Stansted – Market Power Assessment: Non-confidential Version

Civil Aviation Authority

Stansted – Market Power

Assessment:

Non-confidential Version

The CAA’s Initial Views – February 2012

Civil Aviation Authority 2

Civil Aviation Authority 3

Table of contents

Executive Summary ................................................................................................. 5

Rationale and context for the CAA‟s airport market power assessments ...... 5

The extent and nature of Stansted‟s market power ....................................... 5

Way Forward .............................................................................................. 14

1. Introduction .............................................................................................. 15

Rationale and context for the CAA‟s airport market power assessments .... 15

The CAA‟s approach to the assessing airport market power ....................... 16

Way Forward .............................................................................................. 16

2. The markets in which Stansted competes .............................................. 19

Stansted‟s business .................................................................................... 19

The use of market definition ........................................................................ 20

Product market definition ............................................................................ 21

Geographic market definition ...................................................................... 34

The interdependence of demand from different user groups ....................... 46

3. Factors contributing to Stansted’s market position .............................. 49

Introduction ................................................................................................. 49

Market shares ............................................................................................. 49

Airline switching .......................................................................................... 52

Passenger switching ................................................................................... 56

Implications of passenger and airline switching and the multisided

nature of the market.................................................................................... 63

Airport sensitivity to passenger and airline switching .................................. 65

Buyer power ............................................................................................... 69

Entry and expansion ................................................................................... 80

Capacity constraints ................................................................................... 81

Pricing and behaviour ................................................................................. 84

Cargo ......................................................................................................... 92

Long term view ........................................................................................... 92

Overall assessment of market power .......................................................... 94

Civil Aviation Authority 4

Civil Aviation Authority 5

Executive Summary

Rationale and context for the CAA’s airport market power assessments

1. Where airports enjoy high levels of market power, it can be appropriate for

the CAA to apply economic regulation, so as to improve outcomes for

passengers, cargo shippers and airlines.

2. The CAA is committed to ensuring that its regulation of airports promotes

choice and value for consumers, whilst also meeting the better regulation

principles. In order to deliver on this commitment, during 2011 the CAA

embarked upon a project to understand the extent and nature of market

power held by the airports that are currently „designated‟ for price control

regulation: Heathrow, Gatwick and Stansted.

3. This document sets out our work to date to understand the market power

held by Stansted. These views are based on the evidence currently available

to the CAA. As noted above, we expect that additional evidence will become

available before the CAA takes a firm view on the position of each airport.

Reflecting this, there are a number of issues that have been left open and,

whilst the CAA has sought to highlight where the airport might sit on the

market power spectrum, it has not reached a definitive view at this stage.

4. We would like to thank those stakeholders who have engaged with the CAA

during 2011, and invested time and money in providing evidence and

analysis.

The extent and nature of Stansted’s market power

5. The first step in assessing an airport‟s market power is to consider the

markets in which it operates with regard to its various user groups to

establish a frame of reference within which to conduct the analysis.

Market definition

6. Stansted provides infrastructure and infrastructure services to its various

groups of users, including passenger airlines, passengers, and cargo

carriers, and also retailers and car park operators. Each of these user groups

could be considered as different „sides‟ of the airport market; each with its

own characteristics and ability to respond to changes in the price and service

offered by Stansted.

7. In order to understand the market position of Stansted, it is important first to

consider the characteristics of the airport‟s users including the choices

available to them to reduce their use of the airport. As a next step, the

interactions and interdependences between the choices made by these

different airport user groups need to be analysed to reach a view on their

combined impact on the airport.

This treatment of airports as operating in a multi-sided market is described in more detail in the CAA‟s

Guidance on assessing airport market power

Civil Aviation Authority 6

Provision of services to passengers

8. The focus on Stansted on low-cost carriers has implications when defining

the market within which the airport competes for passengers. In particular,

there is a degree of uniformity in the requirements of the various passenger

types at Stansted, and limits on the extent to which the airport can segregate

its offer to passengers or price discriminate between them.

9. We have, therefore, adopted a product market for passengers at Stansted

that covers the provision of services for the reception, processing and

boarding of passengers, without differentiating between passengers based

on their individual characteristics.

10. In order to establish a geographic market for the supply of services to

passengers, we consider a range of evidence, including a number of different

ways to define an airport catchment. This showed that there were significant

overlaps between the catchment areas of Stansted and all of the major

London airports, and that they all draw a significant proportion of their

passengers from the Greater London area. These overlaps, when combined

with the location of the airport and evidence on passenger responsiveness,

supports the view that the relevant geographic market covers the South East

of England, Greater London and includes a number of districts falling in the

East Anglia planning region.

Provision of services to passenger airlines

11. Stansted predominantly serves low cost, short haul, point to point airlines.

The airport provides these airlines with a mix of services, including facilities

and services for the handling of passengers and those for the transportation

of bellyhold cargo, although this is currently not a major requirement of

airlines at Stansted.

12. These services are likely to be available at most mid-sized airports, which

implies that there are a relatively large number of airports within the same

(airline-facing) product market. However, due to the complexity associated

with operating from larger airports, and the difficulty in achieving quick aircraft

turnarounds, we do not consider Heathrow to be a close substitute to airlines

at Stansted, and that it falls outside of the product market.

13. The demand for capacity at Stansted is significantly higher in the early

morning period than throughout the majority of the rest of the day. This

reflects the particular requirement of based aircraft for access to an early

morning slot – in order to deliver a sufficient number of rotations to deliver

high aircraft utilisation – whereas the demand from aircraft operating their

services into Stansted (inbound carriers) is more spread out. Combined with

this, the airport can vary its charging structures in ways that particularly affect

based or inbound carriers, including by varying charges by the time of day, or

through varying overnight parking charges. Although it currently does not do

so.

Most passenger flights at Stansted do not carry cargo. Less than 1% of Stansted‟s cargo is carried

bellyhold (CAA airport statistics).

Civil Aviation Authority 7

14. If airlines are unable to vary their use of the airport at peak, this might

support defining a separate market for the supply of peak capacity. At this

stage, the CAA has not been able to reach a definitive view on this issue –

which is likely to be important to the overall market power assessment – and

has left this issue open at this stage.

15. The geographic market within which Stansted competes for airlines is

determined, to a considerable extent, by the fact that the airport

predominantly serves short-haul, low cost point-to-point carriers. The

business model of these airlines tends to involve basing aircraft at airports of

different sizes across Europe from which they offer point-to-point services.

16. This has a number of implications. First, airport charges are also typically a

higher proportion of their overall cost base than other carriers, suggesting a

higher degree of responsiveness to a given price change than for airlines

with other business models. Second, the largest airlines at Stansted have a

relatively large number of existing bases across Europe, and a large number

of airports to which they already operate, which will tend to reduce the costs

associated with moving capacity away from Stansted. However, the airlines

at Stansted emphasise the importance of Stansted (and of serving London)

to their businesses, and the impact that switching away from Stansted has on

their yields. Airlines have also highlighted the lack of responsiveness to

recent increases in the level of charges at the airport, which is discussed

below.

17. Overall, given the business model of airlines at Stansted and the likely

magnitude of switching costs, airports in the Southeast of England are likely

to be very close substitutes to Stansted (although Heathrow is likely to fall

outside of this market, due to its particular operational characteristics).

Furthermore, it appears that the switching of marginal airline services is likely

to take place over a broader area, such that the geographic market relating to

the supply of services to airlines is likely to be European wide.

Provision of services to cargo carriers

18. Stansted provides infrastructure and infrastructure services to cargo-only

carriers for the air transportation of cargo. This includes the provision of

runway and apron space, as well as providing access to cargo-handlers,

access to warehousing facilities, and infrastructure to allow the efficient

onward transfer of cargo. In general, we consider that the airport operates in

a very broad market, and competes with a number of airports across the UK

and Europe for much of this cargo traffic.

The markets in which Stansted operates

19. We have described the markets in which Stansted operates, looking at the

airport‟s passenger-facing and airline-facing activities. However, in practice,

these two aspects of the airport‟s operations are closely linked and there are

important interactions between passengers and airlines that are likely to

affect the overall assessment of airport market power.

This is discussed in the context of the product market.

Civil Aviation Authority 8

20. In particular, whilst there may be relatively low direct costs associated with an

airline relocating to another airport, the willingness to do so will be affected

by whether there are sufficient passengers at alternative airports and whether

airline switching away from Stansted typically involves airlines accepting a

lower yield. To the extent that airlines are able to switch to nearby airports

and attract many of the same passengers, this may reduce the adverse

impact on yields. Similarly, for passengers to view an airport as a reasonable

substitute they will need to be able to find a suitable alternative flight, which

will often need to be to the same destination. The airlines‟ ability to switch

will, therefore, depend upon passenger decisions, whilst passengers‟ choices

are likely to be affected by those of airlines.

21. There are two characteristics of an airport that may strengthen the interaction

between passengers and airlines: a high concentration of network carriers or

a small airport that is trying to become established. As Stansted does not

currently serve network airlines and is a relatively large and well-established

airport – indeed, survey evidence confirms a high awareness of the airport

amongst passengers – neither of these conditions appear to apply, limiting

any affects on the market power enjoyed by the airport. However, the retail

revenues generated from passengers are likely to have a significant impact

on the airport‟s incentives to raise prices to airlines, due to the adverse

impact that lower passenger numbers have on the profitability of the airport.

Market shares

22. Market shares can provide an indicator of an airport‟s market position. Even

under the narrowest definition, when we limit the market to be short haul

flights from the London area, Stansted does not have a high market share,

when viewed as a stand-alone airport, and certainly below the level at which

there would be a rebuttable presumption of dominance. However, Stansted

and Heathrow are both currently owned by BAA, and combining the market

shares of the two airports gives very high market shares and which on a UK-

wide basis are still as high as 58 per cent.

23. An important aspect of understanding market power at Stansted is to

consider the position of the airport at the early morning peak. Looking only at

peak periods increases Stansted‟s market share to 26 per cent (behind

Heathrow and Gatwick‟s 30 per cent), and a combined share with Heathrow

of 56 per cent. On some measures, the combined share of BAA-owned

airports would support a rebuttable presumption of dominance.

Airline switching

24. The predominant airline business model at Stansted is low-cost, short-haul,

point-to-point. In general, these airlines will have invested less at the airport

than other airline business models, have multiple bases across the UK and

Europe and, due to their more streamlined cost structure, face airport

charges that generally account for a bigger proportion of the total costs than

they do for full-service network airlines. This implies that there will be a

greater incentive, and more ability, for low cost point-to-point airlines to

switch in response to a given increase in airport charges, which is consistent

Civil Aviation Authority 9

with evidence that highlights that the major carriers at Stansted operate

particularly dynamic networks, with the routes flown varying to a significant

degree over time.

25. Even if switching costs are low enough to allow airlines to switch, they must

have appropriate alternative airports to switch to. A based carrier at

Stansted‟s ability to switch to neighbouring airports may be limited by

constrained capacity at other London airports, but those airlines with a

network of bases across Europe also have the option of switching to other,

non-London airports.

26. The likely magnitude of switching costs, and the sensitivity to airport charges,

is mirrored by the evidence of airline switching. In recent months, Stansted

has lost several major airlines to Gatwick (Norwegian, Air Berlin and Air Asia

X), despite their being some capacity constraints at Gatwick. In addition,

easyJet has recently announced its intention to move some of its based

aircraft from Stansted to Southend.

Passenger switching

27. We have considered a range of evidence to understand the extent to which

passengers might switch between airports. Analysis based on airport

catchment areas highlights the significance of catchment area overlaps,

particularly over the most densely populated areas (notably Greater London).

This suggests that a significant proportion of passengers are likely to be

marginal, and able to switch away from Stansted, if a suitable service is

available at an alternative airport.

28. Survey evidence is consistent with this general finding, showing that there

are significant numbers of Stansted passengers who have previously used

another London airport (indicating a degree of willingness to travel to use

these airports) and that stated that they had considered alternative airports to

Stansted. This is consistent with the survey results that indicated a higher

degree of price sensitivity at Stansted and Gatwick, when compared to

Heathrow.

29. However, this apparent willingness to switch between airports relies upon

passengers being able to find a suitable service at an alternative airport. In

this respect, whilst Stansted has fewer route overlaps than other London

airports, 74 per cent of passengers at the airport could fly to their chosen

destination from another London airport. This suggests that a large

proportion of passengers are likely to have the willingness and ability to

switch away from Stansted.

30. Further, as only around 10 per cent of Stansted‟s passengers were on

regularly-served routes served by more than one airline at Stansted, it is

likely that should an airline remove a service from Stansted, the airport would

be more likely to lose the passengers from that route to another airport.

Indeed, the interplay between passenger and airline decisions means that

the loss of business to an airport resulting from an increase in price may be

greater than indicated by the impact on „marginal‟ passengers, as services

Civil Aviation Authority 10

become unprofitable and prompt those who would have used these services

to consider using another airport.

Airport sensitivity to airline and passenger switching

31. In order to understand the impact that passenger and airline switching is

likely to have on the airport, it is useful to consider the volume reduction (or

„critical loss‟) that would render an increase in prices unprofitable. CAA

calculations suggest that to render a 5 to 10 per cent price rise unprofitable

the airport would need to experience a fall in passenger numbers of between

3 and 11 per cent, depending on the assumptions used. For example, if the

price increase applies to the airports overall revenue, a 5 per cent increase

would be rendered unprofitable if the airport lost approximately 900,000

passengers, or approximately 1,000,000 passengers if operating costs are

saved as passenger numbers fall.

32. Easyjet commissioned Frontier Economics to analyse the potential for airline

switching to constrain prices at Stansted, undertaking similar calculations to

those set out above. Frontier‟s paper shows that airlines switching and

passenger switching would be insufficient to render a 10 per cent increase in

charges unprofitable. However, whilst this analysis makes a useful

contribution to the available evidence, the approach taken restricts

passenger and airline switching in a number of ways that are likely to under-

state the actual level of switching. In addition, the data made available to

Frontier introduces a further potential source of bias, in that it is based on

average easyJet route profitability, rather than the profitability of the most

marginal services.

Buyer power

33. Ryanair and easyJet account for a very large proportion of passengers at

Stansted, with 68 and 21 per cent shares respectively. Further, whilst

Stansted is no doubt an important base for both airlines, the airport accounts

for a much lower proportion (17 and 8 per cent) of the airlines‟ passengers.

This implies that Stansted is significantly more reliant on Ryanair and easyJet

than these airlines are reliant on the airport, and that there could be a degree

of buyer power at the airport.

34. Analysis undertaken for Ryanair by RBB disputes the view that there is a

degree of buyer power at the airport. The RBB paper cites the evidence of

actual switching that took place following the increase in airport charges

faced by Stansted when its agreements with the airport expired, emphasises

the importance to buyer power of having an ability to switch, and also argues

that there are significant costs to Ryanair relocating aircraft across its

European network. Whilst we agree that buyer power requires the

combination of scale with the existence of an ability to switch away from the

airport, we consider it likely that both Ryanair and easyJet have the ability to

switch volumes away from the airport. Indeed, the significant number of

bases operated by the two airlines, and recent examples of the airlines

opening new bases – including the example of easyJet‟s switch to Southend

Civil Aviation Authority 11

– highlights that there is a significant degree of potential switching from these

two airlines.

Competitive price level

35. The interpretation of the evidence on airlines‟ historical responses to price

increases is affected by the relationship between historical prices and the

competitive price level. This is particularly important in light of the very

significant increase in prices paid at the airport during 2006 and the relatively

moderate reductions in volumes that followed.

36. In order to understand whether these price rises were increases towards, or

increases above, the competitive price level we have considered a range of

sources of evidence.

37. Ryanair provided evidence that showed that its charges at Stansted are

considerably higher than at many of the other airports it operates from.

However, when Stansted‟s average charges are compared to those at

comparator airports, it reveals that the level of historical charges was

particularly low and that the recent price levels are broadly in line with a

number of potential comparators. On balance, it appears that the increase in

charges around 2006 was an increase towards, rather than above, the

competitive price level – in which case, we would not expect there to have

been a particularly significant response from airlines switching away from the

airport.

Entry and expansion

38. Expansion and/or entry by existing aerodromes may represent a source of

competitive constraint, albeit one that is limited by the cost and timescales

involved. However, there is relatively limited evidence of significant

expansion and/or entry in Stansted‟s market, with the notable exception of

the planned easyJet expansion at Southend Airport, which involves a

relocation of capacity from Stansted to Southend.

39. At the time of writing, plans are also being discussed for the expansion of

Luton Airport.

Capacity constraints

40. It is clear that there is significant capacity off-peak at Stansted, albeit that this

might reduce over time, as demand grows. However, at peak times, there is

very little spare capacity. The period over which these capacity constraint

holds is currently somewhat limited, relating to 2 or 3 hours, but as demand

grows these constraints would be likely to spread to other hours.

41. Capacity constraints at early morning times appear to be a feature at Luton,

Gatwick and a number of other European airports. This lack of early morning

capacity might reduce the ability of airlines to relocate operations from

Stansted, although Gatwick – which is relatively full by European standards –

has been able to accommodate a number of new airline services, including a

number of services switching from Stansted.

Civil Aviation Authority 12

Pricing and behaviour

42. As discussed above, evidence from Ryanair and from other sources paints a

somewhat different picture about the appropriateness of Stansted‟s current

prices. However, in general terms, the available evidence tends to support

the view that the current price level is reasonable. It is also notable that the

airport has flexibility to raise its prices at peak times but has chosen not to do

so.

43. The airport has, however, seen a considerable fall in passenger volumes,

falling below 18 million, which is back down to 2003 levels. This significant

reduction in airline use of the airport follows the increase in charges at the

airport, but also coincides with a recession in the UK, and low levels of GDP

growth across Europe. Consequently, the reduction in traffic at Stansted

could either reflect economic trends or price responsiveness. In this respect,

Ryanair says that the fall in traffic is as a result of Stansted failing to offer

sufficient discounts, which arguably implies that airlines do have choices of

where to expand and at least some ability to switch away when Stansted‟s

relative prices go up.

44. We have also considered whether the airport has responded to the reduction

in traffic, and the availability of substantial capacity off-peak, by offering

aggressive discounts for growth. The evidence provided by Stansted

appears to support the view that it is offering „competitive‟ discount offers to

support growth in off-peak periods.

45. Overall, the evidence currently available supports the view that there is no

SMP relating to the off-peak periods.

Evidence of market power at peak

46. Capacity in early morning periods appears to be particularly important for the

low-cost carriers at Stansted. There are capacity constraints at these peak

periods at Stansted and also at Luton and Gatwick. This lack of capacity at

this crucial time of day may make switching based aircraft away from

Stansted more difficult.

47. Nonetheless, switching may take place, but it is likely to be more costly to

obtain slots at this time. []. Airlines are generally able to obtain the slots

they want through slot trading, albeit that the slots at the more popular times

are likely to be more expensive.

48. Overall, there remains significant uncertainty about the balance of negotiating

power at peak periods, with the incumbent airlines at Stansted placing

particular weight on capacity at these times, whilst there are also constraints

at a number of alternative airports. Set against this is the general flexibility of

the business models of the airlines that use Stansted at peak periods, and

the fact that the airport is not currently exercising its ability to increase prices

in these periods.

49. On the evidence currently available, Stansted‟s market power at peak

periods could amount to a position of SMP. In order to investigate further this

aspect of Stansted‟s market position, we expect to seek additional

Civil Aviation Authority 13

information on airline switching costs, including on the impact that route

profitability might have on these switching costs.

Impact of common ownership with Heathrow

50. As noted above, Stansted and Heathrow are still currently under joint

ownership. It could therefore be argued that Stansted‟s competition for

certain airlines is muted by concerns about the adverse impact on Heathrow.

Stansted disputes this view and argues that BAA is neutral to Stansted‟s

competition with Heathrow.

51. However, we consider that there is a marked contrast between the

aggressive marketing of Heathrow and Gatwick to passengers and the

relatively low-key approach adopted by Stansted, which could be explained

by the continued ownership by BAA.

52. Of course, even if joint ownership were reducing the ability of Stansted to

compete, this is not sufficient to conclude that divestment by BAA would be

sufficient to remove any SMP held by the airport. Rather, that divestment

would lead to a more competitive outcome than continued joint-ownership.

Initial view on the degree of Stansted’s market power

53. At this stage, the supply of capacity at peak times is the most likely source of

any position of substantial market power at the airport, albeit that we remain

concerned that BAA‟s ownership of Heathrow and Stansted could be

reducing the ability of the airport to raise its profile and adopt aggressive

strategies to attract passengers and airlines to the airport.

54. There is, however, a lack of clarity about the ability of airlines to reduce their

use at peak times, and the evidence on pricing at these times. First, peak

capacity appears important to the airlines at Stansted, as it supports the

efficient use of their aircraft, making them reluctant to reduce their use of

peak capacity. This, however, needs to be set against the apparent flexibility

of their business models, which might enable switching to alternative airports.

55. Second, we have seen evidence that supports the view that Stansted‟s

prices, whilst higher than a number of airports across Europe, appear

comparable to those serving major conurbations. Furthermore, Stansted

does not currently price discriminate between different times of day, despite

having the ability to do so.

56. Looking forward, the CAA‟s view on the market power at Stansted will likely

depend upon the following:

 The evidence available on the barriers to airlines reducing their use of

Stansted at peak times, including the impact on airline yields; and

 Whether, and to whom, the airport is sold by BAA.

57. Whilst the second of these factors is outside of the control of the CAA, we

would hope to be able to work with airlines and the airport to obtain better

information to reach a firm view on the former.

Civil Aviation Authority 14

58. Overall, therefore, we see Stansted as enjoying the least market power of the

three airports being assessed and, whilst the evidence is currently not

sufficiently clear to reach a definitive view, it appears that any position of

substantial market power arises from the relative bargaining positions of the

airport and airlines during a relatively narrow peak period. The relative

strength of these positions might change over time and be affected by the

potential ownership changes at the airport, as well as the balance between

demand and available capacity.

Way Forward

59. We welcome stakeholders‟ views on the information presented in this paper.

The CAA is consulting during 2012 on its views, and will continue to work

with stakeholders to develop its analysis and to resolve the issues that have

been left open at this time. Information about the consultation process,

contact information and the dates for a seminar are set out in chapter 1.

Civil Aviation Authority 15

1. Introduction

1.1 This is the non-confidential version of the CAA‟s Initial Views on the degree of

Stansted‟s market power. Excisions from the text are marked with [].

Rationale and context for the CAA’s airport market power assessments

1.2 Where airports enjoy high levels of market power, it can be appropriate for the

CAA to apply economic regulation, so as to improve outcomes for

passengers, cargo shippers and airlines. This regulation currently takes the

form of a cap on the prices charged by the airport and a series of financial

incentives and other obligations to encourage efficient operation, appropriate

service quality and efficient investment.

1.3 Most UK airports are not subject to this form of economic regulation. Airports

are only subject to this regulation if they have (or are likely to have)

„Substantial Market Power‟ and if economic regulation is likely to improve

outcomes. Further, when airports are subject to such economic regulation, it

can take a number of forms, and be tailored to the extent and nature of

market power.

1.4 The CAA is committed to ensuring that its regulation of airports promotes

choice and value for consumers, whilst also meeting the better regulation

principles. In order to deliver on this commitment, during 2011 the CAA

embarked upon a project to understand the extent and nature of market

power held by the airports that are currently „designated‟ for price control

regulation: Heathrow, Gatwick and Stansted. This work also addresses the

Competition Commission‟s view that the CAA should keep competition

between airports under review, and that the economic regulation of Gatwick

and Stansted might need to adapt to facilitate competition.

1.5 These assessments will inform the CAA‟s views on whether these three

airports should continue to be subject to economic regulation, including

whether – under the proposed reforms set out in the Civil Aviation Bill (2012)

– these airports should be required to hold an economic licence. The work

will also allow the CAA to work with stakeholders in developing future

regulation of these airports that protects consumers.

1.6 We set out below the initial findings of our work to date to understand the

market power held by Stansted. These views are based on the evidence

currently available to the CAA. As noted above, we expect that additional

evidence will become available before the CAA takes a firm view on the

position of Stansted. Reflecting this, there are a number of issues that have

been left open and, whilst the CAA has sought to highlight where each airport

might sit on the market power spectrum, it has not reached a definitive view

on the airport at this stage.

Competition Commission, „BAA Airports Market Investigation – Final Report‟, March 2009, paragraph

10.339.

Civil Aviation Authority 16

The CAA’s approach to the assessing airport market power

1.7 In reaching an initial view as to the degree of market power of Heathrow,

Gatwick and Stansted, we have followed the approach set out in the CAA‟s

Guidance on the assessment of airport market power

1.8 There has been extensive stakeholder engagement, including with the three

regulated airports as well as Luton and Manchester, and the main airlines of

Heathrow, Gatwick and Stansted. This engagement has taken the form of:

 meetings with stakeholders to discuss relevant evidence;

 stakeholder feedback and discussion on work in progress;

 the submission of evidence by stakeholders;

 some stakeholders, including both airports and airlines, have

commissioned reports by economic consultancies; and

 the CAA‟s stakeholder workshop held on 15 November 2011 to set out

its emerging views.

1.9 We have also published a number of working papers in 2011:

 on the general market context

;

 catchment area analysis

; and

 passengers‟ airport preferences

1.10 While undertaking the market power assessments, we also sought the advice

of an economic and a legal consultant.

1.11 The initial views expressed in this assessment are based on the evidence

currently available to us. We expect that additional evidence will become

available before we take a firm view on the position of each airport.

Reflecting this, there are a number of issues that have been left open and,

whilst we have sought to highlight where Stansted might sit on the market

power spectrum, it has not reached a definitive view on the airport at this

stage.

1.12 We would like to thank those stakeholders who have engaged with us during

2011, and invested time and money in providing evidence and analysis.

Way Forward

1.13 We welcome stakeholders‟ views on the information presented in this paper.

There are two periods over which interested parties can engage with the

CAA:

CAA Guidance on assessing airport market power

http://www.caa.co.uk/docs/5/Final%20Competition%20Assessment%20Guidelines%20-%20FINAL.pdf

CAA UK Airports Market – General Context September 2011

http://www.caa.co.uk/docs/5/20110905%20Market%20Context-FINAL.pdf

CAA Catchment area analysis October 2011

http://www.caa.co.uk/docs/5/Catchment%20area%20analysis%20working%20paper%20-%20FINAL.pdf

CAA Passengers’ airport preferences – Results from the CAA Passenger Survey November 2011

http://www.caa.co.uk/docs/5/Passenger%20survey%20results%20-%20FINAL.pdf

Civil Aviation Authority 17

 Those wishing to share their initial views with the CAA should aim to

submit any material to the CAA by 24 March 2012, so that these can

inform the CAA‟s next Q6 price control publication, which is scheduled

for April 2012.

 Those wishing to engage on the detail of the competition

assessments are invited to engage with the CAA during 2012, so that

any additional evidence and analysis can be incorporated in an

updated assessment of airport competition, scheduled for publication in

late 2012/early 2013.

1.14 If you would like to discuss the contents of this paper, and the CAA‟s work on

assessing airport competition, in the first instance please contact Alina

Jardine Goad on 020 7453 6229 / alina.jardinegoad@caa.co.uk or Alexander

Dünki on 020 7453 6212 / [email protected]k. You can also contact

Chris Hemsley on 020 7453 6237 / chris.hemsley@caa.co.uk.

1.15 If you would like to discuss the economic regulation of Heathrow, Gatwick and

Stansted, including the Q6 work programme, please contact Richard Moriarty

on 020 7453 6203 / richard.moriar[email protected]k.

1.16 The CAA will also be hosting a seminar, to take forward the work on price

control design for Stansted, on 16 March 2012. The output of these seminars

– and other discussions with stakeholders – will be brought together into a

publication in April 2012, which will frame the debate on the development of

regulation and, where appropriate, support the process of Constructive

Engagement. If you would like to register your interest in the seminars,

please contact Barbara Perata-Smith, on 020 7453 6202 or

Barbara.PerataS[email protected]k.

1.17 There will also be an opportunity to engage on the CAA‟s initial views on

airport market power during 2012, which will also allow the CAA to work with

stakeholders to narrow down those areas of uncertainty that currently exist.

Stakeholders will also be able to submit further evidence – and comment on

the CAA‟s initial views – during this period. At this stage, the CAA expects to

publish its next substantive analysis of airport market power in late 2012/early

2013.

Civil Aviation Authority

2. The markets in which Stansted competes

2.1 This chapter describes Stansted‟s business, its infrastructure and the

customers it serves. It looks at the varying requirements of different customer

groups and then goes on to analyse the product and geographic markets for

Stansted.

2.2 Reflecting the importance of the linkages between Stansted‟s various

customer groups, the final section of this chapter brings together the previous

sections and draws out the implications of these interrelations for market

definition.

2.3 We note that there is a degree of repetition in the evidence cited in this

chapter and in chapter 3. This reflects the fact that both chapters are seeking

to understand the strength of the competitive constraints on Stansted.

Stansted’s business

2.4 Stansted provides infrastructure and infrastructure services to its various

groups of users, including passenger airlines, passengers, and cargo carriers.

Each of these user groups has its own characteristics and infrastructure and

service requirements from Stansted, which are described below.

Provision of products and services to passengers

2.5 Stansted provides infrastructure and infrastructure services to passengers,

including:

 aeronautical services, including security clearance and flight

information and surface access infrastructure; and

 non-aeronautical services, including retail space and car parking.

2.6 Stansted does not currently directly charge passengers for the use of the

airport.

However, it nevertheless supplies a number of services to

passengers in their use of the airport and can vary its offering in ways that

vary the attractiveness of the airport to passengers. It also charges

passengers directly for the use of their car parks.

Provision of products and services to passenger airlines

2.7 Stansted also provides aeronautical and non-aeronautical infrastructure and

infrastructure services to passenger airlines. This includes facilities and

services for the handling of passengers and those for the transportation of

bellyhold cargo, although this is currently not a major requirement of airlines

at Stansted.

2.8 These services include:

 aircraft-related aeronautical services (e.g. facilities for landing, parking,

and taking-off of aircraft);

There is no statutory prohibition on the airport choosing to charge passengers. Indeed, a number of

regional airports in the UK (and a number of major airports overseas) impose charges directly on

passengers, in the form of „Airport Development Fees‟ and „Passenger Facility Charges‟.

Most passenger flights at Stansted do not carry cargo. Less than 1% of Stansted‟s cargo is carried

bellyhold (CAA airport statistics).

Civil Aviation Authority

 airline-facing non-aeronautical services (e.g. information technology

services, check-in desks and Common Use Self-Service (CUSS)); and

2.9 providing access to staff and supporting services procured by airlines (staff

security clearance, staff car parking, access to groundhandlers, access to

catering suppliers, etc).

2.10 We are adopting a working assumption that all of these services should be

included in the product market within which Stansted operates, as the failure

to supply any one of these services would severely hamper the economic

operation of a passenger airline, which means that any market power enjoyed

by virtue of the ownership of the runways, taxiways and terminals would also

be likely to be enjoyed over all of these services.

2.11 Airports also provide other services that are, to a greater or lesser degree,

ancillary to the operation of passenger services, such as office rental to

airlines or tour operators, either within terminals or at nearby locations. We

are adopting a working assumption that the provision of these services forms

part of a wider market for, in this case, office space.

Provision of products and services to cargo carriers

2.12 Stansted provides infrastructure and infrastructure services to cargo-only

carriers for the air transportation of cargo. This includes the provision of

runway and apron space, as well as providing access to cargo-handlers,

access to warehousing facilities, and infrastructure to allow the efficient

onward transfer of cargo.

Provision of access to Stansted’s infrastructure and services to third party service

providers

2.13 Many services for airport users are provided by third party service providers.

For airlines and cargo carriers, these service providers (contracted by them)

include ground- and cargo-handlers respectively, and maintenance and repair

operations.

2.14 For passengers, these service providers include food and drink, and providers

of other retail services (e.g. bureau de change). In order to provide these

services, suppliers typically need to rent terminal space and to obtain access

to the landside and airside facilities for their staff and to bring in stock.

The use of market definition

2.15 Defining the relevant market is usually the first step of any competition

assessment. It provides the context for the analysis by setting out the relevant

set of products and geographic areas which encompass the closest

substitutes for the products and services of interest.

2.16 However, as noted in our guidelines, market definition is not an end in itself,

but rather provides a frame of reference for the analysis of competitive

effects.

We note here that access to the airport, the construction and operation of its terminals and runways

are all controlled by Stansted Airport.

Civil Aviation Authority

2.17 In practice, in differentiated product markets, it can be difficult to draw a line

around a group of products with a varying degree of competitive constraint on

each other to define the market. In such circumstances it is appropriate to

consider the degree of competitive constraints faced by the product in

question in the round, regardless of whether they arise from within or outside

a defined relevant market.

2.18 Markets are generally defined in two dimensions: product and geographic.

The CAA‟s Guidelines set out the basis on which we have carried out the

market definition analysis. It should be noted, in particular, that in the absence

of sufficient data and given the difficulty in establishing the competitive price

level, rather than carry out a quantitative hypothetical monopolist test we have

assessed available evidence on the products supplied by the airport and

airline and passenger preferences using the principles of the test.

2.19 In particular, we have looked at the ability of airlines and passengers to switch

their business away from the airport, the ability of other airports to begin

supplying a substituted product to that of Stansted and the effect of these

factors of the profitability of any price rise by Stansted.

2.20 The CAA‟s approach is set out in more detail in the CAA‟s Guidance on the

Assessment of Airport Market Power.

2.21 The remainder of this chapter considers the product and geographic market

definitions from the perspective of airlines and passengers, looking at both

demand and supply side substitutability. We then draw this evidence together

and consider the effects of the interdependence of demand from the different

user groups and the implications of this on overall assessment of the markets

in which Stansted operates.

Product market definition

2.22 The following section considers each of Stansted‟s main customer groups. It

looks at the characteristics of these groups and whether there are separate

segments within these groups that have different product requirements.

2.23 By looking at the choice sets available to each of Stansted‟s customer groups

(and, if relevant, subgroups) on a product dimension and the ability to switch

between the different alternatives, we have assessed the potential for these

customers to react to an increase in charges by the airport. We have then

suggested product market definitions that include all alternatives that

customers can easily switch between.

Passenger demand side substitutability

2.24 A number of passenger sub-groups can be distinguished using various

characteristics:

 journey purpose;

„Guidance on the assessment of airport market power’, CAA, April 2011

We refer to market segments where we want to describe and delineate different groups of airlines and

passengers that are likely to have distinct features. This is not the same as describing these as

separate economic markets; we assess in this paper to which degree there might be separate markets

for different market segments, or to which degree they might suggest one, albeit differentiated, product

market.

Civil Aviation Authority

 destination; and

 passenger origin (inbound, outbound and connecting passengers).

2.25 We consider the relevance of these subgroups at Stansted and whether they

have varying requirements that might imply separate product market

definitions. We then look at passengers‟ ability to switch away from the

airport.

Journey purpose

2.26 One way to distinguish passengers is by their journey purpose. Figure 1

below shows the passenger breakdown by passenger origin

(UK or foreign

and journey purpose at Stansted (Holiday, Visiting Friends and Relatives

(VFR), or Business).

Figure 1 Passenger breakdown by passenger origin and journey purpose

Source: CAA analysis of the CAA Passenger Survey (2010)

2.27 Stansted serves both business and leisure passengers. Despite the fact that

its airlines do not generally offer a premium service, 16 per cent of its

passengers are business passengers. In addition, there is a recent trend

towards a number of airlines offering additional services that might be

particularly attractive to passengers travelling on business, such as priority

boarding and reserved seating, which may lead to an increase in business

travel on these airlines at Stansted.

2.28 In terms of the business/leisure split, it could be argued that the product

market for business passengers may be narrower than that for Holiday and

VFR passengers if they require the airport to provide different services,

service quality and/or infrastructure that only business passengers require.

However, the majority of the airport‟s services are offered to all passengers,

irrespective of their journey purpose. Indeed, there is no direct way for the

airport to distinguish passengers travelling on business from those travelling

for leisure purposes. To the extent that these two passenger groups have

The CAA‟s survey asks people to identify where they reside. This is used as a proxy to identify

whether passengers are originating from a UK district, or from overseas.

UK BUS

UK HOL

22%

UK VFR

26%

O/S BUS

O/S HOL

16%

O/S VFR

20%

Civil Aviation Authority

different car parking needs, for example business passengers are likely to

have more of a preference for short stay or valet parking, there may be some

potential for price discrimination here, but we do not consider this sufficient

enough to warrant separate market definitions.

2.29 Stansted does provide some services – through third-party suppliers – that

are particularly likely to be attractive to passengers travelling on business,

such as fast-track security and premium lounges. These facilities are not

restricted to business passengers and, given that business passengers can

choose whether to purchase these extra services, there does not appear to

be a strong rationale for identifying a sub-market for the provision of airport

services to business passengers.

2.30 A very large proportion (84 per cent) of Stansted‟s passengers are travelling

for leisure purposes, which can be split into those travelling for a Holiday (38

per cent of Stansted) and for VFR (46 per cent). This highlights that Stansted

has a particularly high proportion of its traffic accounted for by VFR traffic,

relative to other UK airports.

2.31 In general, passengers travelling for VFR purposes are less flexible in terms

of their destination than those travelling on a holiday, who may be able to alter

their destination. However, we have not identified any strong reason why

holiday and VFR passengers would consume different airport services (albeit

that the retail spend between these groups may differ).

2.32 Furthermore, individual services will typically have a mix of Holiday and VFR

(as well as business) passengers, making it difficult for the airport to

discriminate between these passengers. Consequently, there does not

appear to be a strong rationale for identifying a sub-market for the provision of

airport services to either Holiday or VFR passengers.

Destination

2.33 The vast majority of Stansted‟s passengers are short-haul, with only two per

cent travelling to long-haul destinations. This means that to understand the

market power of the airport, it is important to understand the characteristics of

short-haul passengers. Indeed, the very high shares of passengers

accounted for by short-haul flights implies that there is unlikely to be any need

to define a sub-market by destination type.

2.34 This focus of Stansted‟s operations on short-haul services has implications for

the choices available to passengers. In particular, short-haul passengers

typically face a wider choice set than long-haul passengers, as there are more

airports offering short-haul services and are more route overlaps between

short-haul services at different airports. In addition, for those passengers who

are flexible on their destination airport

there are more short-haul

destinations, at more airports, increasing the choice set further.

Some passengers might be seeking a “city break”, “ski holiday” or “Spanish beach holiday” and be

willing to choose between services to a similar type of destination.

Civil Aviation Authority

Passenger origin (inbound, outbound and connecting passengers)

2.35 Stansted provides services to both departing/outbound and arriving/inbound

passengers. The infrastructure services required by inbound and outbound

passengers differ somewhat, as outbound passengers typically have access

to more significant (airside) retail offerings, and also require security

clearance. However, there is a significant overlap in the services offered to

these two groups. Furthermore, given the significant proportion of

passengers who travel through the same airport for both legs of their trip, any

attempt to discriminate between inbound and outbound passengers would not

appear likely to be effective, as it would still affect the overall trip cost faced

by passengers. Reflecting this, we do not consider that there is a need to

distinguish between these groups in terms of the product market.

2.36 In respect of how passengers arrive at the airport, the majority of Stansted‟s

passengers arrive at the airport by surface travel, although there is a small

proportion of passengers who connect between services. These passengers

are „self-connectors‟, as the airlines at Stansted do not offer through-tickets or

formally support connections between services. Reflecting this, passengers

do not therefore require transfer facilities, such as airside security clearance

and systems for connecting baggage.

2.37 This lack of connecting passengers, supports the view that the airport is

competing to attract passengers onto point-to-point services, which is

consistent with views expressed to the CAA by the airport and by airlines.

Ability to switch

2.38 The above discussion highlights that the passengers using Stansted share

common key requirements and that there are factors that limit the airport‟s

ability to differentiate its offer in a way that might support a narrowing of the

passenger-facing product market.

2.39 There are few direct costs to passengers switching between airports, other

than the costs involved in travelling to an alternative airport (a factor that is

considered in the context of the geographic market, below).

2.40 However, the ability of a passenger to switch away from the airport will

depend on whether or not a suitable destination is available from an

alternative airport. Whilst the airport can structure its charges in ways that

can encourage particular services (such as targeted discounts to new

destinations), route choice is determined by the airlines rather than the

airport. Reflecting this, we have not identified a strong argument to narrow

the market in terms of different destinations. The impact of route overlaps on

switching is examined in the final section of this chapter.

Summary – Product market: passengers

2.41 In terms of the airport product, there is a degree of uniformity in the

requirements of the various passenger types at Stansted, and limits on the

extent to which the airport can segregate its offer to passengers or price

discriminate between them.

Civil Aviation Authority

2.42 In summary, therefore, it seems that the appropriate product market for

passengers at Stansted should be the general provision of airport

infrastructure and infrastructure services to passengers. This is likely to cover

a range of aeronautical and non-aeronautical services that are required for

the reception, processing and boarding of passengers.

2.43 However, the choice set for passengers will depend on whether or not a

suitable destination is available from the alternative airport, which we have

considered in the context of the geographic market, below.

Passenger airlines demand side substitutability

2.44 Airlines‟ infrastructure requirements are likely to differ according to their

business model and the type of service they offer. These differing

requirements may affect the choices available to different airlines and the

airports with which Stansted competes.

2.45 In this section we consider different ways of distinguishing between airlines

and whether this has any implications for product market definition, including:

 short-haul vs. long-haul;

 airline business model; and

 based vs. inbound

2.46 We then go on to consider airlines‟ ability to switch.

Short-haul vs. long-haul

2.47 In 2010, 98 per cent of Stansted passengers‟ final destinations were domestic

or European, with just 2 per cent of passengers flying long-haul, a figure that

is likely to be reduced further by the removal by Air Asia X of its services from

the airport.

2.48 The infrastructure needed to operate a short-haul route versus a long-haul

route are broadly similar, in that all airlines have a minimum requirement in

terms of runways, aprons and terminal facilities. However, long-haul routes

are more likely to be served by larger aircraft – requiring a higher specification

of runway – and they are more likely to carry bellyhold cargo, which can add

to the requirements of long-haul operations, relative to those operating short-

haul.

2.49 However, given that the vast majority of routes from Stansted are short-haul,

it is not necessary to consider any split for market definition purposes

between the supply of infrastructure for short-haul and long-haul services, as

it is unlikely to affect the overall assessment of market power.

2.50 Looking ahead, the airport may attract higher volumes of long-haul services,

which might have implications for the appropriate market definition. We

consider the airport‟s future growth, and efforts to attract additional services,

in chapter 3.

Airline business model

2.51 Stansted predominantly serves low cost, point-to-point airlines. Figure 2

below gives the airline business model breakdown at Stansted over the last

Civil Aviation Authority

ten years. This shows that by 2010, at least 94 per cent of Stansted

passengers flew on low-cost carrier (LCC)

airlines, with three per cent flying

charter and three per cent full-service carriers (FSC).

Figure 2 Share of Stansted’s passengers by business model, 2010

Source: CAA Airport Stats, 2010

2.52 Given that Stansted does not currently serve FSC or charter airlines to an

appreciable extent (despite Stansted‟s attempts to attract such airlines – see

chapter 3), this implies that the provision of facilities to these airline business

models is unlikely to determine the airport‟s market power. In the remainder

of this paper, we therefore concentrate our analysis on LCCs (operating short

haul, point-to-point services), since it is the interaction between the airport,

these airlines and their passengers that will determine the extent of

Stansted‟s market power.

2.53 In terms of their infrastructure requirements, LCCs have fewer needs to

FSCs, particularly those operating networks with connecting passengers.

LCCs do not require facilities to accommodate transfer passengers and

usually require more basic terminal facilities than FSCs.

2.54 The LCC business models do, however, rely upon achieving high levels of

utilisation of their aircraft, by reducing the time that aircraft are on the ground.

This can be achieved by efficient turnaround times, short taxiing times and

low levels of airfield delay. These requirements can limit – to a degree – the

airports that no-frills airlines view as being reasonable alternatives to

Stansted. In particular, relatively uncongested airports, or those supporting

reliable and efficient turnarounds, may be close alternatives (for this

dimension of comparability).

2.55 Consequently, congested airports or those with relatively high levels of airfield

delays and relatively low levels of punctuality might not be viewed as being

We generally adopt the term „low-cost carrier‟ to identify airlines, such as easyJet, Ryanair, Monarch

and Jet2 that operate point-to-point services. These airlines are often referred to as „no frills‟ airlines. In

reality, individual airline business models vary, and we use these terms to differentiate these airlines

from „full-service‟ airlines, that offer different ticket classes, support connecting traffic and tend to

operate using a more diverse fleet of aircraft.

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Low-Cost

Full-Service

Charter

Civil Aviation Authority

close alternatives to Stansted. For this reason, we do not consider that

Heathrow is likely to be viewed as a reasonably close alternative airport for

most of the airlines operating from Stansted

2.56 A further factor that can affect the utilisation of aircraft is the ability of airlines

with aircraft based at the airport to leave early in the morning. Early morning

departures increase the likelihood that the airlines are able to achieve a high

number of rotations each day. This may have implications for the ability of

airline to change their usage pattern of Stansted across the day. This is

discussed in more detail below.

Based vs. inbound

2.57 Another way in which airlines can be distinguished is by whether or not they

base aircraft at Stansted overnight, or whether they are based at another

airport, their first service to Stansted being operated inbound.

2.58 As noted above, the no-frills airlines at Stansted are likely to place a particular

premium on high levels of aircraft utilisation, which means that when they

base aircraft at the airport they will also place a premium on the ability to

depart from the airport early in the morning. This preference for early-

morning departures may also be compounded by the preferences of some

passengers (notably those travelling for business) to leave relatively early, in

order to arrive at their destination in time to carry out a full day‟s work,

although the share of business passengers at Stansted is relatively low.

2.59 In contrast, whilst operators of inbound aircraft might also place a premium on

high levels of utilisation, the differing sector lengths involved in flying into

Stansted spreads the arrival times over a broader period of the day, relative to

the usage pattern of the first wave of departing aircraft.

2.60 So while the demand for slots is more spread out for non-based aircraft, a

based aircraft will have a particular requirement for access to an early

morning slot. This raises the question of whether the airport is able to

differentiate its offer between based and inbound aircraft to such an extent

that they should be considered to be separate product markets.

2.61 In principle, the airport can vary its charging structures in ways that

particularly affect based or inbound carriers, including by varying charges by

the time of day, or through varying overnight parking charges. The airport‟s

ability to differentiate between inbound and outbound services will also

depend upon the extent to which the airlines operating these services can

respond to any price change and reduce their use of the airport. In this

context, it is relevant that the main airlines at Stansted operate both based

and inbound aircraft, and have a number of bases across Europe which

would, in principle, allow them to switch outbound operations to be inbound

(and vice-versa).

2.62 However, a switch between outbound and inbound operation changes the

flight timings that can be offered to passengers. To the extent that local

passengers have strong time preferences, this might act as a barrier to

We discuss later the extent to which passengers might still view the two airports as being substitutes.

Civil Aviation Authority

airlines switching between inbound and outbound operations. The relatively

low levels of business passengers at Stansted might suggest a higher degree

of flexibility over departure times than at London City or Heathrow, which both

have particularly high proportions of business travellers (33 and 63 per cent

respectively)

2.63 On the basis of the evidence available, it might be appropriate to distinguish

between the supply of infrastructure to based and inbound operators, in light

of the potential costs associated with switching between based and inbound

operation, and the emphasis placed by airlines at Stansted on operating

based services at peak times. Further analysis of this issue would require

airline data on route revenues, to allow comparisons between outbound and

inbound profitability, so as to understand better the costs of switching an

outbound service to an inbound one. This would allow the CAA to understand

if switching services in this way would undermine an attempt to raise prices

above the competitive level to on or other of these groups.

2.64 These issues are considered in more detail in the context of whether there are

„temporal‟ markets, from paragraph 2.78.

Ability to switch

2.65 The switching costs of airlines are considered in more detail in chapter 3,

however as the ability of airlines to switch away from Stansted will affect the

assessment of the relevant market, we also discuss this issue below (in

respect of those factors that affect the product market) and in the following

section, in the context of the geographic market.

2.66 The airlines at Stansted are predominantly LCC airlines. The infrastructure

and service level requirements of these airlines will tend to be below those of

a FSC airline. As discussed above, LCCs will typically not require facilities to

connect passengers and baggage between flights, or to facilitate the carrying

of bellyhold cargo, or facilities targeted at premium passengers, such as

business lounges.

2.67 Furthermore, the aircraft flown by Ryanair

and easyJet

can operate from a

wide range of airports across the UK and Europe, albeit that some

commercial airports – such as Southampton and London City – have runways

that do not allow some common aircraft types to operate. These factors

mean that the choice set of the major airlines at Stansted is relatively wide

(below we discuss factors that might limit the ability to exercise these choices,

in the context of the geographic market).

2.68 Indeed, Ryanair, the main airline at Stansted, currently also operates at a

number of much smaller airports than Stansted, which highlights that its

minimum requirements are relatively low and below the level of infrastructure

available at Stansted (with its long runway and cargo facilities). This implies

See Figure 14 of „UK Airports Market – General Context‟, CAA, September 2011

As at November 2011: 275 Boeing 737 800s (Ryanair Half Year Results, Nov 2011)

As at 30 September: 167 A319s, 35 A320s, and 2 Boeing 737-700s (easyJet Annual Report 2011)

Civil Aviation Authority

that most mid-sized airports would be appropriate substitutes to Stansted in

terms of the product offering.

2.69 Turning to the impact of an increase in airport charges on the airlines

operating at Stansted, airport charges are generally a much higher proportion

of LCC‟s costs than they are for FCCs, and for those operating network and

long-haul services (such as BA, BMI and Virgin). As a result, the airlines at

Stansted are likely to be more sensitive to an increase in airport charges than

many of the airlines operating at Gatwick or Heathrow.

2.70 Figure 3 shows the share of operating costs accounted for by airport-related

costs for the major UK airlines. It illustrates that these costs constitute about

10 per cent for BA and Virgin („VS‟), which is representative of network/long

haul based carriers, and about 20 per cent for bmi and Flybe („FB‟), which is

representative for carriers with a more regional focus and smaller aircraft,

albeit that BMI has a mix of regional, short-haul and mid-haul operations. In

contrast, airport-related costs account for about 30 per cent for easyJet, an

LCC.

Figure 3 Shares of operating costs of major UK airlines

Source: CAA airline account information, latest available financial years

2.71 Figure 4 shows a summary of easyJet operating costs, which highlights the

importance of costs relating to ground operations (£14.79 per seat flown),

which represents 45 per cent of total operating costs (excluding fuel). This

Ryanair currently operates from 160 airports, with 47 bases. easyJet currently operates from 123

airports, with 19 bases.

This analysis has been taken from the CAA‟s airline statistics, which only cover airlines registered in

the UK.

Figures taken from Table 6 of the 2009/10 airline accounts published regularly on the CAA‟s website:

http://www.caa.co.uk/default.aspx?catid=80&pagetype=88&pageid=13&sglid=13. Airport-related costs

for the purpose of this figure include the following line items: 22, 24, 25, 27. This is likely to include also

costs for services that fall outside the services relevant for this assessment, for example for ground

handling services. Costs charged for relevant services provided by airport operators are therefore likely

to constitute a lower share.

2.7%

4.0%

11.2%

10.4%

18.9%

9.2%

5.3%

7.8%

8.8%

9.6%

4.5%

3.8%

5.9%

4.9%

8.9%

34.6%

21.9%

28.5%

26.2%

15.7%

29.7%

35.9%

18.8%

16.1%

31.0%

19.3%

29.1%

27.9%

33.6%

16.0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

bmi

EZJ

Other

Fuel

Crew and aircraft related costs

(except maintenence and overhaul)

ANS charges

Other airport-related costs

Airport charges

Civil Aviation Authority

compares to the current airport charges at Stansted, which are equivalent to

£7.26 per passenger for 2012/13.

Figure 4 easyJet operating costs (excluding fuel)

Source: easyJet Annual Report, 2011

2.72 The airlines operating at Stansted are able to use a large number of other

airports – in operational terms – and have cost structures that mean that they

are more sensitive to increases in airport charges. In addition, the airlines at

Stansted typically make relatively limited investments in facilities at the

airport, particularly when compared to the investments made by full-service

and network carriers. Indeed, both Ryanair and easyJet stress the flexible

nature of their operations.

25,26

2.73 It appears that the costs involved in switching from one airport to another in

terms of the physical move would be limited. It should be noted, that in terms

of market definition and market power assessments, we only need to consider

a switch at the margin, i.e. it would not be necessary for Ryanair to switch its

whole operation at Stansted (which may be considerably more costly) to

another airport, but just a share of its business.

2.74 Consequently, the main barrier to switching is the ability to operate a

profitable route out of an alternative airport and the impact that switching a

route from Stansted to the next-best airport might have on the yields earned

by the airline. This is discussed further in chapter 3.

2.75 For the purposes of the product market definition, this supports the view that

the airlines at Stansted operate in a broad market, and have the flexibility to

use a large number of airports, albeit that there are a small number of

commercial airports where the infrastructure does not support their

operations, due to the capability of the runways. The operational

characteristics of Heathrow, with its congestion, relatively low levels of

punctuality and longer taxiing and airspace delays, suggests that this airport

is unlikely to be a viable alternative for the airlines currently at Stansted.

„BAA Stansted Airport Charges – April 2012-March 2013‟, BAA

See, for example, interview with Michael O’Leary, retrieved from www.anna.aero, quoted in footnote

75.

“easyJet has built flexibility into its fleet planning arrangements such that it can increase or decrease

capacity deployed, subject to the opportunities available and prevailing economic conditions. The

Company also has flexibility to move aircraft between routes and markets to improve ROCE.”, easyJet

Annual Report (pg 11), 2011.

Civil Aviation Authority

Summary: product market – passenger airlines

2.76 In summary, the product market definition for Stansted from the perspective of

airlines appears to include any mid-sized airport (or larger) with an

appropriate runway. However, airlines use airports in order to gain access to

passengers and require access to a catchment that is sufficiently large and

affluent to support the profitable operation of services. This narrows the set of

airports that might be included in the product market, to include only those

airports that provide access to a relatively large pool of demand, as an airline

would not be able to switch to an airport with a poor catchment even if it had

the appropriate infrastructure to operate there. This interplay between

passengers and airlines is discussed further in the final section of this

chapter.

2.77 There may be a case for distinguishing between the supply of infrastructure

and infrastructure services to based and inbound aircraft, reflecting the

different demand conditions throughout the day, the uncertainty surrounding

the ability to switch between based and inbound, and the ability of the airport

to vary its charges between these two aircraft types. This issue appears

particularly important in the context of the overall assessment of Stansted,

and is considered further below in the context of whether a distinction can be

made between peak and off-peak capacity. Reflecting the current balance of

evidence, the CAA has not yet taken a firm view on this aspect of the market

definition for Stansted.

Temporal markets

2.78 The previous section focused on how the differing airline infrastructure

requirements, might inform the relevant product market for the assessment of

Stansted. This section further considers whether there is evidence to support

a narrowing of the product market for Stansted, to distinguish between the

services supplied to airlines flying at different times of day.

2.79 The nature of the capacity available at Stansted means that, in operational

terms, it is highly substitutable between different aircraft types. The runways

and taxiways are, for example, capable of handling the largest passenger

aircraft currently operating. This means that there are few physical limitations

to the ability of airlines to redeploy their slots between services at different

times.

2.80 However, there are commercial and operational constraints that mean that

certain services at certain times of day are more popular for airlines. In

particular, based airlines at Stansted tend to place particular value on the

early morning slots, in order to deliver a sufficient number of rotations to

deliver high aircraft utilisation, and to make a route profitable.

2.81 This may imply separate markets for the provision of peak and off-peak slots

since airlines currently using morning peak slots could not easily switch to off-

peak slots. Although Stansted does not currently price differentiate by time of

day, other airports do and there are no legal barriers to such a charging

structure. The need to distinguish between peak and off-peak (and adopt a

narrower product market definition) will, therefore, depend upon the ability of

Civil Aviation Authority

airlines to reduce their use of peak capacity in response to price, either by

switching to off-peak periods, or to other airports.

2.82 This issue of the flexibility of airlines at peak is important to the overall

assessment of Stansted‟s market power. The CAA has not been able to

reach a definitive view on the issue, and so we leave this aspect of the market

definition open at this stage. This issue is discussed in more detail in chapter

Cargo carriers

2.83 While Stansted‟s passenger airline customers do not generally carry bellyhold

cargo, 6.8 per cent of Stansted‟s flights are cargo-only flights, making the

airport the UK‟s second largest for cargo-only flights, with an 18 per cent

share of UK cargo-only flights. Stansted‟s cargo customers are predominantly

freighter carriers and less than 1% is carried bellyhold. Table 1 shows

Stansted‟s main cargo customers.

Table 1 Stansted’s cargo customer shares of freighter cargo

FEDERAL EXPRESS CORP

37%

BRITISH AIRWAYS PLC

24%

UNITED PARCEL SERVICE CO

14%

TITAN AIRWAYS LTD

JET2.COM LTD

ABX AIR INC

ASIANA AIRLINES

Other

Source: CAA Airport Stats 2010

2.84 As for passenger airlines, the choice set for cargo carriers will be limited by

their infrastructure and service requirements, and the relevant product market

will be affected by their responsiveness to price changes. In this respect, the

infrastructure requirements of cargo-only flights are likely to affect the choice

set of these airlines, as these services have particular runway and

groundhandling requirements, as well as the need to access appropriate

surface access.

2.85 In respect of the product market, there are a number of airports in the UK that

are capable of accepting cargo-only services, albeit that some impose

operational restrictions on the aircraft that can be used. For example, there

are significant cargo-only operations at East Midlands, Manston, Robin Hood,

Manchester, and a number of other UK airports. As some cargo operations

typically use large aircraft, notably Boeing 747s, the infrastructure

requirements are likely to narrow the product market somewhat. In addition,

the limitations imposed by airports under the Traffic Distribution Rules mean

that there cannot currently be an increase in the cargo-only services at

Heathrow, effectively removing this airport from those that can compete with

Stansted for cargo-only aircraft.

Heathrow will compete, to some degree, with these cargo-only flights, as it is the UK‟s largest cargo

airport, with the majority of its cargo being carried as bellyhold.

Civil Aviation Authority

Provision of infrastructure and access to third-party suppliers

2.86 As noted above, there are a number of services that are provided by third-

parties, under contract to airlines, such as groundhandling services, fuel

providers, catering and maintenance. In addition, there are a number of

services provided by third parties to passengers, such as retail and catering.

The airport provides access and terminal space to these third parties. The

airport also typically controls the overall retail offering within the terminal,

often with a revenue-sharing arrangement.

Suppliers to airlines

2.87 Access to Stansted for these airline suppliers is crucial to the airlines‟

operations, with all possible substitutes at the airport requiring access to the

terminal and apron facilities. For this reason, we consider that the product

supplied by the airport to airlines includes the provision of access to third

party suppliers. Further, any market power enjoyed by virtue of the runway,

apron and terminal infrastructure would also be enjoyed in respect of

providing access to these third-party suppliers.

Suppliers to passengers

2.88 In respect of suppliers of passengers, it is useful to distinguish between retail

services that are related to the passage of passengers through the airport,

where passengers may have low levels of discretion as to whether they might

make a purchase, and those where purchases are highly discretionary.

2.89 For those services where most passengers will have a choice as to whether

to consume the product or service – such as fashion and electronics – we

have identified no reason to suggest that passengers would be any less

sensitive to the price and service offering than they would be in other retail

environments. Further, the suppliers of these goods and services will

generally have other options to providing their services at Stansted, since

they are part of wider retail markets. It is unlikely, therefore, that Stansted

would be able to profitably raise prices to these suppliers, or for the suppliers

to raise prices to passengers, above the competitive level

. Consequently,

we do consider them again in this paper.

2.90 Turning to those services that are less discretionary – which might include the

provision of food and drink, and travel essentials, such as bureau de change

– it is possible that a significant proportion of passengers have somewhat

limited options as to whether or not to purchase these services. This will vary

depending upon the individual service being considered. For example, given

the waiting times that can be involved in air travel, and security restrictions,

many passengers may face little choice but to purchase some food and drink.

In such circumstances, the airport could limit the space available to particular

retail offerings, and/or limit competition between different suppliers, and in

doing so put upward pressure on prices to passengers and rental payments to

the airport (including any revenue sharing arrangements).

The locational characteristics of airports, with their consistent and high footfall might, however, justify

a premium on retail rents over some other retail environments.

Civil Aviation Authority

2.91 It appears, therefore, that if the airport enjoys a position of substantial market

power over the provision of aeronautical infrastructure, it might also enjoy a

similar position in respect of some retail activities.

Supply-side substitutability

2.92 The CAA has adopted the OFT‟s view that “supply-side substitution can be

thought of as a special case of entry – entry that occurs quickly, effectively

and without the need for substantial sunk investment.”

2.93 Airports that are able to quickly begin providing the same infrastructure and

service as other airports in the relevant product market can be included within

that relevant product market.

2.94 However, given the investment and timescales needed to upgrade an airport‟s

facilities, it is unlikely that we would consider this kind of supply-side

substitution quick and effective enough for this type of airport to be

considered in the relevant market. Instead we will consider them as

constraints through potential entry.

Geographic market definition

2.95 The previous section considered the types of airports that are likely to be

included in the same product market definition as Stansted. However, in order

to understand which airports compete with Stansted for its customers, we

must also consider the geographic market in which it operates. The

geographic market will be affected by the airports that each type of user

considers to be reasonable alternatives to using Stansted.

2.96 Reflecting this, to reach a view on the geographic market for Stansted, the

remainder of this section considers the ability of each type of user to relocate

to alternative airports. As with the product market definition, in the following

section, we look at the choice set available to each of Stansted‟s different

user groups (and if relevant, subgroups) on a geographic dimension and their

ability to switch between the different alternatives. We also assess the

potential for these users to react to an increase in charges by the airport.

2.97 We have then suggested geographic market definitions reflecting the

alternatives that each type of user – passengers, passenger airlines and

cargo airlines – can readily switch between.

Passengers

2.98 In past cases, competition authorities have considered that the geographic

market for passengers extends to include at least the south east of England.

This is same geographic scope for the market defined by the Office of Fair

Trading (OFT) in the context of the BAA airports market investigation, while

the Competition Commission left it open

. We have found that the available

Para 3.56 of the CAA‟s guidelines

OFT and CC paper refs http://www.oft.gov.uk/shared_oft/reports/transport/oft882.pdf

Civil Aviation Authority

evidence tends to support a market for passengers that includes at least the

south east of England.

2.99 The ability of passengers to switch to different airports will vary depending on

whether a passenger is inbound or outbound, their journey purpose and their

destination. This implies that the area over which the airport attracts

passengers may vary by passenger type.

2.100 The following sections look at the different passenger types and considers

whether there is any impact on geographic market definition resulting from

different passenger characteristics.

UK resident vs. non-UK resident

2.101 Whether a passenger is resident in the UK or outside of the UK may affect the

choice set available to them on a geographic level. Figure 5 considers the

impact that residence has on passengers‟ average journey times to and from

the four major London airports. As discussed in the CAA‟s working paper on

catchment areas

, throughout this section we have focused on the travel

times accepted by 80 per cent of passengers.

2.102 It shows that UK resident passengers are more likely to travel further to reach

an airport, compared to those residing overseas. Further, the analysis of

Stansted passengers reveals that foreign residents travel between 60 and 75

minutes to/from the airport, compared to between 90 and 120 minutes for UK

residents.

Figure 5 Surface travel time according to residence (4 major London airports)

Source: CAA analysis of the CAA passenger survey 2010 and DfT surface access data

2.103 These differences based on residence could reflect a number of factors,

including the potential greater awareness of alternative airports by UK

residents, and the tendency for a larger proportion of foreign residents to be

For clarity, we refer here to a broad geographic area of England, which includes Greater London, East

Anglia and the „home counties‟. This should be distinguished from the South East planning region,

which does not include Greater London.

„Catchment area analysis – working paper‟, CAA, October 2011

We note that this is a benchmark that has been adopted by other competition authorities, such as the

Competition Commission in its analysis of supermarkets.

See „Catchment Area Analysis – Working Paper‟, CAA, October 2011 (Annex). Figures quoted are

the travel times that relate to the 80

percentile of passengers.

Civil Aviation Authority

travelling to major cities, notably to London, whereas the point of origin of UK

residents is more distributed.

2.104 However, there may be other important differences between UK residents and

foreign residents. For example, UK residents travelling outbound will only be

able to switch to other nearby airports, irrespective of their choice of

destination. In contrast, a foreign resident travelling inbound to the UK might

be able to vary their destination and travel to an alternative airport located in a

different town or country as well as to other nearby airports. This ability might

be offset somewhat by the lower level of awareness of alternative UK airports.

2.105 However, as the airport does not contract directly with passengers, it is

unlikely to be able to distinguish between passengers on the basis of

residence. Further, as discussed above, the airport is unlikely to be able to

successfully differentiate between inbound and outbound passengers, as the

majority of passengers travel back to their original departure airport, which

means that any variation of charges between inbound and outbound has a

common effect on the passenger‟s overall trip cost.

2.106 It is possible, however, that the airport could vary other charges in ways that

indirectly affect UK and foreign (or inbound/outbound) passengers differently,

such as through car parking charges (which are more likely to affect UK

resident/outbound passengers).

2.107 Overall, we do not consider that there is sufficient evidence to distinguish

between UK and foreign (or inbound/outbound) passengers for the purposes

of this assessment.

Short-haul vs. long-haul

2.108 The willingness to travel to use an airport can vary depending upon the

passengers‟ destination. Figure 6 below shows that the surface travel time for

domestic passengers is less than that of short-haul passengers, which is less

than that of long-haul passengers. This pattern might be explained by the

larger proportion of domestic and short-haul passengers‟ overall journey time

that would be accounted for by surface travel, when compared to long-haul

services. The greater availability – and higher level of route overlaps – of

domestic and short-haul services might be a further explanatory factor, as

long-haul passengers are more likely to have to accept longer surface access

journeys in order to reach certain long-haul destinations.

An airline may be able to distinguish in these ways, and may choose to pass through an increase in

airport charges disproportionately to the less price elastic passengers. We examine the relevance of this

in the final section of this chapter.

Civil Aviation Authority

Figure 6 Surface travel time according to flight duration (4 major London

airports)

Source: CAA analysis of the CAA Passenger Survey 2010 and DfT surface access data

2.109 For the purposes of defining the geographic market of Stansted, and

reflecting the large proportion of passengers that travel on short-haul

services, this analysis suggests that a large number of passengers are willing

to accept journey times of up to 90 minutes to reach an airport. However, the

average journey times for some short-haul passengers are much higher than

the figures shown above (which relate to all four major London airports). For

Stansted, UK resident short-haul passengers travel between 90 and 120

minutes to reach the airport, whilst foreign residents travel around 75 minutes

to reach the airport.

2.110 Whilst Stansted might not have any significant long-haul operations, to the

extent that the airport competes to attract long haul passengers to the airport

in the future (and attract airlines to operate these services), it is likely to

compete for passengers over a relatively broad geographic area, reflecting

the tendency for long-haul passengers to be willing to accept longer surface

access journey times than those travelling on domestic or short-haul services.

Journey purpose

2.111 Figure 7 below shows the surface travel time for UK residents at Stansted

travelling short-haul, distinguishing between business, holiday and VFR. It

shows that VFR passengers travelling for around 90 minutes to reach the

airport, less than holiday or business travellers.

See Figure A. 17, „Catchment Area Analysis – Working Paper‟, CAA, October 2011.

Civil Aviation Authority

Figure 7 Surface travel time for UK residents at Stansted

Source: CAA analysis of the CAA Passenger Survey 2010 and DfT surface access data

2.112 This pattern is somewhat different at Stansted than at the other London

airports, where holiday passengers on average travel further to reach an

airport than business or VFR passengers.

One potential explanation, cited

in the CAA‟s Catchment Area Analysis working paper, is that the business

passengers may be travelling to Stansted to fly on one of the large number of

services that are only offered from the airport. The lower average surface

travel times for VFR passengers could also be explained by the concentration

of these passengers‟ point of origin in Greater London and, in particular,

central, north and east London.

2.113 In light of the mixed evidence on average travel times, and the practical

difficulties associated with the airport distinguishing between passengers on

the basis of their journey purpose, it does not appear to be appropriate to

define different geographic markets based on journey purpose. Instead, the

average surface access times of 90 to 105 minutes appear to be appropriate.

Passenger preferences

2.114 The CAA working paper on passenger survey results published in November

looked at which airports passengers considered using or had used in the past,

their reasons for choosing the airport and their responsiveness to a change in

price.

2.115 Figure 8 shows the airports previously used in the past two years and airports

considered as alternatives for the current flight by short haul passengers at

Stansted. It shows that the airports that the passengers at Stansted had

previously used and considered were, perhaps unsurprisingly, Heathrow,

Gatwick and Luton. London City, Manchester and Birmingham were also

previously used by at least five per cent of passengers. This supports the

view that the geographic market for passengers is at least as wide as greater

London.

See Figure 10, „Catchment Area Analysis – Working Paper‟, CAA, October 2011.

„Passengers‟ airport preferences: results from the CAA Passenger Survey‟, CAA, November 2011

Civil Aviation Authority

Figure 8 Airports previously used in the past two years and airports

considered as alternatives for current flight by short haul passengers at

Stansted

Source: CAA analysis of CAA Passenger Survey data – supplementary stated preference question

2.116 Figure 9 below shows that the three main reasons for airport choice at

Stansted were location and surface access, cost and routes and frequency,

with the most popular being location and surface access.

Figure 9 Reasons for airport choice

Source: CAA analysis of CAA Passenger Survey data (January to July 2011 provisional)

Civil Aviation Authority

2.117 These results imply that when passengers are choosing an airport, the

airport‟s facilities and services are not particularly important factors.

Instead, the results suggest that flight availability, cost and the journey time to

reach an alternative airport are particularly important factors, which are also

likely to affect passenger switching costs.

2.118 In terms of informing the geographic market definition (and the market power

assessment), these results are consistent with an approach that considers

measures of surface access journey times, route overlaps and cost.

Catchment area analysis

2.119 A common approach to defining the geographic markets of airports is to look

at catchment areas. The evidence in the CAA‟s working paper on catchment

area analysis suggests that the geographic market for Stansted covers at

least the South East of England (including Greater London), perhaps also

extending to include the East of England.

2.120 Figure 10 shows the 60, 90 and 120 minute travel time isochrones around

Stansted. It shows that within 90 minutes, it is possible to travel across

London or north as far as north Cambridgeshire.

Figure 10 60, 90 and 120 minute surface access travel times for Stansted

Source: CAA analysis of DfT surface access data

Red: 60 minutes; Orange: 90 minutes: Yellow: 120 minutes

2.121 Figure 11 illustrates the actual time travelled by passengers at Stansted. Pale

green and dark green shows the area covered using the travel time within

which 80 per cent of passengers travelled. It is very similar to the chart above,

reflecting the fact that 80 per cent of Stansted‟s passengers travel 90 minutes

or less to reach the airport.

The passenger survey included a number of potential responses relating to airport facilities and

services. These responses, together with a number relating to airline factors, are grouped within „other‟

in the figure.

Civil Aviation Authority

Figure 11 Stansted actual surface travel time catchment area

Source: CAA analysis of the CAA Passenger Survey 2010 and DfT surface access data

Dark green – 70%; Light green – 80%: White – 90% of passengers

2.122 Lastly, Figure 12 shows the historical actual usage by passengers of

Stansted. This was compiled by listing each district in passenger number

order. Those districts encompassing the first 70 per cent of passengers are

shaded dark green, those encompassing the next 10 per cent (from 70 to 80)

are shaded pale green and the districts encompassing the next 10 per cent

(from 80 to 90) are shaded white.

Figure 12 Stansted historical actual usage catchment area

Source: CAA analysis of the CAA Passenger Survey (2010)

2.123 This figure highlights that Stansted has drawn passengers from the Greater

London area, the area of the South East that is north of London and much of

the East of England. The fact that it does not appear to draw strongly from

south or west of Greater London is likely to reflect the additional travel time

Civil Aviation Authority

involved in travelling through or around London, and that passengers would –

other things being equal – prefer to travel to their closest airport.

Catchment area overlaps and the geographic market

2.124 It should be noted that the geographic market is not defined by the catchment

area of an airport, as this shows the current pattern of passengers‟ use of an

airport, rather than describing the geographic area from which the most

important competitive constraints arise. A geographic market can, therefore,

include a broader area, to the extent that airports attracting passengers over

these areas impose important competitive constraints on the airport being

assessed. Further, a narrow catchment area (based on historical usage)

could be a product of active competition from neighbouring airports for the

passengers originating nearby the airport being assessed.

2.125 It should be noted that there are significant overlaps between the catchment

areas of Stansted and all of the major London airports, as they all draw a

significant proportion of their passengers from the Greater London area.

Indeed, the CAA‟s Catchment Area working paper considered the strength of

overlaps with the other London airports.

2.126 The results of this analysis are reproduced in Table 2 below. This table treats

a district as being in a catchment if it contributes to the nearest (by surface

travel time) 80 per cent of passengers using an airport. Districts can,

therefore be in the catchment of multiple airports. For each type of overlap,

the table shows the number of districts, the total number of passengers

originating from that overlap area (across the four airports), the passengers

using Stansted – both in terms of absolute numbers and proportions of the

total – and, in the final column – Stansted‟s share of passengers originating in

each of the overlap areas. More information is set out in the CAA‟s

Catchment Area Analysis Working Paper.

Table 2 Overlaps in Stansted’s surface travel time catchment area

Source: CAA analysis of the CAA Passenger Survey 2010 and DfT surface access data

2.127 This analysis highlights the strength of overlaps with the other London

airports, with 64 per cent of Stansted‟s passengers originating from (travelling

to) a district that falls within the catchment area of all three of Heathrow,

Gatwick and Stansted. Only 5 per cent of Stansted‟s passengers originate in

a district that does not fall within the catchment areas of at least one of these

airports. In addition to these airports, the catchment area of Stansted

Civil Aviation Authority

suggests that the airport interacts significantly with both Norwich and

Southend airports.

2.128 In this context, we also note that the results of the modelling undertaken by

Frontier (for easyJet) using the DfT passenger allocation model, suggests that

passengers responding to price increases at Stansted relocate to Gatwick

(around 40 per cent of those switching), Heathrow (20 per cent) and Luton (16

per cent), with some significant switching to non-London airports (14 per

cent).

This modelling is discussed in more detail in chapter 3.

2.129 This supports the view that the relevant geographic market covers much of

the south east of England, including Greater London and a number of districts

falling in the East Anglia planning region.

Summary: Passengers – Geographic Market

2.130 In summary, there does not appear to be any reason to distinguish between

different geographic markets by passenger types at Stansted.

2.131 Passenger survey evidence suggests that most passengers using Stansted

would be willing to travel up to 90 minutes to use an airport. In combination

with catchment overlap evidence, this would suggest that Stansted‟s

catchment covers the South East of England, and includes Greater London.

2.132 In addition, historical usage shows that Stansted draws a significant number

of passengers from the East of England, implying that the geographic market

for Stansted passengers should be broadened to include a number of districts

in the East of England.

Passenger airlines

2.133 To form a view on the relevant geographic market for the provision of

infrastructure and infrastructure services to passenger airlines, it is useful to

consider the airports to which the airlines using Stansted could switch a

significant proportion of their business in light of a price rise.

2.134 The ability for airlines to switch between different airports on a geographic

dimension will depend, to some extent, on the airline‟s business model,

destinations and whether the aircraft or based at the airport or operates

inbound.

2.135 As discussed above, Stansted predominantly serves LCCs, operating short-

haul, point-to-point services. The business model of these airlines tends to

involve basing aircraft at airports of different sizes across Europe from which

they offer point-to-point services. As discussed above, airport charges are

also typically a higher proportion of their overall cost base than other carriers,

suggesting a higher degree of responsiveness to a given price change.

2.136 Furthermore, we have already highlighted that the largest airlines at Stansted

have a relatively large number of existing bases across Europe, and a large

number of airports to which they already operate:

See http://www.caa.co.uk/docs/5/Frontier%20pre-

Market%20power%20assessment%2006122011.pdf for the results of this modelling. Percentages are

based on the unrestricted modelling for Stansted, comparing the modelled increases at other airports to

the decrease at Stansted.

Civil Aviation Authority

 Ryanair operates from 160 airports, with 47 bases; and

 easyJet operates from 123 airports, with 19 bases.

2.137 This characteristic will tend to reduce the costs associated with moving

capacity away from Stansted, albeit – as noted above – that doing so might

result in a reduction in the revenues earned by the airlines.

2.138 These airlines, and other no frills operators, also have a track record of

moving services between different airports, cancelling under-performing

routes and moving their growth across a number of different European

markets. Given that they already have bases across Europe, it should be

easier for them to switch aircraft between bases than it would be for a network

carrier to relocate to a new base.

2.139 In addition, a number of the services at Stansted operate on an „inbound‟

basis, with their first service arriving at the airport, rather than being based at

the airport overnight. These services are likely to be particularly flexible and

responsive to prices at Stansted, as the costs of moving services between

two destination airports are likely to be somewhat lower than the costs

involved in moving an aircraft between different bases.

2.140 At present, whilst a significant number of services at Stansted operate

„inbound‟, the majority operate on an „outbound‟ basis. These inbound

services are, however, still significant in the context of understanding the

responsiveness of the most marginal services – i.e. those most likely to

discipline the airport‟s conduct.

2.141 There may also be a number of other services where the switching costs

involved are particularly low, such as switching of services between airports in

the southeast of England, where the major airlines already have a significant

presence. In such circumstances, switching capacity between airports is

more likely to result in passengers also switching with the airline, and would

tend to involve lower marketing costs than switching that involves launching a

service that is new to a region.

2.142 This suggests that there is a relative hierarchy in terms of the services that

are most likely to be marginal, and the most responsive to price changes.

Services that can be relocated to alternative London airports are likely to incur

particularly low switching costs, whilst moving inbound services from Stansted

to other airports where the airlines have an existing presence are also likely to

have relatively moderate switching costs. Beyond this, relocating aircraft

based at Stansted to another base is likely to result in higher switching costs,

albeit that the larger airlines at Stansted emphasise that their business

models are designed to be flexible.

2.143 It appears that, given the business model of airlines at Stansted, and the likely

magnitude of switching costs (which are discussed in more detail in chapter 3)

that the airports in the Southeast of England are likely to be very close

substitutes to Stansted (although Heathrow is likely to fall outside of this

Moving aircraft between bases is more likely to result in costs relating to relocation and/or redundancy

of staff, and/or require pilots and cabin crew to be recruited at the alternative base.

Civil Aviation Authority

market, due to its particular operational characteristics).

Furthermore, it

appears that the switching of marginal services is likely to take place over a

broader area. In particular, the number and distribution of existing bases of

Stansted‟s two largest customers – and the large number of European

airports to which they currently operate – means that the geographic market

relating to the airlines at Stansted is likely to be European wide.

2.144 However, there are factors that might limit this switching in practice. One of

these is the potential for the yields that can be earned at Stansted to act as a

barrier to switching, as services moved to other airports may earn a lower

yield. This could reflect the particular value that access to the southeast of

England (and Greater London, in particular) confers on airlines at Stansted.

The CAA has not been able to analyse this issue further at this stage.

2.145 On the basis of the evidence currently available, whilst there is a significant

degree of uncertainty associated with the magnitude of switching costs

relating to the impacts on yields, it seems likely that the airlines at Stansted

would be able to switch aircraft relatively easily between bases in response to

a 5-10 per cent increase in prices above the competitive price level at

Stansted. By switching to existing bases, the risks and costs involved are

likely to be smaller.

2.146 The issue of switching costs is discussed in more detail in chapter 3, where

we also consider the incentives on airlines to switch away from the airport.

Cargo carriers

2.147 We have seen mixed evidence on the likely geographic market for Stansted

from the perspective of cargo carriers:

 Robin Hood Doncaster referred to the importance of truck driver shift

patterns in defining the geographic areas over which an airport

competes for cargo carriers. On this basis, they argued that the market

was largely UK-wide for most cargo.

 East Midlands airport also refers to trucking, mentioning the relevance

of a four hour truck drive and proximity to bellyhold operations at

Heathrow, Biirmingham and Manchester as important factors in why

cargo carriers should choose EMA.

 Manston airport refers to itself as an excellent platform for access to

Europe.

2.148 The CAA‟s assessment of the case for de-designating Stansted also referred

to confidential information that highlighted the importance of intercontinental

cargo flows at Stansted, with a significant proportion of Stansted cargo

travelling to Heathrow and then onward to destinations outside of the

UK/Europe.

2.149 There is currently limited evidence available to the CAA on the extent of the

geographic market for cargo carriers at Stansted. However, the evidence that

This is discussed in the context of the product market.

http://www.eastmidlandsairport.com/emacargo.nsf/Content/WhyEMA

http://www.manstonairport.com/for-business/aviation/freight.html

Civil Aviation Authority

is available appears to support the CAA‟s previous view that the geographic

market is likely to be at least national, with aspects that are competing in an

intercontinental market.

2.150 The time sensitive nature of some cargo might suggest a sub-market for this

cargo. We have not received any evidence to date on this point, but invite

views from Stakeholders.

Supply-side substitutability

2.151 Supply-side substitutability is not relevant to geographic markets in the

context of airports since it is not possible for an airport in one geographic

market to begin supplying passengers in another geographic market without

relocating, opening a new airport or improving surface access. None of which

would be possible in a short time frame.

The interdependence of demand from different user groups

2.152 Each of Stansted‟s user groups could be considered as different „sides‟ of an

airport market; each with its own characteristics and ability to respond to

changes in the price and service offered by Stansted.

2.153 Indeed, reflecting the linkages between them, it is likely that the reaction of

passengers, cargo carriers or airlines to changing prices and/or service

quality will precipitate a reaction from the other (interdependent) users. These

reactions may then act together to affect the profitability of such a change in

price for the airport and, ultimately, determine the extent of the market power

enjoyed by the airport.

2.154 This section presents a summary of the earlier findings and considers the

potential interactions between the passenger-facing and airline-facing sides of

the market and how these might influence the definition of the market within

which Stansted operates.

The supply of infrastructure to passengers

2.155 The relevant product market for assessing Stansted‟s market power in

respect of passengers is likely to cover a range of aeronautical and non-

aeronautical services that are required for the reception, processing and

boarding of passengers, including the provision of car parking and other

surface access to the terminals.

2.156 We saw no reason to distinguish between different geographic markets by

passenger types at Stansted. Passenger survey evidence suggests that most

passengers using Stansted would be willing to travel up to 90 minutes to use

an airport. In combination with catchment overlap evidence, this would

suggest that Stansted‟s catchment covers the South East of England. In

addition, historical usage shows that Stansted draws a significant number of

passengers from the East of England, implying that the geographic market for

Stansted passengers should be broadened to include a number of districts in

the East of England.

Civil Aviation Authority

The supply of infrastructure to passenger airlines

2.157 Overall, in respect of the supply of services by Stansted to airlines, the

relevant product market for this assessment is the provision of infrastructure

and infrastructure services, covering a broad range of aeronautical and non-

aeronautical activities, including the landing and parking of aircraft. Given the

business model of airlines present at Stansted, it is likely that most mid-sized

airports would be able to provide the infrastructure and infrastructure services

to enable these airlines to operate.

2.158 It was noted that there could be a case for adopting a narrower product

market – distinguishing between capacity provided at different times of day –

as operational and commercial factors mean that based aircraft for airlines at

Stansted need to depart early in the morning.

2.159 It appears that, given the business model of airlines at Stansted, and the likely

magnitude of switching costs (which are discussed in more detail in chapter 2)

that the airports in the Southeast of England are likely to be very close

substitutes to Stansted (although Heathrow is likely to fall outside of this

market, due to its particular operational characteristics).

Furthermore, it

appears that the switching of marginal services is likely to take place over a

broader area. In particular, the number and distribution of existing bases of

Stansted‟s two largest customers – and the large number of European

airports to which they currently operate – the geographic market from the

perspective of airlines is likely to be European wide.

2.160 There is, however, a significant degree of uncertainty associated with the

magnitude of switching costs relating to the impacts on yields.

Combining the passenger and airline-facing markets

2.161 This chapter identified a number of areas where airlines and passengers

interact making it more difficult to assess the impact of changes in price. In

particular, in order for passengers to switch to an alternative airport, an airline

must offer the desired route to the passenger. Similarly, in order for an airline

to switch to another airport, there must be sufficient demand from passengers

at that airport for any routes that they offer.

2.162 In particular, whilst there may be relatively low direct costs associated with an

airline relocating to another airport, the willingness to do so will be affected by

whether there are sufficient passengers at alternative airports and whether

airline switching away from Stansted typically involves airlines accepting a

lower yield. To the extent that airlines are able to switch to nearby airports

and attract many of the same passengers, this may reduce the adverse

impact on yields. Similarly, for passengers to view an airport as a reasonable

substitute they will need to be able to find a suitable alternative flight, which

will often need to be to the same destination. The airlines‟ ability to switch

will, therefore, depend upon passenger decisions, whilst passengers‟ choices

are likely to be affected by those of airlines.

This is discussed in the context of the product market.

Civil Aviation Authority

2.163 This interdependence can be particularly important where airlines operate

networks, with passengers connecting between services, as the addition of

services can attract more passengers, which itself encourages airlines to

operate additional services. It is also possible that similar virtuous circles

exist at smaller airports, with relatively low levels of passenger and/or airline

awareness, as new routes can attract new passengers – who are then more

familiar with the airport and likely to use it again – which then can be used in

marketing to airlines.

2.164 Furthermore, airports are unable to identify passenger groups and would

therefore find it difficult to price discriminate between them, but where airlines

would be able to identify these passenger groups as they collect more

detailed information from passengers. Where different passenger groups

have varying needs, it may therefore be possible for airlines to discriminate

between them.

2.165 These varying needs are likely to make some passenger groups more or less

price sensitive than others. This implies that, when an airport raises its

charges to an airline, the airline may pass that increase on disproportionately

to different passenger groups. The ability to price discriminate in this way may

mean that passenger volume losses as a result of price increases are less

than if the increase was passed on uniformly.

2.166 As Stansted does not currently serve network airlines and is a relatively large

and well-established airport – indeed, survey evidence confirms a high

awareness of the airport amongst passengers – we have not identified any

significant airline-passenger interactions that significantly affect the market

power enjoyed by the airport. However, the retail revenues generated from

passengers are likely to have a significant impact on the airport‟s incentives to

raise prices to airlines, due to the adverse impact that lower passenger

numbers have on profitability. The potential for passenger and airline growth

to support each other – and create a form of virtuous circle – could also be

relevant to the consideration of the potential for growth in long-haul services

at the airport.

Civil Aviation Authority

3. Factors contributing to Stansted’s market position

Introduction

3.1 This chapter considers the factors contributing to Stansted‟s market position,

which is then used to support the CAA‟s overall assessment of the degree of

market power enjoyed by the airport.

3.2 The previous chapter considered the markets in which Stansted operates,

with regard to its various user groups, with a particular focus on passenger

airlines, passengers, and cargo carriers. This analysis highlights the

importance of understanding how passenger and airline choices affect the

market power enjoyed by the airport, including how these two customer

groups might interact to discipline (or not) the airport‟s conduct.

3.3 This chapter considers the evidence available on the strength of competitive

constraints faced by Stansted. In line with the CAA‟s Guidance on the

assessment of airport market power

, we will be considering the cumulative

impact of competitive constraints arising from within and outside of the

markets set out in the previous chapter.

Structure of this chapter

3.4 This chapter is structured in a number of sections, starting with a description

of the size of Stansted in its markets. This is then followed by three sections

focusing on the potential impact of customer switching on the airport‟s

incentives; looking at airline switching, passenger switching and how these

two forms of switching interact.

3.5 This is then followed by sections that discuss buyer power, entry and

expansion, the impact of capacity constraints, and that consider the relevance

of evidence on the airport‟s pricing and behaviour. There is also a

consideration of the impact of cargo operations on the airport‟s market

position.

3.6 The final section provides a summary and considers how circumstances

might change over time, and affect the airport‟s market power going forward.

Market shares

3.7 Depending on how the market for Stansted is defined, its market shares will

vary. As shown in Chapter 2, the product and geographic markets for

Stansted vary depending on the user group being considered. This analysis

focuses on the market shares of Stansted compared to the London-area

airports. This reflects the availability of data, notably the survey data (which is

only available for a subset of UK airports). The results set out below will,

therefore, tend to overstate Stansted‟s share of the geographic markets set

out in Chapter 2.

From the passenger perspective

3.8 On the evidence currently available, the market definition for Stansted from

the passenger perspective appears to be airports in the south east and east

CAA, 2011, op. cit.

Civil Aviation Authority

of England offering flights to appropriate destinations. We concluded that

there was no compelling evidence to define the product and geographic

markets further by passenger type.

3.9 Figure 13 below shows the London area airport shares of passengers flying

within Europe. Given that Stansted currently only offers flights to short haul

destinations, these appear to be the most appropriate market shares to

consider. They show that Stansted only has a 21 per cent share of this

market, compared with Heathrow with 37 per cent and Gatwick with 29 per

cent, which is well below the level at which there would be a rebuttable

presumption of dominance.

3.10 However, given that Heathrow and Stansted operate in the same market for

passengers travelling within Europe and they are currently both owned by

BAA, it is the joint market share that is most relevant when considering the

degree of market power held by Stansted, which in this case would be

equivalent to that held by the BAA. The joint market share is 58 per cent,

significantly higher than the level that would lead to a rebuttable presumption

of dominance.

Figure 13 London area airport shares of passengers flying within Europe

Source: CAA airport statistics

3.11 Although it was concluded that at this time, there is no compelling evidence to

suggest further dividing the market by passenger type, we have still

considered a passenger split by journey purpose, in order to assess the

whether this conclusion has any bearing on our assessment of market power.

3.12 Stansted‟s highest share on its own is still only 21 per cent of London

passengers from the UK visiting friends and relatives (VFR). As shown in

Figure 14, these proportions are lower when viewed on a UK-wide basis.

37%

29%

21%

10%

LHR

LGW

STN

LTN

LCY

Civil Aviation Authority

Figure 14 Airport share of passengers by journey purpose and residence 2010

Source: CAA Passenger Survey

3.13 Again, looking at BAA‟s share of this market, the airports account for 22 per

cent of UK holiday traffic (the lowest combined share across the passenger

segments), and 39 per cent of both UK business and UK VFR traffic. By

virtue of Heathrow‟s share of foreign passengers, the combined shares of

foreign business, foreign leisure and foreign VFR are all very high, at 67, 68

and 61 per cent, respectively. All of these shares are significantly higher than

the levels that would support a rebuttable assumption of dominance.

From the airline perspective

3.14 In Chapter 2, we concluded that on the evidence currently available, the

market definition for Stansted from the airline perspective appears to be mid-

sized and larger airports in Europe with a catchment of passengers to

generate a sufficient yield.

3.15 We do not have Stansted‟s share of passengers flying from these airports in

Europe, but it is clearly very low. One way to look at the importance of

Stansted to airlines is to look at the share of an airlines business accounted

for by Stansted. For Stansted‟s biggest customers, Ryanair and EasyJet,

these shares are 17 and 8 per cent respectively, suggesting that Stansted

does not have a large share of airlines‟ business across Europe.

3.16 As discussed in chapter 2, it may be appropriate to consider the provision of

infrastructure and infrastructure services at Stansted in the morning peak as a

separate economic market. Table 3 below gives the shares of passengers

travelling from the five main London area airports at the morning peak. It

shows that by this measure, Stansted‟s market share is higher than its share

of the all day market, although it is still only the third largest player. When we

exclude Heathrow, which we do not consider to be a competitor to Stansted

for airlines, Stansted‟s share rises to 37 per cent. However, we consider

Stansted to be part of a wider European market. We do not have the data to

calculate the shares on a European-wide basis, but Stansted‟s shares would

be likely to drop significantly on this basis.

Civil Aviation Authority

Table 3 London Area airports’ shares of departing passengers at the morning

peak and throughout the day

Share of Departing Pax, 2010

6-8 am local time

All day

Heathrow

30%

52%

Gatwick

30%

25%

Stansted

26%

15%

Luton

12%

London City

Total

100%

Source: CAA Airport Stats

Summary

3.17 On most measures, if we consider Stansted as a stand-alone airport, its

market shares, however the market is defined, would not tend to support the

view that Stansted has substantial market power and would certainly not

support a rebuttable presumption of dominance.

3.18 However, while whilst under joint ownership, the shares of Heathrow and

Stansted should be combined for any sector in which they jointly operate.

BAA‟s share of this sector is much larger than Stansted‟s individual share and

above the level that would support a rebuttable presumption of dominance.

The implications of joint ownership are discussed further below.

3.19 It should also be noted that, as set out in the CAA‟s Guidance on the

Assessment of Airport Market Power, there are aspects of airport markets that

might reduce the reliability of market shares as an indicator of market power.

A number of other factors will affect the assessment of market power, in

particular how easily airlines and passengers can switch between airports and

how easy it is for competing airports to enter or expand. These factors are

discussed further below.

Airline switching

3.20 The annex on airline switching gives details on what affects an airline‟s ability

to switch, including a discussion of switching costs and available alternatives.

It also gives examples of actual switching to and from the London airports.

Ability of airlines to switch away from Stansted

3.21 As discussed above, switching costs will vary by airline business model, type

(or extent) of switching and on whether the switch involves moving based

aircraft.

3.22 The predominant airline business model at Stansted is LCCs, operating short-

haul services. In general, these airlines will have invested less at the airport

than other airline business models and have multiple bases across the UK

and Europe.

In addition, this type of airline business model does not rely on

While we exclude Heathrow as a strong competitor for Stansted‟s current airlines, the two airports do

compete with each other for passengers.

As discussed in paragraph 2.136, Ryanair currently operates from 160 airports, with 47 bases; and

easyJet currently operates from 123 airports, with 19 bases.

Civil Aviation Authority

transfer traffic, which means that these airlines do not face the additional

switching costs associated with the impact of airline network economics.

These factors imply that airlines at Stansted will generally find it easier to

switch away than network airlines.

3.23 The table below summarises the costs involved in switching operations to a

new base for LCCs and for FSCs operating network and long-haul services.

This type of switch would generally incur the highest costs. In contrast, a

switch to an airport where the airline already has a base would incur lower

costs, with the lowest cost switch being to increase the frequency of an

existing route at an existing base. Given that LCCs tend to have bases

across Europe, they have more options as to where to switch; again lowering

switching costs compared to network airlines.

3.24 Illustrating the point further, the Competition Commission refers to evidence

provided by BA and Ryanair.

BA stated that it had sunk very large

investments at Heathrow: to recreate BA‟s maintenance facilities elsewhere

would cost about £1 billion. In contrast, Ryanair said that the capital cost of a

move from Stansted to Gatwick would be approximately £10–£12 million.

3.25 In addition, as discussed in paragraph 2.65, due to the more streamlined cost

structure of LCCs, airport charges generally account for a larger proportion of

the total costs than they do for FSCs. This implies that there will be a greater

incentive for LCCs to switch in response to an increase in airport charges.

Figure 15 Types of costs involved in switching operations (if no existing

operations at the new airport)

Sources: Airline interviews and confidential submissions, Competition Commission BAA airports market

investigation March 2009, Appendix 3.4 Costs to airlines of switching airports

3.26 Even if switching costs are low enough to allow airlines to switch, they must

have appropriate alternative airports to switch to. A based carrier at

Stansted‟s ability to switch to neighbouring airports may be limited by

These network effects, which affect airlines that transport significant numbers of connecting

passengers, arise from the adverse impact that cancelling one service has on the profitability of other

services across the network.

„BAA airports market investigation‟, Competition Commission, March 2009 in Appendix 3.4: Costs to

airlines of switching airports.

Civil Aviation Authority

constrained capacity at other London airports. However, Stansted has more

near-neighbouring airports with capacity than Gatwick. In addition, airlines

with a network of bases across Europe also have the option of switching to

other, non-London airports; although switching costs may be higher for such a

switch.

3.27 EasyJet argues, however, that the switching costs it faces are still too high to

enable them to switch in order to discipline effectively a 10 per cent price

increase. It commissioned analysis from Frontier Economics, which we

discuss further below. One way to test this is to look at actual switching, which

we do in the next section.

Evidence of switching

3.28 While airlines at Stansted are likely to face lower switching costs than airlines

with different business models, this does not in itself indicate whether these

costs are sufficiently small to allow airlines to switch services in sufficient

numbers to discipline a 10 per cent increase in prices. One source of

evidence that can be used to understand likely switching behaviour is if actual

switching has been observed.

3.29 In this respect, Stansted has lost several major airlines to Gatwick recently

(Norwegian, Air Berlin and Air Asia X). It is notable that this is despite their

being some capacity constraints at Gatwick. The combined impact of the Air

Berlin and Air Asia X switches is approximately 3.4 per cent of Stansted

passengers, with the Norwegian moves taking place somewhat earlier and

accounting for approximately 1.6 per cent of passengers at Stansted.

addition, easyJet has recently announced its intention to move some of its

based aircraft from Stansted to Southend.

3.30 []

3.31 These recent switches imply that at current prices, switching is possible and

that airlines do switch.

Analysis of route churn

3.32 A further source of evidence about the magnitude of switching costs is the

extent to which airlines vary their schedules, change the destination of their

aircraft, and vary the number of aircraft based at an airport.

3.33 Figure 16 below shows that Ryanair has increased and decreased routes and

frequencies significantly over the past 10 years. In contrast, easyJet has

made relatively fewer changes in its schedule in recent years, apart from a

large decrease in 2009 as shown in Figure 17.

These calculations are based on comparing traffic volumes in 2010 by Air Berlin and Air Asia X to

total passenger numbers at Stansted. Norwegian did not operate throughout the whole of 2010, and

only operated a limited operation in 2009, so we have compared 2008 figures, when the carrier operated

multiple routes from the airport. It is worth noting that Norwegian has grown at Gatwick, so the

magnitude to the traffic lost from Stansted is arguably greater than that used to calculate this estimate.

easyJet estimates that Southend will attract around 800,000 passengers in its first year (although not

all of these would have travelled via Stansted): http://corporate.easyjet.com/media/latest-news/news-

year-2011/16-06-2011.aspx

Civil Aviation Authority

Figure 16 Routes started/dropped and frequency increases and decreases as a

proportion of total number of routes for Ryanair at Stansted and the total

number of routes (right axis)

Source: CAA Airport Statistics

Figure 17 Routes started/dropped and frequency increases and decreases as a

proportion of total number of routes for easyJet at Stansted and the total

number of routes (right axis).

Source: CAA Airport Stats

3.34 This analysis shows that there is considerable variation in the routes flown

from Stansted, albeit within a longer term trend of growth (Ryanair) and

relative stability overall volumes (easyJet). In particular, it demonstrates that

there is more volatility than for some other carriers – indeed, Figure 18

reproduces similar analysis for BA at Heathrow, which shows much less

volatility in routes, with most of the changes taking the form of variations in

frequency.

-40

-30

-20

-10

100

110

120

130

-30%

-20%

-10%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Freq Decr

Freq Inc

Route Drop

Route Start

Routes

-30

-20

-10

100

-30%

-20%

-10%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Freq Decr

Freq Inc

Route Drop

Route Start

Routes

Civil Aviation Authority

Figure 18 Routes started/dropped and frequency increases and decreases as a

proportion of total number of routes for BA at LHR and the total number of

routes (right axis).

Source: CAA airport statistics

3.35 This analysis highlights that the major carriers at Stansted operate particularly

dynamic networks, with the routes flown varying to a significant degree over

time. It also supports the view that switching costs between airports that are

not bases seem unlikely to represent a significant barrier to airline switching.

This analysis does not, however, establish whether the major airlines at

Stansted face significant costs when switching aircraft away from Stansted,

as the overall number of aircraft based at Stansted has not varied markedly

over time. The examples of switching highlighted above provide some

evidence on this form of switching.

Passenger switching

Ability of passengers to switch away from Stansted

3.36 As discussed above in chapter 2, the main cost to passengers switching

between airports is the difference in travel times and cost of travel to the

different airports. So long as a suitable destination is served by an alternative

airport, a passenger is likely to be able to switch there at little cost other than

those relating to surface travel.

3.37 As stated above, survey evidence suggests that in choosing an airport,

location is the most important factor, followed by the availability of the route

and the cost of the flight. This implies that, whilst passengers are likely to

prefer to travel from their nearest airport, price and route availability are also

important factors. Consequently, there is the potential for passengers to be

reasonably price sensitive, and in particular where the same flight is available

at another local airport.

3.38 It is useful, therefore, to consider evidence from passengers‟ use of the airport

to understand the extent to which passengers are likely to switch between

Stansted and other airports. This is considered in the next section.

-40

-30

-20

-10

100

110

120

130

-30%

-20%

-10%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Freq Decr

Freq Inc

Route Drop

Route Start

Routes

Civil Aviation Authority

Evidence of switching/ potential to switch

3.39 Given that the greatest cost to passengers in switching airport is likely to be

the difference in travel time and cost, in order to assess the likelihood and

potential for passengers to switch, it is useful to consider the catchment area

of Stansted and which airports also draw passengers from within that

catchment area. It is also useful to consider survey evidence on which

airports Stansted passengers have used in the past, the airports they would

consider using and their likely reaction to an increase in fares.

3.40 This section considers catchment area analysis, survey evidence and

modelling undertaken by Frontier, using the DfT‟s passenger allocation

model.

Catchment area analysis

3.41 The catchment area analysis paper considers the catchment areas of

Heathrow, Gatwick, Stansted and Luton under different measures. The

results of Stansted‟s catchment area analysis are included in the section

above on geographic market definition. We will now consider further how the

catchment areas of Stansted‟s nearest airports overlap with each other and

therefore the likely alternative airports for Stansted‟s passengers.

3.42 As highlighted above, most passengers travel up to 90 minutes to get to

Stansted, suggesting that Stansted‟s catchment should include all areas

within 90 minutes drive of the airport. On this measure there is a significant

overlap between Stansted‟s catchment and that of Heathrow, Gatwick, Luton

and London City airports. Indeed, all these airports lie within a 90 minute

travel time of Stansted airport, suggesting that they are alternatives for a

significant proportion of Stansted‟s passengers.

3.43 Furthermore, by looking at how the various London area airports‟ catchments

overlap, it is possible to assess the alternatives available for passengers

located in different areas.

Some passengers, particularly those to the north-east of Stansted would face a longer travel time to

these other airports, as they would need to travel past Stansted to reach them. Other passengers, such

as those in Greater London, are more likely to face a lower increase in overall travel time should they

switch between Stansted and the other London airports.

Civil Aviation Authority

Figure 19 Overlap of districts within 90 minutes

Source: CAA analysis of DfT surface access data

Blue: 1 airport; Light blue: 2 airports; Light red: 3 airports; Red: 4 airports

3.44 Figure 19 shows all the districts in the south east within a 90 minute travel

time of Stansted, Gatwick, Heathrow or Luton or some combination of these.

It shows that passengers travelling from greater London have the greatest

choice (i.e. of all four airports) within 90 minutes travel time.

3.45 These passengers with the greatest choice are likely to be the most marginal

passengers: given that it is easy for them to travel to all four main London

airports, they are likely to be more sensitive to any increase in price.

3.46 As shown in our catchment area analysis paper, this area of greatest overlaps

is also the most densely populated area, suggesting that a significant

proportion of passengers are likely to be marginal. However, the ability of

these passengers to switch between these airports is still going to be

restricted by the availability of routes. We consider this aspect further below.

3.47 An alternative way to assess the strength of overlaps is to consider the

historical usage pattern of each airport, rather than relying on surface travel

times from each district. Looking at the overlap of catchments using historical

usage, as shown in Figure 20 below, shows that the area where all four

catchments overlap shrinks considerably. This is likely to reflect the fact that

passengers generally have a preference for their closest airport, even when

alternative airports are available nearby.

The figure also shows that the

number of districts served by the airports is wider than a 90 minute travel time

suggests, although for these further away districts, passengers appear to

have fewer choices.

These passengers would also have relatively easy access to London City Airport.

Indeed, a lack of overlaps in historical usage should not be interpreted as necessarily indicating a lack

of choice and competition. Competing airports with similar offers might be expected to produce a

pattern of usage where passengers use their closest airports and there are relatively low overlaps

between catchment areas based on historical usage.

Civil Aviation Authority

3.48 Nevertheless, the area with the most overlap is still Greater London, where

passengers from most districts have travelled to at least three airports.

Figure 20 Overlaps of historical actual usage catchment areas using 80%

passengers

Source: CAA analysis of the CAA Passenger Survey (2010)

Blue: 1 airport; Light blue: 2 airports; Light red: 3 airports; Red: 4 airports

Survey evidence

3.49 Another way to assess the ability of and likelihood that passengers switch

away from Stansted is to ask them directly. Our working paper on the

passenger survey results gives a detailed analysis of the survey data.

consider the most relevant analysis for Stansted below.

3.50 Figure 21 below shows the airports previously used and considered by

passengers at Stansted in the last two years. It shows that around half of

passengers at Stansted had previously used Heathrow or Gatwick, with

approximately one quarter considering these airports (although Heathrow was

considered less than Gatwick). Around a quarter of passengers at Stansted

had used Luton previously, with a similar number considering it for their

current journey, making it the second most considered airport after Gatwick.

3.51 These results suggest that there are significant numbers of Stansted

passenger who have previously used another London airport (indicating a

degree of willingness to travel to use these airports) and that stated that they

had considered alternative airports to Stansted. This supports the view that

these three airports should be considered as close competitors to Stansted in

attracting passengers.

„Passengers‟ airport preferences: results from the CAA Passenger Survey‟, CAA, November 2011

Civil Aviation Authority

Figure 21 Airports previously used in the past two years and airports

considered as alternatives for current flight by short haul passengers at

Stansted

Source: CAA analysis of CAA Passenger Survey data – supplementary stated preference question

3.52 The passenger survey also asked passengers what their response to an

increase in price (£10 for short haul) would be. Figure 22 below shows that

over 15 per cent of passengers in each passenger group, except UK business

passengers, would switch away following an increase in price.

It should be

noted that these measures of responsiveness are based on changes in

airfares, and an increase in airport charges of 5 to 10 per cent will not have as

large an impact on airfares. This is discussed in more detail below.

Figure 22 Short haul passenger price responsiveness to an increase in the cost

of using their current airport, by journey purpose and residence

Source: CAA analysis of CAA Passenger Survey data – supplementary stated preference question

The survey asked passengers whether they would respond to changes in airfares. This is not

equivalent to saying that they would respond to a change in airport charges to the same degree, as this

will depend upon the link between airport charges and airfares. The analysis is, therefore, more useful

to consider the relative price sensitivity of passengers rather than it being a direct indicator of

responsiveness to airport charges – which would require further analysis to establish.

Civil Aviation Authority

3.53 The results of this survey highlight the differing degrees of price

responsiveness at the three airports, with Heathrow showing less

responsiveness that Gatwick and Stansted.

3.54 In general, there is no consistent pattern of responsiveness between Gatwick

and Stansted. Foreign Holiday and UK Business passengers appear more

responsive at Stansted than at Gatwick, whilst VFR and UK Holiday

passengers appear less responsive at Stansted. These differences between

Stansted and Gatwick may not, however, be significant.

Modelling of passengers‟ willingness to switch airports

3.55 In its report for easyJet, Frontier estimates the number of passengers that

would leave Stansted in response to a 10 per cent rise in the airport‟s

aeronautical charges.

3.56 Frontier used the DfT‟s passenger demand forecasting model NAPALM

. It

estimated that about 0.69 million passengers (per annum) would want to

switch from Stansted to other airports, in response to a 10 per cent increase

in charges, provided there were no capacity constraints at any of the other

airports. These passengers would switch to Gatwick (0.5mppa), Heathrow

(0.15mppa), Luton (0.11mppa) and other airports.

3.57 In an alternative scenario Frontier assumed that passengers could not switch

to either Heathrow (due to capacity constraints) or London City (due to the

limited number of available routes and constraints on aircraft size), which

reduced the number of passengers willing to switch away from Stansted to

0.61 million (per annum).

3.58 Using the DfT‟s NAPALM model in this way, despite being primarily designed

to estimate long-run passenger demand forecasts rather than short-run

reactions to price increases at one specific airport, is a useful contribution to

attempting to assess the impact on passenger numbers at Stansted resulting

from a unilateral price increase.

3.59 Nevertheless, its methodology and underlying assumptions involve a number

of features

that suggest that the estimated number of passengers is likely to

be a the lower end of the spectrum:

 The calculation assumes a full pass-through of the airport charge

increase to passengers. However, airlines have previously told us that

in the short-run such price increases would be absorbed fully by the

airlines and only incorporated over the following scheduling seasons.

 The model holds total passenger demand fixed, meaning it assumes

that no passengers would leave the market but always switch to an

alternative airport. It might be that a number of passengers, instead of

going to a different airport, might abandon their flight altogether.

The National Air Passenger Allocation Model is a fully-estimated multinomial choice model that is

used to convert unconstrained forecasts of air passenger demand into forecasts of passenger demand

by airport.

While these may be appropriate assumptions for a long-run forecasting model, they are likely to

understate the level of short-run passenger switching away from a single airport in response to a price

increase.

Civil Aviation Authority

 The model treats passengers of FSC and LCC separately; for example,

LCC passengers faced with a price increase at Stansted cannot

substitute to FSC services at Heathrow.

 The model does not seem to allow for route substitution; for example, a

passenger flying to Charles de Gaulle airport can only switch to another

flight to Charles de Gaulle.

 An initial analysis of the model suggests that regional UK airports

would generally lose a lower percentage of their passengers in

response to a similar price rise than would any of the London airports.

Given that regional airports are generally regarded as competitive, we

need to understand better whether a high or low share of estimated

passenger switching can indeed be interpreted as suggesting a high or

low degree of market power, absent any further explanations.

 Furthermore, we do not think it is appropriate to prevent passengers

from switching to Heathrow. While Heathrow‟s runway capacity might

be constrained, there is still capacity to accommodate additional

passengers (through larger aircraft and higher load factors).

3.60 Overall, the modelled responsiveness of passengers to price may appear

relatively high, considering that a 10 per cent rise in the airport‟s revenues

would only constitute a fraction of a passenger‟s total travel costs. However,

a large number of Stansted‟s passengers originate from the Greater London

area, which is the most contested geographical area, and a high proportion of

Stansted‟s passengers travel on routes that are also offered from at least one

other London airport. It might therefore be plausible that passengers could

react even to small differences in price, as the model suggests.

Route availability

3.61 Some passengers will tend to have a fixed destination, so would only consider

airports where flights to their chosen destinations were available.

The

degree to which passengers are willing to switch between destinations will

tend to vary with their journey purpose, with those travelling for business or

for VFR being less likely to switch than those travelling on holiday. For those

less willing to switch destination, the availability of services at different

airports can affect the degree to which passengers will switch between

airports in response to changes in price or service.

3.62 For some services, their destination might be served equally well by a number

of destination airports, some of which might be captured by being considered

as serving the same city (e.g. Rome Fiumicino and Ciampino would both be

captured by a city pair London-Rome), but others might not be captured (e.g.

Valencia, Alicante and Murcia could be considered as reasonable substitutes

to go to the south east coast of Spain). However, assessing overlaps of city

pairs between London airports provides some indication of the level of choice

passengers (with strong destination preferences) currently have between

airports. It should be noted, however, that this is only a snapshot, as airlines

Some passengers might be able to switch between different destinations, for example for another

nearby airport close to their final destination or a different holiday destination all together.

Civil Aviation Authority

regularly change their route network, so that the degree of overlap might

change in both directions over time.

3.63 Figure 23 shows the level of city pair overlaps for short haul flights in the

London area. It shows that 60 per cent of routes available at Stansted are

also available at other London airports. Stansted has the lowest proportion of

overlaps, reflecting the fact that it also serves the highest number of short

haul destinations (provided predominantly by Ryanair), including to a number

of relatively small destination airports.

Figure 23 Number of short-haul and domestic route overlaps between

LON airports, 2010

Source: CAA airport Stats

3.64 However, 60 per cent is still a significant proportion and if we weight this by

passenger numbers on these routes, we see that 74 per cent of passengers

could fly to the same destination from another London airport.

3.65 As the strength of competition is determined by the marginal passengers – i.e.

those most able to switch between airports – this analysis supports the view

that there is considerable scope for passengers at Stansted to switch

between airports, even if passengers have very strong destination

preferences.

Implications of passenger and airline switching and the multisided nature of

the market

3.66 The previous two sections looked at airlines‟ and passengers‟ ability to switch

and evidence of actual switching. Given that airports provide services to, and

earn revenues from, both airlines and passengers, and that airlines also

provide services and earn revenues from passengers, it is important to

consider these interrelationships in assessing an airport‟s ability to exert

market power.

3.67 If an airport raises its prices to airlines, airlines are likely to raise their prices

to passengers (certainly in the long term, if not immediately). In examining

whether or not an airport can profitably raise its prices, it is therefore

important to assess the effect of the rise in price on airlines and passengers

combined.

3.68 Given that airlines at Stansted largely offer short-haul point-to-point flights and

that it is generally accepted that the market for short-haul flights is

competitive, we would expect that prices to passengers are competitive. In

competitive markets, prices will over time reflect costs of supply – in

particular, the forward-looking (or marginal) costs of supply. As such any

increase in airport charges (a marginal cost) is likely to be fed through to

passengers. Any unanticipated changes in airport charges might reduce the

Cities Served Overlaps % overlap LHR LGW STN LTN LCY

LHR 64 55 86% 47 33 28 16

LGW 121 92 76% 65 48 21

STN 131 79 60% 42 16

LTN 69 59 86% 13

LCY 26 24 92%

Civil Aviation Authority

speed with which airfares adjust, as airlines will have announced their

schedules and commenced sales of tickets. However, to the extent that a

price rise is anticipated, there might be a reasonable expectation that these

costs would be incorporated into fleet planning and pricing decisions.

3.69 In this context, it is relevant that the Airport Charges Regulations impose

obligations on larger airports, including Stansted, that require regular

consultation and transparency on charging. These obligations mean that

(absent price control regulation) the airport is required to announce its

charges in advance and thus provide airlines with an opportunity to vary their

fleet planning and pricing before the higher charges apply.

Overall,

therefore, an increase in Stansted‟s charges to airlines is likely to lead to an

increase in ticket prices to passengers.

3.70 As shown above, many passengers flying from Stansted have the choice of a

number of different airports to fly from, with a large proportion (i.e. those

within 90 minutes travel time of all four main London airports) having limited

switching costs. The pool of marginal customers is therefore potentially fairly

large. An increase in prices is therefore likely to induce at least some

passengers to switch to another airport. Indeed, modelling undertaken by

Frontier, using DfT data, indicated that there would be significant passenger

switching between airports in response to changes in airfares (this is

discussed in more detail below).

3.71 For airlines operating marginal routes, this loss of passengers may render

some routes unprofitable. The airline may therefore cease to operate that

route and switch an aircraft to another airport. If no other airline is serving

that route at Stansted, the airport is likely to lose the majority of the

passengers from that route.

3.72 In this respect, it is relevant that in 2010 only around 10 per cent of Stansted‟s

passengers were on regularly-served routes served by more than one airline

at Stansted.

These overlaps were between Ryanair and easyJet for five

popular Spanish and Portuguese holiday destinations; and between Air Berlin

and Germanwings for Hannover and between Pegasus and Turkish Airlines

for Istanbul. This implies most routes at Stansted are only run by one airline

and therefore, if a route was cancelled, Stansted would be more likely to lose

the passengers from that route to another airport. This is most likely to be the

case for VFR and business passengers.

3.73 This interplay between passengers and airlines indicates that the loss of

business to an airport resulting from an increase in price may be greater than

indicated by the impact on „marginal‟ passengers: it may also include

additional passengers on any routes that become unprofitable, who would

then need to travel to a different airport to get to their destination.

For example, the Airport Charges Regulations 2011 requires airports meeting the size threshold to

provide 4 months notice of any intention to vary the level of airport charges.

Routes with at least 100 departures in the year.

Civil Aviation Authority

Airport sensitivity to passenger and airline switching

3.74 As discussed above, to assess Stansted‟s ability to exercise market power,

we consider how likely it is for Stansted to be able to profitably raise prices by

5-10 per cent. The sections above consider the ability of airlines and

passengers to react to a price increase at Stansted by switching away, which

allows us to assess the likely proportions of passengers that may switch in

response to a price rise.

3.75 To understand the airport‟s market power, we need to combine this evidence

about the magnitude of potential switching with an understanding of the

airport‟s sensitivity to reductions in passenger and airline volumes, as it is this

sensitivity that will determine the airport‟s incentives to compete. To do this,

we can estimate the number of passengers that would need to be lost in order

to make a 5 or 10 per cent increase in price unprofitable. This is known as a

critical loss analysis.

3.76 This can then be compared with the number of passengers that would be

likely to switch in the face of such a price rise. If it shows that a greater

proportion of passengers are likely to switch than the number that would

render such a price increase unprofitable (i.e. the critical loss number), then

this implies that the airport does not have substantial market power.

3.77 The calculation involved to estimate the critical loss of passengers is relatively

straight-forward. However, it involves a number of assumptions to be made

about the airport‟s costs and revenues. It is also difficult to summarise the

analysis on passenger and airline switching into a single number, due the

large number or factors that affect this and the difficulty in quantifying them,

as outlined above.

3.78 We have calculated the volume (passenger) loss at which a price increase of

5 or 10 per cent becomes unprofitable, under a certain set of assumptions.

64,65

A price rise can be applied by the airport in different ways, for example across

aeronautical charges or across a combination of both aeronautical and non-

aeronautical charges. The CAA‟s Guidance on the Assessment of Airport

Market Power considers the implications of the fact that airports generate

revenues from both passengers and airlines. As an increase in charges could

result in higher charges to passengers, airlines or both – and could take the

form of higher charges for essential passenger-facing services – it is

important to consider the critical loss that results from a 5 or 10 per cent

increase in airport revenue per passenger, in addition to that which might

result from an increase only to aeronautical revenues.

3.79 Reflecting this, the critical loss figures in Table 4, for 5 and 10 per cent

increases in price are shown for these two different methods. Additionally,

fluctuations in an airport‟s operating costs may be linked to some degree to

Critical loss analysis is also relevant to understanding the relevant market within which an airport

operates. We have chosen to present the information within the assessment of the market power for

ease of exposition. The conclusions are, of course, relevant to the analysis of market definition.

The critical loss is therefore the volume loss that would make the airport indifferent between raising

prices and keeping them at their existing levels.

More information is included in the Annex to this paper.

Civil Aviation Authority

changes in its passenger numbers. As a result, critical losses are calculated

assuming fixed operating costs, and assuming a staff to passenger elasticity

of 0.3 between operating costs and passengers.

That is, operating costs per

passenger fall 30 per cent for each passenger switching away.

Table 4 Critical loss calculations for Stansted

Source: Stansted 2010 Regulatory Accounts

3.80 The second and third columns of Table 4 show the critical loss calculations for

cases where the airport increases only its aeronautical charges by 5 or 10 per

cent, keeping non-aeronautical revenue unchanged.

The required loss of

passenger volume for a price increase to be unprofitable is 494,327 for a 5

per cent increase and 962,662 for a 10 per cent increase of aeronautical

charges, assuming operating costs per passenger are fixed. These volumes

increase to 601,170 and 1,164,115 respectively when a cost saving for each

lost passenger is assumed.

3.81 However, Stansted could, in principle, also increase some non-aeronautical

charges, particularly those charges that are unavoidable for airlines and/ or

passengers. If Stansted is found to have market power, it is likely to have

market power over all services connected to passengers‟ and airlines‟ use of

the airport and therefore the ability to raise prices above the competitive level

for these services.

This staff-passenger elasticity was adopted by the CAA and the Competition Commission in the Q5

price control review, in order to forecast operating costs.

Another way of describing a 0.3 staff to passenger elasticity is that, for a 10 per cent fall in

passengers, the number of staff would be expected to fall by 3% percent.

This methodology was employed in the Frontier Economics report for easyJet, available on the CAA

website: http://www.caa.co.uk/docs/5/rpt-

easyJet%20Competition%20Assessment%20Final%20Report_Abridged.pdf

Increasing aeronautical revenue

SSNIP increment 5% 10% 5% 10%

Passengers 18308000 18308000 18308000 18308000

Aeronautical Revenue 121,600,000 121,600,000 121,600,000 121,600,000

Non-Aeronautical Revenue 97,500,000 97,500,000 97,500,000 97,500,000

Total Revenue 219,100,000 219,100,000 219,100,000 219,100,000

Operating Costs 133,400,000 133,400,000 133,400,000 133,400,000

Current

Aeronautical Revenue per Passenger 6.641905178 6.641905178 6.641905178 6.641905178

Non-Aeronautical Revenue per

Passenger

5.325540747 5.325540747 5.325540747 5.325540747

Total Revenue per Passenger 11.96744593 11.96744593 11.96744593 11.96744593

Operating Costs per Passenger 7.286432161 7.286432161 7.286432161 7.286432161

After price increase

Aeronautical Revenue per Passenger 6.974000437 7.306095696

Non-Aeronautical Revenue per

Passenger

5.325540747 5.325540747

Total Revenue per Passenger 12.29954118 12.63163644 12.56581822 13.16419052

Increase in Revenue 6,080,000 12,160,000 10,955,000 21,910,000

Critical loss with fixed costs 494,327 962,662 871,810 1,664,364

Critical loss with cost savings 601,170 1,164,115 1,055,406 1,995,762

Increasing total revenue

Civil Aviation Authority

3.82 Under the assumption that the airport raises both aeronautical and non-

aeronautical charges, the required losses in passenger volumes to make an

increase unprofitable are 871,810 and 1,664,364 for a 5 and 10 per cent

increase in total revenue per passenger respectively. These figures increase

to 1,055,406 and 1,995,762 if we assume the airport makes a cost saving for

each lost passenger.

3.83 The range of passengers that would need to be lost in order to render a 5-10

per cent price rise unprofitable is therefore 3-11 per cent of total passengers,

depending on the varying assumptions used.

3.84 As discussed above, survey results showed that over 15 per cent of

Stansted‟s passengers would switch away from Stansted following a £10

increase in airfare. This is equivalent to roughly 10 per cent of the average

short haul airfare. However, the airport charge is just a proportion of the total

airfare. So a 10 per cent increase in airport charges would lead to a lower

percentage increase in airfare.

3.85 Evidence available for easyJet above suggests that airport charges account

for between 28 and 45 per cent of the airfare. So a 10 per cent increase in

airport charges that are fully passed through would imply a 3-5 per cent

increase in airfare. Unfortunately the survey did not ask passengers for their

response to this, smaller increase in airfare.

3.86 Given the larger price difference that was used and the accepted flaws of

asking hypothetical questions in surveys, this analysis and the overall

comparisons cannot be relied upon to give a definitive answer on the airport‟s

market power.

However, it is useful insofar as it indicates that the switching

and critical loss figures are of a similar magnitude, indicating that the impact

of a 5-10 price rise is not clear from the currently available evidence.

Frontier analysis

3.87 EasyJet commissioned Frontier to analyse the potential for airline switching to

constrain prices at Stansted, undertaking similar calculations to those set out

above. Frontier‟s paper argues that the costs for easyJet of switching from

one airport to another exceeds 10 per cent of aeronautical charges, implying it

would have no incentive to switch following a 10 per cent increase in these

charges. It also shows that the level of passenger switching following a 10

per cent increase in aeronautical charges (assuming full pass-through) would

not be sufficient to render such an increase in charges unprofitable.

3.88 The report provided a useful analysis of the impact of an increase in prices on

airport profits.

These figures are approximately double those of an increase in aeronautical revenue only. The

explanation lies in the fact that, due to the approximately 55%-45% distribution of total revenue across

aeronautical and non-aeronautical revenue, a 10% (5%) increase in aeronautical revenue per

passenger results in an overall increase of 5.4% (2.6%) approximately in total revenue per passenger.

As a result, increasing total revenue per passenger by the same increment would have a more

considerable impact.

Indeed, the analysis of critical loss and of switching should take place at the competitive price level,

rather than at the prevailing price level, which could further limit the conclusions that can be drawn from

the analysis.

Civil Aviation Authority

3.89 However, there are a number of factors that affect the way that the results

should be interpreted

 The report‟s analysis of airline switching restricts any switch to be

within the London market. However, there is evidence that the airlines

at Stansted compare yields across a European market; which is

consistent with the CAA‟s view on the likely scope of the geographic

market within which the airport is operating when competing for

airlines.

 The analysis of passenger switching also restricts passengers‟ choices.

In particular, the model does not allow for the possibility of passengers

exiting the market following a price rise (i.e. to decide not to travel), and

instead only allows them to switch or stay where they are. It also

assumes that airlines can only switch to other London airports. These

assumptions are likely to lead to an underestimate of the level of

passenger switching.

 The interaction between passenger switching and route profitability

does not incorporate the impact that lower passenger numbers might

have on route profitability (something which is likely to be particularly

complicated to model) For example, if an increase in airport charges

induces a drop in passenger numbers, and that drop in passenger

numbers renders some marginal routes unprofitable and these routes

are not served by other airlines at the same airport, most of the

passengers on those routes will also be lost. The results may therefore

underestimate the loss of passengers from the airport following a 10%

increase in passenger charges (depending on the number of

passengers that would need to be lost to render a route unprofitable).

3.90 The CAA also has some concerns with the input assumptions used for airline

switching costs. In particular, the level of marketing spend and revenue loss

that easyJet has provided to Frontier is that associated with a new route. This

might not be the most appropriate figure, as an airline might be expected to

switch aircraft to existing bases and potentially increase frequency on existing

routes. These forms of switching would tend to result in lower marketing

costs than in the scenario used to generate the switching cost assumption.

3.91 This implies that the estimates relate to a particular form of switching (i.e. to a

different airport to operate a new service) and might, therefore, over-state that

actual marketing costs involved in the most likely form of switching. This

might suggest that the switching costs are too high and the estimated airline

response is lower than would happen in practice.

3.92 In general, the methodology considers two separate approaches for airlines

and for passengers – considering switching from the passenger perspective,

but considering the impact on incentives from the airline perspective – without

taking into account the interplay of the two user groups. However, we

acknowledge that taking into account these externalities in quantitative

It should be noted that these factors are examples of the types of limitations that are typically

experienced when the complexity of markets are simplified to allow a tractable model to be constructed.

Civil Aviation Authority

analysis is very difficult, and that the CAA has not attempted to undertake an

integrated modelling approach. An example of the impact of this approach is

the approach to considering capacity constraints, where the possibility of

switching to Heathrow is removed, on the basis that there is limited capacity

for airlines to switch to Heathrow. This ignores the potential for passengers to

switch to the airport.

3.93 The impact of Frontier‟s approach is that it is likely to underestimate the

degree of switching and overestimate the cost of switching, making it more

likely that they reach a conclusion that switching would not be sufficient to

defeat a 5-10 per cent prise rise.

3.94 Despite this, the report clearly adds to the overall evidence base and is a

relevant piece of evidence to consider in the round. We will also consider

how we might build on the approach adopted by Frontier, in ways that might

address the points that we make above, such as by investigating whether

there are different input assumptions on the adverse impact of switching on

airline yields.

Buyer power

3.95 In recent years, Stansted has had an increasingly concentrated customer

base. As the figure below illustrates, Ryanair accounts for 68 per cent of its

passengers, while EasyJet accounts for a further 21 per cent. This

concentration is already considerably higher than at the other London

airports, and seems likely to increase further in light of the airlines that have

switched away from Stansted during 2011 and easyJet‟s intention to switch a

number of services from the airport to operate from Southend from April 2012.

Figure 24 Airline shares of passengers at Stansted 2010

Source: CAA Airport Statistics

3.96 Furthermore, the shares of Ryanair and easyJet‟s businesses that

passengers at Stansted account for are only 17 and 8 per cent respectively,

as shown in Figure 25.

Civil Aviation Authority

Figure 25 Airline concentration at Heathrow, Gatwick and Stansted

Source: CAA Airport Statistics

3.97 When comparing these two figures, it appears that Stansted is significantly

more reliant on the two largest airlines at Stansted than these airlines are

reliant on the airport. In addition, this position contrasts somewhat with both

Gatwick (where there is approximate symmetry between the airline and

airport figures) and Heathrow (where the major airlines are particularly reliant

on the airport). Further, as airlines have switched away from Stansted and

the major airlines have grown their businesses elsewhere, it appears that the

relative bargaining position has shifted to towards the airlines.

3.98 On the face of it, these figures suggest that, with a credible alternative to

switch to, easyJet and, in particular, Ryanair could have quite considerable

bargaining power with Stansted.

RBB report

3.99 Ryanair submitted to the CAA a report by RBB Economics, commissioned to

look into Ryanair‟s ability to exercise buyer power.

It does this firstly by

estimating an elasticity of demand and looks at the impact on Ryanair

volumes of price rises over the last 4 years. RBB calculates a low elasticity of

demand, which reflects the small reductions in volume that followed the date

at which the agreements with Ryanair expired, and the significant difference

between the prices in these agreements and the airport‟s published tariff.

3.100 However, this result (implicitly) relies upon the assumption that the prices

prevailing at Stansted before 2006 were approximately equal to the

competitive price level. The reliance that can be placed on these elasticity

estimates, and the implications that can be drawn from them, depends upon

whether the increase in prices following the expiry of long-term contracts was

an increase in price towards the competitive level, or was an increase from a

price level that was close to the competitive level.

RBB‟s report is available at: http://www.caa.co.uk/docs/5/rbb%20stansted%20final%20non-

confidential%20version%2029%20Nov%2011.pdf

Airline

Airport

% of total airline

passengers

% of total airport

passengers

British Airways

Heathrow

Gatwick

easyJet

Gatwick

Stansted

Ryanair

Gatwick

Stansted

Virgin

Heathrow

Gatwick

Civil Aviation Authority

3.101 Analysis in the next section suggests that Stansted‟s prices were low relative

to other UK airports between 2001 and 2006 and appear currently to be

broadly in line with a number of comparators. This evidence implies that the

elasticity estimates calculated by RBB relate to the responsiveness of airlines

to an increase in prices towards the competitive level, which would be

expected to be below that relating to an increase above the competitive level.

3.102 RBB also emphasises the importance of London as a base for Ryanair, as

illustrated by the continued significant presence at Stansted despite the price

rises. However, it is difficult to separate this general observation from the

preceding argument that the airline has not responded to the expiry of a

number of long-term contracts. Consequently, the interpretation of the lack of

switching could reflect the relationship between the contract prices and the

competitive price level, and that potential that the increase in prices paid

reflected a movement towards, rather than above, the prevailing competitive

price level.

3.103 Furthermore, whilst we accept that London may have a particular strategic

importance to Ryanair, which might prevent it from removing its entire

operations from Stansted, or from airports in London, this does not

necessarily imply that the airline cannot reduce its use of Stansted, or of other

London airports. Indeed, Stansted – and, more generally, the London airports

– failed to attract significant growth from Ryanair in recent years. This

suggests that if there is a strategic importance to operations that limits

switching away from Stansted, these costs are most likely to relate to the

minimum number of services that Ryanair could operate into London, whilst

maintaining a significant overall presence, rather than relating to the number

of aircraft that need to be based at any particular airport.

3.104 One way to consider the issue further would be to analyse the contribution of

Ryanair‟s Stansted operation, in terms of the profitability of the airline and

whether the airline gains a form of strategic advantage by operating from

London.

3.105 []

3.106 []

3.107 The paper then goes on to consider Ryanair‟s buyer power by looking at its

possibilities to switch away. It presents qualitative analysis, with supporting

evidence, to argue that Ryanair would be unable to switch sufficient ATMs to

constrain Stansted. This analysis does not however repeat the earlier

calculation of an implied elasticity. Consequently, the only elasticity estimate

is calculated from the change in volume that took place following the increase

in charges paid by Ryanair at Stansted, which followed the expiry of long-term

pricing agreements, as discussed above.

3.108 RBB also sets out evidence that Ryanair has grown strongly outside of

London, noting that this highlights that there were profitable routes elsewhere.

This pattern of growth suggests that growth at Stansted was significantly

below that experienced elsewhere and that the pricing change at Stansted led

to the airport losing both 32 per cent of its based aircraft between 2007 and

Civil Aviation Authority

2011 and further growth that ended up being deployed elsewhere (an

additional 100 aircraft). This would imply that the elasticity calculated might

understate the actual responsiveness of Ryanair at Stansted, as it is based

only on the loss of services at the airport, rather than also taking into account

the lost growth from new services.

3.109 RBB considers the role of passenger switching, but also acknowledges that if

Ryanair can switch to other airports – other than those to which passengers

can switch – then this would increase its bargaining power. Ryanair‟s recent

growth in based aircraft has been outside of the UK and outside of the

airports serving London and the South East, which strongly supports the view

that Ryanair can switch to airports other than those considered by RBB,

particularly in terms of switching its growth plans between different airports

across Europe.

3.110 However, RBB dismisses pan-European switching by citing the CC‟s

arguments that this would only take place if profitable route opportunities exist

elsewhere, as airline capacity is fixed.

However, RBB presented analysis of

Ryanair‟s growth outside of Stansted and concluded that “[t]his suggests

there were clearly profitable opportunities outside of London that Ryanair

could have redeployed aircraft to from Stansted.”

RBB goes on to argue

that profitable opportunities will be met through fleet growth rather than

through switching, but there is no evidence to support this. Indeed, it appears

that the response to a change in relative prices in the short to medium term

(which would induce switching of existing aircraft or growth) is being confused

with the longer term response to growing demand for the services provided by

Ryanair across Europe.

3.111 We also note that RBB cites the grounding of aircraft as evidence of high

switching costs. However, the seasonal grounding of aircraft appears much

more likely to be a response to the seasonality of demand across Europe, so

as to better matching capacity to demand. It is not, therefore, clear whether

this is evidence of a lack of ability to switch, rather than being evidence of

lower yields in winter. Furthermore, Ryanair recently emphasised that the

costs of grounding aircraft in winter were reduced by the flexibility of its

business model.

3.112 RBB also refers to analysis of Ryanair route profitability, which is not set out

in its response to the CAA. The CAA will request this information in order to

understand better whether the differing yields across Europe provide

evidence on the likely magnitude of the costs of switching away from

Stansted.

RBB report for Ryanair, page 15.

RBB report for Ryanair, page 5

“The great thing about our model is that we’re so flexible we CAN sit up to 80 aircraft on the ground

and give people holidays, unpaid leave, whatever needs to be done, but NOT lose a bundle of money.

The key thing is growth will continue. We’re taking delivery of 30 aircraft from Boeing in calendar 2012,

which takes the fleet to 300. I think we’ll be parking less aircraft on the ground next winter, partly

because fuel will be smoother, and we’ll have had another year to get our average fares to rise across

Europe.”, interview with Michael O’Leary, retrieved from www.anna.aero

Civil Aviation Authority

3.113 Overall, the RBB paper relies on the evidence of actual switching that took

place following the increase in airport charges faced by Stansted when its

agreements with the airport expired. We do not find the arguments that the

airline cannot switch capacity between European bases to be compelling,

particularly in the light of the fact that the airline operates on a pan-European

basis and its statements that it does indeed switch aircraft between non-

neighbouring airports.

Indeed, easyJet also refers to its network

optimisation across Europe, which includes opening and closing routes to

improve overall financial performance.

3.114 Furthermore, whilst we agree that buyer power requires the combination of

scale and the existence of an ability to switch away from the airport, we

consider it likely that both Ryanair and easyJet have the ability to switch

volumes away from the airport. Indeed, the significant number of bases

operated by the two airlines, and recent examples of the airlines opening new

bases – including the example of easyJet‟s switch to Southend – highlights

that there is a significant degree of potential switching from these two airlines.

3.115 In addition, many of the arguments put forward by RBB depend upon the

relationship between the current and historical prices and the competitive

price level. In particular, if the long-term contracts entered into between

Stansted and Ryanair were below the long-term competitive price level, then

the lack of sufficient switching away from the airport to prevent a profitable

increase in price might simply reflect the fact that the price rise was taking

prices towards the competitive level, rather than above it.

3.116 The next sections consider the evidence available about the competitive price

level at Stansted.

Evidence on the competitive price level

3.117 In the CAA‟s Guidance on the Assessment of Airport Market Power, we

considered a number of ways in which a competitive price level could be

assessed, and stated that the long-term average price level would be

assessed using measures of long-run, forward-looking, cost such as

depreciated replacement or incremental cost.

3.118 The Guidance also noted that:

“... it is important to understand, at a minimum, whether there is evidence

that the prevailing and historical price levels are reasonably close to, or

significantly above or below, the competitive level. Given the difficulties

involved in establishing a competitive price level, it might not always be

possible to derive an accurate measure. Nevertheless, where sufficient and

robust evidence is available to determine a reasonable estimate, the CAA

expects to take full account of it.”

For example, the airline stated “Ryanair will switch these London based aircraft to other European

bases where govts have scrapped tourist taxes and reduced passenger charges, in some case to zero,

in order to grow tourism and traffic.” (retrieved from http://www.ryanair.com/en/news/ryanair-cuts-uk-

winter-capacity-by-16-percent). See also references to the relocation of growth from Dublin to other

European airports in http://www.ryanair.com/en/news/ryanair-confirms-20-percent-cut-in-dublin-summer-

2010-traffic

See, for example, http://www.easyjet.com/common/img/easyjet_results_YE_2007.pdf

Civil Aviation Authority

3.119 Gatwick commissioned economic consultants to estimate measures of Long-

run Average Incremental Cost (LRAIC) based on a number of different

scenarios, basing the capital cost estimates on information on the costs of

expanding Gatwick. In addition, we have considered evidence from

comparative pricing data from other airports. These different approaches are

discussed in turn below.

Long-run average incremental costs (LRAIC)

3.120 Gatwick commissioned the consultancy FTI to provide estimates of Gatwick‟s

LRAIC. Its analysis has been shared with the CAA, and a non-confidential

version using different data sources was presented to other interested parties

at a workshop in early December. The non-confidential version of their report

is available on our website, and a more detailed discussion is included in the

CAA‟s Gatwick – Market Power Assessment: Initial Views.

3.121 FTI produced estimates for three different scenarios, drawing heavily on

information about the cost of expanding Gatwick Airport. This analysis

provided estimates of LRAIC of between £10 and £28/passenger, with the

lower estimates being based on the costs of expanding Gatwick up to the

maximum use of its current runway, and the higher estimates relating to the

costs of building a new runway and associated infrastructure (with these costs

being met by incremental traffic only).

3.122 While the calculation of LRAIC numbers is relatively straightforward in

methodological terms, any estimates are highly sensitive to the assumptions

about the scale of costs involved and the expected scale of future passenger

demand. The combination of the inevitable uncertainty attached to the input

assumptions, and the reliance on cost estimates that are based on expansion

(only) at Gatwick, means that the CAA stated that it is difficult to rely on the

FTI estimates in isolation in the context of the market power assessment of

Gatwick. In the context of Stansted, where the cost of expansion might be

expected to be somewhat different to that at Gatwick, it is difficult to place

much reliance on these estimates.

Comparisons with other airports

3.123 An alternative approach to calculations based on estimating long-run

incremental costs is to consider evidence on pricing at comparable airports.

This form of price benchmarking is complicated somewhat by the difficulty in

clearly establishing the types of airports that are reasonably close

comparators to Stansted. For example, it could be argued that Stansted

should be compared to other airports serving predominantly no frills airlines.

Alternatively, Stansted is located close to a very strong pool of demand, with

relatively good surface access links. The latter view would support a

comparison of Stansted to large, but secondary, airports near to large cities

and/or large holiday destinations.

3.124 We had access to three sources for price comparison data, that we discuss in

turn below:

Civil Aviation Authority

 Revenue, passenger number and flight number data from eight London

and regional airports collected from various sources

 the ATRS Global Airport Benchmarking Report 2011, which surveys

156 airports worldwide, including 45 airports in Europe; and

 data obtained from easyJet and Ryanair about airport charges across

their route portfolios.

3.125 Figure 26 and Figure 27 below show the airport charges revenue per terminal

passenger and per flight for eight UK airports. The charts show that between

2002 and 2007, Stansted was charging significantly below all the other

airports for most of the period. Indeed, the period over which Stansted was

(on average) the cheapest airport per passenger coincides with the period

over which there were long-term pricing agreements at Stansted, including

with Ryanair and easyJet. When these contracts expired, Stansted‟s prices

increased.

Figure 26 Airport charges revenue per terminal passenger

Source: CRI, CAA data, airport annual accounts

Civil Aviation Authority

Figure 27 Airport charges revenue per flight

Source: CRI, CAA data, airport annual accounts

3.126 Stansted‟s prices appear to stabilise in line with Manchester‟s and below

those of Gatwick‟s. Looking at airport charges revenue per terminal

passenger, Stansted‟s prices lie below Manchester‟s, whereas looking at

revenue per flight, they are almost equal. While Manchester serves a greater

mix of airlines than Stansted and offers both long- and short-haul flights,

Stansted provides access to the London catchment area. It is perhaps in line

with expectations that the two airports should have similar charges.

3.127 This comparison across UK airports appears to demonstrate that Stansted

certainly has seen a significant increase in its average charges (on both

bases), and that this increase has taken the airport‟s charges to be above

those at a number of other UK airports. However, this increase – particularly

when viewed on a per passenger basis – appears to be an increase of

Stansted towards the level of charges at comparable airports. The evidence

does not support the view that the increase in charges from 2006 was an

increase above the competitive price level.

3.128 Turning to the second comparison, this concentrates on larger international

airports, and so is particularly relevant if the relevant comparators are

considered to be large international airports, whilst the earlier comparison

provides insight into Stansted‟s comparability with other airports serving low

cost carriers.

3.129 Given the heterogeneity of airports, particularly when comparing different

types of airports across different countries, neither comparison will be

conclusive regarding Stansted‟s charges. However, this information may

provide a useful indicator of whether or not Stansted‟s charges are

significantly above or below those of other airports that we might regard as

useful comparators.

3.130 Figure 28 below shows the ranking of the European airports according to the

combined landing and passenger charge elements of airport charges, for an

Civil Aviation Authority

Airbus A320.

This analysis shows that Stansted‟s charges are below

average, but are above a number of primary (and secondary) airports serving

major European capitals, notably Fiumicino (and Ciampino). The analysis

does, however, show that Stansted‟s charges are below those at a number of

other potential comparator airports, such as Birmingham, Paris Orly, Munich,

and Dusseldorf.

Stansted‟s “off-peak” charges (for the Winter period) are

amongst the lowest in the list.

Figure 28 Combined Landing and Passenger Charges at European

airports for A320, 2010 (in USD)

Source: Air Transport Research Society, Key Findings of 2011 ATRS Global Airport Performance

Benchmarking project, highlights presentation

3.131 While using an A320 with just over 110 passengers per flight as a typical flight

appears reasonable for full-service scheduled short haul airlines, it might be

distorting the relative costs per flight for short haul charter and low cost

airlines, given they transport a higher average number of passengers per

flight. For example, easyJet uses predominantly A320-family aircraft with 156

or 180 seats, and its average load factor in 2010 was 87 per cent, which

corresponds to an average number of passengers of 135 and 156.

3.132 Figure 29 and Figure 30 show that while Stansted has the peak landing

charges for A320 aircraft that are well below the average (across the sample

of airports) and has off-peak landing charges that are amongst the lowest.

Stansted‟s passenger charges for international passengers are close to, but

below, the mean across the group.

It assumes an A320 with 150 seats and a 75% load factor, hence 112.5 passengers. Per passenger

charges are doubled as charges are levied only on departing passengers. The “peak/off-peak”

differentiation for Gatwick, Heathrow and Manchester appears to refer to a differentiation for different

times of the day, whereas for Stansted it refers to a differentiation for different months of the year, with

peak referring to April-October.

These airports are not necessarily the „secondary‟ airport serving the capital city (as Stansted does)

but are airports that serve large cities that are not capital cities.

Civil Aviation Authority

Figure 29 Landing charges for Airbus A320, 2010, Europe

Figure 30 Terminal Charge per passenger on international flights, 2010,

Europe

Source: ATRS Global Airport Benchmarking Report 2011

3.133 The report provides no combined passenger and landing charges for long-

haul aircraft types but only the different landing charges. Whilst not directly

relevant to the issue of the pricing faced by the airlines currently using

Stansted, this analysis provides some useful context for understanding how

the airport‟s pricing to potential long-haul services compares to its peers.

Civil Aviation Authority

Figure 31 Combined Landing and Passenger Charges at European

airports for B767-400, 2010 (in USD)

Source: CAA calculations based on ATRS Global Airport Benchmarking Report 2011

3.134 Again, this analysis suggests that Stansted‟s charges are close to, but below,

the average level across the group of airports.

3.135 While a useful source of information, these data are based on published

tariffs. However, in practice a number of airports might provide specific

discounts to airlines, either based on volume or to incentivise the opening of

previously unserved routes. Such discounts can significantly affect average

per passenger charges an airline pays to an airport, and therefore it would be

relevant to base any comparisons on actual charges paid by airlines.

3.136 EasyJet and Ryanair provided us data about the charges they pay to airports

across their respective networks. They capture not only aeronautical charges

(for landing, parking and passengers) but also charges for other services

received from the airport, including for example check-in desk rentals.

3.137 Figure 32 shows easyJet‟s charges across its network, together with the

average level of charges across the network. Stansted and Gatwick are

highlighted in red.

Figure 32 []

[]

3.138 This analysis shows []

3.139 This analysis relies on the implicit assumption that the airports across

easyJet‟s network are reasonable comparators for Stansted. In contrast to

Ryanair, easyJet has focused its network on serving larger destinations and

airports that have good surface access connections. This suggests that this

analysis is particularly useful if Stansted is viewed as comparable to large, but

secondary, airports near to large cities and/or large holiday destinations.

Civil Aviation Authority

3.140 Ryanair also provided price comparison data – in the form of indexed charges

of airports across its network – []. This price comparison implies that – to

the extent that Stansted should be compared to other airports serving

predominantly no frills airlines – Stansted‟s prices at the higher end of the

spectrum.

Summary

3.141 Ryanair and, to a lesser degree, easyJet stress the flexibility of their business

model, routinely vary their route mix and appear able to move capacity across

their networks in the search of higher yields. The airlines, and in particular

Ryanair, also account for a substantial share of the airport‟s business. Whilst

the two airlines no doubt face some costs associated with switching capacity

away from the airport, the balance of available evidence supports the view

that Ryanair, if not also easyJet, enjoys a degree of buyer power at Stansted.

3.142 However, at Stansted prices are negotiated on an airline by airline basis,

which means that any buyer power enjoyed by Ryanair and easyJet would not

protect the smaller airlines at Stansted and would therefore not be sufficient to

constrain any market power at Stansted with respect to these smaller players.

3.143 Further, whilst think it likely that there is a degree of buyer power, it is not yet

clear whether this would be sufficient to offset any market power found to be

enjoyed by the Stansted over Ryanair and easyJet. To investigate this issue

further we will need to consider the evidence available from both carriers on

the yields available at Stansted and other airports, and will be looking for

additional quantitative evidence to better understand the magnitude of

switching costs at the airport.

Entry and expansion

3.144 The preceding section considered the responsiveness of airlines and

passengers to changes in price, and whether these users would switch in

sufficient numbers to an alternative existing airport to constrain the airport‟s

conduct. However, competitive constraints can also arise from entry and/or

expansion of airports in Stansted‟s market.

3.145 The impact of this form of competitive constraint will be limited by the

magnitude of barriers to entry and expansion. These factors are considered

below.

Barriers to entry and expansion

3.146 The CAA‟s Guidance on the Assessment of Airport Market Power noted that

barriers to entry in airport markets are particularly high and that expansion of

existing airports is more likely to represent a competitive constraint on existing

airports than the threat of entry by an entirely new airport. New airports can

sometimes enter the market, but the investment and lead times involved in

new entry are likely to significantly limit the impact of this form of competitive

constraint.

For example, Robin Hood Doncaster Sheffield airport opened in April 2005, and London City Airport

opened in 1988.

Civil Aviation Authority

3.147 Expansion and/or entry by existing aerodromes, and/or the threat thereof,

may represent a source of competitive constraint.

However, as with de novo

entry, the cost and timescales involved in expanding to accommodate

sufficient switching may still be too great to constrain Stansted‟s prices in the

short to medium term.

Evidence of actual entry or expansion

3.148 One way to understand the nature of barriers to entry and expansion is to

consider the history of entry and expansion in the market. There is very

limited evidence of significant entry or expansion in Stansted‟s market.

However, there are two recent examples of expansion in the form of

Southend and the recent announcement of Luton‟s intention to increase

capacity.

3.149 EasyJet recently struck a deal with Southend to transfer some of its flights

from Stansted to Southend. As a result, Southend is upgrading its facilities

and there is also an expectation that the rail link from London to the airport

will be improved. It is possible that other airlines may begin operating out of

Southend, following easyJet‟s lead, however any operations would be limited

by Southend‟s runway specifications, which, for example, would not currently

support Ryanair‟s fleet. EasyJet services from Southend are expected to

commence in April 2012, and will contribute to a significant reduction in the

easyJet services operating from Stansted going forward.

3.150 In addition, Luton‟s owners (Luton Airport Limited) have announced a process

of consultation with a view to increasing capacity at the airport from circa 10.3

mppa to 18 mppa, although it is unclear at this stage whether and when such

expansion might take place.

3.151 In summary, while relatively small scale expansion is possible, the timescale

required to achieve adequate expansion to accommodate sufficient switching

is too long to constrain prices in the short term, but there is some potential for

expansion projects that act to constrain pricing in the medium term at least

(as illustrated by the Southend example).

Capacity constraints

3.152 The CAA‟s Guidance on the Assessment of Airport Market Power

explained

that scarce capacity can be a normal feature of a market and may not

necessarily lead to market power and monopoly pricing. However, the

underlying causes of scarce capacity might constitute significant barriers to

entry and expansion. Where capacity constraints are caused by barriers to

entry and/or expansion, they are likely to affect the strength of competitive

constraints faced by an airport, and may give rise to market power held by

incumbent airports, particularly where the barriers are non-transitory in nature.

3.153 The context of capacity constraints in the southeast and capacity constraints

at the main London airports are discussed in the annex on capacity

constraints.

For more information see, http://www.futureluton.co.uk/optimisation.html

„Guidance on the Assessment of Airport Market Power‟, CAA, April 2011 (chapter 5)

Civil Aviation Authority

3.154 As Figure 33 below shows, Stansted has very limited spare capacity during a

narrow peak period (with limited prospects for additional capacity at this time

of day). However, it has ample spare capacity off peak through much of the

day.

Figure 33 Stansted slot allocation, Summer 2011 and Winter 2001/2012

Summer

Winter

Source: ACL UK Start of Season Report Summer and Winter 2011

3.155 As discussed above, the major airlines at Stansted have argued that morning

peak capacity is particularly important to the operation of based, low-cost

aircraft.

3.156 As we identified above, the strongest local competitors to Stansted are likely

to be Gatwick and Luton. As Figure 34 shows, Gatwick is capacity

constrained. In particular, Gatwick is full through much of the day during the

summer period and Gatwick is full during the peak morning period. This could

potentially affect airlines‟ ability to switch to these airports at this time of day.

Civil Aviation Authority

Figure 34 Gatwick slot allocation, Summer 2011

Source: ACL Start of Season Report, LGW Summer 2011

3.157 Overall, it is clear that there is significant capacity off-peak at Stansted, albeit

that this might reduce over time, as demand grows. However, at peak times,

there is very little spare capacity. The period over which these capacity

constraints holds is currently somewhat limited, relating to 2 or 3 hours, but if

demand grows these constraints would be likely to spread to other hours.

3.158 However, Gatwick notes that the airport “is exploring the possibility of

declaring 55 movements per hour in peak periods, which is well above the

levels currently being achieved by Stansted (which currently has only 50) or

Luton”.

Gatwick also emphasises the spare capacity available, on both a

passenger and ATM basis, across the London airports. In particular, Gatwick

argues that – when comparing current traffic to potential capacity in 2020 –

there is approximately 30 per cent spare ATM capacity and 40 per cent

passenger capacity.

3.159 Overall, the evidence available suggests that the London airports are

operating close to their capacity at peak times, albeit that steps may be

possible at some of these airports to deliver additional peak capacity to the

market. At this stage, it is not clear whether these capacity increases could

be deployed in a sufficiently timely manner to discipline Stansted sufficiently

at peak times. It is also relevant that demand is forecast to grow at the

London airports, with this demand growth offsetting – to some degree – the

potential for capacity to increase at peak times.

Gatwick submission to the CAA, November 2011.

Civil Aviation Authority

Pricing and behaviour

3.160 This section considers evidence from the pricing and behaviour of the airport,

and how this might inform the assessment of the airport‟s market power.

Significant increase in price in 2006

3.161 In the CAA‟s Guidance on the Assessment of Airport Market Power, it stated

that it agreed with the OFT‟s statement that “where prices are likely to differ

substantially from their competitive levels, caution must be exercised when

dealing with the evidence on switching patterns as such evidence may not be

a reliable guide to what would occur under normal competitive conditions.”

84,85

It can, therefore, be important to understand how prevailing and historical

prices relate to the competitive price level.

3.162 Stansted entered into a number of long-term contracts with its major airline

customers, following the aviation downturn after 11 September 2001. These

contracts were struck at prices that were substantially lower than the

prevailing price cap, with some as low as [] (compared to the current price

cap of around £6.50). When these contracts expired, mostly during 2006, the

average prices paid by the major airlines at Stansted increased substantially.

This increase in prices paid was significant and did not appear to result in a

sufficient reduction in use of the airport to render it unprofitable for the airport.

This has been cited by airlines at Stansted as evidence that the airport enjoys

SMP.

3.163 However, the interpretation of this price rise depends upon whether the price

rise is viewed as an increase above the competitive level or an increase

towards the competitive price level. This is important, as the latter would not

be expected to generate a particularly strong response from customers,

whereas a rise above the competitive level should elicit a strong response at

a competitive airport; the absence of which would be particularly strong

evidence of market power.

3.164 It is difficult to establish with precision whether prices were at „competitive‟

levels or whether the current levels are now close to these levels, but the

following provide some evidence:

 There are some indications that some of the 2001 contract prices were

very low. Indeed, the prices have not been sufficient for the airport to

earn the allowed return on its regulatory asset base for an extended

period, and it is not forecast to do so in the near future.

 In contrast, and as discussed in paragraph 3.136, Ryanair has argued

that Stansted now has the highest charges across its network, other

than Gatwick and Dublin, providing supporting evidence for this view.

Evidence from easyJet is more mixed, indicating that the airport is

approximately „mid table‟ across its network.

For example, product markets could be defined excessively broadly on the basis of the prevailing price

level if this is above the competitive level due to pre-existing market power of the undertaking(s) in the

market. This is referred to as the Cellophane Fallacy, after the case US vs. El Du Pont de Nemours &

Co [1956] 351 US 377.

OFT Market Definition guideline (OFT403) paragraph 5.6.

Civil Aviation Authority

3.165 As discussed in paragraph 3.123, the weighting of this evidence depends

upon the view taken of the appropriate comparator airport for Stansted:

should it be compared to UK and European regional airports supporting LCC

operations; or is the appropriate comparison larger airports, reasonably

connected to large cities, with a mix of LCCs, LCCs and long-haul

operations? The former view might suggest that the „competitive‟ price at

Stansted would not remunerate many of the assets at the airport, certainly not

until the airport faces higher demand relative to available capacity, which

might suggest this is not a normal or sustainable price level. However, these

are the prevailing market conditions for a lot of the LCC sector.

3.166 It appears appropriate for a competitive airport both to discount significantly

where there is spare capacity but also to have the freedom to increase

charges and recoup the costs of efficient investment during peak periods.

Given that there is stronger demand at peak times, this implies that the airport

might be expected to increase charges at these peak times – something that

the current price control allows, so long as charges in off-peak times are

sufficiently low so as to bring the average charges below the price cap.

3.167 This suggests that even if there was a case to „write-down‟ the assets at

Stansted (to better reflect those needed at a no-frills airport) by a very

significant amount – and, say, to halve average charges – it could be argued

that an such an airport would still be acting reasonably by loading the

recovery of this lower asset base towards peak periods. With only two of 18

hours facing constraints, this allocation of asset costs to peak might readily

counter-act even an aggressive write-down of Stansted‟s asset base, and

support the view that the current (peak) prices are reasonable given the

market circumstances.

3.168 More generally, whilst the airport has set its published charges close to the

price cap, it has the ability to increase peak charges above the level of the

price cap. However, Stansted has not chosen to set its charges in this way.

This lack of peak/off-peak pricing throughout the day might support the view

that the airport cannot sustain a price increase at these peak times, for fear of

the airlines at peak switching away from the airport. An alternative

explanation might be that the airport is concerned about the potential for the

customers using peak capacity to challenge the airport‟s freedom to increase

charges in this way. The latter explanation might be consistent with the view

that the major airlines at Stansted enjoy a degree of buyer power.

Reduction in traffic

3.169 In 2010, 18.6 million passengers travelled though Stansted, this is a reduction

of 7 per cent on 2009 passenger numbers and brings Stansted back down to

2003 levels. This significant reduction in airline use of the airport follows the

increase in charges at the airport, but also coincides with a recession in the

UK, and low levels of GDP growth across Europe. Consequently, the

reduction in traffic at Stansted could either reflect economic trends or price

responsiveness.

Civil Aviation Authority

3.170 Ryanair says that the fall in traffic at Stansted is not due to the economic

downturn, arguing that it and easyJet have expanded elsewhere despite the

downturn. Ryanair says that the fall in traffic is as a result of Stansted failing

to offer sufficient discounts. This shows that airlines do have a choice of

where to expand and at least some ability to switch away when Stansted‟s

relative prices go up.

Stansted’s efforts to attract new business

3.171 The reduction in traffic at Stansted has resulted in ample spare capacity off-

peak (at least 10 useable slots in each hour). It might be expected that the

airport would be offering aggressive discounts for growth, in an attempt to

generate some revenue from this capacity.

3.172 As noted above, Ryanair, in particular, has argued that this is not happening

and that Stansted‟s charges are out of line with UK and European comparator

airports. However, the CAA has seen both growth contracts offered to

Ryanair at other UK airports and examples of the offers made by Stansted to

other airlines for growth. The level of discounts offered by Stansted appear –

on the face of it – to be substantial [].

3.173 Nevertheless, these offers have not been sufficient to attract growth, or to

retain the current traffic at the airport. []

3.174 []

3.175 These views would tend to suggest that the presence of Ryanair at the airport

has the effect of discouraging other airlines with a similar business model

from operating at the airport. To the extent that this reduces the ability of the

airport to attract new carriers it would limit the airport‟s market power

(certainly off-peak) and might compound the bargaining strength enjoyed by

Ryanair.

3.176 However, this does not appear to explain all of the apparent reluctance to use

the airport, as Air Asia X would not obviously be concerned about Ryanair‟s

short-haul operations []. Clearly, there are other factors that make

Stansted less attractive to Air Asia X (and by extension perhaps to other

carriers). []

3.177 Overall, the evidence provided by Stansted appears to support the view that it

is offering „competitive‟ discount offers to support growth in off-peak periods.

Indeed, the evidence currently available supports the view that there is no

SMP relating to the off-peak periods.

Evidence of market power at peak

3.178 In general, the balance of available evidence implies that airlines at Stansted

have a choice of airports to switch to, facing limited switching costs other than

any loss in revenue associated with the switch.

3.179 However, Chapter 2 identified the importance of morning departure slots for

based aircraft for the airlines at Stansted. In order to make best use of their

aircraft, this type of airline must depart early in the morning to ensure

sufficient rotations or the aircraft.

Civil Aviation Authority

3.180 As shown above, reflecting the importance of these peak slots, Stansted is

capacity constrained during the morning peak. Similarly, Luton and Gatwick

are also constrained at this time. This lack of capacity at this crucial time of

day may make switching based aircraft more difficult.

3.181 Nonetheless, switching may take place, but it is likely to be more costly to

obtain slots at this time. []

3.182 In Chapter 2, it was concluded that the geographic market for airlines at

Stansted was likely to be European wide. We must therefore also consider

the capacity of other European airports during the peak time. So far we have

only limited evidence on this, which is drawn from publicly available

information.

However the following charts show that it is common for these

other European airports to also be constrained at peak.

3.183 This evidence implies that airlines may find it difficult to switch away from

Stansted during the peak morning hours, suggesting that Stansted could have

market power over these airlines at this time. At this stage we do not have

sufficient evidence to conclude whether Stansted has significant market

power and will be seeking further data in order to be able to carry out analysis

on this issue.

„Impact Assessment of revisions to Regulation 95/93 – Final Report‟, European Commission, March

2011

Civil Aviation Authority

Figure 35 Dublin, Summer 2008, slot demand and allocation

Figure 36 Milan Linate, Summer 2008, slot demand and allocation

Civil Aviation Authority

Figure 37 Rome Fimucino, Summer 2008, slot demand and allocation

Civil Aviation Authority

Figure 38 Palma Mallorca, Summer 2008, slot demand and allocation

Civil Aviation Authority

Figure 39 Vienna, Summer 2008, slot demand and allocation

3.184 It could be argued that if Stansted had market power during the peak period,

we should see higher prices during this time period. Stansted offers significant

discounts off-peak so theoretically it could put prices up at the peak without

exceeding the cap. As discussed above, Stansted does not charge higher

prices during the peak, which might suggest that the airport is unable to

increase charges above the current level.

Financial performance

3.185 Stansted is subject to price cap regulation, which is designed to prevent

airports from earning excessive returns. Consequently, analysis of the

financial performance of these airports is unlikely to provide particularly strong

evidence about their market position, particularly if the airports choose to set

their prices at, or near to, the allowed price cap – something which Stansted

is now doing.

Impact of common ownership with Heathrow

3.186 As noted above, Stansted and Heathrow are still currently under joint

ownership. The discussion of market shares highlighted that, whilst

Stansted‟s market shares were generally not particularly high, when the

combined position of Stansted and Heathrow is considered the market shares

are sufficiently high as to adopt a rebuttable assumption of dominance.

3.187 As discussed in the Competition Commission‟s (CC) inquiry into BAA‟s

ownership of UK airports, the effects of joint ownership could take a number

Civil Aviation Authority

of forms. The CC also concluded that the joint ownership of Stansted and

Heathrow inhibited competition to such a degree that there is a case for the

divestment of Stansted by BAA.

3.188 It could, therefore, be argued that Stansted‟s competition for airlines and/or

for passengers is muted by concerns about the adverse impact on Heathrow.

Stansted disputes this view and argues that any transfers of passengers from

Heathrow to Stansted would quickly be back-filled as a result of the strong

underlying demand at Heathrow, making BAA neutral to Stansted‟s

competition with Heathrow.

3.189 However, there is a marked contrast between the aggressive marketing of

Heathrow and Gatwick to passengers and the relatively low-key approach

adopted by Stansted, which could be explained by the continued ownership

by BAA. Indeed, there is some anecdotal evidence that Stansted scaled-back

its marketing efforts [].

3.190 Of course, even if joint ownership were reducing the ability of Stansted to

compete, this is not sufficient to conclude that divestment by BAA would be

sufficient to remove any SMP held by the airport. Rather, that divestment

would lead to a more competitive outcome than continued joint-ownership.

Separate ownership may be necessary, but it is not sufficient

3.191 Ownership appears likely to change in the near future and the impact of this

will depend on the new owner. This may address concerns on incentives to

compete for traffic. However, it is unlikely to be in sufficient time to provide

compelling evidence on behaviour before a decision on market power is

made.

Cargo

3.192 Whilst Stansted has a strong position in cargo-only flights, we consider it likely

that the geographic market for cargo-only operations is likely to be very broad,

and to be UK-wide (if not wider). Further, Stansted‟s shares of these markets

do not give rise for particular concern.

3.193 However, given the size of Heathrow, when Stansted and Heathrow are

considered together, BAA has a very substantial share of UK cargo and –

through Stansted – a significant share of cargo-only operations.

3.194 At this stage, the CAA has not undertaken sufficient analysis to reach any firm

conclusions in respect of cargo at Stansted, and has not received any

evidence relating to cargo operations.

3.195 Reflecting this, we invite contributions from stakeholders and will continue our

work in this area during 2012.

Long term view

3.196 Stansted‟s market power will, inevitably, vary over time, as its circumstances

change. In this section we consider the factors that appear likely to affect the

airport‟s future market power.

3.197 We discuss the following developments and how they might affect Stansted‟s

market power:

Civil Aviation Authority

 Demand growth and capacity constraints;

 The sale of Stansted; and

 The potential growth of long-haul at the airport.

Traffic forecasts and capacity constraints

3.198 The DfT regularly produces long term passenger forecasts for UK airports. Its

latest forecasts published in August 2011 predict that absent any further

runway expansion by 2030 Heathrow, Gatwick and Stansted will be operating

at full capacity, suggesting there will be excess demand for these three

airports.

3.199 Increasing capacity constraints at Stansted and its competitors could suggest

higher switching barriers for passengers and airlines and therefore a higher

level of market power: There would be less scope for airlines at Stansted to

switch to other airports, and excess demand suggests that switching would

have limited effects on the airport as any vacated slots would be taken up by

other airlines.

3.200 However, in response to tightening capacity at the three largest London

airports, it is possible that other airports, including Luton and neighbouring

regional airports, might emerge as more credible competitors by developing

their infrastructure and service offering. Indeed, Stansted is facing the

prospect of expansion from Southend in 2012, which is expected to have a

direct adverse impact on volumes at Stansted, as easyJet relocates capacity

between the two airports. During 2012 additional information will be available

about the initial success of this expansion plan, as traffic statistics emerge.

3.201 Further, the owners of Luton Airport have announced plans to increase

capacity at the airport, taking the airport towards a passenger capacity of 18

mppa. These two events could lead to a significant increase in the availability

of capacity to compete with Stansted – at two airports with strong catchment

overlaps with the airport – albeit that there is a degree of uncertainty attached

to the success of both projects.

3.202 More generally, and over the longer term, a range of other regional airports

might emerge as stronger alternatives to Stansted, and the other London

airports, as capacity constraints provides a commercial case for expansion at

these airports. For example, Birmingham was recently discussed in the press

as a potentially viable alternative to the large London airports. LCC operators

might also find it more profitable to move any traffic growth increasingly to

other geographic areas across Europe.

3.203 However, even with growing capacity constraints, the London airports, in

separate ownership, might still find it profitable to compete for marginal users,

for example by inducing airlines to use larger aircraft and passengers with

higher retail spend to use their airport instead of their competitors.

3.204 In summary, whilst there is a degree of uncertainty, it appears likely that

capacity constraints are unlikely to ease over time and that demand growth

will tend to increase Stansted‟s market power, albeit that there are projects

that might – if successful – act to erode this position.

Civil Aviation Authority

The potential sale of Stansted by BAA

3.205 Following the recent decision at the Competition Appeals Tribunal, the CC‟s

decision to require BAA to sell Stansted Airport remains in effect, and it

appears increasingly likely that in the near future Stansted will be sold by

BAA.

3.206 In the CAA‟s assessment of Gatwick‟s market power, we discuss the

noticeable impact that the sale of Gatwick had on the airport‟s behaviour. We

consider that it would be reasonable to expect a similar development following

the sale of Stansted. Although the precise impact of the sale is difficult to

predict, we would expect the airport to have greater freedom to innovate in

terms of its service offering and in its commercial arrangements with airlines.

For example, the airport might market itself more aggressively and identify

ways for it to distinguish its offer (to a greater degree) from that available at

Luton, Gatwick and Heathrow.

3.207 More generally, the sale of Stansted would remove concerns about the

airport‟s incentives to compete for airlines and passengers, and remove the

relevance of combined Heathrow-Stansted market shares from the above

analysis. It is difficult to assess the impact of ownership change at this stage,

and it will no doubt be difficult to assess the conduct of an independent

Stansted for some time (due to the limited information available about its

conduct), there is a reasonable prospect that divestment would have a

significant impact on the airport‟s market power.

Potential growth of long-haul at the airport

3.208 Much of the analysis in this paper focuses on the relationship between

Stansted and LCCs operating short-haul services. However, the airport can

accommodate large, long-haul operations. As we note, there is evidence that

the airport has been seeking to attract these services, as yet without success.

3.209 To the extent that the airport successfully attracts significant long-haul

operations to the airport, we would expect to update the analysis, and

broaden the focus to include, for example, market definition that covers long-

haul operations. We have not, however, identified any reason why the growth

of long-haul operations at Stansted would have a significant impact on the

airport‟s market power.

Overall assessment of market power

3.210 This final section presents a summary of the factors that appear to contribute

to Stansted‟s overall market position. It considers evidence on market

shares, airline switching, passenger switching and the airport‟s behaviour,

before presenting an overall conclusion.

Market shares

3.211 Market shares can provide an indicator of an airport‟s market position. Even

under the narrowest definition, when we limit the market to be short haul

flights from the London area, Stansted does not have a high market share,

when viewed as a stand-alone airport, and certainly below the level at which

there would be a rebuttable presumption of dominance. However, Stansted

Civil Aviation Authority

and Heathrow are both currently owned by BAA, and combining the market

shares of the two airports gives very high market shares and which on a UK-

wide basis are still as high as 58 per cent.

3.212 An important aspect of understanding market power at Stansted is to consider

the position of the airport at the early morning peak. Looking only at peak

periods increases Stansted‟s market share to 26 per cent (behind Heathrow

and Gatwick‟s 30 per cent), and a combined share with Heathrow of 56 per

cent. On some measures, the combined share of BAA-owned airports would

support a rebuttable presumption of dominance.

Airline switching

3.213 The predominant airline business model at Stansted is low-cost, short-haul,

point-to-point. In general, these airlines will have invested less at the airport

than other airline business models, have multiple bases across the UK and

Europe and, due to their more streamlined cost structure, face airport charges

that generally account for a bigger proportion of the total costs than they do

for full-service network airlines. This implies that there will be a greater

incentive, and more ability, for low cost point-to-point airlines to switch in

response to a given increase in airport charges, which is consistent with

evidence that highlights that the major carriers at Stansted operate

particularly dynamic networks, with the routes flown varying to a significant

degree over time.

3.214 Even if switching costs are low enough to allow airlines to switch, they must

have appropriate alternative airports to switch to. A based carrier at

Stansted‟s ability to switch to neighbouring airports may be limited by

constrained capacity at other London airports, but those airlines with a

network of bases across Europe also have the option of switching to other,

non-London airports.

3.215 The likely magnitude of switching costs, and the sensitivity to airport charges,

is mirrored by the evidence of airline switching. In recent months, Stansted

has lost several major airlines to Gatwick (Norwegian, Air Berlin and Air Asia

X), despite their being some capacity constraints at Gatwick. In addition,

easyJet has recently announced its intention to move some of its based

aircraft from Stansted to Southend.

Passenger switching

3.216 We have considered a range of evidence to understand the extent to which

passengers might switch between airports. Analysis based on airport

catchment areas highlights the significance of catchment area overlaps,

particularly over the most densely populated areas (notably Greater London).

This suggests that a significant proportion of passengers are likely to be

marginal, and able to switch away from Stansted, if a suitable service is

available at an alternative airport.

3.217 Survey evidence is consistent with this general finding, showing that there are

significant numbers of Stansted passengers who have previously used

another London airport (indicating a degree of willingness to travel to use

Civil Aviation Authority

these airports) and that stated that they had considered alternative airports to

Stansted. This is consistent with the survey results that indicated a higher

degree of price sensitivity at Stansted and Gatwick, when compared to

Heathrow.

3.218 However, this apparent willingness to switch between airports relies upon

passengers being able to find a suitable service at an alternative airport. In

this respect, whilst Stansted has fewer route overlaps than other London

airports, 74 per cent of passengers at the airport could fly to their chosen

destination from another London airport. This suggests that a large

proportion of passengers are likely to have the willingness and ability to

switch away from Stansted.

3.219 Further, as only around 10 per cent of Stansted‟s passengers were on

regularly-served routes served by more than one airline at Stansted, it is likely

that should an airline remove a service from Stansted, the airport would be

more likely to lose the passengers from that route to another airport. Indeed,

the interplay between passenger and airline decisions means that the loss of

business to an airport resulting from an increase in price may be greater than

indicated by the impact on „marginal‟ passengers, as services become

unprofitable and prompt those who would have used these services to

consider using another airport.

Airport sensitivity to airline and passenger switching

3.220 In order to understand the impact that passenger and airline switching is likely

to have on the airport, it is useful to consider the volume reduction (or „critical

loss‟) that would render an increase in prices unprofitable. CAA calculations

suggest that to render a 5 to 10 per cent price rise unprofitable the airport

would need to experience a fall in passenger numbers of between 3 and 11

per cent, depending on the assumptions used. For example, if the price

increase applies to the airports overall revenue, a 5 per cent increase would

be rendered unprofitable if the airport lost approximately 900,000 passengers,

or approximately 1,000,000 passengers if operating costs are saved as

passenger numbers fall.

3.221 Easyjet commissioned Frontier Economics to analyse the potential for airline

switching to constrain prices at Stansted, undertaking similar calculations to

those set out above. Frontier‟s paper shows that airlines switching and

passenger switching would be insufficient to render a 10 per cent increase in

charges unprofitable. However, whilst this analysis makes a useful

contribution to the available evidence, the approach taken restricts passenger

and airline switching in a number of ways that are likely to under-state the

actual level of switching. In addition, the data made available to Frontier

introduces a further potential source of bias, in that it is based on average

easyJet route profitability, rather than the profitability of the most marginal

services.

Buyer power

3.222 Ryanair and easyJet account for a very large proportion of passengers at

Stansted, with 68 and 21 per cent shares respectively. Further, whilst

Civil Aviation Authority

Stansted is no doubt an important base for both airlines, the airport accounts

for a much lower proportion (17 and 8 per cent) of the airlines‟ passengers.

This implies that Stansted is significantly more reliant on Ryanair and easyJet

than these airlines are reliant on the airport, and that there could be a degree

of buyer power at the airport.

3.223 Analysis undertaken for Ryanair by RBB disputes the view that there is a

degree of buyer power at the airport. The RBB paper cites the evidence of

actual switching that took place following the increase in airport charges faced

by Stansted when its agreements with the airport expired, emphasises the

importance to buyer power of having an ability to switch, and also argues that

there are significant costs to Ryanair relocating aircraft across its European

network. Whilst we agree that buyer power requires the combination of scale

with the existence of an ability to switch away from the airport, we consider it

likely that both Ryanair and easyJet have the ability to switch volumes away

from the airport. Indeed, the significant number of bases operated by the two

airlines, and recent examples of the airlines opening new bases – including

the example of easyJet‟s switch to Southend – highlights that there is a

significant degree of potential switching from these two airlines.

Competitive price level

3.224 The interpretation of the evidence on airlines‟ historical responses to price

increases is affected by the relationship between historical prices and the

competitive price level. This is particularly important in light of the very

significant increase in prices paid at the airport during 2006 and the relatively

moderate reductions in volumes that followed.

3.225 In order to understand whether these price rises were increases towards, or

increases above, the competitive price level we have considered a range of

sources of evidence.

3.226 Ryanair provided evidence that showed that its charges at Stansted are