Analysis of the third criterion

This note describes the advice from the CAA, respondents material and the analysis in relation to the third criterion. The decision document summarises the analysis and conclusions in this note.

Given the CAA’s expertise regarding airport regulation issues, and its role as the independent adviser on such issues to the DfT, the premise of our analysis has been to consider whether any comments from respondents to the consultation raise issues that call into doubt the CAA’s advice. Where a respondent has raised an issue or provided evidence that casts doubt on the CAA’s advice, this has been reviewed and the available evidence weighed up to reach a conclusion. If no respondent has raised an issue or provided evidence that casts doubt on the CAA’s advice, the CAA’s advice has only been subject to examination if there is some other reason to consider that it might not be appropriate. At all times, the greatest weight has been given to views supported by evidence.

CAA’s advice

The CAA considered that there appeared to be very limited incremental benefits over and above the benefits that might flow from competition and the application of general competition law. However, the costs and potential adverse effects of ex ante price control regulation appeared to be significant. Therefore, it followed that designation under Section 40 of the Airports Act 1986 would not deliver additional benefits, over and above competition law, which exceed the costs and potential adverse effects of such designation.

The CAA did not attempt a robust quantitative analysis of the potential costs and benefits of regulation to inform consideration of whether criterion three was met, nor did it consider that it could be undertaken.

Respondents’ views

The airlines broadly agreed with the CAA that it was very difficult to develop a robust quantitative assessment of the potential costs and benefits, and no airline provided such a quantitative analysis. The only quantitative analysis that was provided was a high level calculation by Ryanair of the costs if 1 million passengers were displaced from Stansted to East Midlands’ airport as a result of airline decisions to reduce flights at Stansted airport in response to future price increases by BAA.

The airlines agreed with the CAA that the current system of RAB based regulation was not working very effectively, particularly as it related to new investment. However, as the airlines opposed de-designation of Stansted airport, their position was broadly that they would prefer RAB based regulation to a situation of no regulation at all. While the airlines referred to a desire to develop alternative forms of regulation, they provided limited detail as to what such alternative forms of regulation might consist of. To the extent that a proposal for any alternative form of regulation was reasonably well developed, it was a proposal from Ryanair for a default price cap relating to additional assets and services, with a requirement for additional investment to be agreed between the airlines and the airport operator, with the CAA acting as the arbitrator.

While having concerns about the way RAB based regulation worked for new investment, the airlines, and Ryanair in particular, considered that the current price cap had been set too high by the CAA, and that outturn costs had shown that it had been set too high. In particular, actual capital expenditure had been much lower than the amount forecast when the price control was set. Ryanair also noted that the CAA had set a higher price cap in the last price review than was recommended by the Competition Commission.

The airlines considered that examples of poor quality of service at Stansted airport and the difficulties between the airlines and the airport in agreeing new investment plans were examples of why ex ante regulation was required and would provide benefits.

The airlines agreed with the CAA that the direct costs of regulation (CAA, Competition Commission, BAA and airline costs) were small compared to potential indirect costs or benefits from regulation.

Analysis

The assessment of the third criterion requires consideration of the incremental benefits of price cap regulation through designation above those provided by competition law, compared to the costs of regulation. Therefore, to assess whether this criterion is met it is necessary to consider the conclusions regarding the first and second criterion. If it is concluded that Stansted airport alone has substantial market power, then there may be anti-competitive behaviour that needs to be addressed through either competition law or a price cap. If it is concluded that competition law may not be effective in addressing this potential anti-competitive behaviour, then there may be benefits from setting a price cap to address some of all of the potential anti-competitive behaviour. Therefore, the assessment of whether this criterion is met is heavily influenced by the conclusion for the first criterion, that on balance it is likely that Stansted airport alone will acquire substantial market power, and the conclusion for the second criterion, that competition law may not be sufficient to address the risks of anti-competitive behaviour.

The CAA and respondents to the consultation agreed that it was very difficult to quantify the potential costs and benefits of regulation for assessing whether this criterion is met. To recognise this difficulty, the DfT has sought to consider both the qualitative and quantitative assessment of whether the third criterion is met. However, recognising that any quantification is difficult and there is a significant range of uncertainty for many of the numbers (so they are indicative), conclusions have only been reached when the qualitative arguments appear robust. Furthermore, the likelihood that particular outcomes will occur when assessing the relative costs and benefits has been considered.

Investment distortions are especially difficult to quantify because of the uncertainty that surrounds the planning processes for airport infrastructure as well as the long timescales involved. It is often very difficult to determine what causes particular investment decisions to be made in the first place and any subsequent changes to them. For scenarios, where investment outcomes are particularly important, we use indicative assessments of the impact of different probabilities to further our thinking, which are described qualitatively. The numbers that we derive for other impacts are subject to large amounts of uncertainty, and rely heavily on the assumptions made, so have to be treated with caution.

For the purposes of this analysis it is assumed that the CAA applies RAB based regulation because it represents the current approach. Although alternative forms of price cap regulation have been considered, none have been fully developed. However, it is notable that the types of alternatives that have been considered, including a default price cap or a market based price cap, are anticipated to create less distortion than a RAB based price cap. In other words, our approach means that we have cautiously over-estimated the costs of regulation.

Given the uncertainty surrounding likely future events, we have explored different outcomes under alternative scenarios below. The scenarios include what might happen if the airport operator at Stansted airport has no market power, as well as what might happen if it does.

Qualitative analysis

No substantial market power

If Stansted airport on its own does not have, and is not likely to acquire, substantial market power, then there are unlikely to be incremental benefits of regulation over and above those provided by competition law. Further, there are likely to be incremental costs. In addition to the direct costs of regulation, the simple act of regulation changes the ‘rules of the game’ and creates distortions that harm investment and efficiency outcomes. In particular, there may be negative indirect impacts on the cost of capital, price flexibility and speed of responses to events in the market. These type of costs are difficult to quantify in any meaningful manner even though they are likely to be important. The conclusion that it is better to have no regulation where there is no substantial market power is generally accepted for most markets within the economy.

In relation to BAA’s ownership of other airports in addition to Stansted airport, the market power at Heathrow and Gatwick airports is separately addressed through the CAA setting price caps for these two airports. Therefore, if the CAA sets a price control for Stansted airport when it does not have substantial market power on an individual basis, then the costs of regulation outweigh any benefits or, at least, the risk that costs are caused by regulation outweigh any potential benefits.

Further, it seems that different costs will occur depending on what level the CAA sets the price control. If Stansted airport on its own has no substantial market power and the CAA sets the price cap at the ‘correct’ level or too high, no benefits are realised that would not be achieved in a competitive market and the direct costs of regulation are incurred in addition to the indirect costs mentioned above. This conclusion favours de-designation. However, the net cost benefit analysis (“CBA“) will favour de-designation even more if the CAA sets a price control at a level that is too low because it risks distorting outcomes in addition to the other costs of regulation. In particular, it creates the possibility of under investment at Stansted airport or other rival airports, which is to the detriment of customers.

Substantial market power

The scenario where Stansted airport has, or is likely to acquire, substantial market power on its own, needs a separate assessment of the incremental benefits and costs. Given the nature of the conclusions about the protection afforded by competition law under the second criterion, this assessment needs to proceed from the conclusion under the second criterion that there is a risk that the protection afforded by competition law may not be sufficient to avoid detrimental effects where an entity with substantial market power chooses to take advantage of this market power. If substantial market power is held, there is a material risk that there is the opportunity for abusive behaviour occurring that creates detrimental effects that are harmful to consumers, and which are not adequately addressed by competition law. Whether abusive behaviour leading to detrimental effects occurs will depend on the decisions of the airport owner.

Where the CAA sets a price control because Stansted airport has, or is likely to acquire, substantial market power on its own, we need to make assumptions about the appropriateness of the price cap set by CAA. It could be assumed that CAA sets a “perfect” price control, which prevents all potential abuse of substantial market power and does not create any distortions. Under this assumption, and given the conclusion that under the second criterion that competition law may not provide sufficient protection against abuse, it would be appropriate to conclude that the criterion is met, and it is appropriate to designate the airport.

In practice, it is probably not realistic to assume that the CAA sets a “perfect” price control. Other possibilities are that it sets a price control that is either “too high” or “too low”, which were also discussed above. On balance, we would argue that the CAA is more likely to set a price control too high than too low. This principally reflects their duty to promote investment, which suggests that the CAA would err on the side of setting a price control that is higher rather than lower. The requirement for a Competition Commission reference also creates a check and balance in the process for setting the price control that reduces the likelihood that the CAA would set a price control that is too low (or too high). However, it is important to note that a price control can appear adequate when it is set, but turn out to be too low if market conditions change during the price control period, for example, NATS following the downturn in air travel after 9/11.

Table 1 below considers the conclusions that can be drawn under the various scenarios discussed above. The counter factual is determined by reference to the degree of market power that has been assumed. In deciding values to include in each cell, and the sign to attach them, it is important to evaluate whether 1) any of the effects described above are relevant in that scenario and 2) whether its effect will be positive or negative compared to the alternatives.

Table 1: Matrix of outcomes for various market power and price control scenarios

		Market power
		Yes	No
Price control	Yes	1) CAA sets the price control at the correct level. Benefits of protection against potential abuse through ‘excessive’ prices, which may manifest itself as higher prices, reduced investment and / or lower quality service compared to outcomes expected in a competitive market. There are no costs or distortions from regulation, except direct costs. Benefits exceed low direct costs of regulation. C3 is met. 2) CAA sets a price control that is too high. Allows some potential abuse that would not be addressed through competition law, including through prices being too high, quality of service being too low or inappropriate investment decisions. However, it protects against further abuse that competition law would not correct. Nevertheless, there is also some potential distortion of investment decisions through the RAB approach, although arguably these are not likely to be significant and might be already partly accounted for by the potential abuse that is allowed. Direct and indirect costs of regulation are incurred. Further analysis needed to reach a conclusion. One approach might be empirical evaluation. 3) CAA sets a price control that is too low. Inappropriate investment decisions (including at rival airports) and the direct plus indirect costs of regulation are incurred. The benefit is that airlines pay lower airport charges. Further analysis needed to reach a conclusion. One approach might be empirical evaluation.	4) CAA sets the price control at the correct level. Regulation has no benefits but direct costs are incurred. C3 is not met. 5)(Same as 2) CAA sets a price control that is too high. There is no benefit from regulation since competitive forces ensure that prices are lower than the cap. However, direct and indirect costs are still incurred. C3 is not met. 6) (Same as 3) CAA sets a price control that is too low. Airlines pay lower airport charges. However,these lower payments do not represent avoiding potential market power abuse. Hence, there are inappropriate investment decisions (including at rival airports) while the direct and indirect costs of regulation are incurred. C3 is not met.
	No	7) There is the opportunity for abuse through excessive prices. The direct costs of regulation are avoided and the costs of setting a price control discussed above are also avoided but, given the conclusion about competition law, customers are exposed to a significant risk of potential abuse. C3 is met.	8) No abuse. C3 is not met.

The matrix above has identified some scenarios under which the preferred outcome is clear. For other scenarios, further analysis is required to clarify whether the criterion is met or not. The table begins to indicate that if you consider that Stansted airport does not have, and is not likely to acquire, substantial market power on its own, the appropriate strategy is likely to be to de-designate, whereas it is more balanced, but leaning towards designation if you consider that Stansted airport has or is likely to acquire substantial market power. The conclusion that Stansted is likely to acquire substantial market power in the future was reached, on balance, in assessing whether the first criterion is met.

Furthermore, building on the conclusion reached for the second criterion, it appears reasonable to assume that there might be significant incremental benefits from price cap regulation that would not be addressed through competition law if Stansted airport has substantial market power on its own. This could cover excessive pricing, reduced quality and/ or reduced output. The precise magnitude of these impacts, the extent of passthrough of costs associated with this behaviour to passengers and the period over which it would be sustained is again open to question, as higher prices would provide further incentives for new investment, including at alternative airports. Also, not all price increases as well as reductions in quality or output by Stansted airport in the future would necessarily be abusive behaviour causing detrimental effects, and indeed, as discussed when analysing the second criterion, the likely appropriate benchmarks are evolving with Court decisions. However, sustained and material increases above reasonable benchmarks for excessive prices could be regarded as excessive prices.

Quantitative analysis

To carry out an indicative quantification, we need to estimate the impacts described below. In relation to market power at Stansted, we assume that it increases over time, where it is possible to translate this assumption into meaningful estimates.

Indicative values for the potential benefits of price cap regulation through avoiding potential detrimental effects in the form of excessive pricing have been considered, which reflect our conclusion under the first criterion that Stansted airport is, on balance, likely to acquire substantial market power and that competition law may not be sufficient to protect against detrimental effects from abusive behaviour. The values range from relatively small values per passenger (up to £0.5) to larger values (£3-4) per passenger, which generates an indicative range of £55m to £350m. The way the values are derived is explained below, but the values also seem a reasonable range of indicative values given the conclusions for the first and second criterion, and recognising the difficulty of establishing a precise benchmark, estimating passthrough of costs to customers and determining the period over which impacts will be sustained. The way the values have been derived is explained below and these approaches should be considered as means to establish a reasonable range that takes account of the uncertainties about the benchmark, passthrough, period of time, etc.

• Excessive pricing or avoidance of excessive pricing – It is probably the hardest number to calculate as it potentially requires a baseline and some view about when prices might be too high. It may for example be argued that prices up to incremental investment cost are potentially not abusive. Options for establishing a range include the difference between the price cap and Ryanair’s re-calculation of the price cap, although this is very rough and ready. The conclusion of this calculation is determined by the assumptions made. Ryanair calculates that the required revenue yield is £3.82 per passenger in 2003/04 prices while the CAA calculates that it ought to be £4.65 per passenger in 2000/1 prices, which is £4.82 in 2003/04 prices or a £1 difference per passenger. If we assume that the capacity at Stansted is constrained to 25 million passengers for the next five years, this translates to a NPV of £110 million. However, although this number is based on previous market power, which seems to be less than current or future market power at Stansted airport, it is possible that Ryanair have emphasised data that favour its own interests. Alternatively, it does seem reasonable, given the argument above, that the CAA might be expected to be more likely to set a high rather than low price cap. Therefore, a range for the likely impact that allows a 50 per cent difference either way from the figure suggested by Ryanair’s data is £55 million to £165 million. An alternative amount is based on the shadow price data generated by NAPALM, where the shadow prices represent the premium that passengers place on using Stansted as their preferred airport. These data assume that no investment occurs at Stansted airport over the next five year period and might suggest that its market power will increase over time as a consequence. The shadow prices increase non-linearly from £0 in 2008 to £6 in 2012 (but average £3 over the period). Further, the modelling suggests that passengers travelling from Stansted airport each year will increase from 24 million in 2008 to 28 million in 2012. These data suggest a NPV for market power abuse of £350 million. The range does not explicitly consider behaviour beyond this time period, but it might be relevant to the assessment, and this contributes to a view that the range is reasonable.
• The value of quality of service – One option is to use the valuation for quality in the CAA’s proposals for price caps for Heathrow and Gatwick airports, which suggests that up to 10% of revenue can be exposed to quality of service performance (+3% to -7%). Hence quality of service could have a NPV of £100 million over the next 5 years. However, a possible range, drawing on an alternatives value for revenue exposure of 5%, is between£50 million and £100 million. This is a relatively high level range, reflecting a broad brush approach.
• Direct costs – Likely to be a small number, best calculated from considering the range of evidence presented. It was £3.2 million for BAA airports in last price review. A reasonable figure seems to be £1 million for the next 5 years for Stansted airport alone.
• Insufficient or delayed investment – Unfortunately, it is not possible to obtain data for this impact even though we recognise it can be potentially be important. It was noted above the difficulties of assessing the distortion created by regulation on investment decisions. Similar issues arise when considering the potential to calculate a value for the indicative range of potentially abusive behaviour through investment decisions. Some options were considered, including the results from the NAPALM model that was generated from moving investment forward at Stansted relating to maximum use (G1) and a second runway (G2). However, the numbers could not be interpreted in a meaningful fashion.

We have sought to account for and avoid any “double counting” by assuming that all abuse is captured in a price effect and there is no need for separate consideration of abuse due to quality of service or investment decisions. This again suggests that the range should be reasonable because it is not cumulative. There has also been an attempt to calculate one figure to encapsulate the effects of regulatory or distortions.

Table 2: Matrix of costs and benefits for various market power and price control scenarios

		Market power
		Yes	No
Price control	Yes	1) CAA sets a correct price control. Costs = £1 million. Benefits = £55-350 million. Net CBA ˜ +£55-350 million 2) CAA sets a price control that is too high. Benefits: avoid 50 per cent of excessive pricing £28-170 million. Costs: £1 million. Net CBA: inconclusive (because relies upon quantifying distortions to investment) 3) CAA sets a price control that is too low. Benefits: £55-350 million. Costs: at least 1 million. Net CBA: inconclusive (because relies upon quantifying distortions to investment)	4) CAA sets the price control at the correct level. Costs: £1 million. Net CBA = -£1 million. 5)(Same as 2) CAA sets a price control that is too high. Costs: £1 million. Net CBA = -£1 million. 6) (Same as 3) CAA sets a price control that is too low. Costs: at least £1 million. Net CBA: inconclusive (because relies upon quantifying distortions to investment)
	No	7) There is the opportunity for abuse through ‘excessive prices’. poor quality or insufficient investment. Costs = £55-350 million. Net CBA: -£55-350 million	8) No abuse and no regulation. Benefits: £1 million. Net CBA: £ 1 million

The table above indicates that if you consider that Stansted airport has, or is likely to acquire, substantial market power on its own, the dominant view is to continue to designate the airport. If you consider that Stansted airport has not, and is not likely to acquire, substantial market power, then this points to a conclusion that the airport should be de-designated.

To confirm whether the appropriate approach is to de-designate given the conclusion that Stansted airport on its own is, on balance, likely to acquire substantial market power, it is necessary to consider the potential scale of any distortion arising from regulation, and a view taken about the relatively likelihood of price controls being set too high or too low. The CAA and the airlines agreed that RAB based regulation created distortions particularly for new investments. The CAA argued that this could be particularly material given that in the foreseeable future there is expected to be significant new investments at Stansted airport, including a new runway. However, it is very difficult to quantify the potential effect of any distortion and to identify which parts of any distortions arise from regulation rather than other factors that will affect investment decisions, such as the planning process. The DfT considered whether any of its airport models could provide helpful results to inform this type of analysis, but ultimately these models did not allow us to obtain useful results.

Given the difficulties of estimating indicative values for potential investment distortions, the size of the distortions and the risk of them occurring, we have tried to evaluate the scale and likelihood necessary to outweigh the range of potential benefits from regulation that we have identified. The shortfall in benefits associated with regulation compared to its costs is likely to be greatest when the airport operator does not possess market power, and so there are few benefits from regulation to counter-balance the potential distortions, with other scenarios potentially generating greater distortions. Relative to the ranges for potential abusive behaviour, the degree of regulatory distortion that would have to arise, even allowing for the size of potential investments in the future, would be very large to alter the net CBA calculations in the tables above given the conclusion for the first criterion. While there might be some overlap in the net CBA for some scenarios, we consider that it is unlikely that the degree of regulatory distortion would be such to overturn the general conclusion that if it is considered that Stansted airport alone is likely to acquire substantial market power, the appropriate approach is to conclude that the criterion is met. Our assumption that it is more likely that the CAA would set a price cap that was too high rather than too low, due to the CAA’s duty to promote investment also affects the conclusions about which outcome is the most likely.

The analysis suggests that while it cannot be ruled out that the costs of distortions could outweigh the benefits of regulation, the balance of risks suggests that the distortions attributable solely to regulation would have to be very large, even given the scale of investments being considered, to outweigh the range of indicative benefits for regulation. The finding is particularly supported by our conclusions for the first and second criteria. Furthermore, while there might be material distortions, there may be options for the CAA to address some of the risks of distortions through different approaches to price regulation than RAB based.

Conclusion

The third criterion is met because on balance while there is likely to be a material distortion, there may be options for the CAA to address some of the risks of distortions through different approaches to price regulation. Furthermore, given the conclusions regarding the first and second criteria, the potential benefits of regulation given the risks and potential detrimental effects would appear to outweigh any costs associated with the distortions.