UPP and Away with Mergers?
Upward pricing pressure (UPP) tests attempt to capture the upward pricing pressure from mergers based on the closeness of competition (diversion ratios) between the merging parties and the margins of customers recaptured as a result of the merger. This note considers the context and purposes of their use, the data requirements, the limitations, and the extensions or alternative tests. This provides the basis for a discussion about their usefulness as one merger assessment tool, amongst other tools.
Context — merger analysis aims and general approaches
The central question in any merger and from a competition perspective is whether the merging firms constrain each other’s behaviour, such that this constraining factor will be removed post-merger. If one is able to answer this, then it is possible to understand what the counterfactual to the merger is expected to be, and from this, assess whether the merger has any competitive impact that is of concern. In this way, the main task for competition economists is to carry out tests of causality (what impact will the merger have on market outcomes relative to counterfactual?) and quantification (by how much will prices rise, investment or innovation decrease, and efficiencies improve?).
This assessment is generally done by defining the markets in which the merging firms operate. This is often analysed by considering where consumers might switch to in the case of a 5–10% price rise and then including in the market all substitutes that may be found in this way (this is often referred to as the SSNIP[1] test). Following this, one may assess the market power of the merging firms in these markets, assessing whether it will be changed by the merger and if so, whether it is unconstrained by buyer, competitor, or consumer behaviour. Each of these latter points may be tested by carrying out analyses around issues such as the structure of the market, buyer power, price elasticities, and competitor behaviour.
UPP tests — more direct means to understand the internalisation effect
In markets for differentiated goods this exercise may be complex. Products or service may not be perfect substitutes and the market definition and market power exercises described may not always elicit all competitive constraints. In these instances, various other tests have often been used by the European and US competition authorities as initial screening mechanisms in assessing whether mergers will raise any potential competition concerns (these are also sometimes adapted and used further in the process as well). These tests often start with what are known as upward pricing pressure (UPP) tests. UPP tests are a more direct means to understand the internalisation effect described above. These tests have gained popularity over time as — assuming the availability of data to run them — they are relatively simple and can indicate whether a merger poses significant concerns or not. This is illustrated further below.
The test and data requirements
In essence an upward pricing pressure test, in its simplest form, looks as follows:
UPP1 = (P2 — MC2) D12
This means that the upward pricing pressure is determined by multiplying the margin of product 2 ((P2 — MC2) by firm 2, one of the merging parties) with the diversion ratio of customers from product 1 to 2 (i.e. switching between the merging parties).
In precedent cases the diversion ratio is generally found by using survey data and asking the hypothetical SNNIP question or alternatively asking who the consumer would switch to if one of the merging parties was no longer available. The diversion ratio may rely on actual switching data of consumers between the merging parties, if such data exist. Alongside this, the margins used here are generally referred to as contribution margins. Here the selling price less the marginal cost per unit is often relied on.
As a simple illustration, if we consider margins of 25% and a diversion ratio one third, then the UPP in this case may be around 8% in a 4 to 3 merger with unaffected costs. If efficiencies are found to be around 10%, then the merger may not, on balance, pose significant competition concerns according to this preliminary check. If, however there are assumed to be no efficiencies, then the 8% UPP would need to be more carefully scrutinised.
The UPP test may be calculated for each of the merging firms, given that diversion between them may not be symmetric and margins may also differ. The test may also be adapted for multiple equations in the case of multiple products. If such a detailed analysis is preferred, the assumption of switching of consumers not only between the merging parties but also between different types of products would need to be carefully understood. Alternatively, a higher-level approach may be tried initially, relying on indicative margins and overall diversion ratios.
Most importantly, one may note that if the diversion ration between the merging firms is low and/or the margins in question are low, then the merger should not raise significant concerns. In the best-case scenario, there would also be pro-competitive effects in the form of efficiencies.
Limitations of UPP type tests
Despite the benefits of these tests, there are drawbacks to be cautioned against. UPP tests provide an indication of the merger effects (whether there is or is not a potential competition concern), they cannot assess the level of any potential price rise that may stem from a merger (this is why they are generally relied on as screening tests by competition authorities). UPP tests are also static and cannot take into account feedback or other effects. These effects may include the feedback effect of a price increase by one merging party on the price of the other merging party, as well as the reaction of rivals to price changes by the merging parties. For these reasons, it would be advisable to use UPP tests as a preliminary check and/or in conjunction with other assessments of the effect of the merger (as mentioned above). Finally, these types of tests may only be as accurate and useful as the data on which they rely are. If one is dealing with sampled data, any sample bias would need to be considered. Similarly, if one is dealing with industries that have high or low variable costs in relation to fixed costs, this would also need to be taken into account in interpreting the margins used as inputs. Despite these limitations, UPP tests and the concept of starting with an assessment of actual movement of consumers (diversion ratios) between the two merging parties can provide strong initial and direct evidence of any competition concerns with transactions and frame the context for further analysis.
Extensions and alternative to the UPP tests
UPP tests, as mentioned above, may be extended to include a system of equations in the case of multiple products. They may also however be carried out in different forms. One may e.g. move toward gross upward pricing pressure tests (GUPPI) tests, which has the same basis, but is indexed to prices for different interpretation. GUPPI tests measure the value of diverted sales (the UPP as described above) as a share of the pre-merger price. One may also consider tests such as relating to illustrative price rise (IPR) tests. As with UPP and GUPPI tests these use the same inputs but now also require one to assume a demand function (e.g. linear or constant elasticity). IPR tests may also allow for some feedback effects between the merging parties (as described above, these are not accounted for in the general UPP test). Finally, one may also move toward full merger simulation models. Distinct to the above, these can take into account feedback effects from the rest of the market, i.e. incorporating data beyond that from the two merging parties, e.g. competitors’ data. These are more data and resource intensive. Here one would need to write an econometric model that characterises the market in question and in this way estimate post-merger prices.
Alongside the above, there are multiple other econometric tests that may be useful to carry out in assessing factors such as pricing in merger transactions. For example, if there are relevant pre-merger variations in market conditions (e.g. changes in prices with resultant movement of consumers between providers), these may be often be useful to better understand and estimate what the effect of the merger will be.
Looking forward
Looking forward, the usefulness of upward pressure pricing tests should be considered on a case by case basis. As evidenced by the literature, these tests are not intended to allow one to conclude on the competitive effects of a merger but are rather often useful tools to use alongside other competition assessment tools. At the very least the basic concept of diversion ratios may provide significant insights at the start of a merger assessment, eliciting direct evidence of whether or not consumers move between the merging parties and hence whether there is a competitive constraint that is in this way removed by the merger.
[1] Small but Significant Non-transitory Increase in Price
