بررسی هزینه های اقتصادی از محدودیت های ذخیره سازی استراتژیک گاز طبیعی

The European Commission wants to implement a single market for gas. One of the components of this market is a regulated provision for “security of supply” which consists of rules for the implementation and use of a given reserve stock of gas. We investigate the impact of this policy on the profitability of a storage operator, using data from Denmark and Italy. Keeping storage capacity constant, the costs of the strategic stock are around 20% of the value of the storage market for Denmark, and 16% for Italy. This cost is due to the inability to extract arbitrage profits from the captive stock. Furthermore, the strategic storage restriction induces behavior that would virtually never be replicated by a private storage operator in an unconstrained market, in particular in the first 6 months of the year when unconstrained firms empty their reservoirs much faster, suggesting the strategic restriction is unnecessarily distorting the market.

The European Commission wants to implement a single market for gas. One of the components of this market is a regulated provision for “security of supply” which consists of rules for the implementation and use of a given “strategic” reserve stock of gas. Given the forecasted dependence of Europe on imported gas, this policy deserves serious consideration. This paper investigates the impact of this policy on the profitability of a storage operator, and therefore indirectly on the incentive to invest in storage. The cost component we focus on — the loss due to the inability to use the captive stock for arbitrage gains — is largely absent from any academic or policy literature on gas storage, but, as this paper shows, it is far from negligible. In order to implement the creation of strategic storage the policy maker regulates the use of the existing stocks owned and managed by commercial operators. This regulation takes the form of a penalty on the “improper” use of the last x% of total capacity in any storage facility — the strategic part of the stock. Proper uses are determined by the regulator. The policy analysis here is the comparison of optimal behavior and profits of a storage operator with and without the penalties. Regulatory intervention is justified in the presence of externalities. Just like the European Commission, this paper assumes that market failures and externalities exist in connection with the market for natural gas, such that it is important for the policy maker to complement the commercial supply of stored gas with an extra stock to be released in case of emergency. Such positive externalities are likely to be dispersed throughout the economy, while the costs of regulation are potentially borne directly by storage operators.2 It is therefore useful to quantify the impact of regulation on the behavior and profitability of storage operators. In this way, the regulator will be able to design the appropriate compensation package for the social service provided by strategic storage. Another assumption used in this paper is that the storage component of the market for natural gas can be unbundled from production and distribution activities and function in a competitive environment. Although storage facilities are investments with large initial fixed costs, casual observation reveals that a large number of agents participate in the market.3 These assumptions allow us to write a model of optimal storage of gas where the agent (a storage firm) buys gas and stores it when prices are low, and sells gas when prices are high. The model is calibrated to match Italian and Danish storage data. Both of these countries have restrictions on the use of stored gas. Italy sets apart around 38% of its stored gas on a permanent basis, while Denmark imposes that stored gas must at all times be enough to cover the following 60 days of normal consumption. While large, Italy's storage capacity is smaller relative to total consumption than Denmark's, which comfortably meets its 60-day restriction.4 Our calibration for Italian and Danish data yields a cost of regulation of around 16% and 20% of discounted net present value of profits for the respective storage markets. These costs are due to the loss of seasonal arbitrage profit associated with the captured stock. We also find that the strategic storage restriction induces behavior that would virtually never be replicated by a private storage operator in an unconstrained market. In particular, during the first 6 months of the year unconstrained firms empty their reservoirs more rapidly than their regulated counterparts, suggesting the constraint may impose unnecessary distortions. The current paper is a measurement and policy evaluation exercise on natural gas storage which focuses on modelling the cost side of strategic gas policy. To our knowledge there is no close analogue to this exercise in the literature. Our paper is not, however, a structural cost–benefit analysis of regulation with detailed measurement of benefits. Instead, the benefit side is summarized by the price process for gas which includes the event of a large disruption, and the reduced form approach we take allows us to do sensitivity analysis on it.5 We believe that given the events of 2008 in Georgia this is a work of current relevance. The model draws on work by Wright and Williams, 1982a, Wright and Williams, 1982b and Wright and Williams, 1991Byers, 2006, Thompson et al., 2009, Hall and Rust, 2000, Chaton et al., 2007a and Chaton et al., 2007b. These last authors study a problem where storage agents face the possibility of a large disruption and therefore build up a precautionary stock up to a given level and then optimally maintain it. They do not study seasonal arbitrage nor do policy evaluation which is the main focus of this paper. Casassus et al. (2005) study a general equilibrium model of oil as an intermediate input with storage aimed at replicating the asset pricing properties of commodities, but again the seasonal component is absent from their framework. The paper proceeds with a presentation of Danish and Italian data and their seasonal properties. Section 3 introduces the problem using an example. Sections 4, 5 and 6, present, calibrate and simulate the main model. Section 7 extends the example from Section 3 to discuss model caveats and Section 8 concludes.

In this paper we evaluate the economic cost of imposing a strategic storage constraint on a gas storage operator. Such constraints impose a limit on the fraction of the gas stock the firm can use for normal seasonal arbitrage. This captive stock can, however, be used in the event of a well defined catastrophe. The catastrophe event is summarized as a low probability of an abnormally high price of gas. The constraint results in a profit loss, the measurement of which is non trivial due to the non linearity of the problem. We construct a model of storage of natural gas that replicates the behavior of stocks and flows in the Italian and Danish markets. The model is then used to evaluate the impact of strategic storage policy on the value of the firm. Our experiments suggest the impact of strategic storage policy on the value of the firm is very large. Our calibration for Italian and Danish data yields a cost of regulation of around 16% and 20% of discounted net present value of profits for the respective storage markets. Furthermore, imposing a Danish-type cyclical constraint on the Italian market that preserves the average size of the strategic stock provides higher insurance but at a higher cost per unit insured, as more gas is withheld in the winter when it is more valuable, and in turn more gas is released in the summer when it is less valuable. The percentage loss in firm value is not as sensitive to the catastrophe price as it is to the size of the constraint or to the normal variation of the price from summer to winter. The reason is that the biggest part of the loss comes from the inability to exercise the seasonal arbitrage on a significant fraction of the gas in stock. We also find that the strategic storage restriction induces behavior that is hard to replicate by a private storage operator in an unconstrained market suggesting the constraint may impose unnecessary distortions. Finally, the exogeneity of the price and the assumption of a fixed capacity condition the quantitative findings in this paper. Modelling the demand for gas is a difficult but important extension to the present exercise and we discuss some of its challenges. As for endogenizing capacity, given that investments in gas storage are long projects with very large fixed costs, it is acceptable to assume a constant capacity in the medium run.