1 Introduction

Third-party access is one of the key tools used by regulators to stimulate competition in network industries. However, in the postal sector, incumbent operators could refuse to provide access to their networks on competitive terms. As the experience from the electronic communications sectors shows, constructive refusals to deal,Footnote 1 caused by unreasonable or onerous terms, are far more frequent than outright refusals. These could be implemented through a margin squeeze, where a dominant firm offers wholesale access at a price greater than the difference between its retail price and its wholesale cost. Such combination of retail and wholesale prices allows the dominant firm to leverage its upstream position and to make downstream competitors unprofitable. Only a vertically integrated company can engage in a price squeeze as it needs to be able to influence prices, or better the margin between the prices, in two related markets. Without a simultaneous presence in two vertically integrated markets, a company can still engage in predatory or excessive pricing, discrimination, or refusal to deal, but only in one market.

Margin squeeze as an abuse of dominant position has primarily occurred in electronic communications markets, where the national regulatory authorities (NRAs) had already granted mandatory access to legacy networks under sectorial regulation. However, in the last few years, a number of national competition authorities (NCAs) have started to investigate margin squeeze cases in major European postal markets where mandatory access had not always been required under sectorial regulations. These include AGCM (Italy) in 2017, Bundeskartellamt (Germany) in 2015, CNMC (Spain) in 2014, and Ofcom (UK) in 2018.Footnote 2 In any case, the standing of margin squeeze as an independent form of abuse is now well-settled in the EU.Footnote 3 Only a few academics have expressed concern that margin squeeze could compromise the internal coherence of competition law by providing a rule that seems to respond specifically to the problems and needs of the electronic communications sector or, at most, the regulated network industries.Footnote 4 Precisely because of their general applicability, provisions of competition law have to be sufficiently flexible so as to encompass a whole spectrum of anticompetitive practices, independently of the features of the industry in which they may take place. On the other hand, even the US Supreme Court pointed out in its Trinko decision that “antitrust analysis must sensitively recognize and reflect the distinctive economic and legal setting of the regulated industry to which it applies.”Footnote 5

In light of the different economic features of the telecommunications and postal sectors, the related diversity of the respective access regimes, and the digitalization and the emergence of new business models, we wish to explore whether recent margin squeeze investigations in the postal sector reflect its distinctive economic and legal setting. To address this question, the paper examines recent margin squeeze investigations in the postal sector under existing national regulatory frameworks, and, where useful, compares them with landmark European margin squeeze cases in the telecoms sector.

In terms of the structure, the paper is divided in five sections. Following the introduction, which explained the timely relevance of the topic, Section 2 discusses access to the postal and the telecoms networks in light of the technical and economic features that distinguish these sectors. Section 3 reviews margin squeeze cases in the postal sector, and Section 4 compares them with Article 102 TFEU margin squeeze cases in the telecoms sector. Section 5 briefly concludes.

2 Access to the Postal and the Telecom Networks in Light of the Respective Technical and Economic Features

The existence of sector-specific regulation, and in particular the imposition of access requirements on incumbents, is typically justified by the existence of a bottleneck that cannot reasonably be replicated and is necessary for entrants. In telecommunications, for example, the existence of the bottleneck (local loop) necessary to access end users involved substantial sunk costs, which made the duplication of the existing network, at least for the so-called last mile, unfeasible in economic terms. The postal sector, in contrast, does not require a high last-mile sunk costs that would involve extensive civil works and high fixed costs. Instead, most costs are labor-related since mail delivery is a very labor-intensive business (Okholm et al. 2015). This makes the postal network more easily replicable, allowing alternative postal operators to bypass the legacy network to a significant, albeit not full, extent.Footnote 6 The possibility of bypass implies that end-to-end operators could offer more significant price reductions than access-based operators, whose ability to reduce prices is determined, among others, by the point of access they use (Copenhagen Economics 2013). Moreover, “postal infrastructures are somewhat different from other network infrastructures in that much of the transmission is made up of common means of transportation, with no dedicated physical infrastructure” (Parcu and Silvestri 2017).

This difference concerning sunk costs has important implications. In particular, the fact that the postal network is unlikely to be seen as a “sunk cost bottleneck,” since labor costs are not sunk, weakens traditionally invoked justifications for mandatory access (Geradin 2015:528). Lower last-mile sunk costs for the postal network are to be considered a good explanation for the differences between the respective regulatory frameworks of posts and telecoms. Article 11a of the Postal Directive refers to the Member States’ right, but not an obligation, to “adopt measures to ensure access to the postal network under transparent, proportional and non-discriminatory conditions.”Footnote 7 In contrast, the importance of effective access in telecoms was such that before the Access Directive 2002/19/EC was adopted, the European Parliament and the Council decided that it was necessary to lay down conditions for mandatory access already in Regulation 2887/2000.Footnote 8

Revenues in the postal and the telecoms sector are shaped differently due to a different demand structure. In the postal sector, demand is more concentrated and comes primarily from a limited number of large business customers (Okholm et al. 2015). As Geradin (2015) explained, even a small reduction in price could allow competing operators to convince a large mailer to switch away from the incumbent. Such a concentrated demand has another important implication: the inability to compete in the market for bulk mail may have a negative impact on the overall level of competition. This is because large customers are important for alternative postal operators as, due to their size, they help them achieve economies of scale, which are necessary to operate in a sustained and competitive manner.

Also, the traditional distinction between wholesale and retail services differs from that in telecommunications. In the telecommunications sector, competitors mostly seek access to incumbents’ networks, and incumbents have an incentive to discriminate between them and their own retail arm. Also, discrimination is likely to be implied, and not overt, and detecting it in the prevailing context of vertical integration may be difficult as long as wholesale and retail activities are not fully separated. In the postal sector, on the other hand, there are two main different types of access seekers, competing postal operators and large postal users. Postal incumbents may have a clear incentive to discriminate between the two groups, favoring the latter, which makes discrimination easier to detect. To the extent that services provided by these two groups may be equivalent, there is a strong possibility that such a discrimination could violate Article 102 TFEU.

As we will see in the next section, in all analyzed cases, national incumbent postal operators effectively discriminated between competing postal operators and large postal users or between end-to-end and access-only operators. Only the UK examined the conduct of its incumbent as a case of discrimination, whereas Germany, Italy, and Spain decided to frame their cases as margin squeeze violations. To the extent that all cases concerned similar scenarios, the question arises whether it makes any true difference which analytical framework (discrimination or margin squeeze) a national competition authority decides to use.

3 Margin Squeeze Abuses in the Postal Sector

Commercial strategies and the propensity of a given sector to a particular type of abusive behavior depend on economic features of the sector and the sectorial regulation that reflects them. Margin squeeze has become widely known as an independent type of abuse mostly due to the Commission’s Article 102 cases in the telecoms sector, but also the fact that its approach was in a stark contrast with the US Supreme Court’s ruling in the 2009 linkLine case.Footnote 9 The telecommunications sector provided fertile ground for such behavior, as it satisfied the conditions necessary to make margin squeeze feasible and profitable. A vertically integrated incumbent, strongly dominant in an upstream market, faced limited downstream competitionFootnote 10 that depends on access to the incumbent’s upstream bottleneck (even if a given bottleneck fell short of an essential facility, as it was defined in Bronner). Also, despite regulation, incumbents retained the ability to manipulate the margin between wholesale and retail prices.

The postal sector seems to satisfy these conditions necessary to implement a margin squeeze. Yet, abuse of dominant position in the postal sector is typically associated with violations regarding discriminatory pricing and anticompetitive rebates.Footnote 11 Copenhagen Economics (2013:186), based on responses from NRAs and NCAs to a survey covering a 10-year period up to 2011, identified 17 cases concerning discriminatory pricing and 11 concerning conditional rebates, but only 1 margin squeeze.Footnote 12 Still, as mentioned earlier, in the last few years, a number of NCAs started to investigate margin squeeze allegations in the postal sector. All of these cases concerned the market for bulk mail delivery, in which the demand for services comes from business clients – postal operators’ final customers, who send large volumes of mail to their own customers – and from competing postal operators.

Let us start with the 2014 margin squeeze decision of the Spanish CNMC. In the downstream market, the Spanish postal incumbent, Correos, held approximately a 90% market share measured in terms of revenue and 84% in terms of volume, whereas its main competitor, Unipost, held, respectively, 8% and 11%. Correos enjoyed a monopoly as the only provider of wholesale access services. Unipost filed a complaint alleging that Correos offered discounts to its large customers far higher than those it offered to Unipost and other similar competitors for using its postal network. The discounts, based on the volume of consignments, savings, and regularity, were approved by the then existing National Postal Sector Commission. They could reach 16% when granted to alternative postal operators, but went up to 57% when offered to large customers.Footnote 13

According to the CNMC, this difference prevented alternative operators from competing with Correos for the large customer segment. Unipost complained that these discounts were so high that it could never offer its services to those customers without incurring losses, and this according to the authority amounted to an abusive margin squeeze. The CNMC agreed and imposed on Correos a fine of nearly 8.2 million EUR, which Correos appealed before the National Court (Audiencia Nacional). The Court annulled the decision and the fine on the grounds that even though the CNMC had proven the existence of a margin squeeze, it failed to establish abuse as it had not demonstrated that the conduct in question led, “even in a presumed or potential manner” (Audiencia Nacional 2015, p. 6), to the exclusion of competitors.

While an abuse of a dominant position consisting of a margin squeeze does not require the competition authority to demonstrate intent to exclude nor the existence of effective consequences, the authority should prove that exclusionary effects are, at least, possible (Audiencia Nacional, Spain 2015). The ruling of Audiencia Nacional was upheld by the Supreme Court, which dismissed the cassation appeal filed by the State Attorney.Footnote 14 The Supreme Court clarified that, in the present case, the possible existence of exclusionary effects had not been proven, since, as the CNMC itself had acknowledged, Unipost had the capacity to counteract the margin squeeze carried out by Correos, which is why the finding of abuse of dominant position could not be made. In particular, the CNMC found that Unipost had a network that allows it to access 90% of consignments and 70% of the population and that its market share had apparently been increasing. This led the CNMC to the following important conclusion, which was often cited and debated by both Audiencia Nacional and the Supreme Court (2018:7):

Unipost could use its own capacity to reach the target population without resorting to Correos’ network. In these circumstances, Correos’ refusal to sell the use of its network at a price equal to that offered to large customers, in no way prevented it from competing for those customers. It cannot therefore be said that Correos, with its discount policy, prevented its main competitor, an alternative operator, from competing for the large customers’ business in its entirety.

Another margin squeeze decision was handed down in December 2017 in Italy where, following a complaint filed by Nexive, Poste Italiane’s main competitor, the AGCM found that Poste Italiane abused its dominant position in the Italian bulk mail delivery market through several behaviors that together amounted to a single exclusionary strategy (AGCM 2017). These included, in particular, technical non-replicability of the incumbent’s offers, margin squeeze, as well as various types of rebates. According to the AGCM, Poste Italiane prevented its competitors from entering the downstream market for deliveries of bulk mail in extra-urban areas, where only Poste Italiane had the infrastructure to deliver the mail, by charging them wholesale prices for bulk mail delivery services that were higher than the retail prices it offered to its business customers for the same service. As a result, even competitors as efficient as the incumbent could not match prices that it was able to offer to its final customers. The AGCM imposed on the incumbent operator a fine of 23.1 million EUR.

In 2015, the BKartA found that Deutsche Post, holding well above 80% market share for licensed postal services, abused its dominant position by granting four large customers (Deutsche Telekom, Vodafone, Telefónica, and Freenet), through so-called target price agreements, discounts below the prices its competitors had to pay for access to its mail delivery network (Bundeskartellamt 2015). The BKartA, just as the CNMC and the AGCM, found that this conduct constituted an abusive margin squeeze. It hindered alternative operators from competing with Deutsche Post as they were impeded to compete for the large customers in question.

In contrast to the BKartA, the CNMC, and the AGCM, Ofcom investigated certain behaviors by Royal Mail as price discrimination in the first abuse of dominance decision that Ofcom issued against Royal Mail (Ofcom 2018). It concerned the market for bulk mail delivery in the UK, where the incumbent operators had a market share in excess of 98%. Whistl, a subsidiary of the Dutch postal incumbent, TNT, was the “first ever delivery competitor to pose a serious challenge to Royal Mail’s effective monopoly in the delivery of letters.” In its complaint to Ofcom, Whistl argued that in January 2014 Royal Mail proposed a number of changes to the terms and conditions of its price plans Contract Chance Notices (CCN) that would inhibit its ability to compete, requiring it to cease its own delivery operations should it wish to use a lower priced access service (Ofcom 2018, p. 1.14). Ofcom was careful to stress that it was not a case of “a potential example of margin squeeze involving the relationship between Royal Mail’s upstream and downstream prices and costs” and that it was “not seeking to establish whether the prices charged to all customers by a dominant undertaking make it impossible for a competitor at the same level of the dominant undertaking to compete on price” (Ofcom 2018, p. 7.196).

To properly understand the case, it is necessary to consider ex ante regulation and in particular Ofcom’s 2014 consultation on a possible modification of the regulatory access regime. During that consultation, Ofcom found that Royal Mail’s “zonal tilt”Footnote 15 and “price differential”Footnote 16 could discourage entry (Ofcom 2014). To address these concerns, Ofcom put forward some proposals. However, as Whistl ceased its operations in the delivery market in June 2015, Ofcom chose not to move forward with these proposals as it considered them inappropriate and disproportionate given that the entry of “another end-to-end entrant of sufficient scale and scope to provide a significant level of letter delivery competition to Royal Mail in the foreseeable future” was unlikely (Ofcom 2017, p. 5.12). Yet, Ofcom’s abuse of dominant position investigation carried out shortly after the consultation revolved around just one of the two issues, namely, the price differential. It is thus striking that while both the zonal tilt and price differential were considered inimical to entry from ex ante perspective, only the latter was examined ex post in the abuse of dominance case.

In its decision, Ofcom committed itself to demonstrating that Royal Mail intentionally hampered Whistl’s entry in to the bulk delivery market. Although proof of intent is not necessary for establishing an abuse of a dominant position, it can nonetheless be relevant. In the postal sector, incumbent operators will likely invoke intent as a defense, arguing that the conduct in question was motivated by the need to ensure delivery under the Universal Service Obligation (USO). Royal Mail made this argument, as has Poste Italiane.Footnote 17 In both cases, NCAs disagreed. In rejecting Royal Mail’s argument, Ofcom relied on the incumbent’s internal document, which indicated that Royal Mail opted for a strategy that, if implemented, would have led to a lower loss of market share (just 1.4%). While this strategy involved both the zonal tilt and price differential, only the price differential was analyzed by Ofcom in its investigation.Footnote 18

4 What Do Comparisons of Margin Squeeze in the Postal and the Telecoms Sectors Tell Us?

In light of the outcomes of these cases, we would like to concentrate the rest of the analysis on two important issues – essentiality of the incumbent’s network and the difference and relationship between margin squeeze and discriminatory pricing. These are directly relevant for understanding the different way in which Ofcom framed its analysis of Royal Mail’s conduct compared to the other NCAs postal margin squeeze cases.

4.1 Is the Legacy Postal Network an Essential Facility or Just an Important Input?

This is clearly a topical question. If Unipost could rely on its own capacity to reach 90% of consignments and 70% of the population, then another part of the CNMC’s decision that seems to consider Correos’ network to be an essential facility, without explicitly stating so, is perplexing. CNMC, for example, notes that “the postal network managed by Correos is difficult to replicate, since private operators do not enjoy sufficient economies of scale to bear the costs of deploying their own network that could replicate the coverage and capillarity of the Correos’ network” (emphasis added, CNMC 2014). If there was any doubt that “difficult to replicate” may still not mean impossible, in another paragraph, CNMC explicitly states that “it is not economically viable for competitors to replicate the coverage and capillarity of Correos’ postal network.”

In Europe, many commentators supported including essentiality of the input into the margin squeeze test (Crocioni and Veljanovski 2003; Gerardin and O’Donoghue 2005). Also, ERG’s Guidelines on the application of the margin squeeze test to bundles, as well as various passages from Deutsche Telekom and Telefónica decisions, imply that even if the Commission’s telecoms decisions, confirmed by the Courts, do not require the upstream input to be strictly essential, that characteristic may nevertheless play an important role in margin squeeze cases.

The argument re-emerged in the national margin squeeze cases. The Bundeskartellamt’s (BKartA) decision issued in 2015 against Deutsche Post and AGCM’s decision from 2017 are other cases in point. Deutsche Post and Poste Italiane, just as Telefónica, disputed the existence of the abuse arguing that access to their networks is not indispensable, as network can be replicated, and that some competitors had their own delivery networks. Poste Italiane, for example, went so far as to assert, erroneously according to the AGCM, that its network could be duplicated in 99.8% of the cases. As the AGCM (2017, p. 234) pointed out:

the development of alternative networks throughout the national territory, and in particular in the extra-urban areas, have a structural limit which is caused by their non-profitability. In fact, both the geomorphic characteristics and the lack of demand for postal services due to low population density make the development of an additional network economically unsustainable compared to that of Poste Italiane which, in fact, ensures 100% coverage of the postal codes and population only inasmuch as it benefits from public contributions aimed at covering the losses deriving from the provision of postal services in situations of market failure.

Moreover, even the most innovative and efficient competitors, such as Nexive and Fulmine, which have heavily invested in the development of alternative end-to-end networks, covering, respectively, 80% and 75% of the population, cannot develop them further. This led the authority to conclude that “the Poste Italiane’s service is an essential input for competitors to complete their offer. In other words […], Poste Italiane’s services to competitors constitute the ‘indispensable service’ referred to by the law on margin squeeze” (AGCM 2017, p. 234). Despite these observations, the AGCM, just as the BKartA, formally followed the Commission’s position and found that the indispensability of a wholesale product is not required in order to prove potentially anticompetitive effects of a margin squeeze.

However, Advocate General Mazák in its opinion in TeliaSonera argued that when other inputs are available, and therefore a given input is not essential, margin squeeze cannot be implemented simply because a dominant firm’s competitors are no longer dependent on its upstream input.Footnote 19 Such view is consistent with the analysis applied in vertical mergers, where competition authorities have to consider, among others, the ability to substantially foreclose access to input. If alternative operators do not need access to a wholesale product to provide their own service on the retail market, then in case of margin squeeze, they would simply bypass the wholesale product in question. This means that the incumbent’s ability to foreclose access would be just theoretical, and not effective, which is what ultimately matters if foreclosure is to have a significant detrimental effect on downstream competition.

Although the CJEU in its decision did not follow AG Mazák’s position, it has nonetheless acknowledged that “the question whether the wholesale input is indispensable may be relevant when assessing the effects of the margin squeeze.”Footnote 20 Arguably, paragraphs 70–72 of the TeliaSonera ruling imply that if an input is indispensable, and competitors suffer a competitive disadvantage from the squeeze, an anticompetitive effect can be presumed. On the other hand, if an input is not indispensable, the competent authority has to consider whether a margin squeeze is capable of producing anticompetitive effects, as it would seem that such effects would be less likely.

For the CJEU, indispensability is not a prerequisite in margin squeeze cases. Rather, just as in the EU Non-Horizontal Merger Guidelines, it becomes relevant only to the extent it can inform the analysis of the likelihood of foreclosure (Pisarkiewicz 2018). In any event, the concept of essentiality, as defined in Bronner, may be too stringent to be meaningfully applied in continuously evolving telecommunications and postal markets. New business models may change the modes of competition, while technology could allow the bypass of existing bottleneck and drastically change the demand for core services.

4.2 Margin Squeeze and Discriminatory Pricing

As we mentioned earlier, the AGCM, BKartA, and CNMC investigated the conduct of their respective postal incumbents primarily as a margin squeeze, but Ofcom opted for price discrimination. According to Ofcom (2018, p. 7.139(c)), “given the nature of the discrimination in issue, the type of foreclosure effect […], and the prevailing conditions of competition in the market at the time [the] conduct took place, it [was] neither necessary nor appropriate […] to carry out an AEC test (or any other type of price-cost test) in assessing the likely effect of the price differential.”

Margin squeeze may itself be a by-product of price discrimination, which could violate Article 102 TFEU. When it is possible to subsume a particular abuse under different legal labels, competition authorities have to choose the category that in their view will capture best its potential anticompetitive effects. In margin squeeze cases, authorities are mainly concerned about the insufficient profit margin that would inhibit existing and potential competitors from successfully challenging the incumbent’s position. In case of a discriminatory squeeze, discrimination would exist between the incumbent’s own downstream arm and its competitor. In case of the postal sector, however, discrimination may affect just a specific category of the incumbent’s customers, and this is precisely what happened in the Royal Mail case. According to Ofcom, the prevailing concern was that price differential amounted in effect to a penalty on access customers seeking to compete in bulk mail on end-to-end delivery. In the case of Whistl, an end-to-end operator that sought to expand its entry in the bulk delivery mail beyond a very limited number of standard selection codes (SSCs) would no longer be able to rely on a lower nationally priced NNP1, but would have to switch to higher APP2/ZPP3 zonal charges.

According to Ofcom, the impact of the price differential would have been material. In a market such as that of bulk mail delivery, not only competition was already limited, but growth would have to be achieved in the context of a declining market, which makes entry and expansion increasingly difficult over time (Ofcom 2018, p. 7.162). If achieved, however, competition in delivery services could trigger significant benefits as it accounts for the major part of the value chain, in contrast to the much more competitive segment of collection and initial sortation services, which accounts approximately for only 10%. This is why Ofcom saw discrimination by Royal Mail as an attempt to penalize any growth in bulk mail delivery competition beyond a limited scale as a central competition issue.

According to the Revised ERG Common Position on Remedies, which concerned the telecoms sector, margin squeeze, rather than as potentially anticompetitive conduct, can be seen as an anticompetitive effect that can arise in a vertical leveraging context through price-based as well as non-price-based abuses.Footnote 21 The European Regulators Group states that “although margin squeeze also has a behavioral aspect, it is classified as an effect here, as it can be the result of different behaviors of the dominant undertaking” (ERG 2006). This distinction becomes particularly important when the authority has to design remedies, as it will make a difference which particular behavior has led to margin squeeze. If potential harm to competition is clearly related to discrimination, it may be more effective to frame the behavior as such rather than as margin squeeze as the finding of a violation would also imply the obligation to cease anticompetitive discrimination.

As Ofcom’s Royal Mail Access Pricing Review reveals, Ofcom has for a long time been concerned about the impact of discrimination. Although “the USPA Condition […] prohibits undue discrimination in relation to matters connected with access […] this condition would not necessarily prevent Royal Mail from seeking to […] differentiate between operators of varying sizes and mailing profile […] which could result in different prices being available as between access-only and end-to-end operators” (Ofcom 2014, p. 4.48). Ofcom believed that uncertainty faced by an end-to-end competitor as to whether they were likely to be able to rely on access prices on similar terms as access-only operator required urgent intervention. Seen from this perspective, the choice of discriminatory pricing rather than margin squeeze as an analytical framework for assessing the impact of Royal Mail’s behavior is clearly more understandable and justified.

5 Conclusions

Over the years, margin squeeze has become a frequently alleged and analyzed type of abuse in the EU. However, it has been associated mostly with the electronic communications sector, where landmark cases have been investigated at the European level by the DG Competition of the Commission. In contrast, recent major margin squeeze cases in the postal sector were investigated only by national competition authorities. Their decisions show that in the postal sector, margin squeeze was often coupled with, or resulted from, other practices and in particular from price rebates or discrimination. In contrast, in the telecoms sector, margin squeeze was mostly examined on its own. Comparing the decisions of national competition authorities in similar cases confirms that, where it is possible to subsume a particular abuse under different legal labels, competition authorities choose the category that in their view captures best the potential anticompetitive effects of a given conduct considering its nature, the type of foreclosure, and the prevailing conditions of competition in the market. As for the debate about a requirement of essentiality of the incumbent’s input for the competitors, we have noted that while NCAs formally follow the Commission’s stance and do not require it in margin squeeze cases, they nonetheless seem to recognize that legacy postal network often provide incumbents with the kind of advantage which in other contexts has led to the very definition of essentiality.