Traditionally, the distribution charge paid by consumers of electricity connected to a grid is based on the capacity for which they subscribe and the number of kWh withdrawn during a given period. Energy consumption is however actually a poor approximation of the willingness to pay for being connected. Additionally, except for thermal losses, the cost incurred by distribution system operators (DSO) does not depend on the quantity transported. The main raison d’être of the kWh index is that it is indispensable for producers and suppliers. Since it is an existing gauge provided by installed meters, why not use it for distribution?
Paying for kilowatt hours
Up to now, the answer to this question was simply that a price related neither to consumers’ preferences nor to suppliers’ costs gives biased incentives to consumers on where and how much to demand, and to DSOs on where to invest and how to allocate maintenance efforts. Consequently, a price per kWh is not efficient for electricity distribution.
Today, with the development of distributed generation units, mainly photovoltaic panels and wind turbines, there are two additional reasons for not charging distribution services per kWh: the financial equilibrium of DSOs is endangered, and this pricing rule is unfair (or regressive in the lingo of public economics).
When the number of “prosumers” increases – prosumers are both producers and consumers of energy because they have installed decentralized production plants at the consumption node – there is more off-grid consumption, which means smaller payments to the DSO, which in turn entails the obligation to increase the distribution price to recoup the costs. This is true because distribution costs are mainly fixed: installation and maintenance of wires, poles, transformers, meters and monitoring appliances. Then, by denoting Nc as the number of consumers, Qc as their average consumption, N-Nc as the number of “prosumers” and Qp as the quantity they withdraw from the grid, under a billing system based on energy, network users have to pay F=total cost/[NcQc + (N-Nc)Qp] for each kWh withdrawn from the grid. Given that, it is obvious that any increase in the number of prosumers decreases the denominator, so that is pushed up. For consumers who have a roof or a field where they can install PV panels, the larger fee increases the incentive to become a prosumer. But for those who have no alternative source for their electricity consumption, in particular poor consumers, the result is an increase in the distribution bill.
Note that in the former argument, we have supposed that the total cost incurred by the DSO does not change when increases. Actually, on the one hand the DSO must invest in storage and command equipment because it has quality and reliability obligations, whereas local intermittent production combined with “selfishly-adjusted” energy injection or withdrawal creates randomness in inflows and outflows in the area under the DSO’s responsibility. On the other hand, the cost is lowered because congestion and thermal losses may decrease. Note that the effect on thermal losses is not straightforward. Self-consumption decreases thermal losses, which reduces the DSO’s operating costs. Conversely, if prosumers use wind-turbine to sell energy at night without paying distribution fees, local thermal losses are larger than without decentralized production, and distribution costs increase.
Paying for kilowatts
If energy withdrawn from the grid were a correct index for utility and cost, the problem would merely consist in solving a standard second best. To maximize welfare under the DSO’s budget constraint, the rule is price discrimination à la Ramsey-Boîteux: a low price for elastic demand by prosumers, and a high price for inelastic demand by consumers. It may look unfair, but having prosumers pay for a part of the DSO’s cost decreases the bill for consumers. By contrast, the latter group would have to pay for the whole distribution cost if prosumers were discouraged by a high price to buy from the grid.
Nevertheless, one can do better by changing the billing basis. Prosumers want to remain connected due to the intermittency of their local generation: at some moments they want to sell extra production to the grid, while at others they want to buy missing energy from the grid. In other words, their demand is for insurance provided by connection.
On the DSO’s side, the cost is related to the installation and maintenance of fixed equipment. Consequently, it mainly depends on i) the distance to cover and ii) the capacity to provide at each node, that is, power (kW), not energy (kWh).
Charging subscribed kW rather than purchased kWh has the obvious advantage of better reflecting the cost incurred by DSOs. Additionally, in many places the capacity required does not change when a consumer becomes a prosumer. This is not true in regions where demand and local production are positively correlated, in particular where the peak demand is for air conditioning: bright sun simultaneously means more demand and more local production from PV panels, then a lower demand for capacity from the grid. By contrast, in all places where peak demand is at the end of the day, the demand for capacity from the grid does not change when distributed generation from PV panels increases.
Paying for kilometers
The main cost for distributors comes from the length of lines, the number of poles, the power of intermediary nodes (in particular transformers) and the workforce necessary to install, improve, and maintain the infrastructure. In general, when a consumer becomes a prosumer, (s)he does not move. With unchanged location, charging for the distance from the network head reflects both the costs and the intensity of the willingness to be connected. In developed countries, despite the economic rationale explained above, paying for distribution services using a distance index is unconceivable, given the current status of electricity (viewed as an essential commodity) and the fairness obsession implemented under the ‘postage stamp’ principle. By contrast, it can be helpful in underdeveloped regions where local production can be cheaper than a connection to the grid because of the length of lines that would be necessary.
In a recent paper published in Competition and Regulation in Network Industries, we explore the principle of connection to the distribution grid priced in terms of distance and its economic effects on the network size (length of lines) and the quality of service. We determine when it is and when it is not optimal to connect all consumers to the network. Depending on the common costs of quality and on the price differential between a local source and sourcing from the network, the marginal consumer connected to the grid is close to or far from the head of the network. The implementation of first best through linear pricing would entail two drawbacks: i) the net profit of the operator would be negative and ii) the charge paid by each consumer would be proportional to the distance, which could be viewed as discriminatory on legal grounds. To avoid these drawbacks, we study the behavior of a regulated distributor as regards the linear price level and the amount of capital installed. As suspected, under the obligation to balance the budget, the number of consumers for whom the connection to the network is optimal is lower than at first best.
For activities with public service obligations, the distributors face additional regulatory constraints. We then study the implementation of a uniform price across customers (postage stamp) as well as a two-part price. When the operator implements a two-part price with a variable part equal to the marginal cost per kilometer for each consumer, the distortion with first best is slighter than in the non-uniform linear price case. Finally, when the distributor has to serve as many consumers as at first best, we find that a two-part price requires a variable part lower than the marginal cost per kilometer and the financial equilibrium is reached thanks to the fixed part of the price, creating cross-subsidies, from consumers located close to the head of the network to those located at the end of lines. It clearly appears that any form of service obligation concerning size, quality and/or budget balancing creates unavoidable distortions as compared with first best. Under these alternative pricing rules all based on distance, the obligation to accept distributed generation does not create any distortion in the incentive to install local generation plants.
Claude Crampes (Toulouse School of Economics) and Mathias Laffont (Autorité de la Concurrence, France)
P.S. The authors remain solely responsible for analyses and opinions expressed in this paper.
