Three Short Talks
Mark Hansen
Wye River
6/7/07
1
Agenda
‰Schedule Padding Trends (for
BR)
‰LGA Gauge Analysis
‰Analysis of Changes at LGA
after “Temporary Order”
2
Flight Time Indices
‰CPI for Flight Times
‰“Market Basket” of City-Pairs
with Constant Weights
‰Flight Time from Scheduled
Departure to Actual Arrival
‰OAG Time
3
OAG and Flight Time Trend 1995-2006
145
140
130
Total
OAG
125
12 per. Mov. Avg. (Total)
12 per. Mov. Avg. (OAG)
120
Post-9/11: Schedule
stable despite
increasing actual
flight times.
Pre-9/11: Schedule
moves with actual
flight times.
115
110
Date
Jan-06
Jul-05
Jan-05
Jul-04
Jan-04
Jul-03
Jan-03
Jul-02
Jan-02
Jul-01
Jan-01
Jul-00
Jan-00
Jul-99
Jan-99
Jul-98
Jan-98
Jul-97
Jan-97
Jul-96
Jan-96
Jul-95
105
Jan-95
Time in Minutes
135
4
LGA Average Gauge Analysis
‰ Compare Seats/Flight (“Size”) and
Pax/Flight (“Load”) on LGA Flight
Segments with “Comparable” nonLGA Segments
‰ Criteria for Comparability
¾Length
¾Traffic
¾Concentration
5
Data
‰T100 February 2007
‰Airports with over 500K
outbound pax
‰Segments with over 1k pax and
Distance>0
‰2440 Observations
6
Models and Hypotheses
‰ Size=f(pax,miles,HHI,LGA)
‰ Load=g(pax,miles,HHI,LGA)
‰ Hypotheses
¾ ↑pax↑Size (scale economies and diminishing
returns from frequency)
¾ ↑miles↑Size (scale economies greater for longer
distances)
¾ ↑HHI↑Size (fragmentation)
¾ ↑LGA↓Size (babysitting+Air21)
¾ Same for Load
7
Size vs Pax-miles, 2/07
6.5
6
5.5
log(ac-size)
5
4.5
Non-LGA
LGA
4
3.5
3
2.5
2
10
12
14
16
log(pax-miles)
18
20
8
6
Load vs Pax-miles, 2/07
5.5
5
log(ac-size)
4.5
Non-LGA
LGA
4
3.5
3
2.5
2
10
12
14
16
log(pax-miles)
18
20
9
Estimation Results
‰ A 1% increase in:
¾ Pax increases size 0.3% and load 0.3%
¾ Distance increases size 0.4% and load 0.5%
¾ HHI increases size 0.2% and load 0.3%
‰ An LGA segment vs a non-LGA
segment that is otherwise identical
¾ Has 7% smaller aircraft
¾ Has 21% lower pax load
‰ Caveat: based on just one month!!
10
Changes Since 1/1/07
‰Distinctions between slots and
exemptions eliminated
‰Operating authorizations
unrestricted with respect to:
¾Community served
¾Aircraft size flown
‰Use-or-lose remains (80%)
11
Methodology
‰Compared May 2007 and May
2006 weekday flights
‰Only ASPM Individual Flight
Data Available (completed
flights only)
12
Flights by Hub Category
17%
17%
5%
5%
54%
24%
May
2006
L
M
N
S
L
M
N
S
54%
24%
May
2007
13
Flights by Airport
2000
1500
2007 Flights
ORD
1000
MDW
500
CVG
0
0
200
400
600
800
1000
2006 Flights
1200
1400
1600
1800
2000
14
Aircraft Size Distribution
100%
90%
80%
Percent Smaller
70%
60%
2006
2007
50%
40%
30%
2007: 51% of flights 70 seats or less
20%
2006: 49% of flights 70 seats or less
10%
0%
0
50
100
150
Seats
200
250
300
15
Flights by Airline
3000
2500
DAL
2007 Flights
2000
AWI
1500
1000
CHQ
500
0
0
500
1000
1500
2000
2006 Flights
2500
3000
3500
16
Conclusions
‰Scheduled block times
increased before 9/11 but not
much since
‰Aircraft size and loads are
lower at LGA (as of 2/07)
‰Little impact from temporary
order
17