In some types of competition case, the standard presumption has become that fixed costs are of little or no relevance for price setting. In other types of case, fixed costs are presumed to be critical. Lau Nilausen explores these potential inconsistencies and explains that pricing incentives depend on industry and firm-specific factors that practitioners will need to assess for each case individually.
Competition regulators assess the link between prices and costs in different contexts. When considering the impact of mergers or pass-on of cartel overcharges, the standard presumption has become that fixed costs are of little or no relevance for price setting. When assessing potential predation, margin squeeze or exclusivity rebates, the standard presumption has become that fixed costs are critical for price setting.
At face value, these divergent presumptions appear internally inconsistent. Whether or not a rational company should reflect fixed costs in its prices during the normal course of business should depend on industry and firm-specific factors, not the type of assessment undertaken by regulators.
In this article, I explore these potential inconsistencies and explain how economic theory supports a single analytical framework relying on the facts of the market and the characteristics of the firm under assessment to determine which costs are relevant for analysing rational pricing incentives. I focus on the approaches supported by the European Commission (“the Commission”) in case precedent and guidance. This reflects both data availability and that the Commission acts as a reference point for European national competition regulators. I therefore only refer to the Commission as an example to make a broader point.
When assessing commercial pricing incentives, the Commission presumes that pricing reflects only marginal costs
In some contexts, the Commission argues that prices are set by reference to marginal costs. This includes assessments of merger effects and of how cartel overcharges may influence pricing throughout the value chain. These situations involve companies assumed to act in a commercially rational way during their normal course of business. The Commission defines marginal costs as “the cost increment incurred when purchasing one additional input”.
The Commission models the impact of mergers on pricing using, for example, the Gross Upward Pricing Pressure Index (GUPPI). A key input in the GUPPI approach is the margin that the merging parties consider when setting prices. The Commission explains that “For the purposes of the merger simulation analysis, the Commission is primarily interested in the margin measure that best reflects marginal costs, that is, those costs that are usually taken into account by firms when setting prices.”
The Commission’s exclusion of fixed costs from the assessment of pricing incentives is also reflected in its assessment of merger efficiencies. The Commission’s horizontal merger guidelines set out the position that “cost efficiencies that lead to reductions in variable or marginal costs are more likely to be relevant to the assessment of efficiencies than reductions in fixed costs; the former are, in principle, more likely to result in lower prices for consumers”. The Commission further explains that “Generally, fixed cost savings are not given such weight as the relationship between fixed costs and consumer prices is normally less direct, at least in the short run.”
The Commission presents a similar philosophy in the context of pass-on of cartel overcharges. The Commission’s guidance for assessing cartel pass-on (pass-on guidelines) explains that overcharges affecting a buyer’s fixed costs are “less likely to be passed on because such costs typically do not affect the direct purchaser’s price setting, at least not in the short run”. The Commission explains that variable costs are “more likely to be passed on, at least to some extent”, as marginal costs “typically affect the direct purchaser’s price-setting decisions”. This discussion in the Commission’s pass-on guidelines cross-refers to the horizontal merger guidelines.
An alternative approach could have been to consider margins based on long run average incremental costs (LRAIC). However, the Commission explains that such “incremental margins include a proportion of indirect costs and the extent to which these costs vary with [volumes] is prima facie less intuitive. Indeed, there are a number of elements suggesting that incremental margins are less relevant in the operators’ pricing decision, if at all”. The Commission hence explicitly rejects that companies consider LRAIC when setting prices and maintains that companies set prices primarily by reference to “costs that naturally vary” with the relevant volume metric.
The Commission keeps the possibility open that fixed costs in exceptional circumstances may inform price-setting. For example, in its pass-on guidelines it concludes that “it is possible that in certain cases, fixed costs may also be taken into account by an undertaking when determining its prices. If this is the case, it should be demonstrated by the party that supports such a view.” The Commission thereby raises the hurdle of an explicit presumption against the relevance of fixed costs for price-setting and places the onus on the parties to provide evidence capable of overturning the Commission’s presumption.
When assessing potentially abusive conduct, the Commission presumes that pricing reflects variable and fixed costs
The Commission adopts different reasoning when undertaking so-called as-efficient-competitor (AEC) tests to assess predation, margin squeeze, and exclusivity rebates (i.e., abuse of dominance). In these contexts, its practice reflects the view that commercially rational companies should set prices based on the LRAIC concept. LRAIC includes “not only all volume sensitive and fixed costs directly attributable to the production of the total volume of output of the product in question but also the increase in the common costs that is attributable to this activity”.
The Commission’s reliance on LRAIC in the context of predation and margin squeeze cases reflects the logic that “if the revenues associated with the downstream activity fall below LRAIC, a rational and profit-maximizing firm … has no economic interest in offering downstream services in the medium term. It could increase its overall result by either raising downstream prices to cover the additional costs of providing the service or – where there is no demand for this service at a higher price, to discontinue providing the service.”
In the Intel case, the Commission argued in the context of exclusivity rebates that “To be viable in the long run, a company must cover at least the total cost of producing its output. In the presence of high fixed costs, as in the x86 CPU industry, this implies that prices on average must be significantly above marginal costs for a company to cover its total costs and, thus, to remain viable.” In this case the Commission adopted an average avoidable cost (AAC) metric arguing that this was “conservative” given that “Other cost benchmarks which also take into account fixed costs elements may be more appropriate.”
In these contexts, the Commission has hence taken the view that companies rationally set prices by reference to medium to long-term profitability reflecting the recovery also of fixed costs. The rationale is that companies must recover their fixed costs to be economically viable in the long run.
The same pricing incentives are relevant for assessing purely commercial pricing and potentially abusive pricing
Different cost standards could have emerged for the assessment of commercial pricing and for the assessment of abuse of dominance if these represented different types of situations in which companies should rationally act on different incentives. However, the type of assessment that the Commission conducts should not affect its understanding of how companies rationally set prices in relation to costs.
In European jurisprudence, abuse of dominance is defined as the “recourse to methods different from those which condition normal competition in products or services on the basis of the transactions of commercial operators”. An assessment of whether a company’s prices imply a different level of cost recovery than under normal competition necessarily requires a cost benchmark consistent with that used under normal competition. The cost standard should therefore be the same for testing for abuse of dominance as when assessing normal commercial pricing incentives.
The Court of Justice of the European Union explains that “Competition on the merits may, by definition, lead to the departure from the market or the marginalisation of competitors that are less efficient and so less attractive to consumers from the point of view of, among other things, price, choice, quality or innovation.” As competition on the merits can result in pricing which may not be viable for a particular competitor in the long run, finding that prices are below this level does not demonstrate the use of “methods different from those which condition normal competition”. This again implies that the cost standard should be the same for testing for abuse of dominance as when assessing normal commercial pricing incentives.
Applying consistent principles to assess links between prices and costs
There is sound economic logic behind both the marginal cost-based reasoning typically referred to in assessments of commercial pricing and the LRAIC standard typically referred to for assessments of abuse of dominance. That leaves the question of whether there is a way to reconcile the two cost measures and establish in which contexts each cost standard is appropriate.
The Commission notes in the pass-on guidelines that the relationship between prices and costs depends on context. Specifically, the pass-on guidelines explain that “The time frame over which pricing is considered will affect whether costs are categorised as variable or fixed. Generally, economic theory suggests that the longer the relevant time frame, the greater the proportion of total costs that should be considered as variable. In other words, a certain cost category which is viewed as fixed in the short run might be regarded as variable by the firm when considering a longer time frame. When assessing the relevant time frame in a specific case, the court may wish to consider information from the party’s internal documents, e.g., information on the costs that the firms take into account in their own pricing decisions.”
This suggests that the question of whether pricing should reflect marginal costs, LRAIC or something in between boils down to i) the time horizon applied by a company setting prices based on rational commercial incentives, and ii) the costs attributable to its supply of the relevant incremental volumes within that time horizon. Importantly, the rational commercial incentives that determine the relevant time period depend on the specific circumstances of the industry and firm. As such, they need to be assessed and evidenced in each case.
In the discussion below, I refer to the cost of incremental volumes rather than marginal volumes. This is to emphasise that businesses generally develop a standard offering across minimum expected volumes rather than making separate manufacturing and sales decisions for each unit of output individually. An assessment of companies’ pricing incentives should therefore reflect the economics of supplying a commercially meaningful increment of quantity.
I base my analysis on a stylised example that addresses a value chain including R&D, investment in production capacity, production, the manufacturer’s inventories, sale from manufacturer to retailer, the retailer’s inventories, and sale to final consumers. This stylised example will not cover every issue relevant to every industry but will simply illustrate how the link between prices and the relevant measurement of costs depends on the specific circumstances under assessment. To do so, it is easier to work backward through the value chain, starting from pricing to consumers.
Pricing by retailers to consumers
In this example, a retailer supplies consumers using products that the retailer holds in inventory. Moreover, products held in inventory are already paid for by the retailer. Under these assumptions, all costs of the product are sunk at the time of the sale to the consumer. The marginal cost of supplying a particular product to a particular consumer is therefore zero.
The finding that the cost of sales is zero may seem counterintuitive. However, this highlights the importance of articulating all relevant factors before concluding whether a retailer genuinely may set prices on the basis that its products have a cost of zero.
In one scenario, a retailer intends to keep supplying this type of product to consumers for the foreseeable future. Making a sale to one consumer would then trigger a need to replenish inventories by buying a similar number of units from the retailer’s supplier. As the retailer would incur the cost to restock inventories for each unit of sale, the cost of sales would in economic terms equal the cost of buying a new unit. In these circumstances, the retailer should not treat its inventories as sunk costs or its cost of sales as zero.
In another scenario, a retailer does not intend to keep supplying the product in question. As the retailer would not incur any costs to replenish inventories, the costs of all units in stock would be sunk and the marginal cost of a sale would in fact be zero. In these circumstances, any non-zero price that the retailer would be able to receive by selling the products would leave the retailer better off than not making the sale. This explains why retailers may offer steep discounts on products that are discontinued, e.g., on fashion products at the end of the season.
The sunk cost fallacy refers to a situation in which an entity irrationally uses past irreversible investments as justification for future investments rather than assessing these on their own merits. Rational business decisions at any time should therefore reflect only the expected resulting future cash in- and out-flows. However, this forward-looking perspective also has another implication: even though the cost to manufacture an existing unit readily available for sale is sunk, this is irrelevant for an assessment of which costs a company in the business of continuous supply would rationally expect to incur.
The relevant question is therefore not whether the cost of a particular unit of product sold to a particular consumer is sunk; the relevant question is what the replacement cost of the unit is. Units that will not be replaced have a replacement cost of zero and it would be rational for retailers to treat them as such during their normal course of business. For the same reason, prices below replacement costs cannot be presumed to be part of an effort to foreclose competitors as they may simply reflect a legitimate commercial objective to monetise inventories before they become obsolete. For units that will be replaced, the opposite applies.
Pricing by manufacturers to retailers – the treatment of variable costs of production
In this example, a manufacturer supplies retailers with units held in stock. The logic explained above therefore applies again: if a manufacturer intends to keep supplying in the foreseeable future, it will need to replenish its inventories and the cost of sales equals (at least) the variable cost of production. In contrast, if an otherwise identical manufacturer is discontinuing the product in question and just running down inventories, all its costs are sunk and the cost of sales is zero as a matter of economics. The benchmark for how otherwise identical manufacturers set prices during their normal course of business may therefore differ depending on the relevant context.
Pricing by manufacturers to retailers – the treatment of investments in production capacity
Whereas variable costs of production vary directly in response to the number of units produced, investments in production assets involve an upfront payment for the ability to produce a number of units over a longer period of time. These costs are typically excluded from the cost measure used by the Commission to assess commercial pricing incentives but included in the LRAIC measure the Commission typically uses to assess abuse of dominance. As for the discussion of variable costs, the appropriate treatment of production capacity costs depends on the circumstances.
In one scenario, a firm has in a previous period made a single large investment in an asset that has sufficient capacity to satisfy demand at the relevant time and that does not depreciate depending on the quantity supplied. Under these assumptions, the supply of incremental volumes does not create a need for new investment. As the incremental volume supplied does not trigger any investments, investment costs should not inform the firm’s pricing decisions.
In another scenario, the firm is capacity constrained. To supply incremental volumes, the firm would therefore not only have to incur variable costs of production but also invest in additional capacity. In such circumstances, the cost of capacity is directly attributable to the supply of the incremental volume and these costs therefore should inform the firm’s pricing decisions.
Whether or not the marginal cost of supply includes capacity costs is therefore a factual, case specific, question. If the firm invests in capacity during the relevant period, it is likely capacity constrained. Importantly, firms may replace assets on an ongoing basis as the assets reach the end of their useful life. Even a firm that does not grow its total sales may therefore face capacity costs when making incremental sales decisions because the alternative may be to not replace retired assets and instead accept lower outputs.
The correct treatment of capacity costs is subject to further complexities. Firms may invest in assets based not on current market conditions but on expectations many years into the future. The fact that a company undertakes capacity investments during the period of assessment therefore does not necessarily imply that the company was capacity constrained during that period. Also, it may be possible for the firm to sell rather than use the assets for production. The fact that a company does not undertake capacity investments during the period of assessment therefore does not necessarily imply that the company did not incur an opportunity cost associated with its production assets during the period.
Pricing by manufacturers to retailers – the treatment of R&D costs
The relevance of R&D costs also depends on industry and firm-specific context.
Costs incurred to develop a product inherently precede the sale of that product. Moreover, up-front R&D costs are unaffected by whether a product ultimately sells in high or low volumes. This suggests that R&D costs should not constrain pricing once the product has been developed.
However, it may not be rational to set a low price for a product failing to attract the volumes hoped for when undertaking the original R&D investment. For example, a company may keep marketing legacy products in parallel with new products but at different price points. In such circumstances, the desired price point for the new product may constrain the degree to which the company may be willing to discount the legacy product to attract incremental volumes. For companies expecting to engage in continuous innovation, future R&D costs to be recovered by future products may therefore inform its overall pricing strategy also for legacy products.
The Commission’s explanation that “the longer the relevant time frame [for pricing decisions], the greater the proportion of total costs that should be considered as variable” implies that no cost category can be treated as inherently variable or inherently fixed. Whether a given cost category is relevant for the assessment of a company’s pricing incentives therefore does not depend on whether the company refers to these costs as variable or fixed during its normal course of business but instead depends on industry characteristics and the circumstances of the individual firm. This is consistent with “a wealth of survey evidence [indicating] that more than 60 percent of American manufacturing companies do include fixed costs in prices, a practice called full-cost pricing” such that what is normal for one firm may not be so for another one.
It may be the case that the facts of past abuse cases coincidentally aligned such that it was justified to conduct AEC tests based on LRAIC and that the facts of past merger cases coincidentally aligned such that it was justified to assess pricing incentives against marginal costs. However, there is no basis in economic theory or fact for any presumption that the facts and circumstances of future assessments would justify AEC tests based on LRAIC for abuse investigations and tests of the impact of mergers on pricing incentives based on a narrow definition of marginal costs. Rather, economic theory and market evidence support that a firm’s pricing incentives depend on a number of industry and firm-specific factors that practitioners will need to assess for each case individually.
 The views expressed in this article are the views of the author only and do not necessarily represent the views of Compass Lexecon, its management, its subsidiaries, its affiliates, its employees, or its clients.
 Guidelines for national courts on how to estimate the share of overcharge which was passed on to the indirect purchaser (2019/C 267/07), paragraph 159.
 Case M.6497 Hutchison 3G Austria / Orange Austria, paragraph 349 footnote 238; M.6992 Hutchison 3G UK/ Telefonica Ireland, Annex I, paragraph 33; Case M.8792 T-Mobile NL/ TELE2 NL, Annex A, paragraph 126; Case M.7018 Telefónica Deutschland/ E-Plus, paragraph 716; Case M.7612 Hutchison 3G UK/ Telefonica UK, Annex A, paragraph 11; Case M.7758 Hutchison 3G Italy/WIND/JV, Annex A, paragraph 13.
 Case M.8792 T-Mobile NL/ TELE2 NL, Annex A, paragraph 126.
 Horizontal Merger Guidelines, paragraph 80. See also e.g., cases M.6992 Hutchison 3G UK/ Telefonica Ireland, paragraph 781; M.8792 T-Mobile NL/ TELE2 NL, paragraph 893; M.7018 Telefónica Deutschland/ E-Plus, paragraph 945; M.7612 Hutchison 3G UK/ Telefonica UK, paragraph 2535, M.7758 Hutchison 3G Italy/WIND/JV, paragraph 1399.
 Horizontal Merger Guidelines, footnote 107.
 Pass-on refers to the extent to which a buyer of a cartelised product has recovered some or all cartel overcharges incurred by increasing its own prices.
 Guidelines for national courts on how to estimate the share of overcharge which was passed on to the indirect purchaser (2019/C 267/07), paragraph 52.
 Pass-on Guidelines, paragraph 52.
 Pass-on Guidelines, footnote 57.
 Communication from the Commission — Guidance on the Commission’s enforcement priorities in applying Article 82 of the EC Treaty to abusive exclusionary conduct by dominant undertakings (2009/C 45/02), paragraph 26: “LRAIC includes product specific fixed costs made before the period in which allegedly abusive conduct took place. Failure to cover LRAIC indicates that the dominant undertaking is not recovering all the (attributable) fixed costs of producing the good or service in question”.
 Case M.8792 T-Mobile NL/ TELE2 NL, Annex A, paragraph 130.
 Case M.8792 T-Mobile NL/ TELE2 NL, Annex A, paragraph 124.
 Pass-on Guidelines, footnote 57.
 Case COMP/38.784 Wanadoo España v Telefónica, paragraph 319.
 Case COMP/38.784 Wanadoo España v Telefónica, paragraph 321. A similar logic was used in the following cases: Case COMP/35.141 Deutsche Post AG, paragraph 35; Case COMP/C-1/37.451, 37.578, 37.579 Deutsche Telekom AG, paragraph 155; Case COMP/38.233 Wanadoo Interactive, paragraphs 39 and 64; Case AT.39523 Slovak Telekom, paragraphs 860 and 861; and Case AT.39711 Qualcomm (predation), paragraph 870.
 Case COMP/C-3/37.990 Intel.
 Case COMP/C-3/37.990 Intel, paragraph 1036.
 Case COMP/C-3/37.990 Intel, paragraph 1037 and footnote 1355. Intel argued that the Commission’s AAC measure in fact included fixed, unavoidable costs. This issue was not resolved in the appeal process as the General Court rejected the Commission’s AEC test for different reasons and therefore did not engage with this part of Intel’s challenge (Case T‑286/09 RENV, Intel v Commission, paragraph 482).
 Case 85/76 Hoffmann-La Roche, paragraph 91.
 Case C-413/14 P Intel v Commission, paragraph 134.
 Pass-on Guidelines, paragraph 162.
 Principles Of Economics 8th Edition N. Gregory Mankiw, page 275: “Because nothing can be done about sunk costs, you should ignore them when making decisions about various aspects of life, including business strategy.”
 Pass-on Guidelines, paragraph 162.
 Ray, Korok and Gramlich, Jacob, Reconciling Full-Cost and Marginal-Cost Pricing, Journal of Management Accounting Research, Vol. 28, No. 1, Spring 2016, page 27.