This post considers the latest studies on European gas prices, a short history of the market and effective hedging.
A common feature of energy prices is that conditional mean and volatility are driven by seasonal trends due to weather, demand, and storage level seasonalities. Recent work by Martínez and Torró extends the work done by Ederington and Salas (2008) and considers seasonalities in mean and volatility when minimum variance hedge ratios are computed.
Their results show that hedging effectiveness is much higher when the seasonal pattern in spot price changes is approximated with lagged values of the basis (futures price minus spot price). This fact remains true for short (a week) and long (one, three and six months) hedging periods. Furthermore, volatility of weekly price changes also has a seasonal pattern and is higher in winter than in summer. Seasonal hedging strategies, linear regression based strategies, or even a naïve position, perform better than more sophisticated statistical methods.
The foundations of a European Gas market
Liberalisation of the natural gas market began in the UK with the 1995 Gas Act. The following year, the Network Code created the National Balancing Point (NBP) hub enabling third party access to the British gas network. Important changes in the contractual conditions and trading system were introduced in the British natural gas market in2004–2005 after the Enron and TFX collapse, resulting in a new regulation, the Uniform Network Code in 2005. The equivalent of NBP’97 is ZBT’2000 in Zeebrugge and EFET for TTF.
The UK natural gas market is the most liquid market in Europe. The vast majority of gas contracts are over-the-counter but the regulated futures market is growing in importance. The futures British gas market is operated by InterContinental Exchange (ICE) Europe. The ICE natural gas futures contract for NBP was launched in January 1997 and has become the benchmark for natural gas trading in Britain and in continental Europe. Continental gas markets were developed following the path the British had marked. The UK and continentalEurope are linked through The Interconnector — a gas pipeline that connects the UK gas entry point at Bacton to the Belgian port of Zeebrugge (ZEE henceforth). It has been open since 1998 and enables the flow of gas between British and continental markets. Since its launch, UK prices have converged progressively to continental prices (Heather, 2012).
The Title Transfer Facility (TTF) was established in 2003 in the Netherlands. It is the virtual trading point for the Dutch natural gas market. It is the most developed natural gas market in continentalEurope, comparable to NBP for hedging and balancing purposes. TTF is intended to be the leading gas hub in Europe, because of its location in the heart of Europe, LNG import facilities, and storage capacity. Futures contracts on TTF and ZEE are also negotiated on the Intercontinental Exchange (ICE).
European Gas – Trading and Hedging
There are several questions tackled in the literature on risk management in energy markets that relate to the special case of European natural gas markets. Beatriz and Hipòlit (2015) highlight the existence of seasonal patterns in first and second moments of price returns and the implications for futures hedging. In a similar context, Chang et al. (2010) found that futures hedging effectiveness of a covariance model specification can change depending on the market trend (bull/bear) in energy markets (oil and gasoline).
The influence of seasonality in energy prices for hedging purposes has also been studied in Suenaga et al. (2008). In their opinion, seasonal hedging turns out to be quite discretional under strong seasonality in prices. Long spot positions from the peak to off-peak price season would be senseless and, they argue, it is better not to have these positions.
Nevertheless, Beatriz and Hipòlit (2015) argued that risk measures should take into account the predictable seasonal price movements and thus attempted to include the unexpected spot price movement both in the hedging ratio computation and in the measurement of hedging effectiveness. Furthermore, the influence of energy variance seasonality on futures hedging performance had not yet been explored until they checked if hedge ratios and their effectiveness have a significant seasonal pattern.
Finally, as an aside, in the last few years non-conventional shale gas has become abundant and represents a downward pressure on winter prices. Furthermore, the increased number of cooling systems and the growing use of natural gas as a fuel are raising summer prices (Henaff et al., 2013). Both effects reduce price seasonality on the US market. It would be interesting to check if seasonality in mean and volatility persists in European natural gas markets.
European Gas prices
Natural gas spot and futures prices can be directly obtained from Platts and the ICE, respectively. The spot price is computed and published daily for delivery next working day after assignment and is the reference for derivative contracts traded at the ICE and those contracts traded OTC in the respective hub. There is a wide range of natural gas derivative contracts (forward, futures, and options) traded at the ICE with the most important of the regulated contracts being monthly futures, especially the front month contract (the most liquid of all the traded contracts). The vast majority of contracts currently traded are OTC contracts, but futures contracts are becoming more important over time as markets become more liquid and more reliable.
Futures and forwards contracts have coexisted from the beginning, but the majority of trades are through OTC contracts. Futures negotiated at the ICE are increasing its importance and liquidity over time and they represent more than one-third of all gas negotiated at NBP. The ICE trades monthly, quarterly, seasonal, and yearly futures contracts in the three markets (the monthly futures being the most liquid).
The most relevant jumps correspond to events mostly related with geopolitics: the dispute betweenRussia and Ukraine about the price of gas and transit combined with abnormally cold weather (3 March 2005, 22 November 2005, January2009, February 2012) and the Libyan civil war (spring 2011). But the most dramatic shortcoming and peak was during February and March2006 when a cold spell was combined with a fire at the Rough natural gas storage facilities in the North Sea — preventing access to nearly over 80% of total UK storage just as withdrawals from storage were about to begin.
The statistical behavior of futures and spot differences has some significant discrepancies that might be critical obstacles to overcome in order to design a successful hedging strategy. The two most insightful results are that futures have a declining pattern as maturity approaches,and that spot prices are more volatile than futures prices. This disparity produces a lower correlation than usual for linking futures and the spot position to hedge.
Seasonality in basis and volatility
Basis (futures price minus the spot price) has a strong seasonal pattern when convenience yield, weather,and storage costs vary during the year (Wei and Zhu, 2006). Basis is positive in winter and negative in summer. In winter, demand is great and so storage levels decrease and storage costs increase (positive convenience yield), producing a positive basis. In summer, demand for natural gas is lower because of warm weather and storage prices decrease and storage levels increase (negative convenience yield) and the combination of these effects results in a negative basis. Basis and price volatility have a similar seasonal pattern.
Basis and returns volatility are high in winter and low in summer. Furthermore, the basis contains information of those variables (storage levels, weather, demand, and risk premiums) that reflect uncertainty in the natural spot-futures markets necessary to obtain futures prices. Finally, in contrast to spot and price levels where jumps are frequently found, the basis is more stable as the liaison between spot and futures prices is constrained by the arbitrage arguments.
However, the basis has less ability to forecast futures price changes. These results coincide with the Ederington and Salas (2008) approach where spot price changes are partially predictable; but futures prices results agree with the martingale hypothesis in most cases.
A common feature of natural gas prices is that spot price changes are partially predictable due to weather, demand, and storage level seasonalities. Further to this, the volatility of natural gas prices is seasonal. In winter, the active storage management is less flexible and price jump buffers are more difficult than in summer. Moreover, higher marginal cost production, demand inelasticity, and winter weather shocks trigger price jumps that produce a higher volatility than in summer.
Ederington and Salas (2008) adapted the standard minimum variance hedge ratio approach to the case where spot price changes are partially predictable. In this context, they show that the riskiness of the spot position is overestimated, the achievable risk reduction underestimated, and more efficient estimates of the hedge ratios are obtained. In their 2008 paper, they proposed to use the basis (futures price minus the spot price) at the beginning of the hedge as the information variable to approximate the expected spot price change. If futures prices are unbiased predictors of futures spot price, the basis will be a measure of the expected change in the spot price until maturity (Fama and French, 1987).
Recent work by Martinez and Torro (and the basis of this post) found that previous studies had overestimated the unexpected shocks in spot prices as a large part of these shocks can be partially anticipated using the information contained in the basis (between 10% and 30%).
Since a strong seasonality effect also exists in the volatility of spot and futures price returns – which have been significantly higher in winter than in summer – Martinez and Torro also found that seasonal hedging strategies, linear regression based strategies, or even a naïve position, perform better than more sophisticated statistical methods.
Consequently, the riskiness of the spot position is typically overestimated and the achievable risk reduction underestimated. This poor effectiveness is enhanced by the features of most energy commodity prices; The special statistical features of natural gas prices (specifically their high volatility, kurtosis, and the high volatility of the basis) can produce very wrong computed risk reductions.
- Ederington, L.H., Salas, J.M., 2008. Minimum variance hedging when spot price changes are partially predictable. J. Bank. Financ. 32, 654–663
- Fama, E.F., French, K.R., 1987. Commodity futures prices: some evidence on forecast power, premiums, and the theory of storage. J. Bus. 60, 55–74.
- Heather, P., 2010. The evolution and functioning of the traded gas market in Britain. Oxford Institute For Energy Studies, NG 44
- Heather, P., 2012. Continental european gas hubs: are they fit for purpose? Oxford Institute For Energy Studies, NG 63.
- Martínez, Beatriz, Torró, Hipòlit. European natural gas seasonal effects on futures hedging, Energy Economics (2015)
- Chang, C.Y., Lai, J.Y., Chuang, I.Y., 2010. Futures hedging effectiveness under segmentation of bear/bull energymarkets. Energy Econ. 32, 442–449
- Suenaga, H., Smith, A., Williams, J., 2008. Volatility dynamics of NYMEX natural gas fu- tures prices. J. Futur. Mark. 28, 438–463.