Biography for William SwanChief Economist, Seabury-Airline Planning Group. Visiting Professor, Cranfield University. Retired Chief Economist for Boeing Commercial Aircraft 1996-2005 Previous to Boeing, worked at American Airlines in Operations Research and Strategic Planning and United Airlines in Research and Development. Areas of work included Yield Management, Fleet Planning, Aircraft Routing, and Crew Scheduling. Also worked for Hull Trading, a major market maker in stock index options, and on the staff at MIT’s Flight Transportation Lab. Education: Master’s, Engineer’s Degree, and Ph. D. at MIT. Bachelor of Science in Aeronautical Engineering at Princeton. (bill.swan@cyberswans.com)

© Scott Adams

Airline Evolution

William M Swan

Chief Economist

Boeing Commercial Airplanes, Marketing; Retired

Spring 2007

Structure is Destiny

• Structure of costs across airplane sizes

• Structure of fares and reservations

• Structure of route networks and hubs

The Airplanesare Amazingly Similar

• 707– Prototype 707 in 1954– Advanced 707-320

• Seating 189, charter configuration• Speed 600 mph nominal• Altitude 36,000 ft• Range: Transcontinental• Wing 146 ft, length 152 ft, body width 12 ft

• 737– 3rd generation family -600, -700, -800, -900– Largest are -800/-900

• Seating 189/215, charter configuration• Speed 530 mph actual• Altitude 35,000 ft• Range: Transcontinental• Wing 113ft, length 130 ft/138 ft, body width 12 ft

The Significant Changes

• Jets now come from 70-550 seats in size– 100-400 seats if you want to use Boeings– 70-350 seats if you want to use 2 engines

• Ranges now cross the Pacific– How long will people sit?– The world is round – limits useful range

Big Airplanes are Cheaper per SeatConventional Representation (Confusing)

$0

$10

$20

$30

$40

$50

$60

$70

$80

$90

0 100 200 300 400 500 600

Seats

Tri

p C

os

t

Underlying Linear RelationshipWell-Adjusted Presentation (Clear)

$0

$5,000

$10,000

$15,000

$20,000

$25,000

0 100 200 300 400 500 600

Seats

Tri

p C

os

t

Big Airplanes Make You Wait(Cost with Frequency Value Included)

$-

$40.00

$80.00

$120.00

0 100 200 300 400 500 600

Seats

Tri

p C

os

t

Wait Time Cost

Airplane Seat Cost

System Summed Cost

Concepts to Keep

• The denser the route, the cheaper the seats• Not the Same as bigger airline, wider network

– No indication that extensive networks are cheap

• Not the Same as longer flight distance– However, longer the distances are cheaper per Km

• Economies of Airplane Size have Persisted since jet airplanes:

– MIT study in 1971– AA/UA fleet planning 1986– Boeing Study 2001

Ticket Prices

• Yield has declined 2-3%/year since 1971– Representing a 1% annual decline in fares– Further decline due to change in ticket mix– Yield is “cents per kilometer”

• Two kinds of fares– Advance purchase, discount fares– Regular, unrestricted, full fares

• Low Cost Carrier (LCC) pricing– Erosion of full fare levels– Less than meets the eye

0.06

0.07

0.08

0.09

0.10

0.11

0.12

1975 1980 1985 1990 1995 2000 2005

Inte

rna

tio

na

l Y

ield

, 2

00

3$

/Km

Estimated Fares

trend at -2.4%/year

reported yields

Prices, Fares, and Yields

Fare Regulation in US• Why Regulate Fares?

– Economies of Density• Average cost per seat > Marginal cost per seat• Natural monopoly – at least when network is thin

– Mentality of only one (“full”) price• No discount fares, airlines for premium travel only

• How fares were regulated (US case)– Yield (cents per km) fixed

• Independent of range (but longer distances are cheaper per km)• Independent of market density (but denser markets are cheaper per seat)

– Set to cover average costs including return on investment• Consequences of regulating fares this way

– Long haul was immensely profitable– Large markets were immensely profitable– Low value trips were not offered low value tickets (few discounts)– Load factor 50-55% range (today 70%+)

Deregulation

• US deregulation 1978– End of restrictions on starting new nonstops– End of fare regulations

• 1977 Regulated snapshot:– Only ATL and ORD were hubs– JFK gateway for most Europe flights– Regional carriers feeding majors at hub cities – Interline (between airlines) connections– Limited 30% discount advance purchase fares– Fares proportional to distance, no “boarding” cost– Fares independent of market size, no “small” cost

First Response to Deregulation

• Airlines added new nonstop routes– Bleeding traffic off old connecting legs– Reducing head-to-head competition– Making networks thinner but with more links– Filling out hubs

• Prices went up in small, short markets– It took a while unlearn “long, big” paradigm– Smaller communities gained services– Hubs began to develop– Regional carriers merged with majors

• They were always loosing money before, anyway

Evolution of Routes & Networks

• Origin-to-Destination (O&D) flows small– Few pairs big enough for local only service

• Need to combine flows to build size– Get to at least 100 seats per departure– Best layout turns out to be coordinated hubs

• Three Stages of Hubs1. Natural gateways, minimum spanning trees2. Competitive hubs, banked connections3. Continuous hubbing

Most Markets are Small

0%

2%

4%

6%

8%

10%

12%

14%

16%

<3.12

5

<6.25

<12.5

<25 <50<10

0<20

0<40

0<80

0

<1600

1600

+

Passengers per Day One Way

Sh

are

of

RP

Ks Too Small For

Nonstop

Half of Travel is in Connecting Markets

0%

2%

4%

6%

8%

10%

12%

14%

16%

O&D Passengers per Day

Sh

are

of

Wo

rld

RP

Ks

Connecting Markets

Nonstop Markets

Lots of O&D Connections

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

300100

0200

0300

0400

0500

0600

0700

0800

0900

0

10000

11000

12000

St. Mi. Range Block (excludes US domestic O&Ds)

Sh

are

of

O&

D P

as

se

ng

ers

4-leg connect

Double Connect

1-connect

thru

nonstop

Connecting Share of Loads Averages about 50%

0%

10%

20%

30%

40%

50%

60%

70%

80%

0 2000 4000 6000 8000 10000Flight Distance (Km)

Lo

ca

l T

raff

ic S

ha

re o

f O

nb

oa

rd

Network Evolution:Airlines Hate To Compete

• Avoid head-to-head competition– Preferred airline wins big

• First choice on all high-fare traffic– Higher yields

• First choice for whatever low-fare travel is going– Full on off-peak days means higher load factor

– Unstable head-to-head competition– Natural Monopoly

• New Routes = “get your own monopoly”

Networks Develop from Skeletal to ConnectedHigh growth does not persist at initial gateway hubs

Early developments build loads to use larger airplanes:Larger airplanes at this state means middle-sized

Result is a thin network – few links

A focus on a few major hubs or gateways

In Operations Research terms, a “minimum spanning tree”

Later developments bypass initial hubs:Bypass saves the costs of connections

Bypass establishes secondary hubs

New competing carriers bypass hubs dominated by incumbents

Large markets peak early, then fade in importance

Third stage may be non-hubbed low-cost carriers: The largest flows can sustain service without connecting feed

High frequencies create good connections without hub plan

Skeletal Networks Develop Links to Secondary Hubs

Early Skeletal Network

Later Development bypasses Early Hubs

ORD

ATL

DFW

DEN

JFK

LAX

MIA

SFO

Regional and Gateway Hubs in US

Hub Concepts

• Hub city should be a major regional center– Connect-only hubs have not succeeded

• Early Gateway Hubs get Bypassed– Traffic builds early, stays flat in later years

• Later hubs duplicate and compete with early hubs– Many of the same cities served– Which medium cities become hubs is arbitrary– Often better-run airport or airline determines success– Also the hub that starts first stays ahead

Largest Routes are Not Growing as bypass flying diverts traffic

-20%

-10%

0%

10%

20%

30%

40%

50%

60%

ASK growth Frequencygrowth

Airplane sizegrowth

World, 1993-2003

Top 100 Routes

ORD

ATL

DFW

DEN

JFK

LAX

MIA

SFO

Many Secondary Hubs in US

STL

SLC

CVG

PHX

IAH

MSP

DTW PITEWR

SEA

Competition is Rising in Regional Flying

0

10

20

30

40

50

60

1970 1975 1980 1985 1990 1995 2000 2005Her

finad

ahl N

umbe

r of C

ompe

titor

s

N.AmericaAsiaEuropeOther Short

Competition is Rising in Long-Haul

0

5

10

15

20

25

30

1970 1975 1980 1985 1990 1995 2000 2005

Her

finad

ahl N

umbe

r of C

ompe

titor

s

AtlanticPacificAsia-EuropeOther Long

Examples of the 3 Kinds of Hubs

• International hubs driven by long-haul– Gateway cities– Many European hubs: CDG, LHR, AMS, FRA– Some evolving interior hubs, such as Chicago– Typically 2 banks of connections per day – one in, one out

• Regional hubs connecting smaller cities– Most US hubs, with at least 3 banks per day (each way)– Some European hubs, with 1 or 2 banks per day

• High-Density hubs without banking– Continuous connections from continuous arrivals and departures– American Airlines at Chicago and Dallas– Southwest at many of its focus cities

Continuous Hubs

• AA had 12 banks a day at DFW & ORD• Revised so airplanes turn in 25 minutes• Passengers connect in 40-120 minutes• Higher aircraft and gate utilization• Nearly the same connect times as banked• AA connects 50%, and lives by it• WN connects 33%, and tops up with it• Ryanair connects 15%, and fights it

Why Secondary Hubs?Airlines Hate Competition

• Avoid “head-to-head” whenever possible– Preferred carrier wins big

• Gets first choice of premium fare demand• Gets full loads during off peaks• Leaves 2nd choice carrier low yield, high peaking

– Result: Lots of new routes

130

140

150

160

170

180

190

200

210

220

230

1970 1975 1980 1985 1990 1995 2000 2005 2010

Sea

ts P

er A

irpl

ane

1990 FORECAST

2004 data1990 data

US Deregulation

Forecasters in 1990 Were Confused

1.0

1.5

2.0

2.5

3.0

3.5

1970 1975 1980 1985 1990 1995 2000 2005

Dai

ly D

epar

ture

s p

er N

on

sto

p P

air,

ave

rag

e

Nonstop Pairs (index)

Departures/Pair

What We Missed: New Routes

Minot Connects to the World

18:00 Bank Gives Minot 38 DestinationsInbound Bank Outbound Bank

Origin Depart Hub Origin Depart Hub Hub Arrive Destin' Hub Arrive Destin'city time time city time time ==> time time city time time cityONT 1200 1727 DLH 1655 1748 1848 2116 MBSBOS 1505 1728 SAN 1210 1748 1849 2136 CMHSNA 1200 1728 IND 1604 1749 1850 2227 HPNPSP 1210 1729 TUL 1550 1750 1850 2130 AZOPDX 1210 1729 DTW 1700 1753 ==> 1835 2030 MEM 1850 2130 AZOMSO 1355 1730 GRB 1641 1755 ==> 1836 1932 FAR 1850 2215 TYSCWA 1630 1731 MKE 1635 1756 ==> 1837 2159 IAD 1850 900 LGWGFK 1620 1731 SJC 1215 1756 ==> 1838 2159 RDU 1851 2142 DTWRST 1650 1732 RAP 1530 1757 ==> 1839 2209 PVD 1852 2128 FNTSMF 1205 1732 DTW 1705 1759 ==> 1839 2214 GSO 1853 2217 BWIORD 1600 1734 DSM 1650 1759 ==> 1840 2207 BDL 1854 2246 BOSDFW 1510 1735 MSN 1645 1800 ==> 1841 2108 GRR 1855 2255 ORFYEG 1355 1735 MOT 1635 1800 ==> 1842 2139 BUF 1855 2008 MLIYYC 1357 1735 SFO 1220 1800 ==> 1843 2104 OKC 1855 2124 LANABQ 1405 1739 BOI 1415 1804 ==> 1844 2210 ATL 1856 2126 DFWLNK 1615 1740 GEG 1312 1804 ==> 1845 2159 ROC 1857 2158 YYZDCA 1559 1741 ATL 1620 1805 ==> 1845 2022 SBN 1858 2007 GRBSTL 1600 1742 MDW 1635 1809 ==> 1845 2134 DAY 1859 2002 OMALAX 1215 1744 CVG 1655 1809 ==> 1846 2208 CLT 1900 2200 PITYWG 1618 1744 CWA 1715 1815 ==> 1847 2208 DCA 1900 2027 ORDBIS 1630 1747 1847 2253 TPA 1901 2030 MCI

The One Horse in a “One-Horse Town”

Industry Growth is Small Markets

• Virtuous Circle:– Better services: More Value

• Faster connections (add 15% demand for online)• Fewer Stops (add 15% for each lost stop)• Higher frequencies (add 15% for full-day schedule)

– Lower Costs: Lower Prices• Higher traffic volumes mean lower costs• Competitive choices eliminate monopoly pricing

• New “small” markets get new services– Smaller towns, secondary city airports– Grow network from “below”

The First Big Event Nobody Noticed(Deregulation: 1984)

• Peoples Express opened a low cost hub– At Newark (EWR) airport, New York City– Cheap fares, lousy service

• AA discovered PE – Became aware of the extent of PE connects– Responded by matching PE fares

• 70% off full fare (compared to 35% off for SSave)• Capacity only available midweek• AA clearly the preferred choice at matched fares

Results of Big Event• PE went out of business

– Due to “horrendous peaking of traffic”– No midweek loads

• AA found it was making more money– 80% average weekly load factors (not 60%)– Filling previously empty mid-week seats– Selling tickets for half previous discount fares– Revenue Management controlling sales

• Paradigm shift:– Old way was set fares, get load factor

• Weak demand means lower load factor– New way was set load factor, sell to fill

• Weak demand means lower average fare

Results of New Fare Structure

• Marginal seats sold at marginal cost– Breaking “single price” mentality– Economically efficient– Allows full cost recovery from multiple prices

• High full fares pay cost of frequency• Low controlled fares get marginal capacity• Saturday stay, advance purchase discriminate

Southwest and LCCsNew Airlines from Deregulation

• 19 out of 20 start-ups failed• WN (Southwest) succeeded• America West (hubbed) survived (+AirTran)• LCCs had 20% cost advantage from labor• WN had “shuttle technology”

– Engineered for loading and unloading– Reliability from high frequency– Incidental connections high (30%)– Business airline: not “cheap,” Just good– Good employee relations; reasonable wages

The 2nd Big Event Nobody Noticed(Deregulation in 1998)

• Airlines were paying $3/segment booking fees– Computer reservations systems owned by AA, UA– Travel agents hooked to mainframes

• Agents got 8-15% booking fees• Agents got bribes to sell AA, UA, DL….dominant networks

• Southwest refused to pay fee– Was thrown off reservations systems– Continued to sell on internet– No drop in Southwest business– No one noticed

• Majors’ Res systems no longer in control

Consequences of 2nd Event

• Majors became able to reduce commissions– Travel agencies no longer had pricing power– Removed one high-cost part from trip

• Start ups no longer had to pay majors– Previous Res System profits greater than airline’s– Majors owned Res Systems– Majors no longer controlled cost of entry– Majors lost full information about competitor’s prices

• Later consequence: Competitive Pricing– Direct-to-airline bookings made prices hard to monitor– Internet intermediaries compared multiple airlines sites– Cost of information on prices greatly reduced– Now only 18 out of 20 start-ups fail = twice the successes– Majors unable to extract rents to pay pilots’ premiums

The LCC/HCC War• Airline Industry is forever young

– Birth and death process• 38% of ASK service 20 years back, airline gone• 28% of today’s ASKs with new airlines• Index of competition flat to rising

• HCCs have adapted– Labor rates down: wages, rules, retirement– Service quality, costs, and prices down

• LCCs will migrate services– Higher quality: boarding, onboard, reliability– More connections at higher prices– More price differentiation– Higher connecting share

• Who can tell which is which?

Cost Reductions Keep Coming

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

1970 1980 1990 2000 2010 2020 2030

Inte

rna

tio

na

l Y

ield

, 2

00

3$

/Mi

Estimated Fares -1%/yrYield trend

-2.4%/year

reported yields

Better Airplanes, Higher Bypass, Hotter Turbines

Jets for Props

Revenue Management, Higher Load

Facors

Distributiion costs, Lower commissions

LCCs,nonunion pilots,

faster turns?

Lower Airport Costs,

Lower ATC costs?

Future Cost Improvements

Two Choices:1. Regulated Airlines

• Few routes, larger airplanes• Focus on inelastic business demand• Monopoly prices and costs• Permanent Names

2. Competitive Markets• Many routes, smaller airplanes• Innovation, adaptation• Competitive prices and costs• Bankruptcies and Start-ups• Biggest names still survive

Evolution: Part 2Birth and Death

• 38% of the air travel 20 years ago– Was flown by carriers that do not exist today

• 28% of the air travel today– Is flown by carriers that did not exist 20 years ago

• Competition is greater now– By any reasonable technical measure– But only slightly greater. Almost unchanged

• Conclusion: A healthy industry requires– Failure of badly run airlines– Failure of most new start-up airlines– Success of some new start-up airlines

• Overall employment and services should grow

Mergers That Work

• When airlines serve the same airports– Their merger will be a success– Merged airline offers better connections– Better, more valuable service to customers– May eliminate some competition

• Mostly this is a short-haul carrier merging with a long-haul carrier

More Mergers that Work

• Failing airlines are acquired– Majority of employees retained– Majority of airplanes retained– All airports’ still served– Management employees of failing airline gone

• This process is good– Reduces operations in a orderly fashion– Maintains the most service and employees– Avoids losses associated with bankruptcy

Mergers that Seldom Work

• Merging airlines to expand network reach– Overlap of airports only at edges– Makes bigger airline– Does not improve many connections

• Most such mergers have not worked– Difficulties with employee cooperation– Little increase in value or saving in cost– Tendency to retain bad practices

Bankruptcy: How Airlines Fail

• Government airlines do not fail– They just need money, over and over

• Regulated airlines seldom fail– They just don’t improve services– They also may not improve costs

• Competitive airlines do fail– Efficiency comes from eliminating the bad ones– In the “Profit and Loss” system

• The losses are the important thing• A loss comes when it costs more to run the airline

– Than the customers are willing to value the service• Bankruptcy is the way to stop doing things that are losses

Bankruptcy: Social Details

• In the US– Bankrupt means not enough money to pay

• wages; leases; loans; owners of airline– Owners give up all their invested money– Part of loans are given up (“haircut”)– Lease payments may be reduced

• Or airplanes taken away– Wages may be reduced– Airline gets smaller– BUT IT CONTINUES TO EXIST

• After 3rd Bankruptcy and reorganization– Airline may stop operating

Bankruptcy: Social Details

• In Europe– Bankruptcy is more disgraceful– Airline is shut down– Airplanes and airport gates are sold– Employees all loose their jobs– Similar to 3rd bankruptcy of US airline

• Stock holders loose all their money• Bond holders loose almost all their money• Airplanes and employees try to find new airlines

The Hard Problem

• A healthy industry means some airlines fail• Failure is hard on employees• Failure may reduce services• How to make transition smooth

– Most employees get jobs at new airline– Most airplanes are put back to use – Most services are kept operating

• No country has “ideal” Government Policies – Either regulate to avoid failure– Or allow messy bankruptcy

• Arguments about who looses how much money• No incentives to make smooth transitions

• Be the first: Do it “better”

William Swan:

Data Troll

Story Teller

Economist