Rand Mg696

NAT I O NA L DE F E N SE R E SE A R CH
I NS T I T U T E a nd P R OJ E CT AIR F OR CE
THE ARTS
This PDF document was made available
CHILD POLICY
from www.rand.org as a public service of
CIVIL JUSTICE
EDUCATION
ENERGY AND ENVIRONMENT
HEALTH AND HEALTH CARE
INTERNATIONAL AFFAIRS
NATIONAL SECURITY
POPULATION AND AGING
PUBLIC SAFETY
SCIENCE AND TECHNOLOGY
SUBSTANCE ABUSE
TERRORISM AND
HOMELAND SECURITY
TRANSPORTATION AND
INFRASTRUCTURE
WORKFORCE AND WORKPLACE
the RAND Corporation.

Jump down to document6
The RAND Corporation is a nonprofit

research organization providing
objective analysis and effective
solutions that address the challenges
facing the public and private sectors
around the world.
Support RAND
Purchase this document
Browse Books & Publications
Make a charitable contribution
For More Information

Visit RAND at www.rand.org
Explore RAND National Defense Research Institute

RAND Project AIR FORCE
View document details

Limited Electronic Distribution Rights
This document and trademark(s) contained herein are protected by law as indicated
in a notice appearing later in this work. This electronic representation of RAND
intellectual property is provided for non-commercial use only. Unauthorized
posting of RAND PDFs to a non-RAND Web site is prohibited. RAND PDFs are
protected under copyright law. Permission is required from RAND to reproduce,
or reuse in another form, any of our research documents for commercial use. For
information on reprint and linking permissions, please see RAND Permissions.
This product is part of the RAND Corporation monograph series.

RAND monographs present major research findings that address the
challenges facing the public and private sectors. All RAND monographs undergo rigorous peer review to ensure high standards for
research quality and objectivity.
8IZ)BTUIF$PTU
PG'JYFE8JOH
"JSDSBGU3JTFO
".BDSPTDPQJD&YBNJOBUJPOPGUIF
5SFOETJO64.JMJUBSZ
"JSDSBGU$PTUTPWFSUIF1BTU
4FWFSBM%FDBEFT
.BSL7"SFOBt0CBJE:PVOPTTJt,FWJO#SBODBUP
*SW#MJDLTUFJOt$MJPSE"(SBNNJDI
Prepared for the

United States Navy and United States Air Force
Approved for public release;
distribution unlimited
NAT IONAL DE FENS E RES EA RC H I NS TI TUTE and

PR OJ ECT AI R FORC E
The research was conducted within the Acquisition and Technology

Policy Center of the R AND National Defense Research Institute
(NDRI) and the Resource Management Program of RAND Project
AIR FORCE (PAF). Both NDRI and PAF are federally funded research
and development centers sponsored by the Office of the Secretary
of Defense, the Joint Staff, the Unified Combatant Commands, the
Department of the Navy, the Marine Corps, the defense agencies,
and the defense Intelligence Community under contracts W74V8H06-C-0002 and PFA7014-06-0001.
Library of Congress Cataloging-in-Publication Data

Why has the cost of fixed-wing aircraft risen? : a macroscopic examination of the trends
in U.S. military aircraft costs over the past several decades / Mark V. Arena ... [et al.].
p. cm.
Includes bibliographical references.
ISBN 978-0-8330-4312-2 (pbk. : alk. paper)
1. Airplanes, MilitaryUnited StatesCosts. 2. United States. Air Force
Procurement. 3. United States. NavyProcurement. I. Arena, Mark V.
UG1243.W54 2008
358.4'183dc22
2008026145
The RAND Corporation is a nonprofit research organization providing

objective analysis and effective solutions that address the challenges
facing the public and private sectors around the world. R ANDs
publications do not necessarily reflect the opinions of its research clients
and sponsors.
R is a registered trademark.
Copyright 2008 RAND Corporation
All rights reserved. No part of this book may be reproduced in any

form by any electronic or mechanical means (including photocopying,
recording, or information storage and retrieval) without permission in
writing from RAND.
Published 2008 by the RAND Corporation
1776 Main Street, P.O. Box 2138, Santa Monica, CA 90407-2138
1200 South Hayes Street, Arlington, VA 22202-5050
4570 Fifth Avenue, Suite 600, Pittsburgh, PA 15213-2665
RAND URL: http://www.rand.org
To order RAND documents or to obtain additional information, contact
Distribution Services: Telephone: (310) 451-7002;
Fax: (310) 451-6915; Email: [email protected]
Preface
In recent decades, cost escalation for military xed-wing aircraft of all

types has exceeded that of commonly used ination indices, including
the Consumer Price Index, the Department of Defense procurement
deator, and the Gross Domestic Product deator.1 A relatively xed
investment budget (albeit one with cyclical variations) means that the
Services must somehow accommodate higher unit costs. This accommodation may mean buying fewer aircraft than in the past or it may
mean reprioritizing budgets between acquisition and operations and
support.
This monograph explores the causes of this unit cost escalation,
including both economy-driven factors that the Services cannot control and customer-driven factors that they can.
The research was conducted between January 2006 and September 2007 and was jointly sponsored by the Assessment Division, Oce
of the Chief of Naval Operations (OPNAV N81) and by the Principal
Deputy, Oce of the Assistant Secretary of the Air Force (Acquisition), Lt Gen Donald Homan, SAF/AQ, and Blaise Durante, SAF/
AQX.
The research was conducted within the Acquisition and Technology Policy Center of the RAND National Defense Research Institute (NDRI) and the Resource Management Program of RAND Project AIR FORCE (PAF). Both NDRI and PAF are federally funded
research and development centers sponsored by the Oce of the Sec-
This study exclusively examines manned aircraft and data. Unmanned aerial vehicles
(UAVs) are excluded from the analysis.
iii
iv
Why Has the Cost of Fixed-Wing Aircraft Risen?
retary of Defense, the Joint Sta, the Unied Combatant Commands,

the Department of the Navy, the Marine Corps, the defense agencies,
and the defense Intelligence Community.
For more information on RANDs Acquisition and Technology
Policy Center, contact the Director, Philip Antn. He can be reached by
email at [email protected]; by phone at 310-393-0411, extension
7798; or by mail at the RAND Corporation, 1776 Main Street, Santa
Monica, California 90407-2138. More information about RAND is
available at http://www.rand.org.
Project AIR FORCE, a division of the RAND Corporation, is
the U.S. Air Forces federally funded research and development center
for studies and analyses. PAF provides the Air Force with independent
analyses of policy alternatives aecting the development, employment,
combat readiness, and support of current and future aerospace forces.
Research is conducted in four programs: Aerospace Force Development; Manpower, Personnel, and Training; Resource Management;
and Strategy and Doctrine. Additional information about PAF is available on our Web site: http://www.rand.org/paf/
RAND Project AIR FORCE reports that address military aircraft cost estimating issues include the following:
t In An Overview of Acquisition Reform Cost Savings Estimates, MR1329-AF, Mark A. Lorell and John C. Graser use relevant literature and interviews to determine whether estimates of the ecacy
of acquisition reform measures are robust enough to be of predictive value.
t In Military Airframe Acquisition Costs: The Eects of Lean Manufacturing, MR-1325-AF, Cynthia R. Cook and John C. Graser
examine the package of new tools and techniques known as lean
production to determine whether it would enable aircraft manufacturers to produce new weapon systems at costs below those
predicted by historical cost-estimating models.
t In Military Airframe Costs: The Eects of Advanced Materials and
Manufacturing Processes, MR-1370-AF, Obaid Younossi, Michael
Kennedy, and John C. Graser examine cost-estimating methodologies and focus on military airframe materials and manufactur-
Preface
t
t
t
t
t
ing processes. The authors provide cost estimators with factors

useful in adjusting and creating estimates based on parametric
cost-estimating methods.
In Military Jet Engine Acquisition: Technology Basics and CostEstimating Methodology, MR-1596-AF, Obaid Younossi, Mark V.
Arena, Richard M. Moore, Mark A. Lorell, Joanna Mason, and
John C. Graser present a new methodology for estimating military jet engine costs and discuss the technical parameters that
derive the engine development schedule, development cost, and
production costs and present a quantitative analysis of historical
data on engine development schedule and cost.
In Test and Evaluation Trends and Costs for Aircraft and Guided
Weapons, MG-109-AF, Bernard Fox, Michael Boito, John C.
Graser, and Obaid Younossi examine the eects of changes in the
test and evaluation (T&E) process used to evaluate military aircraft and air-launched guided weapons during their development
programs. The report also provides relationships for developing
estimates of T&E costs for future programs.
In Software Cost Estimation and Sizing Methods, Issues and Guidelines, MG-269-AF, Shari Lawrence Peeger, Felicia Wu, and
Rosalind Lewis recommend an approach to improve the utility of
the software cost estimates by exposing uncertainty and reducing
risks associated with developing estimates.
In Lessons Learned from the F/A-22 and F/A-18E/F Development
Programs, MG-276-AF, Obaid Younossi, David E. Stem, Mark A.
Lorell, and Frances M. Lussier evaluate historical cost, schedule,
and technical information from the development of the F/A-22
and F/A-18E/F programs to derive lessons for the Air Force and
other Services to improve the acquisition of future systems.
In Price-Based Acquisition: Issues and Challenges for Defense Department Procurement of Weapon Systems, MG-337-AF, Mark A. Lorell,
John C. Graser, and Cynthia R. Cook document savings and cost
avoidance on government and contractor activities resulting from
the use of price-based acquisition strategies in a manner useful
to the acquisition, planning, and cost-estimating communities,
and generate recommendations for approaches to more accurately
vi
t
t
t
t
t
assessing the potential cost savings and cost avoidance that can be
expected from the wider use of price-based acquisition.
In Impossible Certainty: Cost Risk Analysis for Air Force Systems,
MG-415-AF, Mark V. Arena, Obaid Younossi, Lionel A. Galway,
Bernard Fox, John C. Graser, Jerry M. Sollinger, Felicia Wu, and
Carolyn Wong describe various ways to estimate cost risk and
recommend attributes of a cost-risk estimation policy for the Air
Force.
In Systems Engineering and Program Management: Trends and
Costs for Aircraft and Guided Weapons Programs, MG-413-AF,
David E. Stem, Michael Boito, and Obaid Younossi evaluate
the historical trends and develop a cost-estimating method for
systems engineering and program management, which is one of
the most costly below-the-line items for military aircraft and
guided weapon systems.
In Evolutionary Acquisition: Implementation Challenges for Defense
Space Programs, MG-431-AF, Mark A. Lorell, Julia F. Lowell, and
Obaid Younossi study how to help the Air Force acquisition community formulate policies that anticipate and respond to the prospect of more widespread use of evolutionary acquisition strategies
relying on a spiral development process, as recently mandated by
the Oce of the Secretary of Defense.
In Historical Cost Growth of Completed Weapon System Programs,
TR-343-AF, Mark V. Arena, Robert S. Leonard, Sheila E. Murray,
and Obaid Younossi conduct a literature review of cost growth
studies and provide a more extensive analysis of the historical cost
growth of the completed acquisition programs.
In Is Weapon System Cost Growth Increasing? A Quantitative Assessment of Completed and Ongoing Programs, MG-588-AF, Obaid
Younossi, Mark V. Arena, Robert S. Leonard, Charles Robert
Roll, Jr., Arvind Jain, and Jerry M. Sollinger analyze completed
and ongoing weapon system programs development cost growth
and determine the magnitude of cost growth and show cost
growth trends for the past three decades.
Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix
Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxi
CHAPTER ONE
The Escalation of Aircraft Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

CHAPTER TWO
Data and Price Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

Data Sources and Normalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Sources of Data and Their Content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Technical and Schedule Databases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Data Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Adjustments and Normalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Final Dataset and Systems Represented . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Measuring Cost Escalation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
CHAPTER THREE
Economy-Driven Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Distribution of Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Labor Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
vii
viii
Material and Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Fees and Prots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
General and Administrative Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Material Overhead . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Fees and Prots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Notional Aircraft Comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
CHAPTER FOUR
Customer-Driven Factors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
Quantity Eects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
Cost Improvement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Procurement Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Conguration Eects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Basic Technical Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
Other Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
CHAPTER FIVE
Pairwise Comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Economy-Driven Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
Customer-Driven Factors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
Total Escalation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
CHAPTER SIX
Industry Views on Military Fixed-Wing Aircraft Cost Escalation . . . . . 57

Military Fixed-Wing Aircraft Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Increased Military Utility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
Stealth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
Weight Reduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
Lean Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
Government Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Berry Amendment and Buy American Legislation . . . . . . . . . . . . . . . . . . . . . . . 62
OSHA and Environmental Regulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Antitamper Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
International Trade in Arms Regulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Contents
ix
CHAPTER SEVEN
Options for the Air Force and the Navy to Reduce Fixed-Wing
Aircraft Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Make Fixed-Wing Aircraft Procurement More Stable and Predictable . . . . 68
Stabilize Project Management and Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
Rethink Competition Within the Industrial Base . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
Encourage International Competition and Participation . . . . . . . . . . . . . . . . . . . . 72
Improve the Process of Formulating Requirements and Capabilities . . . . . . 72
Focus Attention on Upgrades and Commercial Derivatives . . . . . . . . . . . . . . . . . 73
Increase the Use of Evolutionary Acquisition Principles . . . . . . . . . . . . . . . . . . . . . 74
Lessons Learned from the F-22A and F/A-18E/F Development
Programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
CHAPTER EIGHT
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
APPENDIX
A. Aircraft Included in the Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

B. Survey of Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Figures
S.1. Contributors to Price Escalation from the F-15A (1975)

to the F-22A (2005) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii
1.1. Cyclical Defense Procurement Outlays, Between Fiscal
Years 1960 and 2008 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2. Annual Quantity of Aircraft Procured, 1974 to 2005 . . . . . . . . . . . . . 3
2.1. Average Unit Procurement Costs for Fighter Aircraft, by
Fiscal Year, 1974 to 2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2. Average Unit Procurement Costs for Electronic Aircraft,
by Fiscal Year, 1974 to 2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.1. Labor, Equipment, and Material Percentages of Weapon
System Cost, 1969 to 2003 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.2. Average Annual Growth for Aerospace Labor Costs,
1989 to 2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.3. Annual Direct and Burdened Labor Costs from a Select
Set of Aircraft Programs, Fiscal Years 1969 to 2003 . . . . . . . . . . . . . 22
3.4. Percentage of Labor Hours by Subcomponent, 1969
to 2003 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.5. Aerospace Labor Productivity, 1987 to 2003 . . . . . . . . . . . . . . . . . . . . . 24
3.6. Material and Equipment Price Escalation, 1986 to 2004 . . . . . . . 26
3.7. G&A Percentage of Total Cost, by Fiscal Year, 1969 to
2003 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.8. Material Overhead Percentage, by Fiscal Year, 1969 to
2003 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.9. Operating Margin for Aircraft Sector, by Year, 1975 to
2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
6.1. Trend in Composite Material Use in Aircraft, 1967 to
2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
xi
Tables
2.1.
2.2.
3.1.
3.2.
3.3.
3.4.
4.1.
4.2.
4.3.
4.4.
4.5.
4.6.
4.7.
5.1.
Average Annual Cost Escalation for Aircraft and Ination

Indices, 1974 to 2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Average Annual Escalation Rate for Unit Procurement and
Flyaway Costs for Various Navy Aircraft, 1974 to 2000 . . . . . . . . 13
Average Distribution of Labor, Equipment, and Material
Costs for a Select Group of Fixed-Wing Aircraft, 1969 to
2003 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Direct and Indirect Annual Escalation Rates for Labor
Subcomponents, 1969 to 2003 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
Material and Equipment Escalation Rates, 1986 to 2004 . . . . . 27
Contributions of the Economic Factors to Cost Escalation
for a Notional Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Cost Improvement Slopes, by Minimum Number of
Annual Buys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Cost Improvement and Production Rate Slopes, by
Minimum Number of Annual Buys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Cost Improvement and Production Rate Slopes with a
Minimum of Five Annual Buys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Aircraft Models Long-Term Production Proles and
Dates of Conguration Change. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
Results of Regression Incorporating Conguration
Change Eects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
Results of Regressions on Technical Characteristics. . . . . . . . . . . . . 45
Results of Regressions on Airframe Materials Complexity . . . . 46
Percentage Contributions to Annual Cost Escalation, by
Economy-Driven Factors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
xiii
xiv
5.2.
5.3.
A.1.
Percentage Contributions to Annual Cost Escalation,

by Customer-Driven Factors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
Percentage Contributions to Annual Escalation Rate . . . . . . . . . . 54
Aircraft Systems and Types Used in the Analyses . . . . . . . . . . . . . . 84
Summary
As with many other military weapon systems, military aircraft have

experienced long-term, unit cost increases that are greater than the rate
of ination.1 These increases, largely driven by the desire for greater
capabilities, appear likely to persist and could have dire implications
for aircraft inventories, particularly given relatively xed defense
investment budgets. Commenting on the continually increasing costs
for aircraft, one industry executive (Augustine, 1986, p. 143) famously
wrote,
In the year 2054, the entire defense budget will purchase just one
aircraft. The aircraft will have to be shared by the Air Force and
Navy 3 days per week except for leap year, when it will be made
available to the Marines for the extra day.
Given increasing costs for military aircraft, relatively xed budgets to procure them, and resulting decreased procurement rates, the
Air Force and the Navy asked RAND to examine the causes of military
aircraft cost escalation. From available data, we calculated cost escalation rates as well as their economy-driven and customer-driven
causes.
For every type of aircraft we examinedpatrol, cargo, trainer,
bomber, attack, ghter, and electronic warfareannual unit cost esca1
Throughout this document, we use the terms price and cost interchangeably. Formally, in
most cases we are referring not to cost but to what cost estimators term as price, that is, the
actual dollars required to buy the system (including all fees and prots). By cost increase (or
cost escalation), we mean the dierences in actual prices paid for aircraft over time and not
the dierence between the estimated and actual values.
xv
xvi
lation rates in the past quarter century have exceeded common ination indices, such as the Consumer Price Index, the Department of
Defense procurement deator, and the Gross Domestic Product deator. This trend is true whether cost escalation is measured using either
procurement or yaway cost. Patterns of cost escalation diered by
aircraftsome showed cost improvement over time, while others
steadily increasedbut, again, all exceeded that for other ination
measures.
We considered two groups of contributors to cost escalation:
economy-driven variables, which include costs for labor, equipment,
and material, and customer-driven variables, which include costs for
providing performance characteristics that the Services want in their
aircraft.
We found that the rates ($/hr) of aircraft manufacturing labor,
in both direct and fully burdened wages, have increased much faster
than other measures of ination. Nevertheless, increased productivity
has meant that overall, labor costs have grown only slightly faster than
ination. Furthermore, the proportion of labor cost in the overall cost
of aircraft has been steadily decreasing (from a prime contractor perspective) as more manufacturing is outsourced. With two exceptions
(specialty metals and avionics systems, such as navigation equipment),
materials and equipment used in aircraft manufacturing have increased
in cost at roughly the same rate as other measures of ination. Altogether, we nd that labor, material, equipment, and manufacturer fees
and prots have helped increase the cost of aircraft about 3.5 percent
annuallywhich is less than the rate of increase for some ination
indices during the same time.
The government can aect the cost of military aircraft in several
ways, particularly through the quantity it demands and the characteristics it species. Although we did not nd a consistent cost improvement
eect stemming from purchases over time in aircraft procurement, we
did nd a procurement rate eect by which higher production rates
helped reduce unit prices. One reason for this may be the economic
leverage from larger purchases that allows manufacturers to invest in
eciency improvements. Other possible reasons are the spreading of
xed overhead costs over more unitsthus reducing average unit price.
Summary
xvii
Another explanation could be more ecient use of labor and tooling

when production rates are higher.
When considering comparison pairs of aircraft, we found that
complexity of the aircraft (performance characteristics and airframe
material) contributed to aircraft cost escalation, often at rates far exceeding those of ination. Figure S.1 shows the contributions of the various factors to cost escalation when comparing an F-15A (1975) to an
F-22A (2005). The chart shows that roughly a third of the overall cost
escalation is due to economy-driven factors. The remainder is due to
customer-driven onesmainly system complexity.
Interviews that we conducted with representatives of aircraft manufacturers conrmed many of these ndings. In particular, these representatives noted that the increased demand for greater aircraft stealth
and reduced aircraft weight contributed to cost escalation. They also
cited government regulations, such as those designed to protect American industry and technology and those for environmental protection
and occupational health as sources of aircraft cost escalation.
Figure S.1
Contributors to Price Escalation from the F-15A (1975) to the F-22A (2005)
12
Annual escalation rate, %
10
Regulatory
Production rate and
learning
8
6
Customer-driven
factors
Complexity
General and
administrative
Labor
Economy-driven
factors
Equipment
Material
0
RAND MG696-S.1
xviii
The Services could choose to address cost escalation in several

ways, some more feasible than others. Improved procurement stability and longer-term contracts could encourage manufacturers to make
investments to increase eciency and cut costs. Fewer change orders
to aircraft may help reduce costs as well. International competition
and participation in the construction of military aircraft could also
reduce costs, although this would likely be opposed by Congress and
might be feasible only for noncombat aircraft. Focusing on aircraft
upgrades in successive model improvements rather than on acquisition
of new aircraft types, as has been done for several aircraft (e.g., the F/A18E/F), could help contain procurement cost escalation, although the
age of some existing aircraft may limit the application of this practice.
At present, the Air Force and the Navy appear to be opting for
fewer aircraft but with the highest technological capabilities. Such a
strategy helps ensure that U.S. aircraft remain far superior to those of
any other military in the world. Maintaining such capabilities, however,
does have a cost. We do not evaluate whether this particular tradeo
is good or bad. We note only that it exists and point out related issues
that the Services will have to address in deciding how to allocate future
appropriations for aircraft procurement.
Acknowledgments
Many individuals contributed to this study and we would like to thank

them. First, we thank Trip Barber of OPNAV N81 for both cosponsoring this study and providing very useful input and guidance along the
way. We would also like to thank Blaise Durante (SAF/AQX) for also
cosponsoring this project and for his long-term support of Project AIR
FORCEs cost research.
We are grateful to the U.S. manufacturers and their parent
organizationsBoeing, Lockheed Martin Corporation, and Northrop
Grumman Corporationfor their time and insight. Particularly, we
thank Laura Russell and Ken Goeddel (Boeing); Richard Janda, Larry
McQuien, and Doug Steen (Lockheed Martin); and Chris Bowie
(Northrop Grumman) for coordinating our interactions with industry
and providing helpful data for the study. We are also grateful to David
Burgess, Donald Allen, and Soroja Raman of NAVAIR for helpful
technical discussions and data on Navy xed-wing programs. Richard
Hartley (SAF/FMC) and Jay Jordan from the Air Force Cost Analysis
Agency provided similar help and guidance for Air Force systems.
Finally, we acknowledge both reviewers of this document: John
Graser (RAND) and Bill Stranges (former NCAD and NAVAIR ocial). Their comments and suggestions greatly improved this work.
xix
Abbreviations
ACEA
APN
B&P
BLS
CAD
CAIV
CAM
CBO
CC
CCDR
CDRL
CFE
CI
CIC
CPI
CTOL
CV
DCARC
DoD
ECI
EVM
EW
Arms Control Export Act

Aircraft Procurement Navy
bid and proposal
Bureau of Labor Statistics
computer-aided design
Cost as an Independent Variable
computer-aided manufacturing
Congressional Budget Oce
conguration change
Contractor Cost Data Report
Contract Data Requirements List
contractor-furnished equipment
cost improvement
cost improvement curve
Consumer Price Index
conventional take o and landing
carrier version
Defense Cost and Resource Center
Department of Defense
Employment Cost Index
earned value management
empty weight
xxi
xxii
FY
G&A
GDP
GFE
HAPCA
IPT
IR&D
ITAR
JSF
LM
MMA
NASA
NATO
NAVAIR
NGC
OPNAV
OSD
OSHA
PA&E
PPI
PR
QC
R&D
RAM
SAR
SEC
STOVL
UAV
USAF
USN
scal year
general and administrative
Gross Domestic Product
government-furnished equipment
Historical Aircraft Procurement Cost Archive
integrated product team
internal research and development
International Trade in Arms Regulations
Joint Strike Fighter
Lockheed Martin
multi-mission maritime aircraft
National Aeronautics and Space Administration
North Atlantic Treaty Organization
Naval Air Systems Command
Northrop Grumman Corporation
Oce of the Chief of Naval Operations
Oce of the Secretary of Defense
Occupational Safety and Health Administration
Program Analysis and Evaluation (OSD)
Producer Price Index
procurement rate
quality control
research and development
radar absorbing materials
Selected Acquisition Report
Securities and Exchange Commission
short take o and vertical landing
unmanned aerial vehicle
U.S. Air Force
U.S. Navy
CHAPTER ONE
The Escalation of Aircraft Costs
Commenting on the continually increasing costs of military aircraft,

Norman Augustine (1986, p. 143) famously observed,
In the year 2054, the entire defense budget will purchase just one
aircraft. This aircraft will have to be shared by the Air Force and
Navy 3 days per week except for leap year, when it will be made
available to the Marines for the extra day.
Augustine based this prediction on costs for individual aircraft that

had grown by a factor of four every decade, with increases more closely
related to the passage of time than to modications in speed, weight,
or technical specications.
The trends that Augustine observed have persisted across time and
weapon systems. Unit costs for weapons have typically grown at least
5 percent annually, with those for advanced weapon systems such as
tactical aircraft growing 10 percent annually (Hellerman, 2003; Kirkpatrick, 2004). Among recent aircraft series, for example, the unit costs
for the F-15 increased from $11.9 million in 1974 (as measured in thenyear dollars) to $54.0 million in 2000.1 Eskew (2000) observed that
the real cost escalation (beyond ination and performance growth) for
military aircraft is about 3 percent per year.
Although design improvements may explain some increases, it is
remarkable that other advances have not helped minimize them. As
1
Expressing this dierence in constant 2006 dollars, the trend is $44.0 million in 1974
versus $58.6 million in 2000.
Dov Zakheim (2005), the former Under Secretary of Defense (Comptroller), noted, such
ndings are not easy to fathom. One might have thought that
more ecient production methods, including computer aided
design and manufacturing, microminiaturization of components,
and the employment of greater computing power, all would have
reduced costs or at least held them level.
These trends, as Augustine would note, have dire implications for

the number of aircraft the U.S. Air Force (USAF) and the U.S. Navy
(USN) can procure. One way to demonstrate this is through the numbers of aircraft that the Department of Defense (DoD) has been able
to procure through recent budget cycles.
In recent decades, defense procurement spending has been cyclical, uctuating since the mid-1960s between $60 billion and $130
billion in constant dollars (Figure 1.1) (Oce of Under Secretary of
Defense, 2007b). After peaking during the late 1960s, outlays decreased
Figure 1.1
Cyclical Defense Procurement Outlays, Between Fiscal Years 1960 and 2008
Procurement outlays (FY08 $B)
140
120
100
80
60
40
20
0
1960
1970
1980
1990
Fiscal year
RAND MG696-1.1
2000
2010
The Escalation of Aircraft Costs
through the early 1970s, increased through the mid-1980s, and

decreased following the end of the Cold War and the rst Gulf War
through the late 1990s. They have increased since then, because of the
Global War on Terror operations, but are expected to decrease again in
the near future.
During this same time, the number of aircraft that DoD has
purchased has cycled with the procurement budget (Figure 1.2), but
with an overall downward trend. For example, when total outlays
troughed in 1975, DoD procured 193 ghter aircraft and 391 total
aircraft. When outlays peaked in 1985, DoD procured 300 ghter aircraft and 509 total aircraft. When they troughed again in 1995, DoD
procured only 24 ghter aircraft and 101 total aircraft. In 2005, when
outlays peaked again, it procured 66 ghter aircraft and 231 total aircraft, or roughly half what it procured annually in the trough of the
mid-1970s and roughly a third what it procured annually during the
last peak of the mid-1980s.
Figure 1.2
Annual Quantity of Aircraft Procured, 1974 to 2005
700
Number of units
600
500
All
400
300
Fighters
200
100
0
1970
1975
1980
1985
1990
Fiscal year
RAND MG696-1.2
1995
2000
2005
2010
To be sure, other variables, such as changing threats and missions, can aect the procurement of aircraft and their composition in
any given period of time. Yet the escalating cost of aircraft and the
downward cycle of procurement rates raise issues about the number of
aircraft DoD will ultimately be able to procure and operate.
The Navy has faced similar issues in procuring ships. Since the
mid-1960s, the cost of ships has increased from 100 to 400 percent
(Clark, 2005; Arena et al., 2006a).
Concerns over these trends led the Navy and Air Force to ask
RAND to address sources of cost escalation in procuring military
xed-wing aircraft. The issues we address are:
t How does escalation in aircraft costs compare with cost escalation in other sectors of the economy?
t What are the sources of any escalation in these costs?
t Can cost escalation be reduced or minimized?
In the next chapter, we examine some measures of cost escalation and their trends. In Chapter Three, we examine economy-driven
sources of cost escalation, or those associated with labor, material, and
equipment, over which the Services have little control. In Chapter
Four, we examine customer-driven sources of cost escalation, including how quantities ordered, changing congurations, and desired technical characteristics all aect aircraft costs. Chapter Five oers some
pairwise comparisons of how both economy-driven and customerdriven characteristics have contributed to aircraft cost escalation. We
present views of some industry representatives on the sources of cost
escalation in Chapter Six. Next, we discuss some options for addressing
cost escalation in Chapter Seven. Finally, we present our conclusions
in Chapter Eight.
CHAPTER TWO
Data and Price Trends
Data Sources and Normalization

Military xed-wing aircraft systems dier widely by size, speed, range,
weight, airframe material composition, length of production run, production rate, and costs. Dierences within aircraft mission (or class)
attack, bomber, cargo, electronic, ghter, patrol, and trainercan be
considerable as well.1 A single aircraft design can have blocks or series
that dier considerably. Even within the same block or series, costs can
increase from tail to tail as newer technology is gradually introduced
on the production line. In this monograph, we distinguish between
aircraft type (e.g., F/A-18) and aircraft series (e.g., F/A-18C/D versus
F/A-18E/F) but not block congurations (because of data limitations).
When assessing cost at the annual buy level, we use a single set of technical characteristics for all aircraft bought in that scal year.
To begin addressing aircraft cost issues, we review three topics in
this chapter. First, we examine available data sources, including their
limitations. Second, we discuss how to measure cost escalation. Third,
we assess how cost escalation for aircraft compares with other measures of cost ination such as the Consumer Price Index (CPI), the
DoD procurement deator, and the Gross Domestic Product (GDP)
deator.
1
DoD has a standard nomenclature for its aircraft using the mission/design/series convention. For example, the F/A-18E/F means it is a ghter/attack mission aircraft, the 18th
in the series of aircraft of that mission designated by DoD, and it is the 5th and 6th series
within that mission and design.
Sources of Data and Their Content
With one exception,2 we analyze total budgeted system cost for aircraft
throughout this document. These costs are labeled Gross P-1 in budget
documents.3 In addition to airframe, propulsion, and avionics costs
(usually referred to as recurring yaway), Gross P-1 includes belowthe-line elements: support equipment, training equipment, publications, and technical data.
We developed an annual cost and quantity database using three
primary sources of information: the Historical Aircraft Procurement
Cost Archive (HAPCA), a Congressional Budget Oce (CBO) (1992)
study which documented cost and quantity data for all the military
services between 1974 and 1994, and P-1 budget documents. HAPCA,
developed by the Naval Air Systems Command (NAVAIR), contains
procurement data (cost and quantity) for aircraft systems procured by
the Navy from the late 1940s to 2000, including subsystem and belowthe-line elements. HAPCA does not contain Navy procurement cost
data for any aircraft past 2000.4 For the Air Force and more recent
Navy programs, we therefore compiled comparable total-system-level
cost data using a combination of the CBO and P-1 budget data. The
resulting overall cost database covered the years 1974 through 2006 for
all Air Force and Navy xed-wing procurements that were not classied. All three data sources were fairly consistent at the top-line level,
In Table 2.2, we explore the dierence in price escalation for procurement versus yaway
costs.
Given the complexity of aircraft and the technologies involved, certain parts may have to
be ordered earlier than other parts of the aircraft to have everything ready to meet the nal
assembly schedule. Recognizing this, Congress often authorizes and appropriates funds for
these long-lead items in a scal year before the funds needed for the rest of the aircraft in
that annual buy. These funds are entitled advance procurement funds and are shown on
the budget documents as such. Although Net P-1 accrues advance procurement funds to
the budget year that the funds were allocated, Gross P-1 accrues advance procurement to
the annual buys that the funds are used to purchase.
HAPCA data included only estimates for scal year 2000. We replaced these with actual
cost data in our database.
which was not surprising given that the HAPCA and CBO dataset
were built from the original P-1 documents.5
Technical and Schedule Databases
To understand the causes of cost escalation, we needed a database with

detailed technical characteristics for each aircraft model. HAPCA contains data on performance and weightincluding cruising and maximum speed, empty and maximum weight, avionics weight, combat
radius, engine thrust, and materials composition6 for both Navy
and Air Force aircraft. However, much of the technical information
HAPCA contains is incomplete and does not document the sources of
information or the underlying assumptions such as operating conditions or maximum speed evaluated.
We expanded these data with gures in published documents,
proprietary-source documents, and publicly available databases. Published documents include a NASA history on modern aircraft (Loftin,
1985), Janes All the Worlds Aircraft (Janes Information Group, annual),
RAND publications (Large, Campbell, and Cates, 1976; Dryden, Britt,
and Binnings-DePriester, 1981; Resetar, Rogers, and Hess, 1991), and
other government-sponsored research (Groemping and Noah, 1977;
Heatherman, 1983). Other documents referenced for technical specications include NAVAIRs Standard Aircraft Characteristics, and
Beltramo et al. (1977). Proprietary-source documents included information acquired from contractor internal documents and presentations. Publicly available databases include Air Force and Navy current
and historical factsheets and those of enthusiast associations.7
5
The P-1 database quantity information for USN matched but the overall cost numbers
were sometimes o by as much as 1 percent because of rounding.
The materials composition database contains the percentage of airframe structure that is
aluminum, steel, titanium, composite, or other.
Air Force factsheets include those available at http://www.af.mil/factsheets/ and http://

www.nationalmuseum.af.mil/factsheets/. Navy factsheets include those available at http://
www.navy.mil/navydata/fact.asp and http://www.history.navy.mil/branches/org4-20.htm.
Enthusiast data are available at http://www.aero-web.org, http://www.aerospaceweb.org/,
and http://www.spacey.net/airplane/. All these Web sites were accessed February 9, 2007.
Data Limitations
We note some limitations to our cost and technical data and analyses.
First, we have limited the analysis to the total cost level. Although the
HAPCA database has a subelement breakout of cost, governmentfurnished equipment (GFE) is added to airframe. This prevented us
from analyzing cost escalation at a subsystem level (e.g., avionics and
airframe) as we originally intended because in many cases large portions of the subsystem costs are GFE. Second, although we are concerned only with production costs, the early scal year buys are likely
to overlap with some research and development dollars as well, leading
to a potential understatement of procurement costs for the rst few
lots. Third, because our technical characteristics database is compiled
from several secondary sources, it is only approximate, representing
average or consensus gures, and not necessarily the results of physical validation or testing. Fourth, procurement quantities include only
those purchased by USN and USAF. Large foreign procurement of
similar variants of some aircraft systems could have cost consequences,
such as those we examine below in which larger total quantities (including foreign sales) can help reduce unit costs.
Adjustments and Normalization
Where appropriate,8 we adjusted costs for ination by using the standard Navy aircraft procurement (APN) deator (Oce of Budget,
2004) to a scal year 2006 basis. We also considered several other
deators in the rst stage of our analysis, but none had substantial
eects on regression coecients or other numerical results.9
Final Dataset and Systems Represented
Our work is multistage, with each stage considering a dierent number

of systems. We considered all aircraft for which we had data in our
analysis of cost escalation trends. In estimating cost improvement and
8
This normalization process was used for our analysis of customer-driven factors described
in Chapter Four. Elsewhere in the document, we use nonnormalized values.
That is, none of the other deators resulted in changes of more than 1 in the second signicant digit in our cost improvement or production rate coecient estimates.
production rate eects on aircraft costs, we assessed only those systems that met certain statistical criteria (described below). In estimating the eects of technical characteristics on aircraft costs, we included
all systems for which we had a complete set of cost and technical data,
regardless of the number of annual buys. We could not nd airframe
materials data for all the aircraft used for the technical characteristics analysis, so our material complexity analysis is based on a further
reduced set of aircraft. Appendix A lists all systems assessed in each
part of our analysis.
Measuring Cost Escalation

We focus on long-term changes in price,10 or what we call cost escalation. We use this term to describe the general changes in price, typically
for a similar item or quantity, between periods of time. It is important
to distinguish escalation from growth. Cost growth is the dierence
between actual and estimated costs. It reects how well we can predict
the cost of a future system. We are not concerned with the quality of
aircraft price predictions; rather, we are studying how the actual price
for an aircraft changed as time passed.11
We quantify the escalation in terms of annual growth rates. We
chose this approach to minimize distortions caused by examining
trends over diering periods of time. If we were to examine the simple
increase in price (i.e., nal to initial cost), our results would depend
on the amount of time between the two values being compared. In
general, longer periods of time would lead to greater price increases.
By calculating annual growth rates, we normalize cost increases to a
common baseline.
Algebraically, we dene annual cost growth as
10
Throughout this document, we use the terms price and cost interchangeably. Formally, in
most cases we are referring not to cost but to what cost estimators term as price, that is, the
actual dollars required to buy the system (including all fees and prots).
11
For examples of cost growth on defense weapon systems, see Arena et al. (2006b).
10
rate =
( Year2 Year1 ) Cost

2
Cost1
1 ,
(2.1)
where
t Cost2 is the cost at Year2, and
t Cost1 is the cost at Year1.
That is, the annual growth rate is a compound function in which
year-to-year increases accumulate. If, for example, Cost 2 is $5 and Year2
is 2004, and Cost1 is $4 and Year1 is 1998, then the resulting annual
growth rate for cost may be calculated as 3.8 percent.12 For cases where
we have more than two observations, we use optimized least squares
regression to calculate an annualized growth rate. The regression
approach ts the natural logarithm of cost (the dependent variable) as a
function of the scal year (the independent variable). The annual rate is
the exponential of the coecient for the scal year, minus one.
Trends
The rst issue we address is how the long-term cost growth for xedwing aircraft compares with other measures of ination. Table 2.1
shows the annual escalation rate in the unit procurement cost13 for
various types of xed-wing aircraft as well as common measures of
ination including the CPI, the DoD procurement deator, and
12
Mathematically, the terms in this example are, Year2 Year1 = 2004 1998 = 6 and Cost 2/
Cost1 = 5/4 = 1.25. The sixth root of 1.25 is approximately 1.0379; subtracting one from this
gives an annual growth rate of 0.0379, or approximately 3.8 percent.
13
We examine unit procurement cost trends as we have the most complete set of cost data
with respect to timeframe. However, below we address the trends in terms of recurring yaway costs as well. As the name implies, yaway costs are costs that directly lead to specic
aircraft units (e.g., hardware, change orders, GFE, and management). Procurement costs
encompass all yaway costs and those indirect costs not associated with a specic aircraft
unit, such as spare parts, data, contractor support, and training equipment, but are necessary
to operate and maintain the eet.
11
Table 2.1
Average Annual Cost Escalation for Aircraft and
Ination Indices, 1974 to 2005
Aircraft Type
Average Annual
Rate, %
Patrol
11.6
Cargo
10.8
Trainer
9.1
Bomber
8.4
Attack
8.3
Fighter
7.6
Electronic
6.7
Ination Index
Average Ination
Rate, %
CPI
4.3
DoD procurement deator
3.8
GDP deator
3.7
the GDP deator. These growth rates represent the rates of increase
between 1974 and 2005, unadjusted for ination. We determined
these escalation rates using the regression approach described above.
Appendix A lists all aircraft included in each aircraft type.
By type, cost escalation for aircraft in the past quarter-century has
varied from about 7 to 12 percent. This rate of escalation is similar to
that seen in Navy ships since 1965 (Arena et al., 2006a). The long-term
escalation rate has also been greater than that for common measures of
ination. Even the rate of increase for electronic warfare aircraft, with
the lowest rate of increase of the types listed above, was above that of
other ination indices.
The ordering of aircraft from highest to lowest rate of increase is
noteworthy. Surprisingly, patrol aircraft top the list, with an annual
cost growth rate more than double that for any ination measure. One
might have anticipated that more technically advanced systems, such
as ghters and attack aircraft, would have the highest rates. The rate for
patrol aircraft is a result of the limited duration of the P-3 program
which dominates the trend for this type. This program ran from scal
12
year 1974 to scal year 1987, a period in which ination indices ranged
from 6.2 to 7.3 percent. This partially accounts for the higher rate of
escalation seen in costs for patrol aircraft than seen for other aircraft
produced in times of lower ination. Cargo aircraft had the second
highest rate of increase, also more than double that for any of the
common measures of ination we note. This, too, is somewhat surprising, given that such aircraft tend to have less complexity, including
fewer mission systems and fewer requirements for avionics and weapons. One reason for the high rate of increase for cargo aircraft may be
the signicant increases made to their capability (e.g., payload rate,
range, speed). We will explore such changes in subsequent chapters.
One question that arose during our early evaluation of the data
was whether unit procurement costs might misrepresent overall trends
in cost escalation. To assess this possibility, we compare, in Table 2.2,
the unit procurement and yaway cost14 trends from 1974 to 2000 using
HAPCA data (which contain Navy aircraft only, as discussed above).
Although the rates of procurement cost escalation in the HAPCA data
diered from those in the P-1 data,15 we found little dierence between
the rates of increase for procurement or yaway costs in the HAPCA
data. This similarity in escalation rates for the two dierent costs suggests
that our results are not biased by using P-1 rather than unit yaway cost.
The rates of escalation were not uniform (or monotonically
increasing) over the 30-year period. Figure 2.1 plots the average unit
procurement cost by scal year for various ghter aircraft models.
Some aircraft, such as F-18E/F and F-22A, show a traditional cost
improvement trend. In the initial years of procurement of these systems, there is a higher average unit cost that decreases exponentially
14
Flyaway costs are generally considered to be more representative of the true hardware
cost because they exclude such items as support, initial spares, and other contractor support
services that might dier greatly by system.
15
We found that trainer aircraft, for example, have a procurement cost growth rate of 13.8
percent annually in the HAPCA data, as shown in Table 2.2, but of only 9.1 percent in the
P-1 data, as shown in Table 2.1. The dierence results because the P-1 data include a broader
set of systems, such as the T-34 and the Joint Primary Aircraft Training System, most with
lower rates of cost increase, that are not in the HAPCA data.
13
Table 2.2
Average Annual Escalation Rate for Unit Procurement
and Flyaway Costs for Various Navy Aircraft,
1974 to 2000
Aircraft Type
Procurement, %
Flyaway, %
Trainer
13.8
14.1
Cargo
13.2
13.0
Patrol
11.2
9.9
Attack
8.2
8.4
Electronic
7.7
7.5
Fighter
6.5
6.2
Figure 2.1
Average Unit Procurement Costs for Fighter Aircraft, by Fiscal Year,
1974 to 2005
Average unit costs (then-year $M)
1,000
100
F/A-18E/F
F/A-18A/B/C/D
F-14A/D
F-15A/B/C/D/E
F-16A/B/C/D
F-5E/F
F-22
10
1
1970
1975
1980
1985
1990
1995
2000
2005
2010
Fiscal year
RAND MG696-2.1
in subsequent years, and eventually levels o. Other aircraft, such as

F-14A/D, display a steady increase in average unit cost in subsequent
scal years, as do electronic aircraft, as shown in Figure 2.2. Still other
14
Figure 2.2
Average Unit Procurement Costs for Electronic Aircraft, by Fiscal Year,
1974 to 2005
Average unit costs (then-year $M)
1,000
100
E-2C
E-3A
E-8B
E-8C
RC-12
TR-1/U2
EA-6B
10
1
1970
1975
1980
1985
1990
1995
2000
2005
2010
Fiscal year
RAND MG696-2.2
aircraft, such as F-16 and F-15, show a mixed pattern with plateaus
whose duration are likely related to model changes.
In contrast to the dierent patterns of cost growth for aircraft,
those observed for naval ships in earlier research (Arena et al., 2006a)
generally followed the traditional cost improvement pattern. This suggests that xed-wing aircraft, in general, are more subject to modication and upgrade over the life of a program. In fact, it is not uncommon for an aircraft to have several planned upgrades over its production
life. These upgrades are typically driven by changes in requirements,
obsolescence issues, or the need to mitigate risk by deliberately incorporating new technologies later in production. For example, the F-16
grew in capability over its production run. Its aircraft empty weight
grew from 15,600 to 19,200 pounds between Block 10 (F-16A/B) and
Block 50 (F-16C/D), its engine was upgraded from an F-100-PW-200
to either an F-110-GE-129 or an F-100-PW-229, and it had numerous
other upgrades in its avionics and mission systems.
15
Ideally, we would next explore the sources of escalation within a

single aircraft program at a more detailed level, such as the airframe,
propulsion, or avionics. For example, in earlier analyses of naval ship
cost escalation, we observed that most of the escalation in the FFG-7
ship class occurred in electronics systems, primarily governmentfurnished equipment. Unfortunately, we were not able to get a consistent
set of data to do such a comparison for aircraft. Although the HAPCA
data do show costs for lower-level work breakdown structures, the
groupings are not as helpful because airframe costs and all contractorfurnished equipment (CFE) are grouped into one category, from
which the separate eects of airframe, CFE, and GFE costs cannot be
discerned.
Summary
No one set of data can oer comprehensive insights on cost and technical characteristics for military aircraft. The most complete set of cost
data are those in P-1 budget submissions, which we use throughout
our analysis. For technical parameters, we use a variety of sources of
publicly available and contractor-provided information. We found that
xed-wing aircraft cost escalation has been about 2 to 7 percent greater
than that for common measures of ination. The trends of increase
seemed to dier by system. Cost escalation data appear to reect
upgrades and improvements that have occurred within programs, as
well as dierences between programs. We turn next to the potential
sources of this escalation, including both economic factors and those
related to system complexity and capability.
CHAPTER THREE
Economy-Driven Factors
Economy-driven factors in aircraft cost escalation are those that the

Services have little direct control over. These include such items as
labor rates, material costs, and equipment costs.1 For example, aircraft
manufacturing wage rates increase over time because of overall changes
in wages and prices throughout the economy, as well as changes in
prevailing wages manufacturers must pay to retain skilled workers. All
these variables are beyond the ability of the Services to control. In this
chapter, we explore how these economy-driven factors have changed
in recent decades, using a combination of data from the Bureau of
Labor Statistics (BLS) and Contractor Cost Data Reports (CCDRs)
for actual military aircraft.
Distribution of Costs
We used a series of CCDRs2 to determine the relative importance of
labor, equipment, and material costs in aircraft procurement. CCDRs
include data on F-14, F-15, F-16, F-18A/B, F-18E/F, A-10, AV-8B,
1
The Services do have choices on which materials to use in an aircraft as well as on the
nature and types of equipment used in it. We consider such variables in the next chapter,
when we discuss how customer-specied performance can aect costs. Here, we consider elements that the Services cannot control directly, such as equipment and material that manufacturers use to build aircraft, as well as costs for manufacturers labor.
CCDRs are provided to the government by contractors and detail the actual costs for the
weapon systems purchased. We obtained these data from the Defense Cost and Resource
Center (DCARC) Web site maintained by the Oce of the Secretary of Defense Program
17
18
C-17, T-45, F-22A, C-2A, E-2C, F-5E, and S-3A aircraft. The selection
was limited to those data readily available on the OSD PA&E Web site,
which archives historical CCDRs. For each program, we selected early,
middle, and late production points for analysis of trends. We limited
our selection to three data points to avoid bias from a few programs
with extensive CCDR histories. We use these data to analyze several
cost issues in this chapter.
Table 3.1 shows the relative distribution of labor, equipment, and
material costs in the CCDR data. These distributions do not include
subcontract costs specically identied in the CCDRs that are typically a mix of labor, material, and equipment, because we were unable
to correct for allocations by prime contracting rms that historically
included subcontracting costs as material costs in their CCDRs. The
percentages are based on the weapon system cost breakouts in the
CCDRs (air vehicle + system engineering and program management
+ integrated logistic support) that are comparable, but not identical, to
the procurement costs shown earlier. (Procurement costs also include
initial spares.)
Labor and equipment have accounted for most costs in these
xed-wing aircraft. These ratios, however, have not remained steady
over time. Figure 3.1 shows the change in the percentages for these
three components over time based on linear regression for the same
CCDR data. The negative regression slope (downward trend) in labor
Table 3.1
Average Distribution of Labor, Equipment, and Material
Costs for a Select Group of Fixed-Wing Aircraft,
1969 to 2003
Element
Percentage
Standard
Deviation, %
Labor
47
16
Equipment
38
20
Material
15
13
Analysis and Evaluation (OSD PA&E). See http://dcarc.pae.osd.mil/default.aspx for more

detail. We obtained a selection of the available reports between 1969 and 2003.
19
percentage is statistically signicant (at the 5 percent level). This trend

is likely due to two causes. First, productivity improvements and the
use of lean manufacturing have helped reduce the amount of labor
needed to perform similar functions.3 Second, aircraft manufacturers
have increasingly outsourced work when it is cost eective to do so, as,
for example, in the machining of simple parts. Such an outsourcing
would shift cost from labor to materials. Third, there has been a shift to
advanced computer-aided design systems that have also been thought
to improve the design process and reduce errors. However, the slopes
for the material and equipment percentages are not signicant to the 5
percent condence level, and therefore we cannot denitively say where
the shift out of labor has been.
Figure 3.1
Labor, Equipment, and Material Percentages of Weapon System Cost,
1969 to 2003
60
Labor
Percentage of total cost
50
40
Equipment*
30
20
Material*
10
0
1960
1970
1980
1990
2000
Fiscal year
*Trend not statistically different.
RAND MG696-3.1
See Figure 3.5 for increases in aerospace labor productivity over time.
2010
20
Below, we discuss in more detail individual economy-driven components of aircraft cost growth, including labor rates, material and
equipment, and other related costs (i.e., general and administrative
(G&A) and fee/prot). Because most CCDRs do not explicitly identify
fees and prot, we rely on publicly available annual reports (10-Ks) to
analyze prot trends.
Labor Rates
As described above, labor costs are a signicant portion of the total
system cost. Thus, changes in labor rates are of importance to understanding the inuence of the economic factors. Figure 3.2 shows BLS
data on the average annual increase for total compensation and direct
wage rates in the aerospace sector (rst two bars of Figure 3.2) compared with wage growth in the durable goods manufacturing industries
Figure 3.2
Average Annual Growth for Aerospace Labor Costs, 1989 to 2005
Average annual rate (1989 to 2005), %
6
5
4
3
2
1
0
Total
compensation
SOURCES: BLS and DoD.
RAND MG696-3.2
Direct
wages
DoD
procurement
deator
CPI
ECI
21
(Employment Cost Index (ECI) from 1989 to 2005).4 It also shows

two comparison indexes over this same timeframethe CPI and the
DoD procurement deator.
A few important trends are evident in this gure. One is that the
total compensation in the aerospace sector grew at a much higher rate
than did direct wage rates. This suggests that the cost growth of health
care and other benets has outpaced general ination. Earlier RAND
research (Arena et al., 2006a) found a similar trend in the shipbuilding industry, in which health care and insurance costs contributed to
faster growth for indirect wage costs. Unlike the shipbuilding industry,
where unburdened labor rates grew at about the same rate as ination
measures, here we see that direct wages have grown slightly faster than
the CPI, DoD deator, or ECI, indicating that direct labor costs for
the aerospace sector have also outpaced other common measures of
ination.
The CCDRs show an even greater annual rate of increase for the
direct and burdened labor costs of military aircraft. Figure 3.3 shows
the burdened and direct labor rates for the same set of aircraft programs used in Table 3.1.5 These data indicate that direct labor rates
have grown at an annual rate of 5.6 percent and the burdened rate has
grown at a slightly higher rate of 5.9 percent (exponential ts). Note
that there is considerably more variability around the t trend line for
the burdened data because other factors, such as business base, 6 are
important for determining this rate.
These labor rates can be further broken down into four subcomponents: engineering, tooling, quality control (QC), and manufacturing. Table 3.2 shows the direct and indirect rate increase for these
four subcomponents. Growth for all four elements, whether direct or
4
The range of dates was limited by data availability from the BLS.
Note that total compensation and burdened labor rates are not the same. Burdened labor
includes total compensation in addition to other indirect costs such as corporate insurance,
maintenance and repair, and depreciation.
Business base considerations are important in determining the burdened costs because
certain xed costs (costs that occur independently of workload such as security costs) get
spread to all the work. With a bigger base to spread these costs, the eect on the hourly rate
is reduced.
22
Figure 3.3
Annual Direct and Burdened Labor Costs from a Select Set of Aircraft
Programs, Fiscal Years 1969 to 2003
100
90
Labor rate ($/hr)
5.9%
Direct
Burdened
80
70
60
50
40
5.6%
30
20
10
0
1960
1970
1990
1980
2000
2010
Fiscal year
SOURCE: CCDRs.
RAND MG696-3.3
Table 3.2
Direct and Indirect Annual Escalation Rates for Labor
Subcomponents, 1969 to 2003
Labor Subcomponent
Direct, %
Indirect, %
Engineering
5.7
6.2
Tooling
6.0
5.9
QC
5.6
5.3
Manufacturing
5.8
5.8
indirect, appears to be similar. In other words, growth in no one element appears to be driving increases in overall labor costs.
Another important issue is whether there has been a shift in the
labor content; in other words, has there been a shift in the relative
hours by labor subcomponent? Figure 3.4 shows trend lines (linear
regressions) for the fraction of the labor hours by subcomponent. We
23
Figure 3.4
Percentage of Labor Hours by Subcomponent, 1969 to 2003
80
Percentage of total hours
70
60
Manufacturing
50
40
30
20
Engineering
Tooling
10
QC
0
1960
1970
1980
1990
2000
2010
Fiscal year
NOTE: Trends are not statistically different.
RAND MG696-3.4
have omitted individual observations for clarity. None of the trends

shown are statistically signicant. This implies that there has not been
an observable change in the labor ratios or a shifting of work between
the labor types that could account for some change in direct rates. The
average (and the associated standard deviation) for the distribution of
hours for the time period shown has been
t
t
t
t
manufacturing, 54 percent (15 percent)

engineering, 25 percent (16 percent)
tooling, 13 percent (7 percent)
quality control, 9 percent (2 percent).7
Some reduction in the percentage of labor costs relative to those

for equipment and material costs might be attributed to greater eciency. Figure 3.5 shows the relative labor productivity (output per
7
Numbers do not add to 100 percent because of rounding.
24
Figure 3.5
Aerospace Labor Productivity, 1987 to 2003
Relative productivity, % (1997 = 100)
125
120
115
110
105
100
95
90
85
80
1985
1990
1995
2000
2005
Year
SOURCE: BLS.
RAND MG696-3.5
hour) as measured by the BLS in the aerospace sector from 1987 to

2003. The overall annual productivity increase is 2.2 percent per year.
Note that the automotive industry in this timeframe saw an approximate 3.2 percent gain in productivity. Over the 16-year span, productivity in the aerospace sector increased by a total of about 40 percent.
This gain in productivity helps to oset the increase in direct and indirect labor rates.
Material and Equipment
Material and equipment costs are the other two economic factors in
aircraft cost growth. As was seen from the earlier split in Table 3.1,
these two elements account for just over half of the weapon system cost
for typical xed-wing aircraft. Metals (e.g., aluminum, steel, titanium)
or composites (e.g., carbon ber, bismaleimide, and thermoplastics) are
mainly used for the manufacture of the airframe (the aircrafts main
25
structure). Modern aircraft increasingly use high-strength, low-weight

materials such as titanium and composites for improved performance.
Composites have the added benet of potentially reducing signatures
(i.e., reducing the possibility of detection by sensors such as radar).
Equipment comprises all systems such as avionics (electronics to control the aircraft), sensors (such as radar), other mission systems (such as
electronic countermeasures, communications, targeting systems, guns,
and missiles), and propulsion systems.
To understand how price changes in these materials and systems
aect cost, we analyze the price escalation for several commodities
from Producer Price Index (PPI) data collected by BLS. These include
t steel mill productsBLS Series ID WPU1017
t aluminum sheet, plate, and foil manufactureBLS Series ID
PCU331315331315
t titanium mill shapesBLS Series ID WPU102505
t thermosetting resins and plastics materialsBLS Series ID
PCU3252113252114
t other aircraft parts and auxiliary equipment manufacturing
BLS Series ID PCU336413336413
t aircraft engine and engine parts manufacturingBLS Series ID
PCU336412336412
t aeronautical, nautical, and navigational instruments, not sending/
receiving radioBLS Series ID PCU3345113345111.
Figure 3.6 shows the average compound growth rate for these
seven components from 1986 through 2004. The horizontal lines
depict growth in the CPI and the DoD procurement deator during
that same time. Growth for most of the components is above the DoD
deator but less than the CPI. Two components, those for navigation
equipment and titanium prices, exceed growth for both comparison
indices. These components do not reect the recent price increases in
metals experienced since 2004, including the 40 percent increase in the
price of steel in 2004 and 2005 and the near doubling of the price of
titanium between January 2005 and April 2006.
26
Figure 3.6
Material and Equipment Price Escalation, 1986 to 2004
Average annual rate (1986 to 2004), %
CPI
3
DoD
2
aTrend
A
en ircr
gi af
ne t
s
N
eq avi
ui ga
pm tio
en n
t
re erm
sin o
s/p set
la tin
st g
ic
s
A
irc
ra
ft
pa
rt
s
Th
Ti
ta
ni
um
lu
in
St
ee
um
la
does not include recent, dramatic price changes.
SOURCE: BLS.
RAND MG696-3.6
To determine indices for material, we use a weighted average of

the rst four PPI components in the list above for each aircraft. We base
the weighting on the proportion of each material in the nal weight
of a typical airframe.8 We recognize that the materials composition of
xed-wing aircraft has dramatically changed over the last few decades.
Nevertheless, we see these changes driven by performance and requirements issues (not a substitution for economic reasons) because the customers desire for performance has aected the choice of materials for
airframes. In the next chapter, we address the cost implications of the
shift toward more advanced, and expensive, materials. For an index of
the cost escalation for equipment, we use a simple average of the last
8
Using nal weight of material composition in the aircraft does not reect production
waste dierences between material types and their associated manufacturing methods and
scrap prices.
27
three PPI components. Table 3.3 shows the annual indexes for material
and equipment.
Table 3.3
Material and Equipment Escalation
Rates, 1986 to 2004
Factor
Material
Equipment
Annual
Increase, %
1.9 to 3.1
2.6
NOTE: We present the possible range

of values because the material escalation
depends in large part on the material
mix of the airframe.
Fees and Prots

The nal set of economy-driven factors that we examine is fees and
prots. Although the government has some control on negotiating fee
levels, it does not set market expectations for the rms or the level of
return the companies must demonstrate to their shareholders. Rather,
the marketplace sets a baseline level of return that must be ultimately
reected in prots that manufacturers earn. Similarly, there are allowable charges that the rms can bill the government. Three such charges
are G&A costs, material overhead, and fees and prots. G&A is fairly
consistently reported in the CCDRs, but material overhead and prot/
fee are not. Below, we review trends in G&A, material overhead, and
fees and prots for aircraft manufacturers.
General and Administrative Costs
G&A costs are allowable charges that cover general corporate expenses
that cannot be attributed to a single program or contract. These costs
typically cover expenses such as corporate management salaries and
benets, legal and accounting costs, and oce supplies, as well as internal research and development (IR&D) and bid and proposal (B&P)
costs. Figure 3.7 shows the fraction of G&A relative to total cost
28
Figure 3.7
G&A Percentage of Total Cost, by Fiscal Year, 1969 to 2003
18
G&A percentage of total price
16
14
12
10
8
6
4
2
0
1960
1970
1980
1990
2000
2010
Fiscal year
SOURCE: CCDRs.
RAND MG696-3.7
from the CCDRs for the 14 programs described above. Despite the
variability, there is a clear and signicant trend (>95 percent condence), with the G&A proportion of total costs increasing approximately 0.2 percent per year.
Material Overhead
Material overhead covers activities such as handling, ordering, and

storage costs for material and equipment purchased on a contract.
Figure 3.8 shows trends in material overhead cost relative to total cost
from the CCDRs for the 14 programs described above. Note that the
trend line is not signicant because of the high variability of the data.
This variability is due to how material activities are charged on a contractsometimes as a direct charge, sometimes in general overhead,
and sometimes as an explicit fee. Typically material overhead is about
3 percent of total aircraft costs.
29
Material overhead percentage of total price
Figure 3.8
Material Overhead Percentage, by Fiscal Year, 1969 to 2003
10
9
8
7
6
5
4
3
2
1
0
1960
1970
1980
1990
2000
2010
Fiscal year
RAND MG696-3.8
Fees and Prots
Fee and prot levels are often inconsistently reported or not reported
at all in the CCDRs. To assess whether these costs have changed over
time, we use the annual reports (10-Ks) for the three major U.S. military aircraft manufacturersBoeing, Lockheed Martin (LM), and
Northrop Grumman Corporation (NGC)and data on operating
margins in their defense aircraft sector. These reports are publicly available on company or Securities and Exchange Commission (SEC) Web
sites. Operating margin is the ratio of overall prot to sales and does
not directly correspond to prot earned on any particular contract. For
example, other sector expenses or investments might oset the total
prot earned on individual contracts.
Analyzing trends in operating margins for these rms is dicult
because of the many reorganizations and mergers they have undergone
in recent decades. Nevertheless, such data are the best data publicly
available. Figure 3.9 displays trends in operating margins over time.
30
Figure 3.9
Operating Margin for Aircraft Sector, by Year, 1975 to 2005
Aircraft sector operating margin, %

(prot/sales)
20
15
10
5
0
5
LM
NGC
Boeing
10
15
20
1970
1975
1980
1985
1990
1995
2000
2005
2010
Fiscal year
SOURCE: Company annual reports.
RAND MG696-3.9
The data are highly variable, particularly in the late 1980s. After 1995,
the operating margins stay within a band of approximately 5 to 15 percent. This narrow range might be a reection of the way DoD negotiates prot levels.
Although operating margins for all three manufacturers have uctuated over time, there is no discernable, consistent trend to be seen.
Rather, operating margins have generally remained between 5 and 15
percent of total sales.
Notional Aircraft Comparisons

How much does each economic factor that we have reviewed contribute
to the total annual rate of increase in the price of xed-wing aircraft?
The answer depends to a certain extent on the characteristics of the
aircraft (e.g., the labor/material/equipment split, the materials of construction) and the timeframe. For illustration, we will use a notional
31
comparison of an aircraft built in 1980 and the same aircraft built in

2000. We assume that the material distribution for this notional aircraft is 55 percent aluminum, 10 percent steel, 30 percent titanium,
and 5 percent composites. Our analysis approach uses a weighted
price index based on the component indexes (i.e., labor, material, and
equipment) described above.9 Table 3.4 lists the contributions of the
economic factors for this example. The total for all the factors is 3.5
percent. In comparison, the CPI rate of increase was approximately
3.8 percent and the DoD procurement deator was 3.3 percent. Thus,
the overall growth resulting from the individual economic factors for
this notional example was similar to the other measures of ination.
Table 3.4
Contributions of the Economic Factors to
Cost Escalation for a Notional Example
Economic Factor
Labor
Contribution to
Annual Rate of
Increase, %
0.8
Material
1.3
Equipment
1.1
Fees and prots
0.2
Total
3.5
Summary
In this chapter, we have explored the contribution of economy-driven
factorsthat is, those largely outside the direct control of the Servicesto cost growth for aircraft. Although we found that labor costs
(both direct and indirect) grew at a rate greater than other measures of
ination, we also saw that gains in productivity oset these increases.
Materials and equipment costs grew at a rate at or below these same
9
The approach is analogous to the way the ination is constructed using component
indexes. Here we use the percentage of value as our weighting factors. For more information
on weighted price indexes, see, for example, Schultze and Mackie (2002) or Statistics Canada
(1995).
32
measures of ination. For a notional example, the overall rate of growth

resulting from the individual economy-driven factors was 3.5 percent,
a rate of increase similar to that for the CPI and DoD deator in the
same period of time. Our notional example, however, assumed that
an identical aircraft was being purchased over time. Undoubtedly, the
complexity and quality of the aircraft (e.g., its performance and eectiveness) that the Services have purchased have increased over time. In
the next chapter, we explore the price implications of the demand for
more complex aircraft.
CHAPTER FOUR
Customer-Driven Factors
Although the Services cannot inuence many of the elements that work
to set aircraft prices, they are, of course, responsible in at least two ways
for the amount of money they spend on aircraft. First, the Services
decide on the number of aircraft they wish to purchase. Second, they
determine the characteristics they want these aircraft to have. Although
the number and characteristics of aircraft may be determined by
threats to the nation that the Services must address, it is still true that
the Services can inuence these variables more than economy-driven
ones aecting price. We examine these customer-driven variables in
this chapter.
For the analysis in this chapter, all unit costs are adjusted to scal
year 2006 dollars using APN procurement deators. Ination adjustment is critical because the eects of ination itself are accounted for
in our previous analyses of the economy-driven factors of labor, equipment, and material.
Quantity Effects
The quantity of aircraft that the Services procure can aect aircraft cost
in two dierent but related ways. First, additional quantities ordered
over time can have a cost improvement eect, in which accumulated
experience in producing the same system year after year helps to reduce
its unit cost. Second, the quantity ordered in any given year has a procurement rate eect that results from changing the lot size of the same
system from one year to the next, with high procurement rates helping
33
34
to reduce unit cost through greater operating eciency and the spreading of xed costs over more units.
Confounding these quantity eects are the inuences of conguration changes. A conguration change is a series change in the production design (e.g., from F/A-18A/B to F/A-18C/D). Such a change may
disrupt production to the point that some quantity eects are altered
or masked through increased cost. We will explore how this series
change eect modies the quantity eects by directly accounting for
the sequence of series variants. (For example, for the F/A-18 model, we
consider the A/B, C/D, and E/F variants as one linked program and
look at the eect on the cost of each series change).
Cost Improvement
The cost improvement (CI) eect means that the eort and expenditure required to produce an aircraft will decrease as the number of
aircraft produced increases. More generally, it can be expressed as the
eect of learning-by-doing on aircraft cost. Mathematically, the cost
improvement eect can be expressed as
CI slope )/ln( 2 )
C t = C1 n ln(
t
(4.1)
where
t
t
t
t
Ct is the cost of the unit at the midpoint of annual buy t

C1 is the cost of the rst unit
nt is the unit number of the midpoint1 of annual buy t
CI slope is the unit cost improvement slope.
For technical reasons, the midpoint of the rst annual buy is assumed to be reached after
one-third of the total units for the rst annual buy have been procured. For subsequent lots,
the midpoint is assumed to be reached when one-half of the total units in an anual buy have
been procured.
35
For Equation 4.1, we have used a lot midpoint formulation because

our cost data are a lot average only; we do not have cost by individual
aircraft tails.2
We assessed Equation 4.1 for each aircraft, many of which had
production runs of three or four continuous annual buys. Table 4.1
presents regression equation results depicting the cost improvement
eect as a function of the minimum number of annual buys of the
aircraft. Scaling by number of annual buys is important because fewer
annual buys mean fewer data points in Equation 4.1, yielding potentially unreliable results. Table 4.1 shows that the mean and median cost
improvement slope, and hence the cost improvement eect, is modest
only a few percentand does not substantially depend on the number
of annual buys. It shows that for programs with at least ve annual
procurementsfor which there were 52 aircraft systemsevery doubling of aircraft unit number leads to a unit cost decrease averaging
3 percent.3
Notice also that for programs with at least six annual buys,
there is essentially no average cost improvement.4 This does not mean
that cost improvement is nonexistent. On the contrary, the standard
Table 4.1
Cost Improvement Slopes, by Minimum Number of Annual Buys
Minimum Number of Annual Buys
3
Mean CI slope
0.98
0.95
0.97
0.99
Median CI slope
0.97
0.94
0.97
1.00
Standard deviation CI slope
0.18
0.15
0.15
0.14
Number of aircraft systems
98
64
52
37
Readers requiring either an introduction or a detailed explication of learning curve analysis chould consult Goldberg and Touw (2003).
A CI slope 0.97 implies that moving from the rst to the second unit reduces unit cost
by 3 percent (1 0.97 = 0.03). A further 3 percent reduction is also seen in the fourth unit,
compared with the second unit, etc.
Mean, median, and standard deviations change little with minimum number of annual
buys greater than six.
36
deviation of 0.14 shows that some systems have cost improvement

slopes much lower than one. Nevertheless, this also does show that
some systems have cost improvement much greater than one or experienced price increases as the number of annual buys increased.
Procurement Rate
Procurement rates may have positive or negative eects on unit price.

A high procurement rate, for example, can help spread xed overhead
costs over more aircraft, thus reducing the average unit cost. Higher
procurement rates may result in greater and more ecient use of existing plant and tooling, also helping to reduce unit costs. Higher procurement rates may also lead to more ecient use of labor through
specialization. Other variables, such as increases in total costs caused
by unforeseen supplier problems, unexpected labor and materials costs,
or failures in implementing new technologies, could lead to higher
costs, which, in turn, could lead to the government cutting procurement rates.
Mathematically, the combination of cost improvement and procurement rate eects can be stated as
C t = C1 n tln(CI slope )/ln( 2) rtln(PR slope )/ln(22 )
(4.2)
where
t
t
t
t
t
t

nt is the unit number of the midpoint of annual buy t
CI slope is the unit cost improvement slope
rt is the number of units procured in annual buy t
PR slope is the procurement rate slope.
Again, we have used the lot midpoint formulation as our cost

data are by lot and not unit. Cost improvement and procurement rate
eects are estimated for all systems in the HAPCA and P-1 databases;
the constants C1, CI slope, and PR slope are uniquely estimated for all
systems.
37
Equation 4.2 looks and functions like Equation 4.1. The additional procurement rate slope term can be interpreted just like the
cost improvement slope: A doubling of the procurement rate leads to a
1 PR slope percentage change in average cost.
Yet, there are several complications with this formulation. Some
of these complications force us to restrict the systems under consideration. Others require that we make arbitrary methodological decisions
that could change the results of the analysis. First, many systems have
limited production runs, but Equation 4.2 cannot be used for systems
with fewer than four scal year purchases. Of the 178 Navy and Air
Force xed-wing aircraft systems (including variants) in the combined
P-1 and HAPCA database, only 52 have at least ve scal year buys.
Second, midpoint quantity and procurement rates tend to be highly
correlated, especially among systems with short production runs.
Because this correlation can lead to statistically misleading results, systems with midpoint and lot size correlations greater than the absolute
value of 0.6 are excluded in the nal analysis of cost improvement and
procurement rate eects.5 These restrictions reduce the number of systems that can be analyzed from 52 to 24.
The restricted data still include a diverse array of Air Force and
Navy programs over the past 50 years, including attack, cargo, electronic, patrol, and training aircraft, but no bombers. Table 4.2 shows
Table 4.2
Cost Improvement and Production Rate Slopes, by Minimum Number of
Annual Buys
Minimum Number of Annual Buys
5
10
Average CI slope
0.97
0.98
0.98
0.99
0.99
0.99
Standard deviation in CI slope
0.13
0.14
0.14
0.14
0.16
0.16
Average PR slope
0.90
0.88
0.88
0.86
0.83
0.80
Standard deviation in PR slope
0.23
0.15
0.15
0.15
0.16
0.14
24
14
14
12
Number of aircraft systems
For example, the F/A-18 series A/B, C/D, and E/F are excluded from this analyis because
of the high degree of correlation between midpoint and lot size.
38
the results of applying Equation 4.2 to these systems and how cost
improvement and procurement rate vary by minimum number of years
in which systems are purchased.
Table 4.3 shows average cost improvement and procurement rate
slopes by aircraft type and by service, along with minimum and maximum slope evident among individual systems we consider.
With and without adjusting for procurement rate eects, there
is little to no cost improvement on average: Slopes center on 96 percent (Table 4.3) to 97 percent (Table 4.1). Cost improvement slopes
range from 74 percent for C-9 to 139 percent for C-130H, with an
overall average of 96 percent; it is notable that these extremes are in
the same category that on average have almost no CI eect: cargo
aircraft. On average, attack, ghter, cargo, and training aircraft have
cost improvement slopes less than one, meaning that unit costs become
cheaper as unit number increases, whereas electronic and patrol aircraft have cost improvement slopes greater than one, meaning that unit
costs actually rise as unit number increases.
Table 4.3
Cost Improvement and Production Rate Slopes
with a Minimum of Five Annual Buys
CI Slope
PR Slope
Attack
0.91
0.87
Cargo
0.99
1.08
Fighter
0.93
0.78
Electronic
1.09
0.73
Patrol
1.09
1.04
Trainer
0.97
0.91
Minimum
0.74
0.61
Mean
0.96
0.89
Maximum
1.39
1.75
Navy mean
0.94
0.87
Air Force mean
1.01
0.94
39
Procurement rate slopes range from 0.61 for F-15E to 1.75 for C-37,
with an overall average of 0.89. This means that, on average, a doubling
of annual procurement quantity yields an 11 percent decrease in unit
cost. Also, Air Force systems have a 7 percent higher cost improvement
slope and 7 percent higher procurement rate slope than Navy systems.
These dierences are almost fully explained by the pull of extreme cargo
aircraft observations: a C-130H CI slope of 1.38, and a C-37 PR slope of
1.75. Without these two data points, the Air Force would generally have
lower cost improvement and procurement rate slopes than the Navy.
In determining average cost improvement curve and procurement
rate eects, we chose to mix the results of regressions without weighting by their statistical signicance. In general, smaller production runs
will yield less statistically signicant results, even though the substantive underlying relationship is present. Yet, we found surprisingly little
dierence between cost improvement and procurement rate slopes in
data samples that included or excluded small production runs. We also
found minimal correlation between these slopes and no evident trend
over time for them.
Conguration Effects
It is possible that separating a single aircraft model into its component
series for quantity analyses neglects the cost consequences of a shared
development and production environment.6 We therefore analyze cost
improvement and production rate slope eects within a single model
by specically incorporating conguration (series change) eects in
our regression analysis.
Only a few aircraft systems have long-term continuous production
of well-dened conguration upgrades. In addition to the well-known
cases of the F-14, F-15, F-16, and F-18 ghters, the NAVAIR database
contains the older but still relevant AV-8B Harrier II, F-4, A-6, and P-3
6
Here, a series change means modifying a number of components of an already existing and successful aircraft, such as the modication needed to make the F-16A/B into the
F-16C/D. This modication process is usually far cheaper than a blank-sheet design of a
new airframe, with new avionics and propulsion.
40
programs. Table 4.4 lists all congurations for the aircraft and the year
they were implemented. Of all the aircraft considered in this subsection,
only the F/A-18 is not included in our analyses above on cost improvement and procurement rate eects (because of correlation issues).
We analyze conguration eects with the following equation:
CI slope )/ln( 2 )
C t = C1 n ln(
rtln(PR slope )/ln(22 )
t
exp(d1 * CC1,t + d 2 * CC 2,t )
(4.3)
where
t
t
t
t
t
t
t
t
t
t

nt is the unit number of the midpoint of annual buy t
CI slope is the unit cost improvement slope
rt is the number of units procured in annual buy t
PR slope is the procurement rate slope
d1 is the rst conguration change coecient
CC1,t is the dummy variable for the rst conguration change
d2 is the second conguration change coecient
CC2,t is the dummy variable for the second conguration
change.
Table 4.4
Aircraft Models Long-Term Production Proles and Dates of
Conguration Change
Model
Service
Original
Conguration
Second
Conguration
Third
Conguration
A-6
Navy
A-6A
1959 A-6E
AV-8B
Navy
AV-8B
1982 AV-8B NAA 1988 AV-8B RAD 1991
F-14
Navy
F-14A
F-15
Air Force F-15A/B
1971 F-14A+
<1974 F-15C/D
1970
1986 F-14D
1988
1979
1988
F-15E
F-16
Air Force F-16A/B
1978 F-16C/D
1983
F-4
Navy
F-4A
1955 F-4B
1960 F-4J
1965
F/A-18
Navy
F/A-18A
1979 F-18C/D
1986 F/A-18E/F
1997
P-3
Navy
P-3A
1961 P-3B
1965 P-3C
1968
41
The conguration change variables indicate whether a procurement lot has progressed to the next redesign or improvement in technology. A lot has a value of 1 if the system produced in that lot has
progressed to the next series, 0 if it has not. The conguration change
coecients should indicate an increase in the cost of the system compared to the original conguration. In Equation 4.3, the conguration
change variables work to increase or decrease ln(C1), the constant.
In Table 4.5, we compare the calculated cost improvement and
procurement rate coecients using Equation 4.3 with conguration
change terms to those calculations using Equation 4.2. The aggregate
result indicates that explicitly modeling conguration change has little
substantive eect on the values for the cost improvement or procurement
rate slopes as shown in the values of Table 4.5. Thus, the fact that we
did not account for conguration changes in our earlier cost improvement and procurement rate analysis did not inuence our results.
For each aircraft in Table 4.5, the leftmost column shows the
system name. The CI Slope column contains the estimate for CI
slope not accounting for conguration change; the CI Slope with
CC column contains the estimate for CI slope accounting for conguration change. The next two columns contain PR slopes without and
then with conguration change taken into account. The next column
contains the correlation between lot midpoint unit number and the
number of units procured in that lot. The last two columns contain
the conguration change coecients: CC1 for the rst conguration
change, CC2 for the second conguration change.
In the results including conguration change eects, the F-4
shows the highest correlation between cost improvement and procurement rate, 0.66. This high correlation calls into question the statistical
validity of the values for the F-4. We therefore present the mean values
with and without this observation included in the sample. Including
conguration change apparently has little to no eect on the mean CI
or PR slopes of the remaining seven programs. Including conguration
change eects results in
42
Table 4.5
Results of Regression Incorporating Conguration Change (CC) Effects
System
CI Slope
CI Slope
with CC
PR
Slope
PR
Slope
with CC
Correlation
Between
CI and PR
CC1
Slope
CC2
Slope
A-6
0.88
0.86
0.71
0.73
0.10
1.13
AV-8B
0.85
0.86
0.72
0.71
0.14
0.97
F-14
0.94
0.93
0.77
0.80
0.43
0.97
1.29
F-15
0.97
1.18
0.66
0.62
0.51
0.60
0.49
0.98
F-16
0.99
0.91
0.94
0.94
0.46
1.45
F-4
0.81
0.72
0.86
0.86
0.66
1.21
3.48
F/A-18
0.98
0.91
0.83
0.87
0.07
0.95
2.02
P-3
1.05
1.02
0.68
0.76
0.41
0.77
1.28
Mean (all)
0.93
0.92
0.77
0.79
1.00
1.59
Mean
(without
F-4)
0.95
0.95
0.76
0.78
0.98
0.87
t little to no change in cost improvement or procurement rate slopes

for the A-6, AV-8B, and F-14 aircraft
t a slight increase in the cost improvement slopes but no eect on
the procurement rate slopes for F-16 and F-4 aircraft
t a attening of the cost improvement slope but a steepening of the
procurement rate slope for P-3 aircraft
t a reversal of the cost improvement slope for F-15 aircraft, with a
negative cost improvement (CI > 1) slope but a persisting steep
procurement rate eect.
The conguration change coecients, CC1 and CC2, are multipliers of average unit cost.7 They are mathematically very simple eects,
but their meaning must be interpreted with great caution. For example,
these equations show that, taking into account CI and PR eects, the
third conguration F-15 is 51 percent less expensive than the original
conguration (or 1, representing the original conguration, less 0.49,
representing the CC2 slope for the third conguration, or 0.51 x 100
percent). But historical costs for the F-15 actually rose 10 percent from
7
CC1 slope and CC2 slope are dened as eCC1 and eCC2, respectively.
43
the last lot of the original conguration to the rst lot of the third
conguration. One can understand the intuition behind Equation 4.3
and conguration eects by considering CI eects, PR eects, and CC
eects together for a specic lot-to-lot comparison.
At the end of the initial conguration, roughly 500 F-15A/B had
been produced, and by the beginning of the third conguration, nearly
950 F-15A/B/C/D had been produced. Hence, at a 1.18 slope, CI eects
should have increased unit costs by 17 percent.8 And the average production rate starting the third conguration (42) was 58 percent less
than the production rate ending the original conguration (97); at a
0.62 slope, PR eects should have increased unit costs by 78 percent.9
Yet actual unit costs rose by 10 percent, meaning that another eect
(conguration change) could be counteracting the CI and PR eects.
A conguration change slope of 0.49 implies a 51 percent decrease in
unit cost as a result of the conguration change. Multiplying these factors together as seen in Equation 4.3 yields 1.17 1.78 0.49 = 1.02;
that is, in this case, there was a large conguration eect bringing costs
down that nearly balanced the cost-increasing CI and PR eects. The
dierence between the predicted cost increase of 2 percent and the
actual cost increase of 10 percent is due to statistical error.
For all seven programs within correlation bounds (i.e., those
for which the correlation between lot midpoint unit number and the
number of units procured in that lot is less than the absolute value of
0.6), the rst conguration change has virtually no eect on cost. This
is shown in Table 4.5 by the mean 0.98 CC1 slope for these programs.
The second conguration change leads to a 13 percent decrease in cost,
after controlling for cost improvement and production rate eects. This
is shown in the 0.87 mean for the CC2 slope among the ve remaining
programs within correlation bounds that had at least two conguration changes.
Equation 4.3 contains ntln(CI slope)/ln(2). We calculate 17 percent by dividing 950ln(1.18)/ln(2) by

500ln(1.18)/ln(2) then subtracting 1.
9 Equation 4.3 contains r ln(PR slope)/ln(2). We calculate 78 percent by dividing 42ln(0.62)/ln(2) by
t
97ln(0.62)/ln(2) then subtracting 1.
44
Basic Technical Characteristics

In addition to assessing learning curve and production rate eects, we
used our technical database to determine how several aircraft-specic
performance variables aect average unit cost over the entire program.
These variables include service, personnel complement, ceiling, range,
airframe weight, thrust (or horsepower), cruising speed, maximum
speed, mission of aircraft, and whether it is carrier-based.
For this analysis, we rst identied variables that are highly correlated with average unit cost. We then identied the subset of variables with high explanatory power, that is, large, statistically signicant
regression coecients but little correlation with one another.
The nal set of variablesempty weight, maximum speed,
whether an aircraft is carrier based, whether the aircraft is electronic,
and whether the aircraft is a bomberare shown in Table 4.6. We analyzed them using the following equation:
C = K EWa MSb CarrierBasedc Electronicd Bombere (4.4)
where
t
t
t
t
t
t
t
t
C is the average unit cost of an aircraft system

K is a constant
EW is the empty weight of the aircraft
MS is the maximum speed of the aircraft
CarrierBased is a dummy variable for carrier-based aircraft
Electronic is a dummy variable for electronic aircraft
Bomber is a dummy variable for bomber aircraft
a, b, c, d, and e are the coecients of the EW, MS, CarrierBased,
Electronic, and Bomber variables.
The parameter values of Table 4.6 help us to understand some of

the technical drivers of aircraft unit cost. For example, a larger aircraft
(all other things being equal) costs more than a smaller one. Not only
is this a reection of having to build a larger airframe but also of more
expensive systems that are put onboard. Notice, however, that the scaling
45
Table 4.6
Results of Regressions on Technical Characteristics
Parameter
Standard Error
t Statistic
p Value
ln (empty weight)
0.91
0.08
11.54
<0.0001
ln (max speed)
0.83
0.14
5.75
<0.0001
Carrier based
0.38
0.17
2.27
0.0256
Electronic aircraft
1.01
0.25
4.06
0.0001
Bomber aircraft
1.14
0.35
3.21
0.0018
11.83
1.20
9.86
<0.0001
Intercept
is less than one. In other words, an aircraft twice as big does not cost
twice as muchit costs 1.88 (2^ 0.91) times as much. Carrier-based
aircraft cost more than land-based ones. This dierence is likely due to
the more extreme operating environment (e.g., arrested landings and
sea environment) that these aircraft face. Electronic aircraft (e.g., P-3,
E-2C, and AWACS cost more than other aircraft types of similar performance. This dierence is due to the additional mission equipment
that these aircraft have. The bomber parameter is driven by the B-2, for
which stealth was an emphasized capability.
These variables help us to assess the relationship between the airframe design and propulsion aspects of the aircraft but do not address
the internal complexity of its airframe. The complexity of the airframe
can be assessed by quantifying its type of materials. The metric we use,
labeled simple, is the share of airframe structure that is neither titanium nor composite material. Unfortunately, only 49 of the 93 aircraft
with complete data on basic technical characteristics have available data
on material composition. To retain as many observations as possible for
analysis, we regressed basic technical characteristics on 93 aircraft as
in Equation 3.4. The residuals of 49 of those (denoted in Appendix A)
are used as the dependent variable to assess the cost eects of materials
composition. Mathematically, this yielded
R* = K Simplea
where
(4.5)
46
t R* are the residuals of regression for Equation 4.4

t Simple is the fractional share of airframe structure that is neither
composite nor titanium
t a is the coecient for the Simple variable.
Table 4.7 presents the results of these regressions. The 1.42 coefcient of Simple indicates that as titanium and composite materials
increase as a proportion of airframe materials (or as simpler materials
decrease), aircraft unit cost increases. Increasing the proportion of titanium and composite materials in an airframe from 0 to 5 percent costs
a 7 percent premium, whereas increasing them from 40 to 45 percent
costs a 13 percent premium.10
Table 4.7
Results of Regressions on Airframe Materials Complexity
Parameter
a
ln(K)
Value
Standard
Error
T Statistic
P Value
1.42
0.43
3.31
0.0014
6.51
1.92
3.39
0.0014
Other Elements
Our analyses do not account for changes in avionics complexity, software implementation, operating and support costs, and longevity. Metrics such as avionics power, weight, or lines of code could serve as proxies for avionics complexity. Mean time between shop visit could serve
as a proxy for operating and support costs. The number of landings or
ying hours permitted between major overhauls could serve as a proxy
for longevity. Data for these variables, however, are available for only
the most recent systemsand are not even applicable to the oldest systems in our database. Even with sucient information on these variables, it is unlikely that the limited number of observations would provide clear results.
10
According to Equation 4.5, cost is proportional to Simplea. Seven percent is derived from
95 1.42/100 1.42. Thirteen percent is derived from 551.42/60 1.42.
47
Summary
In this chapter, we have explored the contribution of customer-driven
factors that are in the direct control of the Services. We found that
technical characteristics (such as airframe weight, maximum speed,
and materials composition) correlate very strongly with unit price,
suggesting that technical complexity is a major driver within the customer-driven factors. We also explored cost improvement and production rate eects. The cost improvement eect was highly variable, and
we observed no general or consistent trends. On the other hand, production rate did show a consistent eect in that the average unit price
was lower for increased production rate. In the next chapter, we explore
the relative importance of all the customer and economic factors by
comparing the dierence between pairs of aircraft.
CHAPTER FIVE
Pairwise Comparisons
In this chapter, we analyze both economy-driven and customer-driven

factors together by comparing aircraft systems recently procured with
same-mission, previous-generation aircraft procured decades ago.
These pairwise comparisons are necessary to complete our analysis.
Cost escalation measured between two systems may be attributable to
diering economic environment and customer choices for each aircraft
being procured. Without picking one aircraft in the past and one in the
present, there is no way to quantitatively explain cost escalation.
We selected eight aircraft pairs for the comparison, at least one
example from each mission type. Where possible, we chose examples
from both Services. Finally, we also chose pairs that spanned the greatest period of time and technical change. For example, we compare the
F-15A to the F-22A rather than the F-15C/D to the F-16C/D. Obviously, many such combinations were possible. The pairs selected are
representative, only.
We will explain dierences in cost escalation between comparisons by specifying the characteristics of each aircraft in comparison pairs and accounting for economy- and customer-driven factors.
Among the individual economy-driven factors we consider are labor,
material, and equipment. Individual customer-driven factors we consider are those relating to technical characteristics, airframe complexity, and production rate. Once these are determined, we aggregate and
summarize the eects of both types of factors on cost escalation and
compare them with actual escalation.
49
50
Table 5.11 presents the eects of individual economy-driven factors on
the eight aircraft pairs. For each pair, we present labor, material, and
equipment escalation as well as G&A escalation as calculated in Chapter Three.
As noted in Chapter Three, labor costs have increased at a rate
greater than ination but, until recently, material and equipment costs
have increased somewhat less so. Nevertheless, as also noted above,
productivity improvements in aircraft manufacturing and outsourcing
practices have reduced the proportion of costs attributable to labor. In
1970, material and equipment represented 45 percent of the cost of an
aircraft; in 2005, they represented 62 percent. How much of cost escalation can these three factors explain?
For each pairwise comparison, we estimated each economydriven factor by examining how representative indices (as identied in
Chapter Three) changed over the time period in question, after weighting each by that factors percentage of total cost. (Analysis of CCDRs
generated annual estimates of the share of cost attributable to labor,
Table 5.1
Percentage Contributions to Annual Cost Escalation, by Economy-Driven
Factors
Comparison
Labor
Material
Equipment
G&A
Total
F-15A (1975) to F-22A (2005)
0.8
1.6
1.0
0.2
3.5
F/A-18A/B (1983) to F/A-18E/F (2003)
0.8
1.0
1.1
0.2
3.1
B-1B (1984) to B-2A (1993)
0.9
0.9
1.1
0.2
3.1
C-130H (1980) to C-17 (2005)
0.8
1.3
1.0
0.2
3.2
E-3A (1975) to E-8C (2005)
0.8
1.4
1.0
0.2
3.4
E-2C (1975) to E-2C (2004)
0.8
1.5
1.0
0.2
3.4
T-34C (1978) to T-6A (2001)
0.8
1.3
1.0
0.2
3.3
T-34C (1978) to T-45TS (2000)
0.8
1.4
1.0
0.2
3.3
Note that rows of tables in this chapter may not sum to the total values presented because
of rounding.
51
material, and equipment.) For example, in 1975, labors share of total

aircraft cost was calculated to be 52.3 percent.
In other words, out of the individual indices for labor, material,
and equipment, we create a composite cost index for each aircraft of
the pair. This composite index depends on the year procured and the
specic materials used in an aircraft. We determine the composite price
index by weighting the individual labor, material, and equipment indices by their shares of total aircraft cost. We then use the resulting composite index values to compute the overall annual cost escalation using
Equation 2.1. The component contributions to annual cost escalation
(i.e., the values shown for labor, material, and equipment) were based
on the ratio of each components annual percentage increases relative
to the overall total.
Of the three economy-driven factors, material contributes the
most to aircraft cost escalation, from 0.9 to 1.6 percent by the pairs we
consider. Equipment had the second biggest inuence, ranging from
1.0 to 1.1 percent, followed by labor, whose eects ranged from 0.8 to
0.9 percent. All three combined resulted in cost growth of 2.9 to 3.3
percent, or only a fraction of the 7 to 12 percent cost growth evident in
the aircraft we examine.
We next examine how customer-driven factors contribute to overall aircraft cost escalation in Table 5.2. Customer-driven factors include such
items as the technical characteristics and use of advanced materials that
we analyzed in Chapter Four. We calculate the customer-driven contributions to total cost escalation dierently from our calculations of
economy-driven contributions to total cost escalation. Rather than
using changes in representative cost indices, we employ the scaling factors derived in the previous chapter to understand how unit cost should
have changed. We examined how changes in empty weight, maximum
speed, carrier-based, airframe materials complexity, and production
rate should have inuenced the average annual cost escalation. For
example, if the maximum speed increased by 20 percent from one air-
52
craft to another, the results in Table 4.6 indicate that the average unit
cost should increase by 16 percent (1.2^ 0.83 1). Again, the reader
is cautioned that the technical complexity measures are associative and
not necessarily causal. Other factors could inuence aircraft cost, but
the terms we have identied in Chapter Four have the strongest statistical relationship.
In Table 5.2, we have also included a term called regulatory factors. This factor of 0.6 percent is based on a 1998 proprietary analysis by the Electric Boat Division of General Dynamics Corporation
that examined the changes in construction hours for four classes of
submarines from the late 1960s through the current Virginia class.
That approach categorized changes in the construction hours into several areas: technical advancement (e.g., performance improvements),
weapon systems (e.g., number and complexity of systems), stealth characteristics (e.g., signatures), survivability, quality assurance, oversight
control (e.g., reporting requirements), and regulatory environment
(e.g., environmental protection laws). For the regulatory factor shown
in Table 5.2, we selected only those areas that were broadly applicable to defense manufacturing, i.e., oversight control and regulatory
environmentin other words, general changes to the manufacturing
environment driven by the government (although not necessarily by
Table 5.2
Percentage Contributions to Annual Cost Escalation, by Customer-Driven
Factors
Comparison
F-15A (1975) to F-22A (2005)
Technical Airframe
CharacterComistics
plexity
0.6
Procurement
Rate
Regulatory
Total
4.3
0.9
0.6
6.4
F/A-18A/B (1983) to F/A-18E/F (2003)
0.1
2.1
0.5
0.6
3.1
B-1B (1984) to B-2A (1993)
5.0
6.1
1.7
0.6
3.5
C-130H (1980) to C-17 (2005)
5.9
1.1
0.5
0.6
7.1
E-3A (1975) to E-8C (2005)
0.6
0.0
1.1
0.6
2.3
E-2C (1975) to E-2C (2004)
0.3
0.3
0.6
0.6
1.9
T-34C (1978) to T-6A (2001)
2.9
0.0
0.4
0.6
3.1
10.5
0.1
0.6
0.6
11.9
T-34C (1978) to T-45TS (2000)
53
one of the Services). Despite the study being focused on submarine

issues, these changes to the general manufacturing environment should
be similar. We converted the values reported in the study to an annual
rate using Equation 2.1.
The contributions to annual cost escalation seen in Table 5.2 are
based on cost indices for each aircraft of the pair using Equations 4.2,
4.4, and 4.5 (and their associated parameters listed in Chapter Four).
The annual cost escalation for Technical Characteristics is based on
Equation 4.4. The Airframe Complexity component is based on Equation 4.5. The annual cost escalation due to procurement rate changes
is based on Equation 4.2. Again, the ratio of the index values was converted to an annual rate using Equation 2.1.
The contributions of customer-driven factors to cost escalation range considerably by comparisonas one might expect given
the variety of aircraft and the dierent system histories. The technical characteristics resulted in a 5.0 percent contribution to cost deescalation for the B-1B and B-2A bomber comparison2 but a 10.5
percent increase for the T-34 to T-45 comparison. Procurement rate
factorsa direct consequence of annual buy decisionsrange from
0.5 percent for the C-130H and C-17 cargo plane comparison to 1.7
percent for the bomber comparison. Airframe complexity contributed most to cost escalation in the F-15A and F-22A comparison and
the B-1B and B-2A comparison; not surprisingly, the F-22A and the
B-2A are the two aircraft in the comparison table containing the most
advanced materials.
Total Escalation
Table 5.3 summarizes economy-driven factors and customer-driven
factors. We also note the contribution of learning over the course of
2
This de-escalation is due to a decrease in the maximum speed of the aircraft. Note, however, that the annual increase resulting from airframe complexity, 6.1 percent, more than
osets this reduction.
54
Table 5.3
Percentage Contributions to Annual Escalation Rate
Comparison
Economy- CustomerDriven
Driven
Factors
Factors
Cost Improvement
Correction
Predicted
Actual
F-15A (1975) to F-22A (2005)
3.5
6.4
0.1
10.0
9.9
F/A-18A/B (1983) to F/A18E/F (2003)
3.1
3.1
0.0
6.2
4.7
B-1B (1984) to B-2A (1993)
3.1
3.5
0.0
6.5
10.7
C-130H (1980) to C-17 (2005)
3.2
7.1
0.2
10.2
12.8
E-3A (1975) to E-8C (2005)
3.4
2.3
0.3
5.4
6.0
E-2C (1975) to E-2C (2004)
3.4
1.9
0.0
5.3
6.4
T-34C (1978) to T-6A (2001)
3.3
3.1
0.1
6.5
6.7
T-34C (1978) to T-45TS

(2000)
3.3
11.9
0.0
15.2
15.9
production, shown in the Cost Improvement Correction column. This

correction adjusts for the fact that, in some cases, we are comparing the
aircraft programs at dierent points in the production run. For example, it would be misleading to compare the rst or second airframe of
an aircraft type to the 300th. This correction term puts the comparison
on a uniform production experience level.
In comparing the aggregate of these estimates with historical
escalation, we see that, except for the bomber comparison, predicted
escalation tracks actual escalation closely. Four of the eight comparisons come within 1 percent of actual escalation; only for cargo and
bomber aircraft did the dierence between actual and predicted escalation exceed 2 percent. The small dierence between the actual total
cost escalation and that predicted by our model of disaggregating the
contributions to economy-driven and customer-driven factors lends
support to the validity of our disaggregation method and subsequent
attribution of cost escalation to each set of factors.
The range of economy-driven factors is narrow, from 3.1 to 3.5
percent. The range of customer driven factors is wide, from 1.9 to 11.9
percent. This implies that the wide variability in cost escalation between
pairs is almost entirely due to diering estimations of customer-driven
factors. Pairwise comparisons with smaller actual escalation have a
55
larger share of that escalation explained by economy-driven factors; the

proportion of escalation explained by economy-driven factors ranges
from one-fth (or 3.3 percent or a total of 15.9 percent) for the T-34C/
T45TS comparison to four-sevenths (or 3.4 percent of a total of 6.0
percent) for the E-3A/E-8C comparison.
These trends all imply that the main source of cost escalation for
aircraft is customer-driven factors or, ultimately, the increased performance and capability that has proceeded from decade to decade. This
increased capability has come, in some cases, at a considerably increased
cost. By contrast, economy-driven factors have remained fairly consistent across the aircraft types. In the next chapter, we explore aircraft
manufacturers perspectives on these and related issues.
CHAPTER SIX
Industry Views on Military Fixed-Wing Aircraft

Cost Escalation
In addition to quantifying the variables that contribute to aircraft cost

growth, we also visited and held lengthy discussions with the major
prime contractors for xed-wing military aircraft production. In this
chapter, we examine what representatives of these rms said are the
most signicant reasons for aircraft cost growth. We also review other
research concerning what the industry is doing to reduce or avoid further cost increase.
Military Fixed-Wing Aircraft Industry

The military xed-wing aircraft industry has undergone considerable
consolidation since World War II. In the 1940s, there were 16 prime
contractors in military aircraft manufacturing, and even in 1990 eight
of these remained (Birkler et al., 2003). Today, there are three prime
aircraft manufacturing contractors with active nal assembly and
checkout production capabilities: Boeing, Northrop Grumman, and
Lockheed Martin.
Boeing produces the F/A-18E/F for the U.S. Navy, the C-17 for
the U.S. Air Force, and the F-15 for foreign military sales. Lockheed
Martin produces the F-22A and C-130J for the U.S. Air Force and
the three dierent variants of F-35, conventional take o and landing
(CTOL), carrier version (CV), and short take o and vertical landing
(STOVL) for the USAF, USN, and U.S. Marine Corps as well as for
57
58
the United Kingdom Royal Navy and the Royal Air Force. Northrop
Grumman, in addition to being responsible for the nal assembly of
the E-2C aircraft for the USN, is a major partner with Boeing, producing the aft fuselage for the F/A-18E/F, and with Lockheed Martin,
producing the center fuselage for the F-35.
Each of these prime contractors responded to a RAND questionnaire, reproduced in Appendix B, and shared their opinions during
interviews on why the cost of xed-wing aircraft has increased and
what can be done about it. We have grouped these views into two categories: increased military utility and government requirements.
Increased Military Utility

The technological revolution in design tools, electronics, and propulsion technology has given aircraft designers the opportunity to create
weapon systems of increasingly greater lethality and capabilities. This
is especially true for combat aircraft such as the F-22A, the F/A-18E/F,
and the F-35, whose complexity is substantially higher than that of
their predecessors. For example, the F-16, the initial lightweight ghter
of the 1970s, had 15 subsystems and thousands of interfaces but less
than 40 percent of its functions managed by software. Today, the
F-35 has 130 subsystems, hundreds of thousands of interfaces, and
more than 90 percent of its functions managed by software. Furthermore, newer ghter aircraft include sophisticated electronic warfare
capabilities, perform with higher thrust engines, and can evade the
enemy radar far better than can legacy ghters such as the F-14, F-15,
and F-16.
Stealth
Most industry observers note that stealth features have been major
contributors to the complexity of combat aircraft. The new class of
combat aircraft is much stealthier and evades enemy radar better than
their predecessors. In addition to the tactical benets that stealth
oers, it also makes the system much more survivable. New computer-
Industry Views on Military Fixed-Wing Aircraft Cost Escalation
59
aided design (CAD) tools allow for three-dimensional design capability that has enabled designers to move away from nonaerodynamic
stealth airframe design such as the F-117 to the highly aerodynamic and
stealthy airframe of the F-22A and the F-35. CAD information can be
directly fed to computer-aided manufacturing (CAM) hardware, such
as ve-axis, high-speed milling machines and automated composite
fabrication tools. These tools can produce parts with extreme dimensional precision and tolerances, which lead to minimal or no appreciable gaps and mismatches, thus eliminating the need for any solid
shims. Gaps and mismatches in the surface increase radar visibility,
so eliminating them is a goal of any stealthy design. Improvement in
radar absorbing materials (RAM) and their application in manufacturing has substantially beneted aircraft stealth. Adding these features
requires additional nonrecurring design labor costs as well as recurring
labor and materials costs, which all contribute to additional overall
costs.
Weight Reduction
Weight has been the enemy of aircraft from the dawn of aviation.
Designers constantly seek to reduce weight to increase performance.
Yet modern aircraft require y-by-wire capabilities and sophisticated
electronics to enable them to communicate in hostile environments,
jam enemy radars, and avoid surface-to-air missiles. These electronics
add weight and require space on the platform. Increased maneuverability and speed also require complex, often heavy, propulsion systems.
The aircraft industry continuously introduces new lightweight
materials and innovative structural designs to military airframes. Chief
among these are composites, many of which are embedded in a resin
matrix. This combination of materials is surprisingly lightweight in
comparison to metals used to make the same parts, but their weight
advantages come at a cost. Younossi, Kennedy, and Graser (2001)
found that parts made from composite materials are generally more
costly to design and manufacture in comparison to metal parts. Figure
6.1 shows that the use of composite materials in aircraft has increased
over time.
60
Figure 6.1
Trend in Composite Material Use in Aircraft, 1967 to 2000
60
V-22 FSD
Structural weight, %
50
Air Force
Navy
Marine Corps
V-22
40
B-2A
30
YAV-8B
A-12
F-35 (CV)
F-35 (CTOL)
F-35 (STOVL)
AV-8B
F-22
YF-22/YF-23
20
F/A-18E/F
F/A-18A/B
10
C-17
F-117
F-111 F-14 A-10 F-15
B-1B
F-16
0
1965
2000
Year
NOTE: FSD = full-scale deployment.

RAND MG696-6.1
Titanium is a lightweight, high-strength material that can withstand high temperatures. As Figure 6.2 shows, manufacturers used
titanium in signicant amounts in the late 1960s to produce the Navys
F-14 and the Air Forces F-15 aircraft and it is being used increasingly
once again. The requirement for speed and high-temperature engines
increased the need for it in the airframe structure, especially in the aft
fuselage (Younossi, Kennedy, and Graser, 2001).
The F-16 and F/A-18A/B aircraft, however, emphasized aordability, so performance requirements were modied to permit the use
of aluminum rather than titanium. More recently, the F/A-18E/F and,
especially, the F-22A aircraft have emphasized performance, leading to
more use of titanium.
In recent years, the commercial demand for titanium has increased
substantially. This new demand is often called the golf club eect
in reference to the use of the metal in golf clubs. Other manufacturers, including those making bicycle frames and tennis rackets, have
also increased their use of titanium. Furthermore, the commercial
61
Figure 6.2
Trend in Titanium Use, 1967 to 2000
60
Structural weight, %
50
Air Force
Navy
Marine Corps
40
F-15
F-22
30
F-35 (CV)
F-14
YF-22
20
F-35 (CTOL)
B-1B
F/A-18E/F
F-35 (STOVL)
F-117
10
V-22 FSD
F/A-18A/B
F-16
AV-8B
V-22
0
1965
2000
Year
RAND MG696-6.2
aircraft industry has increased the use of titanium (Tran-Le and Thompson, 2005). Demand in China has increased also; titanium is used as
an additive in steel production (Tran-Le and Thompson, 2005). This
commercial demand has increased the need for raw material and has
placed a tremendous burden on its suppliers. Every aircraft manufacturer we interviewed during this study raised concerns about the eect
of commercial demand for the metal on aircraft manufacturing costs
and schedules.
Lean Manufacturing
The principles of lean manufacturing were rst introduced by Toyota, a

Japanese automotive manufacturer, to improve the quality and aordability of their vehicles. These principles are a collection of activities
focused on improving the entire enterprise, including
t product development, by incorporating ease of manufacturing
and quality control during the design phase using integrated
product teams (IPTs)
62
t continuous improvement during production, as recommended by

production workers, through Kaizen events
t kitting of the material and tools required during assembly in especially designed Kanbans for ease of access
t design of the assembly for a streamlined product ow
t procurement of material from suppliers with whom the prime
contractor has formed a strategic alliance, eliminating the need
for rigorous quality checks and a large inventory of parts and
materials through just-in-time delivery of materials and parts.
As with the automotive industry, the aerospace industry has
implemented some or all of these initiatives in processing parts and
subassemblies. Nearly all aircraft prime contractors and their major
vendors have embraced at least the principles of lean manufacturing,
realizing some cost savings. Although there is much anecdotal evidence
and many pilot project results, there is no evidence that these principles
were applied to the entire aircraft manufacturing enterprise (Cook and
Graser, 2001). Although lean manufacturing approaches show promise
for future savings, we could nd no evidence in our historical examination that they measurably reduced prices.
Government Requirements
Aircraft manufacturers highlighted four areas of government policy as
potential sources of aircraft cost growth. These were the Berry Amendment and Buy American legislation, Occupational Safety and Health
Administration (OSHA) requirements and environmental regulations,
antitamper requirements, and International Trade in Arms Regulations (ITAR). We discuss each of these below.
Berry Amendment and Buy American Legislation
Two laws restrict how aircraft manufacturers may purchase materials:

the Berry Amendment (10 U.S.C. 2533a) and the Buy American Act
(41 U.S.C. 10a10d). Within the aerospace industry, the Berry Amendment applied mainly to specialty metals, requiring that these materials
63
be smelted in the United States. Recently, the specialty metals provisions was removed from this amendment and placed within 10 U.S.C.
2533b. The Buy American Act, originally passed in 1933 and amended
a number of times since then, also restricts the purchase of material
by aerospace contractors, particularly the purchase of material from
overseas vendors. These two laws require that manufacturers of military aircraft forgo potentially less-expensive foreign sources of products
that meet technical specications in favor of U.S.-made materials. For
certain critical industries, including aerospace, this requirement may
ow down several levels to the subcontractor/supplier level. The laws
have been in eect for decades, but several industry groups are currently seeking relief from some of them. The exemptions they are seeking include
t an exemption for commercial items
t an exemption that would allow manufacturers of dual-use items
to commingle foreign and domestic specialty metals in production, provided they acquired the appropriate quantity of domestically smelted specialty metals
t a de minimis exemption allowing for delivery of an item containing noncompliant specialty metals if they constitute less than 2
percent of the total amount of specialty metals in an item.
OSHA and Environmental Regulations
Industry representatives also cited OSHA and environmental regulations as potential sources of cost growth. OSHA regulations ensure
the health and safety of workers in the aerospace industry and include
measures to help prevent employees from falling o large aircraft as
well as requirements for respiratory gear and proper ventilation around
such volatile chemicals as paints. These regulations could increase costs
if new, exotic materials involve the greater use of volatile chemicals
and, hence, investment in more protection for workers. Information
about the direct eect of these regulations on aerospace manufacturing
is sparse. Environmental regulations that control the emission of volatile chemicals into the atmosphere also mandate the need for an envi-
64
ronmental control facility for hazardous materials. The need for such a
facility increases both direct and indirect costs of the products.
Antitamper Requirements
Antitamper requirements are relatively new policies to protect critical U.S. technologies by preventing their unauthorized access and
duplication. Such measures may include encrypting software, coating
of chips, or attaching explosives to sensitive compartments. Anti-tamper requirements are designed to prevent an enemy from studying and
duplicating the systems on a downed aircraft and subsequently reverseengineering them. Currently, antitamper measures are voluntarily
implemented according to guidelines included in the Defense Acquisition Guidebook. The lack of a standardized denition for critical
technology, however, has made implementation of these requirements
dicult. Antitamper requirements can increase costs by requiring
additional design and manufacturing work and making maintenance
and replacement more dicult.
International Trade in Arms Regulations
ITARs implement the Arms Control Export Acts (ACEA) requirements. They control the transfer of technologies, both military and
commercial, to other nations. The actual regulations may be found
under 22 CFR Parts 120130. Globalization has made their application increasingly dicult in recent years, as manufacturers with globalized supply chains nd it dicult to include their oshore plants
and engineers in production and design decisions. This may increase
expenses by complicating the design process or requiring greater controls on information. Many aerospace industry representatives whom
we interviewed argue that these regulations are unnecessarily restrictive, because many restricted components are already manufactured by
other countries. This regulation also aects what military products can
be marketed and sold through foreign military sales.
65
Summary
Industry representatives we interviewed indicated three areas that have
contributed to cost escalation for xed-wing aircraft: (1) a diminishing
industrial base (both at the prime and supplier levels), (2) increased
military capability, and (3) broader government regulations. The
second point is in agreement with our observations from the previous
chapterthat customer-driven factors heavily inuence the magnitude
of the cost escalation. The other two areas, although important but
dicult to quantify, are also viewed as contributing to price changes.
However, the analysis presented in the chapters above suggests that
these areas are less important to cost escalation.
CHAPTER SEVEN
Options for the Air Force and the Navy to Reduce

Fixed-Wing Aircraft Costs
We did not conduct an exhaustive search of ways that the two Services
might reduce the costs of their xed-wing aircraft, but our interviews
with aircraft manufacturers and other knowledgeable sources elicited
seven preliminary suggestions:
1. Make xed-wing aircraft procurement more stable and predictable
2. stabilize project management and design
3. rethink competition within the industrial base
4. encourage international competition and participation
5. improve the process of formulating requirements and capabilities process
6. focus attention on upgrades and commercial derivatives
7. increase the use of evolutionary acquisition principles.
Some of these ideas are highly speculative and, given the current
scal and legislative environment, have dubious prospects for implementation. Nonetheless, for completeness of discussion, we address
each below, pointing out both positive and negative aspects. We also
discuss lessons learned and documented in previous RAND research
on aircraft development programs. Note that some of these approaches
may have to be addressed early in the acquisition phase to have a meaningful eect on the procurement cost.
67
68
Make Fixed-Wing Aircraft Procurement More Stable and

Predictable
Manufacturers of military aircraft face two problems with respect to
stability. First, initially planned quantities often dier greatly from the
actual quantities produced. For example, the F-22A program initially
anticipated that 659 units would be produced in the initial Selected
Acquisition Report (SAR) dated 1991. By the 2005 SAR, the total
number of aircraft had fallen to 186a reduction of nearly two-thirds.
The C-17 program planned 211 units in the initial SAR dated 1985. By
1990, the quantity had dropped to 121, dropped further to 41 by 1993,
but rose to 181 by 2002.
Second, annual production quantities often vary signicantly
from year to year. For example, the F/A-18E/F scal year buys, excluding those for the rst two years, varied from 30 to 48 units. The C-17
acquisition for the Air Force varied from 6 to 15 units each scal year
(again excluding the initial procurement years). The F-22A procurement prole has ranged from 10 to 24 since the initial development
production. In most of these cases, the production line was designed
and equipped to accommodate a higher production rate than has ever
been realized.
Such variability in procurement can add to the unit cost in two
ways. First, by overestimating the actual quantity that will be produced,
manufacturers scale facilities to a rate of production that will not be
met. Hence when production quantities are lower than planned, manufacturing is not operating at the most economic point; for example,
facilities are less than fully productive and ineciencies are introduced
through potential gaps or idle time. Second, the year-to-year variability
in acquisition makes it dicult for the rms to manage their stas and
suppliers. With work levels varying by 50 percent or more, manufacturers sometimes must choose between releasing sta or placing them
on overhead until the procurement rate increases again. By developing a realistic and steady production plan, the Services would permit
manufacturers to size their production to a more economic scale and
possibly take advantage of the corresponding higher productivity.
Options for the Air Force and the Navy to Reduce Fixed-Wing Aircraft Costs
69
Multiyear buys or advance appropriation could introduce a

greater semblance of stability to aircraft manufacture (Blickstein and
Smith, 2002). The Air Force and Navy could also improve stability by
resisting changes to its xed-wing aircraft procurement plan with each
budget cycle. There are other advantages to multiyear procurements
such as the option to purchase larger quantities of material and equipment (through Economic Order Quantities) to leverage eciencies in
the supply base. The Services have successfully used multiyear procurement to purchase F/A-18E/F, C-130J, AV-8B, KC-10, C-17, F-15,
E-2C, and F-16 aircraft. The proposed savings reported range from 5.5
to about 11.9 percent for ghter and attack aircraft. The range of savings was even wider (5.5 to 17.9 percent) when we include other xedwing aircraft going back to 1982 (Younossi et al., 2007).
This procurement strategy would permit the Air Force, Navy, and
aircraft manufacturers to establish contractual agreements for future
aircraft over several years. Under multiyear procurement, Congress
authorizes all procurement quantities and funding necessary in the rst
year but appropriates funds annually. Because multiyear procurements
establish penalties against the Service for not procuring the specied
number of aircraft, and because Congress rarely backs away from such
an agreement once it is implemented, such agreements give manufacturers greater condence in making investments and allow them to
increase their purchasing leverage with suppliers. Multiyear procurement is advantageous to the aircraft industry because the method provides some advance notication of future procurement by the Air Force
and the Navy, and thereby permits long-term planning and investment
by manufacturerssuch as optimized production schedules, more
cost-eective ways of procuring materials and parts, and avoiding the
need to prepare and negotiate proposals for the future lots (Younossi
et al., 2007).
There are some downsides to the multiyear procurement approach,
however. The savings resulting from multiyear procurement are very
dicult to validate (Younossi et al., 2007). Most of the savings values
are based on estimates and not actual costs. This diculty is part of
the reason the range of savings resulting from multiyear procurement
is so large. A second downside of multiyear procurement is that the
70
approach limits future exibility by the Congress or the Services. For

example, if there were an emerging nancial need during the contract
that required a cut to the procured quantities, the government would
have to pay substantial penalties to realize the reduction. A further
downside is that multiyear procurement is eective only where there
is design stability. Thus, the approach is really applicable only to programs when the production conguration is xed.
Stabilize Project Management and Design

One possible way the Air Force and the Navy could potentially limit
cost escalation is through improved management stability. For example,
the Air Force and the Navy could curtail their current practice of rotating ocers through jobs every three to four years. Because a xed-wing
aircraft may take up to two years to build and a program may stretch
over decades, a single program could have several program managers,
each with diering management methods and philosophies.
Management initiatives might also help reduce the number of
changes during production. Although changes may result from new
military threats or experience with earlier units, they are disruptive and
expensive to the program. For example, we have seen that the total cost
growth on military aircraft averages about 35 percent over the entire
procurement (Arena et al., 2006b). Although not all of this growth
is due to changes, some of it certainly is. During the early years of
the Reagan administration, the Secretary of the Navy insisted on personally approving every change to a ship or aircraft contract. This, in
eect, transferred such decisions away from acquisition ocials and to
his oce. Such drastic measures can control changes.
Rethink Competition Within the Industrial Base

A recurring issue since the end of the Cold War has been the consolidation of the defense industrial base to reect the defense budget
71
reductions during the 1990s. Considerable consolidation has already

occurred within the xed-wing aircraft base, with the loss of Martin
Marietta, LTV, and McDonnell Douglas from the manufacturer base.
Even Northrop Grumman, while remaining in the industrial base,
builds portions of aircraft for Boeing and Lockheed rather than complete aircraft under its own name (except for the E-2). As a result,
the United States now has only two xed-wing ghter/attack aircraft
manufacturers with a full systems integration capability. Prime aircraft
manufacturing rms have noted a similar consolidation in their supplier base.
Given this dramatic consolidation, maintaining a competitive
environment at the prime level remains challenging. Indeed, new
design programs are few and far between. Thus, the loser of a competition faces bleak nancial prospects. The use of competition at the
prime level as a means of cost control will be limited. Given these realities, the Air Force and the Navy might encourage competition at the
subsystem level (e.g., electronics and major subsystems). Yet, it is also
unclear whether this subtier of the industrial base is robust enough
for such a strategy, as some of these vendors are already sole sources
for unique materials and specialized critical components (Oce of the
Under Secretary of Defense, 2007a).
There is an additional issuethat of the stability and suciency of
the design base. With only the three manufacturers mentioned above,
there is also a concern over maintaining a reasonable technical capability for aircraft design, especially as demands for new designs become
less frequent. With the F-35, F-22A, and F/A-18E/F as the only hightechnology aircraft currently in production for the U.S. military, and
no new designs for manned aircraft on the horizon, the United States
must remain concerned about who will design its next-generation aircraft.1 If DoD decides to maintain such a small base, it will come at a
cost.
A few F-16s and F-15s are being produced for foreign military sales.
72
Encourage International Competition and Participation

For security reasons, Air Force and Navy aircraft production is limited
to the United States. This makes sense from a defense perspective, but
it limits the competition and innovation that might be realized from
procuring aircraft in a global market. Allowing the Air Force and the
Navy to buy from foreign companies might help increase competition
and reduce costs. Some of our North Atlantic Treaty Organization
(NATO) allies currently build both military and commercial aircraft,
and companies such as BAE Systems and EADS have established U.S.
companies or counterparts to bid on U.S. aircraft production, such as
the new tanker aircraft. Competition or participation by foreign suppliers might have a positive eect on the U.S. xed-wing aircraft industry. The globalization of the defense industry is, in general, an evolving
process. For example, BAE and Rolls Royce, both UK companies, are
participating in the development of the Joint Strike Fighter (JSF or
F-35). Foreign companies are also increasingly buying U.S. companies.
The purchase of United Defense by BAE Systems is one example. Still,
involving foreign rms in the manufacture of U.S. xed-wing aircraft
could raise issues of access to and control of highly sensitive technology.
Also, as discussed above, the Berry Amendment (10 U.S.C. 2533a), the
Buy American Act (41 U.S.C. 10a10d), and ITAR may limit initiatives and require waivers for participation of foreign rms.
Improve the Process of Formulating Requirements and

Capabilities
The combined aviation capabilities of the U.S. Air Force and the U.S.
Navy have no peer throughout the militaries of the world. Yet, the
United States continues to build aircraft that push the edge of technology and capability. This technical advancement has its price in both
schedule and cost. As a result, the concept of controlling the processes
for new technologies has gained some popularity. For example, the
Navy has instituted a board structure to review and approve initial
73
design and subsequent changes in requirements for new and in-service

aircraft (and ships). At issue is whether each change is worth the cost,
technical risk, and schedule change. These boards also consider the
eects of changes on the contract with the builder. This consideration
was called Cost as an Independent Variable (CAIV) in the mid-1990s
when DoD attempted to view each incremental change in capability in
terms of its cost.
It is too early to tell whether these boards are having the desired
eect. The United States has technical supremacy in the aviation world
today, and this technical edge would allow the Services to reduce the
number of aircraft to meet mission needs and to reduce the risk to
pilots and crews. Decisions on the numbers and capabilities of aircraft
must include both military and business issues that require a high level
of consideration within the two Services. In the case of the JSF/F-35,
there has been some pressure to reduce the buys of aircraft and move
the program further to the right in schedule terms, that is, to delay
some production because of cost growth and increased technical risk.
Focus Attention on Upgrades and Commercial Derivatives

Another way to reduce the costs and technical risks of replacement
aircraft would be to focus on modications or changes to existing
aircrafton an upgrade2 rather than buying new. The Services have
pursued this strategy for some time, as with the B-52 in the Air Force,
the E-2D in the Navy, and the CH-46 in the Marine Corps. However,
the age of existing aircraft sometimes makes an upgrade option less
feasible.
The Services have sometimes used commercially derived, modied airframes for military purposes. For example, the Air Force E-8
Joint STARS uses the B-707 airframe. Further, the Navy multimission maritime aircraft (MMA), a replacement for the P-3 Orion,
By upgrade, we mean either a retrot of an airframe already in the eet (such as the
upgrades to the B-52) or the introduction of a new series (such as the change from E-2C to
E-2D).
74
will be using a B-737 airframe. Also, several commercial candidates

exist for the U.S. Air Force tanker replacement program.
Both Services pursued a continuous cycle of technology improvements during the Cold War to match external threats. With no nearpeer competitor today, they may be able to slow the pace of innovation,
focusing instead on upgrades of existing aircraft to save cost and reduce
technical risk. There is a severe downside to such a strategy, however.
The design and engineering parts of the industrial base cannot simply
lie fallow for a period of time and be expected to sprout anew when
needed. Consolidation of remaining manufacturers and the elimination of any semblance of competition would also probably occur with
this strategy.
Increase the Use of Evolutionary Acquisition Principles3

This concept aims to reduce the time between the identication of new
operational needs and the delivery of an operational weapon system
that meets those needs. Instead of a one-step process to building full
capability, evolutionary acquisition aims to achieve desired capabilities over time, with phases of development divided into increments
of operational capabilities (Lorell, Lowell, and Younossi, 2006). Once
threshold capabilities are elded, feedback from operational experience
can be incorporated into the next increment or phase of development,
thus reducing technical risk and increasing the probability of better
cost estimates.
The benets of this approach include a reduced likelihood of major
research and development (R&D) delays and cost overruns. Drawbacks
include the possibility that contractors may allocate easy development
tasks to early iterations, leaving more dicult tasks for later stages,
in hopes of getting the Service to commit to the project. Managers at
both the contractor and Service levels are more important players in an
evolutionary acquisition process. The exibility in nal requirements
3
For a more complete discussion, see Lorell, Lowell, and Younossi (2006).
75
and capabilities could cause scope and cost creep. Finally, Congress is
not generally happy with a program that is not fully dened at the time
of authorization/appropriation and therefore may be less likely to support this approach when it involves large aircraft programs.
Lessons Learned from the F-22A and F/A-18E/F

Development Programs
In addition to the ideas noted above by aircraft manufacturers whom
we interviewed, we note here some ideas from lessons learned from previous RAND research on aircraft development programs. Specically,
earlier RAND research (Younossi et al., 2005) on the F-22A and F/A18E/F programs may oer some more general lessons on controlling
aircraft cost growth. Among its ndings
t Each program used dierent methods to solicit contractor proposals and to divide the work among contractors in the design
phase. The F/A-18 drew on preexisting relationships among its
contractor base to minimize technology risks. It also implemented
a strong CAIV strategy.
t Concurrent development of new technology created greater technical challenges for the F-22A, whereas incremental improvement
reduced technical risk in the F/A-18E/F. Stealth requirements
accounted for a big part of the dierence between the two programs. The lower risk for the F/A-18E/F may have contributed to
its stable cost and schedule.
t The programs allocated dierent portions of their budgets for
management reserves. The F-22A allocated only 2 percent of its
budget to management reserve whereas the F/A-18E/F maintained
a 10 percent reserve. Thus, the F/A-18E/F was more likely to have
the reserves to x program problems. Given the dierences in the
risk for the two programs, it seems that the F-22A should have
had the higher proportion for management reserves.
76
t General program lessons included

Early, realistic, cost and schedule estimates set programs on the
right path for development.
A stable development team structure, proper team expertise,
clear lines of responsibility and authority, and a lead contractor responsible for the overall program progress are critical to
program success.
An experienced management team and contractors with
prior business relationships help eliminate early management
problems.
Concurrent development of new technology for the airframe,
avionics, and propulsion adds signicant risk.
Reducing the cost and risk of avionics should be a key focus of
the concept development phase. Avionics is a key cost driver of
modern weapon systems.
Preplanned, evolutionary modernization of high-risk avionics
can reduce risk and help control costs and schedules.
Careful monitoring of such major cost drivers as airframe
weight is important. Dramatic changes to the planned airframe
weight over time are an early indicator of problems.
Earned value management (EVM) data should be used to
monitor and manage program costs at the level of integrated
product teams.
Summary
There are several ways to potentially reduce xed-wing aircraft costs,
but few are plausible or palatable given current realities. For example,
neither Congress nor DoD are likely to permit competition to build
U.S. military xed-wing aircraft in the international market. Prior
attempts to balance cost and requirements have met with limited success (e.g., the CAIV concept has been around several years). Focusing
attention on upgrades of existing systems could result in systems that
are less capable or might have higher ownership cost. Although the
nation, the Air Force, and the Navy understandably desire technology
77
that is continuously ahead of current and potential competitors, this

comes at a cost. We do not evaluate whether the cost is too high or low;
we note only that it exists.
CHAPTER EIGHT
Conclusion
We have found that the cost of military aircraft has increased in recent
decades by 7 to 12 percent. This is about twice the rate of common
ination measures during this same time. Given that long-term
defense investment spending, although somewhat cyclical, will remain
relatively constant, this means that the government can aord to buy
fewer, increasingly expensive aircraft. In fact, in peak defense funding years, the government is now purchasing fewer aircraft than it did
during trough years of a few decades ago.
To address the sources of cost escalation in military aircraft, we
analyzed both economy-driven variables, or those over which the Services have little control, and customer-driven variables, or those that
they can inuence. Among economy-driven variables are labor, material, equipment, and manufacturer fees and prots. Although labor
costs have grown slightly faster than ination, these increases have been
oset by gains in productivity. The costs of materials and equipment
grew at rates slightly below that of ination but have recently exceeded
it. The limited data available on fees indicate that changes in these
costs have not accounted for much cost growth, probably as a result
of how DoD establishes fees on its contracts. Altogether, our research
found that the contributions of economy-driven variables to aircraft
cost escalation were slightly below the rate of consumer ination.
By contrast, customer-driven variables, such as the technical characteristics of an aircraft, procurement rates, and complexity of the airframe, have contributed substantially to cost escalation. We found that
customer-driven variables contributed more to cost escalation than did
79
80
economy-driven variables for four of eight comparison pairs of aircraft

that we examined. The economy-driven escalation was fairly consistent
over the timeframe studied and for all pairsroughly 3.5 percent. The
customer-driven contributions varied widelyanywhere from about 2
to 12 percent.
Interviews with aircraft manufacturing industry representatives
conrmed the contribution of customer-driven variables to aircraft
cost escalation. Specic features said to be driving costs include those
for greater stealth of aircraft as well as demand for composite materials that can help reduce aircraft weight. Beyond these specic government requirements, broader government policies, such as those requiring U.S. materials, occupational and environmental regulations, and
safeguards to protect U.S. critical technologies, are felt to contribute to
cost escalation.
We examined several options to reduce cost escalationnone of
them are a panacea. If the Services wished to benet from more competition, they could encourage international competition and participation in the construction of military aircraft. Given likely opposition
from Congress to such an initiative, and hence a very low likelihood
for its implementation, the Services may wish to consider other means
to reduce costs, such as stabilizing procurement rates and management,
or incorporating lessons learned into development programs for some
aircraft.
One more direct way to control cost escalation would be to curb
requirements growth over successive generations of aircraft and focus
on incremental improvements. Such an approach, however, could slow
the pace of innovation and potentially risk losing the technical edge
the United States has over potential threats.
The Services have been moving toward reduced procurement
quantities as a way to stay within annual procurement budgets. These
reductions have helped them continue to procure aircraft that remain
far superior to that of any other military in the world. Maintaining
such capabilities is not, of course, a bad thing. Technological superiority in the air means that the United States can continue to deter and
defeat adversaries. But it does result in increased unit costs, as we have
Conclusion
81
attempted to illustrate. Knowing this can help the Services make the
increasingly dicult choices between individual aircraft capabilities
and total numbers of aircraft.
APPENDIX A
Aircraft Included in the Analysis
In Chapters Two and Four, data limitations required that we use varying subsets of aircraft systems for dierent parts of the analysis. These
subsets are chunks (not random samples) that were selected on the
basis of (1) data availability, (2) time-period suitability, and (3) consistency with the requirements of regression models.1 In Table A.1, we
present the entire universe of aircraft considered in the analysis and
indicate which aircraft form the chunks used in analyses of (1) broad
cost trends, (2) cost improvement slope, (3) cost improvement slope
with procurement rate (PR) slope, (4) basic technical characteristics,
(5) advanced materials, and (6) mission type. In Table A.1, an X
indicates that an aircraft type is used in the analysis described in the
column heading; those analyses can be found summarized in the table
or in the tables listed in that column. Mission types are broad categorizations, as both tanker and transport missions are listed as cargo. For
many current aircraft, the patrol mission has been integrated into the
electronic mission, yet the distinction has been maintained for older
aircraft.
1
See Deming (1950), p. 14: A judgment-sample is planned with expert judgment. A chunk
is dictated by convenience.
83
84
Table A.1
Aircraft Systems and Types Used in the Analyses
Cost Trends CI (at Least
(Tables 2.1,
5 Lots)
2.2),
(Table 4.1),
67 Aircraft 52 Aircraft
Types
Types
X
CI and PR
(at Least 5 Technical CharLots) (Tables
acteristics
4.2, 4.3),
(Table 4.6),
24 Aircraft
93 Aircraft
Types
Types
X
Mission
Type
Attack
A-10A
A-1G
A-1J
A-3A
A-4E
A-4M
A-5A
A-6A
A-6E
10
11
12
A-7D/K
13
A-7E
14
AC130H/U
15
AF-2
16
AV-8A
17
AV-8B
18
B-1A
19
B-1B
20
B-2A
21
B-52
22
B-58
23
C-12F
24
C-130E
Cargo
25
C-130F
Cargo
26
C-130G
Cargo
27
C-130H
Cargo
28
C-130J
Cargo
29
C-137E
30
C-141A
Advanced
Materials
(Table 4.7),
49 Aircraft
Types
Attack
Attack
X
Attack
Attack
Attack
Attack
Attack
A-7A
Attack
A-7B
X
X
X
X
Attack
Attack
Attack
Attack
Attack
Attack
Attack
X
X
Bomber
Bomber
Bomber
Bomber
Bomber
Attack
Cargo
Cargo
X
Cargo
85
Table A.1Continued
(Tables 2.1,
5 Lots)
2.2),
(Table 4.1),
Types
Types
CI and PR
acteristics
4.2, 4.3),
(Table 4.6),
24 Aircraft
93 Aircraft
Types
Types
Mission
Type
Cargo
31
C-17
32
C-2AR
33
C-29A
34
C-32B
35
C-37
36
C-40A
37
C-5A
38
C-5B
39
C-9
40
CT-39E
Cargo
41
CT-39G
Cargo
42
E-2A
43
E-2C
44
45
46
47
Advanced
Materials
(Table 4.7),
49 Aircraft
Types
Cargo
Cargo
Cargo
Cargo
Cargo
X
Cargo
Cargo
Cargo
Electronic
Electronic
E-3A
Electronic
E-4A/B
Electronic
E-6A
E-8B
48
E-8C
49
EA-6A
50
EA-6B
51
EC-121K
Electronic
X
X
X
X
Electronic
Electronic
Electronic
X
X
X
Electronic
Electronic
52
EC-130
Electronic
53
F/A18A/B
Fighter
54
F/A18C/D
Fighter
55
F/A18E/F
Fighter
56
F-104
57
F-111A
58
F-111F
59
F-117
60
F-14A
Fighter
X
Fighter
X
X
Fighter
Fighter
X
Fighter
86
Table A.1Continued
(Tables 2.1,
5 Lots)
2.2),
(Table 4.1),
Types
Types
CI and PR
acteristics
4.2, 4.3),
(Table 4.6),
24 Aircraft
93 Aircraft
Types
Types
Advanced
Materials
(Table 4.7),
49 Aircraft
Types
Mission
Type
Fighter
61
F-14A+
62
F-14D
Fighter
63
F-15A/B
64
F-15C/D
65
F-15E
66
F-15E
(ADV)
67
F-16A/B
Fighter
68
F-16C/D
Fighter
69
F-16N
70
F-22A
71
F-4A
Fighter
Fighter
X
X
Fighter
Fighter
Fighter
Fighter
Fighter
Fighter
Fighter
Fighter
Fighter
72
F-4B
73
F-4E
74
F-4J
75
F-5E
76
HC-130
Cargo
77
LC-130
Cargo
78
KC-10A
79
KC-130F
80
KC-130J
81
KC-130R
82
KC-130T
83
KC-135
84
MC130H
85
OV-1
86
P-2D
87
P-2F
88
P-2H
89
P-3A
90
P-3B
91
P-3C
X
X
Cargo
X
Cargo
Cargo
Cargo
X
X
Cargo
Patrol
Patrol
Patrol
X
X
Cargo
Cargo
Patrol
X
Patrol
Patrol
Patrol
87
Table A.1Continued
(Tables 2.1,
5 Lots)
2.2),
(Table 4.1),
Types
Types
92
P-5B
93
PC-6
CI and PR
acteristics
4.2, 4.3),
(Table 4.6),
24 Aircraft
93 Aircraft
Types
Types
Advanced
Materials
(Table 4.7),
49 Aircraft
Types
Mission
Type
Patrol
Patrol
94
RC-12
95
RA-5C
Electronic
96
S-3A
97
T-2C
98
T-33B
99
T-34C
Patrol
X
Patrol
Trainer
Trainer
Trainer
100 T-38A
Trainer
101 T-39A
Trainer
102 T-44A
Trainer
103 T-45A
104 T-46A
Trainer
Trainer
105 T-6A
(USAF)
Trainer
106 T-6A
(USN)
Trainer
Trainer
Trainer
107 TA-4F
108 TA-4J
109 TAV-8A
110 U-2
111 UC-12B
112 UC-35
113 UV-18
114 V-22
(USN)
115 VT-39E
116 WC-130
Trainer
X
Electronic
Cargo
Cargo
Cargo
Cargo
X
Cargo
Cargo
APPENDIX B
Survey of Industry
RAND was tasked by N81 and SAF/AQ to learn the reason for the
phenomenal cost increases of military xed-wing aircraft over the past
three decades. To that end, we analyzed the information available to
the Navy and Air Force cost analyses organizations.
During our discussions with the aircraft industry leaders, we
hoped to examine how unit procurement costs have changed over time
(i.e., why do aircraft today cost more than they did 30 years ago?).
We also wished to share our preliminary observations with the aircraft
industry leaders so that their insights and data could further inform
our analysis.
Below is a list of seed questions used to jumpstart these meetings.
The list was a starting point only. Other related topics were welcome.
1. In your view, what are the primary sources of military aircraft
cost escalation for the past several decades?
2. How has the complexity of military aircraft evolved? What metrics do you think best capture the evolution in the complexity
of xed-wing aircraft (i.e., material mix, complexity of avionics,
systems integration, low observability, and such)?
3. Are there any changes to contractual, regulatory, and statutory
requirements that you believe may have added to the acquisition
costs of military aircraft over the past six decades (i.e., contract
military specication requirements, quality assurance requirements, OSHA requirements, government oversight or environmental requirements)? If so, how can we quantitatively or qualitatively capture or reect their eects on aircraft costs?
89
90
4. Overhead costs have grown in the past few decades. Do you

have any data that illustrate the increase in overhead costs? Can
you identify the main contributors to the overhead cost growth
(e.g., retirement, health care, and other benets)?
5. Please tell us about the changes in the industrial base that may
have aected military aircraft costs (e.g., diminishing sources of
materials and equipment, lack of competition at the subvendor
level, or ow down of government requirements).
6. Have any government reporting requirements changed in the
past six decades (i.e., earned value, cost data reporting, or
CDRL requirements)? Do you believe that any of these requirements have added to military aircraft costs?
7. What initiatives can the government encourage to reduce the
cost of future military aircraft (e.g., multiyear acquisition,
lean production, contractual incentives for cost reduction, and
such)?
Bibliography
Arena, Mark V., Irv Blickstein, Obaid Younossi, and Cliord A. Grammich, Why
Has the Cost of Navy Ships Risen? A Macroscopic Examination of the Trends in U.S.
Naval Ship Costs Over the Past Several Decades, Santa Monica, Calif.: RAND
Corporation, MG-484-NAVY, 2006a. As of December 6, 2006:
http://www.rand.org/pubs/monographs/MG484/
Arena, Mark V., Robert S. Leonard, Sheila E. Murray, and Obaid Younossi,
Historical Cost Growth of Completed Weapon System Programs, Santa Monica,
Calif.: RAND Corporation, TR-343-AF, 2006b. As of May 30, 2007:
http://www.rand.org/pubs/technical_reports/TR343/
Augustine, Norman R., Augustines Laws, 6th ed., New York: Viking Penguin,
1986.
Beltramo, Michael N., Donald. L. Trapp, Bruce W. Kimoto, and Daniel P. Marsh,
Parametric Study of Transport Aircraft Systems Cost and Weight, Los Angeles, Calif.:
Science Applications, Inc., NASA CR151970, April 1977.
Birkler, John, Anthony G. Bower, Jerey A. Drezner, Gordon T. Lee, Mark A.
Lorell, and Giles K. Smith, Competition and Innovation in the U.S. Fixed-Wing
Military Aircraft Industry, Santa Monica, Calif.: RAND Corporation, MR-1656OSD, 2003. As of May 30, 2007:
http://www.rand.org/pubs/monograph_reports/MR1656/index.html
Blickstein, Irv, and Giles K. Smith, A Preliminary Analysis of Advance
Appropriations as a Budgeting Method for Navy Ship Procurements, Santa Monica,
Calif.: RAND Corporation, MR-1527-NAVY, 2002. As of December 6, 2006:
Clark, Vernon, Testimony Before the U.S. Senate Armed Services Committee,
Washington, D.C., February 17, 2005. As of May 30, 2007:
https://navcms.news.navy.mil/navydata/cno/testimony/clark050217.pdf
Congressional Budget Oce, Total Quantities and Costs of Major Weapon
Systems Procured: 19741993, March 24, 1992.
Cook, Cynthia R., and John C. Graser, Military Airframe Acquisition Costs: The
Eects of Lean Manufacturing, Santa Monica, Calif.: RAND Corporation, MR-
91
92
1325-AF, 2001. As of November 1, 2007:

Defense Acquisition University, Defense Acquisition Guidebook. As of
November 9, 2007:
https://akss.dau.mil/dag/welcome.asp
Deming, William Edwards, Some Theory of Sampling, New York, N.Y.: Dover
Publications, 1950.
Dryden, James, Therman P. Britt, and S. Binnings-DePriester, An Analysis
of Combat Aircraft Avionics Production Costs, Santa Monica, Calif.: RAND
Corporation, N-1685-AF, 1981. As of November 1, 2007:
http://www.rand.org/pubs/notes/N1685/
Eskew, Henry L., Aircraft Cost Growth and Development Program Length:
Some Augustinian Propositions Revisited, Acquisition Review Quarterly, Summer
2000, pp. 209220. As of May 24, 2007:
http://www.dau.mil/pubs/arq/2000arq/eskew.pdf
Goldberg, Matthew S., and Anduin E. Touw, Statistical Methods for Learning
Curves and Cost Analysis, Lithnicum, Md.: INFORMS, 2003.
Groemping, Ralph A., and J. W. Noah, Estimating Aircraft Airframe Costs by
Parametric Methods, Falls Church, Va.: J. Watson Noah Associates, Inc., TR10618-USN, 1977.
Heatherman, Paul E., Historical Aircraft Program Schedule Data, Directorate of
Cost Analysis, Comptroller, Aeronautical Systems Division, Wright-Patterson
AFB, Ohio, November 1983.
Hellerman, Christopher, Controlling Costs in Tactical Aircraft Programs,
Testimony Before the U.S. House of Representatives Government Reform
Committee Subcommittee on National Security, Emerging Threats, and
International Relations, Washington, D.C., April 11, 2003.
Janes All the Worlds Aircraft, Coulsdon, United Kingdom: Janes Information
Group, annual.
Kirkpatrick, David L. I., Trends in the Costs of Weapon Systems and the
Consequences, Defence and Peace Economics, Vol. 15, No. 3, June 2004, pp.
259273.
Large, Joseph P., Harry G. Campbell, and David Cates, Parametric Equations for
Estimating Aircraft Airframe Costs, Santa Monica, Calif.: RAND Corporation,
R-1693-1-PA&E, February 1976. As of May 24, 2007:
http://www.rand.org/pubs/reports/R1693-1/
Loftin, Laurence K., Jr., Quest for Performance: The Evolution of Modern Aircraft,
Washington, D.C.: National Aeronautics and Space Administration, 1985. As of
February 9, 2007:
http://www.hq.nasa.gov/oce/pao/History/SP-468/cover.htm
Bibliography
93
Lorell, Mark A., Julia F. Lowell, and Obaid Younossi, Evolutionary Acquisition:
Implementation Challenges for Defense Space Programs, Santa Monica, Calif.:
RAND Corporation, MG-431-AF, 2006. As of May 30, 2007:
NAVAIR, Standard Aircraft Characteristics, separate documents for AV-8B, C-2,
E-2C, EA-6B, S-3, T-2B, T-34B, T-39D, T-44A, and T-45 systems. As of July 17,
2007:
http://www.history.navy.mil/branches/org4-4.htm
Oce of Budget, Department of the Navy FY 2005 Budget Estimates, March
2004, p. 32. As of February 8, 2008:
http://www.nance.hq.navy.mil/fmb/05pres/databook/DATABOOK.pdf
Oce of the Under Secretary of Defense for Acquisition, Technology & Logistics
Industrial Policy, Annual Industrial Capabilities Report to Congress, Washington,
D.C., February 2007a.
Oce of the Under Secretary of Defense (Comptroller), National Defense Budget
Estimates for FY 2008, Washington, D.C., March 2007b. As of July 17, 2007:
http://www.defenselink.mil/comptroller/defbudget/fy2008/fy2008_greenbook.
pdf
Resetar, Susan A., J. Curt Rogers, and Ronald W. Hess, Advanced Airframe
Structural Materials: A Primer and Cost Estimating Methodology, Santa Monica,
Calif.: RAND Corporation, R-4016-AF, 1991. As of May 24, 2007:
http://www.rand.org/pubs/reports/R4016/
Schultze, Charles L., and Christopher Mackie, eds., At What Price? Conceptualizing
and Measuring Cost-of-Living and Price Indexes / Panel on Conceptual, Measurement,
and Other Statistical Issues in Developing Cost-of-Living Indexes, Washington, D.C.:
National Academy Press, 2002.
Smith, Giles K., A. A. Barbour, Thomas L. McNaugher, Michael D. Rich, and
William L. Stanley, The Use of Prototypes in Weapon System Development, Santa
Monica, Calif.: RAND Corporation, R-2345-AF, 1981. As of May 30, 2007:
http://www.rand.org/pubs/reports/R2345/
Statistics Canada, The Consumer Price Index Reference Paper, Catalogue
62-533-XPB, 1995. As of November 1, 2007:
http://www.statcan.ca/english/concepts/
Tran-Le, Jody, and Stephen Thompson, Chinas Impact on Metals Prices in Defense
Aerospace, Washington, D.C.: Oce of the Deputy Under Secretary of Defense
(Industrial Policy), December 2005 (published April 2006). As of November 1,
2007:
http://www.acq.osd.mil/ip
Younossi, Obaid, Michael Kennedy, and John C. Graser, Military Airframe Costs:
The Eects of Advanced Materials and Manufacturing Processes, Santa Monica,
94
Calif.: RAND Corporation, MR-1370-AF, 2001. As of February 11, 2007:

Younossi, Obaid, David E. Stem, Mark A. Lorell, and Frances M. Lussier, Lessons
Learned from the F/A-22 and F/A-18E/F Development Programs, Santa Monica,
Calif.: RAND Corporation, MG-276-AF, 2005. As of May 30, 2007:
http://www.rand.org/pubs/monographs/MG276/index.html
Younossi, Obaid, Mark V. Arena, Kevin Brancato, John C. Graser, Benjamin W.
Goldsmith, Mark A. Lorell, Fred Timson, and Jerry M. Sollinger, F-22A MultiYear Procurement Program: An Assessment of Cost Savings, Santa Monica, Calif.:
RAND Corporation, MG-664-OSD, 2007. As of July 6, 2007:
Zakheim, Dov, Remarks to Lexington Institute Capitol Hill Conference on
Fighters and the Future of Joint Warghting, Washington, D.C., January 28,
2005. As of December 6, 2006:
http://lexingtoninstitute.org/627.shtml

Rand Mg696

Uploaded by

Copyright:

Available Formats

Rand Mg696

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Rand Mg696

Uploaded by

Copyright:

Available Formats

NAT I O NA L DE F E N SE R E SE A R CH

This PDF document was made available

from www.rand.org as a public service of

the RAND Corporation.

The RAND Corporation is a nonprofit

For More Information

RAND Project AIR FORCE

View document details

This product is part of the RAND Corporation monograph series.

Prepared for the

NAT IONAL DE FENS E RES EA RC H I NS TI TUTE and

The research was conducted within the Acquisition and Technology

Library of Congress Cataloging-in-Publication Data

The RAND Corporation is a nonprofit research organization providing

Copyright 2008 RAND Corporation

All rights reserved. No part of this book may be reproduced in any

In recent decades, cost escalation for military xed-wing aircraft of all

Why Has the Cost of Fixed-Wing Aircraft Risen?

retary of Defense, the Joint Sta, the Unied Combatant Commands,

ing processes. The authors provide cost estimators with factors

Why Has the Cost of Fixed-Wing Aircraft Risen?

The Escalation of Aircraft Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Data and Price Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

Why Has the Cost of Fixed-Wing Aircraft Risen?

Material and Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Industry Views on Military Fixed-Wing Aircraft Cost Escalation . . . . . 57

A. Aircraft Included in the Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

S.1. Contributors to Price Escalation from the F-15A (1975)

Average Annual Cost Escalation for Aircraft and Ination

Why Has the Cost of Fixed-Wing Aircraft Risen?

Percentage Contributions to Annual Cost Escalation,

As with many other military weapon systems, military aircraft have

Why Has the Cost of Fixed-Wing Aircraft Risen?

Another explanation could be more ecient use of labor and tooling

Annual escalation rate, %

Why Has the Cost of Fixed-Wing Aircraft Risen?

The Services could choose to address cost escalation in several

Many individuals contributed to this study and we would like to thank

Arms Control Export Act

Why Has the Cost of Fixed-Wing Aircraft Risen?

The Escalation of Aircraft Costs

Commenting on the continually increasing costs of military aircraft,

Augustine based this prediction on costs for individual aircraft that

Why Has the Cost of Fixed-Wing Aircraft Risen?

These trends, as Augustine would note, have dire implications for

Procurement outlays (FY08 $B)

The Escalation of Aircraft Costs

through the early 1970s, increased through the mid-1980s, and

Why Has the Cost of Fixed-Wing Aircraft Risen?

Data and Price Trends

Data Sources and Normalization

Why Has the Cost of Fixed-Wing Aircraft Risen?

Sources of Data and Their Content

Data and Price Trends

To understand the causes of cost escalation, we needed a database with

Air Force factsheets include those available at http://www.af.mil/factsheets/ and http://

Why Has the Cost of Fixed-Wing Aircraft Risen?

Our work is multistage, with each stage considering a dierent number

Data and Price Trends

Measuring Cost Escalation

( Year2 Year1 ) Cost