To Bid or Not to Bid, That is the Question: The Assessment of Bidding Markets in Merger Control

David Wirth is a consultant at Ashurst LLP. This post is based on a chapter by Mr. Wirth, first published in the International Comparative Legal Guide to: Merger Control 2017 (contributing editors Nigel Parr and Catherine Hammon, Ashurst LLP), published by Global Legal Group. The complete chapter, including footnotes, is available here.

Many markets are characterised by an auction process in which customers issue tenders for contracts and suppliers bid against each other in order to win that contract. Prices in such markets, which are commonly referred to as bidding markets, are typically individually determined for each contract, and the buyer (i.e. the tendering authority) is able to compare offers and negotiate with the different competing bidders.

The nature of bidding markets involve market features and dynamics which differ from ordinary markets (i.e. where prices are set individually by each supplier and customers decide whether or not to purchase the goods or services in question). This, in turn, can have an impact on the nature of the competitive assessment in those markets, particularly in the context of the assessment of mergers between competing suppliers. However, a detailed discussion of bidding markets is noticeably absent from the merger guidelines of a number of competition authorities around the world.

The aim of this post is not to provide a detailed exposition of the economics of auction theory, but to outline the nature of the competition issues that often arise in bidding markets, and to summarise the type of empirical evidence that can be presented to inform the assessment of the competitive effects of mergers in such markets. The rise in the number of markets that involve tender processes (particularly in e-commerce and business-to-business markets) suggests that the economics of bidding markets is becoming increasingly important to merger control and should be given more consideration in the competition authorities’ guidelines.

Unilateral Effects in Bidding Markets

It is noticeable that the merger control guidelines of various competition authorities around the world (including the UK’s Competition and Markets Authority, the European Commission, and Australia’s ACCC) are either completely silent or say very little about the assessment of bidding markets. The US horizontal merger guidelines, however, explain that “anti-competitive unilateral effects in these settings are likely in proportion to the frequency or probability with which, prior to the merger, one of the merging sellers had been the runner-up when the other won the business.” The US guidelines also go on to state that “these effects are likely to be greater, the greater advantage the runner-up merging firm has over other suppliers in meeting customers’ needs.

The approach set out in the US horizontal merger guidelines is consistent with the approach adopted by the UK’s Competition Commission (CC), the predecessor to the Competition and Markets Authority (CMA). In its recent Tradebe Environmental/Sita (2014) report, the CC explained that “to win a tender process the winner must beat the next best offer. Unilateral effects will be most likely where the parties ranked first and second in tenders merge.” The CC went on to state that “the merger is most likely to be harmful in situations where the merger parties were the two most competitive bidders for a customer as this constraint is then removed.” A similar analytical framework has been set out in a number of other UK cases, notably in relation to Capita/IBS (2009) and Stericycle/Ecowaste (2012).

The basic intuition behind unilateral effects concerns in bidding markets is that, unless the merging parties are the two lowest priced bidders for tenders (and are constraining each other in the prices at which they bid for tenders), competition is unlikely to be adversely affected by the merger. This reflects the fact that:

  1. for tenders in which neither party is the lowest priced bidder, a merger that results in the parties now bidding as one cannot result in higher prices as both parties were previously being undercut by a rival (or rivals), and this would not be expected to change as a result of the merger; and
  2. for tenders in which one of the parties is the winning bidder, but a rival is the runner-up, prices cannot be higher if the merging parties bid together as their prices are ultimately being constrained by the prices of the second ranked supplier, which is not expected to change as a result of the merger.

Accordingly, as explained by the US horizontal merger guidelines and in decisions of the CC, it tends to be in situations in which the merging parties are ranked first and second in tenders in which unilateral effects arise (as the price is ultimately determined by the prices of the runner-up). In such situations, the magnitude of the adverse effects will be greater the larger the headroom that exists between the second and third ranked bidder as the merged entity will be constrained, post-merger, by the price of the third ranked bidder.

The relevance of setting out this framework for assessing unilateral effects is that in bidding markets, the usual inverse relationship between the number of competitors and the level of prices does not necessarily hold. For example, a 4–3 merger between two parties that are never the two most competitive bidders is unlikely to give rise to unilateral effects concerns as the parties are not imposing a competitive constraint on each other. Despite this clear conceptual background, the competition authorities have been unable in some cases, particularly at Phase 1, to get comfortable that a merger involving bidding markets does not give rise to competition concerns.

For example, in the recent Xchanging/Agencyport (2015) merger case which was referred to Phase 2 (and cleared), the CMA’s Phase 1 press release stated that “[t]he merger will reduce the number of major suppliers of PAS software for these users from 4 to 3,”and it was explained that the merger may give rise to a lessening of competition as a result of facing competition from one fewer competitor. Despite the same bidding data being presented at Phase 2, the CMA then concluded that: “On the basis of the tendering data that was available to us, the Parties therefore appear not to provide a particularly strong competitive constraint on each other and appear to face competition from NIIT and Sequel, as well as four other suppliers.

Whilst the description above provides a conceptual framework against which the parties’ bidding data can be tested, the US horizontal merger guidelines also explain that the mechanisms and likelihood of anti-competitive unilateral effects can differ according to the bargaining practices used, the auction format, the sellers’ information about one another’s costs and about buyers’ preferences. It is also relevant to consider the capacity of rival bidders—whilst the parties may never have been the first and second ranked bidders in tenders, if the capacity of rivals is limited, then the successful award of a few large contracts can reduce the constraint exerted by these rivals in future tenders and the parties may become more highly ranked.

The framework for the assessment of unilateral effects in bidding markets also raises the prospects of targeted anti-competitive effects from mergers. This reflects the fact that the merging parties may compete against one another for some customers, but not for others, and third parties may only be close substitutes for certain customers. For those customers where the merging suppliers are best and second best placed to serve (i.e. in terms of price and product characteristics), the parties may be able to target those customers with higher prices post-merger, even if they are not the preferred bidders in other tenders.

The Relative (Un)Importance of Market Shares

Views often differ as to the role that market shares play in assessing markets characterised by tenders and, in particular, whether market shares provide a reliable indicator of market power. In its Xchanging/Agencyport reference decision, the CMA stated that: “In most cases it has concluded that market concentration was a relevant parameter of competition even where customers used tenders to appoint their suppliers, and that a reduction in the number of competitors could harm customers.

However, it is widely acknowledged that in markets where companies bid for contracts, and contracts are large, infrequent and long-term, historical market shares (which represent legacy contracts) do not provide a reliable indicator of current market power. At the extreme, in relation to a large one-off contract with a long duration, there will often be only one winner and the successful bidder will have a very high market share as a result of winning the contract. In such situations, competition takes place “for the market” at the bidding stage, which forces prices to remain at the competitive level. Accordingly, the high market share of the successful bidder does not indicate market power, as it fails to reflect the competition that has taken place for the contract at the bidding stage. In such circumstances, it is clear that any reliance on market shares will provide an unreliable and misleading picture as to the nature of competition in the market.

More generally, the greater the size of the contracts, the smaller the number of tenders and the longer the duration, the more likely it is that market shares are a poor proxy of market power. For example, in Siemens/VA Tech (2005), the European Commission stated that “the fact that there is bidding on a market does not in itself allow any conclusion to be drawn as to the intensity of competition to be expected, or as to the significance of market shares as an indicator of possible market power. The key factor is rather the bidding pattern in individual cases. For example, even where there is a small number of credible bidders, particularly intensive competition is to be expected if, in a bidding market, a large proportion of tenders is awarded in a few large transactions […] In this and similar cases, market shares would, in practice, provide very little information on the possible market power of a bidder.

In the context of large and infrequent contracts, a supplier’s market share at any particular point in time is unlikely to reflect accurately the competitive constraint that it will exert on the next “competition” (i.e. the next contract to come up for renewal). Indeed, the winning or losing of one contract can completely change the market share profile of the different suppliers in the market. For example, in its assessment of the acquisition by Northgate of Anite, the OFT concluded that “given the length of the contracts negotiated in this instance the OFT considers it likely that market shares estimated on a legacy basis may give a distorted picture of the recent competitive situation.” The OFT went on to state that “the set of data which relates to newly tendered offers within the last five years, offers the best measure of recent competitive interactions.

Following this reasoning, market shares may provide a more reliable indicator of market power where contracts are relatively small, frequent and with short durations. In these circumstances, market shares are more likely to reflect the current state of competition in the market (i.e. which firms have been successful in the most recent tenders). It may also be the case that, in certain situations, market shares reflect the ability of suppliers to make credible bids in future tenders (e.g. where reputation, experience or brands are important).

Market shares are not, therefore, irrelevant in the assessment of mergers in bidding markets, although they become less relevant the larger and more infrequent the contracts put out to tender. In general, however, an assessment of reliable and up-to-date bidding data from tenders (e.g. over a five-year period) is likely to be more relevant as it allows the closeness of competition between suppliers at the bidding stage to be assessed, which may not be adequately reflected in market shares.

When Two Bidders are Enough

It is often suggested that in bidding markets, the presence of fewer bidders may still secure a more competitive outcome than may otherwise be the case in ordinary non-bidding markets (and indicated by market shares). Economic theory predicts that under certain circumstances, just two bidders is sufficient to generate competitive outcomes. For example, where two firms are identical (in terms of both the product offering and costs) and have full information about each other, economic theory predicts that market outcomes will be very competitive with just two bidders (assuming that there are no capacity constraints). This is because any bid above cost can be expected to be undercut by the other bidder, which will result in the tender being lost.

In Stericycle/Ecowaste (2012), the CC acknowledged (based on a report titled “Bidding Markets” for the CC by Professor Paul Klemperer) that competition in bidding markets can result in “competitive outcomes with fewer firms than might otherwise be the case,”and went on to note that under certain circumstances, “two firms may provide enough competition to give rise to a fully competitive outcome.” The circumstances in which just two bidders are enough to secure a competitive outcome is often referred to as an “idealised bidding market.” The “Bidding Markets” report explains that this can only occur where the following conditions hold:

  1. competition is “winner takes all,”so each supplier either wins all or none of the contract;
  2. competition for contracts is “lumpy,”so that each contest is large (and therefore important) for a supplier relative to its total sales over a period; and
  3. competition begins afresh for each contract (and for each customer), such that when tenders are repeated the outcome of a previous tender does not impact on the outcome of another (Klemperer explains that this roughly corresponds to tenders taking place infrequently).

The intuition here is that in lumpy, infrequent and winner takes all markets, bidders will be forced to bid at their most competitive price in order to stand a realistic chance of winning the tender. Failing to win the tender could have a significant detrimental impact on their business, which forces bids down to a competitive level. Even where the conditions for an idealised bidding market are not met, it may still be possible to reach competitive outcomes with few bidders (but more than two) as acknowledged by the CC in Alpha Flight Group/LSG (2012), which was cleared with just three credible bidders.

There are examples in which 3–2 mergers in bidding markets have been cleared. In 2007, the European Commission cleared the Syniverse/BSG merger despite the number of bidders being reduced from three to two (with the parties having broadly equal shares of supply). The European Commission explained that “[t]he analysis of the bidding data, gathered through the market investigation has shown that BSG and Syniverse were never the only two bidders in a tender. In fact, in the cases where only two clearing houses took part in a tender, they were either BSG and Mach or Mach and Syniverse. This means that in nearly half of the cases, the MNOs determined that the tender process was sufficiently competitive with two bidders, namely Mach and BSG or Mach and Syniverse.” In addition, for tenders in which more than two bidders were invited to bid, the Commission found that the parties were winner and runner-up in only one out of 20 tenders, with Mach being runner-up in all other tenders won by either Syniverse or BSG.

This case is interesting for two principal reasons: firstly, the Commission’s reasoning relied on the fact that customers were often satisfied with inviting only two firms to participate in tenders to secure a competitive outcome (notwithstanding the presence of additional potential suppliers pre-merger), which indicated that two bidders was sufficient; and secondly, in tenders where only two suppliers were invited to bid, the merging parties were never bidding against each other, which indicated that they were not close competitors whilst the largest supplier, Mach, which participated in all tenders, was the main competitive force to both merging parties.

In 2014, the UK’s CC cleared the Tradebe/Sita merger despite this also being a 3–2 merger. Despite acknowledging that a 3–2 merger usually raises concerns, the CC conducted a ranking analysis of 20 tenders that had taken place over the previous four years (relating to around half of all customers) and found that the largest supplier in the market, Stericycle, was either the most competitive or the second most competitive bidder in all cases. The CC found only two examples where Tradebe and Sita were the two lowest price bidders, and these were both outside the relevant geographic market. The CC also found no examples where customers switched between Tradebe and Sita or vice versa. On the basis of this analysis, the CC concluded that “while the merger parties regularly bid in the same tenders, they did not constrain each other’s pricing to a significant extent.

These cases are interesting as they follow the framework for assessing unilateral effects as set out above in focusing on situations where the parties are the first and second ranked bidders in tenders. Whilst these decisions emphasise that the assessment is very much fact specific, they indicate that 3–2 mergers may be acceptable but only in situations where either the parties rarely bid against each other, or, where they do bid against each other, they are hardly ever ranked first and second by customers.

Quantitative Evidence in Bidding Markets

To assess whether a proposed merger may be expected to give rise to unilateral effects concerns, there are a number of different types of analysis that could be undertaken, subject to the relevant data being available. These include the following:

  1. participation analysis;
  2. customer preference (ranking) analysis;
  3. regressions between prices and the number/identity of bidders; and
  4. the assessment of win/loss data (i.e. customer switching).

Each of these quantitative techniques is discussed in turn below.

a) Participation analysis

The simplest type of quantitative analysis that can be undertaken in bidding markets is to consider the extent to which the merging parties participated in tenders, and, in particular, participated in tenders against each other. At its simplest level, this requires information that identifies, for each tender that has taken place over a period of time in the relevant market, the identity of the bidders that took part in the tender.

The intuition behind the analysis is that, if the parties are close competitors to one another, they will tend to bid against each other more frequently in tenders than other suppliers. Conversely, if the parties do not frequently face each other in tenders, then this would suggest that they generally do not compete for the same customers and are therefore not close competitors.

Whilst participation analysis is a technique that is commonly undertaken in merger analysis (and is relatively easy), it does not directly address the concept of unilateral effects in bidding markets. Just because the parties may have bid for a number of the same tenders, it does not follow that they were ranked first and second in such tenders, or that they were imposing a competitive constraint on each other (unless they were the only two bidders). Such a finding would suggest that further analysis (e.g. the ranking analysis discussed below) is required as the participation analysis does not provide any further insight into the relative strengths of the different tender offers or customer preferences. This limitation in the analysis was acknowledged in the Capita/IBS report, where the CC acknowledged that “our analysis of the bidding data largely refers to how many times suppliers bid against each other, which in itself does not demonstrate fully the competitive constraint imposed by software suppliers.

Conversely, if the participation analysis shows that the merging parties do not participate in the same tenders, then more can be inferred as to the likelihood of unilateral effects concerns because this indicates that the parties were not ranked first and second choice by customers for such tenders. Accordingly, whilst participation analysis is not determinative as to which cases give rise to unilateral effects concerns, it can be informative for identifying which cases are unlikely to give rise to unilateral effects concerns, and can usefully be employed at Phase 1 as an initial filter.

For example, in Capita/IBS (2009), which was referred by the OFT to the CC for a Phase 2 investigation, and subsequently resulted in an adverse finding, the OFT found that Capita and IBS had a significant share of newly tendered contracts, with Capita winning 40–50% of contracts and IBS 20–30%. In addition, the parties bid against each other in the vast majority of cases and one or the other of them was successful in 70–80% of occasions when they both tendered. At Phase 2, the CC supplemented this analysis with a customer preference (ranking) analysis (discussed further below), and found that the customer responses listed Capita and IBS as the top two bidders in their most recent tenders more than any other pair of software suppliers (with the parties being listed as the top two bidders for nine out of the 15 tenders for which the CC received information).

In contrast, in the recent Pork Farms/Kerry Foods (2015) merger, which was cleared at Phase 2, the CMA noted that the parties competed head-to-head to supply retailers in only a limited number of cases (i.e. in 11 out of the 60 tenders that had taken place over the previous five years). This means that, irrespective of any ranking analysis, the parties did not compete against each other for over 80% of tenders that had taken place (and were not therefore viewed as being close competitors by a large number of customers). Moreover, the CMA found that in only one tender were the parties identified as the only bidders, which indicated that they were ranked first and second in the tender.

b) Customer preference (ranking) analysis

As noted above, a customer preference (or ranking) analysis can usefully be undertaken in situations in which, as well as the name of the bidders for tenders, the ranking of the losing bidders can be established. However, as explained further below, this often relies heavily on data being made available from customers, which is generally not available to the suppliers that participated in such tenders (and therefore relies on the information being gathered by the relevant competition authority during the merger control process).

The benefit of this type of analysis is that it is directly linked to the theory of unilateral effects concerns mentioned above. In particular, the second placed bidder for each tender can generally be considered to be the bidder that exercises the strongest competitive constraint on the winner of that tender. If the data revealed that one of the merging parties often ranks second behind the other merging party, this provides strong evidence that the merging parties are each other’s closest competitor, and that the merger is likely to lead to higher prices (although, as noted in the US horizontal merger guidelines, the magnitude of the adverse effects will depend on the amount headroom that then exists between the second and third ranked bidder).

As mentioned above, the European Commission conducted a ranking analysis in Syniverse/BSG in relation to the tenders where more than two suppliers were invited to bid. The Commission found that “Syniverse and BSG were only very rarely both the winner and runner-up in the same tender. In particular, out of 20 tenders for which customers’ information was available, BSG and Syniverse were ranked first and second in only one case. With the exception of this one case, Mach was the runner-up in all the tenders that were won by BSG or Syniverse.” This analysis appears to have played an important role in the European Commission deciding to clear what was a 3–2 merger. Similarly, in Tradebe/Sita (2014), the CC conducted a ranking analysis of 20 tenders over the previous four years (relating to around half of all customers) and found that the largest supplier in the market, Stericycle, was always either the most competitive or the second most competitive bidder. The CC found no examples where Tradebe and Sita were the two lowest price bidders within the relevant geographic market.

The CC also conducted a ranking analysis in Stericycle/Ecowaste (2012). In that case, the CC concluded that the parties were each other’s closest competitor on the basis that “Ecowaste South West was ranked in the top three bidders in four of the six tenders. It was the first in one and was the closest competitor to what is now Stericycle in two others. Only one other bidder was ranked in the top two bidders with Stericycle more than once.” The CC also noted that some rival suppliers face capacity constraints, which may limit their ability to bid at the most competitive price in future tenders.

Whilst customer preference (ranking) analysis will often require information from customers, to the extent that the parties are shortlisted as the last two suppliers in tenders (and are aware of such an outcome), then this can be used to infer rankings. For example, in Safenet Inc/nCipher (2006), the OFT concluded that the parties were the closest competitors on the basis that they were regularly shortlisted as the last two suppliers at the tender stage, which indicated that the parties were ranked first and second for those tenders. In contrast, in Northgate/Anite (2008), the OFT concluded that Anite was rarely the only bidder against Northgate, and where it did bid against Northgate, Anite’s tenders were often declined at the shortlisting stage, implying that they rarely (if ever) submitted the two most competitive bids.

One of the main issues with this type of analysis, however, is that it does not necessarily establish the magnitude of any anti-competitive price effects. If customer preferences are such that they are largely indifferent between the second, third and fourth bidders, for example, then the merger between the first and second bidder may be unlikely to generate any substantial anti-competitive price effects as there are a number of other very close substitutes. Accordingly, simply obtaining information on customer tender rankings (i.e. on a discrete choice basis) may not provide the additional insight required as to the closeness of competition from the alternative bidders, which is required to establish the magnitude of the anti-competitive effects.

c) Regressions between prices and the number/identity of bidders

If data on prices, discounts or margins offered by the parties in each tender and the number of bidders are available, it may be possible to use econometrics to examine the relationship between the number of bidders, and the identity of bidders, and the prices bid. This analysis can potentially provide evidence on:

  1. how many bidders are enough to secure a competitive outcome;
  2. which competitors tend to result in lower priced bids in tenders, and, in particular, whether the presence of both merging parties results in lower priced bids (e.g. compared to a situation where only one of the parties bid); and
  3. how the reduction in the number of bidders resulting from the merger will affect prices in the market.

To the extent that prices tend to decrease systematically as the number of bidders increases, this suggests that a merger between two bidders might give rise to unilateral effects, even if the two parties cannot be characterised as particularly close competitors. Such an analysis is similar to, and raises many of the same issues as, a price/concentration analysis in relation to ordinary markets (which is often used to assess prices and levels of concentration across a number of different geographic areas).

A variation on this analysis is to consider whether the parties offer lower prices when both parties bid for tenders compared to situations in which only one of the parties bid. If one of the merging parties is found to offer lower prices when the other merging party participates in the tender (compared to tenders where the other is not present), then this would indicate that the proposed merger may eliminate an important competitive constraint and, therefore, that the merger may give rise to price rises. Alternatively, if the prices offered by each merging party are largely unaffected by the presence of the other merging party, this would indicate the opposite effect (i.e. that the parties exert nothing more than a weak competitive constraint on each other).

Where such price comparisons are being made, it is important that a series of control variables are used to take account of factors other than the number (and identity) of bidders that may influence prices. As tenders are often tailored to the specific customer requirements, for example, it is important that such variations in the quality and size of contracts are taken into consideration. Unless these factors are controlled for in the analysis, it may present a misleading picture of any relationship between the number/identity of bidders and prices (e.g. the lower prices observed may be as a result of contracts having lower volumes, and be unrelated to the number of bidders). Econometric techniques may be able to control for some of these factors, but it will depend on the relevant data being available.

In Oracle/PeopleSoft (2004), the European Commission carried out an econometric analysis of the parties’ bidding data with the aim to investigate whether the level of discounts offered were affected by the number and identity of final-round bidders. The European Commission found that once the size of the contract was taken into account, the number of final bidders did not affect the level of discounts offered, and there was no clear evidence that the presence of a particular competitor led to larger discounts. The Commission noted, however, that the analysis still had its limitations and therefore did not provide proof that the merger had no harmful effects (e.g. due to poor quality data, and the fact that analysing the identity of bidders in the final round only of bidding may provide an incomplete picture of the actual competitive process).

In Syniverse/BSG, the European Commission also conducted an econometric pricing analysis to supplement the evidence from the participation analysis and ranking study. The analysis sought to assess whether the prices that BSG bid varied according to whether or not Syniverse participated in the tenders, whilst controlling for various other factors that may influence prices (such as the contract length, the identity of the incumbent, the year of the tender, etc.). The Commission found that the “prices offered by BSG are unaffected by whether or not Syniverse participated in a tender, which implies that Syniverse does not exert a strong competitive pressure on BSG’s prices.

As with any econometric technique, it is also relevant to consider its potential limitations. In addition to deciding which control variables to apply, it is also important to understand how many and which suppliers the buyer asks to tender (as buyers often limit the number of suppliers invited to tender) as this could affect the measure of concentration. It is also important to understand the extent to which the one fewer bidder arising from the merger can simply be replaced by another bidder. For example, if the customer is largely indifferent between a number of suppliers, it may be able to invite other bidders to participate in future tenders relatively easily, which will not be identified by the analysis.

d) Win/loss switching data

In bidding markets, the closeness of competitors may also be assessed by the extent to which the parties have won/lost contracts from one another. This may provide evidence on churn (how often customers switch between suppliers), as well as the closeness of competition. The advantage of providing win/loss switching data is that it is based on data that is available to the merging parties (in relation to switching between the parties at least), although the overall level of churn in the market is unlikely to be known to the parties.

In Pork Farms/Kerry Foods (2015), the CMA said that it observed eight instances of switching between the parties (in relation to a sample of 60 tender/negotiation processes that had taken place). Contracts lost by Kerry were mostly to suppliers other than Pork Farms, while contracts lost by Pork Farms were equally split between Kerry and other suppliers. The relevance of such switching analysis in this case was that retailers and suppliers confirmed that incumbent suppliers were usually given an opportunity to respond in the event that a rival made a better offer to the retailer, indicating a degree of competitive pressure even where switching did not result.

It should be noted, however, that where long-term legacy contracts are concerned, the switching of customers from one party to the other may not accurately reflect the level of competition that has taken place at the most recent bidding stage. Indeed, the legacy supplier may not even have been invited to bid (or shortlisted) for the new tender. In such situations, the fact that a customer has switched from one merging party (i.e. the legacy supplier) to the other does not necessarily reflect the competitive constraints between the bidders at the most recent tender situation.

Informational Disadvantage of the Merging Parties

Whilst a number of the quantitative techniques set out above are relatively simple, a key feature of the assessment of bidding markets is that the merging parties may often be at an informational disadvantage as much of the key information in relation to tenders will be in the hands of customers. In particular:

  1. the parties will generally only be aware of the tenders in which they participated, which is likely to overstate their position (e.g. in any participation analysis) as it excludes all other tenders that may have taken place in the market where neither party was invited to bid;
  2. in many tender processes, the parties are not aware of the number, or identity, of the other participants. This reflects the nature of the auction process and the way in which tenders are structured. For example, in sealed bid auctions, the parties may not be aware of which other rivals were also participating in the tender (which can affect the ability of the parties to carry out any detailed quantitative analysis in relation to the impact of the number of bidders/identity of bidders on prices); and
  3. the parties will often be unclear as to outcomes of the tender process (i.e. what their ranking was, what prices were bid/discounts offered by other competitors, how their bid faired relative to other bidders, which rivals were considered for a shortlist, etc.). Without such information, it is difficult to carry out any customer preference (ranking) analysis.

This informational disadvantage creates challenges for the parties and their advisers during the merger control process, as it means they do not have full visibility of all the requisite data to inform the competitive assessment. This can also limit the range of quantitative techniques that can be employed to inform the assessment.

The issue of the data limitations of the parties has been recognised by the competition authorities in a number of cases. For example, in Tradebe/Sita, the CC explained that “we also received data relating to past bids from the merger parties and their competitors. This data is likely to be less useful than customer provided data as companies typically have less information regarding who else competed in the tender and how bidders ranked.” The European Commission’s ranking analysis in its Syniverse/BSG decision was also conducted on the basis of “ranking data received from customers,”and the CC’s ranking analysis in Capita/IBS was based on customer questionnaire responses.

Accordingly, as the key holder of the relevant tender information will often be the customer, it is important that the competition authorities identify and address the information asymmetries early in the merger control process, particularly at Phase 1, in order to avoid unnecessary references being made (and decisions being based on partial and incomplete information). It is important, however, that customer responses are, where possible, supported with contemporaneous documents (e.g. formal tender scorecards) so as to improve the credibility of the responses and eliminate gaming strategies by customers in an attempt to block a merger.

Buyer Power in Bidding Markets

The importance of buyer power as a constraint upon the exercise of market power by suppliers is widely recognised, and features in many jurisdictions’ merger control guidelines. In bidding markets, buyers may be able to employ a range of strategies to increase the sophistication of their procurement processes or change the tendering process altogether in order to tilt the balance in their favour. Such strategies could involve: threatening to re-tender contracts, delay tenders, or to delist products to extract lower prices; investing time and effort to identify and develop alternative sources of supply; engaging in multi-sourcing purchasing strategies to reduce the reliance on any one supplier; or designing contracts in such a way to achieve the best possible terms. The availability of large and/or long-term contracts may also be an attractive way to support new entry or expansion.

An assessment of countervailing buyer power played a key role in two recent (2015) Phase 2 cases involving bidding markets in the UK, namely Xchanging/Agencyport and the clearance of Pork Farms/Kerry Foods.

In Xchanging/Agencyport, the CMA satisfied itself that, despite high switching costs and infrequent switching (every 10 years or more), customers are sophisticated and well-informed purchasers. The CMA found that:

  1. customers conducted periodic reviews of the software offerings available in the market to assess whether or not they should switch supplier, and they put considerable effort into the purchasing decision, including obtaining detailed information on the different product offerings (including information from other customers’ experience);
  2. customers were found to assess price, quality and likely levels of servicing over the lifetime of the product when making purchasing decisions, and incentives to find the best option were considered to be strong given the adverse consequences of a poor product choice or implementation; and
  3. some customers opted to delay the time of purchase if they were not satisfied with the offers available, indicating a further strategy for exerting pressure on suppliers.

The emphasis given to this competitive constraint following the Phase 2 investigation is interesting as it reflected the evidence of sophisticated buyers putting significant effort into the buying process. It is also interesting as it did not feature as part of the Phase 1 investigation, perhaps reflecting the importance of Phase 2 interactions with third parties, particularly customers, in allowing more evidence to be gathered.

In Pork Farms/Kerry Foods, a deal where both parties supplied cold and hot pies, sausage rolls and pasties and slices to the food retail sector, the CMA reached the view that buyer power was likely to remain post-merger reflecting: (i) the commoditised nature of the products; (ii) no “must-stock” brands; and (iii) evidence that retailers (including smaller players) can and do switch all or part of their business between suppliers. The CMA found that where suppliers attempted to increase prices, this could provoke customers to issue a tender, or to threaten to move some or all of their volume to a rival.

The CMA also found that customers have various other strategies to apply pressure on suppliers, including:

  1. threatening to split volumes between suppliers and multi-sourcing;
  2. driving volumes between suppliers through the use of promotions; and
  3. threatening to switch volumes away from a supplier in other food categories.

Customers also monitored supply costs, conducted range reviews, and required suppliers to provide evidence to back-up any price increase requests. If the price increase request appeared to be unjustified, retailers confirmed that they would consider delisting and switching to alternative suppliers. The CMA also found that there was some evidence that retailers could support entry by providing a greater degree of certainty of orders to suppliers, and some retailers indicated that they would be willing to encourage entry or expansion.

Again, this finding is interesting given the conclusion reached after Phase 1 that insufficient evidence was received to allow the CMA to conclude that customers had buyer power, with retailers indicating that switching supplier can be difficult. Getting to the bottom of this issue appears to have been difficult for the CMA to accomplish within the Phase 1 timeframe (despite the extensive market testing which did occur) with Phase 2 allowing closer interrogation of third party evidence.

An example of contractual terms being used to generate buyer power in bidding markets was set out in Alpha Flight Group/LSG (2012), which was in relation to the supply of in-flight catering contracts. The CC explained that “airlines may be able to use contractual provisions to resist price increases, either by threatening to terminate the contract and/or by changing the overall value proposition for the caterer by imposing changes to service levels, the product specification or product sourcing.” The CC found evidence that the contractual terms of some in-flight catering contracts allowed airlines to switch provider, and also allowed airlines to change the terms of supply in such a way as to reduce the overall value proposition to the in-flight caterer. The CC concluded that this suggested that the balance of power rested with the airlines rather than in-flight caterers. The CC also found that airlines could leverage their catering requirements and supply contracts across a network of airports (and internationally) in order to improve their bargaining position.

Whilst the assessment of countervailing buyer power often relies upon the buyer’s outside options (i.e. the ability to switch to alternative suppliers, to sponsor entry, etc.), the tendering and contractual framework provides a mechanism in which buyers can impose a disciplining effect on suppliers, particularly in the context of relatively large and infrequent contracts. A consideration for the competition authorities, therefore, is whether, in the context of bidding markets, the market testing at Phase 1 might be more tailored to allow a more detailed investigation into the procurement process and the mechanisms available to customers to exert countervailing power.

Efficiencies in Bidding Markets

A recent UK case, Tradebe/Sita (2014), has raised the possibility of mergers in bidding markets giving rise to efficiencies. Notwithstanding the fact that this was a 3–2 merger, the CC concluded that the merger would give rise to variable cost efficiencies that were rivalry enhancing, which would impact on competition as the joint venture (JV) would be able to bid at lower prices in tenders (i.e. the merger would create a stronger second player in tenders to the current market leader, Stericycle).

Whilst the competitive dynamic in bidding markets suggests that a merger between the lowest and second lowest cost providers would be likely to give rise to unilateral effects concerns, it also means that if a merger between the second and third lowest cost bidders results in variable costs savings, then the combined entity bidding as one will be able to offer a greater competitive constraint to the winning bidder (on the basis that the costs of the second bidder will be lower post-merger).

On the basis of a detailed analysis of the average costs of Tradebe, Sita and the market leader, Stericycle, the CC concluded that Stericycle would win 13 out of 15 customers, and come second where it does not win (on the basis that it was the lowest cost provider). However, by combining the various different plants of Tradebe and Sita, the CC observed that the parties would be able to manage capacity better and to internalise more of the costs that were being outsourced to third parties, which would lead to an immediate reduction in the average treatment costs of the JV. Although this was partly offset by higher transport costs (as the parties own sites were located further away from the customer), the CC concluded that the parties’ average costs would be lower as a result of the merger, which would allow them to bid more competitively in future tenders.

Whilst this case was ultimately not cleared on the basis of efficiencies, as the bidding data (mentioned above) indicated that the parties were never ranked first and second in tenders, it demonstrates how efficiencies from mergers can create a more effective rival in markets involving a bidding process, thereby reducing the gap between the first and second ranked bidder.

The Need for a Forward-Looking Assessment

Although bidding data is often informative as to the nature of competition that has taken place in recent years, it is, by its very nature, historic data. This means that it may not be determinative as to the nature of competition that will take place at the next round of bidding, particularly where tenders are relatively large and infrequent, and where there is scope for innovation and technological advancements in the interim period. In such situations, it may also be relevant to conduct a more forward-looking assessment to see whether major innovations or developments are foreseeable, such that the parties may become closer competitors at the next round of bidding (or that the transaction has dynamic effects due to the potential loss of innovation between the parties).

This point was specifically acknowledged by the CMA in its Xchanging/Agencyport reference decision. In particular, the CMA stated that: “Even in cases where the CC has considered that anti-competitive effects would be most significant where the merging parties were likely to be the two most competitive bidders, it has performed a forward-looking assessment, taking account a variety of other factors including competitors’ business plans and incorporating differences in customer requirements in its assessment. In other words, it has not limited its analysis to a review of past tender data.

The CMA acknowledged, in this case, that the bidding data suggested that the parties were not each other’s closest competitors in the previous five years, but went on to say that it “considers it to be important to assess closeness of competition not only on the basis of recent wins and current offerings, but also on the basis of competitors’ plans for, and investment in, product development and improvement in their competitive offering.” Whilst this is a more speculative theory of harm (as any investment or new innovation does not guarantee the successful establishment of a strong competitive position), this indicates that the parties’ internal documents and business plans will be closely scrutinised to assess whether they would have become closer competitors absent the merger.


Bidding markets often involve market features and dynamics which differ from ordinary markets, with much less reliance being placed on market shares as an indicator of market power, particularly where contracts are relatively large and infrequent. The key focus in the assessment of mergers, however, remains on whether the parties are close competitors and, in particular, whether they are regularly the first and second ranked choices of customers, with the magnitude of the adverse effects dependent on the extent to which the next best alternative(s) are significantly less preferred than the merging parties.

Whilst there is a range of quantitative techniques that can be used to inform the competitive assessment in bidding markets, the limited availability of data is often a difficulty facing the merging parties (with customers generally holding the relevant tender information). It is important, therefore, that the competition authorities are rigorous in their market testing and gather the relevant information from customers early in the process, so as to be able to assess mergers within the appropriate unilateral effects framework, particularly at Phase 1. It is also relevant to “kick the tyres” to test what strategies customers are able to implement so as to maintain power in the tendering process over suppliers as highlighted by the two recent Phase 2 UK cases (Park Farms/Kerry Foods and Xchanging/Agencyport), as well as to consider whether there are any innovations or changes imminent that would have resulted in the parties becoming closer competitors absent the merger.

The complete chapter, including footnotes, is available here.

Both comments and trackbacks are currently closed.