Pricing under uncertainty

Uncertainty concerning future demand is a complicating factor in several respects. It is particularly relevant in determining the capacity and composition of an electrical power system and in formulating a practical pricing policy; see Boiteux and Sta si (1952), Brown and Johnson (1969,1970 and 1973), Turvey (1970), Turvey-Anderson (1977) and Crew and Kleindorfer (1976,1978 and 1979). Available capacity is also subject to stochastic changes, which have to be taken into considera-tion; see Liokas (1983).

We now comment on the implications of uncertainty in regard to pricing policy. The paper by Brown and Johnson (1969) and the discussion following that paper were reviewed in some detail in chapter 2.

Pricing plays an equilibrating role in the market economy. Given this role, an electricity shortage might initially seem curious. In theory there is always a set of prices which will clear the market, even when e.g. the quantity supplied decreases drastically.

However, it is difficult and in some cases impossible to adapt prices to stochastic changes in demand and supply. Nowadays, changes in prices to customers are not used as a means of achieving equilibrium in shortage situations. Instead, electricity is rationed or, if this does not suffice, some users may be disconnected periodically.

Electricity producers, at least in Sweden, make a rather sharp distinc-tion between energy and capacity shortage; see Sanghvi (1982). Energy shortages occur when the electricity system cannot deliver the total energy demand during e.g. a year. Lack of water in a hydro-based system, as in Sweden or France, may cause an energy shortage. Capacity shortage occurs when the available capacity of plants and distribution lines is not sufficient to supply the instantaneous demand for electricity during peak-load hours. An energy shortage may be announced some months in advance, so that customers will have time to adopt. This is not possible when there is a capacity shortage. Thus the possibilities of using prices as a means to clear markets are considerably greater in cases of an energy shortage than when capacity shortage occur; cf.

Andersson-Taylor (1985).

If the electricity price is not adjusted to stochastic changes in supply and demand, substantial gross losses in efficiency will result. These losses are in principle similar to those in Figure 3.1 above for a uniform electricity price as compared to peak-load pricing in cases of complete certainty. The question of whether and to what extent prices should be adjusted to stochastic changes is primarily a practical problem. The optimal degree of fluctuation in electricity prices should be determined on the basis of cost/benefit evaluations of the practical and feasible alternatives for different consumer categories.

There are some exceptions in practice to the rule of not using instan-taneous pricing as a means to achieve equilibrium in shortage situations.

The exceptions are spot markets for exchanges among the power producers themselves and some specific contracts on rebates for "second rate" power on a interruptible basis in extreme peak-load situations to a limited number of large customers. For example, some large customers can be offered contracts on "second-rate" power at a lower price. In return for the lower price, these users have to be prepared to have their power supply interrupted when the risk of a shortage arises (interruptible rates).

Electricité de France has been testing rates which particularly take stochastic changes into account; see Penz (1981). This system is aimed at allowing consumers to assume comparatively greater payment liability for their demand during periods with potential risks of electricity shortages in return for an offer of lower electricity prices during other periods. Customers who choose the new tariff through a prearranged contract may have to pay a considerably higher price during a certain number of periods throughout the year when electricity shortages are anticipated. Such periods are announced in advance, but on relatively short notice. The customers do not know exactly when these periods could occur. They either have to ensure themselves of access to their own reserve capacity by making preparations well in advance or else accept consumption at the substantially higher price. The incentive for choosing the new rate is that the price will be so much lower during the rest of the year that many customers will find it worthwhile.

The use of such pricing alternative can, of course, reduce the require-ments of reserve capacity in the power industry. This new control mechanism can lead to an increase in total energy demand on an annual basis, whereas demand during peak periods, especially during extreme peaks, will be reduced.

In order to guard against interruptions in the supply of electricity, power companies maintain a reserve margin. When determining the reserve margin, the particular pricing and rationing regime to be used in the event of a shortage has to be defined in advance; cf. Visscher (1973) and Crew-Kleindorfer (1978). Some actual price structure, determined in advance, will usually be retained regardless of what happens to demand or to short-term changes in supply conditions.

If electricity prices are set in advance and are not altered continuously, the energy price should include a weighted cost to cover the risk of connecting up power stations with a SRMC higher than that which would apply to the energy type most likely to be utilized. A weighted cost for the risk that a shortage could arise at given prices during peak periods should also be incorporated; see Boiteux and Stasi (1952).

As has been pointed out by Turvey (1968a and 1970, p. 485) and Crew and Kleindorfer (1979, p. 105), a simple trade-off exists between pricing and investments in safety margins. If more reliability is desired, it can be achieved either by raising the price beforehand or by increasing capacity. Moreover welfare gains can be achieved by a discriminating price policy if the irregularity or uncertainty in the demand of different customer groups is considered explicitly when pooling demand in a collective or semi-collective network; see Boiteux and Stasi (1952). Then

the capacity required can also be reduced. In addition, customers can have a reserve capacity of their own simply by choosing e.g. radiators which can be heated by both electricity and oil. Such a choice will to a large extent depend on whether the tariff structure provides them with incentives to do so.