di Jonathan Stern, Director of Gas Research dell'Oxford Institute for Energy Studies

[Estratto dal paper"Is there A Rationale for the Continuing Link to Oil Product Prices in Continental European Long-Term Gas Contract?" di Jonathan Stern (April 2007). Copyright 2007 Oxford Institute for Energy Studies]

One of the arguments of those who support continuing contractual linkage with oil is the empirical evidence from competitive gas markets that – even without such linkage – gas prices tend to follow the dynamics of oil prices.

EVIDENCE FROM THE USA
Key findings of an econometric study (Chart 1) published by the Energy Information Administration in 2006, which analysed WTI crude oil prices and Henry Hub natural gas prices over the period 1990–2006, were that (EIA 2006) “… natural gas and crude oil prices historically have had a stable relationship, despite periods where they may have appeared to decouple. The statistical evidence also supported the a priori expectation that while oil prices may influence the natural gas price, the impact of natural gas prices on the oil price is negligible… oil prices are found to influence the long run development of gas prices but are not influenced by them”. But Foss’ work casts significant doubt that this linkage will dominate the future (Foss 2007). Her assessment of the potential for gas to oil product switching in the USA suggests limitations similar to those raised above in relation to Europe. Indeed she goes further to suggest that gas prices are likely to be much more affected by power demand and price dynamics, than by oil dynamics and that this may maintain US gas prices within a corridor of $3 –6/mmBtu from which they will not be able to escape for more than a brief period for the following reasons:

• Diminishing returns from US and Canadian E&P activity associated with prices above $6/mmBtu;
• Demand elasticity and restructuring will limit opportunities for supply (such as LNG) which is not competitive;
• Structural changes in demand with power becoming the marginal application which will set the gas price. “Petroleum liquids switching capacity will remain at best constant but is more likely to decline as a share of net summer generation capacity”.

In summary, Foss suggests that the relationship between and natural gas prices going forward will be less contingent, a function of both the dominance of natural gas production and drilling over oil, and a reflection of changing patterns of gas use.

EVIDENCE FROM THE UK
A comparison of National Balancing Point (NBP) prices with Brent crude prices (and a range of other oil and oil-linked gas prices (see Chart 2) shows that the relationship has been directionally similar during the period 1999 –2006.
(This chart suggests that there has not only been a strong relationship in terms of historical prices but that forward prices demonstrate a similarly strong relationship).
Econometric work by Panagliotidis and Rutledge covering the period 1996–2003 supports this hypothesis, even prior to the establishment of the Interconnector (IUK) pipeline in 1998 through which the UK is commonly said to ‘import’ Continental European oil-linked prices. (Panagliotidis and Rutledge 2004). But it is not certain whether evidence from the pre-1998 period is still relevant to the modern UK gas market given very substantial changes in the commercial environment since that date.In particular, it may be that the potential for gas customers to switch to oil products was much greater pre-1998, which could have accounted for a closer price linkage with oil.
In a major study of UK gas prices, Wright demonstrates the complexity of these relationships and finds no relationship between front month IPE oil and gas prices, or gasoil and front month gas prices. (Wright 2006,pp.104-07 especially figures 4.21 and 4.22). For the period 1990 –2005, Huggins found changing dynamics between 5-year periods: during the years 1990 –1995 gas prices had a close relationship with oil prices; during the period 1995–2000 gas prices appeared to have no relationship to oil; and during the period 2000–2005, gas prices had a relationship but with a significant time lag. (Huggins 2006).
Futyan’s work on the price impact of the Interconnector (IUK) pipeline finds that prior to the commissioning of the pipeline and during its first year of operation when flows were very low, ‘the NBP followed a clear seasonal pattern, independent of the falling oil price. The seasonal pattern indicated gas-on-gas-competition was driving prices.’ (Futyan, p.34).
For the period 1999–2005, Futyan identified a clear, but changing, correlation between the NBP and oil prices with direct price linkage from 1999–2001 and linkage plus seasonality from 2002–2005 (Chart 3).
Also, during periods when the IUK pipeline was not operating, either due to planned maintenance or because unplanned outages, Futyan found evidence of delinking of NBP from Zeebrugge Hub gas prices (which are dominated by the long-term oil product-linked Continental European contracts). However, these periods were not long – generally one or two weeks. (Futyan, p.36).
During the winters of 2003–04, 2004–05 and 2005–06 gas price volatility showed the market had its own dynamics with actual shortage, as well as ‘market sentiment’ about shortage, creating substantial short-term upward pressure on prices.
No definitive conclusions can be drawn from these episodes, because of their short duration, but they suggest that, freed from Continental European oil linkage, the UK market may have its own dynamics, which would support Wright’s findings.
We are left with some very pertinent questions posed by Wright (2006, p.111): “…can it really be the case that if there were no oil-indexed gas contracts in Europe there would be no link between oil and gas prices? And how can the co-integration between spot oil and spot gas prices found by Panagliotidis and Rutledge be explained from within the oil-indexed contract paradigm?...which now leads us to consider another possibility – that gas traders who are also upstream producers directly link both their whole-sale market and futures/forward trades with the oil price or, alternatively use their trading power in a way which achieves the same effect”.
Thus for the British market, the evidence as to whether there would be a link between NBP and oil prices in the absence of the influence of Continental European oil-linked prices through IUK is not conclusive; but there is reason to believe that at least such a link would be significantly less strong.

EVIDENCE FROM THE CONTINENTAL EUROPE
The DG COMP Sector Investigation Report analysed its very large sample of gas contracts comparing those with prices indexed to gas with those indexed to oil products. The results are statistically less significant than those reviewed in the cases of the USA and UK above, because of the relatively short two-year time period (January 2003–December 2004), but the report (EU Energy Sector Inquiry 2006, paras 285 and 286) concluded: “Long-term contracts indexed to oil are much less volatile than those indexed to hub gas prices. In this case, hub gas prices are below oil indexed prices for the majority of the analysis period, with the exception of the period November to February. The overall level of prices in oil-indexed contracts was higher than in gas-indexed contracts for most of the period. However, the short periods when oil-indexation was cheaper were also the periods of highest volume (in winter). We suspect that, on a volume-weighted basis, there was no clear commercial advantage either way”.
The evidence from the US and British markets from the past decade is that over periods of 5 –15 years, a strong link between gas and oil prices can be empirically observed, but that short-term gas supply and demand dynamics can separate oil and gas prices for a period of months up to several years. A possible hypothesis from an analysis of the past 10 –15 years is that over such a long period, gas buyers and sellers of gas with formal contractual linkage to oil product prices may have experienced similar financial outcomes to those with indexation to Henry Hub or NBP.However, over shorter periods – perhaps lasting as long as a winter or a summer season (or even longer) – prices in contracts linked to oil products could have been fundamentally different – higher or lower – than those reflecting short-term gas supply and demand conditions. With companies increasingly required to ‘mark their contracts to market’ and report their ‘value at risk’ to the financial community on a regular basis, it is doubtful that energy utilities will be able to ignore short-term market conditions if (or when?) hub trading in Continental Europe begins to strengthen. One of the main advantages of retaining the current system of oil product price linkage is claimed to be the avoidance of short-term gas price volatility associated with competitive gas markets. The DG COMP study confirmed that over a two-year period January 2003 – December 2004, long-term contract prices were much less volatile than hub prices. Given that oil prices also tend to be volatile, long-term contracts minimise volatility by averaging prices over a period of months with a lag of several months, to produce quarterly current prices. (Quarterly long-term contract gas prices have been based typically on the average of the prices of indexed products over the previous 6–9 months with a lag of 3 months. There is anecdotal evidence that both the period of the index and the time lags are becoming shorter such that gas current prices begin to reflect more recent oil product prices).
Not everybody believes this to be an advantage. A pure economic view is that volatility produces price signals which show market players the need to, for example, invest in more storage in order to avoid shortages during winter periods.
The problem with this argument, with reference to the UK market, is that by the time those signals became evident in the early 2000s lead times for new supply and storage investments meant that it was not possible to moderate prices during the period 2004–07. The discomfort felt by many British industrial customers during the winter of 2005–06 – when high and volatile prices caused some energy intensive users to close their plants on a temporary or permanent basis – did not persuade them that such signals were a manifestation of a successfully operating market. (Mackenzie 2006).
Moving to gas-indexed prices need not necessarily exclude a mechanism to moderate volatility by adopting a price reference period of several months, just as in current long-term Continental European contracts. The difference would be that the reference period would average spot gas prices, rather than oil product prices, over a period of weeks or months. However, eliminating short-term price volatility could reduce the interest of speculators in this market and hence reduce liquidity.

WHAT COULD BE THE DETERMINANTS OF FUTURE GAS PRICES?
If it is correct that even in liberalised and competitive markets gas and oil prices will remain aligned over the longer term, then abandoning formal contractual linkage between the products should not pose problems for market stakeholders. Thus a conviction that the long-term alignment of gas prices with oil prices seen in the past will be a good guide to the future, should remove the resistance of Continental European market players to moving from oil-indexed contracts to short-term prices set at market hubs. But short-term pricing would create a much more competitive gas market which is not in the interest of dominant incumbent players. And at a time of high oil prices, it will be difficult to persuade incumbents that it is in their interest to make such a change. Put another way, gas market stakeholders and the financial community may not be prepared to accept DG COMP’s conclusion that ‘there is no clear commercial advantage between oil indexed and gas indexed contracts’ (see EU Energy Sector Inquiry 2006, paras 285 and 286). Because although this may be correct over a 5–10 year period, for relatively long periods of months and possibly years, prices of oil and gas may diverge significantly.
Without formal contractual linkage, gas prices will have their own short-term dynamics and surplus or shortage will have a significant impact on the short-term profitability of commercial players.
A key question for the future is, if not oil product prices, what will be the most important determinants of gas prices in Europe? One possible hypothesis (which draws on Foss’ recent work on US gas prices) is that long-term prices will be set by a price band where production and delivery cost define the floor, and alternative fuels in the power sector (because the latter is the principal source of incremental demand for gas) the ceiling.
If gas prices remain outside that band for a period of several years, the consequences will either be insufficient supply or lack of market growth.
Within the corridor, oil price linkage may be observed but will not necessarily have a positive or negative influence.
Because different European countries and regions will have different gas delivery costs, and different power generation alternatives, price dynamics may vary throughout Europe. Spain will have different price dynamics to North West Europe because of that country’s much greater dependence on LNG and pipeline gas from Algeria with a high degree of crude oil indexation, combined with an inability to build coal and nuclear power plants.
In North West Europe,the general level of oil prices will be a very important determinant of how long oil linkage can be sustained, because of the need to retain competitiveness in the power sector where the principal competition is coal and renewable energy (and in some countries possibly nuclear power in the
longer term):

• At oil prices significantly above $30/bbl, oil-linked gas prices are above the ceiling of the price band i.e. uncompetitive in the power sector which, in the longer term, will impact future demand which is heavily dependent on power generation.
  Below $30/bbl, oil-linked gas prices will be competitive with coal (depending on the source and the environmental standards which both fuels are required to meet) and oil price linkage could continue without restricting the market share of gas.
• At oil prices significantly below $20/bbl, i.e. below the floor of the price band, demand for gas-fired power generation will be stimulated but there will be little incentive to provide new supplies from the majority of sources.

Thus when oil prices were in the range of $20 –30/bbl i.e. until around 2003, oil-linked pricing did not present a significant problem for the gas industry. (For much of the 1990s when oil prices were below $20/bbl, costs of production and delivery from the - UK and Norwegian - North Sea, Netherlands, Russia and North Africa were low enough to allow producers significant profits and incentives to continue exploration and development. With exhaustion of reserves in the North Sea and the Netherlands and increased costs of new supplies from non-European sources, this is unlikely to be the case in the 2010s. Stern 2006). Since 2003, and particularly if oil prices significantly above $30/bbl continue for a long period of time – as many expect – oil-linked gas prices will prevent development of gas-fired power generation, thereby eliminating significant demand growth in the majority of European markets.
Another important determinant of gas prices will be whether the market is in shortage or surplus. In a shortage it will be much easier for market players to retain oil price indexation, whereas a surplus of supply could seriously threaten oil price linkage. A combination of high oil prices and a gas surplus would place maximum pressure on market participants to move to alternative indexation. In early 2007 it seemed likely that the UK was entering several years of oversupply during which it could serve as Europe’s ‘clearing house’ for gas, or ‘offshore loading jetty’ for LNG, which could be delivered to UK terminals and flow to Continental Europe using recently expanded pipeline capacity, in the Interconnector (IUK) and BBL pipelines. In this way, existing players and new entrants in Continental European markets will be able to access flexible and competitively priced supplies from the UK. Continental European utilities are able to operate a degree of arbitrage between NBP and long-term oil-indexed prices by reducing nominations of the latter when the prices of the former are more attractive. Those with LNG cargoes and regasification capacity on both sides of the Atlantic will also arbitrage NBP (and other European) prices against Henry Hub. (With appropriate allowances for transportation costs).
Because of this range of new and diverse influences on European gas prices, it seems likely that – at a minimum – the future will be different to the past. But the dominance of 20–25 year long-term oil-indexed take-or-pay contracts between market players who have strong interests in maintaining the status quo, means that the change is unlikely to be rapid. Given the market power of these players and the fact that many contracts in force today will still be in force 10–25 years hence, the most likely outcome is a progressive introduction of spot indices in price indexation formulae leading to greater diversity of indexation applied to individual market circumstances. The speed with which this occurs may depend on two considerations:

• the ability of Continental European buyers to manage the gas surplus emerging in North West Europe in the late 2000s;
• whether dominant incumbents (buyers and sellers) will be satisfied with market conditions where – in North West Europe at least – it will be very difficult to build new gas-fired generation (and hence obtain significant market growth) while retaining oil-linked prices. As long as dominant players – suppliers and buyers – are relaxed about slow or no future market growth, they may continue to be comfortable with oil-linked pricing and resist pressure to make any rapid changes.