German Aviation Research Society Competing Airports – Theory or Reality 19-20 November 2004

1
German Aviation Research SocietyCompeting
Airports Theory or Reality19-20 November 2004

• The Impact of Low-cost Carriers on Regional
  Airports Financial Performance Evidence from the
  UK
• Zheng Lei
• Dr Andreas Papatheodorou
• Dr Edith Szivas
• University of Surrey, UK

1
2
Overview

• The Rise of Low-cost Carriers
• Regional Airports
• Low-cost Carriers and Regional Airports
• Proposed Methodology A Panel Data Analysis
• Some Preliminary Results

2
3
Airline Deregulation in the EU

• Three Aviation Liberalisation Packages
  (1987-1997).
• From 1993, EU airlines became able to fly between
  member states without restriction and within
  member states subject to some controls on fares
  and capacity.
• From 1 April 1997, full cabotage was implemented
  and any airline with a valid Air Operators
  Certificate can operate within the EU at
  market-determined prices (Gillen Lall, 2004).

3
4
The Emergence of Low-cost Carriers

• The first European LCC Ryanair- adopted
  Southwests low-cost, no frills strategy in 1995.
• easyJet the first European airline which
  established as a LCC from its inception.
• In 1998, there were three LCCs, the number was
  jumped to fifteen in 2002 and thirty-one in early
  2004.

4
5
Low-cost Carriers in Europe
5
6
The Structure of the European Airline Industry

• Network carriers still control the majority of
  the market share. The top three network carriers
  accounted for 34 of total AEA RPKs in 2002, with
  the top six accounting for 56 (AEA, 2003).
• Charter airlines serve 45 billion packaged tour
  market, accounting for about 21 of
  intra-European air travel (Binggeli Pompeo,
  2002).
• LCCS has 30 of short-haul market share in the UK
  and Ireland (charter traffic included) European
  wide, less than 10 of capacity, but was forecast
  an annual growth of more than 10 per cent in the
  next 10 years (Dennis, 2004 Travel Tourism
  Analyst, 2003).

6
7
The Competition Strategy in the Airline Industry
(1)

• Cost leadership lower the cost of production and
  compete by offering low rates protected by
  economy of scale facilities the firm builds to
  reduce the unit cost of production.
• Differentiation to make some dimension of an
  airlines activity different from that of its
  competitors, e.g. better schedules, better
  amenities, better services it protects markets
  by establishing a group of loyal consumers.
• Monopoly to restrict the ability of competitors
  to enter selected geographic markets, e.g.
  control over network hubs.

7
Source Sorenson, 1990
8
The Competition Strategy in the Airline Industry
(2)

• The rise of LCCs are characterised by cost
  leadership strategy.
• Cost leadership is an extremely difficult
  strategy in any competitive environment every
  decision that managers make must be guided by the
  principle of cost control.
• It is an even more difficult strategy in the
  airline industry as there are strong indications
  that there are generally no increasing returns to
  scale associated with airline networks.

8
9
Economies of Density

• Although an airline cannot substantially reduce
  unit costs by expanding scale, economies of
  density exist in the airline industry.
• Supply side economies of density an airlines
  unit cost declines when the airline adds flights
  or seats on existing routes, all other things
  held constant.
• Demand side economies of density an airline
  induces higher demand through higher route
  frequencies.
• However, the strategy of focusing on economies of
  density is open to both network carriers and
  LCCs.
• Network carriers explore economies of density
  through hub-and-spoke network, while LCCs achieve
  it from point-to-point operation on dense routes.

9
10
Point-to-point operation vs. Hub-and-spoke
Network (1)

• Advantages of hub-and-spoke
• Cost benefit
• Demand benefit
• Better allocation of capacity under demand
  uncertainty
• Disadvantages of hub-and-spoke
• Increase of circuity, which in turn, negatively
  affects airline costs (e.g. extra fuel
  consumption, extra cruise time, extra fixed
  costs).
• Congestion at the hub airport, which severely
  limits the expansion of the hub-and-spoke
  network.
• Cannot satisfy both business and leisure
  passengers business passenger prefers direct
  flight, while leisure passengers prefer the
  lowest possible fares.

10
11
Point-to-point operation vs. Hub-and-spoke
Network (2)

• LCCs provide lower fares that leisure travellers
  are increasingly demanding and direct flight that
  business passengers prefer.
• The sustainable advantages that LCCs have are
  mainly derived from the adoption of
  point-to-point operation (Gillen Lall, 2004).
• 70 of the cost differences between LCCs and
  conventional scheduled airlines can be attributed
  to their business model choices (Booz Allen
  Hamilton, 2002).
• However, no conclusive findings yet the debate
  is still continuing.
• It is important for LCCs to maintain significant
  cost advantages over network carriers.

11
12
Cost advantages of low-cost carriers on
short-haul routes
Assumes 100 per cent direct sales and none
through agents for low-cost airlines. Source
Doganis (2001)
12
13
Non-airport Related (1) Higher Seating Density

• By removing business class and reducing the seat
  pitch and gallery, LCCs can significantly
  increase the number of seat for sales in their
  aircraft.
• However, the business class can bring higher
  yields to traditional airlines, thus the real
  cost advantage is likely smaller.
• Moreover, this strategy is not well protected
  from the manoeuvres of network carriers, e.g. the
  removal of business class by Aer Lingus on
  certain intra-European routes.

13
14
Non-airport Related (2) Distribution Costs

• By cutting off travel agents and using e-tickets
  LCCs could achieve 11 cost advantage over
  traditional airlines.
• This strategy was adopted by conventional
  scheduled airlines as well
• - Commissions paid by network carriers to
  travel agents are
• now typically 1 on short-haul
  flights
• - E-tickets are widely used by network
  carriers as well
• - Network carriers also seek ways of bypassing
  GDSs by streamlining their Internet booking
  facilities
• Therefore, the cost savings from distribution
  channels are unlikely to be sustainable.

14
15
Non-airport Related (3) Lower Staff Salary and
Smaller Administration Costs

• By entering the industry with non-union labour,
  the LCCs were able to gain an initial cost
  advantage.
• This advantage, however, will be eroded as the
  large carriers won labour concessions, while LCCs
  come under pressure to increase pay.
• Similarly, when LCCs grow up, smaller
  administration costs that they enjoyed will be
  difficult to maintain.

15
16
Airport Related (1) High Aircraft Utilisation

• Regional airports help LCCs to enhance their
  ability to sustain high employee productivity and
  reliable on time performance.
• High aircraft utilisation is achieved through the
  greater aircraft and employee productivity from
  fast turnarounds in the uncongested regional
  airports.

16
17
Airport Related (2) Low Airport Charges

• LCCs benefit enormously from low airport charges
  and station costs offered by regional airports.
• For example, Frankfurt/Hahn costs Ryanair 4.25
  per departing passenger and there is no landing
  fee in contrast a 737 operator at Frankfurt/main
  pays 13 per departing passenger and a landing
  fee of about 1.75.

17
18
Major Sources of LCCs Sustainable Competitive
Advantages

• When network carriers learn how to compete with
  LCCs over time, LCCs non-airport related cost
  advantages will be difficult to maintain.
• Given the importance of business passengers and
  of network feed, it is unlikely for network
  carriers to choose secondary airports and
  traditional airlines would find it difficult to
  copy the high utilisation concept while
  maintaining complex satisfactory flight
  connections (Doganis, 2001).
• In sum, the sustainable competitive advantages
  that a LCC has mainly derived from the adoption
  of point-to-point service operating from regional
  airports.

18
19
Overview

• The Rise of Low-cost Carriers
• Regional Airports
• Low-cost Carriers and Regional Airports
• Proposed Methodology A Panel Data Analysis
• Some Preliminary Results

19
20
Airports From the Public Utility to Business
Entity

• As strategically important asset, airports,
  traditionally, have been administered and
  controlled either by central or local government
  or state appointed body.
• Trend towards commercialising and privatising
  airports from 1980s.
• Rationales of the states to commercialise and
  privatise airports
• To avoid the financial burden
• To improve efficiency
• To reduce the cost and maximise revenue
• To improve customer service and quality
  standard

20
21
The UK Story

• Airports Act 1986
• All airports with turnover of over one million
  pounds in two of the previous three years had to
  be transformed into limited companies.
• Part one of the Airport Act privatised seven
  airports of the British Airport Authority (BAA).
• Part two of the Airport Act commercialised 16
  municipally owned airports.

• Shares were initially being held by the original
  local authorities, but public borrowing
  restrictions prompted them to seek being
  partially or fully privatised (Graham, 2004)
• By 2000, a total of 45 airports had been set up
  as companies

21
22
Ownership
22
23
Ownership (continued)
23
Source Modified from Humphreys and Francis
(2002)
24
Over time, non-aeronautical Revenue is becoming
more and more important

• Pressure from airlines and governments to limit
  the increase in aeronautical revenue.
• The motivation of increasing non-aeronautical
  revenue to maximise total profit makes airports
  have incentives not to increase aeronautical
  revenue (Starkie, 2002).
• Therefore, airports have seen the income they
  derived from aeronautical activities decline over
  the past two decades.

24
25
Non-aeronautical revenue shares at Selected UK
Airports
25
Source Graham, 2004
26
Structure of the UK Airport Industry

• Oligopolistic market four big players (Heathrow,
  Gatwick Manchester and Stansted) control 66 of
  market share, while the other 56 airports only
  share the rest 34.
• Ownership concentration Three big airport groups
  control over 82 of market share (BAA 64
  Manchester Airport Plc 12 TBI 6).
• Geographical imbalance 5 London airports control
  59 of market.

26
27
Market Share of the UK Airports
All other airport
All other London airport
Manchester
Heathrow Gatwick
27
Source UK CAA
28
Economies of Scale

• Economies of scale in the airport industry
• Less than 300,000 WLU - average unit cost of
  15
• Between 300,000 and 2.5 million WLU average
  unit cost of 9.4
• Between 2.5 25 million WLU average unit cost
  of 8 (ICAO
• study, adopted from Graham, 2003)
• Diseconomies of scale in the airport industry
• When reaching a point, it is increasingly
  difficult and expensive for a large airport to
  design, build and operate facilities that
  co-ordinate (spatially and functionally)
  activities across and expanding area (Starkie,
  2002).
• Empirical study by Pels (2002) demonstrates that
  a number of large European airports shows this
  was the case
• In sum, regional airports have stronger
  incentives to grow in order to reach the critical
  mass of traffic.

28
29
Characteristics of Regional Airports

• Serving defined area/catchment
• Alternative to major airports
• Dependent on intra-Europe traffic
• Light financial regulation
• Vital to local economy

29
30
Challenges Facing Regional Airports

• Lack of locational advantage and characterised by
  thin traffic as a result of economic poverty,
  high unemployment, small-scale enterprises and
  sparse population in the peripheral regions where
  they are based.
• Difficult for conventional scheduled airlines to
  stimulate sufficient traffic.
• Overlapping of catchement areas.
• Abolition of intra-EU duty free sales.
• Limited ability to generate enough
  non-aeronautical revenue as a result of small
  volume of passengers.

30
31
Overview

• The Rise of Low-cost Carriers
• Regional Airports
• Low-cost Carriers and Regional Airports
• Proposed Methodology A Panel Data Analysis
• Some Preliminary Results

31
32
The Impact of LCCs on Regional Airports
Performance

• LCCs have begun to offer life to smaller airports
  and tourist destinations since they have searched
  for locations away from major congested hub
  airports and the low fares they offered have
  stimulated traffic from regional points (Barrett,
  2000).
• LCCs can make regional airports work in a way
  that traditional airlines cannot by bringing in
  passengers from
• a much wider catchment area (Barrett, 2000).

32
33
LCCs increase the catchment area for regional
airports the case of East Midlands

• Original Catchment Area Enlarged
  Catchment Area

60 minutes accounts for 87 of passengers 7.3
million people
After the entry of LCCs in 2002 the average drive
time now is 120 minutes covering some 24
million people offering huge potential for
future growth
33
Source East Midlands Airport 2004
34
LCCs Impact on Stansted Airports Financial
Performance
No. of Passenger (mil.)
(mil.)
Low cost carrier passenger
Total number of terminal scheduled passenger
Airport operating profit
34
Source UK CAA
35
Comparison of Traffic Type in Large-sized UK
Regional Airports in 2002/03 with 1995/96
Number of passenger (000)
35
Manchester
Stansted
Gatwick
Source UK CAA
36

Comparison of Traffic Type in Medium-sized UK
Regional Airports in 2002/03 with 1995/96 (1)
Number of passenger (000)
Bristol
Cardiff
Leeds Bradford
Liverpool
36
Source UK CAA
37
Comparison of Traffic Type in Medium-sized UK
Regional Airports in 2002/03 with 1995/96 (2)
Number of passenger (000)
Belfast
East Midlands
Luton
Newcastle
37
Source UK CAA
38
Comparison of Traffic Type in Medium-sized UK
Regional Airports in 2002/03 with 1995/96 (3)
Number of passenger (000)
Glasglow
38
Aberdeen
Birmingham
Edingburg
Source UK CAA
39
Comparison of Traffic Type in Small-sized UK
Regional Airports in 2002/03 with 1995/96
Number of passenger (000)
39
Teesside
Bournemouth
Exeter
Source UK CAA
40
Conflicts between LCCs and Regional Airports

• Attracting LCCs is an appealing way for airport
  management to attempt to improve their financial
  performance.
• The adoption of the cost leadership strategy and
  lack of economies of scale in the airline
  industry force LCCs constantly seeking ways to
  reduce the cost.
• For LCCs, airport charges make up a larger
  percentage of total costs. Having pushed all
  other cost inputs to minimal levels, the only
  cost variable that they can further reduce is
  that of airport charges.
• LCCs have always been successful in forcing
  regional airports revenue down by negotiating low
  charges LCCs claim that their passengers provide
  more non-aeronautical revenue sources to regional
  airports than traditional airlines, hence the
  need for airport to trade off a reduction in
  aeronautical revenue s in return for extra
  non-aeronautical revenues (Barrett, 2004).

40
41
Impact of Airport Charges on Total Operating
Costs of Intra-European Air Services for selected
airlines in 1999
41
Source Doganis (2002)
42
The Bargaining Power of LCCs over Regional
Airports

• An airport generally has most market power over
  network carriers
• connecting services, while having least
  market power over LCCs
• point-to-point operations (Starkie, 2002).
• LCCs do not rely on transfer passengers.
• LCC passengers are usually price sensitive they
  are willing to bypass the nearest airport and
  travel longer distance to the airport if the
  fares are low enough.
• Therefore, LCCs have incentives to seek out
  airports that will minimise their operating and
  station costs and then create the market around
  them.
• Consequently, LCCs have more scope for switching
  operations between different airports in order to
  reduce the costs.

42
43
Countervailing Power of Regional Airports

• Regional airports with locational advantages,
  e.g. Luton Airport.
• Airports with more than one LCCs and have more
  balanced traffic.
• Airports take proactive measures in building
  transportation links (e.g. bus services) with
  neighbouring areas which are not served by a LCC
  to build an attractive catchment area (Gillen
  Lall, 2004).
• However, there are more regional airports than
  LCCs. The competition for LCCs customs put
  regional airports in a disadvantaged position.

43
44
Crucial to Measure the impact of LCCs on Regional
Airports Financial Performance

• Relationships between LCCs and airports are
  getting tougher and tougher (Graham, 2004).
• LCCs are taking advantage of their bargaining
  power and sometimes push for a price below the
  airports marginal costs.
• However, LCCs are also risking becoming the
  victim of their own success, e.g. the case of
  Ryanair in Charleroi.
• Important to quantify the impact of LCCs and
  their passengers on regional airports financial
  performance.

44
45
Overview

• The Rise of Low-cost Carriers
• Regional Airports
• Low-cost Carriers and Regional Airports
• Proposed Methodology A Panel Data Analysis
• Some Preliminary Results

45
46
The Classical Linear Regression Model
yit ai b'iXit eit

i 1, N regions t 1,T time periods X K
vectors of non-constant explanatory variables y
Dependant variable, which varies across i regions
over t time periods a intercept and b slope
coefficients both variables are allowed to vary
across regions e the error term, which is
identical and independently distributed for all i
and t
46
47
Advantages of Using Panel Data

• By combining time series of cross-section
  observations, panel data give more informative
  data, more variability, less collinearity among
  variables, more degrees of freedom and more
  efficiency (Baltagi, 1995).
• By studying the repeated cross section of
  observations, panel data are better suited to
  study the dynamics of change.
• Panel data can better detect and measure effects
  that simply cannot be observed in pure
  cross-section or pure time series data.
• But, the choice of an appropriate model depends
  inter alia on the degree of homogeneity of the
  intercept and slope coefficients and the extent
  to which any individual cross-effects are
  correlated with the explanatory variables.

47
48
Definition of Variables (1)

• NAR ƒ(Lccpax, Charterpax, Fullservicepax)
• 
  
• Where
• NAR Non-aeronautical revenue (000) in
  airport i
• Lccpax Total number of terminal passengers
  carried by low-cost
• carriers in airport i.
• Charterpax Total number of terminal
  passengers carried by charter airlines in
  airport i.
• FullservicepaxTotal number of terminal
  passengers carried by full service scheduled
  airlines in airport i.

48
49
Definition of Variables (2)

• AR ƒ(Lccpax,
  Charterpax, Fullservicepax, Lccflights,
• ?
  
  ?
• Otherflights, Freight)
• 
  
• Where
• AR Aeronautical revenue (000) in airport
  i.
• Lccpax Total number of terminal passengers
  carried by low-cost carriers in airport i.
• Charterpax Total number of terminal
  passengers carried by charter airlines in airport
  i.
• FullservicepaxTotal number of terminal
  passengers carried by full service scheduled
  airlines in airport i.
• Lccflights Total number of flight operated
  by low-cost carriers in airport i.
• Otherflights Total number of flights
  operated by other carriers in airport i.
• Freight Total tonnage of air cargo and
  mail carried by all carriers in
• airport i

49
50
Data Source and Software

• Airport revenue data was based on annual Airport
  Statistics published by the Centre for the
  Studies of Regulated Industries, UK, ranging from
  1995/96 to 2002/03 and covering 21 regional
  airports.
• Passengers and flights data were obtained from
  the UK Civil Aviation Authority (CAA). All the
  data were adjusted to financial years, i.e. from
  the 1 April to next years 31 March.
• The software used was TSP.

50
51
Basic Empirical Models

• NARit ai ß1Lccpaxit ß2Charterpaxit
  ß3Fullservicepaxit eit
• 2. ARit ai ß1Lccpaxit ß2Charterpaxit
  ß3Fullservicepaxit ß4Lccflightsit
  ß5Otherflightsit ß6Freightit eit

51
52
Overview

• The Rise of Low-cost Carriers
• Regional Airports
• Low-cost Carriers and Regional Airports
• Proposed Methodology A Panel Data Analysis
• Some Preliminary Results

52
53
Descriptive Statistics for Model 1
Variable Mean
Std.Dev. Minimum Maximum
Cases NAR 26476.7
46741.7 1181.0 247600.0
168 Lccpax 660614.7
1772666.2 0
14538522.0 168 Charterpax
1681345.2 2777612.6 3391.0
11246100.0 168 Fullservicpax
2297914.8 3859685.9 0
20468790.0 168
53
54
Correlation Matrix for Independent Variables

• LCCP CHAP
  OTHP
• LCCP 1
  .02295 .01972
• CHAP .02295 1
  .90723
• OTHP .01972 .90723
  1

54
55
 
Estimation Results of Pooled OLS, Fixed and
Random Effects Models for Model 1
Notes Figures in parentheses are t values.
indicate significance at the 1 levels
55
 
56
Discussions (1)

• The F value of 6.584, which is less than the
  critical F value (10.996), indicates that the
  cross-sectional and time series data can be
  pooled.
• The LM value of 63.439, is greater than the
  critical value at the 1 level of significance
  it indicates that the the null hypothesis that
  the fixed effect model is appropriate is clearly
  rejected.
• The calculated Hausman statistics is 0.17 which
  is less than the critical value at the 1
  significant level. This results implies that the
  null hypothesis of no correlation cannot be
  rejected that is the model has random effects.

56
57
Discussions (2)

• All variables are significant at 1 level and
  have the
• expected signs.
• Each LCC passenger adds 4.90 to the airports
  non-
• aeronautical revenue, while each charter
  passenger and
• full service airlines passengers bring 6.73
  and 6.97
• respectively to the airport.
• Claims that LCC passengers might bring more non-
• aeronautical revenue to the airport are not
  supported.

57
58
Descriptive Statistics for Model 2
Variable Mean
Std.Dev. Minimum Maximum
Cases AR 27861.5 34423.5
819.0 144525.0 168 Lccpax 660614.7
1772666.3 0 14538522.0
168 Charterpax 1681345.2 2777612.6
3391.0 11246100.0 168 Fullservicepax
2297914.8 3859685.9 0
20468790.0 168 Lccflights 6233.5
15979.5 0 123 536.0
168 Otherflights 53358.9
54376.0 2000.0 253099.0
168 Freight 44868.0
74878.5 0 335690.0
168
58
59
Correlation Matrix for Independent Variables

• LCCP CHAP
  OTHP FRE LF OF
• LCCP 1 .02295
  .01972 .33575 .99634 .02827
• CHAP .02295 1
  .90723 .72221 .02519 .90405
• OTHP .01972 .90723 1
  .76116 .02424 .95099
• FRE .33575 .72221
  .76116 1 .35172 .72548
• LF .99634 .02519
  .02424 .35172 1 .03713
• OF .02827 .90405
  .95099 .72548 .03713 1

59
60
Estimation results of Pooled OLS, Fixed and
Random effects models for Model 2
60
Notes Figures in parentheses are t values, ,
, indicate significance at the 10, 5 and
1 levels respectively
61
Discussions (1)

• For the case (i) results, F value of 20.5, which
  is greater
• than critical F value (10.6), indicates that
  the cross-
• sectional and time series data cannot be
  pooled.
• Problems of multicollinearity between the
  passenger
• variables and flight variables.
• After dropping variables of Lccpax, Chapax and
• Fullservicepax, the data can be pooled.

61
62
Discussions (2)

• LM test and Hausman test indicate the model has
• random effect.
• Both variables of LCCflight and Otherflight are
  significant
• at 1 level and have expected signs
• Although the variable of freight is significant
  at the 10
• level the sign has wrong direction.

62
63
Summary

• The sustainable competitive advantages that a LCC
  has mainly derived from the adoption of
  point-to-point service operating from regional
  airports.
• Regional airports have strong incentives to grow
  in order to reach the critical mass of traffic.
• LCCs can make regional airports work in a way
  that traditional airlines cannot.
• However, conflicts are emerging between LCCs and
  regional airports.
• Preliminary empirical results demonstrate that
  LCC passengers have significant impact on
  regional airports non-aeronautical revenue, but
  not as big as charter and full service airlines
  passengers.

63
64
Further Research

• To re-estimate a model which can better measure
• LCCs impact on regional airports Aeronautical
  
• Revenue.
• To study the airport specific effects and time
  effects.
• To test whether the presence of Ryanair or
  easyJet at
• an airport has significant different impact on
  airports
• AR and non AR.
• To estimate a third model to aggregate AR and
  non
• AR.

64
65
Thank you for your attention. Any comments?
65