"Energy Only" – der ewi-Blog

As pointed out by previous research by the Oxford Institute for Energy Studies (OIES), e.g. Petrovich (2015)¹, gas prices in Europe have been overall well aligned in recent years. However, there have been bottlenecks in the European gas infrastructure in 2014 that led to a decoupling of the prices between specific gas hubs. In a joint analysis with OIES, those historically observed bottlenecks are analyzed with ewi ER&S’s TIGER model in order to understand the drivers of the bottlenecks. The study was published in September 2016 both by OIES and ewi ER&S with the title “European gas grid through the eye of the TIGER: investigating bottlenecks in pipeline flows by modelling history”.

The following bottlenecks/price de-couplings are especially in the focus of the analysis:

  • Bottleneck between the French hubs PEG Nord and PEG Sud
  • Bottleneck between the German hub NCG and the Austrian hub CEGH
  • Bottleneck between the German hub NCG and the Italian hub PSV

TIGER is a cost minimizing gas market model with a high resolution of the European gas infrastructure. Within the analysis, the model is calibrated to reproduce the gas flows of the year 2014 based on publicly available data. Daily historic demand data for the major European gas markets, daily historic LNG send-outs in Italy and France, daily major import flows to Europe, e.g. flows from Ukraine to Slovakia, and yearly production data are used as an input to the model. Tariffs for shipping gas across borders are based on ACER² implicitly assuming “steady” flows. The output of the model is the benchmark of the least cost gas flows within the European grid.

If a bottleneck exists not only in the historic flow data, but also in the cost optimal simulation, this means that the bottleneck has a physical nature. If, however, persistent differences between the modelled flows and the historic flows occur, this can be a hint for suboptimal capacity usage in the reality. The study aims to explore the drivers of a deviation between simulation and reality.

In the case of the French hubs, the model returns a congestion between the hubs when LNG is scarce in Southern France. A good match between modelled and real flows is observed in France indicating the physical nature of the congestion. Therefore, further transmission capacity within France will be needed in order to create a single price for natural gas in France.

As can be seen in Figure 1, the TIGER simulation exploits capacity on the NCG-Switzerland-PSV pipeline system more heavily compared to actual historic flows. In reality, the utilization rate of this pipeline system exceeded 80% only in September and October 2014. The simulation, however, returns a 100% utilization rate from April to November. This result indicates that non-physical barriers to trade could be the driving force behind the persistent premium paid for gas in Italy compared to the markets in North Western Europe. The price premium (about 1.6 €/MWh) appears to be higher than the estimated cost of transporting gas from the German NCG to the Italian market zone. Evidence for reasons underlying the non-physical bottleneck cannot be given due to the confidentiality of contractual agreements. However, it can be supposed that the non-physical bottleneck is related to contractual agreements, especially to long-term capacity bookings. It is estimated that more than 80% of the transmission capacity is booked long term by ENI³. It may be difficult to make this capacity available to other shippers if ENI does not use/nominate it, since the transmission capacity allocation in Switzerland is conducted by private auctions and EU rules on contractual congestion are not mandatory.
gas_grid_figure1Figure 1: Simulated and historical utilization rate from NCG to Switzerland in 2014 (%)
(Source: EWI TIGER Simulation, ENTSO-G Transparency Platform, SRG)

Figure 2 illustrates that the simulated gas flows from Germany to Austria at the Oberkappel interconnection point (IP) are lower than the historically observed flows. In reality, the capacity was almost fully used from April 2014 on. The simulation shows a high utilization only in October 2014. This difference between reality and simulation can be explained against the background of the increased NCG-Switzerland-PSV utilization in the model. By shipping gas through Switzerland, the model reduces the need for transits from Germany via Austria to Italy. In a sensitivity analysis, the NCG-Switzerland-PSV capacity is lowered to 70% of its full capacity in an attempt to simulate the non-physical barriers to ship gas from North West Europe down to the higher priced Italian market. As a consequence, the IP in Oberkappel would be highly utilized, as can be seen in Figure 3. With reduced Swiss capacity, the route through Austria is required to supply the Italian market. This outcome indicates that the observed physical bottleneck between Germany and Austria might disappear if the capacity in Switzerland would be used optimally.
gas_grid_figure2Figure 2: Simulated and historical flows from NCG to CEGH in 2014 (GWH)
(Source: EWI TIGER Simulation, ENTSO-G Transparency Platform)

gas_grid_figure3Figure 3: Flows from NCG to CEGH (% firm technical capacity), capping NCG-to-Switzerland transmission capacity at 70% of existing firm technical capacity (left) and without any restriction on NCG-to-Switzerland capacity (right)
(Source: EWI TIGER Simulation, ENTSO-G Transparency Platform)

The study confirms that physical and non-physical barriers to trade were still present in European natural gas markets in 2014. For the French hubs, a physical bottleneck is the likely reason for the price decoupling. In the case of NCG and PSV, a non-physical bottleneck persists that also influence the capacity situation between NCG and CEGH, because arbitrage between the gas markets of Germany and Italy can either occur by shipping gas through Switzerland or by shipping gas through Austria. In a follow up project, OIES and ewi ER&S will analyze possible future bottlenecks under different gas market scenarios, e.g. in scenarios in which more LNG volumes than today will come to European gas markets.


¹ Petrovich, B., Do we have aligned and reliable gas exchange prices in Europe?, OIES Comment Paper, April 2016. (Link)

² ACER, Annual report on the results of monitoring in the internal electricity and natural gas markets in 2014, November 2015. (Link)

³ Source: Case COMP/B-1/39.315- ENI – Commitments submitted to the European Commission, dated February 4, 2010, p. 32-33. Available at: http://ec.europa.eu/competition/antitrust/cases/dec_docs/39315/39315_2670_8.pdf. The Italian Antitrust Authority in September 2012 wrote that the share of transmission capacity on the Transitgas pipeline allocated to ENI on the basis of long term ship-or-pay agreements equals about 85-95% of total capacity (Source: http://www.agcm.it/trasp-statistiche/doc_download/3287-a440chiusura.html, p. 11)