Appendix D: Access Pricing Rules
We present here a brief summary of the various regulatory rules that might be considered for controlling the interconnection or access price of the vertically-integrated monopolist.
First-Best Pricing and Short Run Marginal Cost
There is general agreement among economists that where the state can transfer money to the monopolist at no cost, where there are no relevant capacity constraints, and where the regulator has perfect information about costs, the first-best outcome can be achieved by setting the price of the monopolist's output equal to the marginal cost of producing that output.
Typically, however, pricing at marginal cost (also called first best linear pricing) does not provide sufficient revenue to cover fixed or common costs. In some circumstances it may be possible to recover sufficient revenue to cover these costs by permitting the firm to engage in price discrimination and/or charge nonlinear tariffs.
If this option is not feasible, the first-best outcome cannot be achieved. Instead, the best outcome that can be achieved is the 'second-best' outcome in which prices provide sufficient revenues to cover costs. In particular the prices set by the regulator must exceed marginal costs by a margin sufficient to generate enough revenue to cover the 'fixed' costs of the firm.[125]
Second-Best Pricing With Multiple Products
Where the monopoly firm produces more than one product, the second-best pricing problem of the regulator is more complex. The principle is the same (prices should exceed marginal costs by an amount sufficient to generate enough revenue to cover the fixed costs of the firm), but a new problem arises: how are these mark-ups to be distributed between the various products? How should the fixed or common costs be allocated to ensure that the monopolist's revenue is sufficient to fully cover costs?
Traditionally, regulators have used non-economic methods for allocating these fixed costs, such as 'fully distributed cost' methods. Although pragmatic, these methods have attracted significant criticism for being arbitrary and unprincipled and will not be considered further here.
One approach to allocating the fixed costs was suggested by Ramsey (1927). Ramsey demonstrated that it is efficient to place higher markups over marginal cost on those products which are most inelastic. This minimises the 'deadweight loss' that arises from second-best pricing.
Ramsey prices are not necessarily 'sustainable' in the sense that the optimal price calculated from the Ramsey formula may be so high as to induce a firm to enter to produce that product alone. We might further constrain the second-best prices by requiring that they be 'subsidy-free'.
Incremental Costs
Faulhaber (1975) defined a set of prices as passing the 'incremental cost test' and therefore as 'subsidy-free' if, for each product (or set of products) the revenue from that product (or set) was sufficient to cover the incremental cost of producing that product (or set). Faulhaber demonstrated that if the prices are set so that the monopolist's revenue just covers its costs, we would expect to see entry if and only if the prices violated the incremental cost test. The reason is that if the firm failed to earn sufficient revenues on any subset then some other subset must be earning revenue in excess of stand alone costs and which is an invitation to entry.[126]
An important consequence of this result is that if the monopolist is to be regulated so that overall, its revenue just matches its costs, it is undesirable to set any price the monopolist can charge below average incremental cost (AIC) as this will force the monopolist to charge above stand-alone cost in some other market. A price above stand-alone cost invites inefficient entry and undermines the ability of the monopolist to fully cover its costs. For this reason, in many contexts, the long-run average incremental cost (LRAIC) (and not marginal cost) is seen as the appropriate lower bound on access prices.[127]
An access price set at LRAIC for a particular product prevents the monopolist from earning any contribution towards the fixed or common costs on that product. In contrast, in a context in which the final-product prices of the monopolist are known and sufficient to cover all costs, an access price given by the BW rule ensures that the monopolist is able to fully preserve any existing mark-ups over costs on all its products. Many authors have noted that the prices given by LRAIC and the BW rule represent lower and upper bounds on the set of reasonable prices.
An alternate framework for allocating the fixed costs over the various products is given by the 'peak load pricing' model of Steiner (1957) in which it is efficient to price at marginal cost in all periods except the peak period in which the price is set at long-run marginal cost. In the Steiner model this yielded a mark-up sufficient to cover all the fixed costs.
As in the single-good monopolist case, to a certain extent, the problem of cost allocation can be overcome by the use of non-linear pricing schedules. It may be possible, for example to price at marginal cost, but to charge an up-front 'entry fee' to each customer. If the entry fee generates sufficient revenue to cover the fixed costs, the first-best outcome can be achieved.
Capacity Constraints
Often, the access provider will face capacity constraints. For example, the maximum quantity of gas that can be carried down a single pipeline is, essentially, determined by the physical characteristics of the pipeline.[128]
In the presence of capacity constraints, some means of rationing the capacity is necessary. In general, it is economically efficient to use prices to ration capacity. If it is feasible, it will in general "be desirable to swiftly adjust the 'spot' price for access as demand changes so that the price is always used as the sole rationing mechanism . . . To enable such adjustments to occur, there would need to be a functioning spot market for access with open sale and purchase of access rights. [However] it is possible that, for many access facilities, the costs of establishing the necessary mechanisms to use spot prices as the only rationing device would be sufficiently high as to outweigh any potential benefits."[129]
An alternative to a spot market is some form of 'peak load' pricing. If demand varies in a stable pattern over time it may be possible to use a model such as that suggested by Steiner (1957) (mentioned above). In this model the access price is set at long run marginal cost (LRMC) in the period with the peak demand and at short run marginal cost in other periods. When demand is random, it is not possible to define a 'peak' period in advance, so 'peak load' pricing is infeasible. A less desirable, but occasionally adopted, alternative is to ration solely on a non-price basis such as 'first come first served'[130].
In a capacity-constrained system the incentives for the owner of the access facility to expand capacity will often be muted compared to the socially desirable incentives. Economic efficiency is served by expanding capacity whenever the peak price exceeds the LRMC. However the natural monopoly facility owner will only be willing to expand capacity voluntarily if the marginal revenue from the expansion exceeds the marginal cost. If the regulator chooses to set price equal to SRMC in off-peak periods, while allowing price to ration demand in peak periods, the access provider will seek to restrict capacity in order to raise the price in peak periods.
Asymmetric Information
In general, a regulator needs to know the cost function of the access provider and, often, the demand function for access. The regulator will not usually have perfect access to this information and will rely, at least in part, upon the owner of the access facility to provide that information. Even where information about costs can be demanded by the regulator, it is in general impossible to tell (without extensive investigation) if that information has been subject to manipulation or if the access owner has operated efficiently or indulged in cost-padding. In these cases, the available information will reflect distorted, rather than efficient, cost levels."[131]
Asymmetric information affects the design of regulatory mechanisms in two ways that relate to whether the costs of the firm are fixed in advance or depend upon cost-minimisation effort on the part of the firm. If the costs of the firm are fixed in advance and not observable by the regulator there may be incentives for the access provider to report inflated costs. If the costs of the firm are observable, the facility owner "will have little incentive to exert the effort and care necessary to maintain lower costs if this simply leads to a lower regulated return."[132]
Where the information received by the regulator is distorted the regulator will need to 'bribe' the regulated firm to reveal its costs. This might be achieved by permitting the regulated firm to keep some of the monopoly rents that it may extract. For example, suppose the regulator sets prices on the basis of the firms announced costs, and the regulator does not know whether the firm's costs are high or low. If the regulator simply sets the regulated prices equal to the announced costs the regulated firm will have no incentive to announce its true costs. If the low cost firm is to announce that it has low costs, it must be rewarded by a higher price, for doing so.[133]
There has been much recent research into the question of the optimal regulatory mechanism in the presence of both these problems. For example Laffont and Tirole (1993) derive a mechanism in a context in which the regulator can observe costs (but not effort) perfectly ex post. In their model the regulated firm is subject to a contractual arrangement that provides incentives for the regulated firm not just to expand its own output but also the output of the competitors.[134] Unfortunately, these and other more sophisticated incentive-compatible regulatory models appear to be rather impractical.
Interconnection Of Two-Way Networks
The discussion so far has highlighted the fact that the appropriate interconnection price, in general, will depend upon a number of industry, institutional and regulatory factors. In the specific context of the interconnection between two-way networks (i.e., networks where the product is able to be transmitted in both directions simultaneously, such as telecommunications and postal networks) the best interconnection rule is very simple. .
This is illustrated by using the example of two telephone networks For each network, natural monopoly considerations will typically preclude the duplication of the rival's network. Therefore, if customers in the two networks are to communicate with one another, some form of interconnection is essential.
In this example, Network A is extensive and Network B is currently non-existent [135]. If B cannot duplicate A's network, A has a monopoly over his/her network. If B wishes to compete for the business of a new subscriber, B will have to arrange for the new subscriber to be able to make calls into A's network. In the absence of controls A could, by jacking up the access price to its network, force B out of the market for this new subscriber.
In the next example Networks A and B are roughly the same size. If the new subscriber joins Network A he will wish to make calls into and receive calls from Network B. If A attempts to force B out of the market for this customer by raising the interconnect price, (thereby reducing the value to the customer of joining B's network) B can retaliate by doing the same to A. Both networks lose. This is a bilateral monopoly. The undesirable welfare effects of A's monopoly are offset by B's countervailing power.
If transactions costs are small enough the networks will realise that it is not in their mutual interest to restrict calls from each others' network. The only agreement which does not involve such exercise of monopoly power is one in which the parties agree to carry each other's calls at cost or at some non-discriminatory rate.
Such arrangements are common at the international level, between national telecommunications carriers. National carriers typically agree to terminate each other's calls at a fixed rate per minute (the 'accounting rate') that is the same for calls terminating in each network. At the end of a fixed period the carriers settle for the difference in the number of call-minutes. This sort of interconnection pricing rule is known as a reciprocal pricing rule, or 'reciprocity'.
This logic suggests an additional principle that might apply in the context of two-way networks. This principle states that where there is a bilateral monopoly (as in the case of interconnection for local access), the appropriate pricing rule is one which would arise if the parties had equal bargaining power.
The rule of 'reciprocity' is straightforward to implement. It is characterised by (a) no fixed charges and (b) charges based upon costs or, if the networks are based on similar technologies[136], equal charges (per call or per minute) for terminating calls in each network.
From a regulator's point of view these charges should be sufficiently high to fully cover all of the costs of terminating a local call. In the presence of sizeable fixed costs, as in the telecommunications industry, the definition of the 'cost' of a local call is inherently arbitrary. However that cost is likely to be less than the full per minute call charges (which cover the costs of both 'ends' of the call, not just the terminating end)[137]. Therefore, as a guideline, the usage charges should be no more than the full costs of terminating a call.
If the networks are relatively similar in terms of their customer base, technology and pricing structure, the imbalance of calls is likely to be very much smaller than the total number of calls passing across the interconnection. In such circumstances the absolute level of the interconnection call charges matters very little as, on average, the net interconnection charges paid by either network is zero.
In fact, in such circumstances it is often the case that the negotiating parties themselves recognise that the transactions costs of monitoring and billing the total number of calls that pass across the interconnection do not justify the revenue received on the imbalance. In these circumstances, it is not uncommon for the parties to voluntarily choose to terminate each other's calls for free ('bill and keep'). From a public policy perspective it is not necessary to regulate for such a rule. Once the principle of reciprocity is established, the parties will adopt such a rule if it is in their mutual interest.
If the networks do not have similar customer profiles, there may tend to be a permanent (although possibly small) imbalance in the number of calls terminating in each network. The larger this imbalance, the more significant is the interconnection revenue as a fraction of total revenue and, consequently, the more important is the level of the interconnection charges.
The rule of reciprocity has the advantage of being simple to implement and not informationally demanding. In addition, it is 'robust' because:
- the charges for the imbalance of the calls is likely to represent a tiny fraction of the total revenue received by each network; and
- the size of the regulated interconnection charge is unlikely to have significant distortionary consequences on the decisions of the regulated firms.
However, it should be emphasised that the rule of reciprocity cannot be implemented in a circumstance where one of the parties can choose to take just one 'direction' of a customers calls (such as all the out-bound calls, leaving all the in-bound calls to the other network operator). In such a circumstance, the imbalance of calls between the networks will not tend to be a small fraction of the total calls between the networks.
Back to Top