What are common cost models used for determining interconnection tariffs and how do they deal with common costs?

[Response by Eric P. Chiang, May 2009]

The appropriateness of cost-based pricing methods in interconnection to improve and sustain efficiency and competition is well justified. However, many important issues arise in the way costs are used to price interconnection. For example, how are costs measured and to what extent are common costs (or overhead) allocated?

Primary costing models include fully-distributed cost and incremental cost methods. Jamison (2006, 2008) provides a thorough discussion and analysis of these and other costing methods.

A key difference between cost models is how prices are developed using cost data. A top-down approach1 takes the existing cost structure of a group of services, and allocates the costs incurred in producing each product (using accounting records). This usually leaves a large amount of common costs to be allocated to each product, using various cost-causation methods (what product causes what cost). A bottom-up approach2 does not use actual accounting data, but rather a model to estimate the costs of producing each product using the most efficient means of doing so with current and estimated future technology.

Fully-distributed cost: The fully-distributed cost (FDC) method is a top-down costing approach that was the dominant form of cost measurement over the past several decades. It is still used in many countries today, primarily due to its simplicity in that costs are apportioned to the services which incur the cost. The main advantage to using FDC is that the method uses accounting methods that tend to be more easily verifiable for audit purposes. Secondly, at least for incumbents, FDC often embeds a contribution to common costs.

FDC assumes that there exist some accounts that can be specifically allocated to a single service, while other accounts are classified as common or overhead cost to two or more services. To allocate common costs, input coefficients3 are usually developed as parameters to be estimated when dividing common costs among groups of shared inputs. For example, assume there are two services, A and B. Total costs for the production of the two services is 100, of which 20 is attributable to service A and 30 is attributable to service B. The remaining cost of 50 is considered common costs. If the input coefficient is 0.5 for each service, then the common costs are shared equally, that is 25 for each service. The cost of each service is therefore 45 for service A and 55 for service B.

Though the FDC process seems simple, in reality cost causation is rarely clear from accounting data, and thus cost allocations often include an element of arbitrariness. For example, accounts for which costs are to be determined may be arbitrary in their development; further, these accounts often use subjective methods to allocate the costs to various services.

Given these difficulties, there are many drawbacks of using FDC, including:

  1. The arbitrary nature of determining what constitutes direct costs and common costs
  2. The inability to use historical cost data to capture reduced common costs resulting from new technologies, which can result in higher interconnection costs
  3. The inability to appropriately allocate costs when changes in traffic patterns occur, causing some services to receive less allocated costs for providing the same service levels

While the general trend is for countries to move away from FDC, many countries including some members of the European Union continue to rely on FDC. While there is no published comprehensive survey on costing methods used by countries, the alternative use of incremental costing is seen as an effective means of promoting competition in the face of rapidly changing technology.

Incremental cost: Incremental cost methods are a type of bottom-up approach that arguably can improve and facilitate entry and competition. Incremental costing is a form of marginal cost pricing, with the distinction that incremental costs measure the additional cost of providing an increment (instead of just one additional unit), which can be a service or a network element.

One issue arising with incremental cost is how it accounts for fixed and common costs. This is addressed using variations of the long-run incremental cost (LRIC) method, which uses estimated forward-looking costs of producing services over a longer horizon.

Variations of LRIC: Various forms of LRIC have been adopted in different countries, such as TELRIC, TSLRIC, and LRAIC. But as countries undergo regulatory reform, incremental costing remains a common goal to be achieved once accurate cost studies become available. The following summarizes the variations of LRIC in terms of how they address common costs.

The total element long-run incremental cost (TELRIC)4 method was developed in the U.S. as an approach to calculating prices based on the increment of the unbundled elements used to provide the service. In other words, it measures the incremental cost of adding or subtracting a network element from a hypothetical efficient system using current technologies. In addition, the method includes a reasonable allocation of forward-looking common costs, which allows the incumbent to recover a share of the fair value of their inputs in a competitive market over the long term.

The total service long-run incremental cost (TSLRIC)5 method is similar to TELRIC in that it also estimates forward-looking common costs; however, incremental cost is based on the service, not the element. In other words, TSLRIC measures the incremental cost incurred of providing a service vs. not providing a service. This method is essentially the same as one developed by the European Union called long-run average incremental cost (LRAIC)6 in that it measures the incremental costs of providing the total service, including service-specific fixed costs.

Despite the differences found in the LRIC methods, there exist some common features. Interconnection prices are generally lower under LRIC compared with other cost-based methods, including FDC, as well as most non-cost-based approaches. LRIC methods provide just enough compensation for the incumbent to provide the necessary inputs to the entrant, including a fair return on investment. There are cases, however, where interconnection prices can be lower than costs; e.g., when positive network externalities exist, such that lower interconnection prices are offset by higher in-network revenues. Another example occurs when tariffs are highly unbalanced, such that interconnection prices reflect the subsidized retail prices.



Cost Concepts for Utility Regulators
University of Florida, Department of Economics, PURC Working Paper, 2006.
Jamison, Mark A.

Methods for Increasing Competition in Telecommunications Markets
University of Florida, Department of Economics, PURC Working Paper, 2008.
Jamison, Mark A.


  1. Top-down approach: a general strategy for allocating costs using existing accounting data. It takes the existing cost structure of a group of services, and allocates the direct costs incurred in producing each product. The remaining common costs are allocated to each product using various cost-causation methods such as input coefficients.
  2. Bottom-up approach: a general strategy for allocating costs that does not rely on existing data, but rather on a model that estimates the cost of producing each service using the most efficient means with current technology.
  3. Input coefficients: the percentage of the joint or common costs of a shared input that gets attributed to each service that uses the shared input.
  4. Total element long-run incremental cost (TELRIC): an approach to calculating costs based on the unbundled elements used to provide the service. It measures the forward-looking incremental cost of adding or subtracting a network element from a hypothetical efficient system built from scratch using current technologies.
  5. Total service long-run incremental cost (TSLIRIC): an approach to calculating the total forward-looking incremental cost of providing a service in a hypothetical efficient system built from scratch using current technologies.
  6. Long-run average incremental cost (LRAIC): essentially the same as total service long-run incremental cost (TSLIRIC).