Reforming the ATC System
to Promote Economic Efficiency
Dorothy Robyn
The Brattle Group
NEXTOR National Airspace System Performance Workshop
Asilomar Conference Center
Pacific Grove, CA
September 5-6, 2007
Agenda
Symptoms
Problem #1: Governance
Problem #2: Financing
Conclusions
1
1
Key Points
Flight delays, outdated technology and rising unit costs are
symptoms of the underlying problem with our ATC system
The problem itself is one of faulty incentives—a result of the current
a) governance structure and b) financing system
ATC is a 24/7 service “business” being run out of a traditional
government agency
Passenger-tax financing (together with weight-based landing fees) is
perverse, and contributes directly to flight delays
The federal government can substantially reduce delays by “pricing”
the scarcity of airways and runways. However:
o
o
FAA has failed to make the efficiency case for user fees
DOT’s Rates & Charges Policy remains major (perceived) impediment
to airports’ use of congestion pricing/slot auctions/flat charges
2
2
I. Symptoms
3
3
Symptoms—Flight Delays
• Flight delays are the most visible symptom of ATC problem
Figure 1
Flight Delays
120,000,000
100,000,000
H o urs
80,000,000
60,000,000
40,000,000
20,000,000
0
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
Year
Source: ASQP and T100 Domestic Segment
4
4
Symptoms—Flight Delays
Flight delays are a drag on the economy
o
o
Cost to airlines
Cost to pax
$7.7B (2006)
$3B-$13+B
Published delay statistics understate true costs
o
o
o
Flight cancellations
Missed connections
Airline schedule-padding (see Fig. 2)
5
5
Symptoms—Flight Delays
Figure 2
Increase in Flight Travel Times
18
Air Time
16
Ground Time
14
12
10
8
6
4
2
0
1977
1979
1981
1983
1985
1987
1989
1991
1993
1995
1997
1999
2001
2003
2005
Year
Source: Steven A. Morrison and Clifford Winston, “Delayed! U.S. Aviation Infrastructure Policy at a Crossroads,” in
Aviation Infrastructure Performance: A Study of Comparative Political Economy, edited by Winston and Gines de Rus
(Brookings Press, forthcoming).
6
6
Symptoms—Flight Delays
Proximate cause of increase in delays is decline in aircraft size
Figure 3
Average Seats per Aircraft in the U.S. Commercial Domestic Fleet
140
120
100
80
60
40
20
0
1998
1999
2000
2001
2002
2003
2004
2005
2006
Source: Department of Transportation, Federal Aviation Administration, FAA Aerospace
Forecasts, Fiscal Years 2004-2015 and FAA Aerospace Forecasts, Fiscal Years 2007-2020.
7
7
Symptoms—Flight Delays
Figure 4
Change in Average Seat Size by Airport – All Departures
200
180
160
-15.5%
140
120
-12.7%
-0.1%
-6.8%
-11.4%
9.9%
100
-18.9%
-4.2%
-10.4%
80
60
40
20
0
ATL
DFW
EWR
IAD
IAH
July 2002
8
JFK
LGA
ORD
PHL
July 2007
8
Symptoms—Flight Delays
As a result of downgauging, the increase in departures has exceeded the
increase in passengers
Figure 5
Domestic Scheduled Air Traffic Departures & Passenger Traffic
12.0
700
Passengers
(Right scale)
600
10.0
500
400
Depatures
(Left scale)
6.0
300
(Millions)
(Millions)
8.0
4.0
200
2.0
100
0.0
Year
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
Source: Department of Transportation, Bureau of Transportation Statistics, U.S. Air Carrier Traffic Statistics
9
9
Symptoms—Outdated Technology
A second symptom is our outdated ATC technology
Aging technology limits capacity and raises costs
o
o
o
Imprecise radar wide separation of aircraft
Terrestrial navaids inefficient air routes
Aging equipment expensive maintenance
Chronic problems with modernization:
“Most FAA modernization projects have a record of (1) promising
more capability than they ultimately deliver, (2) being completed
later than promised, and (3) costing far more by the time they are
completed than the initial cost estimates.”
“Reforming the Federal Aviation Administration: Lessons from Canada and the United Kingdom,” by
Clinton V. Oster, Jr., with the assistance of John S. Strong, IBM Center for the Business of
Government (2006), p. 14
10
10
Symptoms—Flat Productivity & Rising Unit Costs
• Symptom #3 is the FAA/ATO’s flat productivity and rising unit
costs
Figure 6
Controller Productivity & ATC Unit Costs
160
3,500
140
3,000
2,500
100
2,000
Expense Per Instrument Operation
80
1,500
60
1,000
Instrument Operations per Controller
120
40
500
20
0
19
83
19
84
19
85
19
86
19
87
19
88
19
89
19
90
19
91
19
92
19
93
19
94
19
95
19
96
19
97
19
98
19
99
20
00
20
01
20
02
20
03
20
04
20
05
20
06
0
19
82
Expense per Instrument Operation ($2006)
Instrument Operations Per Controller
Fiscal Year
11
11
II. Problem #1—Governance
12
12
Problem #1—Governance
As a traditional government agency, FAA is not wellsuited to operating a 24/7, high-tech, service “business”
FAA faces more complex incentives than does an
ordinary business
o
Who is the customer?
o
Federal budget rules and constraints
o
Russ Chew’s three “value chains”
Congress as de facto customer
Pressures to overstate benefits and understate costs of large
investment projects
Lack of technical oversight exacerbates principal-agent problems
Incentives to maximize output
13
13
Economists’ Rx: Move ATO out of FAA
Most economists favor moving ATO out of FAA
o
o
Efficiency rationale
Safety rationale
Less agreement on where to put it, since it remains a monopoly
o
o
o
Independent government corporation (e.g., USATS)
User-owned cooperative (NavCanada)
Private corporation with rate regulation (NATS)
Key question: Where and to what extent can system take advantage
of competition?
o
Current examples
o
Contract towers
Flight Service Stations
ADS-B Contract
What’s possible longer term? History of network industries: technology
creates potential for competition in what was once a natural monopoly.
14
14
III. Problem #2—Financing
15
15
Problem #2—Financing
Tax financing directly encourages overuse of scarce ATC
capacity
o
o
Airlines pay only indirectly
Small aircraft contribute less than large aircraft, even though cost
to serve them is same
DEN-PHX:
o
A320 contributes $1498
50-seat RJ contributes $502
Most important, it ignores congestion costs
Business jets contribute far less than the burden they impose
DEN-PHX:
6-pax Gulfstream pays only $133
Weight-based landing fees compound problem
16
16
Problem #2—Financing
Basing charges on ticket value/aircraft weight made sense
at one time
o
o
Form of Ramsey pricing
Appropriate way to cover costs absent congestion
But now that approach is perverse—and it contributes
directly to delays
Size should not matter!
17
17
Problem #2—Financing
Tax financing encourages inefficiency on FAA’s part as
well
o
o
ATO lacks basic customer feedback
No incentive to provide value-added services because it can’t
charge for them
Tax financing impedes investment
o
o
Tax revenues don’t signal where/how much investment is needed
Investment gets shortchanged because it’s on-budget
18
18
Economists’ Rx: Efficient (Marginal Cost) Pricing
A shift from taxes to efficient prices (cost-based user
fees) could have far-reaching benefits:
o
o
o
More efficient use of ATC system—fewer delays (allocative
efficiency)
More efficient provision of ATC (productive efficiency)
More efficient investment in innovation and new technology
(investment efficiency)
19
19
Economists’ Rx: Prices and Allocative Efficiency
In short run, prices would lead to more efficient use of
scarce ATC capacity
Ideally, fees would vary by time of day/season to reflect
delay costs
But there’s case for starting with a flat, per-flight charge
o
o
Congestion is due in part to current, perverse financing
mechanism
A flat charge would reduce congestion by eliminating bias
toward small a/c
In short, first the cake (flat fee); then the icing
(congestion charge)
20
20
Economists’ Rx: Prices and Productive Efficiency
Prices would encourage more efficient provision of ATC
services by FAA
o
o
Users can compare value to cost even where choices are limited
FAA could compare costs and revenues to decide which services
(or facilities) to expand or contract
FAA could offer—and users could buy—value-added
services:
o
o
An airline could pay more to get priority in system (or pay less
for reduced service)
Several airlines could buy extra services during peak hours
Adoption of fee-for-service by FAA’s Logistics Center
had powerful impact on producer (and user) behavior
o
o
Paying customers provided more direct feedback
Logistics Center took the customer feedback more seriously
21
21
Economists’ Rx: Prices and Investment Efficiency
With FAA moving to NextGen, most important benefit of
prices may be to guide investment and facilitate
technology adoption
o
o
Market signal as to what users are willing to pay for
Key tool to help FAA overcome equipage problem
NavCanada charges less for aircraft equipped with datalink
Non-price options (e.g., offering preferred airspace) can achieve
same goal but with greater difficulty
More generally, prices could facilitate deals with and
between user groups
With prices/user fees, FAA would have option of private
borrowing
22
22
IV. Conclusions
23
23
Conclusions
Flight delays, outdated technology and rising costs are the symptoms,
not the problem
The underlying problem is:
o
o
A governance structure that forces the ATO to deal with multiple “value
chains”
A financing system that guarantees overuse of the system because users
don’t pay their real cost
No easy governance fix as long as ATC is a monopoly, although
incremental steps to increase user input and accountability are key
By contrast, financing fix is straightforward—at least intellectually:
o
o
o
o
Moving to efficient prices for ATC is the single most important thing
the federal government could do to promote aviation
Prices would reduce delays and facilitate investment
Prices would even address governance by creating direct link between
users and ATO
Prices are equitable: users pay their costs—no more, no less
24
24
Conclusions, cont.
In sum, efficiency case for pricing ATC—i.e., the case for
introducing economic incentives—is powerful:
o
“Moral exhortation doesn’t change people’s behavior. Prices do.”
(Austan Goolsbee)
Why, then, has no one made that case for user fees as part of the
FAA reauthorization debate? In a summer when delays have
become a national nightmare, why has the FAA not argued for user
fees and congestion pricing as the immediate response?
Instead, FAA’s case for user fees has been all about cost
recover/revenue adequacy:
o
o
o
We need NextGen to meet expected demand.
To finance NextGen, we need a stable, predictable revenue source.
User fees are stable and predictable.
Similarly, ATA has ignored efficiency; its case for user fees has been
all about equity—making business jets pay their fair share
25
25
Conclusions, cont.
Maybe it’s not too late
Airport congestion pricing may be an easier sell than user
fees
o
o
o
Airports have the authority
Public officials and press “get” congestion pricing
There are ways to address airline opposition
Time is right: the federal government can substantially
reduce delays – and do so before next summer
Remember: “A crisis is a terrible thing to waste.”
26
26