the european natural gas sector between regulation and
TRANSCRIPT
The European Natural Gas Sector –
Between Regulation and Competition
by
Marcus Stronzik
A thesis submitted in partial fulfilment
of the requirements for the degree of
Doctor of Philosophy
in Economics
Approved, Thesis Committee
Prof. Gert Brunekreeft, Jacobs University Bremen (Chair)
Prof. Colin Vance, Jacobs University Bremen
Prof. Tooraj Jamasb, Heriot-Watt University, Edinburgh
Date of Defense: July 2, 2012
School of Humanities and Social Sciences
Preface I
Abstract
One of the major goals of European energy policy is the establishment of a single and
competitive internal gas market by 2014. In 2009, the European Union (EU) adopted the third
gas Directive on common rules to foster market integration across Europe. When the EU
started the liberalisation process in 1998, vertically integrated gas companies, acting as
(regional) monopolies and serving the whole value chain, were the dominant organisational
structure throughout Europe.
This PhD thesis explores in four related articles whether the efforts of restructuring the
European natural gas sector have been successful in creating a regulatory framework that is
well-suited to provide for a level-playing field for market players leading to competitive market
outcomes. Access conditions to infrastructure facilities (pipelines and storage) are of special
interest. The first paper evaluates a specific regulatory measure, i.e. ownership unbundling of
transmission system operators as the strongest form of vertical separation. Applying dynamic
estimators on an unbalanced panel of 18 EU countries, it turns out that the more modest
legal unbundling reduces natural gas end-user prices, whereas ownership unbundling shows
no impact. Looking at natural gas wholesale prices it is assessed if markets are able to
generate efficient price signals. The second article tests for the spatial no arbitrage condition.
Based on price developments at two German hubs and the nearby Dutch market,
cointegration analysis and a state space model with time-varying coefficients are used.
Though market efficiency in terms of information processing has increased, price differentials
are only partly explained by transportation costs pointing at capacity constraints. In the third
paper, the relationship between natural gas storage utilisation and price patterns at three
major European trading points is investigated (test of the intertemporal no arbitrage
condition). The results reveal that market performance differs substantially from the
competitive benchmark. The last article puts the empirical results of the previous papers into
a broader policy context and deduces policy recommendations.
Preface II
Acknowledgements
This work would not have been possible without the help of many others. The following list is
without any claim of completeness. I would like to express my special gratitude to Gert
Brunekreeft and Christian Growitsch for inspiring discussions, constructive criticism, and
support throughout the whole period. Both played a vital role in triggering this piece of
research and keeping it alive. I greatly enjoyed the joint work.
Furthermore, I would like to thank Tooraj Jamasb and Colin Vance for taking over the
supervision of my thesis. A big “thank you” goes to the other co-authors, Anne, Margarethe
and Rabindra, and my colleagues at WIK for countless and fruitful discussions.
I appreciate financial support from the German Federal Network Agency. Several
discussions with staff members enriched and sharpened my understanding of regulatory
framework conditions. Moreover, I thank colleagues from national and international
universities, institutes, and numerous conference meetings for providing ideas, criticism, and
comments.
Last, not least, I would like to thank my friends and family for their endless support and
encouragement. Petra, Mom and Dad, I would like to dedicate this work to you.
Preface III
Declaration on joint authorship
The thesis is based on four main papers of which three are joint work. For each paper
separately, the contributions of the various authors are laid out in the following.
The first paper, “Ownership unbundling of gas transmission networks – Empirical evidence
from a dynamic panel approach” (Chapter 2), is joint work of Christian Growitsch and Marcus
Stronzik. For a first draft, Marcus Stronzik wrote the core text and developed the panel
dataset. Estimations and the interpretation of results were carried out jointly. This first version
was presented by Marcus Stronzik at two conferences with full paper submission and a
double-blind review process. Based on a data update, Marcus Stronzik revised estimations
(e.g. extending the approach to system GMM) and the text. The long run effects were added
by Christian Growitsch. The paper is a result of in-depth discussions on the topic among both
authors over the whole period of writing.
For the second paper, “Price convergence and information efficiency in German natural gas
markets“ (Chapter 3), the basic idea was jointly developed by Christian Growitsch and
Marcus Stronzik. Rabindra Nepal elaborated a first preliminary draft. Marcus Stronzik
completely re-estimated (e.g. incorporating transaction costs and extending both approaches
to an error correction model) and re-wrote the paper. Christian Growitsch provided
substantial supervision throughout the whole period of writing. The paper was presented by
Marcus Stronzik at one conference with full paper submission and a double-blind review
process.
For the third paper, “Does the European natural gas market pass the competitive benchmark
of the theory of storage? Indirect tests for three major trading points” (Chapter 4), the
contributions correspond to the ordering of authors of the publication
Stronzik/Rammerstorfer/Neumann (2009) in Energy Policy. Based on the idea developed by
Marcus Stronzik, Anne Neumann and Margarethe Rammerstorfer provided the basic data.
Estimations and paper writing were done by Marcus Stronzik with the collaboration of the two
co-authors.
The fourth paper “Europe on its way to a single natural gas market: How far have we come?”
(Chapter 5) is single-authored by Marcus Stronzik.
Preface IV
Table of contents
Abstract I
Acknowledgements II
Declaration on joint authorship III
Table of contents IV
Figures VII
Tables VIII
Abbreviations IX
Nomenclature XI
1 Introduction 1
1.1 Research questions 2
1.1.1 Sub-question 1: Evaluation of ownership unbundling 3
1.1.2 Sub-question 2: Test of the spatial no arbitrage condition 5
1.1.3 Sub-question 3: Test of the intertemporal no arbitrage condition 7
1.2 Basic methodological approach 8
1.3 Outline of the thesis 9
References 11
2 Ownership unbundling 14
Ownership unbundling of gas transmission networks – Empirical evidence from
a dynamic panel approach 14
2.1 Introduction 14
2.2 Related literature 16
2.3 Data 19
2.4 Empirical analysis 22
2.5 Conclusions 25
Appendix 1: The US case 27
Appendix 2: Descriptive statistics 29
References 32
3 Spatial no arbitrage condition 35
Price convergence and information efficiency in German natural gas markets 35
3.1 Introduction 35
3.2 Institutional design and recent developments in Germany 36
Preface V
3.3 Literature review 38
3.4 Econometric methodology 39
3.4.1 The law of one price and transmission charges 39
3.4.2 Cointegration tests 40
3.4.3 Time-varying coefficient model 41
3.5 Data 43
3.6 Results 45
3.6.1 Cointegration analysis 45
3.6.2 Time-varying coefficient 47
3.7 Conclusions 50
References 52
4 Intertemporal no arbitrage condition 54
Does the European natural gas market pass the competitive benchmark of the
theory of storage? Indirect tests for three major trading points 54
4.1 Introduction 54
4.2 Literature overview 55
4.3 Empirical model 56
4.3.1 Test on relative price variations 57
4.3.2 Test on market performance 58
4.4 Data 59
4.5 Results 62
4.5.1 Relative price variation 62
4.5.2 Market performance 65
4.6 Conclusions 66
References 68
5 Policy paper 70
Europe on its way to a single natural gas market: How far have we come? 70
5.1 Introduction 70
5.2 Overview of regulation history 72
5.3 Market access 75
5.3.1 Unbundling 75
5.3.2 Wholesale 81
Preface VI
5.3.2.1 Commodity 82
5.3.2.2 Pipeline capacity 87
5.3.3 Storage 92
5.4 Security of supply 94
5.4.1 The EU concern 94
5.4.1.1 Risk of supply disruptions 95
5.4.1.2 Risk of high import prices 99
5.4.2 Investment-related regulations 103
5.5 Conclusions 109
References 112
Declaration 119
Preface VII
Figures
Figure 1: Import dependency of the European gas sector 2
Figure 2: Natural gas end-user prices for households 5
Figure 3: Spot price differentials for three major European hubs 6
Figure 4: End-user prices [USD2000/107 kcal] 19
Figure 5: Logarithmic day-ahead spot prices (€/MWh) 44
Figure 6: Price convergence [β] 48
Figure 7: Information efficiency [ ] 50
Figure 8: Spot and futures prices for delivery at NBP (log) 60
Figure 9: Value chain of the gas sector 71
Figure 10: Interim assessment 73
Figure 11: Physical gas flows 2009 95
Preface VIII
Tables
Table 1: Overview of empirical studies 18
Table 2: Natural gas end-user prices and the impact of regulatory reforms 24
Table 3: Natural gas end-user prices and the impact of
regulatory reforms (including the US) 28
Table 4: Descriptive statistics and information on the unbundling status
of gas TSOs 29
Table 5: Correlation matrix of regulatory indicators 30
Table 6: Correlation matrix and auxiliary fixed effects estimation
for control variables 31
Table 7: Transmission charges 45
Table 8: Unit root tests 45
Table 9: Long-run cointegrating equations (ML estimation) 46
Table 10: Results of the time-varying coefficient models 47
Table 11: Unit root tests 61
Table 12: Granger causality 61
Table 13: Summary statistics of daily convenience yields 62
Table 14: Correlation analysis for convenience yields 63
Table 15: Estimation results for convenience yields 64
Table 16: Estimation results for 6-month and 12-month bases 65
Table 17: Current access conditions in EU Member States 79
Table 18: European gas hubs vs. Henry Hub 83
Table 19: Gas infrastructure investment projects 97
Preface IX
Abbreviations
ACER Agency for Cooperation of
Energy Regulators
ADF Augmented Dickey Fuller
AIC Akaike Information Criterion
AR Autoregressive
ARCH Autoregressive Conditional
Heteroskedasticity
GARCH Generalised ARCH
BNetzA Bundesnetzagentur
CAPEX Capital Expenditures
CBM Coalbed Methane
CEGH Central European Gas Hub
CR Concentration Ratio
DSO Distribution System Operator
EEPR European Energy Programme
for Recovery
EEX European Energy Exchange
EIP Energy Infrastructure Package
ENTSOG European Network of
Transmission System Operators
for Gas
ERGEG European Regulators Group for
Electricity & Gas
ETCR Indicators of Regulation in
energy, transport and
communications
EU European Union
FCFS First Come First Served
FE Fixed Effects
FERC Federal Energy Regulatory
Commission
FID Final Investment Decision
FRA France
FSU Former Soviet Union
GDP Gross Domestic Product
GECF Gas Exporting Countries’ Forum
GGPLNG Guidelines of Good TPA
Practice for LNG System
Operators
GGPSSO Guidelines of Good TPA
Practice for Storage System
Operators
GMM Generalised Method of
Moments
GPL Gaspool
GSE Gas Storage Europe
IEA International Energy Agency
IP Interconnection Point
ISO Independent System Operator
ITO Independent Transmission
Operator
IV Instrumental Variable
kcal kilocalorie
Preface X
KPSS Kwiatkowski Phillips Schmidt
and Shin
ktoe kilotonne of oil equivalent
kWh kilo Watt-hour
LNG Liquefied Natural Gas
LR Likelihood Ratio
LSDVC Bias-Corrected Least-Squares
Dummy Variable
MW Mega Watt
MWh Mega Watt-hour
MOI Market Opening Index
n.i.i.d. normally, independently and
identically distributed
NBP National Balancing Point
NCG NetConnect Germany
NYMEX New York Mercantile Exchange
OECD Organisation for Economic Co-
operation and Development
OPEC Organization of Petroleum
Exporting Countries
OTC Over-the-Counter
p.a. per annum
PEG Points d’Echange de Gaz
PP Phillips Perron
PSV Punto die Scambio Virtuale
RE Random Effects
REMIT Regulation on Energy Market
Integrity and Transparency
RIIO Revenue set to deliver stronger
Incentives, Innovations and
Outputs
RoR Rate of Return
SIC Schwarz Information Criterion
TPA Third Party Access
nTPA negotiated TPA
rTPA regulated TPA
TSO Transmission System Operator
TTF Title Transfer Facility
TYNDP Ten Year Network Development
Plan
UIOLI Use It Or Lose It
UIOSI Use It Or Sell It
UK United Kingdom
US United States
VAR Vector Autoregressive
VECM Vector Error Correction Model
WACC Weighted Average Cost of
Capital
WTI West Texas Intermediate
ZEE Zeebrugge
Preface XI
Nomenclature
Chapter 2
Indices:
i country
t time period (years)
Variables and parameters:
coefficient (scalar)
,, vector of coefficients
error term
* long run multiplier
unobserved heterogeneity
R vector of regulatory indicators
X vector of oil prices (log) [USD2000/barrel]
y end-user price households (log) [USD2000/107kcal]
Z vector of control variables
Chapter 3
Indices:
ji, hub
t time period (days)
Variables and parameters:
coefficient for short-term adjustment speed of gas prices (loadings)
coefficient for long-run relationship of gas prices (cointegrating vector)
, error term
2 variance
c constant
d dummy for direction of gas flows
P gas spot price [€/MWh]
netP spot price net of transmission charges [€/MWh]
Preface XII
netp log spot price net of transmission charges [€/MWh]
jiTC transmission charges for gas flows from i to j [€/MWh]
(.)E expectation operator
. first difference operator
Chapter 4
Indices:
i quarter of a year
t time period (days)
T maturity (months)
Variables and parameters:
coefficient
F natural gas futures price [€/MWh]
Q quarterly dummy
oil oil price [€/MWh]
r interest rate
S natural gas spot price [€/MWh]
u error term
y convenience yield
Introduction 1
1 Introduction
The creation of genuine internal energy markets for electricity and gas is one of the
European Union’s priority objectives. The existence of a competitive internal gas market is
seen as a strategic instrument in terms both of giving consumers a choice between different
companies supplying energy at reasonable prices, and of making the market accessible for
all suppliers. The introduction of competition down- (supply to end-users) as well as
upstream (import/wholesale) of gas networks has been triggered - at least to some extent -
by insights from economic theory, namely the theory of contestable markets (Baumol et al.
1982). Regulation should be limited to those parts of the economy where a natural
monopoly1 is accompanied by irreversible costs. Only in such a situation, can a monopolist
gain supra-normal profits. Without irreversible costs, the monopolist is disciplined by potential
competitors who will enter the market once the incumbent tries to raise prices above the
competitive level. With regard to gas, the condition generally holds for the pipeline
infrastructure for transporting and distributing gas, while other parts of the value chain should
be open to competition.
In the 1990s, vertically integrated gas companies, acting as (regional) monopolies and
serving all parts of the value chain, were the dominant organisational structure throughout
Europe. In order to open parts of the gas sector for competition, a restructuring was
necessary. One of the first liberalisation steps on the European level was the gas Directive of
1998 (98/30/EC), which consisted of a quite lax option manual for Member States. Besides a
stepwise opening of the end-user market with a target of 30% until 20102, Member States
could choose between regulated or only negotiated third party access (TPA) to the grid.
Concerning vertical separation of utilities, accounting unbundling3 was set as a minimum
standard. This new regulation resulted in differing rules across Europe4, a still fragmented
market with hardly any competition up- or downstream. The European gas Directive of 2003
(2003/55/EC) tightened the bottom lines of the regulatory framework. Full opening of both
end-user markets, households as well as industrial consumers, should have been achieved
by July 2007. Regulated TPA became mandatory and legal unbundling5 was set as the
minimum standard. Having investigated the energy sector quite in depth, the European
Commission was still unsatisfied with market outcomes, especially in the gas sector
(European Commission 2007). The Commission criticised low switching rates especially in
the household sector and found the market highly concentrated and still dominated by
incumbents. The so-called sector inquiry triggered a political debate that led to the third gas
Directive agreed upon in 2009 (2009/73/EC). An Agency on the European level has been
established fostering cooperation between national energy regulators. The Directive tightens
the minimum standard for vertical separation of energy companies once again and aims at
1 A natural monopoly means that the market is served at least cost by only one company. This situation is
characterised by subadditive cost functions. It should be noted that a natural monopoly today does not mean that it will persist in the future. Market dynamics such as a shift in demand or technological progress can lead to different cost structures in the future. See e.g. Knieps (1997).
2 This target meant that by 2010 30% of consumers should have a choice between different suppliers. 3 Accounting unbundling means that the utility has to set up separate accounts for their different services. 4 E.g., while some countries opted for negotiated TPA (like Germany), others implemented regulated TPA. 5 Legal unbundling means that the different services have to be operated by separate companies which can
still belong to the same owner.
Introduction 2
improving market transparency and promoting investments into cross-border transmission
capacities.
1.1 Research questions
The leading research question for the thesis is whether these efforts throughout Europe of
restructuring the gas sector have been successful in creating a regulatory framework that is
well-suited to provide for a level-playing field for market players leading to competitive market
outcomes.
The importance of a well-functioning European gas market is highlighted by Figure 1 with two
main developments. The black line depicts domestic gas production in the EU-156 in percent
of total final gas consumption. The blue line shows the share of total final energy
consumption based on gas. While indigenous production is relatively shrinking indicating an
increasing import dependency of the EU, the importance of gas as an energy source is
growing. According to the IEA (2010) reference scenario for the EU, gas demand is projected
to increase by 0.7% p.a., while indigenous production will decrease by 3.1% p.a. due to
limited European gas reserves. Thus, Europe’s dependency on imports is projected to rise to
around 80% or 90% by 2030.
Figure 1: Import dependency of the European gas sector
Source: Own calculations based on IEA (2009a).
6 EU-15 has been chosen due to missing values for a lot of countries out of EU-27.
0%
20%
40%
60%
80%
1980 1985 1990 1995 2000 2005
Consumption Production
Introduction 3
This translates into the necessity of transporting an increasing amount of gas from outside
the EU through the network to the main consuming regions inside the EU.7 Getting the
regulatory framework wrong and providing gas at too high prices to end-users will reduce
social welfare substantially. Primarily, regulators have to address two topics:
the level of network charges in order to prevent network operators from earning
supra-normal profits and
the access to networks in order to provide a non-discriminatory framework in which all
market participants have the same conditions for using the pipelines.
The thesis focusses on the second issue as it is more important in the current context for two
reasons. First, regarding end-user prices, network charges usually have only a minor cost
share of less than 20%.8 Second, during recent years many regulators have concentrated on
the first issue and introduced new ratemaking provisions while paying less attention to the
actual access conditions. This means that a lot of work has already been done with regard to
ratemaking, particularly during the preparation phases of these regulations. Only very
recently, regulators have begun to touch access conditions. The thesis tackles this issue by
sub-dividing the leading research question of a well-functioning natural gas market into three
sub-questions, described in the following.
1.1.1 Sub-question 1: Evaluation of ownership unbundling
For third parties, the network cannot be by-passed and is an essential facility to their
business. Thus, it constitutes a monopolistic bottleneck. A network operator that is also
active in the downstream and/or upstream markets might have an incentive to discriminate
against third parties through cross-subsidisation or non-tariff measures (cf. Perry 1989 and
Vickers 1995). Therefore, the vertical structure of the gas sector is crucial concerning market
outcomes. During the political discussions preceding the third gas Directive the strengthening
of the unbundling standards was one of the most debated topics. While the Commission has
argued in favour of the most rigid form of vertical separation, ownership unbundling9, some
Member States (e.g. Germany and France) as well as industry have opposed these efforts,
doubting the benefits. The Directive now allows for both legal as well as ownership
unbundling, and leaves the choice to Member States. Though final agreement has been
reached, the question of whether ownership unbundling is superior to legal unbundling
remains on the political agenda. In both cases, transmission and other gas services have to
be operated by legally separated entities. These companies can have the same owner under
legal unbundling, while ownership unbundling requires an operation by different owners.
7 It is expected that liquefied natural gas (LNG) will play an increasing role in the future European energy mix.
Nevertheless, LNG will not be considered here in order to keep the focus as straight as possible. Furthermore, today conventional pipeline transport is still dominant. E.g., 2010, LNG represented around 30% of total gas import capacities with the remaining 70% referring to gas transmission (ERGEG 2010). For the role of LNG and its future perspectives see e.g. Rüster and Neumann (2009) as well as Rüster (2010).
8 See e.g. BNetzA (2009). Cost shares differ across countries due to tax regimes and across end-user
groups. 9 The network operator is not allowed to own companies (or substantial parts of them) operating services in
the down- or upstream markets.
Introduction 4
Ownership unbundling is preferable if it increases social welfare. This is the case if an
introduction leads to lower end-user prices compared to a situation when ownership
unbundling is absent.10 Economic theory gives little guidance as ambiguous results are
reported (cf. Laffont and Tirole 1993, Vickers 1995 and Buehler 2005). On the one hand, it is
argued that the stricter vertical separation is designed, the less incentive a network operator
has to discriminate between affiliated undertakings and third parties. Thus, competition on
down- and upstream markets will be fostered and risk of vertical foreclosure reduced. On the
other hand, ownership unbundling may lead to a loss of economies of scope or double
marginalisation. E.g., retail firms or shippers may have better information concerning future
demand and supply conditions easing network planning if all information lies within one
company. Regarding double marginalisation, ownership unbundling and oligopolistic or
monopolistic structures in the down- or upstream markets may result in a double mark-up
billed to consumers. Since vertically integrated or only legally unbundled companies
maximise their profits over the whole holding, the mark-up occurs only once. The net welfare
effect depends on model assumptions such as the degree of competition in the downstream
market, the demand function, the role of investments or the design of access price
regulation.
While some countries have already introduced ownership unbundling for gas transmission
system operators (e.g. Netherlands, Spain and UK), others stick to legal unbundling (e.g.
France). Figure 2 depicts the development of natural gas end-user prices for households in
the four mentioned countries indicating that there is no straightforward relationship between
ownership unbundling and the evolution of end-user prices. Prices tend to converge, but with
Spain still having prices roughly 30% above the others in 2007. France, the Netherlands and
the UK are more or less at the same level but with differences in the development over time.
At the beginning of the considered period Dutch households paid the lowest prices for gas. In
2007, end-users in France and the UK were better off.
Therefore, whether ownership unbundling is a proper tool to foster competition in the gas
sector remains an open question. This leads to the first sub-question addressed in the thesis:
Sub-question 1
Are countries with ownership unbundling performing better than countries without? In
particular: Has the introduction of ownership unbundling led to a reduction in end-user
prices?
10 Note that lower end-user prices are not a sufficient condition for higher social welfare as investments may be
negatively affected (cf. Buehler et al. 2004, 2006).
Introduction 5
Figure 2: Natural gas end-user prices for households
Source: Own calculations based on IEA (2009b). Note: Prices adjusted for Purchase Power Parity in constant USD2000.
1.1.2 Sub-question 2: Test of the spatial no arbitrage condition
While the first sub-question analyses one specific regulatory measure, the following two sub-
questions broaden the scope, turning the focus upstream to the wholesale market. In many
European countries the liberalisation process has resulted in the separation of gas from the
pure transportation service allowing gas to become a tradable commodity (like oil). To handle
gas like a financial contract increases flexibility of market participants along the value chain
and allows new players (e.g. financial intermediaries) to enter the market.11 Nowadays, gas
is traded at various market places throughout Europe, either at exchanges like the European
Energy Exchange (EEX) or at virtual trading points, so called hubs. Only if these market
places are able to generate competitive price signals can the sector exploit the full flexibility
potential. Non-competitive price formations translate into higher costs in moving gas along
the value chain, thus lowering social welfare.
Figure 3 shows the spot price differentials between three major European trading points, the
Dutch Title Transfer Facility hub (TTF), Zeebrugge (ZEE) in Belgium and the National
Balancing Point (NBP) in UK, with NBP as the most liquid trading place in Europe serving as
a reference. Price differentials have narrowed indicating some degree of convergence, which
goes along with an increase in trading volume at most hubs. The higher the liquidity of a
market is the more efficient the market is expected to perform. Concerning gas markets,
11 In light of the financial and economic crisis, the reader might give his/her own judgement whether this is
positive or negative. At least with regard to economic theory, a situation with high flexibility is usually preferable over a situation with low flexibility. The abuse of flexibility is more an issue of poorly designed framework conditions (e.g. intransparent market rules).
0
200
400
600
800
1989 1994 1999 2004
US
D/1
07 k
ca
l
FRA NDL UK SPA
Introduction 6
usually the churn rate is taken as an indicator for liquidity, measuring the ratio between
traded and physically delivered volumes. The corresponding churn rates are 15 for NBP, 4.5
for Zeebrugge and 4 for TTF. Compared to the world’s largest and most mature gas market,
the Henry Hub in the US, with a churn rate of around 350, European trading volumes are still
much lower. However, European gas wholesale markets may exhibit an efficient pricing
behaviour.
Figure 3: Spot price differentials for three major European hubs
Source: Own calculations based on Heren data. Note: Logarithms of day-ahead prices have been taken.
Competitive, thus efficient markets are characterised by the absence of (persistent) arbitrage
opportunities, i.e. market participants are not able to earn riskless profits by simultaneously
entering into transactions in two or more markets (Fama 1970). Once such an opportunity
exists, a rational actor takes advantage of it, levelling out price disparities very quickly.
Particularly, arbitrage freeness has to be fulfilled in two dimensions, in a spatial as well as in
a temporal context. Spatial arbitrage opportunities mean price disparities for the same
product traded at regionally distinct locations at the same point in time. Intertemporal
arbitrage points at price differences for a product traded at one market place but at different
points in time.
Regarding spatial arbitrage, if a market participant observes different prices at two distinct
locations, he will buy gas at the cheap market place and transport it to the high-price region
as long as the price differential is large enough to cover transportation costs. Therefore, the
-1.500
-1.000
-0.500
0.000
0.500
1.000
1.500
Oct-0
5
Apr-0
6
Oct-0
6
Apr-0
7
Oct-0
7
Apr-0
8
Oct-0
8
Eu
r/M
Wh
TTF vs. NBP
ZEE vs. NBP
Introduction 7
spatial no arbitrage condition for efficient markets claims that prices of identical products
traded at regionally distinct locations should differ in transaction costs only, the so called law
of one price (cf. Baulch 1997). Concerning natural gas markets, transaction costs are mainly
reflected by charges shippers have to pay for gas transmission.
One prerequisite for arbitrage is that the regulatory environment provides for sufficient
access to the network meaning that free capacity is available to the arbitrageur if required.
Otherwise, in addition to transportation charges other costs would be present in the market
preventing shippers from trading and leading to lower overall social welfare. These additional
costs would indicate capacity constraints that might be due to technical or contractual
reasons. While the former is a sign of underinvestment, the latter points at hoarding of
capacity rights by incumbents. Transmission capacity is often booked long in advance and
has to be nominated when shippers actually want to use it. Hoarding of capacity rights
means unused capacity which is not available to other market participants signalling the
abuse of market power. Both issues, the underinvestment (especially with regard to cross-
border transmission capacity) as well as unused capacity rights, have been raised by the
European Commission in their sector inquiry (European Commission 2007). This leads to the
second sub-question:
Sub-question 2
Do European gas markets - on the wholesale level – meet the spatial no arbitrage condition?
In particular: Have any improvements been achieved concerning spatial arbitrage through
certain changes in the regulatory framework?
1.1.3 Sub-question 3: Test of the intertemporal no arbitrage condition
Since natural gas is a storable commodity, not only the network should be efficiently used but
also storage facilities to avoid unnecessary welfare losses. The efficient use of storage is
closely related to intertemporal arbitrage. If prices are low and market participants expect
them to rise, gas will be bought at the spot market and put into storage. Later, when prices
are high, gas will be withdrawn from storage and sold at the spot market. Therefore, storage
facilities provide market players with some additional flexibility. Various storage opportunities
exist serving different needs for flexibility. While depleted oil and gas reservoirs as well as
aquifers have relatively low deliverability and injection rates serving seasonal arbitrage,12 salt
caverns with high rates can also be used for short-term arbitrage.13
Concerning intertemporal arbitrage, the theory of storage provides for the corresponding
market condition of efficient pricing behaviour. The theory shows that filling quantities are
determined by the equivalence of marginal storage cost and the price spread defined as the
12 As a large part of natural gas is used for heating purposes, demand follows a seasonal pattern with peaks
during winter. 13 Please note that several other types of storage exist. E.g., the network itself can be used to store gas, the so
called line-pack. For an overview of the various types of natural gas storage and their characteristics see Grewe (2005).
Introduction 8
difference in spot and futures prices (Working 1949). This condition only holds as long as
futures prices do not fall below spot prices, which cannot be assured in the long run.
Consequently, Brennan (1958) has included an additional factor, the convenience yield, that
measures an implicit stream of benefits a consumer of a commodity (such as natural gas)
receives from storage.14 These benefits for the holder of inventories arise because the
stored product may depict an input for further production or it may increase the ability to meet
unexpected future demand. Therefore, the intertemporal no arbitrage condition claims that
the return from purchasing the commodity today and selling it for delivery later (the so-called
basis) equals the interest forgone by storing the commodity plus marginal storage cost less
marginal convenience yield from an additional unit of inventory (cf. Williams and Wright
1991).
Though the European Commission emphasises the necessity for independent storage
operators and the increase in transparency of available capacities to third parties as
preconditions for efficient market operations, only non-binding guidelines for storage
operators are in place on the European level (ERGEG 2011). Compared to pipelines, storage
has much lower economies of scale, and essential facility characteristics are less clear (cf.
e.g. Creti 2009). The current gas Directive leaves Member States with the option of
implementing a negotiated or regulated access regime for storage. As a result, access
regimes vary across Europe. While, e.g., Belgium implemented regulated third party access,
Germany opted for a negotiated regime. The Netherlands and the UK decide on a case-by-
case basis whether regulated or only negotiated third party access is applied. However,
availability of storage capacity is crucial for market players if they want to gain from temporal
flexibility. This leads to the third sub-question:
Sub-question 3
Do European gas markets - on the wholesale level – meet the intertemporal no arbitrage
condition? In particular: Are there any differences across Europe?
1.2 Basic methodological approach
The thesis is concerned with the relation between regulatory framework conditions of the
European natural gas sector and actual market outcomes. According to the leading research
question, it is analysed if access regulations of essential facilities are properly set to provide
for competition in the up- or downstream markets. The first sub-question asks whether a
specific regulatory measure, ownership unbundling, has influenced end-user prices. The
other two sub-questions are related to whether price developments at the wholesale level
meet certain conditions for an efficient use of infrastructure. Sub-question 2 refers to
transmission pipelines, whereas sub-question 3 tackles storage. Since econometric methods
provide for a powerful toolbox testing whether a market performs competitively or not and
why, the basic methodological approach is empirical modelling. From economic theory
hypotheses about certain relationships are deduced, one expects under the assumption of
14 The original idea of convenience yield was introduced by Kaldor (1939).
Introduction 9
perfect competition. Afterwards, the validity of these hypotheses is tested using actual
market data. In other words: observed market data is tested against a theoretical benchmark
indicating competitive market behaviour. Deviations from expected results hint at market
imperfections caused by a poorly set regulatory framework.
While the US natural gas market has been studied quite extensively using empirical
approaches (cf. e.g. De Vany and Walls 1993 and 1996, Serletis 1997, Serletis and Rangel-
Ruiz 2004, Modjtahedi and Movassagh 2005, Serletis and Shahmoradi 2006, Cuddington
and Wang 2006, Wei and Zhu 2006 as well as Mu 2007), literature for the European context
is quite rare. To the best of my knowledge, none of the specific sub-questions has been
tackled by existing literature so far. Therefore, the thesis aims at closing some lacunae in
empirical research of energy markets.
1.3 Outline of the thesis
Sub-question 1 is addressed in Chapter 2. As mentioned above, the European Commission
has intensively discussed the mandatory separation of natural gas transmission from
production and services. However, economic theory is ambiguous on the price effects of
vertical separation. Since some countries already have established ownership unbundling
(e.g. UK) while others have not (e.g. France), the question is analysed empirically applying
panel data analysis. In particular, the effect of ownership unbundling of gas transmission
networks as the strongest form of vertical separation on the level of end-user prices is
investigated. Different dynamic estimators as the system generalised method of moments
(system GMM, Blundell and Bond 1998) and the bias-corrected least-squares dummy
variable (LSDVC, Bruno 2005) estimator are applied on an unbalanced panel out of 18 EU
countries over 19 years. The chosen approach allows to avoid the endogeneity problem and
to estimate the long-run effects of regulation. A set of regulatory indicators is introduced such
as market entry regulation, ownership structure, vertical separation and market structure.
Furthermore, structural and economic country specifics are accounted for. Among these
different estimators, consistent results are produced. Ownership unbundling has no impact
on natural gas end-user prices, while the more modest legal unbundling reduces them
significantly. Moreover, third-party access, market structure and privatisation show significant
influence with the latter leading to higher prices.
Chapter 3 deals with the test of the spatial no arbitrage condition for the European natural
gas market (sub-question 2). In 2007, Germany changed network access regulation in the
natural gas sector and introduced a so-called entry-exit system. The re-regulation’s effect on
competitiveness remains to be examined. To study the development of natural gas spot
prices at two major trading hubs in Germany, cointegration analysis (Johansen 1988, 1991)
and a state space model with time-varying coefficients (Kalman 1960) are applied. To
analyse information efficiency in more detail, the state space model is extended to an error
correction model. Furthermore, German prices are compared to those at the nearby Dutch
Title Transfer Facility (TTF) hub, which thus serves as a kind of benchmark. Transaction
costs are explicitly considered. Overall, results suggest a reasonable degree of price
Introduction 10
convergence between the corresponding hubs. However, allowing for time-variant
adjustment processes, price differentials are only partly explained by transportation costs.
Persistent price differences indicate capacity constraints. Nonetheless, market efficiency in
terms of information processing has increased considerably.
Sub-question 3, the test of the intertemporal no arbitrage condition, is covered by Chapter 4.
A comparative analysis of the relationship between natural gas storage utilisation and price
patterns at three major European trading points is presented, the Dutch TTF, the Belgian
Zeebrugge and the National Balancing Point (NBP) in the UK. Due to limited data availability,
two indirect tests are used, developed by Fama and French (1987, 1988) and usually applied
in other commodity markets. To model market efficiency, the no arbitrage condition is
imposed as provided by the theory of storage. The results reveal that while operators of
European storage facilities realise seasonal arbitrage, substantial short-term potentials
remain unexploited. Thus, overall market performance differs substantially from the
competitive benchmark of the theory of storage.
The final Chapter 5 turns back to the leading research question whether regulatory efforts
throughout Europe have been successful in creating a framework fostering competition in the
natural gas sector. Results of the previous chapters are put into a broader policy context
trying to highlight what Europe has actually achieved on its way to a single internal gas
market and what the remaining areas are, regulators should concentrate on in the future. An
overview of the most recent developments in regulations is provided focussing on market
access conditions as well as on security of supply issues. Overall, regulatory efforts aiming at
improved wholesale competition are quite promising. The increased use of market-based
capacity allocation and congestion management procedures, as required by the third
legislative package, should reduce the risk of market foreclosure and boost liquidity. On the
other hand, attempts to address large-scale investments into gas infrastructure are not well
designed, leaving scope for further improvements.
Introduction 11
References
Baulch, B. (1997), Transfer Costs, Spatial Arbitrage, and Testing for Food Market Integration,
American Journal of Agricultural Economics 79(2), 477-487.
Baumol, W.J., J. Panzer and R. Willig (1982), Contestable Markets and the Theory of Industry
Structure, New York.
Blundell, R. and S. Bond (1998), Initial conditions and moment restrictions in dynamic panel data
models, Journal of Econometrics 87, 115–143.
BNetzA [Bundesnetzagentur] (2009), Monitoringbericht 2009, Bonn.
Brennan, M.J. (1958), The Supply of Storage, American Economic Review 48, 50-72.
Bruno, G.S.F. (2005), Approximating the bias of the LSDV estimator for dynamic unbalanced panel
data models, Economics Letters 87(3), 361-366.
Buehler, S. (2005), The Promise and Pitfalls of Restructuring Network Industries, German Economic
Review 6(2), 205-228.
Buehler, S, A. Schmutzler and M.-A. Benz (2004), Infrastructure quality in deregulated industries: is
there an underinvestment problem?, International Journal of Industrial Organization 22(2),
253-267.
Buehler, S., D. Gärtner and D. Halbheer (2006), Deregulating Network Industries: Dealing with Price-
Quality Tradeoffs, Journal of Regulatory Economics 30(1), 99-115.
Creti, A. (2009), The Economics of Natural Gas Storage – A European Perspective, Springer, Berlin.
Cuddington, J.T., and Z. Wang (2006), Assessing the degree of spot market integration for U.S.
natural gas: evidence from daily price data, Journal of Regulatory Economics 29(2), 195-210.
De Vany, A., and W.D. Walls (1993), Pipeline Access and Market Integration in the Natural Gas
Industry: Evidence from Cointegration Tests, The Energy Journal 14(4), 1-19.
De Vany, A., and W.D. Walls (1996), The Law of one Price in a Network: Arbitrage and Price
Dynamics in Natural Gas City Gate Markets, Journal of Regional Science 36(4), 555-570.
ERGEG [European Regulators Group for Electricity & Gas] (2010), ERGEG 2010 Status Review of the
Liberalisation and Implementation of the Energy Regulatory Framework, Document No. C10-
URB-34-04, December 7, Brussels.
ERGEG [European Regulators Group for Electricity & Gas] (2011), Assessment of Capacity Allocation
Mechanisms and Congestion Management Procedures for effective Access to Storage and
Proposals for the Amendment of the Guidelines of Good Practice of Storage System
Operators (GGPSSO), Document No. E10-GST-14-03, February 2, Brussels.
European Commission (2007), Inquiry pursuant to Article 17 of Regulation (EC) No 1/2003 into the
European gas and electricity sectors (Final Report), Communication from the Commission,
COM(2006) 851 final, January 10, Brussels.
Fama, E. (1970), Efficient Capital Markets: A Review of Theory and Empirical Work, Journal of
Finance 25(2), 383–417.
Fama, E.F. and K.R. French (1987), Commodity Futures Prices: Some Evidence on Forecast Power,
Premiums, and the Theory of Storage, The Journal of Business 60(1), 55-73.
Introduction 12
Fama, E.F. and K.R. French (1988), Business Cycles and the Behavior of Metal Prices. Journal of
Finance 63, 1075-1093.
Grewe, J. (2005), Auswirkungen der Liberalisierung auf die Erdgasspeicherung – Eine ökonomische
Analyse für den deutschen Erdgasmarkt, Münster.
IEA [International Energy Agency] (2009a), Natural Gas Information, Paris.
IEA [International Energy Agency] (2009b), Energy Prices & Taxes, Paris.
IEA [International Energy Agency] (2010), World Energy Outlook 2010, Paris.
Johansen, S. (1988), Statistical Analysis of Cointegration Vectors, Journal of Economic Dynamics and
Control 12(2/3), 231-254.
Johansen, S. (1991), Estimation and Hypothesis Testing of Cointegration Vectors in Gaussian Vector
Autoregressive Models, Econometrica 59(6), 1551-1580.
Kaldor, N. (1939), Speculation and Economic Stability, Review of Economic Studies 7, 1-27.
Kalman, R.E. (1960), A New Approach to Linear Filtering and Prediction Problems, Transactions of
the American Society of Mechanical Engineers, Journal of Basic Engineering, Series D 82(1),
35-45.
Knieps, G. (1997), Phasing out Sector-Specific Regulation in Competitive Telecommunications, Kyklos
50, 325-339.
Laffont, J.J. and J. Tirole (1993), A Theory of Incentives in Procurement and Regulation, Cambridge,
MA.
Modjtahedi, B. and N. Movassagh (2005), Natural-gas Futures: Bias, Predictive Performance, and the
Theory of Storage, Energy Economics 27, 617-637.
Mu, X. (2007), Weather, storage, and natural gas price dynamics: Fundamentals and volatility, Energy
Economics 29, 46–63.
Perry, M.K. (1989), Vertical integration, determinants and effects, in: Schmalensee, R. and R.D. Willig
(eds.), Handbook of Industrial Organization 1, Amersterdam, 183-255.
Rüster, S. (2010), Recent Dynamics in the Global Liquefied Natural Gas Industry, Working Paper, TU
Dresden.
Rüster, S. and A. Neumann (2009), Linking Alternative Theories of the Firm - A First Empirical
Application to the Liquefied Natural Gas Industry, Journal of Institutional Economics 5(1), 47-
64.
Serletis, A. (1997), Is there an East-West Split in North American Natural Gas Markets? Energy
Journal 18(1), 401-414.
Serletis, A., R. Rangel-Ruiz (2004), Testing for common features in North American energy markets,
Energy Economics 26(3), 401-414.
Serletis, A. and A. Shahmoradi (2006), Futures Trading and the Storage of North American Natural
Gas, OPEC Review 30(1), 19-26.
Vickers, J. (1995), Competition and Regulation in Vertically Related Markets, The Review of Economic
Studies 62(1), 1-17.
Introduction 13
Wei, S.Z.C. and Z. Zhu (2006), Commodity Convenience Yield and Risk Premium Determination: The
Case of the U.S. Natural Gas Market, Energy Economics 28, 523-535.
Williams, J.C. and B.D. Wright (1991), Storage and Commodity Markets, Cambridge.
Working, H. (1949), The Theory of Price of Storage, American Economic Review 39, 1254-1262.
Ownership unbundling 14
2 Ownership unbundling
Ownership unbundling of gas transmission networks – Empirical
evidence from a dynamic panel approach15
2.1 Introduction
Ownership unbundling has been one of the most important issues in the discussions
preceding the third legislative package for European energy markets. With ownership
unbundling as the strictest regulatory regime of vertical disintegration, the company that
owns and operates the transmission assets is fully separated from the rest of the system,
meaning that it has no further business activities in retail or production and import.16 On the
one hand, the European Commission strongly argued in favour of ownership unbundling for
gas and electricity transmission networks, preferring it to legal unbundling, which had been
the minimum standard in the previous Directive. Having investigated the gas and electricity
sectors, the Commission was dissatisfied with market outcomes, especially in the gas sector.
The status of unbundling of transmission system operators was identified as one major
obstacle to a well-functioning market environment (European Commission 2007, 2008). On
the other hand, several transmission system operators (TSOs) and countries (e.g. Germany
and France) opposed these efforts of the Commission, doubting the economic benefits and
raising juridical arguments against ownership unbundling. The package was adopted in
summer 2009 and has entered into force in 2011 allowing now for both legal and ownership
unbundling, and leaving the choice to Member States. Though a final agreement on the
package has been reached, the question as to whether ownership unbundling is superior to
legal unbundling remains on the political agenda. Therefore, the problem has not been
solved.
From an economic point of view, ownership unbundling would be preferable if it increased
net social welfare. An intuitive indicator for a positive welfare effect is if the introduction of
ownership unbundling leads to lower end-user prices.17 Economic theory gives little
guidance as ambiguous results are reported (cf. e.g. Laffont and Tirole 1993, Vickers 1995,
Buehler 2005, Höffler and Kranz 2011). On the one hand, it is argued that the stricter vertical
separation is designed, the less incentive a network operator has to discriminate between
affiliated companies and third parties. Thus, competition on down- and upstream markets will
be fostered and the risk of vertical foreclosure reduced. On the other hand, ownership
unbundling may lead to a loss of economies of scope and thus of operational efficiency. The
15 This chapter draws on Growitsch and Stronzik (2011). An earlier version of the paper has been presented at
the 15th
International Conference on Panel Data and the 2009 Verein für Socialpolitik conference. 16 The weakest form is accounting unbundling, meaning that the utility has to set up separate accounts only for
different services. Legal unbundling requires that the different services have to be operated by separate companies which can still belong to the same owner.
17 Please note that lower end-user prices are not a sufficient condition for higher social welfare as investments
may be negatively affected (cf. e.g. Buehler et al. 2004, 2006). For the purpose of this paper, we concentrate on retail prices, however.
Ownership unbundling 15
net welfare effects, however, are sensitive to model assumptions and depend on, for
instance, assumptions about the intensity of competition in the downstream market, on the
demand function, on the role of investments or on the design of access price regulation.
Thus, the impact of vertical separation remains an empirical question. Since some countries
have already established ownership unbundling (e.g. UK) and others have not (e.g. France),
the question can be answered empirically. For that, we construct a panel of 18 EU18
countries spanning a time interval of 19 years. In order to isolate the effect of ownership
unbundling, we gather data from national regulators and introduce a dummy variable which
explicitly differentiates ownership unbundling from any other form of vertical integration or
separation. In addition, we analyse the effect of other regulatory instruments as the degree of
market opening, public or private ownership of the transmission system operator (TSO),
negotiated or regulated third-party access (TPA) and whether any restrictions to market entry
exist upstream. As regulatory reforms might take some time to become price effective (e.g.
Joskow, 2008), we also calculate the dynamic effects in terms of long run multipliers.
Additionally, we control for differences among countries’ economic performance, the
structure of the gas sector (e.g. gas export and import) and the oil price, as in several
countries gas contracts are indexed to oil prices.
In order to analyse the effect of ownership unbundling on retail prices19, we face various
econometric challenges, i.e. (I) cluster correlation, (II) a potential endogeneity bias and (III) a
small sample bias. Therefore, we (I) apply a static fixed effects estimator with robust
standard errors (Froot 1989 and Williams 2000), and (II) the System Generalised Method of
Moments estimator (system GMM, Blundell and Bond 1998) which uses lagged variables as
instruments. (III), a bias-corrected least-squares dummy variable estimator (LSDVC, Bruno
2005) allows us to compensate for the small sample bias in Instrumental Variable estimation.
This threefold estimation strategy ensures more robust results than a single estimator
strategy.
Our paper contributes to the literature analysing the effects of reforms in energy regulation
on final customer prices in three ways. First, this paper is the first that explicitly models
ownership unbundling. Second, in contrast to comparable previous studies we address the
potential endogeneity problem of regulatory reforms. Third, our paper is the first to identify
the long-run effects of reforms by calculating dynamic multipliers, addressing the natural lag
between the introduction of a regulatory regime and its effect on end-user prices.
The remainder of the paper is organised as follows. We start with an overview of existing
studies that empirically assess regulatory reforms in the energy sector. Section 2.3 presents
the dataset and some descriptive analyses with a focus on regulatory indicators. The applied
econometric approaches and the results are discussed in Section 2.4. Finally, we conclude.
18 In an earlier version of this paper we included other OECD countries as well. By interviews of the designated
regulatory authorities it turned out that none of these countries has introduced ownership unbundling for gas TSOs, but show ambiguous regulatory structures. The special case of the US is discussed in Appendix 1.
19 We use the terms end-user prices for household customers and retail prices synonymously throughout the
text.
Ownership unbundling 16
2.2 Related literature
A change of the regulatory framework of an industry, such as the introduction of ownership
unbundling in the gas sector, is justified if it improves social welfare. In many countries, both
energy sectors – electricity and gas – have passed through a radical liberalisation process.
Countries have chosen different approaches not only with regard to implemented measures
but also with regard to speed and timing. However, empirical studies on the price effect of
the liberalisation process in the energy sectors are rather rare (an overview is provided in
Table 1). The three studies on gas will be at the centre of the following discussion.
Copenhagen Economics (2005) has been the first investigating the determinants of natural
gas end-user prices. Copenhagen Economics has developed its own indicator capturing
regulatory reforms, the so-called Market Opening Index (MOI) for 14 European countries.
The indicator is scaled between 0 and 1, with 1 indicating full market opening. The
explanatory variables are modelled with a lag of one year and a fixed effects estimator is
applied. Privatisation and competitive tariff structures tend to decrease end-user prices.
However, regulation of these prices shows an increasing effect, quite the opposite of what is
usually intended. The indices which cover information on the unbundling of networks are not
significant. Due to the high level of aggregation, it is difficult to identify effects of a single
regulatory measure like the introduction of ownership unbundling. Therefore, results remain
vague. Furthermore, the construction of the indicator lacks transparency and seems to be
based – at least partly – on individual judgements.
Like the majority of studies of liberalisation in developed countries (OECD or EU), a recent
paper by Brau et al. (2010) relies on the OECD database, indicators of regulation in energy,
transport and communications (ETCR), as it is publicly available and provides for
consistency. This database covers a broad range of regulatory areas, i.e. entry regulation,
public ownership, vertical integration and market structure. Each of these four main
indicators consists of three sub-indicators.20 As the weighting procedure as well as the
coding provided by the OECD data is somehow arbitrary and results are sensitive to these
issues, Brau et al. advocate the application of sub-indicators and the adjustment of the actual
coding in light of the specific research question to be analysed while most previous studies
rely on the aggregate indicators and the OECD coding.21 The authors compare two sources
for residential end-user prices in the gas sector – Eurostat and the International Energy
Agency (IEA) – and find hardly any evidence of beneficial effects of regulatory measures on
European end-user prices. Instead, privatisation tends to increase prices. However, Brau et
al. neglect two important sub-indicators: the existence of market barriers for entrants and the
20 The OECD database provides regulatory indicators for seven network industries, i.a. electricity and gas, for
29 OECD countries. The overall structure of the indicators is the same across sectors. Concerning gas, each of the 12 sub-indicators is coded 0, 3 or 6 with the highest score of 6 indicating the most restrictive conditions regarding competition. For aggregation equal weights are used except for the indicator for vertical integration. See Conway and Nicoletti (2006) for further details.
21 Looking at two earlier studies on the electricity sector by Steiner (2001) and Hattori and Tsutsui (2004), with
the latter being more or less a re-evaluation of the former, gives some indication regarding this line of argument. While Hattori and Tsutsui define the unbundling variable to be 1 if legal unbundling has been established, Steiner already regards accounting unbundling as a form of vertical separation leading to the detection of reverse effects for unbundling and the introduction of a power exchange.
Ownership unbundling 17
regime of third-party access (TPA), both of which are crucial for market entry conditions.
Therefore, estimations may face an omitted variable problem. Furthermore, the authors rely
on the three OECD sub-indicators of vertical integration which do not explicitly cover
ownership unbundling of TSOs. This will be discussed in the next section.
Concerning the empirical methods applied, Table 1 shows that dynamic panel approaches
like GMM or LSDVC have become popular as they are generally capable of dealing with the
endogeneity problem of dynamic price processes. Reforms of the regulatory framework not
only influence end-user prices or investments, they might also be driven by these
variables.22 Indeed, only Cambini and Rondi (2010) effectively model regulatory indicators as
endogenous regressors in the dynamic setting.23 Their approach is different to ours,
however, as they concentrate on the influence of incentive regulation on investment
behaviour using firm-level data, not controlling for ownership unbundling or focussing on end-
user prices. Also, they do not report the long run effects of regulation.
To sum up, economic theory reports ambiguous effects concerning the pros and cons of
ownership unbundling which calls for an empirical assessment. To date, no empirical study
has isolated the effect of ownership unbundling and evaluated its long-run effect on natural
gas retail prices. With this study we fill this gap by trying to support these discussions with
empirical evidence and answering the question as to whether or not this action has brought
about beneficial outcomes.
22 Regarding electricity, this has been explicitly analysed by Nagayama (2009). He defines one single indicator,
the so-called liberalisation model, subsuming information on various areas of regulation, such as e.g. price regulation and consumer protection. Applying an ordered probit model he shows that the selection of a certain liberalisation model by a country is influenced by electricity prices.
23 Brau et al. (2010), for example, apply system GMM, but only lagged dependent variables are instrumented.
Potential endogeneity of regulatory indicators is not tackled.
Ownership unbundling 18
Table 1: Overview of empirical studies
Study Coverage List of Variables Empirical Model
1) Success Factors
2)
Countries Sector(s) Time Dependent Regulatory Indicators
Steiner (2001) 19 OECD Electricity 1986-1996
End-user prices industry
Ratio of end-user prices
Sector performance (capacity utilisation)
OECD database3)
Power Exchange
Time to privatisation
Time to liberalisation
Factor analysis
FE, RE
Price levels
Third-party access (-)
Power exchange (-)
Unbundling (-)
Hattori and Tsutsui (2004) 19 OECD Electricity 1987-1999
End-user prices industry
Ratio of end-user prices
OECD database3)
Power exchange
Time to privatisation
Time to liberalisation
FE, RE
Price level
Market opening (-)
power exchange (+)
Unbundling (+)
Alesina et al. (2005) 21 OECD Utilities, communications and transport
1975-1998
Investments
Capital stock
OECD database3)
Lagged, squared
Interaction terms
Difference GMM (Arellano-Bond 1991)
Entry liberalisation (+)
Privatisation (+)
Increasing returns
Copenhagen Economics (2005)
14 EU Electricity and gas 1993-2003
End-user prices industry MOM database
Lagged
Factor analysis
FE
Privatisation (-)
Tariff structure (-)
Regulation of end-user prices (+)
Zhang et al. (2008) 36 developing countries
Electricity 1985-2003
Generation per capita
Installed capacity per capita
Generation per employee
Capacity utilisation
Private ownership (generation)
Market share of three largest generators
Regulatory governance index
Interaction terms
FE Market share of three
largest generators (+)
Nagayama (2009)
78 developed and developing countries
Electricity 1985-2003
Electricity price
Households
Industry
Liberalisation model
Ordered probit
FE, RE
IV
Power price drives liberalisation
Liberalisation (+)
Brau et al. (2010) 12 EU Gas 1991-2007
End-user prices households
Eurostat
IEA
OECD database3)
System GMM
(Blundell-Bond 1998)
Ambiguous
Privatisation (+)
Cambini and Rondi (2010) 5 EU Electricity and gas (firm-level data; 23 TSOs and DSOs)
1997-2007
Investment rate
Incentive regulation
X-factor
Allowed cost of capital
FE
2SLS
Difference GMM (Arellano-Bond 1991)
Incentive regulation (+)
X-factor (-)
WACC (+)
Sen and Jamasb (2010) 19 Indian states
Electricity 1991-2007
Efficiency (plant load, network losses)
End-user prices industry and households
Independent regulatory body
Unbundling of networks
Open network access
Privatisation
LSDVC (Bruno 2005) Regarding prices: open
network access (-)
Nillesen and Pollitt (2011) New Zealand Electricity (firm-level data on 28 DSOs)
1995-2007
Unit operational costs Ownership Unbundling FE, RE Ownership unbundling
(-)
Source: Own compilation. 1)
Estimators: FE (fixed effects), RE (random effects), GMM (generalised method of moments), IV (instrumental variable), 2SLS (two stage least squares) and LSDVC (bias-corrected least-squares dummy variable).
2) Main identified effects. (+) signals an increasing and (-) a decreasing effect on the dependent variable(s).
3) OECD indicators of regulation in energy, transport and communications (ETCR).
Ownership unbundling 19
2.3 Data
In order to analyse the effect of regulatory reforms on end-user prices with a special focus on
ownership unbundling we construct a panel consisting of 18 EU countries encompassing
1989 through 2007. While the starting date has been chosen because no major changes in
gas market regulations have occurred before 1990 in the considered countries, the choice of
countries has been made for data availability reasons.24
As market liberalisation is a step-wise process, typically starting with opening up the segment
of large industrial customers for competition and ending with residential consumers (Joskow
2008), we focus on end-user prices for households since they indicate the effects of
regulatory reforms best. We use net-of-tax natural gas prices in real terms of the database
Energy Prices & Taxes published by the International Energy Agency (IEA) expressed in
USD2000/107 kcal. Figure 4 shows the price dynamics in the considered EU countries with
residential end-user prices being compared to industry prices (for descriptive statistics see
Appendix 2).25
Figure 4: End-user prices [USD2000/107 kcal]
Source: Own calculations based on IEA price data.
24 The analysis has an EU focus and is based on the OECD database, indicators of regulation in energy,
transport and communications (ETCR) that provides information for 19 EU Member States (out of a whole sample of 29 OECD countries). With no data on household gas prices available for Sweden leaves us with 18 countries.
25 All price data will enter estimations in log-form. The absolute values in Figure 4 are for visualisation
purposes only.
0
500
1000
0
500
1000
0
500
1000
0
500
1000
200520001990 1995 1990 1995 2000 2005
1990 1995 2000 2005 1990 1995 2000 2005 2000 20051990 1995
Austria Belgium Czech Republic Denmark Finland
France Germany Greece Hungary Ireland
Italy Luxembourg Netherlands Poland Portugal
Slovak Republic Spain United Kingdom
Price households Price industry
year
Ownership unbundling 20
The overall picture across countries is rather heterogeneous. Not only do price levels differ
substantially, but the development over time also varies considerably. Southern European
countries, i.e. Greece, Italy, Portugal and Spain, are among the ones with the highest
residential end-user prices and the largest gap between households and industry, though the
difference in most of them has diminished over time. While almost all Member States have
higher residential end-user prices than for industry, the eastern European countries that
joined the EU in 2004, reported the opposite at the beginning of the period.
In order to account for changes in regulatory regimes of the considered countries we rely on
the OECD indicators for regulation in energy, transport and communications (ETCR) as it
secures data consistency. The ETCR dataset provides yearly time series up to 2007.
Following the arguments of Brau et al. (2010) we use information of the sub-indicators
instead of applying the main indicators: besides the more or less arbitrarily chosen weighting
structure26, using aggregate indicators leads to a loss of information. Moreover, effects of a
distinct regulatory measure cannot be identified. Since this paper analyses the effects of
ownership unbundling, we concentrate in the setup of our regulatory indicators on the
information provided by the basic questions underlying the overall OECD indicator and
amend this data where necessary. With regard to the coding of the ETCR dataset, Brau et al.
(2010) point to the problem of the arbitrary cardinalisation of often categorical variables into
the 0-6 scale that may influence results. Therefore, if not otherwise mentioned, we apply
discrete variables instead of the 0-6 scale. All regulatory indicators are coded in a way that
higher values correspond with conditions usually being assumed to be more restrictive
regarding competition. Thus, positive signs are expected in the estimations.
A first set of indicators deals with conditions of market entry. Concerning third-party access
to the transmission network, the discrete variable TPA is introduced with 0 indicating
regulated access, 1 negotiated access and 2 otherwise. In order to measure the effect of
giving customers access to alternative suppliers so that they have the choice among different
offers, we create the continuous variable Liberal. As information on the actual degree of
market opening is provided, the variable is designed as 100% minus the given percentage.
An increasing share of customers who can choose among several suppliers should lead to
an increase of the competitive pressure in this segment resulting in lower end-user prices.
Limitations on access to production or import markets are covered by the discrete variable
Barriers.27
For the privatisation efforts of the various countries ETCR provides for three variables
indicating the percentage of the sector owned by the government. As these indicators are
highly correlated and in order to circumvent multicollinearity, the given information is
condensed to one single variable Ownership by simple averaging and discretising thereafter.
Since private entities are generally assumed to perform more efficiently, the indicator is 0 if
26 See Footnote 20. 27 In particular, it is asked if any regulations are in place that restrict the number of competitors allowed to
operate a business in at least some upstream markets.
Ownership unbundling 21
the public share is less than 25%, 1 if the share is below 50%, 2 if it is less than 75% and 3
otherwise.28
The main issue of this paper – ownership unbundling of gas TSOs – is not covered by the
OECD database. Therefore, we create a new dummy OU_TSO indicating whether a country
has introduced ownership unbundling (OU_TSO = 0) or not (OU_TSO = 1). While five
countries have established ownership unbundling after having legally unbundled their gas
TSOs first, Poland and Portugal directly shifted from vertically integrated utilities to ownership
unbundling. Therefore, to distinguish whether lower end-user prices are caused by the most
rigid form of vertical separation or just by breaking up formerly vertically integrated gas
utilities, we introduce a second dummy VI_TSO. This variable becomes 0 if at least legal
unbundling has been implemented and 1 otherwise.29 From the OECD indicators for vertical
separation only the one for gas distribution is added to the variable list (VI_DSO).30 Since
only two characteristics are actually reported in the ETCR dataset, the variable spoils down
to a dummy with 0, indicating a (not further specified) form of separation of gas distribution
system operators (DSOs) and 1 otherwise.
The last set of indicators deals with market structures in the competitive parts of the value
chain, i.e. production/import (MS_Prod) and supply (MS_Supp).31 ETCR provides for ranges
indicating the importance of the largest player. The corresponding two variables are coded
as follows: a 0 indicates that the largest utility has a share of less than 50 % in the relevant
sector, 1 with a share between 50 and 90%, and 2 otherwise. We expect retail prices to
increase with market concentration.
In order to avoid misspecifications of our model, we consider several control variables which
may influence end-user prices for households, i.e. sector-specific factors such as the total
amount of gas supplied to customers, the indigenous gas production, natural gas imports
and exports as well as macroeconomic indicators like the GDP, all expressed as per capita
variables. Due to the close relation between oil and gas, the oil price is additionally taken into
account. We use West Texas Intermediate (WTI) prices to map the oil price because gas
prices are expressed in USD and WTI has been widely applied in international comparisons
(cf. e.g. Brown and Yücel 2007).
Concerning the development of natural gas end-user prices, it can often be observed that
industry prices adjust faster to these sector-specific and macroeconomic factors than
residential end-user prices. Therefore, industry prices may function as a kind of transmission
28 A correlation matrix for the various indicators, including the four main indicators of the ETCR dataset, can be
found in Appendix 2. The created variable Ownership is highly collinear to each of the original three OECD sub-indicators for public ownership.
29 The information on both dummy variables has been collected from various benchmarking reports of the
European Commission and checked by personal interviews with the designated regulated authorities. Both variables show no problematic correlation. Details on both dummies are provided in Appendix 2.
30 The other two ETCR indicators covering the separation of the production and the supply segment are
dropped for two reasons. First, correlations with indicators for entry regulation are rather high facing the risk of multicollinearity. Second, vertical separation is usually discussed in the sense of essential facilities with the potential of being a natural monopoly. While this is true for gas pipelines (transmission and distribution), production and supply are clearly competitive parts of the value chain. See e.g. Gordon et al. (2003).
31 The third category of the OECD, the market structure in gas transmission, has been neglected due to the –
at least - questionable relevance and data inconsistencies.
Ownership unbundling 22
belt between these factors and the prices for households. An auxiliary fixed effects
regression reveals that industrial prices are indeed significantly influenced by per-capita
GDP, the oil price and per-capita gas production but not by residential end-user prices and
per-capita natural gas import.32 Therefore, natural gas imports as well as end-user prices for
industry are chosen as control variables with the latter serving as a kind of instrument for
GDP and natural gas production. Regarding the industrial gas prices, a negative (and
significant) coefficient would indicate a higher demand elasticity, whereas a positive sign
would just show a level effect on residential end-user prices.33 However, as the oil price is of
special importance for the gas market, it additionally enters the set of control variables. Not
only are several gas contracts in Europe still indexed to oil, but oil is also a close substitute
for gas in the heating market; this is of particular interest for the residential end-user market.
For annual data it is quite common to use also the one-year lagged oil price as the
adjustment process of gas prices is lagged by around six months (cf. e.g. Siliverstovs et al.
2005).
2.4 Empirical analysis
To analyse the effect of ownership unbundling on retail prices, certain econometric problems
have to be tackled. The first one is that longitudinal price data might be biased by
autocorrelation or cluster-correlation. Approaches that assume independence of
observations – like the traditional fixed effects and random effects estimator – generally
underestimate the true variance leading to inflated t-statistics. To avoid possibly
corresponding inconsistency problems we use a robust variance estimator (cf. Froot 1989
and Williams 2000). This estimator allows for heteroscedasticity, both between and within
clusters, and for serial correlation.
A second econometric challenge is that causality between the institutional change in
regulation and prices might not be unique, but reciprocal. Then, explanatory variables are
correlated with the error term (endogeneity problem). In order to address endogeneity, we
apply the System Generalised Method of Moments estimator (system GMM, Blundell and
Bond 1998), which uses lagged (endogenous) variables as instruments. We model all
regulatory indicators as endogenous variables.34 Furthermore, we allow gas imports to be
endogenous as causality between the dependent variable and imports can be assumed to be
bidirectional. Using the standard treatment for endogenous variables, lags of order two and
higher for the transformed equation and lag one for the levels equation are specified (cf.
Roodman 2006, 2008). The remaining regressors are treated as strictly exogenous
instrumenting themselves.
32 A preceding correlation analysis shows that gas supply, production as well as exports are highly collinear.
Therefore, only the production variable enters the regression. Results for both, the correlation analysis as well as the regression, are presented in Appendix 2.
33 This design also captures the effect, if industrial end-user prices gain more from regulatory reforms (higher
demand elasticity) which has been shown by Steiner (2001) and Hattori and Tsutsui (2004) for electricity. 34 Modelling the indicators strictly exogenous leads to a rejection of the Sargan test indicating that regulatory
reforms – at least partly – have been driven by high end-user prices for households.
Ownership unbundling 23
Third, our data set is rather small, as it covers only 18 countries. Nickel (1981) has shown
that the conventional least-squares dummy variable (LSDV) estimator for dynamic panels is
inconsistent for a finite time horizon T and a large number of cross-sectional dimensions N
(Nickell 1981). Furthermore, with only 18 countries we likely face the problem of a small
sample bias. Therefore, consistent Instrumental Variable (IV) estimators like Anderson and
Hsiao (1982) and Generalised Method of Moments (GMM) estimators like Arellano and Bond
(1991) might be biased (Kiviet 1995). Kiviet (1995, 1999) as well as Bun and Kiviet (2003)
provide for techniques to approximate the small sample bias but deliver consistent results
only for balanced panels. With the unbalanced nature of our panel, adopting those
corrections would lead to a high loss of information as they would in fact require discarding
the cross-sections or time series causing the imbalance. Therefore, we apply the corrected
least-squares dummy variable (LSDVC) estimator by Bruno (2005) whose approach avoids
the aforementioned shortcomings. The small sample bias is approximated via bootstrapping.
For our empirical analysis, we model the log end-user price for residential costumers yit of
country i and period t as a function the regulatory indicators Rit, of the current and one year
lagged oil price Xit, and of other control variables Zit including industry prices and gas
imports. ηi captures unobserved heterogeneity across countries and εit is the error term
satisfying the usual assumptions.
A dynamic setting additionally allows for some gas price dynamics by introducing one and
two years lagged end-user prices as further explanatory variables.35
(2-1) itiitititititit ZXRyyy 2211
.
Given the lagged endogenous variables, we can calculate the long run effect of regulation on
household prices in a partial adjustment model:
(2-2) )1/( 21
* .
The results of our three model specifications evaluating the impact of regulatory reforms, and
ownership unbundling in particular, on end-user prices of residential customers are listed in
Table 2. The parameter values indicate the short-run effects (impact multipliers). Overall, we
obtain robust results throughout all three estimators. In none of our models does ownership
unbundling (OU_TSO) seem to have any significant effect on retail prices.36 The same holds
for vertical separation of distribution networks (VI_DSO). However, countries which at least
legally unbundled transmission networks from the rest of the value chain (VI_TSO) indeed
show lower retail prices. This holds especially true on the long run, where the effect
approximately doubles. Apart from that, only one regulatory variable, third party-access
(TPA) has the expected sign. Countries without regulated TPA have significantly higher
natural gas prices than others. Again, this effect is twice as high in the long run.
35 A twice lagged dependent variable is required to obtain valid results of the Arellano-Bond AR(2) test for the
system GMM and LSDVC estimator to prevent autocorrelation in the error terms.
36 One might argue that the result is mainly driven by the low within variation of the dummy. The application of a robust variance random effects estimator, which additionally takes into account cross-sectional variation, leads to the same result: OU_TSO remains insignificant and VI_TSO significant.
Ownership unbundling 24
Table 2: Natural gas end-user prices and the impact of regulatory reforms
Log Price_hh Fixed effects One-step system GMM LSDVC (BB)
Log Price_hht-1
0.684***
(0.061)
0.682***
(0.056)
0.709***
(0.067)
Log Price_hht-2 -0.180***
(0.023)
-0.170***
(0.023)
-0.183***
(0.046)
TPA
Long-run multiplier
0.045**
(0.020)
0.092**
(0.036)
0.047**
(0.019)
0.097**
(0.035)
0.047**
(0.023)
0.098**
(0.049)
Liberal
Long-run multiplier
0.023
(0.042)
0.047
(0.086)
-0.029
(0.045)
-0.059
(0.031)
-0.018
(0.041)
-0.038
(0.086)
Barriers
Long-run multiplier
-0.003
(0.017)
-0.007
(0.034)
-0.005
(0.018)
-0.010
(0.036)
-0.003
(0.020)
-0.006
(0.043)
Ownership
Long-run multiplier
-0.033*
(0.017)
-0.067*
(0.035)
-0.033*
(0.018)
-0.067*
(0.037)
-0.026*
(0.015)
-0.056*
(0.032)
VI_TSO
Long-run multiplier
0.082**
(0.028)
0.165***
(0.051)
0.082**
(0.029)
0.167***
(0.054)
0.082**
(0.032)
0.173**
(0.072)
OU_TSO
Long-run multiplier
0.052
(0.030)
0.104
(0.060)
0.049
(0.032)
0.100
(0.066)
0.042
(0.040)
0.089
(0.083)
VI_DSO
Long-run multiplier
0.009
(0.014)
0.018
(0.029)
0.008
(0.016)
0.017
(0.132)
0.008
(0.030)
0.017
(0.063)
Log Oil 0.168***
(0.036)
0.284***
(0.069)
0.164***
(0.046)
Log Oilt-1 0.109**
(0.038)
-0.052
(0.129)
0.102**
(0.039)
MS_Prod 0.123***
(0.041)
0.112**
(0.044)
0.108**
(0.051)
MS_Supp -0.080*
(0.042)
-0.073
(0.050)
-0.075**
(0.034)
Import_pc -0.111*
(0.057)
-0.126*
(0.068)
-0.113
(0.080)
Log Price_ind 0.097**
(0.041)
0.104***
(0.033)
0.102**
(0.047)
Year dummies yes yes yes
R squared (within) 0.852 -- --
Arellano-Bond AR(2) test -- 0.339 0.771
No. of observations 195 195 195
Notes: Standard errors in parentheses are robust to heteroskedasticity and to within group serial correlation. LSDVC: initialised with the Blundell and Bond (1998) estimator (BB) and standard errors calculated via bootstrapping (1,000 runs); AR(2) test with regard to first step BB estimation. Standard errors for long run effects have been calculated with the delta method. ***, **, * denote significance at 1%, 5% and 10%.
Ownership unbundling 25
Neither market liberalisation (Liberal) nor the existence of limitations on access to production
or import markets (Barriers) have significant price effects. Interestingly, a higher share of
publicly owned companies (Ownership) reduces natural gas retail prices. This result
corresponds to previous empirical studies (e.g. Alesina et al. 2005 and Brau et al. 2010).
Looking at the control variables, the oil price is significant in all estimations. The positive sign
is consistent with expectations as the oil price should control for the fact that many contracts
still link the gas price to the development of the oil price. The lagged adjustment process
embedded in long-term gas contracts is represented by the significance of the lagged oil
price. The overall effect, which is here the sum of Log Oil and Log Oilt-1, is of similar
magnitude across all approaches. End-user prices are raised by about 2.8%if the oil price
increases by 10%.
Concerning effects of market structures, the results for our market concentration variables
are ambiguous. As expected, prices rise with an increase in market concentration on the
wholesale level (MS_Prod). At the same time, market concentration in natural gas service
(MS_Supp) tends to reduce retail prices. This counterintuitive result is caused by two effects.
First, countries with relatively low end-user prices at the beginning of our observation period
report higher initial concentration rates (e.g. eastern European countries). Second, a
reduction of MS_Supp quite often coincides with an increase in retail prices (e.g. France,
Hungary, Ireland and The Netherlands). However, results should be treated with care as this
variable exhibits only a very low within-variation. Import volumes (Import_pc) influence retail
prices in 2 out of 3 models as expected: prices drop with increasing supply. However, the
variable is only weakly significant. The trade-off between industrial and residential end-user
prices as shown by Steiner (2001) as well as by Hattori and Tsutsui (2004) cannot be
confirmed in our estimation. Indeed, the positive signs of the variable Log Price_ind indicate
that high industrial prices go along with higher household prices.
To sum up, all three of our estimators, the FE estimation with robust standard errors, the
system-GMM and the dynamic LSDVC, lead to similar results. None of the approaches find a
significant effect of ownership unbundling on the level of end-user prices for residential
customers.
2.5 Conclusions
Ownership unbundling of gas transmission system operators (TSOs) as the most rigid form
of vertical disintegration from competitive and regulated operations is still an issue in the
discussions about a suitable future regulatory framework of the European gas market. The
controversial debate preceding the third legislative package has led to a compromise that
allows now for both legal and ownership unbundling, and leaves the choice to Member
States. However, the problem of whether or not the rigid ownership unbundling is
economically more beneficial than other forms of unbundling has not yet been solved. While
economic theory reports ambiguous results, we have answered the question empirically by
analysing the developments of natural gas end-user prices in different countries of which
Ownership unbundling 26
some already have established ownership unbundling (e.g. UK) while others have not (e.g.
France).
Analysing an unbalanced panel of 18 EU countries over 19 years with a number of static as
well as dynamic estimators reveals no evidence for a price-decreasing effect of ownership
unbundling. However, the breaking-up of formerly vertically integrated TSOs with at least
introducing the more modest legal unbundling has resulted in lower end-user prices.
Furthermore, third-party access and privatisation show significant influence with the latter
leading to higher price levels. Their effect even doubles in the long run. These results are
consistent across our estimations.
From a policy point of view, our results do not support a further separation of the different
stages of the natural gas value chain. On the contrary, as countries which at least legally
separated transmission networks already seem to have reaped the economic benefits, a
further tightening of unbundling rules for gas TSOs does not seem to be economically
reasonable.
However, as most of the countries in our analysis have established ownership unbundled
gas TSOs only very recently, our results should be treated with care. In order to arrive at a
clearer picture on this issue a re-evaluation might be carried out in few years. An interesting
topic for further research could be the empirical analysis of the impact of different forms of
vertical separation of transmission networks on the investment level in the European gas
sector. Regarding the dynamic efficiency, a debate is currently going on that is very similar to
the one that has been at the centre of this paper.
Ownership unbundling 27
Appendix 1: The US case
Regarding the European debate on TSO unbundling, some authors mention the US
interstate pipelines as an example for ownership unbundling in the gas sector and refer to
FERC Order No. 636 of 1992 (cf. e.g. Glachant 2011: 60f.). Therefore, neglecting the US
might bias results presented in the core text underestimating the effect of ownership
unbundling.
Without any doubt, FERC Order No. 636 can be regarded as one of the major regulations
towards a competitive gas market in the US, as it took away the market power of existing
interstate pipelines and effectively removed them from having any dominant role in gas
commodity markets (cf. e.g. Jensen 2007). Furthermore, interstate pipelines were separated
in ownership terms from gas distribution early on, initiated by the Holding Company Act of
1935.37
With regard to FERC Order No. 636, the Federal Energy Regulatory Commission (FERC)
directed pipeline gas marketing affiliates to transfer title to gas sales upstream at so-called
pooling points. Downstream of these pooling points, i.e. in the actual pipeline, all gas is
owned by shippers. This means that the operator of an interstate pipeline (or any affiliated
company) no longer owns the gas transported through his trunk line. On the other hand,
affiliates of interstate pipeline companies are still allowed to transport gas through other
interstate pipelines not owned by that company. Thus, even the US have not implemented
full ownership unbundling in European terms. Several gas holdings still control both,
interstate pipelines as well as supply companies. Except for distribution, only legal
unbundling is required (cf. e.g. Ascari 2011). The Notice of a proposed rulemaking, issued by
FERC on April 7, 2011, might serve as a kind of additional proof. This Notice is concerned
with the bidding of interstate pipeline affiliates in open season procedures for pipeline
capacity (see FERC 2011).
In the European context, the application of the provisions of FERC Order No. 636 would
constitute an additional element of the functional unbundling rules and might best be
classified as a kind of “pipeline-specific ownership unbundling”.38
Even if one follows the arguments of the authors referring to US interstate pipelines as an
example for ownership unbundling and attaching the US to the panel, overall results do not
change significantly. As presented in Table 3, the variable VI_TSO for legal unbundling
remains significant, while ownership unbundling (OU_TSO) still has no impact on natural gas
end-user prices.39
37 For a historical overview of US gas market liberalisation see e.g. Makholm (2007). 38 For further information with regard to the role of functional unbundling in the European context see
European Commission (2010). 39 Lung-run multipliers are omitted.
Ownership unbundling 28
Table 3: Natural gas end-user prices and the impact of regulatory reforms
(including the US)
Log Price_hh Fixed effects One-step system GMM LSDVC (BB)
Log Price_hht-1
0.680***
(0.061)
0.679***
(0.056)
0.717***
(0.070)
Log Price_hht-2 -0.175***
(0.025)
-0.166***
(0.025)
-0.183***
(0.057)
TPA
0.028
(0.020)
0.029
(0.020)
0.032*
(0.018)
Liberal
0.036
(0.041)
0.045
(0.046)
0.027
(0.039)
Barriers
-0.004
(0.019)
-0.006
(0.021)
-0.003
(0.023)
Ownership
-0.037**
(0.017)
-0.036*
(0.018)
-0.029**
(0.014)
VI_TSO
0.072**
(0.027)
0.069**
(0.030)
0.073**
(0.036)
OU_TSO
0.024
(0.032)
0.020
(0.034)
0.015
(0.039)
VI_DSO
0.004
(0.020)
0.002
(0.021)
0.003
(0.025)
Log Oil 0.145***
(0.036)
0.245***
(0.072)
0.141***
(0.038)
Log Oilt-1 0.065
(0.048)
-0.077
(0.117)
0.057
(0.038)
MS_Prod 0.112***
(0.035)
0.103**
(0.038)
0.095**
(0.047)
MS_Supp -0.074*
(0.041)
-0.070
(0.048)
-0.068**
(0.032)
Import_pc -0.091
(0.054)
-0.105
(0.068)
-0.096
(0.074)
Log Price_ind 0.108**
(0.040)
0.114***
(0.033)
0.116**
(0.049)
Year dummies yes yes yes
R squared (within) 0.841 -- --
Arellano-Bond AR(2) test -- 0.152 0.528
No. of observations 212 212 212
Notes: Standard errors in parentheses are robust to heteroskedasticity and to within group serial correlation. LSDVC: initialised with the Blundell and Bond (1998) estimator (BB) and standard errors calculated via bootstrapping (1,000 runs); AR(2) test with regard to first step BB estimation. ***, **, * denote significance at 1%, 5% and 10%.
Ownership unbundling 29
Appendix 2: Descriptive statistics
Table 4: Descriptive statistics and information on the unbundling status of gas TSOs
Country Log Price_hh a Log Price_ind
a Log Oil
b Import_pc
c
TSO Unbundling d
Legal Ownership
Austria 5.896 (0.142)
5.121 (0.086)
-- 0.703 (0.169)
2002 --
Belgium 5.917 (0.113)
5.004 (0.156)
-- 1.161 (0.218)
2001 --
Czech R. 6.037 (0.234)
6.099 (0.190)
-- 0.664 (0.106)
2006 --
Denmark 5.943 (0.247)
-- -- 0.000 (0.000)
2001 2005
Finland 4.833 (0.224)
4.731 (0.128)
-- 0.606 (0.120)
-- --
France 5.963 (0.098)
5.175 (0.240)
-- 0.550 ((0.088)
2004 --
Germany 5.814 (0.090)
5.096 (0.120)
-- 0.725 (0.126)
2005 --
Greece 6.051 (0.124)
5.556 (0.181)
-- 0.094 (0.102)
2007 --
Hungary 5.799 (0.206)
5.924 (0.285)
-- 0.685 (0.198)
2000 2006
Ireland 6.092 (0.183)
5.594 (0.387)
-- 0.392 (0.362)
-- --
Italy 6.143 (0.071)
5.182 (0.079)
-- 0.689 (0.230)
2002 --
Luxembourg 5.639 (0.126)
-- -- 1.701 (0.569)
-- --
Netherlands 5.800 (0.161)
4.968 (0.181)
-- 0.556 (0.421)
2004 2005
Poland 6.068 (0.503)
5.844 (0.133)
-- 0.176 (0.031)
-- 2005
Portugal 6.804 (0.024)
5.840 (0.095)
-- 0.135 (0.146)
-- 2006
Slovak R. 5.758 (0.462)
5.919 (0.181)
-- 0.955 (0.117)
2006 --
Spain 6.431 (0.093)
5.372 (0.185)
-- 0.341 (0.217)
2001 2006
UK 5.851 (0.146)
5.034 (0.291)
-- 0.108 (0.111)
1996 1997
All 5.891 (0.419)
5.419 (0.473)
3.248 (0.353)
0.569 (0.469)
-- --
Notes: For the first four variables mean values as well as standard deviations (in brackets) are presented. a End-user prices for households (hh) and industry (ind) in USD2000/10
7 kcal (base year 2000). Source: IEA.
b WTI oil price in USD2000/barrel (base year 2000). Source: IEA.
c Gas imports (ktoe) per capita. Source: IEA.
d Date of implementation (legal enactment). Source: Benchmarking Reports of the European Commission as well
as interviews of the designated regulatory authorities.
Ownership unbundling 30
Table 5: Correlation matrix of regulatory indicators
Indicator Entry
(ETCR) TPA Liberal Barriers
Owner. (ETCR)
PO_ Prod
PO_ TSO
PO_ DSO
Owner- ship
VI (ETCR)
VI_ Prod
VI_ Supp
VI_ DSO
VI_ TSO
OU_ TSO
Struc.
(ETCR)
MS_ Prod
MS_ Supp
MS_ TSO
Entry (ETCR) a 1
TPA b .826 1
Liberal b .849 .817 1
Barriers b .738 .350 .339 1
Ownership (ETCR) a .419 .202 .282 .403 1
PO_Prod b .428 .223 .264 .448 .894 1
PO_TSO b .380 .175 .211 .425 .901 .940 1
PO_DSO b .245 .094 .081 .289 .767 .768 .780 1
Ownership .368 .163 .198 .420 .907 .969 .975 .812 1
VI (ETCR) a .550 .582 .575 .255 .215 .214 .118 .122 .149 1
VI_Prod b .469 .429 .421 .312 .147 .157 .060 .056 .085 .832 1
VI_Supp b .723 .676 .768 .348 .331 .293 .233 .148 .238 .710 .446 1
VI_DSO b .082 .248 .179 - .111 .041 .050 .011 .089 .042 .607 .146 .305 1
VI_TSO .697 .630 .684 .389 .351 .332 .251 .225 .271 .734 .598 .711 .330 1
OU_TSO .419 .306 .418 .226 .272 .263 .132 .231 .184 .419 .393 .441 .095 .544 1
Structure (ETCR) a .463 .341 .520 .300 .520 .477 .442 .169 .431 .371 .303 .435 .100 .397 .244 1
MS_Prod b .519 .398 .551 .271 .526 .506 .415 .295 .418 .431 .398 .472 .075 .516 .378 .792 1
MS_Supp b .284 .235 .343 .196 .394 .337 .327 .099 .321 .325 .219 .291 .218 .265 .179 .855 .473 1
MS_TSO b .297 .157 .340 .264 .296 .278 .310 -.029 .279 .070 .065 .263 -.128 .125 -.032 .722 .376 .480 1
Note: a Aggregate indicator of the OECD database ETCR;
b Sub-indicator of the OECD database ETCR.
Ownership unbundling 31
Table 6: Correlation matrix and auxiliary fixed effects estimation for control variables
A. Correlation matrix
Variable Log Price_ind Log Oil Supply_pc Prod_pc Export_pc Import_pc GDP_pc
Log Price_ind a 1.000
Log Oil b .325 1.000
Supply_pc c -.267 -.036 1.000
Prod_pc c -.302 -.085 .848 1.000
Export_pc c -.250 -.065 .784 .939 1.000
Import_pc c .102 .103 .229 -.232 -.051 1.000
GDP_pc d -.670 .082 .336 .275 .247 .091 1.000
B. Fixed effects estimation e
Dep. Var.
Expl. var. Log Price_hh Log Oil Prod_pc Import_pc GDP_pc Year dummy R squared (within) No. of obs.
Log Price_ind -0.002
(0.059)
0.555***
(0.092)
-0.227**
(0.102)
0.074
(0.269)
-0.616***
(0.166) yes 0.559 233
Notes: Standard errors in parentheses are robust to heteroskedasticity and to within group serial correlation. ***, **, * denote significance at 1%, 5% and 10%. a End-user prices for industry in USD2000/10
7 kcal (base year 2000). Source: IEA.
b WTI oil price in USD2000/barrel (base year 2000). Source: IEA.
c Total gas supply (Supply_pc), gas production (Prod_pc), gas exports (Export_pc) and gas imports (Import_pc) all in ktoe per capita. Source: IEA.
d Gross domestic product in 10
4 USD2000 (base year 2000) per capita. Source: IEA.
e Based on a fixed effects estimator with robust standard errors (Froot 1989 and Williams 2000). Standard errors in parentheses are robust to heteroskedasticity and to within group
serial correlation. ***, **, * denote significance at 1%, 5% and 10%. The counterintuitive result of the negative impact of GDP per capita is caused by a cross-sectional (eastern European countries with a relatively small GDP per capita tend to have higher end-user prices) and a time series effect (decreasing end-user prices over time, e.g. in Ireland).
Ownership unbundling 32
References
Alesina, A., S. Ardagna, G. Nicoletti and F. Schiantarelli (2005), Regulation and Investment, Journal of
the European Economic Association 3(4), 791-825.
Anderson, T.W. and C. Hsiao (1982), Formulation and Estimation of Dynamic Models Using Panel
Data, Journal of Econometrics 18, 47-82.
Arellano, M. and S. Bond (1991), Some Tests of Specification for Panel Data: Monte Carlo Evidence
and an Application to Employment Equations, Review of Economic Studies 58, 277-297.
Ascari, S. (2011), An American Model for the EU Gas Market?, EUI Working Papers No. RSCAS
2011/39, Robert Schuman Centre for Advanced Studies, Florence School of Regulation,
Florence.
Blundell, R. and S. Bond (1998), Initial Conditions and Moment Restrictions in Dynamic Panel Data
Models, Journal of Econometrics 87, 115–143.
Brau, R., R. Doronzo, C.V. Fiorio and M. Florio (2010), EU Gas Industry Reforms and Consumers’
Prices, The Energy Journal 31(4), 163-178.
Brown, S.P.A. and M.K. Yücel (2007), What Drives Natural Gas Prices?, Working Paper 0703, Federal
Reserve Bank of Dallas.
Bruno, G.S.F. (2005), Approximating the Bias of the LSDV Estimator for Dynamic Unbalanced Panel
Data Models, Economics Letters 87(3), 361-366.
Buehler, S. (2005), The Promise and Pitfalls of Restructuring Network Industries, German Economic
Review 6(2), 205-228.
Buehler, S, A. Schmutzler and M.-A. Benz (2004), Infrastructure Quality in Deregulated Industries: Is
there an Underinvestment Problem?, International Journal of Industrial Organization 22(2),
253-267.
Buehler, S., D. Gärtner and D. Halbheer (2006), Deregulating Network Industries: Dealing with Price-
Quality Tradeoffs, Journal of Regulatory Economics 30(1), 99-115.
Bun, M.J.G. and J.F Kiviet (2003), On the Diminishing Returns of Higher Order Terms in Asymptotic
Expansions of Bias, Economics Letters 79, 145-152.
Cambini, C. and L. Rondi (2010), Incentive Regulation and Investment: Evidence from European
energy utilities, Journal of Regulatory Economics 38(1), 1-26.
Conway, P. and G. Nicoletti (2006), Product Market Regulation in the Non-Manufacturing Sectors of
OECD Countries: Measurement and Highlights, OECD Economics Department Working
Paper, No. 530, Paris.
Copenhagen Economics (2005), Market Opening in Network Industries: Part II: Sectoral Analyses,
Copenhagen Economics for DG Internal Market.
European Commission (2007), Inquiry pursuant to Article 17 of Regulation (EC) No 1/2003 into the
European gas and electricity sectors (Final Report), Communication from the Commission,
COM(2006) 851 final, January 10, Brussels.
European Commission (2008), Progress in creating the internal gas and electricity market, Report
from the Commission to the Council and the European Parliament, COM(2008) 192 final,
April 15, Brussels.
Ownership unbundling 33
European Commission (2010), Interpretive Note on Directive 2009/72/EC Concerning Common Rules
for the Internal Market in Electricity and Directive 2009/73/EC Concerning Common Rules for
the Internal Market in Natural Gas – The Unbundling Regime, Commission Staff Working
Paper, January 22, Brussels.
FERC [Federal Energy Regulatory Commission] (2011), Bidding by Affiliates in Open Seasons for
Pipeline Capacity, Notice of Proposed Rulemaking, Docket No. RM11-15-000, April 7,
Washington, D.C.
Froot, K.A. (1989), Consistent Covariance Matrix Estimation with Cross-Sectional Dependence and
Heteroscedasticity in Financial Data, Journal of Financial and Quantitative Analysis 24, 333-
355.
Glachant, J.-M. (2011), A Vision for the EU Target Model: The MECO-S Model, EUI Working Papers
No. RSCAS 2011/38, Robert Schuman Centre for Advanced Studies, Florence School of
Regulation, Florence.
Gordon, D.V., K. Gunsch and C.V. Pawluk (2003), A Natural monopoly in Natural Gas Transmission,
Energy Economics 25(5), 473-485.
Growitsch, C. and M. Stronzik (2011), Ownership Unbundling of Gas Transmission Networks –
Empirical Evidence, EWI Working Paper No. 11/07, Institute of Energy Economics at the
University of Cologne (EWI), Cologne.
Hattori, T. and M. Tsutsui (2004), Economic Impact of Regulatory Reforms in the Electricity Supply
Industry: A Panel Data Analysis for OECD Countries, Energy Policy 32(6), 823-832.
Höffler, F. and S. Kranz (2011), Legal Unbundling can be a ‘Golden Mean‘ between Vertical
Integration and Vertical Separation, International Journal of Industrial Organization 29(5), 576-
588.
Jensen, J. (2007), Gas Pricing: North America, Energy Charter Secretariat (Ed.), Putting a Price on
Energy: International Pricing Mechanisms for Oil and Gas, Brussels, 99-125.
Joskow, P. (2008), Lessons learned from Electricity Market Liberalization, The Energy Journal 29(2),
9-42.
Kiviet, J.F. (1995), On Bias, Inconsistency and Efficiency of Various Estimators in Dynamic Panel Data
Models, Journal of Econometrics 68, 53–78.
Kiviet, J.F. (1999), Expectation of Expansions for Estimators in a Dynamic Panel Data Model; Some
Results for Weakly Exogenous Regressors, in: Hsiao, C., K. Lahiri, L.-F. Lee und M.H.
Pesaran (eds.), Analysis of Panel Data and Limited Dependent Variables, Cambridge
University Press, Cambridge, 199-225.
Laffont, J.J. and J. Tirole (1993), A Theory of Incentives in Procurement and Regulation, MIT Press,
Cambridge, MA.
Makholm, J.D. (2007), Seeking Competition and Supply Security in Natural Gas: The US Experience
and European Challenge, Paper presented at the 1st CESSA Conference in Berlin on May 31,
National Economic Research Associates Inc. (NERA), Boston, MA.
Nagayama, H. (2009), Electric Power Sector Reform Liberalization Models and Electric Power Prices
in Developing Countries: An Empirical Analysis using International Panel Data, Energy
Economics 31(3), 463-472.
Nickell, S.J. (1981), Biases in Dynamic Models with Fixed Effects, Econometrica 49, 1417-1426.
Ownership unbundling 34
Nillesen, P.H.L. and M.G. Pollitt (2011), Ownership Unbundling in Electricity Distribution: Empirical
Evidence from New Zealand, Review of Industrial Organization 38(1), 61-93.
Roodman, D. (2006), How to Do xtabond2: An Introduction to “Difference” and “System” GMM in
Stata, Working Paper No. 103, Center for Global Development, Washington D.C.
Roodman, D. (2008), A Note on the Theme of Too Many Instruments, Working Paper No. 125, Center
for Global Development, Washington D.C.
Sen, A. and T. Jamasb (2010), The Economic Effects of Electricity Deregulation: An Empirical
Analysis of Indian States, EPRG Working Paper 1001, University of Cambridge, January.
Siliverstovs, B., G. L’Hégaret, A. Neumann and C. Von Hirschhausen (2005), International Market
Integration for Natural Gas? A Cointegration Analysis of Prices in Europe, North America and
Japan, Energy Economics 27(4), 605-615.
Steiner, F. (2001), Regulation, Industry Structure and Performance in the Electricity Supply Industry,
OECD Economic Studies 32, 143-182.
Vickers, J. (1995), Competition and Regulation in Vertically Related Markets, The Review of Economic
Studies 62(1), 1-17.
Williams, R.L. (2000), A Note on Robust Variance Estimation for Cluster-Correlated Data, Biometrics
56, 645-646.
Zhang, Y.-F., D. Parker and C. Kirkpatrick (2008), Electricity Sector Reform in Developing Countries:
An Econometric Assessment of the Effects of Privatization, Competition and Regulation,
Journal of Regulatory Economics 33(2), 159-178.
Brau, R., R. Doronzo, C.V. Fiorio and M. Florio (2010), EU Gas Industry Reforms and Consumers’
Prices, The Energy Journal 31(4), 163-178.
Copenhagen Economics (2005), Market Opening in Network Industries: Part II: Sectoral Analyses,
Copenhagen Economics for DG Internal Market.
Hattori, T. and M. Tsutsui (2004), Economic Impact of Regulatory Reforms in the Electricity Supply
Industry: A Panel Data Analysis for OECD Countries, Energy Policy 32(6), 823-832.
Steiner, F. (2001), Regulation, Industry Structure and Performance in the Electricity Supply Industry,
OECD Economic Studies 32, 143-182.
Spatial no arbitrage condition 35
3 Spatial no arbitrage condition
Price convergence and information efficiency in German natural gas
markets40
3.1 Introduction
The creation of an integrated competitive natural gas market throughout Europe is one of the
European Union’s priority objectives. The introduction of the European gas Directive
(98/30/EC) and the EU ‘Acceleration Directive’ (2003/55/EC) have brought fundamental
changes in the natural gas sector across many European countries. As such, the natural gas
industries have been transformed from vertically integrated monopolies to more competitive
structures. While some countries, notably including the UK and the Netherlands, have been
relatively progressive in the liberalisation process, others such as Germany have moved
more cautiously. Germany did not effectively open its natural gas market until the EU
Directive 2003/55/EC had been transposed into national law.
The German Energy Law (Energiewirtschaftsgesetz), introduced in July 2005, aimed at
facilitating market entry in the energy sector and at establishing competition in wholesale and
retail energy markets. Subsequent regulatory rulings (mainly the
Gasnetzzugangsverordnung) established regulated third party access based on an entry-exit
system. As of October 2007, this new institutional design has become mandatory for all
transmission network operators (TSOs) in the natural gas sector monitored by the German
energy regulator (Bundesnetzagentur).41 The implementation of the entry-exit system has
led to the establishment of regional natural gas wholesale markets in the form of different
virtual trading points (hubs) in so-called balancing zones or market areas for trading natural
gas in Germany. At the two major German hubs, NetConnect Germany (NCG) roughly
covering the South of Germany and GASPOOL (GPL) covering Northern Germany, a liquid
market for natural gas seems to have evolved. Whether these two sub-national markets are
yet competitive remains to be empirically tested. However, analysing the competitiveness
only of the two major German hubs might be misleading, inasmuch as it neglects the
possible importance of other connected European gas trading locations. The Netherlands is
the largest gas market adjacent to Germany. The Dutch Title Transfer Facility (TTF) hub can
be considered to be one of the most liquid wholesale gas trading hubs in continental Europe
besides Zeebrugge. Given the proximity and the liquidity of the TTF, the Dutch gas market
can serve as a (competitive) benchmark for the German natural gas spot markets.
In theory, competitive connected markets should show equal prices for a certain good (law of
one price). Such markets share a common long run equilibrium price and can be said to be
40 This chapter is based on Growitsch et al. (2010). The explanations regarding the regulatory framework
conditions refer to mid-2010. More recent developments are discussed in Chapter 5. 41 In the first year after the Gasnetzzugangsverordnung had been enacted network operators had the choice
between the introduction of the new entry-exit system and the retention of the old system based on path dependent network fees.
Spatial no arbitrage condition 36
economically integrated. Persistent price differences beyond transaction costs should not
exist. Likewise, the market’s response to external shocks and corresponding reversion to the
competitive equilibrium should be fast, i.e. information should be processed efficiently (Fama
1970).
In order to evaluate the development of the German natural gas wholesale market’s
competitiveness, we study the price convergence between different market areas and the
development of market efficiency. To analyse whether the prices at the different hubs have a
common long run equilibrium, we use cointegration analysis (Johansen 1988, 1991). One
implicit assumption of cointegration analysis is that the structural relation among the prices is
fixed over the considered time period; however, due to the on-going changes of the
regulatory framework in Europe and Germany and different mergers of Germany’s gas
TSOs, market integration could also be a gradual and on-going process. In a second step,
we therefore examine the convergence path of the natural gas spot market prices and the
degree of market integration estimating a state space model using a Kalman filter (Kalman
1960). In contrast to cointegration analysis, this estimator allows time-varying coefficients
and thus explicitly accounts for possible dynamic structural changes. Furthermore,
introducing a constant into the estimation equation allows us to indirectly test for capacity
constraints. Finally, the state space model is extended to an error correction model to
analyse how efficiently markets respond to new information. The time-varying nature of this
approach allows us to draw conclusions as to how market efficiency evolved over time, and
whether changes of the regulatory framework have led to improved market competitiveness.
The paper is structured as follows. Section 3.2 describes the institutional design of the
German natural gas market. Section 3.3 discusses the previous literature on international
natural gas wholesale market development. The econometric methodology is laid out in
Section 3.4. Data description follows in Section 3.5. Section 3.6 includes the estimation
results and their interpretation. Finally, Section 3.7 concludes with potential policy
recommendations.
3.2 Institutional design and recent developments in Germany
In 2005, a new Energy Law (Energiewirtschaftsgesetz) was enacted in Germany to
transpose the European Directive 2003/55/EC into national law. Its major purpose is to
develop and establish competition in the German energy sector. Concerning third party
access to the natural gas networks, a system charging regime based on simple entry and
exit charges (entry-exit system) was imposed. In October 2006, the transmission network
operators and the regulator concluded an agreement about the institutional design of the new
regime. Since October 1st 2007, the entry-exit system has become mandatory for all TSOs.
The agreement initially divided Germany into 19 entry-exit zones (also called market areas or
transmission system zones). Meanwhile, due to several poolings, the number of zones has
decreased to six, three for L-gas and three for H-gas (as of October 2009).42 NetConnect
42 H-gas is high caloric natural gas primarily delivered from Norway and Russia to Germany. L-gas is a low
caloric natural gas and has a lesser energy content than H-gas. As low caloric gas plays only a minor role in
Spatial no arbitrage condition 37
Germany (NCG) initiated a major pooling which became operational on October 1st 2008 and
combined the former areas of E.ON and Bayernets.43 While NCG covers the South of
Germany, GASPOOL as the second major market zone is located in the northern part of
Germany.44 The core of this area had already been established in 2006 as a result of a
cooperative arrangement between BEB45, StatoilHydro and DONG Energy. In July 2008,
Gasunie, operating the Dutch transmission network, has taken over the transportation
services of BEB. The third market zone is run by RWE.46 In April 2009, the German
government announced its aim of reducing the number of entry-exit zones to one for H-gas
and L-gas respectively, in order to increase wholesale market competition and liquidity.
The entry-exit system requires that the natural gas shippers book capacity at the relevant
entry and exit points separately; hence, the fees to be paid for the transportation of natural
gas (so called entry and exit charges) should no longer be based upon the distance between
the entry and exit points (also known as the contractual path) as previously practiced in
Germany. The abolition of such a ‘path-based’ charging system was meant to promote price
transparency because shippers need not obtain individual quotations for each separate
customer, thereby reducing pricing complexity. Also, the trading possibilities at multiple hubs
as a result of an entry-exit system should deliver a competitive price signal to the German
natural gas market as a whole. The entry-exit system should facilitate both domestic as well
as cross-border transports for third parties thus encouraging market entry and eventually
competition. Furthermore, the market redesign aims at increasing the flexibility and comfort
in booking procedures inasmuch as no capacity reservation is required for the individual
pipeline sections used to fulfill transport contracts. In sum, consumers and distributors were
intended to benefit from increasing gas-to-gas competition as a result of the implementation
of entry-exit after gas market liberalisation.
The reform’s economic success faces certain risks, however. A high number of market areas
complicates market operations. Also, dominant players may continue to operate in the zones
of their network operators’ affiliate,47 while new market entrants may be deterred. The
transmission of natural gas via the network may become expensive due to pan-caking and
also may be impossible due to congestion and grandfathered capacity rights held by the
incumbents. Such market barriers could effectively rule out the aim of achieving competition
and liquidity in the natural gas sector even with the introduction of the entry-exit regime.
Germany, our focus is solely on H-gas. Roughly speaking, each of the three market zones for H-gas, NCG, GASPOOL and RWE, also operates an L-gas network.
43 In October 2009, GRTgaz Deutschland, ENI and GVS joined NCG. 44 GASPOOL has been established in October 2009 with ONTRAS and Wingas joining the cooperation of
Gasunie, StatoilHydro and Dong Energy. Due to several renamings of this area over the considered period and to avoid confusion, we use the current name GASPOOL of this market zone throughout the paper.
45 BEB, owned by Shell and ExxonMobil, was one of the frontrunners as they introduced an entry-exit system
already in 2004. 46 In preparation of selling their transmission network, RWE handed over the assets to Thyssengas, a 100%
subsidiary of RWE. 47 In Germany, the network operators are legally unbundled but not in terms of ownership.
Spatial no arbitrage condition 38
3.3 Literature review
Different international studies have been carried out on natural gas markets integration and
price convergence as an aftermath of market liberalisation. However, the methodology to
account for market integration and price convergence differs across studies. Using
cointegration analysis, Walls (1994), De Vany and Walls (1993, 1996) as well as Serletis
(1997) found that the opening of network access in the aftermath of FERC Order 436 in 1985
led to greater market integration as prices across different locations converged in the North
American natural gas markets. Likewise, King and Cuc (1996) examined the degree of
pairwise price convergence using a time-varying coefficient approach in the North American
natural gas spot markets confirming the results of the cointegration analyses. Serletis and
Rangel-Ruiz (2004) showed that the main driver for North American natural gas prices is the
price trend at the Henry Hub. Applying a vector error correction model (VECM), Cuddington
and Wang (2006) found different degrees of market integration across regions. While the
East and Central regions are highly integrated, the Western market is only loosely connected
to common price trends.
In the European context, Asche et al. (2002) applied cointegration techniques to test for the
law of one price across the French, German and Belgian market using monthly natural gas
import prices. Their results show an integrated gas market where prices across the regions
considered follow a similar pattern over time. Using a Kalman filter, Neumann et al. (2006)
study the price relation between the UK (National Balancing Point) and the Belgian spot
market (Zeebrugge). They conclude that prices between these two markets are fully
converged.
Further empirical studies have investigated price relations for natural gas either between
different continents or between gas and other commodities in a certain region. Among the
first group are e.g. Ripple (2001), Siliverstovs et al. (2005) and Neumann (2009). All studies
find evidence for an increased price convergence across continents. The relationship of gas
prices to other commodities has been analysed quite extensively, e.g. by Asche et al. (2006)
and Panagiotidis and Rutledge (2007) for the UK, and by Hartley et al. (2008) and Brown and
Yücel (2009) for the US.
The current study differs from the existing literature on natural gas price convergence and
market integration in three ways. First, we focus our analysis on the competitive effect of a
new network access regime on a sub-national gas market level by looking at the dynamic
price interactions between two major German gas spot markets. Moreover, these regional
market developments are put into a European context by relating them to the Dutch spot
market. Second, we apply both cointegration analysis and a time-varying coefficient
approach. Extending the time-varying coefficient approach to an error correction model
enables us to draw conclusions not only about price convergence, but also about how
efficiently new information is absorbed by the market and how this information efficiency has
evolved over time. Third, we explicitly control for transportation costs, which have been
neglected to a certain extent in the previous literature.
Spatial no arbitrage condition 39
3.4 Econometric methodology
In this section, we (1) deduce empirical criteria for measuring German natural gas market
competitiveness and its development, and (2) present the corresponding econometric
methodology. To be able to test the law of one price, we adjust the regional natural gas spot
prices at GASPOOL (GPL), NetConnect Germany (NCG) and the Dutch Title Transfer
Facility (TTF) hub for transmission charges. We employ cointegration analysis in order to test
whether prices tend towards a common long-run equilibrium price. We use these results as a
first indication as to whether markets are integrated or not. In a second step, we estimate a
time-varying coefficient model using the Kalman filter to study price convergence over time.
Here, we develop a time-varying error correction model to identify the development of
information efficiency.
3.4.1 The law of one price and transmission charges
The starting point for our analysis is the spatial arbitrage condition for efficient markets,
which says that prices of identical products traded at regionally distinct locations should differ
only in transaction costs (law of one price). In natural gas markets, the most noteworthy
transaction costs are the network fees that shippers have to pay for gas transmission.
Accounting for these transmission costs, freedom from arbitrage for gas traded at different
locations is assured if the price in the exporting region ( tiP , ) plus the transmission costs
)( ,tjiTC equals the price in the importing region ( tjP , ). The spatial arbitrage condition can
be generalised as follows:
(3-1) tjititj TCPP ,,, ,
where equality holds only if trade between the two regions occurs. If the price differential is
strictly less than the cost associated with gas transport, market participants have no incentive
to trade. In an entry-exit regime, these costs in one direction consist of the exit fee for the
exporting region and the entry fee for the importing region. As regards continental Europe,
transmission charges are direction-specific, i.e. in general ijji TCTC . Due to this
asymmetric characteristic, the arbitrage condition in (3-1) can be reformulated as:48
(3-1a) tijtijtjtjitjiti TCdPTCdP ,,,,,, ,
1, tjid if tjtjiti PTCP ,,, and 0, tjid otherwise;
1, tijd if titijtj PTCP ,,, and 0, tijd otherwise;
i,j = GPL, NCG, TTF.
48 With this approach we modified a model suggested by Zachmann (2008) who analysed convergence of
European electricity spot prices controlling for the outcome of capacity auctions.
Spatial no arbitrage condition 40
Physical gas flows are only possible in one direction at a time. As no information is publicly
available about the actual flows between the considered regions, we include a set of dummy
variables d mapping potential flows. A dummy variable di is one if the price P in region i plus
the costs of shipping the natural gas to region j (TCi→j) are less than or equal to P in region j
and vice versa. These dummies enable us to determine the relevant transmission costs as a
function of the actual price differential. It should be noted that both dummies can be zero at
the same time, but both dummies can never be unity at the same time. The former situation
prevails if spot price differentials are too low to exceed the transmission charges, in which
case trading gas between the two regions would be unprofitable.
With an adjusted spot price (ex transmission charges) tjitjiti
net
ti TCdPP ,,,, and using
log prices ( )log( ,,
net
ti
net
ti Pp ), Eq. (3-1a) becomes
(3-1b) net
tj
net
ti pp ,, ,
i,j = GPL, NCG, TTF.
If this condition is violated, markets are neither fully competitive nor integrated. However, real
world complexities in trading imply that several factors may exist leading to deviations from
this condition. The hypothesis to be tested is that the establishment of an entry-exit system
has removed some of the causes of inefficiency, and that prices converged as a result
thereafter.
3.4.2 Cointegration tests
A precondition for cointegration analysis is that the time series considered have a unit root.
Therefore, we test for unit roots using the Augmented Dickey Fuller (ADF, Dickey and Fuller
1979) and the Kwiatkowski Phillips Schmidt and Shin (KPSS, Kwiatkowski et al. 1992) tests.
While ADF is based upon the null hypothesis of a unit root, the KPSS is based upon the null
hypothesis of stationarity. Given that unit root tests sometimes face the problem of poor
power properties, double testing mitigates the risk of a false conclusion.
Applying the Johansen test (Johansen 1988, 1991) to natural gas spot price series of two
regions, one expects exactly one cointegrating relationship if these regions have a long run
equilibrium. The corresponding two dimensional Vector Error Correction (VEC) model is:
(3-2)
1
1
1
l
k
t
net
ktk
net
t
net
t ppp ;
where Δ is the first difference operator, net
tp is the vector of the two spot prices,
),0.(...~ diint , is a ( 22 ) matrix of the form β , with comprising the
cointegrating vector and representing the corresponding loadings. While β coefficients
show the long-run equilibrium relationship between price levels, coefficients measure the
adjustment speed towards equilibrium. The closer β is to one, the better economically
Spatial no arbitrage condition 41
integrated the markets are. A high absolute value for in turn indicates a high speed of
price adjustment and a more efficient market.
According to Eq. (3-1b), we expect 1,1 . In order to control for transaction costs other
than transmission fees, we allow for a constant in the cointegrating relationship; however, if
the law of one price holds, this constant should be (close to) zero.
The Johansen approach assumes a constant cointegrating vector over time. As pointed out
by several studies, e.g. King and Cuc (1996) and Kleit (2001), the cointegration relationship
does not shed any light on the dynamics of possible price convergence or divergence.
Against the background of the dynamic regulatory framework in Europe as pointed out
earlier, the assumption of a fixed relationship between spot prices over time might be
problematic. Following the line of argument of Barrett (1996), Baulch (1997a, 1997b) as well
as Barrett and Li (2002), we use results from cointegration only as a kind of pre-test as to
whether markets are integrated or not.
3.4.3 Time-varying coefficient model
An approach with time-varying coefficients can overcome the drawbacks of cointegration
analysis and can account for the dynamics of parallel price developments from regionally
distinct markets. The introduction of a time-varying coefficient into the linear relationship of
prices enables us to analyse the path of price convergence or divergence over time.
Recalling Eq. (3-1b) and introducing a constant cij, again reflecting costs associated with
trades between the two regions and not already covered by transmission charges, we can
formulate the following state space model:
(3-3)
ttt
t
net
tjtij
net
ti pcp
1
,,,
where ),0.(...~ 2
diiNt and ),0.(...~ 2
diiNt are white noise processes and t is the
vector of unobservable coefficients at time t.
t represents the strength of price convergence across regions. If 0t , it implies that
there is no relation between the natural gas spot prices, and that markets were completely
decoupled. If prices have converged and markets are perfectly integrated and competitive,
t should be equal to one. Furthermore, as prices (net
tip , ) are already adjusted for
transportation costs, we expect cij to be negligible. If cij were substantial, it would tend to
imply that significant other costs in addition to the transportation charges were present in the
market preventing shippers from trading and, therefore, prices from converging.
Economically, a constant gap between the price series considered might thus indicate
permanent capacity constraints. Such constraints could be of technical nature as a result, for
instance, of underinvestment. On the other hand, contractual capacity constraints might
Spatial no arbitrage condition 42
suggest hoarding of capacity rights by incumbents,49 and might signal third party
discrimination or even the abuse of market power.
The state space model is estimated using the Kalman filter (Kalman 1960).50 This technique
processes the data in two consecutive steps. It first estimates t by using available
information until period t-1. In a second step, the estimates of t are updated by
incorporating prediction errors from the first step (to time t-1) to compute values for time t.
Applying the Kalman filter provides information for cij and for t at each point in time, and
thus enables us to obtain detailed information on the common development of prices.
In employing the Kalman filter, it is important to determine the initial variances for t and t
as well as of the expected value of 0 .51
Calibrating net
j
net
i
p
pE
1,
1,
0 1)( , )(1.0 ,
2 net
tipVar and 000,1/22
provides suitable
noise reduction and signal preservation.
Finally, we use the framework of time-varying coefficients to formulate an error correction
model in the following way:
(3-4)
ttt
t
net
tj
net
titij
net
ti ppcp
1
1,1,, )(,
where t measures the time it takes to bring the system back towards equilibrium after new
information appears on the market. The larger the absolute value for t , the higher the
speed of price adjustment and the more efficiently information is converted into price signals.
Therefore, the time-varying framework enables us to draw conclusions not only on how
efficiently prices adjusted to new information, but also how efficiency evolved over time.
Since the entry-exit regime was introduced in order to ease gas transmission and foster gas-
to-gas competition, we expect the absolute value of t to increase over time.
The error correction model specification needs a new calibration of the starting values for the
Kalman filter:
net
j
net
i
net
i
pp
pE
1,1,
2,
0 0)(
, )(01.0 ,
2 net
tipVar and 22
.
49 Transmission capacity is often booked very long in advance and has to be nominated when shippers
actually want to use it. Hoarding of capacity rights means unused capacity which is not available to other market participants.
50 For further details see Harvey (1987). 51 Usually, the maximum likelihood function has several local maxima. Therefore, inadequately chosen starting
points can lead to undesirable results. Exaggerated values of 2
v would lead to the inclusion of short-term
behaviour making it difficult to distinguish random shocks from structural relationships. In contrast, setting the variance too low would ignore significant developments in the convergence process over time.
Spatial no arbitrage condition 43
Setting the expected value of 0 to zero assumes inefficient information processing at the
beginning of the observation period. This seems especially plausible, since market
participants seem to need time to adopt to a new regulatory regime.52
3.5 Data
The aim of this paper is to test for market integration and price convergence between the
major two entry-exit zones in Germany, namely GPL and NCG. Additionally, we analyse
price relations with respect to the Dutch trading hub TTF, which is well connected with
Germany and is a major European gas trading point. For the German trading hubs, we have
used the day-ahead spot market settlement price for natural gas as publicly obtained from
the European Energy Exchange (EEX),53 while data for TTF has been obtained from
Energate. Daily price data is preferred over weekly or monthly data because lower
frequencies can lead to temporal aggregation problems if used to study the price adjustment
process (Taylor 2001). The use of high frequency data should better capture the markets’
reactions to ongoing regulatory and market reforms, thereby facilitating the investigation of
market integration. The observation period lasts from the 1st of October, 2007, when the
mandatory introduction of the entry-exit system came into operation to the 30th of September,
2009.54 The prices have been transformed into logarithmic form because spot market prices
for natural gas tend to be highly volatile which may bias our results.
It can be seen from Figure 5 that the day-ahead prices for natural gas were volatile
throughout the time period considered for all market areas. The prices became more volatile
during the last quarter of 2007, while the volatility declined from the first quarter of 2008 to
the third quarter of 2008. From the first quarter of 2009, prices witnessed a steep decline.
The development is likely – at least to some extent – to be explained by the falling prices for
crude oil.55 Looking at the price differentials for the three pairs of market zones considered,
no persistent price divergence is obvious. The development of price differentials roughly
parallels the development of price volatility; the higher the volatility, the larger the magnitude
of the observed gaps. Prices at NCG are often slightly higher than at TTF and GPL.56 All in
all, the price series show a rather similar development pattern over time57 leading to the
52 Growitsch and Weber (2008) show exact this in a paper on the re-design of the German balancing power
market. 53 Although only 10% of the total volume is traded at EEX (90% OTC), we have chosen EEX data as it is the
only publicly available price series covering a longer time period without structural breaks. E.g., OTC prices for NCG are only available since October 2008.
54 Price data for both German market areas prior to October 2007 hardly exist and face the problem of low
reliability. The considered time period covers two so called “gas years”, starting with the heating season in October and ending in September. For each price series this results in 494 observations (trading days).
55 Gas prices in mainland Europe are often index-linked to that of crude oil with some time lag. While the price
for Brent peaked on 3rd
of July, 2008, GPL prices reached their maximum on 23rd
of September 2008. NCG and TTF prices had their peak both on the 17
th of September, 2008. This amounts to a lag of roughly two
and a half months. 56 For roughly 75% of the observations NCG prices lie above spot prices at GPL and TTF. Consequently, this
ratio for GPL and TTF is 50%. 57 This is confirmed by the descriptive statistics which are available from the authors upon request. E.g., the
standard deviation as a proxy for volatility is around 0.40 €/MWh (in log terms) for all three regions.
Spatial no arbitrage condition 44
expectation of highly integrated markets, and thus values close to one for the coefficient in
Eqs. (3-2) and (3-3).
Figure 5: Logarithmic day-ahead spot prices (€/MWh)
Notes: Right Y-axis: Log spot prices for GPL, NCG and TTF (upper lines) Left Y-axis: Differential of log prices for GPL vs. NCG (black line), GPL vs. TTF (red line) and NCG vs. TTF (blue line).
Information on transmission charges were obtained from the websites of the relevant TSOs
as well as in interviews of the shippers.58 For each possible relation between the three
considered markets, we chose a representative connecting point. In order to get consistent
data, we converted the corresponding capacity-based entry and exit charges expressed in
[€/(kWh/h)/a] into [€/MWh]. We assume that a transport of one MWh of natural gas
corresponds to holding a capacity of one MW. The TSO have changed their fees at different
points in time. The resulting development of transmission charges over time is summarised
in Table 7. While the cost of transporting natural gas within Germany has decreased quite
substantially (by about 10%), fees for cross-border transmission have remained more or less
the same.
58 Due to the ongoing reorganisation of the German TSOs and market zones, consistent historical data on
transmission charges is not publicly available.
0
0.5
1
1.5
2
2.5
3
3.5
4
-0.300
-0.200
-0.100
0.000
0.100
0.200
0.300
0.400
0.500
0.600
GPL vs. NCG GPL vs. TTF NCG vs. TTF GPL NCG TTF
Spatial no arbitrage condition 45
Table 7: Transmission charges
Direction of transport
Connecting point
Transmission charges [€/MWh]
01/10/2007 01/10/2008 01/12/2008 01/04/2009 01/07/2009
GPL NCG Bunder Tief
0.557 0.523 0.512 0.497 0.497
NCG GPL 0.619 0.582 0.565 0.534 0.534
GPL TTF Oude Statenzijl
0.420 0.420 0.420 0.410 0.400
TTF GPL 0.409 0.409 0.409 0.394 0.408
NCG TTF Bocholtz
0.387 0.399 0.386 0.386 0.398
TTF NCG 0.486 0.498 0.484 0.484 0.490
3.6 Results
To answer the question of whether the introduction of an entry-exit regime has led to more
competitive market conditions, first we present the results of the cointegration analysis, and
then the results of the time-varying coefficient approaches.
3.6.1 Cointegration analysis
To check whether the price series fulfil the precondition for cointegration analysis, we test for
unit roots. The results from both ADF and KPSS, displayed in Table 8, provide a clear
picture. All time series have a unit root and are I(1) as first differences are stationary.
Table 8: Unit root tests
Natural gas spot prices (log)
Region ADF KPSS
Level First difference Level First difference
GPL -1.012 -21.353*** 1.950*** 0.182
NCG -0.852 -23.738*** 1.926*** 0.231
TTF -0.949 -23.034*** 2.005*** 0.210
Notes: Tests include a constant but no time trend. For ADF, the lag length is selected according to Schwarz Information Criterion (SIC). For KPSS, bandwidth has been chosen according to Newey-West using the Bartlett Kernel. The provided numbers denote the t-ratios for ADF and the LM statistic for KPSS. *, **, *** indicate significance at the 10, 5 and 1 %-levels.
Next, we estimate Eq. (3-2) applying the approach developed by Johansen (1988, 1991).
Each pairwise price relation has exactly one cointegrating term indicating that long-run
equilibria do exist.59 Price series share a common stochastic trend and hence will not drift
59 Both unrestricted cointegration rank tests, Trace as well as Maximum Eigenvalue, lead to equivalent results.
Spatial no arbitrage condition 46
apart greatly in the long run. Table 9 shows the main results of the corresponding VEC
models.
Table 9: Long-run cointegrating equations (ML estimation)
Region Cointegrating equation LR test
Constant ]1,1[ 0
GPL-NCG -0.983*** (0.010)
GPL -0.155** (0.076) -0.038
(0.028) 3.071*
4.115**
NCG 0.158** (0.073)
4.581**
GPL-TTF -0.997*** (0.007)
GPL -0.332*** (0.064) -0.014
(0.021) 0.227
26.049***
TTF 0.188*** (0.064)
8.343***
NCG-TTF -1.011*** (0.010)
NCG -0.342*** (0.076) 0.0166
(0.030) 1.328
19.470***
TTF 0.025
(0.076) 0.106
Notes: Lag length to map short-run dynamics selected according the Schwarz Criterion using an unrestricted VAR. Numbers in brackets report standard errors. Numbers for the LR test denote the χ2-statistics. *, **, *** indicate significance at the 10, 5 and 1 %-levels.
All coefficients are very close to one. The likelihood ratio test in the penultimate column of
Table 9 tests for the restriction of the cointegrating vector being ]1,1[ , meaning that a
1% price change in region i is accompanied by the same price change in region j. Only in the
case of the two German areas this indication of very strong market integration has to be
rejected at a 10% level. The insignificant constant in the long-run cointegrating equation
signals that no other significant transaction costs in addition to the already captured
transmission charges and therefore no persistent price differential exist. Looking at the error
correction coefficient we see significant bi-directional price adjustments for GPL-NCG and
GPL-TTF. The relationship of natural gas spot prices between GPL and TTF is stronger
since the level of significance and adjustment speed are higher. While for GPL-NCG only
around 16% of an external shock is absorbed within one period (i.e. one trading day), GPL
prices adjust for 33% of an imbalance with TTF prices. The corresponding half-lives, defined
as the time in which a marginal change in the stationary component becomes half of the
initial jump and computed as )1ln(/)5.0ln( , are 4.1 days and 1.7 days respectively. The
stronger interrelation between GPL and TTF might be due to the fact that both networks are
now run by the same TSO, Gasunie. The asymmetric price adjustment in the latter case, with
GPL prices adjusting faster than TTF prices, indicates that the TTF as the larger and more
liquid market is the leading market for GPL. The same argument holds for the NCG-TTF
results, with only the NCG adjusting to deviations from equilibrium. A likelihood ratio test
restricting TTF to zero confirms that TTF is weakly exogenous for NCG. These results
support the hypothesis that the Dutch TTF can be considered as a kind of reference or
leading market for both German market areas.
Spatial no arbitrage condition 47
To sum up, results from cointegration analysis provide evidence of market integration across
the two German market areas considered in this study; however, one has to interpret these
results with care. The presence of cointegration does not necessarily imply the stability of the
estimated β parameter. Also, the long-run β coefficient may not stay constant over time, as
several structural changes have occurred in the natural gas markets within the period
considered.
3.6.2 Time-varying coefficient
The time-varying coefficient approach is particularly suitable for accounting for these
structural changes. Table 10 presents the main results of the analysis of market integration
through price convergence (Eq. (3-3)), as well as the outcomes of the error correction model
(Eq. (3-4)) which gives insights into the development of information efficiency.
Table 10: Results of the time-varying coefficient models
Region
Price Convergence
[Eq. (3-3)]
Information Efficiency
[Eq. (3-4)]
Constant Constant
GPL-NCG 0.971***
(0.003)
0.068**
(0.027)
-0.684**
(0.291)
-0.002
(0.003)
GPL-TTF 0.923***
(0.003)
0.155***
(0.028)
-0.878***
(0.244)
-0.000
(0.002)
NCG-TTF 0.916***
(0.004)
0.178***
(0.026)
-0.653**
(0.296)
0.006**
(0.003)
Notes: For the coefficients the final state is provided. Numbers in brackets report the root mean square error for the coefficients and standard errors for the constant. *, **, *** indicate significance at the 10, 5 and 1 %-levels.
In the final state, GPL and NCG show the highest degree of price convergence (0.971), but a
lower speed to adjust to changes. This means that highly integrated markets are not
necessarily the most efficient markets. Comparing the results with the previous cointegration
analysis, what is most striking is the significance of the constant in all three estimations of
Eq. (3-3). Allowing for a time-varying specification reveals a price differential that goes
beyond transmission charges. The two pairs with the Dutch TTF have an additional price gap
roughly 2.5 times higher than the national price relation between GPL and NCG. This gap is
most likely caused by capacity constraints, thus indicating an additional scarcity. Therefore,
the higher values for GPL-TTF and NCG-TTF point at tighter cross-border constraints. As
one of the major complaints concerning the natural gas market - raised not only by market
entrants but also by energy regulators - is the insufficient amount of transmission capacity
available to market participants, the result of the state space approach is much more
consistent with market observations than the one of cointegration analysis in the previous
Spatial no arbitrage condition 48
sub-section.60 Due to the assumption of constant price relations over the considered period,
the VECM failed to demonstrate this finding. In fact, the averaging characteristic of
cointegration analysis may lead to overestimated degrees of price convergence.
Figure 6 shows the development of the coefficients over time. All three pairs of market
zones started with a rather high degree of price convergence (above 0.94)61 that increased
moderately after the introduction of the entry-exit regime in October 2007. The coefficients
for GPL-NCG and GPL-TTF peaked around July 2008 which coincides with the takeover of
BEB by Gasunie. A plunge of the coefficient for both price relations can be observed right
after the establishment of NCG as a merger of E.ON and Bayernets (October 2008). While
prices between NCG and GPL started to converge again, the gap between GPL- and TTF-
prices increased even further. Concerning NCG and TTF, the price gap decreased more or
less steadily until the first quarter of 2009, but increased thereafter.
Figure 6: Price convergence [β]
While the coefficient of GPL-NCG swung back to the initial level of price convergence by
September 2009, the other two relations dropped significantly below their starting value. All
60 The issue of available transmission capacity is raised in nearly every monitoring report on the German
natural gas market. The paucity of secondary markets for unused capacity rights amplifies this problem. Meanwhile. the German Ministry of Economics and Technology has revised the order dealing with network access (Gasnetzzugangsverordnung) enforcing shippers to market unused capacity rights.
61 The high value at the beginning is not sensitive to the assumption of 1)( 0 E . Setting 0)( 0 E
reveals similar results.
0.91
0.93
0.95
0.97
0.99
Oct-0
7
Jan-
08
Apr-0
8
Jul-0
8
Oct-0
8
Jan-
09
Apr-0
9
Jul-0
9
GPL-NCG GPL-TTF NCG-TTF
Spatial no arbitrage condition 49
in all, none of the measures expected to foster competitiveness is associated with a
significant increase in price convergence.62
However, as pointed out above, market integration in terms of price convergence has to be
distinguished from market efficiency. Therefore, the question remains whether the observed
degree of price convergence is sufficient to allow efficient adjustment to new information. The
development of information efficiency is depicted in Figure 7. The coefficient of the error
correction model of Eq. (3-4) indicates how fast prices turn back to equilibrium once new
information has appeared. The higher the absolute value of , the faster prices adjust to
new information and the more efficient markets are. The price relation between GPL and
NCG shows the lowest efficiency over most of the time period considered. Very shortly after
the entry-exit system became mandatory, prices adjusted pairwise between NCG and TTF to
around 40% of new information within one day, which is twice as much as the pairwise price
adjustment between GPL and NCG (Figure 7).63 The faster information processing between
NCG and TTF at the beginning of the observation period might have been due to higher
trading volumes at NCG compared to GPL.64 After two years, all price relations exhibit half-
lives of less than one day.65 The increase in efficiency is accompanied by a growth in trading
volume. The churn rate of NCG, e.g., measuring the ratio between traded and physically
delivered volumes, has risen from 1.6 in September 2007 to 2.7 in September 2009, thus
getting closer to the churn rate of the Dutch TTF of around 4.66 For GPL-TTF and NCG-TTF,
a major increase in the speed of information processing can be observed during the fourth
quarter of 2008, thus shortly after the creation of NCG.
The zero value of between GPL and NCG around January 2009, which indicates that
prices did not adjust to new information in any way, can be explained by the gas conflict
between Russia and Ukraine. On the 1st of January, 2009, Russia stopped gas delivery to
the Ukrainian transmission network completely. Russia accused the Ukraine of consuming
gas illegally that had been designated for transit to Germany and other European countries.
Since the EU imports around 25% of its gas from Russia, with the main transit route via the
Ukraine, the suspension of delivery resulted in gas shortages in Central and Southeast
Europe, which necessitated a rearrangement of gas flows. While cross-border gas transport
only changed the direction, the rearrangement within Germany led temporarily to inefficient
pricing behaviour.
62 One might argue that the formations of the β coefficients are mainly explained by the oil price movement as
a common exogenous factor driving the gas prices. This argument seems to be implausible, since the decline of price convergence between NCG and TTF started later than in the other two cases. Furthermore, the oil price has increased already since beginning of 2009. Taking into account the time lag of roughly 2.5 months, β should have envisaged an upswing in all three price relations during the second quarter of 2009.
63 Due to the small but significant constant (see Table 10) NCG-TTF never completely returns to equilibrium. 64 95% of natural gas trading volume at EEX is related to the market area of NCG, only 5% to GASPOOL. 65 Comparing these values with the results of the previous sub-section shows again the averaging
characteristic of the VECM. 66 The number of 1.6 is related to the market zone of E.ON as a precursor of NCG. Compared to the British
National Balancing Point (NBP) as the most liquid hub in Europe, churn rates in Continental Europe are still significantly lower. NBP has a churn rate of around 15.
Spatial no arbitrage condition 50
Figure 7: Information efficiency [ ]
Although prices have diverged rather than converged since the mandatory introduction of the
entry-exit system, information efficiency has increased significantly over the last two years.
As Barrett and Li (2002) point out, market integration has to be distinguished from efficiency
issues in spatial price analysis. Regarding the price relations between the two major German
market zones for natural gas as well as the connection to the Dutch TTF, markets are
sufficiently integrated to provide for an improved processing of new information. Prices adjust
within roughly one trading day.
3.7 Conclusions
The aim of this paper was to study the development of market integration and efficiency in
Germany after the mandatory introduction of an entry-exit network pricing regime in the
natural gas market. We applied Johansen’s cointegration analysis and a time-varying
coefficient approach (Kalman filter) to test for price convergence. Moreover, the state space
model was extended to an error correction model to analyse how fast prices adjust to new
information, i.e. how efficient the markets are.
Using price series for the two major German market areas, NetConnect Germany (NCG) and
GASPOOL (GPL), as well as data from the Dutch Title Transfer Facility (TTF) hub as a
competitive benchmark, we explicitly accounted for transportation costs in order to test the
spatial no arbitrage condition. Results of the Johansen approach show a level of price
convergence close to one between all three locations with the Dutch TTF – as the more
mature market – leading the pricing behaviour at both German hubs. However, the Johansen
-1.00
-0.80
-0.60
-0.40
-0.20
0.00
0.20
Oct-07 Jan-08 Apr-08 Jul-08 Oct-08 Jan-09 Apr-09 Jul-09
GPL_NCG GPL_TTF NCG_TTF
Spatial no arbitrage condition 51
method seems to overestimate price convergence as it assumes fixed price relations over
the two years period of the study, and thus fails to account for changes in the market
environment. The time-varying coefficient model overcomes these drawbacks and reveals
lower levels of convergence, which are nonetheless sufficient to provide for an improved
processing of new information. Since the mandatory introduction of the entry-exit system,
information efficiency has increased significantly, especially after the merger of the two
market zones resulting in the foundation of NCG in October 2008. Prices adjust to new
information within roughly one trading day. Only during the gas conflict between Russia and
the Ukraine (resulting in gas shortages in Central and Southeast Europe at the beginning of
the year 2009) was the information efficiency at the two German hubs severely low. Cross-
border information processing was not affected by these supply disruptions.
However, we found a persistent price differential between the markets not explained by
transportation costs. These price differences indicate capacity constraints. They exist
between any of the markets, but are 2.5 times higher across the German border than
between the two German zones. Overall, the wholesale market has developed well over time
but still seems to lack competitiveness to a certain extent. One important reason could be
blocked or congested transportation capacity (contractual constraints through capacity
hoarding). Thus, establishing effective and transparent rules concerning entitlements to open
network access for third parties is and remains necessary in order to benefit from a fully
liberalised market in Germany.
Spatial no arbitrage condition 52
References
Asche, F., P. Osmundsen and R. Tveteras (2002), European Market Integration for Gas? Volume
Flexibility and Political Risk, Energy Economics 24(3), 249-265.
Asche, F., P. Osmundsen and M. Sandsmark (2006), The UK Market for Natural Gas, Oil and
Electricity: Are the Prices Decoupled? Energy Journal 27(2), 27-40.
Barrett, C.B. (1996), Market Analysis Methods: Are Our Enriched Toolkits Well Suited to Enlivened
Markets? American Journal of Agricultural Economics 78(3), 825-829.
Barrett, C.B., and J.R. Li (2002), Distinguishing Between Equilibrium and Integration in Spatial Price
Analysis, American Journal of Agricultural Economics 84(2), 292-307.
Baulch, B. (1997a), Transfer Costs, Spatial Arbitrage, and Testing for Food Market Integration,
American Journal of Agricultural Economics 79(2), 477-487.
Baulch, B. (1997b), Testing for Food Market Integration Revisited, Journal of Development Studies
33(4), 512-534.
Brown, S.P.A., and M.K. Yücel (2009), Market Arbitrage: European and North American Natural Gas
Prices, The Energy Journal, Special Issue, 167-185.
Cuddington, J.T., and Z. Wang (2006), Assessing the degree of spot market integration for U.S.
natural gas: evidence from daily price data, Journal of Regulatory Economics 29(2), 195-210.
De Vany, A., and W.D. Walls (1993), Pipeline Access and Market Integration in the Natural Gas
Industry: Evidence from Cointegration Tests, The Energy Journal 14(4), 1-19.
De Vany, A., and W.D. Walls (1996), The Law of one Price in a Network: Arbitrage and Price
Dynamics in Natural Gas City Gate Markets, Journal of Regional Science 36(4), 555-570.
Dickey, D., and W. Fuller (1979), Distribution of the Estimators for Autoregressive Time Series With a
Unit Root, Journal of the American Statistical Association 74(366), 427-431.
Fama, E.F. (1970), Efficient Capital Markets: A Review of Theory and Empirical Work, Journal of
Finance 25(2), 383-417.
Growitsch, C. and C. Weber (2008), On the electricity reserves market redesign in Germany, CNI-
Working Paper No. 2008-01, Berlin.
Growitsch, C., M. Stronzik and R. Nepal (2010), Price Convergence and Information Efficiency in
German Natural Gas Markets, Paper presented at the 25th Annual Congress of the European
Economic Association (EEA), August 23-26, Glasgow.
Hartley, P., K. Medlock and J. Rosthal (2008), The Relationship of Natural Gas to Oil Prices, The
Energy Journal 29(3), 47-65.
Harvey, A.C. (1987), Applications of the Kalman Filter, Truman F. Bewley, (ed.): Advances in
Econometrics, Fifth World Congress, Vol. 1, 285-313.
Johansen, S. (1988), Statistical Analysis of Cointegration Vectors, Journal of Economic Dynamics and
Control 12(2/3), 231-254.
Johansen, S. (1991), Estimation and Hypothesis Testing of Cointegration Vectors in Gaussian Vector
Autoregressive Models, Econometrica 59(6), 1551-1580.
Spatial no arbitrage condition 53
Kalman, R.E. (1960), A New Approach to Linear Filtering and Prediction Problems, Transactions of
the American Society of Mechanical Engineers, Journal of Basic Engineering, Series D, 82(1),
35-45.
King, M., and M. Cuc (1996), Price Convergence in North American Natural Gas Spot Markets, The
Energy Journal 17(2), 17-42.
Kleit, A.N. (2001), Are Regional Oil Markets Growing Closer Together? An Arbitrage Cost Approach,
The Energy Journal 22(1), 1-15.
Kwiatkowski, D., P.C.B. Phillips, P. Schmidt and Y. Shin (1992), Testing the Null Hypothesis of
Stationary against the Alternative of a Unit Root, Journal of Econometrics 54(1-3), 159–178.
Neumann, A. (2009), Linking Natural Gas Markets – Is LNG Doing its Job? The Energy Journal,
Special Issue, 187-199.
Neumann, A., B. Silverstovs and C. von Hirschhausen (2006), Convergence of European Spot Market
Prices for Natural Gas? A Real-Time Analysis of Market Integration Using the Kalman Filter,
Applied Economics Letters 13(11), 727-732.
Panagiotidis, T., and E. Rutledge (2007), Oil and gas markets in the UK: Evidence from a
cointegrating approach, Energy Economics 29(2), 329-347.
Ripple, R. D. (2001), US West Coast Petroleum Industry in the 1990s: Isolated or Globally Integrated,
Oil, Gas and Energy Quarterly 50, 105-139.
Serletis, A. (1997), Is there an East-West Split in North American Natural Gas Markets? Energy
Journal 18(1), 401-414.
Serletis, A., R. Rangel-Ruiz (2004), Testing for common features in North American energy markets,
Energy Economics 26(3), 401-414.
Siliverstovs, B., G.L. L´Hegaret, A. Neumann and C. von Hirschhausen (2005), International Market
Integration for Natural Gas? A Cointegration Analysis of Prices in Europe, North America, and
Japan, Energy Economics 27(4), 603-615.
Taylor, A.M. (2001), Potential Pitfall for the for the Purchasing Power Parity Puzzle? Sampling and
Specification Biases in Mean-Reversion Tests of the Law of One Price, Econometrica 69(2),
473-498.
Walls, W.D. (1994), Price Convergence across Natural Gas Fields and City Markets, The Energy
Journal 15(4), 37-48.
Zachmann, G. (2008), Electricity wholesale market prices in Europe: Convergence? Energy
Economics 30(4), 1659-1671.
Intertemporal no arbitrage condition 54
4 Intertemporal no arbitrage condition
Does the European natural gas market pass the competitive benchmark
of the theory of storage? Indirect tests for three major trading points67
4.1 Introduction
The discussion among European regulators and the public debate over securing energy
supplies underscore the fundamental importance of using natural gas storage to counteract
supply and demand disruptions, balance the system and provide additional flexibility. The
European Union’s third legislative package clearly states the need for a competitive, efficient
natural gas market, but where the past decade focussed on the development of a level
playing field for operators mainly in (long-distance) transportation and LNG imports, other
parts of the value chain, such as the market for storage, remained untouched. The EU’s
legislative package was triggered by a European-wide sector inquiry finding that available
storage capacity not booked or otherwise reserved, and therefore accessible to market
participants, is rather scarce and sometimes non-existent (European Commission 2007).
Evidence from the US, where a futures market for inventories exists at NYMEX, indicates
that the time has passed when gas storage was considered solely as backup inventory or as
a seasonal supply source (Hirschhausen 2008). Today in the US, natural gas is both, a
highly traded commodity and a profitable asset in risk management.
Approximately one third of Europe’s total natural gas consumption is used in power
generation (IEA 2008). The remaining demand is characterised by seasonal usage, i.e.
heating and cooling. Storage that is sited near or adjacent to electricity dispatch and
balancing facilities provides optimum flexibility to meet seasonal peaks. In this context, the
EU’s legislative package emphasises the necessity for independent storage operators and
arrogates an increase in transparency of available capacities to third parties as preconditions
for efficient market operations.
Both, non-discriminatory access and reliable, timely market information contribute to the goal
of creating a truly competitive market. In general, competitive markets are characterised by
the law of one price, i.e. they do not provide arbitrage opportunities (in a temporal or spatial
context).
The starting point for the analysis presented in this paper is the interdependency of natural
gas spot and futures market prices linked to the use of storage. The theory of storage
proclaims: (1) price signals influence the operation of storage facilities and infrastructure
investments when a competitive market environment exists and (2) development of natural
gas storage capacities and efficient adjacent markets will reduce volatility of spot prices. The
theory also demonstrates that the return from purchasing the commodity today and selling it
67 This chapter draws on Stronzik et al. (2009). The explanations regarding the regulatory framework
conditions refer to mid-2009. More recent developments are discussed in Chapter 5.
Intertemporal no arbitrage condition 55
for delivery later (the so-called basis) equals the interest forgone by storing the commodity
plus marginal storage cost less marginal convenience yield from an additional unit of
inventory (the latter is defined as the benefits accruing from storage).
This paper investigates the application of the theory of storage to the European gas market
using two indirect tests developed by Fama and French (1987 and 1988), as there is limited
availability of inventory data for Europe. The first test (1988) is based on the relative variation
of spot and futures prices. In the second test (1987) seasonal dummies instead of inventory
data are used to capture variations in the marginal convenience yield. It allows us to study
overall market performance by analysing whether the basis varies with nominal interest rates
corresponding to different maturities of various futures contracts, and to verify the existence
of seasonality in the basis.
The remainder of this paper is organised as follows. Section 4.2 presents a literature
overview of empirical applications for natural gas markets and the key seminal papers on the
theoretical background of the theory. The two indirect approaches are introduced in Section
4.3 and testable hypotheses are derived. Section 4.4 describes the data set with some
stylised facts. Empirical results and their interpretation are presented in Section 4.5, and
summaries and conclusions appear in Section 4.6.
4.2 Literature overview
Increasing international trade and the development of global markets have put commodity
price determination back on the agenda. The theory of storage (Working 1949) shows that
filling quantities are determined by the equivalence of marginal storage cost and the price
spread defined as difference in spot and futures prices. This condition only holds as long as
futures prices do not fall below spot prices, which cannot be assured in the long run.
Consequently, Brennan (1958) includes an additional factor - the convenience yield
measuring an implicit stream of benefits that a consumer of a commodity (such as natural
gas) receives from storage.68 These benefits for the holder of inventories arise because the
stored product depicts an input for further production as well as the ability to meet
unexpected future demand. Based on this, the properties of the convenience yield were
brought into the centre of research.
For example, French (1986) and Fama and French (1987, 1988) derive implications of a
convex marginal convenience yield in terms of futures and spot price variances and
correlations. They illustrate that for a high level of inventory, contemporaneous spot and
futures prices show similar variances and therefore high correlation. Hence, lower inventory
levels imply that the variance of spot prices exceeds the variance of futures prices,
consequently leading to a lower correlation between both prices. More recently, Cho and
McDougall (1990) and Ng and Pirrong (1994) discuss these results and show that the
convenience yield is inversely related to the level of inventory.
68 The original idea of convenience yield was introduced by Kaldor (1939).
Intertemporal no arbitrage condition 56
The majority of recent studies applying the theory of storage to the natural gas industry
examine the North American market. Susmel and Thompson (1997) analyse the relationship
between commodity price volatility and investment in US storage facilities during natural gas
market deregulation. With respect to a switching ARCH model with two states and two
autoregressive terms they show that investments in additional storage facilities are followed
by an increase in volatility. Wei and Zhu (2006) use a bivariate GARCH model to estimate
different risk premiums for the US market. While the dependence of estimated convenience
yields on other explanatory variables confirm the theory of storage, it does not hold for all
resulting risk premiums. Dincerler et al. (2005) and Khan et al. (2005) provide additional
evidence for the dependency of commodity futures prices upon inventory levels with a
special focus on mean-reverting behaviour for various commodity markets in the US,
including natural gas. The predictions of the theory of storage are confirmed for the North
American natural gas market between 1990 and 2002 by Serletis and Shahmoradi (2006).
However, Modjtahedi and Movassagh (2005) find only partial support for the cost-of-carry
theory of the basis determination in their analysis of US data from 1993 to 2004.
In a first application to the European market, Haff et al. (2008) find similar results for the UK
natural gas market with a non-linear effect of storage on the relationship between spot and
futures prices. While Modjtahedi and Movassagh (2005) detect a negative risk premium for
the US, Haff et al. (2008) show the opposite for the UK. Nevertheless, European evidence is
rare.
In Europe, natural gas market participants often face constrained access to storage facilities,
such that it is also useful to survey the literature on pricing behaviour under capacity
constraints. Prior research applies game theory to the strategic behaviour of agents, i.e.
Kirman and Sobel (1974), Maskin and Tirole (1988) or Lang and Rosenthal (1991). The
effect of capacity constraints on collusion and market efficiency is analysed in e. g., Dixit
(1980), Spulber (1981), Brock and Scheinkman (1985), Davidson and Deneckere (1986),
and more recently, Compte et al. (2002) and Kovenock and Dechenaux (2003).
This paper presents the first comparative analysis of major European trading points applying
the theory of storage and using the two indirect tests developed by Fama and French (1987,
1988). With the limited availability of inventory data for natural gas storage in the emerging
European market for the considered period of 2005 - 2008, the indirect approach is well-
suited to provide market insights.69
4.3 Empirical model
The present value model is the most basic description for rational asset pricing: the price of
the asset equals all current and discounted future expected payoffs and benefits from holding
the asset. This implies that different assets with equal expected payoffs in the future must
have equal prices as long as the market is efficient. Consequently, for commodities with
actively traded futures the following no arbitrage condition must hold:
69 Nevertheless, our findings are checked for robustness by incorporating storage levels for the last year.
Intertemporal no arbitrage condition 57
(4-1) ttt YSrF )1( .
The futures price tF of the consumption good (natural gas) must equal the spot price St, plus
forgone interest r, reduced by additional benefits Yt which represent a measure for the
convenience yield net of warehousing cost.70 Expressing the net marginal convenience yield
as a percentage rate yt of the spot price and releasing the assumption of linear rates, the no
arbitrage condition takes the form71
(4-2) Tyr
t
T
t eSF ,
with rtT = r(T - t) and yt
T = y(T - t) being the corresponding rates at time t for the considered
period, we get
(4-3) T
t
T
tt
T
t yrSF lnln ,
where the term on the left side forms the basis, defined by the return from purchasing the
commodity at t and selling it for delivery at a future date T. As a result, the differences in
prices are completely determined by forgone interest and net convenience yield.
According to Pindyck (2001), the convenience yield as benefit from holding the commodity is
a function of several factors such as price volatility and actual storage levels.72 Demand for
storage should increase with price volatility driving up the value of storage due to a greater
need to buffer uncertainty in future production or consumption.73 In turn, convenience yield is
negatively related to current stock levels, i.e. the higher the level of stored goods, the less
value gained from storing an additional unit. We note that weather conditions and the price of
oil may also influence the value of gas storage. Since natural gas is often used to produce
heat and heat demand highly correlates with changes in temperature, itself highly stochastic,
we can expect that natural gas demand is also driven by temperature and should therefore
follow a similar stochastic.74 Oil is a close substitute for natural gas in the heating market and
will also impact the convenience yield for natural gas.
4.3.1 Test on relative price variations
In a first step, we evaluate the no arbitrage condition of Eq. (4-3) with respect to variations of
spot and futures prices in different states of storage activity. We use an indirect approach
developed and first applied by Fama and French (1988), but also considered in Serletis and
Shahmoradi (2006). The main advantage of this method is that it does not require inventory
data (quantities).
70 Throughout the paper we assume constant warehousing cost. 71 A more technical explanation is provided in e.g., Geman (2005). 72 Since we express the convenience yield as a rate, the price level of the commodity as a further factor
mentioned by Pindyck (2001) can be neglected. 73 This is shown in Brennan (1958) and Telser (1958). 74 In the following analysis we neglect weather because the considered trading points are locally separated.
This implies that market efficiency and integration could not be driven by a common weather variable. For more about weather and its impact on storage and the natural gas market in the US see for example Mu (2007).
Intertemporal no arbitrage condition 58
The no arbitrage condition in an efficient market defines the net marginal convenience yield
as the difference between interest rate and basis
(4-4) t
T
t
T
t
T
t SFry lnln .
Using the inverse correlation between the marginal convenience yield and the storage level
derived from theory, the sign of the right hand side of Eq. (4-4) becomes a proxy for the
inventory level (in relative terms) of a (purely) competitive market. Given the observed spot
and futures prices as well as the interest rate, negative ytT (i.e. low benefits from storage)
should therefore correlate with a relative high storage level.
Commodity markets are usually expected to be backwardated most of the time, such that
spot prices exceed discounted futures prices, which goes with a positive net marginal
convenience yield. Since we consider a period of three “gas years”75, a larger share of
positive observations is expected.
The inverse relation between convenience yield and storage levels together with an assumed
convexity of the slope of the marginal convenience yield allow the deduction of further
hypotheses on the relation between spot and futures price variations. The convexity implies
that an additional unit of inventory leads to a larger reduction in marginal convenience yield if
the current level of inventory is relatively low. Therefore, a positive sign of Eq. (4-4) should
correlate with spot price variations that are higher than those in futures contracts. This, in
turn, leads to the expectation of higher variability in marginal convenience yield during
periods of relatively low inventory levels, i.e. a positive sign of Eq. (4-4).76 Similarly for a
negative coefficient, changes in spot and futures prices should be approximately equal,
implying only a moderate variability of the convenience yield.
4.3.2 Test on market performance
The second approach following Fama and French (1987) is also based on Eq. (4-3) and tests
the overall performance of the natural gas market. We regress the basis on the different
variables of the no arbitrage condition.
(4-5) tttttt
T
tt
T
t uoilQQQQrSF lnlnln 6
4
5
3
4
2
3
1
21 .
The net marginal convenience yield ytT is approximated by quarterly dummies Qt
i which
depict the seasonal pattern of natural gas demand. The quarterly dummies equal one if the
75 A “gas year” starts on October 1 and is divided into winter (October-March) and summer seasons (April-
September). Natural gas is usually withdrawn from storage in winter months when demand and prices are high, and injected in summer months when supply exceeds actual demand. The so-called “shoulder seasons” e.g., March and April, in some locales such as Texas that experience rapid changes in weather can also affect supply and demand.
76 Spot price volatility in the electricity market is usually higher if generation is close to overall installed capacity
(Cartea et al. 2008). The corresponding “capacity constraint” in the storage market is natural gas availability to buffer against price fluctuations.
Intertemporal no arbitrage condition 59
corresponding futures contract matures during that period.77 In an efficient market, the
convenience yield will reflect these seasonalities. The oil price (oilt) is included as an
additional factor possibly influencing the value of storage. Moreover, the residuals ut are
modeled as AR(1)-process to control for existing autocorrelations.
We expect the seasonal dummy coefficients to have significant explanatory power. High
winter and low summer demand create arbitrage opportunities that market participants can
exploit in an efficient market. Having controlled for seasonalities, the basis should vary one-
for-one with the nominal interest rate, expecting β1 to be significant and close to 1.
Disregarding other conceivable reasons for market imperfections (e.g., market power at the
wholesale level), a β1 far from 1 implies that storage users do not fully exploit arbitrage
opportunities.
4.4 Data
We use daily data for spot and futures prices from the three major trading points, National
Balancing Point in the UK (NBP), the Dutch Title Transfer Facility (TTF), and Zeebrugge in
Belgium (ZEE) as provided by Heren.78 The data covers the period from October 2005 to
September 2008, i.e. three “gas years”. We focus on the analysis of futures with six-
(futures6) and 12-month maturities (futures12). The corresponding spreads between futures
and spot prices are basis6 and basis12. For oil prices, we use daily spot prices for Brent. For
risk-free interest rates, we use the daily EURIBOR rates for six- and 12-month maturities.
Figure 8 illustrates the logarithmic spot and futures prices for NBP. Futures12 prices are
regularly well above spot prices. Market situations with strong backwardation are observed in
winter 2005/2006. The two peaks in spot prices were caused by shortages in production in
Norwegian natural gas fields accompanied by relatively low temperatures across Europe.
Overall, lower spot prices are observed during the summer season of a “gas year”. The drop
in spot prices down close to zero in October 2006 resulted from a temporary oversupply
caused by opening a new pipeline which connects Norway and the UK. Spot prices reacted
moderately to the close of Rough, the largest storage facility in the UK, due to a fire on
February 16, 2006. Such events indicate the importance of sufficient transmission capacity in
addition to storage facilities for a well-functioning European gas market. In winter, the prices
of futures with six-month maturity are well below the futures prices with longer maturity, while
in summer the opposite is true. Rather similar pictures occur for TTF and Zeebrugge which
are therefore not explicitly highlighted.
77 We use quarterly dummies to map seasonality as they fit best. Qt
2 and Qt
3 represent the summer season of
a “gas year”, and the other two are winter dummies. For example, Qt2 indicates that the considered futures
contract matures during the second quarter of the year (April to June). 78 All prices are expressed as logs. Zeebrugge and NBP price data (p/Therm) is converted into €/MWh using
daily exchange rates and 1 therm = 29.3071 kWh. The factor for oil is 1 bbl = 1.6303 MWh.
Intertemporal no arbitrage condition 60
Figure 8: Spot and futures prices for delivery at NBP (log)
Prior to our empirical analysis, we show the most important properties of the time series
considered. Table 11 gives the results of the unit root analysis. The Augmented Dickey Fuller
(ADF) and the Phillips Perron (PP) tests indicate non-stationarity of oil prices, interest rates
and futures prices. Spot prices as well as basis6 and basis12 are stationary in levels.79
Given the physical interconnection of the hubs, we refer to Granger (1969) in order to
investigate bidirectional interdependencies between hubs and products, as shown in Table
12.
We observe that NBP and ZEE dominate price developments at TTF.80 While oil prices
impact natural gas futures prices at all three hubs, only spot prices at TTF are affected.
These results suggest a closer relation between NBP and ZEE. In the following section we
therefore expect similar results for NBP and ZEE when testing the no arbitrage condition.
However, the specific impact of oil on our results is less clear due to asymmetric effects on
spot and futures prices.
79 All time series with unit root in levels are stationary in first differences. 80 Neumann et al. (2006) confirm these results.
0.6
1.1
1.6
2.1
2.6
3.1
3.6
4.1
4.6
[Eu
ro/M
Wh
]
Spot Futures6 Futures12
Intertemporal no arbitrage condition 61
Table 11: Unit root tests
Natural gas prices
Variable NBP ZEE TTF
ADF PP ADF PP ADF PP
Spot (log) -3.09** -3.02** -3.06** -2.92** -2.85** -2.87**
Futures6 (log) -1.52 -1.55 -1.49 -1.52 -1.37 -1.46
Futures12 (log) -1.91 -1.94 -1.97 -1.99 -1.85 -1.84
Basis6 (log) -3.71*** -3.35** -3.61*** -3.27** -3.21** -3.78***
Basis12 (log) -4.82*** -5.11*** -4.65*** -4.82*** -4.78*** -5.26***
Common variables
ADF PP
Oil (log) -1.24 -1.10
Euribor6 -1.18 -1.25
Euribor12 -1.39 -1.46
Notes: Tests use a constant but not a time trend. For ADF, the lag length is selected according to Schwarz Information Criterion (SIC). For PP, the lag length is determined by referring to Bartlett kernel with Newey-West bandwidth. The provided numbers denote the t-ratios. *, **, *** indicate significance at the 10, 5 and 1 %-levels.
Table 12: Granger causality
A. Between hubs
Variable Spot (log) Futures6 (log) Futures12 (log)
NBPZEE 0.229 2.125 1.450
ZEENBP 3.063** 2.938* 0.473
NBPTTF 24.722*** 5.926*** 4.868***
TTFNBP 3.291** 1.025 0.754
ZEETTF 26.693*** 6.787*** 4.858***
TTFZEE 2.967* 1.282 0.346
B. Relation with oil (log)
Variable Spot (log) Futures6 (log) Futures12 (log)
OilNBP 1.028 4.685*** 7.314***
OilZEE 1.196 4.769*** 7.111***
OilTTF 3.142** 5.426*** 8.827***
Notes: Lag length is set to 2. Numbers shown denote the F-statistic of the Granger causality test. *, **, *** indicates significance at the 10, 5 and 1 %-levels.
Intertemporal no arbitrage condition 62
4.5 Results
We now turn to the results of our two empirical models which are described separately in the
following two subsections.
4.5.1 Relative price variation
The relative price variations allow us to assess the performance of natural gas storage by
comparing the observed market outcome to a competitive benchmark, given by the
intertemporal no arbitrage condition. As previously mentioned, net marginal convenience
yield in competitive markets is reflected by corresponding commodity prices and can be
calculated according to Eq. (4-4). For the two maturities, yt6 and yt
12, the first two panels of
Table 13 show the number of observations (Panel A) and average values of the convenience
yield (Panel B) for the three hubs ordered by the sign of the yield. Contrary to Serletis and
Shahmoradi (2006) who report at least a more or less equal share for the US,81 we observe
a dominance of negative net marginal convenience yields for both maturities. The longer the
maturity, the stronger the disparity. This translates into benefits attached to natural gas
storage which are lower than actual storage costs for the majority of time.
Table 13: Summary statistics of daily convenience yields
yt
NBP Zeebrugge TTF
A. Number of observations
negative positive All negative positive All negative positive All
6 468 288 756 485 270 755 497 258 755
12 644 112 756 653 102 755 710 46 755
B. Average values
6 -0.769 0.197 -0.401 -0.669 0.190 -0.362 -0.407 0.113 -0.210
12 -0.498 0.128 -0.405 -0.456 0.134 -0.376 -0.401 0.051 -0.373
C. Standard deviations of changes
6 0.708 0.175* 0.736 0.604 0.180* 0.644 0.212 0.116* 0.311
12 0.688 0.149* 0.675 0.600 0.157* 0.596 0.394 0.067* 0.341
D. Ratios of standard deviation of per cent futures price changes to standard deviation of per cent spot price changes
6 0.473 0.215+ 0.380 0.513 0.188
+ 0.388 0.486 0.257
+ 0.422
12 0.353 0.194+ 0.312 0.289 0.405
+ 0.331 0.260 0.677
+ 0.341
Notes: * indicates rejection of the null hypothesis of equal variances at the one percent level. + indicates rejection of the null hypothesis of equal ratios at the one percent level.
81 Geman and Ohana (2009) confirm this for the US natural gas market.
Intertemporal no arbitrage condition 63
Panels C and D in Table 13 illustrate price variations. The convexity of the convenience yield
with respect to inventory levels implies that high levels should correlate with only moderate
variations in ytT, which is not the case for the data under consideration. At all hubs and for
both maturities, standard deviations of changes in ytT are larger when the net convenience
yield is negative. Furthermore, in competitive markets of storable commodities, demand
shocks create more independent variations of spot and futures prices when inventory is low.
In this context, spot price changes exceed futures price changes. Panel D reports the ratios
of standard deviations of (percent) futures price changes to the standard deviations of
(percent) spot price changes. Meeting expectations, the ratios are higher in the case of
negative net marginal convenience yields (indicating relatively high storage levels) with two
exceptions, the twelve-month maturities at ZEE and TTF.
Consequently, the data confirm that shocks create larger variations in near-term than in
longer-term futures. If all observations for each trading point are taken into account, ratios
are slightly higher for six-month maturities.
Using actual data on inventory levels allows us to investigate the expected inverse
relationship between inventory level and net marginal convenience yield.82 Table 14 reports
the corresponding correlations of prices at the trading hubs as well as other conceivable
factors influencing the value of storage with net marginal convenience yield.83 The oil price is
presented in first differences due to its unit root characteristic, and price volatility is defined
as 1loglog2 tt PP with Pt being the corresponding spot price of oil or natural gas.
Table 14: Correlation analysis for convenience yields
NBP ZEE TTF
Variable y6 y12 Level [%]
y6 y12 Level [%]
y6 y12 Level [%]
Level [%]
0.024
(0.821)
0.447
(0.000)
- 0.067
(0.532)
0.464
(0.000)
- 0.141
(0.183)
0.286
(0.006)
-
Volatility gas
-0.189
(0.077)
-0.191
(0.075)
0.274
(0.010)
-0.199
(0.063)
-0.208
(0.052)
0.200
(0.061)
-0.115
(0.276)
-0.271
(0.009)
0.151
(0.152)
Volatility oil
0.048
(0.659)
0.071
(0.510)
0.180
(0.094)
0.038
(0.727)
0.046
(0.672)
0.180
0.094)
0.048
(0.654)
0.064
(0.550)
0.218
(0.038)
Oil (dlog)
0.122
(0.257)
-0.059
(0.583)
-0.198
(0.065)
0.116
(0.280)
-0.056
(0.605)
-0.198
(0.065)
0.127
(0.230)
-0.054
(0.614)
-0.105
(0.322)
Notes: Number in brackets report p-values.
82 Storage data for aggregated major European hub areas is published on a weekly basis by the industry
association Gas Infrastructure Europe (GIE) since the beginning of 2007. On October 1, 2007, GIE changed from a four-hub to an eight-hub representation. The values before and after the change are only partly comparable, leading to structural breaks in the series.
83 For the period of study only the UK experiences an increase in capacity. For Belgium and the Netherlands,
the figures remain constant (see IEA 2008).
Intertemporal no arbitrage condition 64
Contrary to expectations, we find a positive correlation of inventory level (in percent) with
twelve-month maturity yields, and that natural gas price volatility correlates negatively with
both convenience yield approximations (yt6, yt
12). Moreover, natural gas price volatility
correlates positively with inventory (level in percent) at NBP and ZEE.
However, as mentioned in the literature section, the convenience yields should increase with
price volatility and decrease with inventory. Moreover, high spot price volatility should be
observed during times of low stocks due to the reduced buffering function of storage.
Neither the oil price nor its volatility exhibits a significant correlation with respect to ytT; the
influence of oil on inventory levels, however, is clear. Concerning the interrelation between
the two price volatilities, we can find neither a significant correlation nor causality in a
Granger sense. The former also holds when time lags are considered.
In the next step, the impact of several factors on the net marginal convenience yield is
analysed. For the oil price, the current as well as the one period lagged value are included as
explanatory variables. Error terms are modeled as an AR(1) process. Except for the 12-
month maturity at TTF, volatility of natural gas spot prices is the only factor showing a
significant negative impact on the dependent variable, thus contradicting expectations (see
Table 15).84 Concluding, most of our findings do not confirm predictions of the theory of
storage.
Table 15: Estimation results for convenience yields
NBP ZEE TTF
Variable y6 y12 y6 y12 y6 y12
Level [%] -0.146
(0.121)
0.158
(0.157)
-0.122
(0.110)
0.149
(0.189)
-0.073
(0.122)
-0.015
(0.099)
Volatility gas -0.671***
(0.253)
-0.638**
(0.288)
-0.745***
(0.255)
-0.556*
(0.318)
-0.923**
(0.414)
-0.521
(0.456)
Volatility oil -0.278
(0.733)
-0.051
(0.736)
-0.506
(0.820)
-0.291
(0.747)
0.067
(0.624)
0.219
(0.523)
Oil (dlog) 0.347
(2.020)
-1.448
(1.405)
0.682
(1.879)
-1.016
(1.260)
1.602
(1.944)
-0.202
(1.147)
Oil (dlog, -1) 0.161
(1.330)
0.738
(1.262)
0.351
(1.220)
0.734
(1.090)
0.202
(1.186)
0.657
(1.012)
Adj. R2 0.824 0.821 0.833 0.845 0.829 0.580
Notes: *, **, *** indicates significance at the 10, 5 and 1 %-levels. The number in brackets reports standard errors.
The following section presents the results from the test on overall market performance
covering the entire sample from October 2005 to September 2008.
84 When considering the only “gas year” which is completely covered by inventory data (2007/08), results do
not change significantly.
Intertemporal no arbitrage condition 65
4.5.2 Market performance
Based on Eq. (4-5), we test the overall market performance with respect to the daily basis for
each considered maturity and trading place, i.e. basis6 and basis12.85
The difference in spot and futures prices for all three trading points is well explained (see last
row in Table 16). While oil has hardly any effect on the bases, the seasonal patterns are
evident in all cases. Bases at TTF reveal a slightly clearer seasonality, which is mainly due to
a much higher volatility at NBP and ZEE observed during winter 2005/2006. The estimated
β-coefficients are usually lower for summer dummies than for Q1 and Q4. In case of basis6
this is expected since winter spot prices are compared with futures prices, reflecting market
expectations of the upcoming summer season. Nevertheless, no such obvious explanation
regarding basis12 exists. The lowering effect of summer dummies on the basis indicates,
compared to the winter cycle, a higher convenience yield during the second and third quarter
of a year. Considering the findings from the previous sections, especially the inverse relation
between price volatility and convenience yield, we can also ascribe the effect to higher price
volatilities in winter.
Table 16: Estimation results for 6-month and 12-month bases
NBP ZEE TTF
Variable Basis6 Basis12 Basis6 Basis12 Basis6 Basis12
r (d) 18.248
(14.787)
21.737*
(12.629)
20.693
(14.796)
30.575***
(11.656)
-10.860
(17.171)
11.426
(11.387)
Q1 0.412*** 0.524*** 0.378*** 0.500*** 0.478*** 0.541***
Q2 0.285* 0.090 0.284* 0.081 0.113* 0.217***
Q3 -0.209 0.015 -0.215 0.010 -0.035 0.140***
Q4 0.264** 0.503** 0.235** 0.476* 0.413*** 0.342***
Oil (dlog) -0.371* -0.320 -0.276 -0.234 -0.121 0.111
Oil (dlog, -1) -0.222 -0.150 -0.172 -0.112 0.083 -0.123
Adj. R2 0.939 0.864 0.939 0.872 0.913 0.835
Notes: *, **, *** indicates significance at the 10, 5 and 1 %-levels. Standard errors (number in brackets) reported for interest rates r (first difference) only.
The result for the interest rates is less intuitive. Here, the EURIBOR is only significant for
basis12 at NBP and ZEE with estimated coefficients far from one.86 For TTF, the interest
rate provides no impact.
85 The appropriate lag structure for oil and natural gas prices is determined by a VAR (p) analysis based on
Akaike Information Criterion (AIC). 86 Taking the first difference of the bases does not change this result. Analysing separate “gas years” shows
lower coefficient values in the last year for basis12. Nevertheless, values are still well-above one. Estimated coefficients are only significant in the third year for all hubs and in the first for ZEE only.
Intertemporal no arbitrage condition 66
Therefore, given the seasonal influence, we can confirm our first hypothesis of the indirect
performance test, i.e. storage facilities realise seasonal arbitrage. On the other hand, the
results with respect to the interest rate hint at substantial arbitrage potentials that are not
exploited by market players.87 This does not necessarily imply that the trading market is
inefficient, but that lack of storage facilities or restricted access to these may exist.
Possible obstacles concerning the appropriate use of storage in the European natural gas
market are often mentioned by the European Commission and European regulators (i.e.
limited access to infrastructure; insufficient information; missing secondary markets for
unused capacities). Additionally, a non-transparent market with high transaction costs
hinders participants from detecting arbitrage opportunities or even making it impossible to
exploit identified arbitrage potentials. In this context, high transaction costs can also partially
explain negative net marginal convenience yield for most periods, especially when
warehousing costs include storage fees and transactional costs.
Assuming market clearance on the storage market, net marginal convenience yield reflected
in natural gas prices is identical for both incumbents and entering firms. Given that the
incumbent already holds capacities, i.e. faces lower transaction costs, she would tend to
ascribe a lower value to natural gas on hand, unlike a new shipper. An owner’s strategic
considerations could potentially assign more value to storage than merely operational value.
The phenomenon of higher natural gas price volatility when inventory is high has also been
identified for the US natural gas market over the last 15 years (Geman and Ohana 2009) and
is related to literature dealing with pricing behaviour under capacity constraints in electricity
markets. Storage capacity might not only be constrained in case of low inventory levels but
also when facilities are filled up with gas which makes injections more difficult. Hence, a
positive correlation between storage levels and volatility could imply that the upper constraint
is more binding. However, this does not explain the inverse relationship between net
marginal convenience yield and natural gas price volatility.
4.6 Conclusions
This paper has assessed the performance of natural gas storage by testing the observed
market outcome against a competitive benchmark, the intertemporal no arbitrage condition.
The analysis is based on indirect tests by Fama and French (1987, 1988) and allows us to
study the performance of three major European trading points.
First, we use net marginal convenience yield, i.e. the benefit from natural gas in stock less
warehousing costs, as a proxy for inventory. As a robustness check, available information on
actual inventory levels covering only half of the whole period from October 2005 to
September 2008 is considered. We find the results less intuitive, and largely contradicting
expectations deduced from theory. Most remarkably, net marginal convenience yield appears
to be inversely related to the volatility of natural gas spot prices.
87 For the UK, Haff et al. (2008) show that even after incorporating inventory levels the no arbitrage condition
must be rejected.
Intertemporal no arbitrage condition 67
Second, we introduce seasonal dummies to map storage levels controlling for seasonality in
convenience yield and for a one-for-one relation between the basis, i.e. the difference
between futures and spot prices, and the risk-free interest rate. While seasonality is found at
all three hubs, the interest rate exhibits a significant influence only at the National Balancing
Point (NBP) in the UK market and the Belgian Zeebrugge. However, coefficients for both are
far away from one. Nevertheless, these markets perform slightly better than the Dutch Title
Transfer Facility hub (TTF). Furthermore, our analysis confirms the significant role of the
price of oil in the European natural gas market.
The indirect tests indicate a fairly high arbitrage potential, especially in the short term, that
remains unexploited by market participants, and that hints at market imperfections. Given the
limited access to infrastructure, insufficient information available to market participants and a
missing liquid secondary market for these products across Europe, we suggest that the
results may be best explained by a lack of transparency that prevents players from detecting
arbitrage opportunities, and by the high transaction costs that prevent them from exploiting
already identified potentials.
It will be interesting to observe how various national and Europe-wide regulatory initiatives
will help to move the market towards the competitive benchmark. If these developments are
successful, we will likely observe a move away from the more technical utilisation of storage
facilities (seasonal arbitrage) to a more service-oriented operation and the introduction of
derivative products, such as virtual storage.
Although this paper focuses on the actual usage of existing storage capacity, we suggest
that the findings have implications regarding investment activities for capacity expansion.
Lower transaction costs together with increased market transparency will drive up net
marginal convenience yield and lead to a better utilisation of existing facilities, thereby
reducing ceteris paribus the need for new infrastructure. On the other hand, improved
arbitrage conditions might increase demand for storage. Nevertheless, players should use
existing storage capacities efficiently before considering investment in new facilities.
Future research should incorporate additional explanatory variables (e.g. weather),
especially with regard to the marginal convenience yield. Storage levels also should be
addressed in future analysis as soon as available information spans a longer time period.
This will allow studying the impacts of specific regulatory measures.
Intertemporal no arbitrage condition 68
References
Brennan, M.J. (1958), The Supply of Storage, American Economic Review 48, 50-72.
Brock, W. A. and J. Scheinkman (1985), Price-setting Supergames with Capacity Constraints, Review
of Economic Studies LII, 371-382.
Cartea, A., M. G. Figueroa and H. Geman (2008), Modelling Electricity Prices with Forward Looking
Capacity Constraints, University of London, Commodities Finance Centre, Working Paper,
London.
Cho, D.W. and G.S. McDougall (1990), The Supply of Storage in Energy Futures Markets, The Journal
of Futures Markets 10, 611-621.
Compte, O., F. Jenny, and P. Rey (2002), Capacity Constraints, Mergers and Collusion, European
Economic Review 46, 1-29.
Davidson, C. and R. Deneckere (1986), Long-run Competition in Capacity, Short-run Competition in
Price, and the Cournot Model, Rand Journal of Economics 17(3), 404-415.
Dincerler, C., Z. Khokher and T. Simin (2005), An Empirical Analysis of Commodity Convenience
Yields, Working Paper, June, University of Western Ontario.
Dixit, A. (1980), The Role of Investment in Entry Deterrence, Economic Journal 90, 95-106.
European Commission (2007), Inquiry Pursuant to Article 17 of Regulation (EC) No 1/2003 into the
European Gas and Electricity Sectors, COM(2006) 851 Final, January, Brussels.
Fama, E.F. and K.R. French (1987), Commodity Futures Prices: Some Evidence on Forecast Power,
Premiums, and the Theory of Storage, The Journal of Business 60(1), 55-73.
Fama, E.F. and K.R. French (1988), Business Cycles and the Behavior of Metal Prices. Journal of
Finance 63, 1075-1093.
French, K.R. (1986), Detecting Spot Price Forecasts in Futures Prices, The Journal of Business 59,
39-54.
Geman, H. (2005), Commodities and Commodity Derivatives: Modeling and Pricing for Agriculturals,
Metals and Energy. Wiley Finance.
Geman, H. and S. Ohana (2009), Forward Curves, Scarcity and Price Volatility in Oil and Natural Gas
Markets, Energy Economics 31(4), 576-585.
Granger, W. W. J. (1969), Investigating Causal Relations by Econometric Models and Cross-spectral
Methods, Econometrica 37(3), 424-438.
Haff, I. H., O. Lindqvist and A. Løland (2008), Risk premium in the UK Natural Gas Forward Market,
Energy Economics 30(5), 2420-2440.
Hirschhausen, Ch. von (2008), Infrastructure Regulation, Investment and Security of Supply: A Case
Study of the Restructured US Natural Gas Market. Utilities Policy 16(1), 1-10.
IEA [International Energy Agency] (2008), Natural Gas Information, Paris.
Kaldor, N. (1939), Speculation and Economic Stability, Review of Economic Studies 7, 1-27.
Intertemporal no arbitrage condition 69
Khan, S., Z. Khokher and T. Simin (2005), Inventories, Convenience Yields and Mean Reversion,
Working Paper, October, University of Western Ontario.
Kirman, A. P. and M. J. Sobel (1974), Dynamic Oligopoly with Inventories, Econometrica 42(2), 279-
287.
Kovenock, D. and E. Dechenaux (2003), Endogenous Rationing and Tacit Collusion in Price-quantity
Super Games, manuscript, Purdue University.
Lang, K. and R. W. Rosenthal (1991), The Contractors’ Game, Rand Journal of Economics 22(3), 329-
337.
Maskin E. and J. Tirole (1988), A Theory of Dynamic Oligopoly, II: Price Competition, Kinked Demand
Curves, and Edgeworth Cycles, Econometrica 56(3), 571-599.
Modjtahedi, B. and N. Movassagh (2005), Natural-gas Futures: Bias, Predictive Performance, and the
Theory of Storage, Energy Economics 27, 617-637.
Mu, X. (2007), Weather, Storage, and Natural Gas Price Dynamics: Fundamentals and Volatility,
Energy Economics 29(1), 46-63.
Neumann, A. B. Siliverstovs, and C. von Hirschhausen (2006), Convergence of European Spot Market
Prices for Natural Gas? A Real-Time Analysis of Market Integration Using the Kalman Filter.
Applied Economics Letters 13(11), 727-732.
Ng, V.K. and S.C. Pirrong (1994), Fundamentals and Volatility: Storage, Spreads, and the Dynamics
of Metal Prices, The Journal of Business 67, 203-230.
Pindyck, R. S. (2001), The Dynamics of Commodity Spot and Futures Markets: A Primer, The Energy
Journal 22(3), 1-29.
Serletis, A. and A. Shahmoradi (2006), Futures Trading and the Storage of North American Natural
Gas, OPEC Review 30(1), 19-26.
Spulber D. F. (1981), Capacity, Output and Sequential Entry, American Economic Review 71(3), 503-
514.
Stronzik, M., M. Rammerstorfer and A. Neumann (2009), Does the European natural gas market pass
the competitive benchmark of the theory of storage? Indirect tests for three major trading
points, Energy Policy 37(12), 5432-5439.
Susmel, R. and A. Thompson (1997), Volatility, Storage and Convenience: Evidence from Natural Gas
Markets, The Journal of Futures Markets 17(1), 17-43.
Telser, L. G. (1958), Futures Trading and the Storage of Cotton and Wheat, Journal of Political
Economy 66(3), 233-255.
Wei, S.Z.C. and Z. Zhu (2006), Commodity Convenience Yield and Risk Premium Determination: The
Case of the U.S. Natural Gas Market, Energy Economics 28, 523-535.
Working, H. (1949), The Theory of the Price of Storage, American Economic Review 39, 1254-1262.
Policy paper 70
5 Policy paper
Europe on its way to a single natural gas market: How far have we
come?88
5.1 Introduction
One of the major goals of European energy policy is the establishment of a single and
competitive internal gas market. In 2009, the European Union (EU) adopted Directive
2009/73/EC on common rules for the gas sector to foster market integration across Europe.
The 2009 Directive is already the third of its kind and repeals the second gas Directive from
2003 (2003/55/EC). When the EU started the liberalisation process with the adoption of the
first Directive in 1998 (98/30/EC), vertically integrated gas companies acting as (regional)
monopolies and serving the whole value chain were the dominant organisational structure
throughout Europe. Joskow (2007: 89), in his well-known and often cited Beesley Lecture,
describes the challenge of the EU to achieve its goal of a well-functioning internal gas
market: “The liberalization of the electricity and natural gas sectors involves a complex
institutional transformation from industries composed of vertically integrated regulated
monopolies (…) to industries with unregulated competitive segments (…) and regulated
(primarily) monopoly transmission and distribution network segments. For liberalized systems
to work well it is necessary to implement sound market institutions and market designs for
the competitive segments, vertical and horizontal restructuring, unbundling of competitive
and regulated network services, and a compatible regulatory framework to govern the
regulated network segments. Poorly performing network segments can undermine the
performance of the competitive segments and adversely affect supply security directly and
indirectly through their effects on the performance of competitive (…) markets.”
As indicated in Figure 9, the main areas of regulatory concern are located in the midstream
segment of the value chain. While gas transport (transmission and distribution) is usually
regarded as the classic representative of an essential facility with natural monopoly
characteristics accompanied by irreversibility (cf. e.g. Gordon et al. 2003 and Hirschhausen
et al. 2007), storage and terminals for the regasification of liquefied natural gas (LNG) are
less capital-intensive compared to pipeline investments, with weaker economies of scale (cf.
e.g. Bertoletti et al. 2008).89 As storage and LNG provide market participants with some
additional flexibility to cope with demand uncertainty (e.g. through intertemporal arbitrage or
peak shaving), they are of great importance for a well-functioning natural gas market.
Therefore, storage and LNG terminals are – at least to some extent – also considered when
88 The paper is based on information on the on-going liberalisation process available by end of October 2011. 89 Beyond storability and some flexibility through LNG, natural gas shows characteristics quite different from
electricity. Long-distance transportation costs (per energy unit transported) are lower resulting partly in pipe-to-pipe competition, travels are normally of a broader scale leading to higher cross-border exchanges, and energy flows are more predictable implying less strict balancing requirements (Ascari 2011).
Policy paper 71
designing the regulatory framework.90 Upstream (e.g. production and import) and
downstream activities (e.g. supply to end-users) have no natural monopoly characteristics
and thus, should be open to competition. The same holds for wholesale where trading takes
place.
Figure 9: Value chain of the gas sector
Source: Own depiction partly based on Rüster (2010).
The question is how the EU copes with the challenge of restructuring the natural gas sector
as described by Joskow in his Beesley Lecture. Is the third gas Directive pointing into the
right direction? Is the EU on track to a competitive internal gas market that is intended to be
completed by 2014, only three years from now?91
The remainder of the paper is organised as follows. Next, an interim assessment is
presented roughly covering the first two gas Directives together with some introductory
remarks regarding the third legislative package on energy markets, thereof Directive
2009/73/EC is just one part. This is followed by a more detailed description of the most
recent developments of European gas market regulations with a special focus on two topics.
Section 5.3 analyses market access conditions, while Section 5.4 focuses on security of
supply issues. Finally, concluding remarks on the status of the European gas market are
presented together with a summary of the main policy recommendations.
90 Since storage facilities and LNG terminals are distinct from pipeline infrastructure, both elements are shaded
in Figure 9. 91 The target of 2014 for completion of the integrated European gas market has been announced by the
European Council (2011).
Exploration&
Production
Processing
Transmission• international• national
StorageTrade (wholesale)• commodity• capacity• long-term contracts
DistributionEnd-users (retail)• industry • households
ShippingLiquefaction(LNG-terminal)
Regasification(LNG-terminal)
Upstream Midstream Downstream
LNG
Policy paper 72
5.2 Overview of regulation history
In the following, the evolution of European natural gas market regulations before the third
gas Directive is outlined on the basis of the OECD database “Indicators of regulation in
energy, transport and communications” (ETCR). For several OECD countries ETCR provides
annual data of indicators, which measure changes in national gas market regulations. The
national and regional (average score of relevant countries) indicators shown in Figure 10
cover four main areas, i.e. entry regulation, ownership structure, vertical integration and
market structure. In turn, each of these four main indicators consists of three sub-indicators.
Each sub-indicator is scaled between 0 (full market deregulation) and 6 (corresponding to the
most restrictive conditions regarding competition). E.g., third party access (TPA) as a sub-
indicator of entry regulation distinguishes between regulated, negotiated and no TPA, of
which regulated TPA is usually regarded as the most competition-friendly framework
condition. Thus, if a country establishes regulated third party access, a score of zero is
assigned to this sub-indicator. To aggregate the sub-indicators to the overall regulatory
indicator equal weights are used except for the sub-indicators for vertical integration. Hence,
the national indicators of gas market regulations are also scaled between 0 and 6. A score of
6 means that a country has not set up any kind of gas market deregulation yet. The
indicators are updated by the OECD on an irregular basis. The current database provides
time series data until 2007, roughly covering the first two European gas Directives. Since
ETCR has a few drawbacks (e.g. an arbitrary weighting structure and the negligence of TSO
unbundling), I use it only for a first indicative assessment of regulatory steps taken until
2007.92 Developments thereafter and a more thorough discussion of European natural gas
sector regulations are presented in Sections 5.3 and 5.4.
Until the end of the 1990s, regulation in Europe was solely driven by national authorities, with
frontrunners regarding gas market liberalisation such as the UK and laggards like France, but
with hardly any harmonisation across countries (see UK and FRA in Figure 10). Compared to
this uncoordinated pre-phase of market liberalisation, noticeable improvements towards
competitive conditions at the European level have not been achieved until the adoption of the
gas Directive in 1998 (bold black line EU Gas in Figure 10).93 The Directive consisted of soft
regulatory requirements, which Member States had to transpose into national law. Next to a
stepwise opening of the end-user market with a target of 30% until 201094, Member States
could choose between regulated or only negotiated TPA to the grid. Concerning vertical
separation of utilities, accounting unbundling was set as a minimum standard, meaning that
utilities merely needed to set up separate accounts for their different services. This regulation
resulted in vastly differing rules across Europe. E.g., while some countries opted for
92 For a comprehensive description of the indicator see Conway and Nicoletti (2006). A more in-depth discussion of the major drawbacks can be found in Brau et al. (2010) and Growitsch and Stronzik (2011). The development of the indicator for the US may serve as an example for the limitations of ETCR, especially with regard to the coverage of vertical separation. The upward shift in 1998 is merely due to data issues. For two sub-indicators of vertical integration, less favourable specifications are reported in the OECD database which contradicts market observations. Especially the two federal regulations, FERC Order No. 636 in 1992 and FERC Order No. 637 in 2000, are regarded as the most important steps for a competitive US gas market that fully blossomed only after 2000 (cf. e.g. Jensen 2007, Makholm 2007 and 2011).
93 Average of EU Member States covered by ETCR. 94 This target meant that by 2010 30% of consumers should have a choice between different suppliers.
Policy paper 73
negotiated TPA (like Germany), others implemented regulated TPA. Therefore, the Directive
left the market highly fragmented with hardly any competition up- or downstream. The
European gas Directive of 2003 tightened the bottom lines of the regulatory framework. Full
opening of end-user markets – for households as well as industrial consumers - had to be
achieved by July 2007. Regulated TPA became mandatory and legal unbundling for
transmission network operators was set as the minimum standard. Though integrated gas
utilities had to set up separate companies for their different services, networks and
competitive segments (like production, import, wholesale and retail) could stay within the
same holding.
Figure 10: Interim assessment
Source: Own calculations based on the OECD database “Indicators of regulation in energy, transport and communications” (ETCR) .
Shortly after the adoption of the second Directive and due to the successful completion of the
accession process, the EU welcomed 12 new Member States, mainly located in Eastern and
South-East Europe.95 That meant that not only the implementation of more competition-
friendly regulations had to be advanced in the EU-15, but also that 12 additional members
had to be integrated into the on-going process of market liberalisation.96 As indicated by
Figure 10, the development has not slowed down and further improvements towards
95 Cyprus, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Malta, Poland, Slovak Republic and Slovenia
joined in 2004, Bulgaria and Romania in 2007. 96 With regard to the EU, the new Member States (based on Eurostat data) account for roughly 13% of natural
gas demand, 9% of gas production, 20% of population (corresponding to potential end-users) and 25% of the territory (corresponding to potential network dimension).
Policy paper 74
competitive framework conditions can be observed. On average, the new Member States
(EU East) have caught up with traditional EU countries like Austria, Belgium, Luxembourg,
the Netherlands and Germany (EU Central). Nevertheless, timing and speed of implementing
regulatory measures differ substantially between countries and regions. While structural
differences regarding the regulatory framework have narrowed across the EU, they have not
completely vanished.
Despite the improvements throughout Europe, there is still a considerable gap to the US,
which hosts the most mature and competitive natural gas market worldwide. Moreover, gas
market regulations in Europe (EU Gas) lag roughly five years behind the liberalisation
process of the European electricity market (EU Ele). These less favourable judgements
about the status of the internal European gas market achieved by 2007 are underpinned by
the sector inquiry of the European Commission (2007). Having investigated the energy
sectors quite in depth, the Commission was still unsatisfied with market outcomes, especially
in the gas sector. The Commission criticised low switching rates especially in the household
sector and found the market highly concentrated and still dominated by incumbents.
Concerning wholesale, limited transmission and storage capacities available to market
participants as well as a general lack of transparency were identified as major obstacles to a
well-functioning market environment.
The sector inquiry triggered a political debate that led to the adoption of the third legislative
package with three main elements regarding the gas sector. While Directive 2009/73/EC
tightens the minimum standard for vertical separation of energy companies once again and
aims at improving market transparency and promoting investments into cross-border
transportation capacities, regulation No. 715/2009 is mainly concerned with access rules for
transmission networks, storage and LNG facilities. Moreover, the European Network of
Transmission System Operators for Gas (ENTSOG) has been formally established under
regulation No. 715/2009. The two main duties of ENTSOG are the elaboration of a European
network code and the 10-year network development plan (TYNDP).97 The third element of
the package is the establishment of the Agency for Cooperation of Energy Regulators
(ACER), which takes over most of the work previously carried out by the European
Regulators Group for Electricity and Gas (ERGEG). ACER was launched on March 3rd, 2011.
Furthermore, several initiatives have commenced aiming at accelerating the efforts towards
the internal gas market. Currently, various framework guidelines are under development (e.g.
on capacity allocation mechanisms, congestion management procedures and balancing) or
under review (e.g. for storage system operators). All activities are coordinated by the Madrid
Forum, a platform of relevant stakeholders concerned with gas market regulation (Member
State governments, the Commission, industry and consumers). The Forum also hosts the
work on the Energy Infrastructure Package (EIP), which deals with network investments and
the discussions about a vision for the future design of the European gas market, the so-
called target model.
97 The network code is expected to be finalised by early 2012. The TYNDP has been published in February
2011 and shall be updated on a regular basis.
Policy paper 75
In the next two sections, developments of the regulatory framework since 2007 are analysed
regarding their capability of closing the gap to the US, as indicated in Figure 10. The main
focus is on market access conditions, especially with regard to wholesale, and security of
supply issues.
5.3 Market access
5.3.1 Unbundling
Since gas networks can be classified as monopolistic bottlenecks, the unbundling regime for
these pipelines is crucial regarding accessibility of the European natural gas market. The
question of how vertical separation of transmission system operators (TSOs) should be
arranged was of particular interest in the discussions preceding the third legislative package.
The Commission argued strongly in favour of the most rigid form, ownership unbundling,
stressing the benefits like circumventing the risk of market foreclosure. Ownership
unbundling means that the network operator is not allowed to own companies (or substantial
parts thereof) operating in the competitive segments of the value chain. Several Member
States (e.g. Germany and France) opposed those efforts of the Commission doubting the
benefits, raising juridical arguments, and pointing at potential drawbacks like the loss of
economies of scope. As a compromise, the Directive now allows for three options: Besides
ownership unbundling, Member States have the choice between the independent
transmission operator (ITO) and the independent system operator (ISO). The ITO approach
is close to the arrangements under the 2nd Directive, but requires a more stringent functional
unbundling.98 In the ISO case, the network is split up into the actual assets, which can still
belong to a gas holding company, and the pure network operation, which has to be covered
by an ownership unbundled undertaking. Therefore, ISO is kind of in between ownership
unbundling and the ITO approach. However, since network ownership can be with the gas
holding, ISO (like ITO) is a kind of legal unbundling.
From an economic point of view, a specific vertical structure will be preferable if it is welfare-
superior to other modes. On the one hand, it is argued that the stricter vertical separation is
designed, the less incentive a network operator has to discriminate between affiliated
companies and third parties, which are active in the competitive segment. Thus, competition
on down- and upstream markets will be fostered and the risk of vertical foreclosure reduced.
On the other hand, vertical separation may lead to a loss of economies of scope and to a
double mark-up billed to customers (cf. e.g. Laffont and Tirole 1993 and Buehler 2005). The
latter problem, the so-called “double marginalisation”, assumes one-part tariffs and occurs if
competition on the adjacent markets is imperfect. As is well known from literature (cf. e.g.
Laffont et al. 1998 and Wright 2002), the problem is weakened through two-part pricing,
which is commonly practiced in gas retail and to a lesser extent at the wholesale level.
However, the problem only vanishes completely if the variable part equals marginal cost. 98 E.g., according to Art. 17 of 2009/73/EC, ITO has to be fully equipped with the necessary staff and assets,
which was not required under the second Directive. Shared services between the holding and the network operator are no longer possible.
Policy paper 76
Accounting for two-part tariffs, Vickers (1995) compares vertical integration with separation
under access price regulation and finds mixed welfare results depending on the actual
demand and market conditions, and the degree of economies of scale related with the
competitive activities. Under imperfect competition (Cournot) and fixed entry costs, vertical
integration is preferable with linear demand, while a unit-elastic demand function favours
vertical separation. With regard to the natural gas sector fixed entry costs might be less
crucial due to relatively low start-up requirements, while potential gains from increased
competition might be substantial given the starting point of highly concentrated wholesale
and retail markets. According to Vickers (1995), this points into the direction of vertical
separation. On the other hand, natural gas wholesale and retail activities incorporate quite
considerable portfolio effects (see e.g. ERAA 2004, EFET 2005, Coquet 2007 and Glachant
2011). Therefore, companies with a large customer base potentially have a strategic
advantage over small firms. High economies of scale would favour vertical integration.
Höffler and Kranz (2011) extend the analysis and further distinguish between legal and
ownership unbundling. Moreover, the authors consider non-tariff discrimination such as
discriminatory information flows, undue delays, overly complex contractual requirements and
the like. Assuming imperfect downstream competition, legal unbundling leads to higher
output levels and thus higher social welfare than ownership unbundling or vertical integration.
This is mainly due to the fact that under legal unbundling the incumbent calculates with the
true input costs and not – as under complete separation – with the higher network access
charge. Therefore, the incumbent is willing to expand output through a more aggressive
pricing strategy in the downstream segment with setting lower prices than under ownership
unbundling. The lower the degree of competition in the adjacent markets, the higher the
potential for the incumbent to influence downstream prices. Thus and contrary to Vickers
(1995), highly concentrated wholesale and retail markets in this context favour legal
unbundling over vertical separation. The results also hold under two-part tariffs.
With regard to dynamic efficiency, Buehler et al. (2004) find that incentives for investments
are higher under vertical integration than under separation. Pipeline investments are usually
highly asset-specific, which means that they are irreversible and imply high sunk costs.
Furthermore, network investments are often associated with positive externalities in the
adjacent segments, e.g. through higher service quality. If a network operator cannot keep the
rents accruing from these vertical externalities, it will underinvest; like a separated
monopolist, which ignores these positive effects on profits in the down- or upstream markets.
Investments also cause secondary effects on retail prices, which may reinforce or weaken
the vertical externality. The advantage of vertical integration vanishes, if perfect competition
for the downstream market is assumed, because it leads to zero downstream profits. The
vertical externality argument is also eliminated under non-linear network charges.
The same arguments generally apply, if legal and ownership unbundling are compared
(Höffler and Kranz 2011). In the case of legal unbundling, the incumbent at least partially
internalises the vertical externalities. However, the authors find mixed welfare results
comparing various kinds of investments and vertical structures. This is due to the fact, that
Policy paper 77
less strict unbundling rules tend to lead to higher levels of investments but not necessarily to
higher levels of output, which determine the net welfare effect in their setting.
Pollitt (2008) provides further arguments in favour and against strict unbundling rules, which
are mostly related to transaction costs and scope economies. For example, ownership
unbundling may reduce transaction costs by facilitating the creation of more efficient price
signals. On the other hand, costs accruing from higher coordination and contract
renegotiation requirements may be increased.
To sum up, economic theory gives little guidance regarding the choice of a specific
unbundling regime for the European gas sector as ambiguous results are reported. Given
highly concentrated gas markets, one may expect high welfare gains due to increased
competition caused by the implementation of ownership unbundling. At the same time, the
current situation may favour less strict unbundling rules, because the incumbent can
potentially better influence output levels, which may increase social welfare. Furthermore,
investment activities may be fostered due to the vertical externality effect.
The picture of mixed results does not change if empirical evidence is considered. Growitsch
and Stronzik (2011) provide an overview of existing literature, which empirically analyses the
effects of energy market liberalisation with only a few studies explicitly looking at vertical
separation. While, for example, Steiner (2001) detects a negative effect of network
unbundling on electricity end-user prices, Hattori and Tsutsui (2004) identify the opposite.
Concerning gas, neither Copenhagen Economics (2005) nor Brau et al. (2010) find any
significant influence of network unbundling on retail prices. Growitsch and Stronzik (2011)
are the first study that distinguishes different unbundling modes. While ownership unbundling
of TSOs has no impact on natural gas end-user prices, the more modest legal unbundling
reduces them significantly.
Looking particularly at scope economies, the main argument against strict unbundling rules,
Michaels (2006) reports that 11 out of 12 reviewed papers detect benefits to vertical
integration. On the other hand, Meyer (2011) shows that coordination losses may not be as
high as expected. Based on US data for the electricity sector, he calculates a cost increase
between 2% and 5% if transmission and distribution networks are separated from the other
parts of the value chain, which would correspond to ownership unbundling. These synergy
losses are rather small compared to cost increases of more than 15% associated with the
separation of electricity generation.
Turning back to the political debate, most of the countries that have legally unbundled their
TSOs under the second Directive have expressed their preference for the ITO model,
whereas nine Member States (Denmark, Hungary, The Netherlands, Poland, Portugal,
Romania, Spain, Sweden and the UK) have already implemented ownership unbundling (see
Table 17). Though different unbundling modes will likely co-exist, the agreed compromise
seems well-suited for the European context. Neither economic theory nor empirical analysis
show a clear advantage of ownership unbundling. Compared to the provisions of the second
Directive, the more stringent functional unbundling requirements for ITO as the least strict
Policy paper 78
form of vertical separation amongst the three options should lead to TSOs acting more
independently from parent companies than before. “Light” operators, which are highly
dependent on services provided by the holding, are no longer possible. It can be seen as an
indication for the evolution of more independent TSOs, even in the absence of a mandatory
ownership unbundling, that some large gas holdings have already sold off their transmission
networks (or parts thereof). While the selling of the transmission network by the German BEB
to the Dutch TSO, Gasunie, in 2008 was a pure management decision, the two more recent
cases of the German RWE and the Italian ENI have been triggered by preliminary
investigations of the European Commission regarding a potential abuse of market power.
Both companies agreed on a pipeline sale to avoid cartel action of the European
Commission.99
Regarding the European debate on TSO unbundling, several authors mention the US
interstate pipelines as an example for ownership unbundling (see e.g. Glachant 2011: 60f.).
Without any doubt, FERC Order No. 636 can be regarded as one of the major regulations
towards a competitive gas market in the US, because it took away the market power of
existing interstate pipelines and effectively removed them from having any dominant role in
gas commodity markets (Jensen 2007 and Makholm 2007, 2011). In 1992, the Federal
Energy Regulatory Commission (FERC) directed pipeline gas marketing affiliates to transfer
title to gas sales far upstream at so-called pooling points. Downstream of these pooling
points, i.e. in the actual pipeline, all gas is owned by shippers. This means that the operator
of an interstate pipeline (or any affiliated company) no longer owns the gas transported
through his trunk line. On the other hand, affiliates of interstate pipelines companies are still
allowed to transport gas through other interstate pipelines not owned by that company.
Several gas holdings still control both, interstate pipelines as well as supply companies
(Ascari 2011). Thus strictly speaking, even the US have not implemented full ownership
unbundling in European terms. Nevertheless, US regulations have led to results, which are
very similar to those expected under ownership unbundling. In the European context, the
application of this commodities clause could constitute an additional operational constraint
for TSOs under the frame of legal unbundling and might be best classified as “pipeline-
specific ownership unbundling”. Subject to legal feasibility, this measure might be an option
for further regulatory action, if it should turn out that current unbundling requirements for
European TSOs are not sufficient.100
99 Note that the Dutch TSO Gasunie is unbundled in ownership terms and that the buyers in the other two
cases are financial institutions (the Australian investment bank Macquarie for RWE and the Italian bank Cassa Deposit e Prestiti for ENI). Furthermore, since December 2010 ENI’s share of corporate capital of transmission or distribution companies is restricted to 20% by Italian law. Before, ENI as the largest gas holding in Italy held around 52% of Snam Rete Gas, the Italian TSO (Colombera 2010). In October 2011, E.ON announced its plans to sell Open Grid Europe, E.ON’s German gas TSO. The announcement coincides with current investigations of the European Commission regarding discriminatory behavior of several incumbents, e.g. E.ON.
100 Initiated by the Holding Company Act of 1935, US interstate pipelines were separated in ownership terms from gas distribution early on, which is regarded as a crucial pre-step for the success of FERC Order No. 636 (see e.g. Makholm 2007). Contrary to the US, Europe has not established such a split of transmission and distribution networks, which might be a precondition.
Policy paper 79
Table 17: Current access conditions in EU Member States101
Country
Unbundling Wholesale Retail
TSO DSO Tariff
model
Congestion
management
Capacity
allocation Balancing
Market
opening Exemption
Austria 7(0) 20(0) Yes(14) 3 NA 4 2,3 100%
Belgium 1(0) 18(5) No 1,2 4,5,6,7,8 1 1,2,3 100%
Bulgaria 1(0) 28(0) Yes(28) 3 7 1 1,3 100%
Czech Rep. 1(0) 79(0) Yes(73) 2,3 2,6,7 3 2 100%
Denmark 1(1) 3(0) No 2 5,6,7 1,2,3 2,3 100%
Estonia 1(0) 26(0) Yes(25) 1 NA NA NA 100%
France 2(0) 25(0) Yes(22) 2 1,2,5,6,7 1,3 1,3 100%
Germany 18(1) 695(0) Yes(667) 2 6,7 1,2,3 1,2,3 100%
Greece 1(0) 3(0) No NA 5 1 2 86%
Hungary 1(1) 10(0) Yes(5) 1 1,2,5,6,7 1,2,3,4 1,2,3 48.2%
Ireland 1(0) 1(0) No 2 6 1 2 100%
Italy 3(1) 263(140) Yes(209) 2 2,6,7 3 3 100%
Lithuania 1(0) 6(0) Yes(5) 4 7 1 1 100%
Luxembourg 1(0) 4(0) Yes(4) 1 NA 1 1 100%
Netherlands 1(1) 10(8) No 2 5,7 1 2 100%
Poland 1(1) 6(0) Yes(1) 3 2,5,6,7,8 1 2,3 100%
Portugal 1(1) 11(0) Yes(7) 2 1,5 4 3 94%
Romania 1(1) 38(2) Yes(36) 3 NA 1 3 100%
Slovak Rep. 1(0) 46(0) Yes(45) 2 6,7 1 2 100%
Slovenia 1(0) 18(0) Yes(18) 4 2,6,7 4 1 100%
Spain 14(1) 22(0) No 1,2 1,2,5,6,7 1,2,3,4 2,3 100%
Sweden 2(2) 5(0) Yes(5) 2 NA 4 2 100%
UK 1(1) 18(15) No 2 4,5,6,7 1,2,4 NA 100%
Source: Own compilation based on European Commission (2011a). Notes: NA = not available Unbundling: TSO: Total number and number of ownership unbundled TSOs in brackets. Remaining TSOs are legally unbundled. DSO: Total number and number of ownership unbundled DSOs in brackets. Exemptions for small DSOs (usually less than 100,000 customers) are reported in the last column with the number of exempted DSOs in brackets. Tariff model: 1 = coupled entry-exit, 2 = de-coupled entry-exit, 3 = point-to-point, 4 = postage stamp; Congestion management: 1 = auction, 2 = pro rata, 3 = lottery, 4 = buy back, 5 = use it or lose it (UIOLI), 6 = secondary market, 7 = interruptible capacity, 8 = use it or sell it (UIOSI); Capacity allocation: 1 = first come first served (FCFS), 2 = auction, 3 = pro rata, 4 = capacity goes with the customer (Rucksack-principle); Balancing: 1 = buying on the regular market, 2 = contracting, 3 = storage.
A similar development towards more independent distribution system operators (DSOs) is
less likely to occur. Compared to transmission, rules for vertical separation of distribution
101 The following countries are omitted: Cyprus and Malta do not have a developed gas market; Finland and
Latvia have received a derogation from gas market regulations.
Policy paper 80
networks are less strict. Furthermore, the majority of countries make use of Art. 26 (4),
exempting small companies with less than 100,000 customers from the requirement to
legally separate the distribution network (see Table 17). Therefore, “light” DSOs, poorly
equipped with staff and assets and highly dependent on services provided by the holding
company, are still feasible, at least under the exemption rule.
These regulations bear the risk that distribution companies still act as vertically integrated
utilities. As these companies provide an essential input into the downstream market of gas
supply to end-users in which most of them also compete, the incentive to discriminate
against third parties through cross-subsidisation or non-tariff measures has not been
completely removed.
The less favourable unbundling rules for DSOs could be one reason for the unsatisfying level
of activity on European retail markets. Although almost all countries have fully opened gas
supply to competition, retail markets are still highly concentrated with very low levels of
supplier switching.102 In a panel data analysis of the European liberalisation process,
Growitsch and Stronzik (2011) detect no significant effect of vertical separation of distribution
networks on residential end-user prices. On average, DSO unbundling measures have
neither adversely affected retail competition nor led to major improvements.
A comprehensive margin squeeze test across Europe could be a first step to get a better
understanding of what actually causes the poor outcomes of European retail market
competition. Energy regulators may claim that this is not necessary, because network tariffs
are controlled by them. This argument seems short-sighted due to the well-known
phenomenon of information asymmetry between the regulator and the network operator,
which is usually only partly solved by regulation. In telecommunications with a longer history
of network regulation, these tests are regularly applied for similar questions in the context of
infringement procedures under Art. 82 of the European Treaty. It is checked whether a
vertically integrated firm prevents its downstream competitors from achieving an
economically viable price-cost margin (see e.g. European Commission 2005).103 In other
words, it is tested whether a certain market (area) supplied by a vertically integrated
company is contestable for other market participants. An application to the European retail
market for natural gas can help to identify the factors that are responsible for the current
situation: Is it the exemption of small distribution network operators, lax overall unbundling
rules, or other factors? The former is the case if market areas of exempted DSOs show a
significantly lower contestability than other market areas. If no significant difference at low
contestability levels is detected, unbundling requirements should be tightened. A third result
is that contestability is not a problem at all, which would point at other reasons (e.g. a low
demand elasticity). As the test implicitly assumes competitive retail and wholesale prices, it
102 The latest benchmarking report of the European Commission (2011a) shows that for almost all EU countries
market shares of the three largest companies (CR3) are above 50%, which is the threshold level in competition law indicating market dominance. With regard to switching, almost all Member States report rates below 10%. Only in Denmark (14.4%), Hungary (21.6%) and Italy (33.6%), end-users have changed suppliers more frequently.
103 In order to determine the possible margin for retail companies, the wholesale price and network charges are
deducted from the retail price. A negative margin would clearly indicate that newcomers are prevented from market entry. The market is then not contestable.
Policy paper 81
has to be applied with care. Price distortions in these adjacent segments might bias test
results.104
One example for price distortions is end-user price regulation, which is still quite common
across Europe: Regulated gas prices are applied in 16 countries for households and in 13
Member States for non-household consumers. The regulatory measure is usually targeted at
protecting certain groups of customers against burdensome energy costs. As price
regulations tend to act as a barrier to entry for new entrants and have rather an increasing
than a decreasing effect on retail prices (Copenhagen Economics 2005), the intended social
benefits are highly questionable. A financial aid by the government, granted to certain
vulnerable groups of customers, would avoid these problems, and at the same time allow for
customer protection.
Retail competition might also be affected by the accessibility and competitiveness of the
wholesale market as supply companies make direct use of it (e.g. as a point of price
reference or for portfolio balancing or risk management purposes). Whether the European
wholesale market is capable of producing competitive price signals is discussed in the next
section.
With regard to unbundling of network operators, an urgent need for changes of current
provisions is not obvious. TSO unbundling rules seem appropriate to foster sufficiently
independent pipeline operators. If this expectation turns out wrong, “pipeline-specific
ownership unbundling” might be an option for further regulatory action. This provision would
disallow TSOs to own gas transported through their trunk lines without the necessity to fully
ownership unbundle them. Whether poor retail competition is caused by the lower
unbundling requirements for DSOs is not clear yet. To get a better understanding, a
European-wide margin squeeze test is suggested.
5.3.2 Wholesale
Regarding wholesale, basically two types of contracts can be distinguished, i.e. commodity
contracts and capacity contracts. For example, if a market player wants to supply Russian
gas to a German client, it has to buy the commodity, i.e. a certain amount of gas demanded
by the customer, and to arrange the transport from Russia to the German delivery point.
Thus, in addition to the commodity the shipper has to contract transportation (e.g. pipeline)
capacity according to the contracted gas in terms of volume and time. Both commodity and
capacity can be bought short- and long-term. Following Art. 2 of regulation No. 715/2009,
long-term means contracts with a duration of more than a year and short-term otherwise.
104 The problems on European retail and wholesale markets, described in the following, do not render the
suggested margin squeeze test meaningless. Carefully interpreted, the test gives at least first insights into the contestability of European retail markets and the relation to the given DSO unbundling rules.
Policy paper 82
Therefore, short-term contracts encompass seasonal services, any kind of spot market
activity and even intraday trading.105
5.3.2.1 Commodity
One indication for a well-functioning wholesale market is the evolution of central market
places, at which natural gas is actively exchanged using standardised contracts. These so-
called hubs are usually the main place where short-term commodity trading is organised.
Typically, trades are concluded bilaterally between parties (Over-the-Counter, OTC), either
directly or by the involvement of brokers.106 A key characteristic of hubs is that it has to be
possible to move gas into and out of the market. Moreover, accessibility to storage facilities
is crucial to provide flexibility services for the management of the volume risk, which accrues
from time-varying and weather-related demand profiles.
The hub can be of either physical or virtual nature. A physical hub is directly linked to the
underlying network and corresponds to a single physical location of the network, i.e. a flange
or a place where several transmission pipelines cross each other. Virtual hubs are trading
platforms, which encompass the transmission network of a certain region or even an entire
country. Gas can be present at any point within the covered zone to be exchanged at the
virtual hub.
In order to trade gas, shippers require access to hubs. Thus, a market player who wants to
ship gas from outside to sell it to another shipper at the hub additionally needs to book a
corresponding amount of pipeline capacity. Likewise, the buyer of the commodity has to
acquire pipeline capacity to transport the gas from the hub to the final destination.
Consequently, commodity trades are generally accompanied by capacity contracts.107 While
a physical hub requires access to the single location, for virtual hubs access to any border
point of the overall market zone is sufficient. On the other hand, as gas is freely allocated
within the trading region of a virtual hub, flows become less predictable. Hence, larger zones
covered by a virtual hub tend to lead to lower predictability. Therefore, a higher flexibility at
virtual hubs and lower barriers to access for shippers come at the cost of higher balancing
requirements for TSOs and a higher risk of congestion (or lower amounts of marketable firm
capacity) for shippers.
Generally, neither of the two approaches is clearly preferable. During the on-going process of
opening up the European natural gas markets for competition a few trading places have
been established, which are mostly organised as virtual trading points (see Table 18). On the
other hand, almost all market places in the US are organised as physical hubs with the Henry
105 Note that many other contract types do exist in the natural gas market (e.g. derivatives like forwards, options
and swaps). 106 If bids and offers are put anonymously, the place is called gas exchange like the German EEX and the
Dutch APX-ENDEX. Natural gas products traded at these exchanges are based on gas deliveries at a certain hub. For example, the EEX provides spot market products with possible deliveries at TTF, GPL and NCG.
107 No matter how many paper trades are created, usually one physical transaction exists that underlies these paper trades. At least, the net balance of all paper trades of a certain day has to be settled physically.
Policy paper 83
Hub as the most prominent example. While the US rely on distance-based tariffs for
transmission capacity, Europe has opted for entry-exit regimes. The European entry-exit
system is based on balancing zones covering a certain pre-defined geographical area. Thus,
its structure is similar to virtual hubs. Therefore, one reason for the different hub approaches
in the US and Europe may be the underlying tariff model for charging transportation
capacity.108 Further incidence is provided by Austria, which is about to change from a point-
to-point model (i.e. distance-based) at the transmission level to an entry-exit approach.
Contemporaneously, CEGH will be transformed from a physical into a virtual hub.
Nevertheless, each hub approach generally works with both tariff models for pipeline
capacity. Zeebrugge may serve as an example, which is not expected to be converted into a
virtual hub, although Belgium relies on entry-exit.
Table 18: European gas hubs vs. Henry Hub
NBP1)
ZEE2)
TTF3)
PSV4)
PEG5)
GPL6)
CEGH7)
NCG8)
Henry
Country UK BEL NDL ITA FRA GER AUT GER US
Launch 1996 1999 2003 2003 2004 2004 2005 2007 1988
Type virtual physical virtual virtual virtual9)
virtual9)
physical virtual physical
Churn rate (ca.)
15 4.5 4 NA 1.2 2.2 3.0 3.5 350
Trading activity/
liquidity
good slightly - decreasing
stable hardly existent
very poor
poor poor increasing
nearly perfect
Source: Own compilation based on websites of ICIS Heren and of hub operators and quarterly reports on the European gas markets (European Commission, DG Energy).
Notes: NA = not available Hub abbreviations:
1) NBP = National Balancing Point,
2) ZEE = Zeebrugge,
3) TTF = Title Transfer Facility,
4) PSV
= Punto di Scambio Virtuale, 5)
PEG = Points d’Echange de Gaz, 6)
GPL = Gaspool (formerly BEB), 7)
CEGH = Central European Gas Hub (formerly Baumgarten),
8) NCG = NetConnect Germany (formerly E.ON Gas Trading,
EGT). 9)
rather used as a physical balancing platform than for trading.
The various locations across Europe have achieved different stages of maturity and liquidity.
The National Balancing Point (NBP) in the UK is not only the first hub that has been
established, it also outperforms the other European hubs concerning trading activity and
liquidity. The churn rate at NBP, measuring the ratio between traded volumes and the
amount of gas physically delivered to customers in the region covered by the hub, is around
three to five times higher than at any other European hub like Zeebrugge, the Dutch Title
Transfer Facility (TTF) and NetConnect Germany (NCG). While Zeebrugge is slightly losing
importance, NCG is expected to catch up with TTF in some time.109 The explanations so far
imply that only four EU Member States have established market places showing at least
some modest short-term trading activity, which hints at a severe lack of market liquidity
108 Transportation capacity contracts are discussed in Section 5.3.2.2. 109 In 2006, when Germany decided to implement an entry-exit regime, the country was divided into 19 separate
balancing zones. Meanwhile, most of these zones have merged into NCG. It will be interesting to see if the French PEG develops in a similar way. Very recently, the formerly three French market zones North, West and East merged.
Policy paper 84
across the EU-27. The most liquid European hub, NBP, has a churn rate that is still an order
of magnitude lower than at the Henry Hub in Texas, the maturest gas market worldwide.
Instead of making use of short-term trading, European natural gas supplies are
predominantly delivered under long-term commodity contracts. In 2009, around 75% of
pipeline supplies were contracted long term (Melling 2010). Only the UK (80%) and the
Netherlands (40%) offered larger parts of their gas production to spot markets. Supplies from
third countries, that accounted for around 60% of EU pipeline supplies, were hardly ever
channelled via hubs. For example, only 2% of Russian gas production was contracted at
short notice.
Long-term contracts link sellers and buyers for a long period into a bilateral monopoly,
generally for 15 to 25 years, during which both of them have strictly defined obligations. The
reasoning for the existence of these contracts is based on transaction cost economics
(Williamson 1979). When a transaction entails one party committing capital that has little
value for other uses (like natural gas wells), the other party has a strong incentive to
appropriate rents arising from the bilateral relationship through opportunistic behaviour. If full
vertical integration is not feasible, long-term contracts provide a device to avoid this hold-up
problem.
Essential characteristics of current long-term gas contracts in Europe are take-or-pay
provisions and a widely used price indexation to oil. The take-or-pay contracts require buyers
to pay for a pre-specified minimum quantity of gas irrespectively of whether the gas is
actually taken. In return, the producer guarantees the delivery. The oil indexation is aimed at
protecting the buyer of the contract against prices above those levels for the main competing
fuel. This implies a risk sharing with the buyer bearing the volume risk and the producer the
price risk. In recent years, contracts have been adjusted in a way that they increasingly allow
for flexibility (Hedge and Fjeldstad 2010). For example, take-or-pay clauses have been
augmented by swing provisions, which means that the buyer has the right to nominate the
daily off-take within a specified range. This allowed range has been considerably broadened
over the last years, which enables the buyer to better react to market conditions by taking
higher volumes when it is most favourable and vice versa. Furthermore, arrangements
usually allow for the renegotiation of pricing terms every three to five years and if market
conditions have changed significantly. In this regard, the adequacy of the predominant oil-
indexed pricing is currently discussed between some large EU wholesalers, like E.ON and
GDF, and gas exporters from third countries. The buyers are urging producers to link prices
to gas spot prices instead of insisting on long-term commodity contracts that shadow the
fluctuations of oil. The discussions are mainly driven by the current market conditions. Since
2009, Europe’s gas market has been experiencing a time of oversupply with spot prices well
below parity with oil product prices. The situation is expected to persist until 2015 (Eurogas
2010).
From an economic point of view, there is no clear-cut answer to the problem of pricing terms,
since valid arguments exist for both positions (cf. e.g. Stern 2007a, 2009). For example, one
argument in favour of oil indexation is a relatively lower price volatility of the underlying,
Policy paper 85
which reduces the seller’s risk stimulating investments into gas wells. In turn, supply security
may be improved.110 On the other hand, it can reasonably be questioned, whether the
original rationale for oil-linked prices still persists, which assumes oil and gas being close
substitutes. In particular, the environmental agenda of the EU has resulted in a move away
from oil towards gas-fuelled equipment, which contemporaneously has reduced the number
of installations capable of switching fuels. Actual market information for long-term commodity
contracts reveals a positive correlation between the maturity of spot markets and the usage
of gas price indexation (Melling 2010): Over 40% of long-term contracts are indexed to gas in
the UK, 5% in Western Europe and hardly any contract in Eastern Europe.
Also controversially debated is how long-term contracts affect competition. According to Allaz
and Vila (1993) long-term contracts tend to favour competition by inducing producers to
behave more aggressively in the spot market. This leads to lower spot market prices, which
are anticipated by market players motivating them to obtain a leadership position by selling
long-term contracts before going on the spot market. Consequently, overall competition is
increased. On the other hand, Liski and Montero (2006) find that in an infinitely repeated
oligopoly the introduction of long-term contracts allows companies to sustain collusive profits
that otherwise would not be possible to achieve. The result is mainly due to the repeated
game structure, which allows for effective punishment if players deviate from collusion.
However, both lines of argument may not be well suited in the current context, because they
assume the existence of liquid spot markets and ask, how competition is affected if contract
markets are introduced. The question for the European gas market is rather the opposite:
How can spot markets develop with contract markets already in place? More crucial in the
European context seems that long-term contracts may be used by EU incumbents as a
barrier to entry on the selling side (cf. e.g. Creti and Villeneuve 2005). Since almost all
existing long-term commodity contracts are hold by major European gas companies like
E.ON and ENI and short-term liquidity is low, it may be difficult for entrants to obtain sufficient
amounts of gas to serve their clients. Furthermore, gas exporters from third countries may be
concerned about the counter-party risk of contracting with commercial entities that are
usually small and thus might go bankrupt more easily during contract duration. This might
reduce the willingness of gas producers to sign contracts with market entrants. On the other
hand, new players increase the value of the well owner’s alternative sales possibilities, which
lowers the necessity to lock in long-term contractual relationships, because the risk of
opportunistic behaviour is reduced (Crocker and Masten 1985).
The last hypothesis is backed by empirical analysis. Based on a sample of 311 long-term
commodity contracts, Hirschhausen and Neumann (2008) find that contracts tend to be
shorter in natural gas markets that are moving towards a more competitive institutional
setting. Moreover, contract durations are positively related to the asset specificity of the
underlying investment. Unfortunately, the authors neither distinguish different parameters of
the institutional framework nor account for causalities within the exogenous variables. The
explanatory variables covering the institutional framework are country dummies for the UK
and the US and a time dummy for the EU assuming more competitive framework conditions
110 Supply security is discussed in more detail in Section 5.4.
Policy paper 86
after 1998 (i.e. the first gas Directive). Thus, certain regulatory measures leading to shorter
contract durations cannot be identified. The second drawback is related to the fact that a
higher value of alternative sales possibilities also means a lower asset specificity. If the
original transaction is terminated by the holder of the long-term contract, the producer can try
to attract other buyers to sign a new contract. The larger the corresponding market and the
higher the number of potential contractors, the lower the risk of contract conditions being less
favourable than the ones of the terminated contract, and thus the lower the asset specificity
(cf. e.g. Williamson 1987).111 Therefore, Hirschhausen and Neumann (2008) use
explanatory variables, of which the institutional framework variables may not only have an
impact on the endogenous variable, i.e. contract duration, but also on the other exogenously
modelled variable, i.e. asset specificity. This may cause problems of statistical interference
(e.g. an inflated t-statistic).
These asset specificity relationships not only raise statistical concerns, they also leave
regulators in a dilemma concerning the question of long-term commodity contracts and
competition as pointed out by Brunekreeft and Guliyev (2009). On the one hand, liquid spot
markets as a representative of a competitive market environment reduce the necessity for
long-term contracts and thus result in a reduced risk of market foreclosure. On the other
hand, existing contracts may be used by incumbents to deter new players from entering the
market, which prevents spot markets from developing and contractors from agreeing on
shorter durations.
Generally, the EU regards interventions in the definition of contractual arrangements
unnecessary and favours market-based solutions, which seems justified, because long-term
contracts have proved to be a viable tool to overcome the moral hazard problem inherent in
gas production. Furthermore, contract designs are continuously adjusted by (re)negotiations
between contracting parties hinting at the ability of market players to react to changing
market conditions. The main problem is likely associated with the resale of long-term
contracted gas. In this regard, one of the few EU actions was the abandoning of the so-called
destination clause, which restricts the possibilities of buyers to resell gas outside their
respective territories. The EU has clearly stated that the clause is against European
Competition Law and negotiated its elimination with contracting parties on a case-by-case
basis (Talus 2011). National measures against a potential market foreclosure due to long-
term commodity contracts included cartel action like in Germany or gas release programmes
like in the UK, Spain and Italy. In 2006, the German competition authority restricted the
maximum duration of gas resale contracts of major incumbents with their clients to two or
four years. The more dependent a customer was on the incumbent’s supplies, the shorter the
allowed contract duration. In 2010, the restriction was lifted, since major revisions of network
111 Williamson (1987: 55) distinguishes four types of asset specificity, i.e. site specificity, physical asset
specificity, human asset specificity and dedicated assets. The mentioned relationship corresponds to the type of dedicated assets, which refers to investments dedicated to a certain trading partner that would otherwise not be made. These investments are not redeployable due to a limited size of the market. Thus, the salvage value outside of the exchange relationship (i.e. the asset value in its next best use to another user) is increased if more potential contractors exist. Site specificity pertains to the decision of the buyer and the seller to locate their operations within physical proximity of each other. Physical asset specificity regards investments in equipment with lower value in alternative uses due to certain design characteristics. Human asset specificity addresses the specialisation of skills which arises from learning-by-doing.
Policy paper 87
access conditions had been decided in the meanwhile (Bundeskartellamt 2010). Gas release
programmes, instead, require buyers of long-term contracts to sell off parts of their
contracted supplies. Since incumbents are forced to provide some of their contracted gas
directly to the market, the approach appears more appropriate to address the liquidity
problem of the European commodity market and might be more promising to attract new
market players. A boost to short-term liquidity can help to solve the mentioned dilemma of
European regulators, at least partly. Therefore, regulatory authorities may consider gas
release programmes to a far greater extent than before. The reason, why the UK has been
successful, whereas the programmes of Spain and Italy more or less failed, should be
studied in more detail and has to be left to further research. Pipeline access has possibly
played a significant role. While network access conditions are regarded as sufficient in the
UK, Spain and Italy are clearly laggards in this regard. In order to supply gas to a certain
customer, the supplier requires access to both, the commodity and the capacity.
The US experience suggests that a system with liquid spot markets may significantly reduce
the overall importance of long-term commodity contracts. In the US, most gas is now traded
through predictable and reliable physical hubs, which constitute well established market
places (Makholm 2006, 2007). However, one has to be careful, when transferring US
experience to Europe, especially with regard to commodity contracts. While US gas is
supplied by a large number of small domestic producers, the EU relies to a large (and
increasing) extent on gas imports, which are supplied by big gas holdings from third
countries. Therefore, long-term contracts might play a greater role in the European context,
once liquid spot markets will have emerged. Since players from two different countries with
distinct legal systems are involved, there is possibly a greater necessity for contractual
arrangements. Furthermore, foreign suppliers might still have limited access to the European
hubs.
5.3.2.2 Pipeline capacity
An internal European gas market requires sufficient access to pipelines to enable market
participants to ship the commodity to locations, where it creates the highest possible welfare.
According to Art. 14 (1c) of regulation No. 715/2009, TSOs shall offer both long- and short-
term capacity products to network users.112 Basically, shippers have two opportunities to
acquire these products, the primary market and the secondary market. Primary capacity is
directly contracted between the TSO and the network user, whereas secondary capacity is
traded between shippers. Since capacity is contracted (booked) in advance, shippers have to
inform the TSO, if and how they will actually use it (i.e. actually transport gas) on a certain
day (nomination). Thus, already booked but not nominated capacity is traded on secondary
markets.113
112 The definition is the same as for commodity contracts. 113 This does not necessarily mean that only short-term capacity is traded in secondary markets. If shippers
contracted capacity long term on expectations that turn out wrong, they can also offer their long-term contracts (or parts thereof) to the secondary market.
Policy paper 88
Based on a regular market monitoring, ERGEG has identified several obstacles, capacity
markets are still suffering from (ERGEG 2010a, European Commission 2011b). Most of
these weak points can be subsumed under three broader categories: First, looking at the
access conditions at the wholesale level reported in Table 17, a great variety of approaches
becomes obvious that Member States have implemented with regard to the overall tariff
model, congestion management procedures, capacity allocation mechanisms and balancing
systems. The lack of harmonisation of market rules and operational procedures in the
different countries leads to market segmentation and higher transaction costs, which
constitutes a barrier to entry, especially for small market participants. Second, the diverse
structure of existing rules also reduces market transparency. For cross-border trades e.g.,
the information provided on both sides often do not match.114 Furthermore, decisive
information for shippers is sometimes missing completely, especially regarding actual
network utilisation.115 The third major problem is insufficient interconnection capacity, which
can be either caused by physical constraints or by contractual congestion. The latter means
that the pipeline is fully booked, but a network user does not offer its excess capacity to the
market, although other shippers have expressed their demand. Whether the capacity is in
fact strategically withheld (i.e. excess capacity) or unused due to valid reasons (e.g.
uncertain and highly weather-related demand profiles of clients that have to be supplied
without interruptions) is often hard to detect (cf. e.g. Kwoka and Sabodash 2011).
Limited installed interconnection infrastructure is of special concern for the Iberian
Peninsulas, Finland and the Baltic Republics, with the latter two still not connected to the rest
of Europe and thus totally dependent on Russian supplies. Central and Western Europe are
affected by contractual congestion rather than physical constraints (ERGEG 2011a). First
come first served (FCFS) is still the dominant capacity allocation criterion in the primary
market throughout Europe. To back their long-term commodity contracts, incumbents usually
also purchase pipeline capacity long term. Since the booking occurs usually at the time when
the pipeline is built, FCFS privileges access by incumbents’ long-term contracts. Moreover,
instruments incentivising anti-hoarding behaviour of capacity holders are hardly
implemented, which hampers market entry, since booked but unneeded capacity is not
offered to other parties. Consequently, secondary markets for pipeline capacity, if at all
implemented, severely lack liquidity. Furthermore, the absence of market-based procedures
(e.g. auctioning of primary capacity) leads to an inefficient utilisation of existing network
capacity, because relevant price information is not disclosed to the market. Adequate price
signals for network scarcities are hardly existent. This prevents markets from adjusting
smoothly to new information and leaves arbitrage potentials unexploited as shown by
Growitsch et al. (2010). Looking at price relations between the Netherlands and Germany,
114 A monitoring of 21 major interconnection points has revealed that less than 25% of the selected points have
harmonised capacity allocation mechanisms and congestion management procedures in place (ERGEG 2011a). The same holds for balancing procedures. Based on an EU-wide survey, KEMA (2009) has identified several obstacles for cross-border trades, e.g. a lack of market-based mechanisms, incompatible products as well as the use of different balancing periods and tolerances on both sides of the interconnection point.
115 An interesting example is associated with storage. Over the last years, storage operators changed several
times the methodology for publication of inventory data. The resulting breaks in time series data make it rather impossible to build up expectations how the market reacts to new information (see Rammerstorfer and Stronzik 2011).
Policy paper 89
the authors additionally find empirical evidence for capacity constraints even for these well-
integrated markets, which are likely due to contractual congestion.
The efforts of the various European institutions touched upon at the end of Section 5.2 are
mainly targeted at these drawbacks. Whether they will actually lead to significant
improvements depends on the final rules, which are yet to be determined. In the following, a
few points are highlighted that are of particular importance for a well-functioning European
gas market.
With regulation No. 715/2009, major steps have already been taken. Besides the
establishment of guiding principles for TSOs as well as for storage and LNG-terminal
operators regarding capacity allocation mechanisms and congestion management
procedures, Art. 13 of EC 715/2009 requires all Member States to switch to a de-coupled
entry-exit regime and prohibits any form of contract-path dependent gas pricing. De-coupled
means that entry and exit capacities can be independently contracted, combined and used at
any point in the network. There is no linked contract path between individual points and
capacities are freely assignable within the region covered by the network.116 The possibility
of an independent reservation of entry and exit capacity with the concurrent abolition of the
formerly existent distinction between transit and domestic transmission will reduce entry
barriers as the need for over-reservations of bundled entry and exit capacities for gas transits
is lowered. A further improvement is that EC 715/2009 has been set into force by comitology,
meaning that the regulation has become directly effective without the necessity of a long-
lasting process by Member States transposing it into national law. Therefore, different tariff
models are no longer a question of different national regulations; they have become a
question of compliance with existing regulations. After all countries are in compliance, the
same tariff model will be used across Europe.
With shrinking indigenous production and increasing demand Europe will become
increasingly dependent on gas supplies from outside the EU. As major consuming regions
are often far away from the corresponding import points, gas transports over long distances
crossing several borders will gain importance. Under the current framework, for each market
zone that gas is transported through, shippers have to book corresponding entry and exit
capacities. Bundled capacity products, shifting gas from one trading region to another market
zone via a single contract, will simplify booking procedures, reduce transaction costs, and
hence facilitate market entry.117 Moreover, bundled products designed as hub-to-hub
capacities will also result in a better standing of hubs, thus fostering liquidity at these market
places.
116 By contrast, the entry-exit system is termed coupled, if some exit or entry points are subject to locational
restrictions for individual connections (e.g. entry point A can only be contracted together with exit point D). The more restrictions are imposed, the more the entry-exit regime turns into a point-to-point model. For further details see e.g. KEMA (2009).
117 A kind of a bundled product has been implemented recently by the Dutch TSO and the German NCG market
zone allowing for gas quality overlapping balancing. In both areas, two gas qualities co-exist, the high caloric H-gas and the low caloric L-gas. The quality overlapping balancing actually leads to a close link of formerly separated balancing zones, thus simplifying booking procedures.
Policy paper 90
Bandulet et al. (2010) show a high potential for significant improvements concerning
competitive framework conditions at the wholesale level associated with the harmonisation of
capacity allocation criteria. Market-based allocation of primary capacity rights via auctioning
is generally better suited to provide for a level-playing field between incumbents and
newcomers than the currently favoured FCFS principle. Auctioning off all available capacity
at entry and exit points may collide with needs for long-term capacity. Shippers that have
contracted gas long term have a reasonable interest in long-term capacity reservations to
secure their gas supplies. If the availability of long-term pipeline capacity is uncertain,
network users might be dis-incentivised to lock in long-term commodity contracts as well.
Furthermore, pipelines are also characterised by a high asset specificity, which calls for long-
term capacity contracts (cf. e.g. Hirschhausen and Neumann 2008). A regulation that
requires TSOs to reserve the whole capacity for mid- or short-term auctions would neglect
the asset specificity of pipelines and gas wells, which may put both kinds of gas
infrastructure projects at risk.118 Therefore, only a certain part of overall capacity should be
reserved for mid- and short-term capacity auctions. The actual percentage is open for
debate.119
At least of the same importance is a European-wide implementation of anti-hoarding
incentives like the use it or lose it (UIOLI) or the use it or sell it (UIOSI)) principle, which
should be made mandatory for all capacity holders. The basic idea is that capacity, which is
not nominated by the capacity holder, is subsequently offered to the secondary market
through auctions or an OTC bulletin board. UIOSI forces the capacity holder directly to offer
unused capacity rights to the market, whereas UIOLI has to be accompanied by a
corresponding obligation for the TSO. Under a pure UIOLI not nominated capacity just drops
back to the TSO; it is not assured that the capacity is subsequently marketed to the
secondary market. Both measures are capable of significantly reducing the risk of market
foreclosure. Moreover, the introduction of UIOLI/UIOSI will probably boost liquidity on
secondary markets, thus initiating a self-reinforcing process. Turning currently hardly existent
secondary markets into well-functioning trading places with high liquidity, thus providing
reliable and competitive price signals, will disclose opportunity costs of unused capacity to
the market. This further reduces hoarding incentives. While auctioning of primary capacity
might be restricted by existing contracts, the application of UIOLI/UIOSI should not be
associated with major legal problems.
Another regulatory measure that has not yet been considered is market coupling through the
use of implicit auctions. Market coupling is already partly implemented in the electricity
sector, which has improved market integration. However, market coupling requires a
functioning short-term trading (day ahead) with harmonised rules in the corresponding
market areas. Hence, it might be an option, once further progress has been made regarding
gas wholesale.
118 For further discussions see also Section 5.4.2. 119 Germany, e.g., has set a limit of 65% for capacity sold long-term, meaning that at least 35% are reserved for
mid- or short-term capacity needs.
Policy paper 91
A remarkable difference to the US is how market-related information is managed and
provided to the different players in Europe in order to achieve a sufficient level of
transparency. According to Art. 16 (1) of the third gas Directive, the TSO shall limit the
amount of information made public if a publication would risk harming legitimate commercial
interests of supply companies shipping gas on the pipelines concerned. Especially market
entrants complain about the level of information actually provided. Among other things,
entrants demand the publication of the identity of primary and secondary capacity holders,
the amounts of unused capacity and detailed forecasts of available capacity (see e.g.
European Commission 2007). This indicates that confidentiality likely outweighs the need for
the market to be fully informed. FERC has chosen the opposite position. FERC Order No.
637 from 2000 requires a comprehensive and immediate provision of all transactional
information on quantities and prices as well as on the identities of the involved shippers.
Users of the regulated network have no right to secrecy (Olsen 2005). In Europe, the
disclosure of information is clearly driven by regulation and not by the market. Up to now,
TSOs have just moved as far as regulations required them to go, sometimes even not that
far. Due to provisions like Art. 16 (1) of 2009/73/EC and the lack of voluntary action of TSOs,
regulators have to specify each tiny piece of information that TSOs should provide to the
market. If it turns out that the current efforts will not sufficiently improve transparency, a
fundamental shift in information policy requirements might be advisable at a later stage.120
Summing up, the various EU regulations and proposed framework guidelines indicate that
most of the described crucial issues are addressed, at least to a certain extent. Therefore,
the overall regulatory process points into the right direction and can form the basis for
substantial improvements of the European gas market. Whether the measures are sufficient
to lead to a liquid market is an open question and depends on the actual design and
implementation of the various provisions like e.g. the percentage for short- and medium-term
capacity auctions and the actual auction design. Furthermore, subtle elements, such as the
rules for nomination procedures, the gas day used for balancing and the actual definition of
capacity products, may also play a role. For example, poorly designed congestion
management procedures may unnecessarily increase network charges or lead to lower
amounts of available firm capacity (e.g. if the TSO relies too much on interruptible capacity),
which would lead to market imperfections. However, a discussion of these issues is beyond
the scope of this paper.
One of the most crucial success factors seems to be, how the EU will be able to balance the
conflicting interests of short- and long-term demand, both for capacity and commodity.
Liquidity on commodity spot markets and secondary capacity markets are likely key to
sufficiently reduce potential market foreclosure incentives of long-term contract holders. In
120 The chosen design concerning the disclosure of information assumes that market needs can be mapped by
regulation. Comparing the dynamics of markets to the speed of political processes, this assumption is at least debatable. In this context, ERGEG carried out a public consultation process on transparency requirements (ERGEG 2010b) trying to collect market views. One outcome of this stakeholder consultation is the Regulation on Energy Market Integrity and Transparency (REMIT), which has been adopted by the European Parliament in September 2011 and is entering into force in 2012. REMIT contains prohibitions of market manipulation and insider trading on energy wholesale markets (exchange- and OTC-based). It gives ACER an important role with respect to market monitoring and data collection. However, enforcement of the provisions remains with Member States.
Policy paper 92
order to overcome the regulatory dilemma of creating liquidity, buyers of long-term contracts
should be forced by regulatory action to offer parts of their contracted volumes to the market.
Gas release programmes for the commodity market and the establishment of UIOLI or UIOSI
for capacity are promising approaches. Once, liquid markets will have developed, buyers will
face opportunity costs that should induce a more competitive market behaviour.
5.3.3 Storage
Demand for natural gas is characterised by a strong seasonal pattern, because a large part
of gas is used for heating purposes. Heating demand is low during summer and high during
winter. Storage facilities provide market participants with the corresponding intertemporal
flexibility. If prices are low and market participants expect them to rise, gas will be bought at
the spot market and put into storage. Later, when prices are high, gas will be withdrawn from
storage and sold at the spot market. Therefore, gas is usually injected into storage during
summer and withdrawn during winter.
In order to make use of storage, shippers have to book corresponding amounts of storage
capacity. Access conditions of the European storage market121 are distinct from pipeline
regulations.122 The third gas Directive calls for a less stringent legal unbundling, allowing for
“light” storage operators, and leaves Member States the choice to opt for negotiated (nTPA)
or regulated third party access (rTPA). For small storage operators, Art. 33 of the Directive
grants exemption from access regulations. Contrary to DSOs, no specific threshold size is
determined. Furthermore, major new facilities can apply for exemptions under Art. 36, if they
successfully demonstrate that the site operates in a competitive environment, i.e. if sufficient
substitutes are available for potential storage users (e.g. other storage facilities, interruptible
gas to power generation facilities, non-congested interconnection capacity, flexibility in import
contracts and indigenous production swing). Decisions about exemptions are left to Member
States.
Seven countries have opted for rTPA, while nTPA has been chosen by eight Member States.
The Netherlands and UK decide on a case-by-case basis whether rTPA or only nTPA is
applied. According to GSE (2010), around two thirds of the EU-wide installed working gas
capacity is operated under an nTPA regime. Major obstacles are highly concentrated storage
markets, which are dominated by incumbents, and a long-term booking of storage capacity
leading to insufficient access conditions for market participants other than incumbents (cf.
e.g. RAMBOLL 2008 and ERGEG 2011c).123 These drawbacks are confirmed by Stronzik et
121 Generally, storage is an integral part of the European gas market. Whether storage is regarded as a market
in its own or as part of the wholesale market is a matter of definition of the various market segments. Since the previous two sections have focused on pipeline-bound gas and storage is different from pipelines in terms of cost structures and regulations, it is dealt separately.
122 Regulations for LNG, the other major flexibility source and linked to the wholesale market, are rather similar
to storage. While the US classify LNG terminals as competitive infrastructures, the EU regards LNG as essential facilities (ERGEG 2011b). Thus, most of the following explanations for storage apply to LNG as well. One of the main differences to storage is that the Directive calls for an rTPA regime.
123 RAMBOLL (2008) calculates for all relevant Member States (i.e. with existing storage facilities) a Herfindahl
Index well above 1,800, which is the threshold level applied by US antitrust law pointing at a very high concentration.
Policy paper 93
al. (2009), who find strong empirical evidence for a deviation of actual market outcomes from
the competitive benchmark. Even for markets with at least some modest trading activity (i.e.
NBP, TTF and Zeebrugge), the authors detect a high arbitrage potential that remains
unexploited, which hints at market imperfections. Storage facilities are still operated in a
more technical sense to cope with seasonal demand fluctuations and are not service-
oriented (e.g. virtual storage).
Although gas storage has much lower economies of scale than pipelines, access regulation
seems justified under current conditions. Storage operators often hold a de facto monopoly,
and entry barriers are rather high (cf. e.g. Creti 2009 and KEMA 2010). Major underground
storage facilities are restricted to certain areas where adequate geological conditions exist.
Furthermore, sites can often be reached by only one single (congested) network.
However, it is not all clear if the distinction between rTPA and nTPA is reasonable. It may be
argued that storage can operate in either a competitive or a non-competitive market
environment. If market conditions are deemed competitive, the facility should be exempted
from any access regulation. Therefore, only one regulatory approach would be required.
Since nTPA has shown to fail, like in Germany under the first Directive, rTPA should be
applied. This line of argument assumes that a clear distinction can be made between the two
possible states of the market environment. EU experience suggests that this is often difficult
to be shown in practice, because many soft parameters, like e.g. the definition of the relevant
market and assumptions about demand elasticity and available substitutes, significantly
affect the final decision (European Commission 2005). Furthermore, the argument draws on
an example for nTPA that might be misleading, since Germany implemented the industrial
self-regulation with a weak regulatory threat.124 Brunekreeft (2004) shows that nTPA
combined with a strong regulatory threat, by means of ex-post antitrust intervention or the
credible proclamation by the legislator to introduce high-powered regulations in case of
companies’ misbehaviour, can work effectively inducing a voluntary price cap. Therefore, the
question if the approach of the third Directive allowing for rTPA and nTPA could be justified
remains open.
Although the decision on the exemption from regulated tariffs may be difficult for a single
case, at least the same rules and criteria should be applied across Europe. Currently,
regulations fall short of a clear guidance for national regulators on how to assess market
power. This leads to the application of different criteria in different Member States, which are
often not comparable (ERGEG 2009). Thus, ACER should develop a corresponding
guideline in order to improve transparency and achieve harmonised decisions.125
Furthermore, it is currently not obvious how major existing storage facilities can be exempted
from TPA regulations at all. Market conditions might have changed substantially since the
first decision on the applicable TPA regime. Thus, an existing facility may now operate in a
perfectly competitive environment. The third Directive (as well as under its predecessors) is
124 For further discussions of the German case see e.g. Brunekreeft (2001) and Growitsch and Wein (2005). 125 In the US, storage is also basically treated as an essential facility with the possibility for exemptions. In 2006,
FERC issued Order No. 678 establishing clear and predetermined criteria for obtaining market-based rates for storage services.
Policy paper 94
missing a corresponding provision. Extending Art. 36 to existing sites should overcome this
drawback.126
Transparency, congestion management, and capacity allocation requirements are tackled by
the Guidelines of Good TPA Practice for Storage System Operators (GGPSSO).127 The
current non-binding status should be transposed into a binding one. Finally, the introduction
of bundled products might be advisable, at least for an intermediate period of time until
functioning secondary markets will have evolved. Usually, a storage user does not only
require access to storage. In order to transport gas to and from the site, the shipper
additionally seeks for access to the network. Thus, congestion in the adjacent transmission
network might additionally block access to storage. Bundled products, linking storage to
pipeline capacity, could help to overcome this obstacle.
5.4 Security of supply
5.4.1 The EU concern
Currently, around 60% of European gas demand is covered by imports (IEA 2010). In the
IEA reference scenario for the EU, demand is projected to increase by 0.7% p.a., while
indigenous production will decrease by 3.1% p.a. due to limited European gas reserves.
Thus, Europe’s dependency on imports will rise to around 80% to 90% by 2030. Due to this
expected increase in import dependency and sometimes difficult political circumstances in
countries exporting gas to the Europe, one of the major concerns of the EU with regard to the
gas market is supply security.
Security of supply can be defined in various ways. Stern (2002), for example, differentiates
between short-term and long-term adequacy of supply and infrastructure to transport gas to
the demand regions and between operational (e.g. extreme weather events) and strategic
security (i.e. catastrophic default of infrastructure or supply sources). De Jong et al. (2006)
define the supply security risk as a shortage in energy supply. They distinguish between
relative shortages (i.e. a mismatch in supply and demand including price increases) and
partial or complete disruptions of energy supplies. In the following, I concentrate on supply
disruptions (i.e. a physical non-delivery of gas) and the more strategic aspect of high import
prices (i.e. prices above competitive levels). Both are usually among those supply security
issues, which are regarded as particularly crucial in the European context (see e.g. European
Commission 2011b).
126 Another option could be to subsume the application for exemption of existing sites under Art. 33. The
provision calls for access regulation only when the facility is “technically and/or economically necessary for providing efficient access to the system for the supply of customers”. In an interpretive note, the Commission could specify an operational procedure and corresponding criteria according to Art. 36 requirements.
127 Accordingly for LNG: Guidelines of Good TPA Practice for Liquefied Natural Gas System Operators
(GGPLNG).
Policy paper 95
5.4.1.1 Risk of supply disruptions
Figure 11 provides an overview of the main gas flows at the European level for 2009.
Regarding EU gas imports, the Russian Federation is the largest single supplier accounting
for ca. 40%, followed by Norway with a share of around 30% of EU imports. A little more than
10% is supplied by North African pipelines via Spain and Italy. LNG amounts to nearly 20%,
mainly shipped to Spain, France and the UK. Important LNG suppliers are Qatar, Trinidad
and Tobago and Algeria. As large gas reserves are located in Russia, North Africa, the
Middle East and in the Caspian region,128 the EU expects these countries to become more
important for European gas imports in the future. Such imports, however, far from being seen
as the solution to European gas security, are almost universally seen as a the problem.
European governments regard several of these gas exporting countries as unreliable
partners with Russia being a special case that they pay close attention to. Events like the
2011 civil unrest in North Africa and the Russian supply disruptions in 2006 and 2009,
caused by conflicts on gas deliveries between Russia and Ukraine and between Russia and
Belarus, may serve as prominent examples.
Figure 11: Physical gas flows 2009
Source: DECC (2010: 90).
128 According to IEA (2010), the largest proven reserves can be found in Russia with 44.9 tcm, Iran with 29 tcm
and Qatar with 25.2 tcm. For comparison, the UK has a proven reserve of 0.3 tcm.
Policy paper 96
ERGEG (2010a) has analysed the exposure of existing networks to gas import cuts and
found out that only eight Member States have a sufficient infrastructure in place to cope with
such supply disruptions. The Baltic countries are among the most vulnerable Member States
as they rely on Russian gas deliveries to nearly 100%. Moreover, they are hardly connected
with the European gas grid and have no national storage facilities.
The poor interconnectivity of certain European regions has also become obvious through the
increased availability of cheap LNG in Western Europe since the end of 2009. The rapid
development of unconventional gas129 in North America has significantly reduced US
demand for imported LNG. As a result, more cheap LNG has been available for the
European market. While Western Europe has profited from lower gas prices, Central and
Eastern Europe only received small amounts of that additional gas supply. Consequently, the
price differential between these regions has increased significantly (European Commission
2011b).
To reduce the vulnerability to supply disruptions, the EU has initiated a process consisting of
a bunch of measures recently implemented or to be established in the near future. These
measures are targeted at diversifying import routes and enhancing existing infrastructure.
Under the Energy Infrastructure Package (EIP), the European Commission (2010) has
developed a blueprint for an integrated energy network that puts forward a method for
strategic network planning at the EU level and identifies major investments needs (so-called
priority corridors). Regarding gas, three corridors have been determined. The Southern
Corridor is aimed at bringing gas from the Caspian Basin, Central Asia and the Middle East
to the EU. The linkage of the Baltic, Black, Adriatic and Aegean Seas is the main objective of
the North-South Corridor in Central Eastern and South-East Europe. Via the North-South
Corridor in Western Europe, internal bottlenecks shall be removed and short-term
deliverability increased. To foster investments, the Commission established the European
Energy Programme for Recovery (EEPR) that co-finances selected energy projects that meet
the priorities identified under the blueprint.130 Based on the experiences with the 2009
Russian supply disruptions, the EU repealed the Security of Supply Directive from 2004 with
regulation No 994/2010. It has been designed to strengthen the prevention and crisis
response mechanisms. The regulation requires Member States to establish a risk
assessment and preventive action plan to address the risks identified. In particular, Art. 6 of
regulation No 994/2010 calls for taking into account the so called (n-1)-criterion. This criterion
describes the ability of the gas infrastructure of a certain area to satisfy total gas demand in
the event of disruptions of the single largest element of the infrastructure. Member States are
obliged to have implemented necessary measures by December 2014.
129 Unconventional natural gas resources include the categories tight gas, shale gas, coalbed methane (CBM)
and hydrates. According to IEA (2011), unconventional gas currently makes up about 60% of marketed production in the US. Though world unconventional gas resources are estimated to be as large as conventional resources (around 400 tcm), the picture of European import dependency will probably not change significantly, at least in the long run. EU’s share of world unconventional gas resources is estimated to be in the range of its share regarding conventional gas (around 4 to 5%). Only for Poland, IEA reports major efforts to develop its shale gas resources.
130 By the end of 2010, nearly the whole budget of four billion € has been committed, with 1.4 billion € allotted to
gas infrastructure projects (European Commission 2011c).
Policy paper 97
Another important part under the EIP is the responsibility of ENTSOG to regularly publish a
10-year network development plan (TYNDP) taking into account planned investment
projects. ENTSOG (2011) distinguishes between projects that have already received a final
investment decision (FID) and those with pending decisions (non-FID). Expected capacity
additions are substantial, e.g. LNG send-out capacity will be increased by more than 80%
(FID plus non-FID projects) until 2020 compared to existing levels (see Table 19). A large
share of LNG and storage projects are located in Southern Europe, especially in Spain.
Around one quarter of transmission investments are co-financed by EEPR. Roughly 40% of
projects improving interconnection between Member States (EU-IP) are dedicated to reverse
flow capacity additions, thus targeted at increasing network flexibility.
Table 19: Gas infrastructure investment projects
Transmission Storage LNG
FID Non-FID FID Non-FID FID Non-FID
Number of projects1)
40 (13) 53 (9) 26 (2) 22 (0) 11 (1) 20 (0)
Capacity additions2)
EU-IP: 12.7%
Import: + 14.6%
EU-IP: 44.5%
Import: + 27.2% +26.7% +25.5% +23.4% +62.6%
Source: Own compilation based on ENTSOG (2011). Notes: 1)
Number in brackets: projects co-financed by EEPR, 2)
2020 vs. today; transmission: interconnection capacity between Member States (EU-IP) and import capacity; storage: maximum withdraw capacity; LNG: send-out capacity.
Regarding pipeline import capacity, three major projects are under development at different
stages. Of these, only Nord Stream has received FID. The pipeline connects Siberian gas to
Europe via the Baltic Sea.131 The first line of Nord Stream has a capacity of 27 bcm and has
gone alive by the end of 2011. A second line with another 27 bcm is due one year later. The
other two, Nabucco and South Stream, are regarded as competing projects since they are
capable of connecting the EU to the gas reserves of the Caspian Basin (namely
Turkmenistan and Azerbaijan). Nabucco is projected to be half the size of South Stream
(Nabucco with 30 bcm vs. 60 bcm of South Stream). While the South Stream project is
mainly driven by Gazprom, Nabucco is backed by the EU as it circumvents Russia.132
Although both are non-FID projects, the implementation of Nabucco is usually regarded as
less likely (see e.g. Bilgin 2009 and Dieckhöner 2010). While South Stream is projected to
start 2015, the scheduled starting date of Nabucco has already been postponed from 2015 to
2017.
ENTSOG (2011) modelling results indicate that the overall gas supply potential will be
sufficient to meet increasing demand requirements. Nevertheless, three bottleneck regions
have been identified that will not fulfil the (n-1)-criterion under Art. 6 of regulation No.
994/2010, i.e. Denmark-Sweden, the Balkans and Poland-Lithuania. This paper is not the
131 The pipeline enables Europe to import Russian gas by circumventing the transit countries Ukraine and
Belarus, the Russian counterparts in the 2006 and 2009 supply disruptions. 132 On September 12, 2011, the EU has adopted a mandate to negotiate a legally binding treaty between the
EU, Azerbaijan and Turkmenistan (see European Commission 2011d).
Policy paper 98
place to explore the ENTSOG results any further or to evaluate whether the measures to
enhance supply security are sufficient,133 but two points are worth being highlighted.
European policy concerning the threat of supply disruptions is driven – at least to some
extent – by the assumption that imports are less secure than domestically produced gas
supplies. Stern (2007b) demonstrates that this assumption is not backed by empirical
evidence. Serious security incidents have also been caused by failures of indigenous
supplies or facilities.134 Thus, increased supply security cannot only be achieved via a
diversification of import routes, likewise it can be assured by a diversity of indigenous
facilities such as pipelines, LNG terminals and storages. Supply disruptions in Eastern and
South-East Europe caused by the 2006 and 2009 pipeline cuts of Russian imports would
have been less severe, if the European gas infrastructure had been more flexible and
capable of re-directing gas flows to these regions.135 Furthermore, these disruptions were
not caused by conflicts between Russia and the EU. Usually, countries exporting gas to
Member States have not an interest not to sell gas and to distort the complex diplomatic
relations with an important and reliable customer. Due to relatively high per capita income
levels, the EU’s willingness to pay is often higher compared to other world regions. For
example, Russian average sales prices for domestic sales are only 25% of the average price
that EU countries pay. In other countries of the former Soviet Union (FSU, mainly Ukraine)
prices on average amount to 77% of the EU level. Although non-EU levels have risen over
the last decade (from 12% to 25% for domestic sales and from 52% to 77% for FSU
countries), those gaps will take time to narrow further (Pirani 2011).
The second point is associated with market access conditions. Framework conditions that
facilitate transparency, remove barriers to market entry and provide for arbitrage
opportunities can help to reduce the risk of supply disruptions (see e.g. Hirschhausen et al.
2010). In a well-functioning market environment supply shocks will lead to price differentials,
which can be levelled out by arbitrage. Thus, markets can smoothly adjust to demand or
supply shocks by re-directing gas flows. Persisting price spreads between different regions
point at bottlenecks of the existing infrastructure. This incentivises investments, because the
investor can absorb scarcity rents. On the other hand, gas supply secured by abundant
transportation capacity reduces the risk of (regional) market power (cf. e.g. Cremer and
Laffont 2002 and Meran et al. 2010). In turn, a reduced risk of price distortions enhances
market transparency. Therefore, market access conditions and supply security in terms of a
reduced risk of supply disruptions are somehow interrelated and not independent of each
other.
133 Security of supply is a result of the complex interplay of future developments of supply, demand and
infrastructure conditions. These conditions are affected by numerous parameters that are accompanied by high uncertainties since long time horizons have to be taken into account.
134 The contamination of the Interconnector between the UK and Belgium in 2002 and the fire at UK’s Rough
storage facility may serve as prominent examples. Both events led to persistent supply disruptions. 135 Even within Germany, the second largest gas consuming EU country and not directly affected by supply
disruptions, this capability wasn’t assured. As shown by Growitsch et al. (2010), information efficiency between the two major German market zones, NCG and Gaspool, nearly disappeared during the 2009 supply disruptions.
Policy paper 99
As in particular discussed in Section 5.3.2.2, current access conditions of the European gas
market are far from being perfect. The overall market environment can be rather classified as
non-transparent. Consequently, the lack of market-based capacity allocation and congestion
management procedures, for example, does not only lead to an insufficient use of existing
infrastructure, it also affects price signals with regard to further investment needs. Tariffs do
usually not reflect costs. Thus, relevant information about existing scarcities in terms of
infrastructure bottlenecks is not disclosed to market participants, which likely impedes the
attraction of private capital to gas infrastructure projects. Furthermore, with incumbents
usually being better informed than newcomers, it is not surprising that current investments
are mostly carried out by incumbents. As scheduled EU measures hint at the establishment
of more market-based access rules, information asymmetry between incumbents and market
entrants may be softened in the future and more private capital attracted.
5.4.1.2 Risk of high import prices
Only a few countries supply gas to the EU. In most cases, the gas is mainly provided by a
single company like the Algerian Sonatrach and the Norwegian Statoil. The most prominent
example is Gazprom, the Russian gas giant. Gazprom is the world’s largest gas company
with the majority of shares controlled by the Russian government. In 2010, the company
owned 18% of the world’s gas reserves and 70% of Russian reserves. Furthermore,
Gazprom accounted for 78% of Russian and 15% of global gas production (Gazprom 2011).
Thus, EU gas companies face strong counterparts when importing gas. The primary
objection of the EU in this context is that import prices may be exaggerated due to
asymmetric bargaining power (cf. e.g. Brunekreeft and Guliyev 2009 and Ochssée 2010).
The high and increasing import dependency is also one of the key differences to the US. The
US relies almost entirely on indigenous production. The situation in the US will likely persist,
at least in the medium term. The US unconventional gas reserves are around twice as high
as the conventional ones (IEA 2011). Hence, limited conventional reserves are sufficiently
replenished by unconventional gas. Furthermore, Canada as a potential exporter to the US
has very similar gas market regulations (cf. e.g. Makholm 2007, 2011).136 Instead,
companies from third countries exporting gas to the EU very often face regulations in their
home market, which are distinct from the EU approach. Moreover, these companies are
usually controlled by foreign governments, which may mix economic with more strategic and
political goals. Therefore, the vulnerability to high import prices is a challenge for the
European market, which the US does not face.
A secondary concern of the EU is that foreign companies may use extra profits from gas
sales for forward integration into the European market cross-subsidising their own
downstream activities and thus distorting competition in the EU (cf. e.g. Brunekreeft and
Guliyev 2009 and Ochssée 2010). This objection is underpinned by Gazprom’s efforts to
136 For the potential Mexican gas exporter, PEMEX, any dominant position comparable to Gazprom can be
safely assumed away, since Mexico accounts only for around 5% of North American gas reserves (BP 2011).
Policy paper 100
expand its presence in all spheres of the European gas sector. Gazprom acquired assets of
gas companies in many European countries ranging from production fields in the North Sea
to retail services. Moreover, Gazprom is involved in a vast number of joint ventures and has
established strategic partnership agreements with leading European gas companies like ENI,
Gaz de France, E.ON and OMV (Victor 2008 and Gazprom 2011). The foreign activities of
Gazprom focus on the large gas consuming countries in Europe like Germany, UK, Italy,
Poland and Turkey. In 2010, roughly 60% of Gazprom’s foreign gas sales were dedicated to
these countries (Gazprom 2011).
One of the EU responses is Art. 11 of 2009/73/EC, which aims to prevent a massive sell-out
of strategic EU energy assets to companies from non-EU Member States. The rule is
particularly targeted at Gazprom and has become known under the term “Gazprom clause”
(Cottier et al. 2010). The provision sets out special certification requirements for transmission
system operators from third countries. National regulatory authorities shall notify the
Commission if non-EU nationals want to buy shares in a European gas TSO. TSOs owned
by non-EU companies are required to comply with all the same conditions as EU operators
under Art. 9, in particular unbundling. Additionally, regulatory authorities shall refuse
certification if security of supply is jeopardised.
Regarding the primary objection of excessive import prices, the establishment of an EU gas
purchasing agency is discussed (Christie 2011). Among others, the debate has been
encouraged by a joint declaration of the President of the European Parliament Jerzy Buzek
and the former President of the European Commission Jacques Delors on the need to create
a European energy community (European Parliament 2010). One element of their proposal is
to engage in coordinated energy purchasing activities on the EU level. The intention is to
strengthen the bargaining power of the EU through the creation of large firms (or even a
single buying authority) against dominant exporters to shift economic rents from outside the
EU to the inside.
The basic concept is well known in economic literature as countervailing power (or bilateral
monopoly) (Galbraith 1954). If companies, that compete horizontally on the same market,
cannot be disciplined, bargaining power on the counter-side of the market may provide this
discipline. On the other hand, competition in other parts of the value chain may be
hampered. If the buyer with countervailing power has also a selling side, bargaining power
may be transferred further downwards (cf. e.g. Ungern-Sternberg 1996 and Dobson et al.
2001).
Proponents of countervailing power often argue that the risk of a gas producers’ cartel to be
similar to oil, which would underpin the need to strengthen the bargaining power of the EU.
They refer to the Gas Exporting Countries’ Forum (GECF) and claim GECF to be equivalent
to the Organization of Petroleum Exporting Countries (OPEC). GECF was founded in 2001
and has eleven members including Algeria, Bolivia, Egypt, Equatorial Guinea, Iran, Libya,
Nigeria, Qatar, Russia, Trinidad and Tobago and Venezuela, which together account for
around two thirds of global gas reserves. However, the current structure and functioning of
the gas markets are rather different from oil (see e.g. Ochssée 2010). Contrary to oil,
Policy paper 101
exporters’ short-term abilities to limit gas production are constrained by the predominance of
long-term take-or-pay contracts, which implies that cooperation is unlikely to involve the
control of output and the influence on prices in the same manner as the OPEC does.
Therefore, the likeliness of an OPEC-like gas cartel is small, at least in the short run.
Collusion may also be weak in the long run, because interests among GECF members seem
rather heterogeneous. Russia that mainly relies on pipeline-bound gas might be interested in
regionally split markets, which are not contested by others. On the hand, Qatar with the
world’s third largest proven gas reserves has opted for LNG, which is more flexible and
globally traded due to its lower transportation costs. Therefore, Qatar might have less
interest in split markets. Furthermore, despite the expected higher importance of LNG in the
future, LNG will likely not turn GECF into an OPEC-like cartel with an increased scope for
collusion, since long-term contracts are also popular for LNG.137 Thus, bargaining power of
gas producers at the global level might not be as high as it is often postulated, which
challenges the need for countervailing power. Paradoxically, the EU efforts towards more
competition and more liquid spot markets may enhance the potential of a gas cartel, because
they might imply a move away from long-term contracts (see Section 5.3.2.1).
According to Brunekreeft and Guliyev (2009), countervailing power may generally be a high-
risk strategy in the European context, because detrimental effects in the internal gas market
(i.e. reduced competition) likely offset any benefits from lower import prices. Even with
countervailing power created by big players, European bargaining power may be small
relative to the major foreign suppliers such as Gazprom. Outside options of Russia (e.g.
China) seem more credible than those of the EU, which leaves Russia with the better
bargaining position. Moreover, Gazprom can effectively by-pass any countervailing power
through forward integration into the EU market. If companies like Gazprom invest in the EU
downstream market, Europe may also profit due to the creation of new and strong
competitors. As shown by Bolle and Ruban (2007), a forward integration of Gazprom may, in
fact, improve rather than decrease EU’s social welfare, because the mark-up on import
prices is likely compensated by substantially lower downstream prices as a result of
increased competition.
Instead of the creation of countervailing power, diversification appears to be better suited to
cope with the risk of high import prices. This approach tries to reduce the dependence on
one source by opening up the option for other sources. Measures like the expansion of LNG
capacity and the diversification of pipeline import routes increase the potential to switch to
other suppliers and thus improve the EU’s bargaining position. At the same time, competition
may be fostered, because new players might be attracted (e.g. LNG from countries not yet
present on the European market). The stability of collusion could be additionally undermined
and the scope for competition further increased if the EU were able to create excess
capacity.
137 In 2010, around 80% of global LNG was delivered under long-term contracts (IGU 2010). Although LNG
terminals are less capital-expensive and have a lower asset specificity than pipelines, new investments continue to be underpinned by long-term contracts.
Policy paper 102
Supply security has – at least partly – public good characteristics (Bohi and Toman 1993 and
Arnold and Hunt 2009). For example, if Nabucco (if ever built) effectively reduces Russia’s
bargaining power, gas imports via other routes will also benefit from lower import prices (non-
excludability). More difficult to determine is the extent to which the good is non-rival. If the
bargaining power of Russia is non-linear in additional EU gas demand with a certain
threshold level, non-rivalry exists until the threshold level is reached. Beyond this level,
consumption of others is affected (rival) and the good turns into a common good. A similar
reasoning applies to supply disruptions. The better the interconnectivity of the European gas
market, the higher the possibility that benefits do not only occur, where they have been
generated. This phenomenon is well-known as network externality (cf. e.g. Katz and Shapiro
1985 and Midthun et al. 2009).138
Due to these public good characteristics, private capital likely underinvests in supply security
measures instead of creating excess capacities. This opens up the scope for state
interventions. Against this background, the direct support of the Nabucco project by the EU
and the co-financing under the EEPR appear somehow justified. Nevertheless, the EU
should be aware that its backing of Nabucco might be interpreted differently by third
countries. Explaining the difference to Russia, which is often accused of combining its role as
a major energy supplier with political purposes, will be a difficult task for the EU.
Furthermore, the primary goal of the EU should be to set up proper framework conditions
with sufficient investment incentives to initiate private capital-driven activities to the largest
extent possible. Political interventions should only support investments where necessary, i.e.
where reasonable, i.e. private capital-driven, investments in supply security would not have
taken place otherwise.
As illustrated before, the whole planning process of the European gas infrastructure is, to a
large extent, led by the EU, which hints at a lack of private capital-driven investments beyond
the public good characteristic of supply security measures. One reason has already been
identified with access conditions that currently lead to a rather non-transparent market
environment impeding private investments (see Section 5.4.1.1). Another reason could be
investment-related regulations, which provide for insufficient incentives. This problem is
tackled in the following section.
Summing up, security of supply and competition are sometimes regarded as conflicting
goals. As has been argued in the previous two sections, this does not need to be case.
Therefore, the EU concerns about supply security should not be used to argue in favour of
any measures, which potentially hamper competition (like less strict unbundling rules for gas
networks aimed at the creation of countervailing power). Instead, the diversification of import
routes and indigenous facilities is capable of sufficiently improving supply security without a
high risk of impeding competition. While the former mainly addresses the risk of high import
prices by reducing the bargaining power on the counter-side, the latter is primarily targeted at
138 Note that the problem shares similarities with network congestion. Congestion can be regarded as a
negative externality, which is internalised through congestion charges network users have to pay. Supply security is a positive externality and more difficult to internalise, because the whole EU is concerned (or at least larger European regions). Thus, one of the main differences is the geographical scope of the problem.
Policy paper 103
supply disruptions by improving interconnectivity. Competition on the European gas market
can even be fostered by these measures if the EU is able to create excess capacity
(internally and at import points). However, due to the public good characteristics of supply
security enhancing investments, additional support would be required.
5.4.2 Investment-related regulations
The profitability of TSOs is strongly affected by ratemaking provisions. Most of the EU
Member States have established some kind of cap regulation (price- or revenue-based),
which is often also referred to as incentive regulation.139 Compared to rate-of-return (RoR)
regulation, the other basic approach, cap regulations tend to provide lower investment
incentives, especially with regard to network extensions (see e.g. Armstrong and Sappington
2006 as well as Brunekreeft and Borrmann 2010). This is due to the Averch-Johnson effect,
which favours capital-intensive investments under RoR at the expense of X-inefficiencies.140
Contrary to Europe, US interstate gas pipelines operate under RoR. Furthermore, FERC
grants an allowed rate of return that is roughly one to two percentage points higher than the
European average.141 Furthermore, high-powered regulations (price or revenue cap) tend to
be associated with higher risks that a regulated company is exposed to than low-powered
RoR regulations (Pedell 2006). On the other hand, many EU countries with high-powered
regimes have implemented a so-called “building-blocks” approach. Under building blocks,
capital expenditures (CAPEX) are not included in cost reduction requirements and actually
treated as being RoR-regulated. Moreover, most Member States have – to a certain extent -
reacted to the problem of delayed recognition of CAPEX for ratemaking.142 E.g., Portugal,
France and the UK apply a planned cost approach, thus evading the problem completely.
According to EURELECTRIC (2011), only in Germany, the Netherlands and Slovakia the
problem of delayed recognition remains.143 Last but not least, a few countries have
139 I focus on TSOs as they can be regarded as the most crucial essential facility regarding European
interconnectivity. Concerning ratemaking regulations, I refer to KEMA (2009), since the study provides the most recent and consistent overview for European TSOs. Changes in some countries, which have occurred in the meanwhile or are expected to be implemented in the near future, have not led or will not lead to a significant shift of the overall picture. E.g., the UK is going to change to a new regulatory framework, the so-called RIIO model (Revenue set to deliver stronger Incentives, Innovations and Outputs). Although output-oriented, RIIO still relies on incentives. According to compliance with targets agreed upon with the regulator, the system operator receives a bonus or malus (Müller 2011).
140 The problem of incentives to underinvest into quality under cap regulations (see e.g. Spence 1975) seems
less important for gas TSOs. Since natural gas is declared as a dangerous good throughout Europe, network operators already have to fulfil a lot of safety requirements making pipeline interruptions less likely. The aspect of product quality is covered by shipping contracts and constitutes an essential part of them. Finally, quality of service regulations are usually considered at the DSO level.
141 Hirschhausen (2008) estimates an average weighted cost of capital (WACC) for US interstate pipeline
projects between 1996 and 2003 at 11.6%. Gordon and Makholm (2008) report that the average allowed return for US DSOs has declined from 11.35% in 2000 to 10.35% in 2007. According to KEMA (2009), the European average WACC lies somewhere between 6.5% and 7%. While the US uses nominal pre-tax rates, Europe mostly applies real pre-tax rates. Taking this information into account and assuming an average US inflation rate of around 2.5% p.a. over the last 10 years, the corresponding WACC-spread amounts to around one or two percentage points.
142 The delayed recognition is a common problem of ratemaking, i.e. actual costs enter allowed revenues with a
certain delay. Due to the (intended) decoupling of costs and revenues over the regulation period of usually three to five years, the problem is more severe under incentive than under RoR regulation (in case of increasing costs).
143 Assuming that measures equally apply to electricity and gas.
Policy paper 104
established special provisions to foster new investments. E.g., France and Italy apply
investment premiums in terms of a higher allowed rate of return for new infrastructure.
Since most of these investment-friendly measures have been implemented only recently, at
least TSO ratemaking has not really stimulated major investments into infrastructure
extensions so far. Looking at the planning process of larger projects, however, the
substantial delays caused by national approval procedures seem to create much higher
barriers to investment than ratemaking. The Commission (2011b) reports an average time
lag of approximately 12 years between the final investment decision and the construction of
the infrastructure. The situation probably worsens when more than one Member State is
involved. Thus, only those projects are implemented that have expected returns well above
the allowed rates. During the last decade, the US have managed to significantly reduce the
delay caused by regulatory action. Due to a simplified approval process with lower
obligations for TSOs to provide proof for the economic need of a project, the typical time
period required to plan and construct a large-scale pipeline extension dropped from five
years to two (Makholm 2007).
On October 19, 2011, the European Commission (2011e) announced a proposal for a new
regulation intended to streamline the approval process of large trans-European infrastructure
projects. But the proposal falls short of creating clear roles and responsibilities. Instead of
giving ACER a strong mandate to take over the oversight and regulation of trans-European
infrastructure projects, it just assigns a low-powered coordinating function to ACER and the
European Commission. Pipelines crossing several Member States, like Nabucco or South
Stream, are comparable to US interstate pipelines. Thus, ACER could be the European
counterpart of FERC. Assigning to ACER the corresponding power and responsibilities would
help to achieve harmonised decisions regarding the regulatory treatment of trans-European
infrastructure projects. Instead, regulatory decisions remain with national regulators.
Therefore, it is not very likely that the new regulation, once implemented, leads to a similar
reduction in delays as in the US.
In fact, currently most large-scale pipeline extensions are (or will be) granted exemption
under Art. 36 of Directive 2009/73/EC (EREGEG 2009). Since almost all of these projects
are developed by consortia consisting of major incumbents,144 granting exemptions from
TPA and strong unbundling obligations may create new barriers to market entry in particular
due to the role of long-term contracting. Long-distance gas pipelines that connect remote
production fields to larger consuming regions can be classified as asset-specific investments
like gas production as discussed in Section 5.3.2.1. Thus, long-term contracts, regarding
pipelines in terms of capacity, are generally a viable tool to overcome the inherent risk of
opportunistic behaviour and reduce transaction costs. On the other hand, if the pipeline
investor is closely linked to the holder of long-term capacity contracts and does not have any
necessity to open up the infrastructure for third parties, the market bears a high risk of being
foreclosed.
144 Regarding the three trans-European pipeline projects mentioned above, the main participating companies
are: (1) Nabucco: OMV, RWE and MOL; (2) South Stream: Gazprom, ENI and OMV; (3) Nord Stream: Gazprom, Wintershall, E.ON, Gasunie and GDF Suez.
Policy paper 105
FERC has solved this conflict by introducing “pipeline-specific ownership unbundling”, as
explained in Section 5.3.1. The provision has removed TSOs (or their affiliates) from having
any dominant role in capacity or commodity trading (Makholm 2006). In the US, pipelines
operate under an open access (i.e. TPA) regime, and an RoR regulation is applied to primary
capacity. Most interstate pipelines have been (and still are) built relying on long-term capacity
contracts. Unused capacity is actively traded in secondary markets. Regarding commodity,
the importance of long-term contracts has decreased significantly together with the
development of well-established and liquid spot markets. In contrast, the EU tries to address
the problem by granting TPA exemptions and imposing obligations on investors, which are
specified on a case-by-case basis. These obligations include provisions regarding the
application of open season procedures, limits on capacity reserved for pipeline investors, the
establishment of bulletin boards for secondary capacity trading, the implementation of
UIOLI/UIOSI principles etc.145 Thus, while FERC has set a clear frame within which the
market adjusts, the EU has opted for case-specific regulatory interventions where guiding
principles hardly exist.
Against this background, the question can be raised as to whether the extensive use of TPA
exemptions in the European context for new gas infrastructure projects is justified. Usually,
the intention of incumbents to build new pipelines, like E.ON with Nord Stream, is to supply
new clients and thus expand their downstream sales activities. The high asset specificity of
pipelines means that investment costs are highly irreversible once the investment is made.
Access regulation opens up the network for entrants that can wait to see if the investment is
successful before they enter, e.g. if sufficient gas demand can be attracted supplied by the
pipeline. Entrants seek access only in good states, and the incumbent’s profits will be
reduced due to increased downstream competition. Therefore, access regulation leads to an
asymmetric risk sharing, with the incumbent bearing all of the down side risk but the
incumbent and entrants sharing the upside risk. If the investor expects regulatory
opportunism, i.e. an access regime that does not account for this asymmetry with ex-post
tariffs oriented towards short-run marginal cost, investment incentives are distorted.
Especially if market entry is costless, the incumbent delays investments (Woroch 2004). One
possibility to overcome the problem is to grant TPA exemptions, also known as access
holidays (Gans and King 2004), which allow the incumbent to temporarily earn monopoly
profits as a compensation for the asymmetric risk sharing. According to this line of argument
the EU approach would be reasonable.
However, the basic problem gets more complicated if e.g. other incumbents may also invest,
market entry is not costless or entrants have the possibility to enter into a pipe-to-pipe
competition. The effect on investment incentives (reduced or even increased) mainly
depends on market imperfections at the downstream level and the quality of the access
regulation.146 The European context offers some contra arguments, which challenge the
current approach.
145 See ERGEG (2009) for further details.
146 See Guthrie (2006) for an overview. Following Guthrie, an interesting distinction with regard to the three major European pipeline projects can be derived. Nord Stream would constitute a “waiting game”, because
Policy paper 106
First, the European downstream market is characterised by a high market concentration with
most incumbents holding a dominant position (see Section 5.3.1). From a welfare
perspective, it might be argued that the incumbent is already compensated for the high ex
ante risk, at least partly, by mark-ups on retail prices paid by its existing customer base.
However, investment incentives are determined by changes at the margin. Providing pipeline
access is generally associated with opportunity costs for the incumbent due to foregone
profits from sales lost to the entrants (Armstrong et al. 1996 and Sappington 2005). The
opportunity costs are likely to be higher with imperfect competition prevailing at the retail
level, because the induced downstream competition may lead to an additional margin
reduction on sales to existing customers.147 In turn, high opportunity costs will delay
incumbent’s investments if the investor is not adequately compensated. Roughly speaking,
the incumbent has more to lose under low retail competition if its pipeline investment is
exposed to access regulation.148 Hence, incumbents are more likely prompted to ask for
TPA exemptions, which corresponds to current market observations. Therefore, regulators
face a dilemma similar to the one outlined for long-term contracts (see Section 5.3.2.1).
Under current market conditions imposing strong access regulations on potential large-scale
pipeline projects may impede investment incentives of incumbents due to possibly high
opportunity cost. Consequently, the prospects of excess capacity might be reduced, which
would affect supply security and competition in the long run (see Section 5.4.1). On the other
hand, granting exemptions may also hamper the development of competition in the long run
due to a higher risk of market foreclosure. Thus, if regulators follow their current approach,
their efforts of creating competitive framework conditions may be counteracted.
Second, the reasoning in favour of access holidays is based on the assumption that network
operators are vertically integrated or at least not sufficiently unbundled from their sales
activities. As discussed in Section 5.3.1, in the future European TSOs are likely to act more
independently from their parent companies (if not completely ownership unbundled). The
intention of unbundling is to cut off completely the link to retail services. With perfect
unbundling, investment decisions of TSOs are made without considering effects on affiliated
gas sales. Consequently, the problem of asymmetric risk sharing disappears and investment
incentives become solely a ratemaking issue. RoR may be considered first, because it tends
to favour capital-intensive investments as outlined at the beginning of this section. The
pipeline business is not characterised by rapid technological progress. Therefore, regimes
that allow the regulated firm more flexibility in setting its tariffs, like e.g. price caps (see
Guthrie 2006), may not be required. Furthermore, RoR is well understood, well-known to the
sector and has demonstrated its usefulness in the US context.
the absence of any other competing project points at large cost advantages, which is underpinned by the Gazprom lead. On the other hand, the two competing projects South Stream and Nabucco might be classified as a “preemption game” due to a smaller cost difference. In the first case, investment timing would not have been influenced by others, whereas in the second case the presence of the competitor (Nabucco) may accelerate the leader’s (South Stream) investment.
147 Assuming perfect retail competition, the sales margin on existing customers would be zero, and thus additional opportunity cost would not occur if access is allowed.
148 This argument is closely linked to the incentive of vertically integrated companies to underinvest in order to create scarcity and increase transportation charges above marginal cost (Joskow and Tirole 2005).
Policy paper 107
Companies that make large lumpy investments take on large risks, since future demand is
highly uncertain. Therefore, these projects are often exposed to a higher risk of turning out
unsuccessful, which ratemaking should take into account; e.g. by a higher allowed rate of
return than for ordinary projects. In essence, investments like Nabucco incorporate an option
for the sector as a whole on uncertain future demand (see e.g. Evans and Guthrie 2006). If
expectations turn out wrong, the project might be abandoned and thus includes an
abandonment option.
If current unbundling requirements turn out to be insufficient and the problem of asymmetric
risk sharing persists, further regulatory action towards stricter vertical separation should be
taken (e.g. “pipeline-specific ownership unbundling”) instead of making extensive use of TPA
exemptions. Even if ratemaking overstates the project risk ex ante, granting TSOs moderate,
but regulated rents seems less risky than probably much higher rents generated through
market foreclosure under TPA exemptions. Furthermore, the capital bias of RoR could work
towards the creation of excess capacity, which would improve supply security and might help
to foster competition.
Third, according to Art. 36 of the third Directive TPA exemptions are especially targeted at
those investments that improve supply security. Thus, the extensive use of this provision is
likely – at least to some extent – motivated by the concerns that are related to the increasing
import dependency of the EU. Given the assumption is true, the approach shares similarities
with the one discussed for countervailing power: A monopoly is created aimed at enhancing
the security of European gas supplies. Due to the public good characteristics of supply
security measures, granting the possibility to earn monopoly rents can be interpreted as an
indirect compensation for the externalities the project creates. While countervailing power
has been argued to be a high-risk strategy (see Section 5.4.1.2), the risk of market
foreclosure in case of security of supply driven TPA exemptions is completely unnecessary.
The same benefit, i.e. project viability, can be achieved by direct financial support (e.g. via
EEPR co-financing) or by a higher allowed rate of return under RoR.
The problem of large-scale trans-European pipeline investments clearly offers a lot of scope
for future research. For example, proponents of access holidays stress the asymmetry
between the opportunity for entrants to obtain access for short periods and the long-term
investment commitment of the incumbent assuming that parts of the incumbent’s supply are
displaced by entrants (e.g. Pindyck 2004). Regarding gas, short-term accessibility may be
limited, because a large part of pipeline supplies is already contracted long term. On the
other hand, studies on long-term contracts postulate that producers will be reluctant to make
investments unless they are assured of long-term access to pipeline capacity, and pipeline
operators will refuse to make investments unless producers are willing to commit reserves on
a long-term basis (e.g. Creti and Villeneuve 2005). In a world of unbundling, incentive
structures of infrastructure investors may change due to changing economies of scope.
Shippers could increasingly act as intermediaries between pipelines and wells, which does
not alter the necessity of a physical connection between both infrastructures, but it may
modify the management of this connection. The number of potential contractors affects asset
specificity (see Section 5.3.2.1), which may render long-term contracting (pipeline and well)
Policy paper 108
less important. Consequently, the market foreclosure argument would also lose importance.
Last but not least, the role of companies from third countries, such as Gazprom, may lead to
different conclusions.
Nevertheless, the EU strategy regarding major infrastructure projects with TPA exemptions
extensively granted by Member States seems not the right way under current conditions.
Due to highly concentrated downstream markets dominated by incumbents, access holidays
will likely lead to the persistence (or exacerbation) of entry barriers and thus counteract
current EU efforts towards a single and competitive European gas market. Instead, ACER
should be assigned with the power and responsibility to develop a proper regulatory
framework that is better fitted to overcome current investment barriers. In this regard, a
regime based on RoR-regulated tariffs may serve as a starting point for further
considerations. The US experience has shown that an RoR approach with adequate rates of
return is able to attract sufficient private capital. Furthermore, RoR seems well-suited to
address some of the problems specifically associated with large-scale infrastructure projects,
such as the different risk exposure. Allowed rates of return could be adjusted to individual
project requirements based on an in-depth risk assessment and clear grounds of guiding
principles, which are the same across Europe.
Summing up, in addition to the identified problems of a non-transparent market environment
caused by access conditions (Section 5.4.1.1) and the public good characteristics of supply
security measures (Section 5.4.1.2), investment-related regulations possibly also contribute
to the lack of private capital-driven investments into the European gas infrastructure.
Regarding insufficient investment incentives, time lags associated with the approval process
of large-scale projects likely play a much greater role than ratemaking provisions. The
attempts of the EU to address this problem are not very promising. The proposal for a
simplified approval procedure should be adapted in the sense of assigning to ACER a clear
role and appropriate responsibilities for trans-European infrastructure projects. Instead of an
extensive use of Art. 36 of Directive 2009/73/EC by national regulatory authorities granting
access holidays, ACER should develop a corresponding framework, which might be based
on an RoR regulation. This type of ratemaking seems to cope well with some of the specific
characteristics associated with large-scale infrastructure projects (e.g. risk exposure).
Granting TSOs moderate, but regulated rents appears more advisable than facing the risk of
probably much higher rents generated through market foreclosure. However, further
considerations are necessary to get a better understanding of the underlying problems.
Meanwhile and to facilitate information disclosure, the EU could use the planning procedures
mentioned in Section 5.4.1.1 for tendering certain projects to the market.149
149 For further information regarding tendering procedures see e.g. Glachant (2011: 43f.). The tendering should
only be used for an interim period of time. Otherwise, the European gas market faces the risk of turning into a state-directed economy.
Policy paper 109
5.5 Conclusions
While the US have already several years of experience of running the gas business like a
competitive market activity, Europe can still be classified as an emerging market. Retail
competition is not very well developed, showing low levels of customers switching to other
suppliers. Wholesale is bound to long-term contracting. Short-term commodity trading is far
from being liquid, secondary markets for pipeline and storage capacity are hardly existent.
Arbitrage potentials remain unexploited. The market severely lacks transparency, relevant
information is not disclosed to market participants via corresponding price signals.
Interconnectivity suffers from contractual congestion through capacity hoarding and
insufficient investments. Last but not least, investment activity is dominated by incumbents.
Although some progress was achieved through the first two European gas Directives, the
third legislative package of 2009 together with the accompanying efforts seem better suited
to overcome some of the identified obstacles. The current regulations tighten minimum
requirements and specify operational procedures as well as market rules in more detail.150
The scope for diverging, country-specific approaches has been significantly reduced, which
will probably lead to an improved level of harmonisation across Europe. Before 2009, gas
market regulations were mostly just a copy of corresponding provisions for the electricity
sector. Nowadays, specific characteristics of the gas sector are increasingly taken into
account, e.g. the different nature of pipeline capacity contracts.
Having looked at the unbundling rules for network operators, no urgent need for changes has
been identified. Although they have been one of the most controversial issues in the
discussions preceding the third package, TSO unbundling provisions seem appropriate to
lead to sufficiently independent pipeline operators. If this expectation turns out wrong, an
adjustment of the unbundling rules is recommended, such as disallowing TSOs to own gas
transported through their trunk lines. This would constitute a kind of “pipeline-specific
ownership unbundling” without the necessity to fully ownership unbundle TSOs. It
corresponds to the application of the commodities clause under FERC Order No. 636 in the
US, which is regarded as a key step towards a competitive US natural gas market. Whether
poor retail competition is caused by the lower unbundling requirements for DSOs is not clear
yet. To get a better understanding, a European-wide margin squeeze test is suggested.
The EU efforts with regard to wholesale also mainly point into the right direction, i.e. towards
more competitive framework conditions (e.g. the European-wide implementation of de-
coupled entry-exit systems). A crucial success factor is how the EU will be able to balance
the conflicting interests of short- and long-term demand for both capacity and commodity. To
account for the highly asset-specific gas infrastructure and to preserve investment incentives,
only a certain percentage of overall pipeline capacity should be auctioned at short notice.
Potential market foreclosure incentives of long-term contract holders are circumvented best
by gas release programmes for the commodity market and the establishment of UIOLI or
150 By the time of writing, work on several provisions (e.g. framework guidelines and the European network
code) is still on-going. Thus, the assessment is preliminary and assumes that rules are decided as they are currently in the offing.
Policy paper 110
UIOSI principles for capacity. Together with a market-based allocation of primary capacity
and the introduction of other products facilitating the tradability of pipeline capacity (e.g.
bundled hub-to-hub capacity) gas prices will likely better reflect market conditions (e.g.
infrastructure bottlenecks), disclosing information to market participants that has not been
available so far. However, the promising outlook for the wholesale segment might be
constrained by poor transparency obligations for TSOs. According to Art. 16 (1) of the third
gas Directive, confidentiality clearly outweighs market needs for information. So far,
information disclosure of TSOs has been mainly driven by regulation. Thus, if current efforts
do not sufficiently improve market transparency, a change of information policy requirements
may be advisable.
Further weaknesses of the current regulatory framework are associated with the adjacent
storage and LNG markets. The non-binding guidelines GGPSSO and GGPLNG, which
address transparency, congestion management and capacity allocation issues for storage
and LNG, should be made mandatory. Major existing storage facilities seem to lack the
possibility of TPA exemption, if they proof to operate in a competitive environment. To allow
a level-playing field with new storage, Art. 36 of the third Directive should be extended to
existing sites.
One of the key differences of the European market compared to the US is the high and
increasing import dependency. Nevertheless, the corresponding EU concern about supply
security should not be used to justify regulatory measures that impede competition. Instead,
the diversification of import routes and indigenous gas infrastructure seems capable of
sufficiently improving supply security without distortive effects on competition in other parts of
the value chain. While the former mainly addresses the risk of high import prices by reducing
the bargaining power on the counter-side (e.g. the Russian gas giant Gazprom), the latter is
primarily targeted at supply disruptions by improving interconnectivity.
Regarding investment incentives, most crucial is how the EU addresses trans-European
pipeline projects, which currently suffer from great delays caused by intransparent national
approval procedures. The current approach of granting TPA exemptions under Art. 36 of
Directive 2009/73/EC faces a high risk of creating new barriers to entry. Instead, a revision of
the current proposal for a new regulation intended to streamline the approval process is
strongly recommended. ACER should be assigned a clear mandate to take over the
oversight and regulation of these projects. The US example has shown that an RoR
approach with adequate rates of return might be able to attract sufficient private capital to
large-scale gas infrastructure projects. RoR seems to cope well with some of the specific
characteristics of these projects (e.g. risk exposure). Granting TSOs moderate, but regulated
rents appears more advisable than facing the risk of probably much higher rents generated
through market foreclosure. Furthermore, facilitating infrastructure investments may also
improve wholesale competition, because abundant capacity reduces the risk of market
power.
According to Joskow (2010), to evaluate regulatory reforms in the real (i.e. imperfect) world,
one has to balance the costs of imperfect markets against the costs and benefits of
Policy paper 111
regulatory imperfections (e.g. due to an imperfectly informed regulator). The analysis of this
paper has illustrated that the EU has significantly advanced on that balance. Nevertheless,
large scope for further improvement remains. It is very unlikely that the EU will meet its target
of an integrated European gas market by 2014. Current regulations have been adopted
without a common understanding of what final set-up of the European gas market to aim at.
The discussions on the so-called “gas target model” started after the third package had
already been decided on.151 Possibly, North-West Europe will be well-integrated by 2014,
since markets are already more developed than in other European regions, and
interconnectivity suffers from contractual congestion rather than from missing investments.
Still, this scenario presumes timely and proper completion of the work on outstanding
regulations (e.g. on framework guidelines and the European network code) and timely
implementation of the third package by Member States. Regarding the latter, transposition
deficits observed by the European Commission (2011a) are not very promising.152
Overall, the EU seems more or less on track towards competitive framework conditions.
Certainly, further efforts on the regulatory framework are inevitable to catch up with the US
gas market. It will probably take years, if not decades, to achieve a market performance
throughout Europe similar to the US.
Regarding areas of future research, large-scale investment projects with the involvement of
companies from third countries, like the Russian gas giant Gazprom, are likely one of the
most demanding and urgent topics, because it pulls together many individual aspects, such
as e.g. supply security, long-term contracts and access regulation, which are of special
concern for a well-functioning European gas market. The interplay of the various issues has
not been well studied yet. A better understanding would be of utmost importance to establish
a proper framework that provides sufficient investment incentives without hampering
competition. Another question is concerned with the usefulness of the distinction between
nTPA and rTPA for storage operators, which requires further considerations. With regard to
long-term contracts, it could be interesting to study previous gas release programmes in
order to identify relevant success factors.
151 For the current state of on-going discussions on the gas target model see e.g. CEER (2011). 152 By June 1, 2011, no Member State had notified transposition measures to the Commission. Moreover,
several countries still do not comply with obligations already set out under the second Directive.
Policy paper 112
References
Allaz, B. and J.-L. Vila (1993), Cournot Competition, Forward Markets and Efficiency, Journal of
Economic Theory 59(1), 1-16.
Armstrong, M., C. Doyle and J. Vickers (1996), The Access Pricing Problem: A Synthesis, Journal of
Industrial Economics XLIV(2), 131-150.
Armstrong, M. and D.E.M. Sappington (2006), Regulation, Competition and Liberalisation, Journal of
Economic Literature 44(2), 325-366.
Arnold, S. and A. Hunt (2009), National and EU-Level Estimates of Energy Supply Externalities, CEPS
Policy Brief No. 186, April, Centre for European Policy Studies, Brussels.
Ascari, S. (2011), An American Model for the EU Gas Market?, EUI Working Papers No. RSCAS
2011/39, Robert Schuman Centre for Advanced Studies, Florence School of Regulation,
Florence.
Bandulet, M., J. Büchner, W. Elsenbast, O. Flörcken and O. Obert (2010), Macro-Economic Effects
Regarding Congestion Management in Europe, Study on behalf of the Bundesnetzagentur
(Germany), the Commission de Régulation de l’Energie (France) and E-Control (Austria), E-
Bridge Consulting GmbH, May 11, Bonn.
Bertoletti, P., A. Cavaliere and A. Tordi (2008), The Regulation of Access to Gas Storage with
Capacity Constraints, Working Paper, University of Padua, February, Padua.
Bilgin, M. (2009), Geopolitics of European natural gas demand: Supplies from Russia, Caspian and
the Middle East, Energy Policy 37(11), 4482-4492.
Bohi, D.R. and M.A. Toman (1993), Energy security: externalities and policies, Energy Policy 21(11),
1093-1109.
Bolle, F. and R. Ruban (2007), Competition and Security of Supply: Let Russia Buy into the european
Gas Market! Discussion Paper No. 258, European University Viadrina Frankfurt (Oder),
Department of Business Administration and Economics, Frankfurt (Oder).
BP (2011), Statistical Review of World Energy 2011, London.
Brau, R., R. Doronzo, C.V. Fiorio and M. Florio (2010), EU Gas Industry Reforms and Consumers’
Prices, The Energy Journal 31(4), 163-178.
Brunekreeft, G. (2001), Negotiated Third-Party Access in the German Electricity Supply Industry,
Economia Della Fonti Di Energia E Dell’Ambiente 44, 31-54.
Brunekreeft, G. (2004), Regulatory Threat in Vertically Related Markets: The Case of German
Electricity, European Journal of Law and Economics 17(3), 285-305.
Brunekreeft, G. and J. Borrmann (2010), The Effect of Monopoly Regulation on the Timing of
Investment, Bremen Energy Working Papers 01, January, Bremen.
Brunekreeft, G. and F. Guliyev (2009), Gas Supply Security and the Competitiveness on the European
Gas Market, Jepman, C.J. (ed.), Gas Market Trading, Energy Delta Institute, Groningen, 68-
123.
Buehler, S. (2005), The Promise and Pitfalls of Restructuring Network Industries, German Economic
Review 6(2), 205-228.
Policy paper 113
Buehler, S., S. Schmutzler and A. Benz (2004), Infrastructure Quality in Deregulated Industries: Is
there an Underinvestment Problem? International Journal of Industrial Organization 22, 253-
267.
Bundeskartellamt (2010), Bericht über die Evaluierung der Beschlüsse zu langfristigen
Gaslieferverträgen, June 15, Bonn.
CEER [Council of European Energy Regulators] (2011), CEER Vision Paper for a conceptual model
for the European gas market, Document No. C11-GWG-74-03, March 2, Brussels.
Christie, E.H. (2011), The potential for an EU Gas Purchasing Agency, Christie, E.H., P.K. Baev and
V. Golovko (eds), Vulnerability and Bargaining Power in the EU-Russian Gas Relations, FIW-
Research Reports 2010/11 No. 03, March, Vienna, 42-61.
Colombera, M. (2010), Energy Law in Italy: Implementing the Third Energy Package and the Climate
Change Package in Italy, in: Smith, M. (ed.), European Energy Review 2010, London, 69-73.
Conway, P. and G. Nicoletti (2006), Product Market Regulation in the Non-Manufacturing Sectors of
OECD Countries: Measurement and Highlights, OECD Economics Department Working
Paper, No. 530, Paris.
Copenhagen Economics (2005), Market Opening in Network Industries: Part II: Sectoral Analyses,
Technical report, Copenhagen Economics for DG Internal Market.
Coquet, P. (2007), Opening of the Gas and Electricity Markets to Retail Competition - What lessons
can be learned from our European neighbours?, Capgemini Policy Paper, Paris.
Cottier, T., S. Metteotti-Berkutova and O. Nartova (2010), Third Country Relations in the EU
Unbundling of Natural Gas Markets: The “Gazprom Clause” of Directive 2009/73 EC and WTO
Law, Working Paper No. 2010/06, May, Swiss National Centre of Competence in Research,
Bern.
Cremer, H. and J.-J. Laffont (2002), Competition in gas markets, European Economic Review 46, 928-
935.
Creti, A. (2009), The Economics of Natural Gas Storage – A European Perspective, Springer, Berlin.
Creti, A. and B. Villeneuve (2005), Longterm Contracts and Take-or-Pay Clauses in Natural Gas
Markets, Energy Studies Review 13(1), 75-94.
Crocker, K.J. and S.E. Masten (1985), Efficient Adaptation in Long-term Contracts: Take-or-Pay
Provisions for Natural Gas, American Economic Review 75, 1083-1093.
De Jong, J., H. Maters, M. Scheepers and A. Seebregts (2006), EU Standards for Energy Security of
Supply, Report No. ECN-C-06-039/CIEP, Energy research Centre of the Netherlands /
Clingendael International Energy Programme, June, Petten / The Hague.
DECC [Department of Energy & Climate Change] (2010), Energy Trends, Quarter Bulletin, Issue
December, London.
Dieckhöner, C. (2010), Simulating security of supply effects of the Nabucco and South Stream
projects for the European natural gas market, EWI Working Paper No. 10/07, December,
Institute of Energy Economics at the University of Cologne, Cologne.
Dobson, P., R. Clarke, S. Davies and M. Waterson (2001), Buyer power and its impact on competition
in the food retail distribution sector of the European Union, Journal of Industry, Competition
and Trade 1(3), 247-281.
Policy paper 114
EFET [European Federation of Energy Traders] (2005), EFET response to Francois Lamoureux, DG
TREN, on questions about progress in EU gas liberalisation and the state of the markets,
June 30, Amsterdam.
ENTSOG [European Network of Transmission System Operators for Gas] (2011), Ten-Year Network
Development Plan 2011-2020, February 17, Brussels.
ERAA [Energy Retailers Association of Australia] (2004), An Australian Wholesale Gas Market – Its
Justification, Framework and Governance, Wholesale Gas Standing Working Group,
September, Sydney.
ERGEG [European Regulators Group for Electricity & Gas] (2009), European Regulators’ Experience
with Article 22 exemptions of Directive 2003/55/EC, Document No. E08-GIF-02-03, March 10,
Brussels.
ERGEG [European Regulators Group for Electricity & Gas] (2010a), ERGEG 2010 Status Review of
the Liberalisation and Implementation of the Energy Regulatory Framework, Document No.
C10-URB-34-04, December 7, Brussels.
ERGEG [European Regulators Group for Electricity & Gas] (2010b), Existing transparency
requirements for natural gas, Public Consultation, Document No. E10-GWG-68-03,
September 8, Brussels.
ERGEG [European Regulators Group for Electricity & Gas] (2011a), Monitoring Report 2010 on
capacity allocation mechanisms and congestion management procedures at selected
interconnection points, Document No. E10-GMM-11-05, February 2, Brussels.
ERGEG [European Regulators Group for Electricity & Gas] (2011b), Final ERGEG study on
congestion management procedures & anti-hoarding mechanisms in the European LNG
terminals, Document No. E10-LNG-11-03b, April 12, Brussels.
ERGEG [European Regulators Group for Electricity & Gas] (2011c), Assessment of Capacity
Allocation Mechanisms and Congestion Management Procedures for effective Access to
Storage and Proposals for the Amendment of the Guidelines of Good Practice of Storage
System Operators (GGPSSO), Document No. E10-GST-14-03, February 2, Brussels.
EURELECTRIC (2011), Regulation for Smart Grids, February, Brussels.
Eurogas (2010), Long-Term Outlook for Gas Demand and Supply 2007-2030, May, Brussels.
European Commission (2005), On the Application of Article 82 of the Treaty to Exclusionary Abuses,
DG Competition Discussion Paper, December, Brussels.
European Commission (2007), Inquiry pursuant to Article 17 of Regulation (EC) No 1/2003 into the
European gas and electricity sectors (Final Report), Communication from the Commission,
COM(2006) 851 final, January 10, Brussels.
European Commission (2010), Energy infrastructure priorities for 2010 and beyond – A Blueprint for
an integrated European energy network, Communication from the Commission to the
European Parliament, the Council, the European Economic and Social Committee and the
Committee of the Regions, COM(2010) 677 final, November 17, Brussels.
European Commission (2011a), 2009-2010 Report on Progress in Creating the Internal Gas and
Electricity Market, Commission Staff Working Paper, June 9, Brussels.
European Commission (2011b), The Internal Energy Market – Time to Switch into Higher Gear, Non-
paper, DG Energy, Brussels.
Policy paper 115
European Commission (2011c), Report from the Commission to the Council and the European
Parliament on the Implementation of the European Energy Programme for Recovery,
COM(2011) 217 final, April 20, Brussels.
European Commission (2011d), EU starts negotiations on Caspian pipeline to bring gas to Europe,
Press Release, September 12, Brussels.
European Commission (2011e), Proposal for a Regulation of the European Parliament and of the
Council on guidelines for trans-European energy infrastructure and repealing Decision No
1364/2006/EC, COM(2011) 658 final, October 19, Brussels.
European Council (2011), Conclusions of the European Council (4 February 2011), General
Secretariat of the Council, Document No. EUCO 2/1/11-REV 1, March 8, Brussels.
European Parliament (2010), Full text of the Buzek and Delors declaration on the creation of a
European energy community, Press Release, May 5, Brussels.
Evans, L.T. and G. Guthrie (2006), Inventive Regulation of Prices when Costs are Sunk, Jounral of
Regulatory Economics 29(3), 239-264.
Galbraith, J.K. (1954), Countervailing power, American Economic Review 44(2), 1-6.
Gans, J.S. and S.P. King (2004), Access Holidays and the Timing of Infrastructure Investment,
Economic Record 80(248), 89-100.
Gazprom (2011), Annual Report 2010, Moscow.
Glachant, J.-M. (2011), A Vision for the EU Target Model: The MECO-S Model, EUI Working Papers
No. RSCAS 2011/38, Robert Schuman Centre for Advanced Studies, Florence School of
Regulation, Florence.
Gordon, D.V., K. Gunsch and C.V. Pawluk (2003), A natural monopoly in natural gas transmission,
Energy Economics 25(5), 473-485.
Gordon, K. and J.D. Makholm (2008), Allowed Return on Equity in Canada and the United States: An
Economic, Financial and Institutional Analysis, NERA Economic Consulting, February, Boston,
MA.
Growitsch, C., M. Stronzik and R. Nepal (2010), Price Convergence and Information Efficiency in
German Natural Gas Markets, Paper presented at the 25th Annual Congress of the European
Economic Association (EEA), August 23-26, Glasgow.
Growitsch, C. and M. Stronzik (2011), Ownership Unbundling of Gas Transmission Networks –
Empirical Evidence, EWI Working Paper No. 11/07, Institute of Energy Economics at the
University of Cologne (EWI), Cologne.
Growitsch, C. and T. Wein (2005), Negotiated Third Party Access – An Indsutrial Organisation
Perspective, European Journal of Law and Economics 20, 165-183.
GSE [Gas Storage Europe] (2010), GSE’s 2010 Survey on Capacity Allocation Mechanisms and
Congestion Management Procedures: Main findings and conclusions, Document No.
10GSE072, September 7, Brussels.
Guthrie, G. (2006), Regulating Infrastructure: The Impact on Risk and Investment, Journal of
Economic Literature XLIV (December), 925-972.
Hattori, T. and M. Tsutsui (2004), Economic Impact of Regulatory Reforms in the Electricity Supply
Industry: A Panel Data Analysis for OECD Countries, Energy Policy 32(6), 823-832.
Policy paper 116
Hedge, K. and E. Fjeldstad (2010), The Future of European Long-term Natural Gas Contracts, Bergen
Energi, September 1, Bergen.
Hirschhausen, C. von (2008), Infrastructure, regulation, investment and security of supply: A case
study of the restructured US natural gas market, Utilities Policy 16(1), 1-10.
Hirschhausen, C.von, A. Neumann and S. Rüster (2007), Competition in Natural Gas Transportation?
Technical and Economic Fundamentals and an Application to Germany, Working Paper WP-
GG-21b, Dresden University of Technology, July, Dresden.
Hirschhausen, C. von and A. Neumann (2008), Long-Term Contracts and Asset Specificity Revisited:
An Empirical Analysis of Producer-Importer Relations in the Natural Gas Industry, Review of
Industrial Organization 32(2), 131-143.
Hirschhausen, C. von, F. Holz, A. Neumann and S. Rüster (2010), Supply Security and Natural Gas.
Lévêque, F., J.-M. Glachant, J. Barquin, C. von Hirschhausen, F. Holz and W.J. Nuttal (eds.),
Security of Energy Supply in Europe: Natural Gas, Nuclear, and Hydrogen, Edward Elgar,
Northampton, MA, 3-20.
Höffler, F. and S. Kranz (2011), Legal Unbundling can be a ‘Golden Mean‘ between Vertical
Integration and Vertical Separation, International Journal of Industrial Organization 29(5), 576-
588.
IEA [International Energy Agency] (2010), World Energy Outlook 2010, Paris.
IEA [International Energy Agency] (2011), Are We Entering a Golden Age of Gas?, World Energy
Outlook 2011,Special Report, Paris.
IGU [International Gas Union] (2010), World LNG Report 2010, Oslo.
Jensen, J. (2007), Gas Pricing: North America, Energy Charter Secretariat (Ed.), Putting a Price on
Energy: International Pricing Mechanisms for Oil and Gas, Brussels, 99-125.
Joskow, P. (2007), Supply Security in Competitive Electricity and Natural Gas Markets, Robinson, C.
(Ed.), Utility Regulation in Competitive Markets: Problems and Progress, Edward Elgar,
Cheltenham, 82-116.
Joskow, P. (2010), Market Imperfections versus Regulatory Imperfections, CESifo Dice Report 8(3), 3-
7.
Joskow, P. and J. Tirole (2005), Merchant Transmission Investment, Journal of Industrial Economics
LIII(2), 233-264.
Katz, M.L. and C. Shapiro (1985), Network externalities, competition, and compatibility, American
Economic Review 75, 424-440.
KEMA (2009), Study on Methodologies for Gas Transmission Network Tariffs and Gas Balancing Fees
in Europe, Report for the European Commission, December, Arnhem.
KEMA (2010), Study on Regulation of Tariffs and Quality of the Gas Distribution Services in the
Energy Community, August, Bonn.
Kwoka, J. and V. Sabodash (2011), Price Spikes in Energy Markets: “Business by Usual Methods” or
Strategic Withholding, Review of Industrial Organization 38(3), 285-310.
Laffont, J.J. and J. Tirole (1993), A Theory of Incentives in Procurement and Regulation, MIT Press,
Cambridge, MA.
Policy paper 117
Laffont, J.J., P. Rey and J. Tirole (1998), Network Competition I: Overview and non-discriminatory
pricing, RAND Journal of Economics 29, 1-37.
Liski, M. and J.-P. Montero (2006), Forward trading and collusion in oligopoly, Journal of Economic
Theory 131(1), 212-230.
Makholm, J.D. (2006), The Theory of Relationship-Specific Investments, Long-Term Contracts and
Gas Pipeline Development in the United States, Paper prepared for the workshop on Energy
Economics and Technology on April 21 in Dresden, National Economic Research Associates
Inc. (NERA), Boston, MA.
Makholm, J.D. (2007), Seeking Competition and Supply Security in Natural Gas: The US Experience
and European Challenge, Paper presented at the 1st CESSA Conference in Berlin on May 31,
National Economic Research Associates Inc. (NERA), Boston, MA.
Makholm, J.D. (2011), Real Unbundlings: The Foundation for a Competitive Gas Market in the United
States, Working Paper, May, National Economic Research Associates Inc. (NERA), Boston,
MA.
Melling, A.J. (2010), Natural Gas Pricing and Its Future – Europe as the Battleground, Carnegie
Endowment for International Peace, Washington, D.C.
Meran, G., C. von Hirschhausen and A. Neumann (2010), Access Pricing and Network Expansion in
Natural Gas Markets, Zeitschrift für Energiewirtschaft 34, 179-183.
Meyer, R. (2011), Vertical Economies of Scope in Electricity Supply – analysing the Costs of
Ownership Unbundling, PhD-Thesis, Jacobs University Bremen, April 12, Bremen.
Michaels, R.J. (2006), Vertical Integration and the Restructuring of the U.S. Electricity Industry, Policy
Analysis Paper No. 572, Cato Institute, July 13, Washington, D.C.
Midthun, K.T., M. Bjørndal and A. Tomasgard (2009), Modeling Optimal Economic Dispatch and
System Effects in Natural Gas Networks, The Energy Journal 30(4), 155-180.
Müller, C. (2011), New regulatory approaches towards investments: a revision of international
experiences, WIK Discussion Paper No. 353, April, Bad Honnef.
Ochssée, T.A.B. von (2010), The Dynamics of Natural Gas Supply Coordination in a New World:
Cooperation or competition between gas-exporting countries from a Russian perspective,
Clingendael International Energy Programme, The Hague.
Olsen, W.P. (2005), Secrecy and Utility Regulation, The Electricity Journal 18(4), 48-52.
Pedell, B. (2006), Regulatory Risk and the Cost of Capital, Springer, Berlin.
Pindyck, R.S. (2004), Mandatory Unbundling and Irreversible Investment in Telecom Networks, NBER
Working Papers No. 10287, National Bureau of Economic Research, February, Cambridge,
MA.
Pirani, S. (2011), Liberalisation Heralds Change in the Gas Market, Russian Analytical Digest N0. 100,
July 26, 10-15.
Pollitt, M. (2008), The arguments for and against ownership unbundling of energy transmission
networks, Energy Policy 36(2), 704-713.
RAMBOLL (2008), Study on natural gas storage in the EU, Final Report, October, Esbjerg.
Policy paper 118
Rammerstorfer, M. and M. Stronzik (2011), Processing weather and storage information – the case of
the European natural gas market, Mimeo, Vienna University of Economics and Business,
Institute for Finance, Banking and Insurance, Vienna.
Rüster, S. (2010), Recent Dynamics in the Global Liquefied Natural Gas Industry, Working Paper No.
WP-RM 19, Dresden.
Sappington, D.E.M. (2005), On the Irrelevance of Input Prices for Make-or-Buy Decisions, American
Economic Review 95(5), 1631-1638.
Spence, A.M. (1975), Monopoly, Quality, and Regulation, Bell Journal of Economics 6(2), 417-429.
Steiner, F. (2001), Regulation, Industry Structure and Performance in the Electricity Supply Industry,
OECD Economic Studies 32, 143-182.
Stern (2002), Security of European natural gas supplies, Chatham House Report, July, Royal Institute
of International Affairs, London.
Stern, J. (2007a), Is There a Rationale for the Continuing Link to Oil Product Prices in Continental
European Long-Term Gas Contracts?, Working Paper No. NG 19, Oxford Institute for Energy
Studies, April, Oxford.
Stern, J. (2007b), The New Security Environment for European Gas: Worsening Geopolitics and
Increasing Global Competition for LNG, Updated version of Working Paper No. NG 15, Oxford
Institute for Energy Studies, December, Oxford.
Stern, J. (2009), Continental European Long-Term Gas Contracts: is a transition away from oil
product-linked pricing inevitable and imminent?, Working Paper No. NG 34, Oxford Institute
for Energy Studies, September, Oxford.
Stronzik, M., M. Rammerstorfer and A. Neumann (2009), Does the European natural gas market pass
the competitive benchmark of the theory of storage? Indirect tests for three major trading
points, Energy Policy 37(12), 5432-5439.
Talus, K. (2011), Long-term natural gas contracts and antitrust law in the European Union and the
United States, Journal of World Energy Law and Business 4(3), 260-315.
Ungern-Sternberg, T. von (1996), Countervailing power revisited, International Journal of Industrial
Organization 14, 507-520.
Vickers, J. (1995), Competition and Regulation in Vertically Related Markets, The Review of Economic
Studies 62(1), 1-17.
Victor, N. M. (2008), Gazprom: Gas Giant Under Strain, Working Paper No. 71, January, Program on
Energy and Sustainable Development, Stanford University, Stanford, CA.
Williamson, O.E. (1979), Transaction Costs Economics: The Governance of Contractual Relations,
Journal of Law and Economics 22, 233-260.
Williamson, O.E. (1987), The Economic Institutions of Capitalism – Firms, Markets, Relational
Contracting, Paperback Edition, New York.
Woroch, G.A. (2004), Open Access Rules and Equilibrium Broadband Deployment, Madden, G. and
R. Cooper (eds.), Frontiers of Broadband, Electronic and Mobile Commerce, Heidelberg, 221-
246.
Wright, J. (2002), Access pricing under competition: An application to cellular networks, Journal of
Industrial Economics 50, 289-315.
Declaration 119
Declaration
I, Marcus Stronzik, confirm that this dissertation is my original work and a product of my own
research endeavours. It includes outcome of work done in collaboration with Christian
Growitsch, Rabindra Nepal, Anne Neumann and Margarethe Rammerstorfer as declared in
the preface. All sentences, passages or illustrations quoted in this dissertation from other
people's work have been specifically acknowledged by clear cross-referencing. A full list of
the references employed has been included.
I further declare that this dissertation is not and has not been submitted at any other
university for review.
Marcus Stronzik
Date of Submission: April 11, 2012