental influences from density and grid complexity on your regulated
company
We note that for those TSOs (and only for those TSOs) for which selected
Capex break has been applied this efficiency score should be considered
as the relevant one (and not the DEA score after Capex break)
Base model with
selected Capex
break
In e3grid2012 we introduced a further DEA outlier analysis – the so called
“selected Capex break” – to ensure that the efficiency frontier spanned by the
peer companies sets feasible cost targets that are not unduly influenced by the
absence of historic investment data
This measure illustrates to what extent your regulated company is affected by
this additional outlier analysis. The increase in efficiency indicates that the TSOs,
on which selected Capex break has been applied, influenced the efficiency score
of your regulated company
Source: Frontier/Sumicsid/Consentec
In
Table 2,
we illustrate the efficiency scores of your regulated company for the
different steps towards the development of the base model. We show the
development of your regulated company and the average efficiency. This allows
assessing the importance of the steps on the efficiency score of your regulated
company compared to the whole sample of compared TSOs.
Table 2.
Development of the base model
Model
Unit Cost (before Call Z) excl.
outliers
Unit Cost (after Call Z) excl.
outliers
DEA NDRS (NGrid) without
selective capex break
DEA NDRS (Composite
variable) without selective
capex break
Base model without capex
break
Your company's efficiency
89%
Average efficiency
61%
89%
62%
89%
64%
80%
77%
88%
84%
Base model
100%
86%
Source: Frontier/Sumicsid/Consentec
In addition DEA allows information on the weighting of the cost drivers that
determine the efficiency scores. The higher (lower) the weights for one cost
driver is the higher (lower) is the importance of this cost driver on the efficiency
score of your regulated company. For example, a weight of 50% for
NormalisedGrid indicates that the efficiency score of your regulated company is
NRA specific report –
Summary Data e3grid2012 for Energinet.DK
 PDF to HTML - Convert PDF files to HTML files
1285018_0005.png
Confidential
July 2013
|
Frontier Economics
5
driven by 50% from NormalisedGrid. We note that by including weight
restrictions we rule out that the efficiency score is only driven by one single cost
driver.
Figure 1
illustrates the output weights for your regulated company and allows
assessing the importance of the three cost drivers:
NormalisedGrid;
Densely-populated area; and
Value of weighted angular towers
for your regulated company.
Figure 1.
DEA input and output weights*
100%
90%
80%
NGTotex: Your company
70%
60%
Env.densely.populated.area:
Your company
50%
40%
30%
20%
10%
0%
Lines.share_totex.angle.vsum:
Your company
Source: Frontier/Sumicsid/Consentec
* excluding 1 outlier and 1 company with selective Capex break
Note: Companies are sorted in descending order of NGTotex.
Sensitivity analysis
In the following we discuss sensitivity analysis we have undertaken in relation to
the base model.
NRA specific report –
Summary Data e3grid2012 for Energinet.DK
 PDF to HTML - Convert PDF files to HTML files
1285018_0006.png
6
E3grid2012
|
July 2013
Confidential
Table 3.
Sensitivities of base model
Model
DEA NDRS
(unrestricted)
3
Description
In this sensitivity we relax the weight restriction applied in DEA analysis and
calculate a model without weight restrictions. In this model we let DEA determine
the importance of the three cost drivers (NormalisedGrid, densely-populated
area, and value of the weighted angular towers) endogenously. This may imply
that the efficiency score is mainly determined by one cost driver with relative low
importance (as suggested from the regression based cost-driver analysis)
This measure indicates the impact from the weight restrictions in the base model
on the efficiency score of your regulated company. An improvement in efficiency
indicates that in particular densely-populated area and/or the value of weighted
angular towers increase their weights in the calculation of efficiency
DEA NDRS (weight
restrictions based
on confidence
intervals)
In this sensitivity we introduce stricter weight restrictions compared to those
applied in the DEA base model. We use the upper/lower bounds from the
confidence intervals for the output parameters estimated from the cost-driver
analysis. The range of the upper/lower bounds lies inside the +/-50% range we
apply in the base model. This implies stricter weight restriction compared to the
base model, which should have an adverse effect in the efficiency scores
This measure indicates the impact from stricter weight restrictions on the
efficiency score of your regulated company. The size of the reduction of the
efficiency scores indicate the impact from reducing the +/-50% range from the
base model – in particular for densely-populated area and the value of weighted
angular towers – on your regulated company.
DEA NDRS (PPI)
PPI sensitivity
Opex efficiency
AdjTotex sensitivity
In this sensitivity we assess the impact from using the Producer Price Index
(PPI) instead of the CPI for indexation of the investment stream on the efficiency
scores in the base model
In this sensitivity we assess the efficiency scores only for Opex. We have done
so by adjusting the Totex figure by replacing the companies’ Capex by the
NormalisedGrid Capex. This allows focussing on the efficiency of the Opex by
using the same output parameters in the DEA model. As the change of costs
may also have an impact on the coefficients from the cost-driver analysis we
have adjusted the coefficients for the weight restriction, as well. We have kept
the range around the adjusted coefficients at +/-50%
This measure gives an indication on the Opex efficiency of your regulated
company. We note that the changes in the efficiency score compared to the base
model will be substantially affected by the adjustment of the coefficients
estimated from the cost-driver analysis, as the relative importance of densely-
populated area and value of the weighted angular towers decrease
Source: Frontier/Sumicsid/Consentec
In
Table 4
we illustrate the efficiency scores of your regulated company for the
sensitivities on the base model. We show the result for your regulated company
and the average efficiency. This allows assessing the impact form the sensitivities
on the efficiency score of your regulated company compared to the whole sample
of compared TSOs.
3
For further details on the model specifications we refer to: Frontier/Sumicsid/Consentec,
e3grid2012
- European TSO Benchmarking Study,
Section 7, July 2013.
NRA specific report –
Summary Data e3grid2012 for Energinet.DK
 PDF to HTML - Convert PDF files to HTML files
1285018_0007.png
Confidential
July 2013
|
Frontier Economics
7
Table 4.
Sensitivity analysis
Model
DEA NDRS (unrestricted)
Your company's efficiency
100%
Average efficiency
91%
DEA NDRS (weight restriction
confidence intervals)
100%
85%
PPI sensitivity (base model)
82%
84%
Adjusted Totex sensitivity (new
weights)
86%
86%
Source: Frontier/Sumicsid/Consentec
We have further undertaken so-called second stage analysis. The purpose of a
second stage analysis is to ensure that we have appropriately specified the best
model using the available data. The second stage analysis indicates that none of
the tested parameters serves as an additional explanatory for the identified
inefficiencies.
In addition, second-stage analysis allows assessing the
impact from size
on the
efficiency scores, e.g. to test if large companies tend to get a low efficiency score.
We used the NormalisedGrid as proxy for size and regressed the efficiency
scores from our base case on this variable. The results indicate that size has no
impact on the efficiency scores.
Unit cost scores – Opex and Capex information
Figure 2
illustrates the position of your regulated company compared to other
TSOs based on unit costs scores (Totex, with and without outliers). A score of
100% indicates that the company is fully efficient based on this measure.
NRA specific report –
Summary Data e3grid2012 for Energinet.DK
 PDF to HTML - Convert PDF files to HTML files
1285018_0008.png
8
E3grid2012
|
July 2013
Confidential
Figure 2.
Unit cost scores (without Capex break)
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
1
2
3
4
5
6
7
8
9
10
11
Unit Cost Score (excl. outlier): Your
company
Unit Cost Score (incl. outlier): Your
company
12
13
14
15
16
17
18
19
20
21
Source: Frontier/Sumicsid/Consentec
Note: Companies are sorted in descending order of unit cost score (excl. outlier).
Unit Cost scores can also be calculated for:
Unit cost Capex (Capex/Normalised Grid); or
Unit cost Opex (Opex/Normalised Grid).
This gives an indication on the position of your company with regard to Opex
and Capex (Figure
3)
and may serve as further indication of the Opex/Capex
efficiency in addition to the Opex efficiency illustrated above. A score of 100%
indicates that the company is fully efficient based on these measures.
NRA specific report –
Summary Data e3grid2012 for Energinet.DK
 PDF to HTML - Convert PDF files to HTML files
1285018_0009.png
Confidential
July 2013
|
Frontier Economics
9
Figure 3.
UC scores for Opex and Capex
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Unit Cost Score Capex (excl.
outlier): Your company
Unit Cost Score Opex (excl.
outlier): Your company
Source: Frontier/Sumicsid/Consentec
Note: Companies are sorted in descending order of unit cost score (excl. outlier).
Dynamic results
In the following we illustrate the cost development of your regulated company
compared to the other TSOs and your unit costs.
Opex unit costs
Unit Cost (Costs/NormalisedGrid) may serve as a first rough indication on the
cost level of your company.
Figure 4
shows the development of the Opex unit
costs (after correction for Call Z) from 2007-2011 for your regulated company
compared to the average (unweighted average) of other companies. Year 2007
4
is
used as a reference year [100%]. If the growth index is above (below) the blue
line the costs of your company grew more (less) compared to the other
companies. We note that these are relative developments of your regulated
company compared to the others, without taking into account the absolute level
of the starting point. Hence, if your regulated company starts from a high
absolute cost level compared to the other companies, it may still be inefficient
even when the relative decrease of unit costs is higher compared to the others.
4
Year 2007 or the earliest year available.
NRA specific report –
Summary Data e3grid2012 for Energinet.DK
 PDF to HTML - Convert PDF files to HTML files
1285018_0010.png
10
E3grid2012
|
July 2013
Confidential
Figure 4.
Development of Opex unit costs for 2007-2011 (without Capex break)
Your company
120%
Average
100%
100%
100%
89%
97%
90%
80%
82%
Unit Cost (Opex) index
80%
60%
60%
48%
40%
20%
0%
2007
2008
2009
2010
2011
Source: Frontier/Sumicsid/Consentec
Capex unit costs
Figure 5
shows the development of the Capex unit costs from 2007-2011 for
your regulated company compared to the unweighted average of other
companies. Year 2007 is used as a reference year [100%]. If the growth index is
above (below) the blue line the costs of your company grew more (less)
compared to the other companies. We note that these are relative developments
of your regulated company compared to the others, without taking into account
the absolute level of the starting point. Hence, if your regulated company starts
from a high absolute cost level compared to the other companies, it may still be
inefficient even when the relative decrease of unit costs is higher compared to the
others.
NRA specific report –
Summary Data e3grid2012 for Energinet.DK
 PDF to HTML - Convert PDF files to HTML files
1285018_0011.png
Confidential
July 2013
|
Frontier Economics
11
Figure 5.
Development of Capex unit costs for 2007-2011 (without Capex break)
Your company
120%
109%
104%
100%
93%
88%
80%
Average
111%
107%
100%
103%
100%
Unit Cost (Capex) index
60%
40%
20%
0%
2007
2008
2009
2010
2011
Source: Frontier/Sumicsid/Consentec
Part B: Data summary
The data summary is provided in several parts:
Asset summary; and
Benchmarking model data summary.
We note that a fuller data summary is attached in Excel format. The Excel file
also includes background calculations.
Asset summary
In
Table 5
we summarise the assets and towers data of your regulated company
provided by your regulated company in response to the Call X.
NRA specific report –
Summary Data e3grid2012 for Energinet.DK
 PDF to HTML - Convert PDF files to HTML files
1285018_0012.png
12
E3grid2012
|
July 2013
Confidential
Table 5.
Assets (piece count) and tower data from Call X
Asset (piece count)
Lines
Cables
Circuit ends
Transformers
Compensation devices
Series compensation
Control center
Towers
Other installation
Off-shore assets
Suspension
Angular
Wood
Steel
Guyed
Total
Source: Frontier/Sumicsid/Consentec
Unit
km
km
#
#
#
#
#
#
#
#
2,073
1,047
375
89
36
9
3
4,898
8
0
Towers
4,161
737
0
4,898
0
4,898
Unit
#
#
#
#
#
#
Model data Summary
In
Table 6
we summarise the model data for the benchmarking model applied in
the R2 report. The selected parameters are based on the cost-driver analysis and
the DEA calculations. The DEA calculations in the e3grid2012 report are based
only on the year 2011.
Total cost
– Total cost consists of the sum of Opex and Capex and is based
on the data your regulated company has provided in response to Call C.
5
NormalisedGrid
– NormalisedGrid is a key cost driver, as the physical
asset base is a natural driver for maintenance and investment costs.
6
Densely populated area
– Areas of high population density may require
more complex routing of transmission lines, combining of multiple circuits
on one tower in order to save land, etc. Hence, it is reasonable to assume
5
For a detailed description we refer to the Excel file “e3grid2012_R2_CAPEX_OPEX_Explanation”
for your company.
For a detailed explanation how this parameter is calculated, we refer to the Excel file “asset list_R2”
for your company.
6
NRA specific report –
Summary Data e3grid2012 for Energinet.DK
 PDF to HTML - Convert PDF files to HTML files
1285018_0013.png
Confidential
July 2013
|
Frontier Economics
13
that some cost impact is explained by densely populated areas, be it
alternatively or complementary to other parameters.
7
Value of weighted angular towers
– Angle towers are required whenever a
transmission line needs to deviate from a straight route. As angle towers
need to sustain higher (lateral) forces, they require more material and are
thus more costly. Therefore, the weighted share of angle towers can be
interpreted as a proxy parameter representing the cost impact of topography
or high population and/or load density.
8
Table 6.
Benchmarking model data summary
Name
Total Cost
NormalisedGrid
Densely-populated area
Value of weighted angular towers
Unit
€
€
km2
€
2011
101,654,326
181,305
509
5,481
2010
101,516,257
169,310
509
5,472
2009
104,378,873
147,209
509
5,472
2008
101,358,135
146,725
509
5,472
2007
Source: Frontier/Sumicsid/Consentec
Part C: Descriptive Analysis – Summary
In this section we provide a summary of
the operating scale of the companies in the dimension of the parameters
which are also used in the benchmarking model;
development of the cost base of the company; as well as
measures of complexity of grid structure.
Operating scale
In the following we provide some descriptive analysis of your company
compared to other companies in the sample.
Table 7
illustrates the values of
your company for the base model as well as key statistics about the sample of
TSOs in e3grid2012 for the year 2011.
7
For a detailed explanation how this parameter is calculated we refer to the Excel file
“e3grid2012_R2_calculation_of_density_area_assignment_of_NUTS”.
For a detailed explanation how this parameter is calculated we refer to the Excel file “asset list_R2”
for your company.
8
NRA specific report –
Summary Data e3grid2012 for Energinet.DK
 PDF to HTML - Convert PDF files to HTML files
1285018_0014.png
14
E3grid2012
|
July 2013
Confidential
Table 7.
Summary of data and comparison
Name
Total Cost
NormalisedGrid
Densely-populated area
Value of weighted angular towers
Unit
€
€
km2
€
Your company
Mean
101,654,326 334,187,238
181,305
354,505
509
5,206
5,481
39,072
Median
173,058,468
250,800
3,665
25,095
St.Dev
481,458,861
443,319
6,869
53,228
Source: Frontier/Sumicsid/Consentec
Complexity of grid structure
The complexity of grid structure matters as a more complex operating
environment may imply higher specific asset costs. Complexity dimensions
considered in the study include:
The complexity of network topology – in particular the use of angular
towers as opposed to suspension towers (and routes built in straight
lines)
9
; and
the population density in the serviced areas
10
.
These are dimensions which have been found to be relevant from an engineering
perspective and they are also found to be statistically significant cost drivers (see
e3grid2012 report).
Complexity of network topology
The share of angular towers and line length per tower gives an indication on the
complexity of the grid structure.
Less complex structure (upper left hand quadrant)
– A TSO located in
the upper left quadrant tends to be characterised by a less complex grid
structure. Companies in this quadrant are characterised by long overhead
lines between towers and towers built in straight routes.
More complex structure (lower right hand quadrant)
– By contrast, a
TSO located in the lower right quadrant tend be characterised by a complex
grid structure. The distance between towers is low and more angular towers
are used and routes are thereby less straight.
The other two quadrants indicate situations where a company faces complexity in
one dimension (e.g. few angular towers or short distance between towers), but
not the other.
9
For a detailed explanation how this parameter is calculated we refer to the Excel file “asset list_R2”
for your company.
For a detailed explanation how this parameter is calculated we refer to the Excel file
“e3grid2012_R2_calculation_of_density_area_assignment_of_NUTS”.
10
NRA specific report –
Summary Data e3grid2012 for Energinet.DK
 PDF to HTML - Convert PDF files to HTML files
1285018_0015.png
Confidential
July 2013
|
Frontier Economics
15
Figure 6
illustrates the position of your company compared to other TSOs with
regard to the complexity of the network topology.
We note that the figures serves as an indication for the grid complexity and
include the caveat that
the line length per tower may differ from the distance between towers
as it is based on circuit (not route) length data; and
the share of weighted angular towers is illustrated by a percentage
value
11
(which has to be differentiated from the value used in the model
calculations).
Figure 6.
Share of weighted angular towers and line length per tower
0.80
Overhead line per tower
Your company
0.40
0.00
0.0
0.2
0.4
Share of weighted angular towers
Source: Frontier/Sumicsid/Consentec
Population density
Areas of high population density may require more complex routing of
transmission lines, combining of multiple circuits on one tower in order to save
land, etc. Hence, it is reasonable to assume that some cost impact is explained by
densely populated areas, be it alternatively or complementary to other
parameters.
11
Value of weighted angular towers divided by the sum of Totex weighted lines.
NRA specific report –
Summary Data e3grid2012 for Energinet.DK
 PDF to HTML - Convert PDF files to HTML files
1285018_0016.png
16
E3grid2012
|
July 2013
Confidential
Densely-populated area
is defined by the size of the area with a population density of
more than or equal to 500 inhabitants/sqkm. In terms of geographic granularity
the NUTS3
12
regions from Eurostat have been used for the countries (or regions
within countries in cases where several TSOs operate in one country) included.
Figure 7
illustrates the share of densely-populated area for your company
compared to the other companies.
Figure 7.
Share of densely populated area
densely-populated area: Your company
100%
90%
80%
70%
60%
intermediate/thinly populated area: Your company
50%
40%
30%
20%
10%
0%
Source: Frontier/Sumicsid/Consentec
Note: Companies are sorted in descending order of densely-populated area.
Frontier Economics Limited in Europe is a member of the Frontier Economics network, which
consists of separate companies based in Europe (Brussels, Cologne, London & Madrid) and
Australia (Melbourne & Sydney). The companies are independently owned, and legal
commitments entered into by any one company do not impose any obligations on other
companies in the network. All views expressed in this document are the views of Frontier
Economics Limited.
12
Eurostat,
Regions in the European Union, NUTS 2006 / EU 27,
2007.
NRA specific report –
Summary Data e3grid2012 for Energinet.DK