Research into gas flexibility services A REPORT PREPARED FOR DTE

(1)

(2)

(3)

Research into gas flexibility services

Executive summary... 1

1 Introduction ...7

1.1 Objective of the study...7

1.2 Approach to the study...7

1.3 Structure of the report ...8

2 Overview of the Dutch gas sector ... 11

2.1 Gas flows ... 11

2.2 Gas trading and the balancing regime... 13

3 Market definition... 17

3.1 Demand for gas flexibility services... 17

3.2 Sources of flexibility ... 18

3.3 Market definition ... 22

3.4 Market definition conclusions... 26

4 Market structure ...27

4.1 Approach... 27

4.2 Capacity shares... 30

4.3 Pivot analysis ... 32

4.4 Sensitivity analysis... 40

4.5 Cost aspects of flexibility sources... 45

4.6 Impacts of policy changes ... 48

5 Conclusions ... 51

Annexe 1: Selected references ...53

Annexe 2: Market structure...55

(4)

ii Frontier Economics | May 2008

Tables & figures

Research into gas flexibility services

Figure 1: GTS high pressure Dutch gas transportation network... 13

Figure 2: Assessment of whether GasTerra is pivotal in the L-gas market, 2006 36 Figure 3: Assessment of whether GasTerra is pivotal in the H-gas market, 200637 Figure 4: Assessment of whether GasTerra is pivotal in the combined H-gas and L-gas market, 2006... 38

Figure 5: Assessment of whether GasTerra is pivotal in the combined L-gas and H-gas market in 2011... 40

Figure 6: Combined L-gas and H-gas pivot analysis 2011: reference (conservative) vs. sensitivity (lagged infrastructure)... 42

Figure 7: Combined L-gas and H-gas pivot analysis 2011: reference vs. sensitivity (cold winter) ... 44

Figure 8: Indicative merit order curve for L-gas flexibility in 2006 based on capacities... 46

Figure 9: Indicative merit order curve for H-gas flexibility in 2006 based on capacities... 47

Figure 10: Pivot analysis L-gas 2006... 58

Figure 11: Pivot analysis H-gas 2006... 60

Figure 12: Pivot analysis GTS System 2006 ... 61

Figure 19: Pivot analysis GTS system 2009... 70

Figure 20: Pivot analysis GTS system 2010... 71

(5)

Table 1: Dutch gas balance, 2006 ... 12

Table 2: Storages connected to GTS as of 2006... 20

Table 3: GasTerra’s capacity shares over time ... 31

Table 4: L-gas flexibility capacity for 2006 compared to Frontier’s 2005 Report 35 Table 5: Pivotal hours of GasTerra over time (rounded)... 39

Table 6: L-gas and H-gas flexibility capacity in 2011... 40

Table 7: L-gas flexibility capacity 2006... 55

Table 8: H-gas flexibility capacity 2006... 59

Table 9: L-gas and H-gas flexibility capacity 2006 ... 61

Table 10: Overview of peak demand forecast ... 62

Table 11: Production capacity assumptions ... 62

Table 12: Infrastructure development assumptions (in mcm/h) ... 63

Table 13: L-gas flexibility capacity 2009 to 2011 ... 66

Table 14: H-gas flexibility capacity 2009 to 2011 ... 68

Table 15: GTS System flexibility capacity 2009 to 2011 ... 70

Table 16: Capacity shares 2006 ... 75

(6)

(7)

Executive summary

In its method decision of November 2005 (regulating the method by which GTS is obliged to offer flexibility services), NMa observed that Gasunie Trade & Supply (now GasTerra) at that time had a dominant position in markets for flexibility services. The method decision was based on a report of Frontier Economics, Research into Gas Flexibility Services (dated March 2005, with a subsequent note updating the analysis in the report dated September 2005) and respondents’ comments to that report. The method decision became effective on 1 January 2006 and applies until the end of 2008.

NMa/DTe must decide whether a method decision is required to regulate flexibility services for the period after 2008. Within that framework it is necessary to establish whether after 2008 GasTerra has a dominant position in the market(s) for gas flexibility services. To this end, NMa/DTe has commissioned Frontier Economics to undertake a study into gas flexibility services covering the period 2009-2011. The period beyond 2011 is uncertain due to possible market developments and this study does not address dominance in that period.

This report presents Frontier Economics’ findings on dominance. These findings are based on analysis of public data and data of the gasmonitor 2006 plus careful consideration of respondents’ submissions to DTe’s consultation on regulation of flexibility services for gas (published 24 January 2008). Frontier presented its initial findings on dominance in the report, Research into gas flexibility

services – initial findings for market consultation, January 2008, which was published as part of DTe’s consultation.

We apply a three-step approach to the study:

• define the product and geographic dimensions of the relevant markets (Section 3);

• explore the structure of the relevant markets (Section 4); and

• draw conclusions about whether GasTerra is dominant in the markets for gas flexibility services (Section 5).

MARKET DEFINITION

The purpose of a market definition is to identify the competitive constraints faced by an undertaking, by assessing whether the undertaking enjoys economic power in relation to the goods or services it supplies. To identify market boundaries we apply the logic of the hypothetical monopolist (SSNIP) test. The result of this step of analysis is a definition of the relevant markets in two dimensions:

• a product dimension; and

(8)

2 Frontier Economics | May 2008

Executive summary

Before applying the SSNIP test, we set out how we measure the quantity of flexibility services, and explore characteristics (such as quantity, cost and control) of the potential sources of supply and demand for flexibility services relevant to the Netherlands.

In applying the logic of the SSNIP test to the product dimension, we draw the following conclusions:

• predictable and unpredictable flexibility do not form distinct markets;

• there is no clear cut separation between flexibility of different durations (e.g. hourly, daily, weekly and seasonal durations). Since the ability for different sources of flexibility to supply the segments of different durations varies, we assess the market structure for each segment separately;

• the H-gas and L-gas systems form, at least at times, distinct markets and that flexibility services may be transferred between the two markets to the extent and direction allowed by the capacity of quality conversion and the gas flow (before flexibility is applied to the flow). Future developments to quality conversion capacity and access rules could potentially alter this finding. Therefore, we also explore with our structural analysis the hypothesis that H-gas and L-gas flexibility services form part of the same market; and

• physical flexibility forms a distinct market from commercial plus physical flexibility. However, due to a lack of information, we consider only the market for physical flexibility services.

In applying the logic of the SSNIP test to the geographic dimension, we draw the conclusion that there are not regional markets for flexibility services within the Netherlands and that the markets for flexibility services are likely to extend beyond the borders of the GTS system to the extent allowed by physical cross-border capacity.

MARKET STRUCTURE

In considering whether GasTerra is dominant in one or more of the relevant markets for gas flexibility services, we investigate the market structure. We then use the results of the market structure analysis to draw conclusions about dominance.

We apply two quantitative approaches to assessing the structure of the relevant markets for gas flexibility services:

• analysis of capacity shares; and

• pivot analysis.

(9)

Capacity shares consider the potential of different providers of flexibility services to compete. Capacity shares unlike market shares do not reflect market outcomes. Therefore, we investigate the merit order (i.e. relative costs) for the provision of flexibility services to the relevant markets and use this to make inferences about the relationship between capacity shares and market shares in a competitive market. This relationship is important when using capacity share information to help form conclusions about dominance.

The structure of costs for the provision of flexibility services implies that in a competitive market, GasTerra’s market share would likely be higher than its capacity share in the L-gas market, the H-gas market and the combined H-gas and L-gas market.

Capacity shares do not provide any information about the extent to which there might be a capacity overhang or shortfall in the relevant market. For this reason we also undertake an assessment of whether GasTerra is pivotal in the relevant markets.

A pivotal position implies that GasTerra would be able to price flexibility freely as it would be guaranteed a certain market share due to lack of competing capacity. Therefore, a pivotal position combined with a high capacity share is a strong indication of dominance. In those markets for which GasTerra’s competitive market shares are likely to exceed its capacity share, a high capacity share could also be an indication of dominance even if GasTerra were not pivotal.

With the conservative reference case, GasTerra’s capacity shares in the L-gas market and the combined H-gas and L-gas markets are significantly higher than 50% over all years in the period 2009 to 2011, for all durations of flexibility services. GasTerra’s capacity shares in the H-gas market are above 50% in 2009, are above 50% in 2010 (with the exception of the seasonal segment) and fall below 50% in 2011. The trend for each of the L-gas, combined H-gas and L-gas, and H-gas markets is that GasTerra’s capacity share declines over time.

GasTerra’s capacity shares tend to be higher in segments with longer duration (i.e. seasonal flexibility), with the biggest difference being between the segments for hourly and daily flexibility. This trend is reversed for the seasonal segment of flexibility services in the H-gas market.

Even though there is excess capacity relative to the level of physical demand, we find that GasTerra is pivotal in the L-gas and the combined L-gas and H-gas markets in the conservative reference case throughout the period 2009 to 2011. The number of hours in a year for which GasTerra is pivotal declines over time. In the L-gas market it declines from 3100 hours in 2009 to 1200 hours in 2011. In the combined L-gas and H-gas market it declines from 3500 hours in 2009 to 800 hours in 2011. GasTerra is not pivotal under the conservative reference case for the H-gas market during the period 2009 to 2011.

(10)

Executive summary

The effect of a delay to infrastructure development is to increase GasTerra’s capacity shares for 2011 (relative to the conservative reference case) in the combined L-gas and H-gas market by between 5 and 9 percentage points and in the L-gas market by between 2 and 4 percentage points. The increase in GasTerra’s capacity shares in the H-gas market was similar to the increase in the combined L-gas and H-gas market.

The number of hours in 2011 for which GasTerra is pivotal also increases with a delay to infrastructure development. In the L-gas market it increases from 1200 to 1500 hours and in the combined L-gas and H-gas market it increases from 800 to 2500 hours. GasTerra remains not pivotal in the H-gas market in 2011.

A demand side driven sensitivity which shows the effect of a cold winter had no impact on capacity shares (since it is a demand side effect) but it increased the number of hours for which GasTerra is pivotal in the combined L-gas and H-gas market in 2011 from 800 to 1500.

ASSESSMENT OF DOMINANCE

The high capacity shares of GasTerra in the L-gas and combined L-gas and H-gas markets throughout the period from 2009 to 2011 in the conservative reference case raise concerns about dominance. The relatively low incremental cost of flexibility from Groningen is likely to mean that, if behaving competitively, GasTerra’s market shares in the L-gas and the combined L-gas and H-gas markets would also be high since they would in all likelihood exceed the capacity shares. The European Court of Justice would presume dominance, absent evidence to the contrary, if GasTerra’s market shares exceeded 50%. When we combine the capacity share results, our understanding of costs and the results from the pivot analysis, we find strong evidence that GasTerra is likely to be dominant in the L-gas and combined L-gas and H-gas markets throughout the period 2009-2011.

GasTerra’s high capacity shares in the H-gas market again raise concerns of dominance because of the low incremental cost of flexibility controlled by GasTerra in the H-gas market. The capacity shares therefore indicate concerns of dominance in the H-gas market in 2009 across all segments and in the hourly, daily and weekly segments in 2010. This is despite GasTerra being found not pivotal in the H-gas market. We find less evidence of dominance for 2011. Therefore, we conclude that if the relevant markets comprised the separate L-gas and H-gas systems that GasTerra is likely to be dominant in the L-gas market throughout the period 2009-2011 for all flexibility durations and dominant in the H-gas market in 2009 and in shorter duration segments in 2010. Conversely, if the relevant market comprised the combined L-gas and H-gas system that GasTerra is likely to be dominant in the combined L-gas and H-gas system throughout the period 2009-2011 for all flexibility durations.

(11)

(12)

(13)

1 Introduction

1.1 OBJECTIVE OF THE STUDY

In its method decision of November 2005 (regulating the method by which GTS is obliged to offer flexibility services), NMa observed that Gasunie Trade & Supply (now GasTerra) at that time had a dominant position in markets for flexibility services. The consequence of this observation was that, in accordance with Article 10a of the Gas Act, GTS as TSO must offer flexibility services. In turn, GasTerra must offer flexibility services at reasonable tariffs and conditions to GTS.

The method decision was based on a report of Frontier Economics, Research into

Gas Flexibility Services (dated March 2005, with a subsequent note updating the analysis in the report dated September 2005) and respondents’ comments to that report. The method decision became effective on 1 January 2006 and applies until the end of 2008.

To address market uncertainty, NMa/DTe must decide whether a new method decision is required to regulate flexibility services for the period after 2008. Within that framework it is necessary to establish whether after 2008 GasTerra has a dominant position in the market(s) for gas flexibility services. To this end, NMa/DTe has commissioned Frontier Economics to undertake a study into gas flexibility services covering the period 2009-2011. The period beyond 2011 is uncertain due to possible market developments and this study does not address dominance in that period.

This report presents Frontier Economics’ findings on dominance. These findings are based on analysis of public data and data of the gasmonitor 2006 plus careful consideration of respondents’ submissions to DTe’s consultation on regulation of flexibility services for gas (published 24 January 2008).1_Frontier

presented its initial findings on dominance in the report, Research into gas flexibility

services – initial findings for market consultation, January 2008, which was published as part of DTe’s consultation.

1.2 APPROACH TO THE STUDY

Demand for gas flexibility arises for two main reasons:

• to meet predictable variations in the gas demand and supply balance, for example, due to predictable variations over the course of a year/week/day; and

• to meet unpredictable variations in the gas demand and supply balance, for example, due to unpredicted temperature variations or temporary production problems over the course of a year/week/day.

(14)

Introduction

Gas flexibility services can provide an increase or decrease in gas supply with some notice period and for a period of time. The commercial and, to a lesser extent, the physical demand for flexibility are dependent not only on the variations described above but also on the gas network access and balancing arrangements. These rules, for example, determine over which periods gas imbalances are measured, which imbalance tolerances are acceptable and beyond which tolerance level an imbalance charge arises. In the Netherlands, transmission users’ gas imbalances are determined on a daily, four hourly and hourly basis. These arrangements drive a need for daily, within day and hourly gas balancing and hence use of flexibility by individual transmission users. Therefore, even if the system as a whole was net balanced, individual users would still require flexibility services.

Longer term flexibility services are also required in order to meet fluctuations in gas demand and supply that affect longer periods of time, e.g. 1 week, 1 month and 1 season.

An analysis of dominance requires a two step approach:

• define the product and geographic dimensions of the relevant markets; and

• explore the structure of the relevant markets.

In exploring the structure of the relevant markets when assessing dominance, we have followed legal advice received by DTe to look at the ability to raise prices. Cases of dominance are rarely clear cut. Therefore, in our study we gather a body of evidence that serves to form an opinion about the nature of the market and the position of GasTerra and other players in the market over the period 2009, 2010 and 2011.

Our analysis is conservative. This is a pragmatic approach to help understand the robustness of our analysis and to limit the need for debate as to whether intervention is required in the case that we find GasTerra dominant. This does not imply that DTe must legally base its method decision on conservative analysis as to whether GasTerra is dominant.

1.3 STRUCTURE OF THE REPORT

This document is the Final Report, which contains our views as to the market definition for gas flexibility and the position of GasTerra in the market. The remainder of this document is set out as follows:

Section 2 provides an overview of aspects of the Dutch gas industry relevant to this study.

Section 3 sets out our approach to defining the market and develops our conclusions about the extent of markets for gas flexibility.

Section 4 provides our view as to the structure of the relevant flexibility markets as defined in Section 3.

(15)

(16)

(17)

2 Overview of the Dutch gas sector

In this section we provide a brief overview of the Dutch gas sector, focussing on those aspects with specific relevance to this study.

2.1 GAS FLOWS

The Dutch natural gas industry has been operating since the discovery of the Groningen field in 1959. Over time, production from Groningen has declined in favour of gas production from other, smaller, fields. Groningen remains the single largest source of Dutch gas production with a 50% share of Dutch production in 2006.2 In addition, Groningen contains the bulk of Dutch developed gas reserves, with 947 bcm3_{(or 81%) out of a total of 1163 bcm for}

the country as at 1 January 2007.4

GasTerra is contracted to take 100% of the output from the Groningen field. Additionally, GasTerra is obliged to offer to take all gas produced from the small fields (i.e. the production fields other than Groningen). At the same time producers have the option to sell their gas to other shippers.5_{Nevertheless,}

GasTerra takes approximately 85 % of the small fields’ production.6

Gas from Groningen consists of low calorific gas (L-gas).7 H-gas is imported and both H-gas and L-gas are exported. Overall, the Netherlands is a net gas exporter. In addition, there is a net flow of gas from the H-gas to the L-gas system, which was approximately 25 bcm in 2006.8_{Table 1 sets out the Dutch}

gas supply and demand balance for 2006.

2 _{See TNO (2007), p.32.}

3 _{Gas volumes are in units of Groningen equivalent (Geq) volumes, i.e. 35.17 MJ/m}3_.

4 _{By including undeveloped gas fields, the total share of Groningen decreases slightly towards 75%,}

i.e. 1046 bcm out of a total of 1398 bcm. Source: TNO (2007), p.13.

5 _{Rules in Respect of the Transmission and Supply of Gas (Gas Act), Section 54.}

6 _{Calculation based on Nma/Dte (2007-2) and information from GasTerra’s website}

(www.gasterra.nl).

7 _{Several different qualities of gas are conveyed in GTS’ network. Low calorific gas comprises L-gas,}

G-gas and G+-gas (collectively referred to as L-gas). High calorific gas comprises H-gas.

(18)

Overview of the Dutch gas sector

H-gas (bcm Geq) L-gas (bcm Geq) Total (bcm Geq) Imports 23.3 - 23.3 Production 39.1 35.6 74.7 Quality conversion (from H-gas) - 25.1 - Total supply 62.4 60.7 98.0 Indigenous consumption 12.7 29.6 42.3 Exports 24.6 30.4 55.0 Quality conversion (to L-gas) 25.1 - - Total consumption 62.4 60.0 97.3

Table 1: Dutch gas balance, 2006

Source: NMa/DTe (2007-2) Note: figures may not add due to rounding, differences between storage levels at the start and end of the period, losses etc

GTS, the gas system operator, operates two high-pressure gas transportation systems in the Netherlands: an L-gas system and an H-gas system. These systems run in parallel, as illustrated in Figure 1, and are connected by 17 blending/nitrogen stations and two air separation units.9_{Blending and nitrogen}

injection allow conversion from H-gas to L-gas, but not in the opposite direction. All of the local distribution companies, which serve smaller customers, are connected to the L-gas system. Large customers take gas directly from either the L-gas or H-gas system. Since small customers tend to have higher variation in demand than large customers, the L-gas system has more variable demand than the H-gas system.

The GTS system is connected to gas systems in Belgium and Germany, to the Danish system though the offshore pipeline, Nogat, and to the Norwegian offshore network through Emden. In addition, the BBL pipeline, which allows flows from the GTS system to the UK, began commercial operations in December 2006.

9 _{In addition, Delta has a single gas blending facility that is dedicated to serving the Delta distribution}

(19)

Border point H-gas network

L-gas network

Figure 1: GTS high pressure Dutch gas transportation network

Source: GTS

There is also a small regional H-gas gas network operated by Zebra in the Netherlands. This network provides gas from the Bacton-Zeebrugge Interconnector via the Zelzate Zebra entry point (capacity 16.8 mcm/d in 2006)10

to customers in the south west of the Netherlands. This network is currently not directly connected to the GTS system.11 For this reason, we do not consider the Zebra system further as part of this study.12

2.2 GAS TRADING AND THE BALANCING REGIME

Since 1 January 2003, transmission tariffs have been applied on an entry-exit basis. Entry, exit and converter station capacity is booked on a first come first served basis. However, use it or lose it provisions are in place as an anti-hoarding measure. Capacity must be booked at least two days (in the case of daily contracts) prior to delivery. However, nominations of booked capacity are possible until 13:00, 1 day prior to flow and renominations are allowed up to 2 hours prior to flow. Secondary trading is allowed for all capacity products. Gas flows and booked capacities of each shipper must be in balance. However a tolerance is applied to each shipper, with the proportional size of the tolerance

10 _{Source: GTE (2005).}

11 _{We understand that a limited number of customers are connected to both the GTS and Zebra}

systems but that switching between the Zebra and GTS systems is not straightforward due to differences in gas quality.

12 _{Our treatment of Zebra is conservative. If we were to include the demand of customers connected}

(20)

Overview of the Dutch gas sector

dependant on the shipper’s portfolio size. Smaller shippers are granted a proportionally wider tolerance than large shippers. Also, the tolerance is dependent on temperature. During cold days less tolerance is available for the shippers. The tolerance is assigned for hourly, cumulative (four hours) and daily flows. If a shipper has an imbalance, e.g. his hourly offtake is greater than his hourly injection plus tolerance, he will be charged an imbalance penalty. This penalty is related to the day-ahead price of gas traded at the TTF.13

Additional tolerance may be purchased from GTS, tolerance may be traded between parties and parties may also trade gas at the Title Transfer Facility (TTF) in order to manage their individual gas balances. Contracts traded on TTF are defined using a day as the basic unit of time, with a flat delivery profile within the day.14

We understand that the long-term booking of quality conversion capacity is problematic for many market participants. Until the end of 2010 no capacity is available making it difficult for parties without conversion capacity and with an H-gas portfolio to gain access to sell L-gas. We understand also, that in practice the use it or lose it and secondary trading arrangements do not mitigate this concern.

TTF is the virtual trading point of the Netherlands and has developed well as a market place since Frontier’s 2005 Report.15 Nearly all indicators show an increase in traded volumes and numbers of trades and/or participants, with the exception of the L-gas trading segment. Liquid trading of L-gas does not exist, for example, only 2% of the total traded volumes on TTF relate to the three L-gas qualities (G, G+ and L). Almost all (99%) of the trades on TTF are made off exchange through brokered or non-brokered bilateral deals. Standard product trading on the energy exchanges, APX or Endex, plays only a small role.

The traded volumes on TTF in 2006 accounted for 6.5% of the yearly gas supply (demand plus exports) and the churn rate was 11.3, which means that each gas unit was re-traded about 11 times. Although trading has increased over time, it remains low compared to more mature trading markets like Britain’s NBP or Henry Hub in the US.

Most of the trading at TTF has focussed on forward contracts. In this segment, TTF is the second most important marketplace for gas in Europe, after NBP. TTF is therefore becoming a gas price benchmark for North-West Europe. This is in part due to a lack of a representative German virtual trading point, and the tendency for Zeebrugge to more reflect the gas price situation of the British market.

Some issues remain with TTF as a trading hub. In particular, the lack of L-gas liquidity and the fact that the bulk of supply never enters the TTF. Several changes to the gas system are expected that are likely to improve the functioning

13 _{See www.gastransportservices.nl.}

14 _{Within day contracts, with an hour as the basic unit of time, may also be traded on TTF. In}

practice, within day contracts are rarely traded.

(21)

(22)

(23)

3 Market definition

In this section we first describe the demand for flexibility services before identifying the potential sources of gas flexibility services to the Netherlands. We then apply the hypothetical monopolist test to different possible markets for gas flexibility services in both a product and geographic dimension. Finally, we draw conclusions about the definition of the relevant markets for gas flexibility services in the Netherlands.

3.1 DEMAND FOR GAS FLEXIBILITY SERVICES Physical demand

The reliable operation of the gas network requires that the system operator, GTS, maintains pressure in the gas transportation network within certain limits. This in turn requires that the injections of gas onto the network are balanced with the offtakes of gas from the network over short periods of time, e.g. over a few hours.

The latter implies that the injection rate of gas onto the system must change with the rate of gas offtake. This requires a supply of gas flexibility, i.e. the ability to change the rate at which gas is injected onto the system. As mentioned previously, longer term flexibility services are also required in order to meet fluctuations in gas demand and supply that affect longer periods of time, e.g. 1 week, 1 month or 1 season.

As described in Section 1, demand for flexibility for the system as a whole is derived from predictable and unpredictable variations in demand16_{or supply. If}

the variation in demand is predictable, the sources of flexibility could be coordinated to perfectly match the demand variation. However, if the demand for gas is unpredictable, sources of flexibility must also be able to be called upon in case demand for gas is greater or less than anticipated.

The overall system demand for flexibility services is what we term “physical” demand for flexibility services. The physical demand for gas flexibility services would come from GTS, who is ultimately responsible for maintaining system security in its role as system operator, and shippers who are incentivised to maintain their imbalances within certain tolerances.17

16 _{Export consumption is included in the analysis along with Dutch consumption because together}

they comprise the quantity demanded to be met by supply from Dutch producers and from imports. To exclude export consumption from the demand side of the analysis would understate the quantity demanded and result in an underestimate of the level of pivotality. Most of exports are done by GasTerra under the order of long-term contracts. However, as conditions are under a regular short-termed renegotiation process and point of delivery is located outside the Netherlands, GasTerra has access to flexibility derived from these volumes. Nevertheless we consider the possibility of re-imports of these volumes back to the Dutch market by foreign import companies, but as long as delivery point is not TTF those volumes are constraint by import (and storage) capacities. See Annexe 1 for detailed information how re-imports are included in the analysis.

17 _{Shippers have a legal requirement to meet the demand of their small customers for a daily}

(24)

Market definition

Commercial demand

In addition to the overall system requirements to balance injections and offtakes, individual shippers also have incentives to balance their own injections and offtakes due to their obligations under the Transmission Services Conditions. The aggregate shipper demand for flexibility services is what we term “commercial” demand for flexibility services.

In a perfectly liquid market for trading flexibility services with perfect foresight, the commercial demand for flexibility services would equal the physical demand for flexibility services. Absent a perfect market, the commercial demand for flexibility services will exceed the physical demand.

A final point to note is that a physical source of flexibility services could supply either physical or commercial demand for flexibility. However, a purely commercial source of flexibility cannot supply physical demand for flexibility. For example, TTF does not provide physical flexibility per se. Rather TTF facilitates the reallocation of flexibility between system users.

Measures of flexibility

As described in Frontier’s 2005 Report, the demand for flexibility can best be described in terms of the flow rate of gas (mcm/h) and in terms of the volume of flexibility (mcm) able to be delivered over some period of time, e.g. one week, one month etc.

Since neither method is necessarily better than the other as a tool to assess dominance, we apply both measures of flexibility services for this study. When considering capacity shares of different market segments for flexibility services, we employ the approach of estimating the volume of flexibility able to be delivered over a period of time. We therefore consider the difference between the maximum rate of flow over the time period and the average rate of flow over the time period (i.e. “maximum to average”).18_{We employ the flow rate and load}

duration curve in order to assess the proportion of time for which GasTerra is pivotal to meet demand.

3.2 SOURCES OF FLEXIBILITY

We briefly describe key characteristics of each potential source of flexibility service to the Netherlands.

Linepack

The high-pressure transport pipeline network has some ability to store gas, called linepack. GTS uses linepack to meet a short term (e.g. over a few hours) imbalance between supply and demand on its network. Linepack is supplied by

18 _{We have considered other measures of flexibility volumes but find they provide no additional}

(25)

the H-gas system and by the L-gas system. Based on GTS figures linepack amounts to about 14 mcm in the L-gas system and 4.7 mcm in the H-gas system.19

Linepack has the ability to respond rapidly to flexibility requirements although it may supply only low volumes of gas. Typically, linepack would be consumed and replenished on a daily cycle.

Providing linepack is a low cost source of flexibility available to the system operator as only 0.5%20_{of the gas volume is lost in running the compressor}

stations to increase linepack volumes.

In effect shippers acquire linepack from GTS through the tolerance band for imbalances. Since the tolerance band is related to flows, the quantity of linepack provided by GTS to shippers is higher in winter than it is in summer. This is because gas flows are higher in winter than in summer. During very high flow periods, the ability for the system to provide linepack is reduced. For this reason, the tolerance band is reduced during periods of very low temperatures. GTS allows spare tolerance to be traded between shippers.

Storage

Storage can be used to provide flexibility by taking gas off the GTS system and injecting it into the storage facility and by withdrawing gas from the storage facility and entering it onto the GTS system.21

The feasible instantaneous flow rates of gas to and from a storage facility are defined by its injection and withdrawal rate, respectively. The volume of flexibility provided by a storage facility is determined by its usable storage volume, i.e. working gas volume. The key characteristics of storage facilities are summarised in Table 2.22

19 _{See GTS (2006). These volumes are those theoretically available. Due to technical reasons and a}

need to hold reserves, total system volume and, in particular, the hourly flexibility capacity is likely to be lower. See Annexe 1 for further explanations.

20 _{Estimate from Frontier’s 2005 Report.}

21 _{In the context of storage, we use the term “injection” to describe gas entering the storage facility}

and “withdrawal” to describe gas being taken out of the storage facility.

22 _{All capacities and volumes are expressed in Groningen equivalent. Therefore all H-gas data is}

(26)

Market definition

Location Gas quality Type of storage Owner Production capacity [mcm/h Geq] Working gas [mcm Geq]

Grijpskerk H-gas depleted field NAM 2.70 1770

Norg L-gas depleted field NAM 2.13 3000

Alkmaar L-gas depleted field BP23 a.o. 1.50 500

Maasvlakte L-gas LNG Gasunie 1.30 72

Kalle

(GER) H-gas aquifer RWE 0.47 254

Epe (GER) L-gas salt cavern Essent 0.40 186

Table 2: Storages connected to GTS as of 2006

Source: Companies’ websites.

The LNG storage at Maasvlakte is able to respond very quickly to flexibility requirements, i.e. within an hour. However, the cost of using the facility is high with around 20% of the gas volume lost through the injection and withdrawal cycle24_{plus any transportation and entry/exit charges to the GTS system. We}

understand that GTS currently contracts for the entire capacity of the LNG storage facility and holds it as a reserve to meet demand during very cold spells. The underground gas storage sites respond less quickly than the LNG facility, requiring as much as 10 hours to attain the maximum injection rate from idle at Grijpskerk. When on stand-by, response rates are around 1 hour or less. The short run marginal cost associated with using underground storages is in the range of 2% of the value of the gas injected and then withdrawn, plus any transportation and entry/exit charges to the GTS system. This is higher than the short run cost of using linepack or Groningen to provide flexibility services but significantly lower than the short run cost of using the LNG storage.

The use of gas storage may be constrained by withdrawal and injection seasons. In addition, cavern storage could potentially be used for more than one full injection and withdrawal cycle within a season.

23 _{In 2007 Taqa, a company from the United Arabian Emirates, acquired BP’s Dutch upstream}

business including the share in Alkmaar.

(27)

Production swing

The Groningen field is a major source of production swing. Groningen is an L-gas field and comprises the vast bulk of L-L-gas production in the Netherlands.25

Production from the small fields is largely H-gas.

Through its legal obligations under the Gas Act, GasTerra controls the bulk of the production swing (100% of the gas produced from Groningen and 85% of the gas production from the small fields).

The flow rate of gas production can be changed within an hour to the extent allowed by the physical production capacity. The short run marginal cost of varying production is low and is largely related to the gas required to heat the gas as its pressure drops in flowing from the production field and into the GTS system.

The volume of flexibility able to be delivered by a production field is not constrained by working gas volume as it would be for, say, a storage facility.26 This is one of the reasons why Groningen is the key source of seasonal flexibility in the Netherlands.27

Quality conversion

The gas blending stations and nitrogen injection plants allow gas to be converted from H-gas to L-gas, but not the reverse. In practice, gas flexibility can be transferred in two directions since the hourly demand for gas in the H-gas and L-gas systems is relatively poorly correlated. An increment in L-L-gas flexibility and a decrement in H-gas flexibility can be provided if the quality converter stations are ramped up. Conversely, a decrement in L-gas flexibility and an increment in H-gas flexibility can be provided if the quality converter stations are ramped down. Technically, converter stations can respond quickly (within an hour) to a need for a change in flow rate.

Interruptible customers

Interruption of gas consumption is an asymmetric source of flexibility supply in that it can be used to reduce excess gas demand but not to address excess supply. Interruptible contracts are most appropriate for the provision of short-term needs for flexibility and would not be particularly appropriate for longer periods.

25 _{In this report, we do not differentiate between L-gas production from the small fields and L-gas}

production from Groningen.

26 _{However, there is a political constraint on Groningen production. Production from Groningen is}

restricted to a maximum of 425 bcm over the ten year period starting from 2006. Source: www.minez.nl.

27 _{In the long run, the pressure in the Groningen field is expected to decline due to decreasing}

(28)

Market definition

Typically interruption clauses are included in delivery contracts for power station and large industrial customers as well as in export contracts. However only limited information is available about whether these clauses are used in practice.28 Imports

Flexibility services may also be imported from sources physically located outside the Netherlands.29

Some underground storage facilities located in Germany are directly connected to the GTS system. Other sources of flexibility are connected to a gas network adjacent to the GTS network and can therefore deliver flexibility services to the GTS systems to the extent allowed by the lesser of (i) the flexibility capacity of the facility itself and (ii) the capacity of the border point between the GTS system and the neighbouring system.

The cost of delivering flexibility from sources that are not directly connected to the GTS network include the underlying cost of operating the flexibility source, the cost of entry/exit onto the system to which the source is directly connected, and the cost of entry/exit between the GTS and neighbouring systems.

TTF

The title transfer facility (TTF) is a virtual delivery point that allows shippers to reassign gas contracts. In this respect TTF does not provide physical flexibility. Rather, TTF helps to allow physical flexibility to be reassigned among shippers and is therefore a source of commercial flexibility.

3.3 MARKET DEFINITION

3.3.1 Approach

The purpose of a market definition is to identify the competitive constraints faced by an undertaking, by assessing whether the undertaking enjoys economic power in relation to the goods or services it supplies. In particular the market definition serves as a pre-requisite for identifying “those actual competitors of the

undertakings involved that are capable of constraining their behaviour and preventing them from behaving independently of an effective competitive pressure”.30

The relevant market must be defined in two dimensions:

28 _{We discuss this later in the Annexe.}

29 _{Following the logic of Frontier’s 2005 Report, we focus on countries that are directly connected to}

the Dutch system. In contrast to the 2005 Report, we include the UK as a possible source of flexibility due to the recent commissioning of the BBL pipeline between UK and the Netherlands. For 2006, we do not take this connection into account as the BBL pipeline began operating in December 2006. We do not consider countries that are not physically connected to the Netherlands – while they might supply commercial flexibility services to the Netherlands through swaps, countries not physically connected to the Netherlands are unlikely to supply physical flexibility services to the Netherlands.

30 _{European Commission Notice on the definition of the relevant market for the purposes of}

(29)

• a product dimension; and

• a geographic dimension.

To identify market boundaries it has become common practice to apply, as far as is practicable, the hypothetical monopolist (SSNIP)31_{test. The SSNIP test asks}

whether a hypothetical monopolist with control over the candidate market would be able profitably to impose a small but significant and non-transitory increase in price (above the competitive price level). The price increase used is typically 5%-10%. The test is applied first to a narrow definition of the market and then, if the test is not met, the market definition is broadened progressively in terms of product definition and geographic scope until such a price increase by a hypothetical monopolist appears both feasible and profitable. In this section we apply the logic of the SSNIP test in order to define the relevant markets.

3.3.2 Product market

Based on the characteristics of the Dutch gas sector, the key issues in defining the relevant product markets for flexibility services are:

• whether markets for predictable and unpredictable demand for flexibility are part of the same market or are separate markets;

• whether markets defined by different durations of time period are part of the same or separate markets, e.g. hourly, daily, weekly and seasonal flexibility;

• whether there are separate markets for L-gas and H-gas or an integrated market; and

• whether separate markets may be defined for physical flexibility and purely commercial flexibility.

Predictable and unpredictable flexibility

We consider whether the same sources of flexibility are available to serve predictable and unpredictable demand, or at least a sufficient proportion of sources can serve either demand, so as to form a credible chain of substitution. Within a very short notice period (e.g. of substantially less than 1 hour) linepack is the only source of flexibility services able to respond. Therefore, sources of flexibility other than linepack could not meet demand for unpredictable flexibility on very short timescales. All sources could however meet demand for predictable flexibility on very short timescales.

For longer timescales, all sources of flexibility could to some extent meet both predictable and unpredictable flexibility demand. The notice period to provide flexibility would not constrain the ability of a source to meet unpredictable demand. Nor do we believe that other factors would be a practical constraint on substitution. For example, gas storage could only meet unpredictable flexibility

(30)

Market definition

demand to the extent that gas had previously been injected into the storage facility. In reality a shipper is likely to hold some gas in storage specifically to respond to the possibility of an unpredictable demand for flexibility.

Outside the very short timescales dominated by linepack, essentially all of the relevant forms of flexibility have an adequately short response time. Therefore, we do not see an operational reason to distinguish markets for predictable and unpredictable demand.

Flexibility requirements of different durations

We consider whether sources providing flexibility over one duration could also be used to provide flexibility over another duration.

Flexibility provided over a short duration could in part meet the flexibility needs over a longer duration. For example, gas withdrawn at a storage facility over the period of a week could help to meet the high demand for gas over the winter season. In contrast, flexibility provided over a long duration cannot meet the flexibility needs over a shorter duration. For example, a constant flow rate of gas withdrawn from a storage facility over the period of a week would not help to meet variations in the demand for gas within that week, e.g. over a day.

However, any attempt to raise price in the daily segment for flexibility could be expected to induce supply side substitution from those whose assets could either provide daily or hourly flexibility.

In practice, there would appear to be no clear cut division between the flexibility segments of different durations because of the substitution possibilities that exist across the different timeframes. However, the level of demand and the ability of different sources to supply (e.g. due to volume constraints) to the different durations vary.

Gas quality

The L-gas system is technically capable of importing flexibility via quality conversion facilities from the H-gas network. The H-gas system is capable of substituting for flexibility transferred to the L-gas system by calling upon other H-gas sources of flexibility.

Conversely, the H-gas system is technically capable of importing flexibility from the L-gas system as a result of a change in the flow rate of gas through the converter facilities.

(31)

The physical extent to which flexibility may be transferred between the two systems is constrained by the average flow rate and the maximum capacity of the conversion facilities. For example, the maximum conversion capacity in 2006 was about 6.4 mcm/hour, compared to a peak demand of 17 mcm/hour in the L-gas market.

We conclude that there are likely to be two distinct flexibility markets for L-gas and H-gas, with the ability to transfer flexibility between the two systems limited by the capacity of quality converter stations. We are aware of future developments to quality conversion capacity and allocation rules that could potentially increase the availability of conversion capacity and reduce the direct cost of using that capacity. For this reason, we also assess the possibility of the relevant market being defined as the combined L-gas and H-gas systems.

Physical and commercial flexibility

As we described above, it is clear that physical flexibility provides shippers with a substitute for commercial flexibility. If a hypothetical monopolist of the commercial flexibility market raised price, he would lose significant market share to physical sources. This implies that there is a relevant market that comprises commercial flexibility and physical flexibility.

However, we also noted that commercial flexibility is not a substitute for physical flexibility. This implies that there is also a separate market which is for pure physical flexibility alone.

As with Frontier’s 2005 Report, we do not have adequate data to analyse the combined market of physical and commercial flexibility fully. Consequently, we focus on the physical markets for flexibility.

3.3.3 Geographic market

With respect to the geographic market definition we consider whether the market for flexibility is wider than the Netherlands, is restricted to the Netherlands or whether there are localised markets within the Netherlands.

The Dutch imbalance arrangements for the GTS system do not differentiate imbalances on a locational basis. Therefore, location does not constrain shippers’ choice between flexibility sources in different locations within the Netherlands. In addition, we understand that transmission constraints within the Netherlands, if any, are few and far between and do not have an incidence that warrants the treatment of sub-national markets.

Specific rules, prices, nomination procedures and available cross-border capacity for cross-border access may limit the ability of foreign sources of flexibility services to substitute for the provision of flexibility services in the Netherlands. GTS is connected to several German networks where flexibility services are offered. This implies that flexibility services outside the GTS system may be able to compete with flexibility services within the GTS system.

(32)

Market definition

constrained by existing cross-border capacity. The physical cross-border capacity is an upper bound on the actual capacity available to deliver flexibility across the border - shippers with rights to the cross-border capacity could decide not to use their rights and not to on-sell them for others to use.

3.4 MARKET DEFINITION CONCLUSIONS

In terms of product market definitions, we conclude that:

• predictable and unpredictable flexibility do not form distinct markets;

• there is no clear cut separation between flexibility of different durations. Since the ability for different sources of flexibility to supply the segments of different durations varies, we assess the market structure for each segment separately;

• the H-gas and L-gas systems form distinct markets and that flexibility services may be transferred between the two markets to the extent allowed by the capacity of quality conversion. Future developments to quality conversion capacity and access rules could potentially alter this finding. Therefore, we also explore with our structural analysis the hypothesis that H-gas and L-gas flexibility services form part of the same market); and

• physical flexibility forms a distinct market from commercial plus physical flexibility (and we consider only physical flexibility in the remainder of this report).

In terms of geographic market definitions, we conclude that the market(s) for flexibility services are likely to extend beyond the borders of the GTS system to the extent allowed by physical cross-border capacity.

(33)

4 Market structure

In this section we analyse the structure of the markets for flexibility identified in the previous section.

In considering whether GasTerra is dominant in the markets for gas flexibility services, we firstly investigate the market structure (this section). We then use the results of the market structure analysis to draw conclusions about dominance (next section). All assumptions for the quantitative analysis of market structure are documented in the Annexe.

In preparing the assumptions and analysis presented in this Final Report, Frontier has considered carefully all submissions made as part of DTe’s consultation on regulation of flexibility services for gas (published 24 January 2008). Frontier’s Initial Report was published as part of DTe’s consultation and parties were invited, inter alia, to comment on the analysis and assumptions presented by Frontier in that report.

4.1 APPROACH

Analytical approach to assessing market structure

Market structure is a key input into an assessment of dominance. The European Court of Justice has defined a dominant market position as:32

“A position of economic strength enjoyed by an undertaking which enables it to prevent effective competition being maintained on the relevant markets by affording it the power to behave to an appreciable extent independently of its competitors, customers and ultimately of consumers.”

Competition authorities generally consider that an undertaking will not be dominant unless it has substantial market power, which can be thought of as the ability profitably to sustain prices above competitive levels or restrict output or quality below competitive levels.

An undertaking’s market share is an important factor in assessing dominance but does not determine dominance in its own right. The European Court has stated that dominance can be presumed in the absence of evidence to the contrary if an undertaking has a market share persistently above 50%.33_{We understand that}

Dutch competition law is modelled on EU competition law and that due regard should therefore be given to precedence from the European Court of Justice and to the Court of First Instance.

We apply two quantitative approaches to assessing the structure of the relevant markets for gas flexibility services identified in the previous section:

• analysis of capacity shares; and

(34)

Market structure

• pivot analysis.

These are two separate ways of assessing the markets for gas flexibility services. Each method has certain advantages, as described below, which is why we apply them both in assessing dominance.

We do not have sufficient data on actual use of gas flexibility from different sources to analyse market shares based on use, in particular when making a forward looking assessment. Therefore, we analyse capacity shares for flexibility services.

Capacity shares consider the potential of different providers of flexibility services to compete. Capacity shares unlike market shares do not reflect market outcomes but rather market structure. Therefore, we investigate the merit order (i.e. relative costs) for the provision of flexibility services to the relevant markets and use this to make inferences about the relationship between capacity shares and market shares in a competitive market. This relationship is important when using capacity share information to help form conclusions about dominance. Capacity shares do not provide any information about the extent to which there might be a capacity overhang or shortfall in the relevant market. For example, conclusions about dominance drawn from a high capacity share in a market with a large capacity overhang would tend to overstate actual dominance in the market.34 For this reason we also undertake an assessment of whether GasTerra is pivotal in the relevant markets identified in Section 3.

In undertaking the capacity share and pivot analysis, we ascribe a source of flexibility to the party who is the owner of the source or who has long term contractual rights to the source. Details of how we have ascribed flexibility sources to different players are set out in the Annexe.

Forward looking assessment

The focus of our study is a three year outlook of the position of GasTerra in the markets for flexibility services.35 We present our results for the structural analysis for 2006 since this is the most recent period for which historic market data is available. We compare our findings for 2006 to our findings set out in Frontier’s 2005 Report (which relates to the years 2003/2004) to understand how the market structure has evolved over time. We then assess the market structure over the forward looking period 2009, 2010 and 2011. Details of the assumptions underlying our analysis are presented in the Annexe.

34 _{The same holds for other measures based on capacity shares like the Hirschmann-Herfindal-Index}

(HHI). We do not calculate HHI as it would require detailed information about all shippers in the market. Furthermore, capacity shares of GasTerra turn out to be sufficiently high (see Table 3) to induce HHIs in a range which would raise concerns independent of the remaining market structure.

35 _{2009 to 2011 is the period that DTe has asked us to assess. In addition, we have not considered an}

(35)

When we undertake a forward looking analysis of market structure, the assumptions and results become more uncertain the further forwards we look. Uncertainty over the timing of market developments, whether market developments will occur, the extent of market developments such as the entry of new capacity and the likely future allocation of sources of flexibility services to different market players all increase the further forwards one looks. For this reason, our analysis of the market structure for 2009 is more certain than for 2010, which is in turn more certain than for 2011.

In undertaking the analysis of market structure we have attempted to be conservative, i.e. to err on the side of making GasTerra’s market position smaller than it actually is. While we note that in the case of certain specific assumptions it may have been possible to have been even more conservative than we were, on balance we believe our analysis to have been conservative.

Why results differ from Frontier’s 2005 Report

The extent of the relevant product and geographic markets are unchanged from Frontier’s 2005 Report. Therefore, the assessment of market structure is applied to the same relevant markets as in Frontier’s 2005 Report. However, the results of Frontier’s 2007 study differ from Frontier’s 2005 Report for the following reasons:

Improved data or information: in 2005 the data available to us were not as extensive as now. In particular, the responses to questionnaires for the gasmonitor provide an important source of information and increase the insights we have of the market.

• For example, this affects our assumptions about GasTerra’s share of quality conversion. In Frontier’s 2005 Report we assumed conservatively that GasTerra had no direct control over quality converters. In the case of Frontier’s 2007 study, we have information about the quality conversion bookings of GasTerra and the total capacity offered by GTS, which allows us to derive a more realistic capacity allocation than we did for Frontier’s 2005 Report.

• We apply a different treatment regarding the control over linepack than was applied in the 2005 Report. In Frontier’s 2005 Report we allocated control of linepack to non-GasTerra players on the basis that any reduction in entry or exit flows by GasTerra would automatically be accompanied by a re-allocation of tolerance away from GasTerra. In Frontier’s 2007 study, we also follow this logic. However, more detailed information has become available to us about GasTerra’s position (e.g. data from the gasmonitor and insights from the market consultation). As a result of GasTerra’s long term bookings at entry and exit points and its long term contractual flow commitments, we assume a portion of linepack is controlled by GasTerra.

(36)

Market structure

demand during cold spells. Therefore, Maasvlakte is not available to provide flexibility services to meet normal demand.

Market developments: during the three year period between the reference years used for Frontier’s 2005 Report (2003) and Frontier’s 2007 study (2006) market developments have occurred. For example, in 2005 Essent commissioned the Epe cavern storage which is located in Germany. This is included in the quantitative analysis for 2006 since it is directly connected to the GTS system.

Methodological modifications: in Frontier’s 2005 Report capacity shares were reported for winter and summer separately. We restrict ourselves to a yearly approach to aid the clarity of this report. Our experience from the analysis applied in Frontier’s 2005 Report is that this simplification to the analysis is unlikely to detract from the strength of the conclusions from this study. Some of the changes in data or assumptions strengthen the position of GasTerra (e.g. the realistic shares of converter bookings or the exclusion of Maasvlakte LNG storage) while others increase the capacity available to non-GasTerra parties (e.g. new storages and increased linepack due to new information published by GTS). All parameters and assumptions of the quantitative analysis are detailed in the Annexe.

4.2 CAPACITY SHARES

We base our analysis of capacity shares for the years 2006 and 2009-2011 on the maximum flow minus the average flow definition of flexibility. This definition is attractive because it reflects the idea that flexibility is needed to balance a situation with unexpectedly high demand. The UK Competition Commission’s36

use of this definition for its investigation of the Rough storage facility implies that the approach we adopt is consistent with a recent regulatory precedent. We therefore estimate available capacity to supply gas flexibility in each market segment (hourly, daily, weekly and seasonal) as the difference between the maximum and average capability.

The capacity share analysis treats the different market segments as though they are not mutually exclusive, i.e. a unit of capacity could supply flexibility into any market segment simultaneously. This approach has the effect of maximising the deliverability of both GasTerra and competitors into each market segment.

Supply of flexibility in the hourly segment is calculated as the difference of capacity (maximum flow) and average utilisation (average flow). For segments with longer durations, with the exception of volume limited sources of flexibility supply, we scaled up hourly flexibility according to the length of the period (e.g. we multiplied hourly flexibility by 24 to get to daily flexibility). Storages are constrained by their working gas volume (or twice their working gas volume in

(37)

the case of cavern storages37_{). Linepack may provide flexibility only within a day}

and is therefore restricted to the hourly segment. Furthermore, we consider interruption of contracts up to two days per year. A detailed description of all data underlying the capacity share analysis is set out in the Annexe.

Table 3 shows our estimated capacity shares for the two relevant markets, i.e. the H-gas system and the L-gas system. We also show capacity shares for the case of a combined H-gas and L-gas system (labelled “GTS”).

GasTerra controls the majority of the capacity in the market. For example, in 2006 it accounted for 80% of the capacity that can potentially be used to supply hourly flexibility in the L-Gas market.

For all four L-gas market segments and in all five years we observe that GasTerra exhibits very high capacity shares, up to 86% percent in the seasonal segment of 2006. This fact is driven by GasTerra’s control of storage capacity, control of

37 _{As explained in the Annexe, we conservatively assume that cavern storages are able to cycle their}

storage twice within a season (i.e. begin the season with full storage, empty the storage, refill it and empty it again). 2006 2009 2010 2011 L-Gas 80% 70% 65% 60% H-Gas 65% 56% 50% 44% Hourly flexibility GTS 77% 70% 65% 57% L-Gas 85% 74% 69% 66% H-Gas 66% 57% 50% 45% Daily flexibility GTS 82% 74% 69% 62% L-Gas 86% 75% 70% 67% H-Gas 68% 59% 52% 46% Wee k ly flexibility GTS 84% 76% 71% 64% L-Gas 86% 76% 72% 68% H-Gas 66% 55% 47% 41% Seasonal Flexibility _GTS _{86% 77% 74% 63%}

Table 3: GasTerra’s capacity shares over time

(38)

Market structure

much of the flexibility made available through the conversion facilities, and control of the Groningen field.

GasTerra’s capacity shares in the H-gas market, while significantly lower than in all corresponding L-gas markets, remain above 50% for the years 2006 and 2009. GasTerra’s capacity shares decline over time within all market segments. In 2010 capacity shares are greater or equal to 50% in all segments with the exception of seasonal flexibility and in 2011, capacity shares are below 50% for all segments. The reduction in GasTerra’s capacity shares over time is mainly due to additional converter capacities in 2010 and additional import capacities.

The view of the combined H-gas and L-gas markets strongly resembles the analysis of the L-gas market. This results from the special role of converters which contribute significantly to non-GasTerra players’ supply of flexibility in the H-gas market. Converters merely shift flexibility between markets and are therefore excluded as a source of flexibility in the analysis where we assume that the entire GTS system forms a single market.

GasTerra’s capacity shares tend to increase with the duration of the market segment (e.g. moving from hourly to daily to weekly durations etc). This is because volume constrained sources of flexibility, e.g. linepack, contribute less than proportionately to the longer duration segments. This trend is reversed however for seasonal H-gas flexibility. Storages, for which GasTerra has a disproportionately high stake, are constrained by working gas volume in contrast to flexibility from imports and conversion.

Our findings from the capacity share analysis can be summarized as follows: GasTerra’s capacity shares in the L-gas market and the overall GTS system are significantly higher than 50% over all years.38_{Capacity shares in the H-gas}

market fall below this threshold in 2010 (seasonal only) and 2011; GasTerra’s capacity shares decline over time; and

GasTerra’s capacity shares tend to be higher in segments with longer duration (i.e. seasonal flexibility), with the biggest difference being between the segments for hourly and daily flexibility. This trend is reversed for seasonal flexibility in the H-gas sector.

4.3 PIVOT ANALYSIS

GasTerra is considered pivotal if market requirements for flexibility cannot be met without the use of assets controlled by GasTerra. In contrast to capacity shares, pivot analysis takes account of information about the level of demand relative to the level of capacity in the relevant market. For this reason, where

38 _{We note that the relevant market cannot be both the separate L-gas and H-gas markets and the}