You are on page 1of 16

Dr. M.

Hamada
Commodity Price Risk II
University of Geneva

The role of storage in European Gas Market


Panagiotis Lianos
The purpose of this paper is to analyze and target a viable gas storage investment in Europe. In
order to assess the viability of such a project, overviews of the gas and storage markets are essential.
Moreover, the impact of gas storage on the power market in Europe is examined since the power
generators, along with industrial users account for more than 2/3 of European gas demand (IEA, 2007)1

After having analyzed the dynamics of supply and demand, the analysis will follow a geographical
breakdown modeled after the main price hubs in Europe. Later the flexibility element of gas storage is
discussed and the position of storage in the gas value chain is presented. Derived demand for gas
coming from power generators is also analyzed leading to conclusions about the threefold relationship
between gas storage, gas prices and power prices. Lastly, an extensive analysis of the potential viability
of a gas storage investment is made. Alternative valuation methods are assessed and conclusions are
drawn for the needed considerations for an investment decision.

When analyzing demand, emphasis is placed on the effect of temperatures to gas price
fluctuations. Households tend to strongly influence gas prices in a direct and indirect manner.
Effectively, during the low temperature winter months, households consume electricity for heating.
With gas being the main input for electricity generation derived demand from households tends to
create the seasonality observed in gas prices. Simply put, gas is cheaper during hot summer months and
more expensive during the colder winter months. However, as the power market is involved in this
assessment, it is important to note that power has its own capacity constraints and fundamentals that
dictate its prices. It is for this reason that a close examination of the spark spread is made since the
transformation cost from gas to power is never constant and it can even take negative prices! (In which
cases power generators may consider purchasing power to meet their deliveries rather than producing it
themselves).

Landscape of gas storages in Europe

Before talking about gas storages in Europe, it is important to briefly present the evolution of its
pricing methodology. Gas in Europe is priced based on hubs. Since 2005, hubs in continental Europe “are
used as the basis for gas supply contracts, leading to the development of markets where gas is priced
based on competition, rather than priced based on alternative fuels”2 (Harris, Jackson; 2005) As
historically, gas was found and extracted along with crude oil, it was considered sensible to be priced on
the highly liquid oil contracts. Later, the price valuation for gas was made in a similar manner as the
multiples approach used to value financial assets. By looking at the alternative inputs to gas as fuel oil or
gas oil, gas was trading at a premium or discount to such oil products. Eventually, gas markets were
deregulated in enough European countries and alongside with the building of pipeline and storage
facilities hubs were created. Hubs serve the purpose of providing a ‘fair’ pricing reference that would
represent local market dynamics while being broad enough to sustain a certain level of liquidity.

1
http://www.iea.org/stats/gasdata.asp?COUNTRY_CODE=30
2
http://www.kingstonenergy.com/eugas0805.pdf
Part of the gas infrastructure in Europe involves pipelines and storage facilities. Exhibit 1
provides an overview of the type of storages in Europe as well as the investments that have been made
in storage over the past 4 years.

EXHIBIT 1
T yp e o f s t o r a g e p r o j e c ts

Aquifer
8%
LNG Peak Shaving
2%

Salt Cavity
26%

Reservoir
64%

St o r a g e i n v es t m e n t b y c o u n t r y

20000

15000

Jul-07
Jan-08
Mcm

Jun-08
10000 Feb-09
Jan-10

5000

0
Y

IA
C

IA

IN

K
M

IA

LY

IA
IA

D
IA

L
A
N
R

U
C

D
LI

N
IU

B
N
R
R

A
A

G
A
A

LA
B

R
T

P
T

A
LG

IT

U
M
M

G
A

LA
U

LA
S

S
U

T
R

N
R
N
P
U

S
H
L

O
E

R
P
F

U
E
E

E
A

T
B

R
E

O
G
R

H
B

LI

P
H

T
C

E
E

N
Z
C
As the graph depicts, the UK is the leader in gas storage investments. However leaving the UK aside, it is
noticeable that European countries have been making significant and increasing investments in
developing new gas storages. Data from the portal for European gas infrastructure (gie.eu.com) suggests
that by 2012 an additional 47,766 Mcm will be added in storage. By 2015 this is expected to increase by
another 5% and after 2015 an additional 40% or 20,210Mcm of storage are expected to be constructed.
This fact alone witnesses the increasing importance of gas storage in Europe.

By looking at Exhibit 2 it is visible that European gas storage injections and withdrawals tend to
follow a similar pattern across the continent. Seasons of high storage utilization tend to be the cold
winter months and the warm summer months, essentially periods with extreme temperatures.
Nevertheless, the high utilization rates occur for entirely different reasons. The high utilization during
summer has to do with the cheaper summer gas prices where effectively traders, industrial users and
power generators are trying to lock in the highest calendar spread possible. Whereas the same market
participants, store during the high consumption period of winter weeks in order to increase their ability
to optimize their portfolios when they need to meet gas delivery obligations.

Exhibit 2

Regional Storage 2010


120%

100%
Capacity Utilization

80%

NW Europe
60%
North Europe
SE Europe
40%
S Europe
20%

0%
2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42
Week

How gas storage can help increase flexibility in the gas market

Fundamentally, gas storage in Europe assists in adjusting the supply push and the demand pulls
that can be created in regional markets. Briefly, when there is oversupply of gas in a particular market
the excess supply can be stored for future consumption. As European gas is priced on hubs, regional
economics should theoretically dictate price levels. One could argue that oversupply should drive prices
down. Nevertheless, in the case of existence of multiple storage facilities a ‘bidding war’ could
potentially be created between capacity holders aiming to purchase cheap gas. Such a war would create
a new economic dimension which would not allow for dramatic price drops. It is for exactly this reason
that the existence of gas storages tends to decrease the volatility of spot prices. It is important to note
that referral to spot prices covers the day ahead as well as the front month contracts. Since spot and
month ahead prices tend to react to economics and follow similar price patterns, day ahead contracts
are used as spot indicators in this study.

Evidently, Exhibit 3 demonstrates that prices within a country can differ from one location to
another, however the price deviation is limited when compared to cross border price differentials.
Probably the most significant element that created the need for several hubs in different locations are
the several infrastructural bottlenecks that exist.

Exhibit 3

Spot PEG prices


35
30
25
EUR/MWh

20
15
South
10
North
5
0
13-May-08

13-May-09

13-May-10
13-Sep-08
13-Nov-07
13-Jan-08
13-Mar-08

13-Nov-08
13-Jan-09
13-Mar-09

13-Nov-09
13-Jan-10
13-Mar-10

13-Jul-10
13-Jul-07
13-Sep-07

13-Jul-08

13-Jul-09
13-Sep-09

By looking at the spot price fluctuations over the past 3 years in the major European gas hubs (Exhibit 4)
it is clear that each region (represented by a hub) tends to react to different economics.
Exhibit 4

Major EU hub spot


35 2.5

30
2
25

GBP/Therm
EUR/MWh

20 1.5 TTF

15 North
1
ZBT
10
0.5
5

0 0
12-Jul-06 12-Jul-07 12-Jul-08 12-Jul-09 12-Jul-10

Restricted pipeline capacity indicates a limitation in how much gas can be delivered in certain
geographical locations. This is why some countries, as Germany and France, can have more than one
pricing hubs, representing the existence of different s&d economics. Exhibit 5 graphically indicates how
storage is positioned in the value chain and how it can increase flexibility in the supply push.

Exhibit 5
To expand on the argument raised earlier regarding the different reasons why storage is mostly
used during extreme temperature periods one could say the following. During the summer months, the
objective is to identify the highest calendar spread. In other words, asses the forward curve behavior
and try to identify the time when the Winter-Summer spread is the highest. This way, the expected cash
flow to be generated from storing would be the highest as well. During winter months however, the
objective of holding gas in storage is two folded. Firstly, it has to do with the availability of a cushion for
times of need, when the infrastructural limitations would not be enough to cover demand. To avoid
potential defaults on deliveries, withdrawals can be made to meet these obligations. The second
objective has to do with optimizing the use of gas storage.

Gas is no exception to the general commodity price behavior of higher volatility in the near
months than the further away. Since the spot prices are very volatile, companies that store gas use the
stored gas for when the volatility of gas prices is against their favor. So companies that consume the gas,
as large industrials and power generators can buy gas of the market until a certain price and can use the
stored gas when they consider that buying spot is uneconomical. On the other hand, storage users as
trading companies would be on the opposite spectrum, keeping gas in storage until spot prices exceed a
certain level where they would try to make as many spot sales as possible.

Gas storage is indeed a flexibility instrument that allows sellers and buyers to hedge their
positions by enabling them to use their stored supplies when they deem optimal. With an expectation
for high increased demand for gas across Europe, new projects will allow for that flexibility to expand. If
there were no new investments in storage, existing storage capacities would not be enough to cover
demand in the years to come, and this would affect the behavior of gas prices and the forward curve.

Impact of gas storage on gas prices

As discussed earlier, gas storage allows for decreasing volatility in spot deliveries for gas.
Evidentially, the trend seems to be that gas injections tend to drive prices up as they create shortages of
supplies in the market. On the other hand, withdrawals allow for additional supplies that would allow
for smoother volatility and lowered prices. Indicatively, Exhibit NUM8 shows the reaction of gas prices
for different levels of aggregate storage balances. For France, the average price of the two pricing hubs
(PEG South and PEG North) is used and presented in relation to the sum of stored quantities in all the
country. Similarly for the Netherlands, the TTF day ahead contract is used in relation to all gas stored in
the country. The reason for using spot contracts is that as mentioned before, the spot prices are the
ones affected the most by storage. This is because optimization of storage is not done at the further out
months of the curve but at the highly volatile spot and near deliveries.

As the Exhibit 6 indicates, during high lower spot prices, storage balances tend to remain high
and withdrawals are made when spot prices increase.
Exhibit 6

Netherlands: Storage-TTF
24 1700
22 1500
20 1300
18 1100
16 900 Eurohub Storage
14 700
TTF
12 500
10 300

France: Storage-PEG Avg


23 7600

21 6600

19 5600

17 4600
PEG Storage
15 3600
S/N PEG Avg
13 2600

11 1600

9 600
1-Jan-10 1-Feb-10 1-Mar-10 1-Apr-10 1-May-10 1-Jun-10 1-Jul-10

Impact of gas storage on power market

One of the main uses of gas in Europe is for power generation. Power in continental Europe is
traded as baseload, peak and off peak. Typically, baseload prices represent full days, peak are usually
8am to 8pm and off-peak 8pm to 8am; all of which represent consumption patterns. As Exhibit 7
indicates, spot baseload, peak and off-peak power prices tend to behave similarly and for this reason
baseload shall be used in the further computations.
Exhibit 7

French Peak-Off Peak-Base


140
120
100
80
60 RTE-base
40
20 RTE PEAK
0 RTEOFF

12-Apr-10
12-Apr-09

12-Dec-09
12-Mar-09

12-May-09

12-May-10
12-Aug-09

12-Oct-09

12-Mar-10
12-Jan-09

12-Jun-09

12-Jun-10
12-Nov-09

12-Jan-10
12-Feb-09

12-Jul-09

12-Sep-09

12-Feb-10

12-Jul-10
As mentioned earlier, the market power has its own economics and the single most important
element dictating their appetite for gas is the margin they can make from transforming spot gas into
spot power. This differential is the spark spread and generally, the higher spark spreads signal increased
demand for gas. As an indication of how variable the spark spread can be, Exhibit 8 shows the spark
spread for the French market using baseload power and average PEG prices. It is interesting to note that
the spread can take negative values. This is a strong indication of how gas storage would be useful for
the power generator. If the generator had gas in storage, he would not be forced to operate at a loss but
continue enjoying a margin. Of course, there are logistical issues with deliverability and additional
pipeline constraints to bring the gas from storage to the plant when needed, but in principle, storage
functions as a risk management instrument for power generators.
Exhibit 8

Spark Spread: RTE-Avg PEG


100

80

60

40
Spark
20

12-Dec-09
12-Jun-09

12-Jun-10
12-Apr-09
12-May-09

12-Jan-10

12-Apr-10
12-May-10
12-Oct-09
12-Jan-09

12-Mar-09

12-Aug-09

12-Mar-10
12-Feb-09

12-Jul-09

12-Sep-09

12-Nov-09

12-Feb-10

12-Jul-10
-20

It is therefore interesting to observe the relationship between a country’s aggregate storage


balance and the power prices. Exhibit 9 indicates that there is no direct relationship between power
prices and storage levels. Using France as an example once again, it is evident that power prices do not
have a tendency to follow gas storage levels. This supports the argument raised earlier that power and
gas markets function separately. There is however a need to acknowledge that the two markets share
many demand drivers, for the most part temperature levels tend to have a similar effect on both power
and gas. However, the market responses are different. Power producers need to respond immediately
to increases in demand and they do not have the luxury of storage. Moreover, their sole alternative to
purchasing gas in case of need is to purchase power from other markets. Lastly, it is worth mentioning
that even though power is traded in time blocks, the concept of hubs is not as developed as in the gas
markets. In power there is only one reference price per country hence not capturing infrastructural
capacity limitations.

The indirect relationship between gas storage and power prices can be observed by looking at
the value chain for gas. As aforementioned, gas storage relates to gas prices as do power prices.
However no direct conclusion can be drawn for power prices by gas storage levels and vice versa.
Exhibit 9

France: Storage-RTE Baseload


7000 100
6000
80
5000
4000 60
PEG Storage
3000 40 RTE Base
2000
20
1000
0 0
1-Jan-10 1-Feb-10 1-Mar-10 1-Apr-10 1-May-10 1-Jun-10 1-Jul-10

Put yourself in the shoes of an investor willing to take a stake in gas storage
What are the main questions that you should ask to assess the viability of your investment

In order to assess the viability of a storage investment, one would first need to identify the
intrinsic value of the storage in question. In order to define intrinsic value it is important to broadly
define the different transformations that gas can take. Exhibit 10 shows that as long as storage costs are
lower than the calendar spread between a near month contract and a further out month contract, there
is a potential arbitrage opportunity. However, when leasing or making a toll agreement for a gas storage
facility, the lease price or toll fee tends to factor in this spread. Effectively, a positive calendar spread
stands for a market in contango whereas the opposite scenario would indicate a market in
backwardation. The intrinsic value of storage is purely derived from the forward curve. For example
today’s intrinsic value of a storage at North Germany (which is priced after NetConnect hub) for 6
months would be Dec-10 contract – Aug-10 contract which would be EUR 2.85/MWh. Exhibit 11
provides the forward curves for main European hubs priced in late July.

To be more precise, gas storage has an intrinsic and an extrinsic value. The intrinsic value of
storage is the aforementioned calendar spread. This value can be ‘locked in’ when selling a forward
calendar spread. On the other hand, the much more volatile nature of the spot prices creates an
extrinsic value for gas storages. Effectively, the extrinsic value is the value derived from the
management of spot sales and purchases to be injected and withdrawn to and from the storage facility.
By using econometric models and a dynamic programming approach, an attempt to optimize this value
is made. In other words, the dynamic programmer will look at spot prices and try to purchase and inject
additional gas in storage when he considers spot gas is relatively cheap. Also, using such an approach,
when for some reason there are unexpected demand shocks that drive spot prices up for short time
periods, the programmer will attempt to withdraw as much gas as possible to take advantage of the
sudden increase in the storage’s extrinsic value.

Exhibit 10

Transportation(space transformation):
-Pa: Price at A (sale location A)
-Pb: Price at B (purchase location B)
-Cp: Cost of pipelin use
Opportunity when Pa-Pb>Cp

Storage (time transformation)


-Pt: Price at near month t
-Pt+n: Price at further month t+n
-Cs: Cost of Storage
Opportunity when Pt+n-Pt> Cs

Processing (Form transformation)


-Gt: Gas price at time t
-Pt: power price at t (assuming instantenuous transformation)
-Ct: Cost of tranformation (including potential tolling costs)
Opportunity when Ptt-Gt>Ct

Exhibit 11

EU Hubs Forward Curves


31
29
27
25 TTF
EUR/MWh

23 NCG

21 PEG NORTH

19 PEG South

17
15
6-Aug-10 6-Aug-11 6-Aug-12 6-Aug-13 6-Aug-14 6-Aug-15 6-Aug-16
Notably, in a gas valuation, the two measures of value have different characteristics. On one hand, the
intrinsic value can be known at anytime and is simply derived from the calendar spread. Whereas the
extrinsic value cannot be easily priced as the spot market is a lot more volatile and making assumptions
on spot price movements is a lot more difficult and results vary more than for a calendar spread.

As a result, an investor looking to purchase a gas storage would aim for the best price. In order
to find that price the investor would want to look at when the market tends to be in contango and when
in backwardation. As discussed before, gas is a highly seasonal commodity. This characteristic derives
from its demand and further creates seasonality in prices. As Exhibit 12 shows, the 1-month, 3-month
and 6-month spreads tend to be almost identical for different hubs. This is a sign on the integrated
infrastructure in continental Europe where open markets allow for easy cross border trade to cover any
potential imbalances that tend to arise in one geographical area compared to another. Moreover, it is
evident that the 1 month spread is positive during the Aug-Feb season; the 3 month during the Sep-Mar
season and the 6 month during the Nov-Jun season. In other words, this is supporting evidence for the
market being in contango during high demand periods and in backwardation during low demand
periods.

Exhibit 12

1 month spread
3

2
TTF
1
NCG
0
PEG NORTH
May-11

May-12

May-13

Jan-14
May-14

May-15

May-16
Jan-11

Jan-12

Jan-13

Jan-15

Jan-16

Sep-16
Sep-10

Sep-11

Sep-12

Sep-13

Sep-14

Sep-15

-1

-2

3 month spread
4
3
2
TTF
1
NCG
0
PEG NORTH
Mar-12
Mar-11

Mar-13

Nov-13
Mar-14

Mar-15
Jul-15

Mar-16
Nov-10

Jul-11
Nov-11

Jul-12
Nov-12

Jul-13

Jul-14
Nov-14

Nov-15

Jul-16
Nov-16

-1
-2
-3
6 month spread
5

0 TTF
Feb-11
Jun-11
Oct-11
Feb-12
Jun-12
Oct-12
Feb-13
NCG

Jun-13
Oct-13
Feb-14
Jun-14
-5

Oct-14
TTF

Feb-15
Jun-15
Oct-15

Feb-16

Jun-16
PEG NORTH

Oct-16
Financial valuations are typically done by either comparing the asset value to prices of other
values with similar characteristics or by looking at the discounted present value of the expected future
cash flows from that asset. However, none of these two approaches would work for the purposes of a
valuation of a gas storage facility.

The multiples approach, comparing the storage in question to other storage values and
multiplying by a discount or a premium would fail for the following reasons. First of all, as seen at the
beginning of the study, there are various kinds of storages so in order to be precise one would need to
compare the asset to similar facility types which would decrease available comparable assets by a lot.
Furthermore, as for the hubs, different economics apply to different geographical locations for the
storage. Quantifying the infrastructural limitations to create a premium or a discount would not be a
highly subjective task which would lead to constant disagreement over the storage purchaser and seller.
But most importantly, there are only few acquisitions of gas storages in Europe and for the most part,
transaction prices are not transparent. For all the above reasons, such an approach would be rejected.

Discounting expected future cash flows would be an interesting approach, but still highly
subjective. As already discussed, quantifying the intrinsic value of the storage could be done by looking
at the forward curve. However, the forward curve would only go out for 6 years and hence assumptions
would need to be made for the behavior of the curve further out as gas storage has a much longer
investment horizon than 6 years. The issue then would become the identification of assumptions to be
made for the evolution of the forward curve further out. Additionally, provided that these assumptions
are agreed upon, further assumption would need to be made for the pricing of the extrinsic value of the
storage. Such assumptions would be highly correlated with the way dynamic programing is made.
Nevertheless, dynamic programming is done on a daily basis with live price data feeds. To add to the
complications, to value the extrinsic value different sets of price levels would need to be set and
multiplied by expected inventory levels to come up with potential expected cash flows. It is easily
evident that the variables are many and highly volatile hence leading to rejecting such an approach.

One could now observe that the issues and problems in question are similar to the ones for
pricing an option. As a result using a Black Scholes option pricing model, one would be able to identify
and quantify the expected cash flows to the most accurate value possible.Before treating a gas storage
facility as an option one would need to assess the strategies that a storage owner would follow to
maximize their income from owning such a facility.

A gas storage owner would aim to follow one of two strategies. Either to sell the contango
spread or sell puts on that same spread. Using TTF spreads as an example Exhibit 13 one can note that
the owner would want to lock in the spreads at their highest values depending on his planning of when
to sell the gas. It is visible that the longer the calendar spread, the higher the potential income from that
spread. This would be the case if the owner decided to sell the forward spread.

Alternatively, an owner can sell put options of gas and limit his downside potential by covering
his delivery obligations with gas from the storage when the option is out of the money. Assuming that
the owner repeats this strategy for the life of the investment, a value could be created by the aggregate
of the premiums that would be collected during that time. In order to identify at what strike price the
owner would sell these options, a Magrabi equation would be used for different calendar spreads on the
forward curve to find the different strikes at which the option could be sold. Discounting the sum of
these premiums to a present value would potentially provide an accurate figure of the value of the
storage.

Exhibit 13

TTF Calendar Spreads


5
4
3
2
1 1 month
0 3 month
Feb-11

Feb-12

Feb-13

Feb-14

Feb-15

Feb-16
Aug-13

Aug-14
May-11

May-12

May-13

May-14

May-15

May-16
Aug-11

Aug-12

Aug-15

Aug-16
Nov-11

Nov-12

Nov-13

Nov-14

Nov-15

Nov-16

-1 6 month
-2
-3
-4
-5
To conclude, it is evident that valuing a storage is highly related to variables as gas and power
prices, infrastructure availability, possibility of substitution and also regulatory changes. An investor
would have to consider at least all the above before moving forward with making a decision. Clearly,
storages are healthy investments in sustainable seasonal contangos year over year but still involve a
high element of risk in their pricing as overpaying would create need for alternative use of the facility.
Such alternatives would include toll and/or agreements that would potentially capture the fluctuation of
market dynamics from the original expectation.

You might also like